Room 211 Friday Mar. 06 - SCaLE 23x

This talk critiques the current model of technology foundations, arguing for a collaborative approach that leverages shared resources and expertise across the community. The speaker suggests that instead of each foundation duplicating efforts in areas like governance, event management, and community development, there should be centers of excellence where foundations can specialize based on their strengths. By connecting with these specialized groups, organizations can reduce their workload and improve sustainability. The talk also emphasizes the importance of community support in open-source software development and discusses the need for higher education institutions to engage more with the open-source community. The session concludes with a call for signing an open letter to reclaim digital agency in academia and the launch of initiatives aimed at promoting collaboration in educational technology.

I want to take that idea that Terry talked about and sort of blow it up. He was saying foundation 2.0 is uh focused less on infrastructure and developer support and that's still needed but needs to expand to include broader community support. And why should all of the foundations be providing the same services? Is there a way where we can just like open source software projects we can

leverage the developments across the community and integrate those across uh foundations. So current model is you have all these foundations doing all the same thing or you have you know peer foundations and open source commercial partners uh contracted consultants universities they're all providing governance IP community development event management you know all the things that foundations should do what if instead oh here's all the things that foundations

do right so what are the costs and the demands on the the the individuals within these supporting organizations to provide all this stuff. That's a lot of work for foundations that are trying to maintain, you know, have issues around sustainability, right? All these foundations of projects are trying to remain uh sustainable and yet there's all of this and it's growing. So instead of everyone doing this, why

don't we have centers of excellence where individual foundations who have expertise in one area whether that's a technology or um uh some sort of outputs or you know resources that they've just cultivated for a while and it sort of took off and I'll give some examples in a minute. Why can't we all connect and we partner with those groups so that when I need intellectual property or

policy work I can reach out to one group when I need event management and um coordination or marketing and communications or grant writing or whatever it might be I can find individual foundations that have expertise in that. So what we end up is looking more like this. So we all these are all different types of organizations and they all have sort of vin diagram like areas of

overlap where yeah this is like I don't know let me I'm making this scale is the because I'm at scale scale is the best conference organizers in the world why do I need conference organizers why don't I call up the scale folks and say hey as a sustainability fundraising option the money I would have spent on hosting and putting together my own conference with folks who aren't

as talented skilled, experienced as your staff. I'll take that $30,000. I'll give it to you and you organize it. That becomes a sustainability model for uh foundations and organizations that have expertise. I'm relieved of having to duplicate that effort and it's just like calling a library or installing, you know, it's what's the difference or connecting through an API. What it's the same model. Why don't we do

this? Why don't we treat our foundations just like we treat our development? So we have all of these services that people expose that then other foundations can take advantage of and there might be multiple maybe Linux Foundation does a great job at um conferences too and if scale is busy we can reach out to the next one and so we get this model where everyone's sharing um

their expertise to for the entire open-source ecosystem. So at open congress last year in Brussels we had this discussion and two areas came to light. It's like let's have like a test and so public policy open source initiative in Eclipse have excellent public policy folks. So why do I need to go out and hire public policy people as the executive director of Apario when I can just

call them up and say hey I've got either a job or the community comes together and here's an issue that the entire open source community would benefit from. And an example is the response to the CRA in Europe that impacted everyone. We all got together, worked through public policy at um at Eclipse and OSI and that was a resource that I could have never developed and find

experts and respond in time. I could leverage what was already out there. The second is security audits with software bill of materials and things like that coming out. Um why do we all need to hire experts to do audits which might you can we even keep them busy the whole time? I mean our release cycles all different but why not use a shared resource that one organization

has expertise in they can just focus on being experts in that and go on. So for Apario it might be oh we with grant writing we have folks on universities that write grants all the time. So, if you're um I don't know what projects you're with, but if you're with project X, Y or Z and you want to apply to Sovereign Tech Fund or an NSF grant

or Sloan Foundation or Ford Foundation, whoever it might be, contact us. We'll help you. You spend less because you don't have to do the whole thing. We make a little money. It becomes a sustainability channel for the foundation. So, hopefully you now understand how open source and hired is aligned. You've seen some opportunities and now how can you engage? Here's just take a picture of this. I

can't go over all these, but like I said, these are all the things that different students, companies, and projects may want to how they might like to engage with Apario and higher education. Um, you don't have to work through Apario. I'd be happy to just hand you off to someone else. Apario is the only open-source uh software foundation dedicated to um edte. So, we do have lots

of connections, but if you want to work with the university specifically, I'd be happy to make connections there. But if you find something of interest, let me know. Uh, one last thing, uh, we just this Monday launched an open letter campaign uh to, to, uh, ask higher education to, uh, re-engage with the open source community and reclaim their digital agency. Um right now too many too much

of higher education is just seeding their technology direction to big companies. Um digital sovereignty is a big issue. Data protection personal privacy and so on especially with students. So uh we have about a hundred signatures there. I've grabbed some big names. Deb Nicholson from Python, Ed at Python, David Wy, uh the open content person, Jack Seuss, folks. And it's up to a hundred now, but um from

when I made the slide, I'd love for you to go to the open letter and sign it. Um you could do it as an individual, not necessarily on behalf of your organization. And finally, we're having an event tonight. Uh so if you if I haven't given you a coupon for a free drink and you'd like to have you or your and a friend or two uh come

join us, we'd love to have you. Any questions? All right. I'm out of time. >> Okay. >> Oh, yeah. Thank you. I don't know if it works, but they said to repeat the questions. [snorts] >> yeah, go ahead. >> Actually, this mic is working. more than a year when I was working on a NSF project uh post uh >> which community development and uh nurturing the innovations

and the two big uh challenges uh which is trust because protecting the innovation and the IP is is what hindering the research uh because they are doing novel things. So how do you address that? And the second thing is about opensource sustainability in term of uh uh maintainers and developers getting paid because almost majority of internet infrastructure is built on top of opensource libraries and softwares. But

what is your take and reflections about the uh getting uh funding the maintainers and uh developers? >> Well, I think I'm we're we're going to have to have a a a couple cocktails later tonight. So, the two questions were um uh protecting novel ideas, right? Uh through open development practices and then sustainability uh getting paid, right? So, unfortunately, I dismiss both those things. So, if you're not

interested in sharing your work because you think it's too novel, then don't make it open source. But I would argue that innovation and development and co-creation is better served through open and transparent collaboration. Um, so if it doesn't fit your your expectations, if you have a business model, I know if you're here for the slide where I said there is no business model and open source and

I said it was a byproduct and I had um the slide of academic use, research use, and administrative use. That's not your main job. your main job is is teaching getting research and so on. So the open source is a byproduct of a uh not your a product that you're going to go sell. So I think in higher ed it's a different model. So there definitely is

the tech transfer office and if you want to commercialize your software through uh proprietary licensing that's an option that you have. Um so that's one thing. Um, as far as um the other Oh, sustainability. Um, uh, what was the specific question? Oh, >> um, right. I had a had another example. But, um, I think that that that uh, you know, you can't have it both ways. you

can't why would I as a developer contributor want to give to your project intellectual property that you're then going to go sell and so on. So I think you can't have it both ways. You can't be the canvas of of higher education or the Illusian. I have a a just a slide of of the Illusian uh website GitHub repository in the read me. It says we love

collaboration. Send us all your fixes. Well, why if I was a campus CIO would I pay my staff to develop fixes and contribute back that I then have to relicens next year those fixes. So, I I think it's sort of the sustainability model isn't about funding. I don't think there I think the idea of of soft money won't last, right? And so, I think higher ed's in

a unique position is that that the work that's being done in open source is that byproduct. you're primarily there in a research endeavor and you're not there to produce software. If that's what you're if you're going to grad school to start a business, then do that. If you're going to grad school to work in an environment that's inherently about peer review and collaborative co-creation around research, then

do that. I you can't do both in my opinion. Maybe somebody else has a different thought. >> Yeah. Uh I I would just like to point out that there are million different business models for open-source and free software. Okay? And so what you really need to do is as you determine your project, you have to think about what is the business model that you want to have.

And there really is no reason why you can't have a business model that allows open-source and free software to happen at the same time that you have a way of making money off of it. Okay. In fact, the the history is is wrote with closed source projects that failed because they never reached the added the the volume necessary to sustain them. And that's where open-source really has

an advantage. Okay? So, you really just have to think about your business model. There's a million articles out there that you can look at or read or books about it. You need to think about that. But you can make money with software. There's a company called Red Hat that proved that. >> So, I completely agree that you can make money with open source software, but you don't

make that money by selling a software license. You make that money through services that support the the software. So, Red Hat doesn't I mean, they're selling you support services. They're allowing you to access tools that are built specifically by them to support the Red Hat Enterprise Linux. So I agree that you can make money off of there's a great research by uh I forget his name out

of uh Molson school of business in uh at McGill or Concordia I forget in in uh Montreal and at when I was at OSI we worked with them to identify open source business models or business models that work with open source. So I completely agree that there's tons of companies that are making money and through the use of open source software but it's not by licensing open

source software. So if your business model is to create a really bitching re you know software that came out of your research selling that it's not open source. You might sell services training uh data collection and analysis using that all sorts of things but you're not going to sell the software. Was there another >> I'd like to I'm not sure if Okay, there we go. I'd like

to kind of transition from the commercial side and present the flip side. Uh I work for a company that has an open source history. My project is uniquely closed source because I I work for Canonicle. Uh I'm the lead of the Canonical Academy project. So we're making exams. I can't open source my exams because then people can't, you know, anyway. So what I'm trying to do is

create an open-source component for the community alongside it. And what I've been looking into is actually open sourcing educational units, you know, RLOS's, reusable learning objects. Why, and this is something I found at a lot of institutions. Why aren't we discussing open-sourcing education itself getting rid of LMS's and leaning into XAPPI, leaning into Tim can things where we don't actually need an LMS. We just you have

a unit, you need a a way to present it. Well, we know how to make web pages. You need a place to store it. We know how databases work. And you need a way to keep a grade book. We know how authentication and logging works. Why are we still talking about LMS's in this day and age? Why are we still talking about SCORM when we decided in

2015 that SCORM was antiquated? Is is there something I am missing being outside of the higher ed conversation? >> No, I think the reason that we still hear about LMS's is because there's billion-dollar companies selling LMS's to universities. But if if for instance a company that has made money in in open source were able to create we have an open documentation academy for contributors doing docs. I'd

like an open education academy for contributors who create educational models. If there were a company willing to support that is that feasible or is is this just a dumb idea that [laughter] everybody else has >> I have thoughts but somebody I think somebody wants to respond in the audience so we'll I was just going to start with a response for the middle ground that he was talking

about. There are I don't know which hexagon Patrick it is, but there are organizations to make money. It depends on how you define making money. If you're a faculty member and want to go through the the retention, tenure and promotion process, then if the institution supports a contribution software, you know, in addition, >> I would love that. >> Well, we have that with like R open side.

Now we have PI open side. I don't know if there's a Julio, but so there are mezo hexagons in them. to answer your question, I'm gonna be honest with you. Um, we have hollowed out in especially in California where I work in higher ed, we have hollowed out tech development. If you look at org charts in a university, most of those people don't know anything about any

part of technology. Even in IT departments, some parts of network management is outsourced in some cases. >> There they're just there's just enough. This isn't an excuse. It's a leadership choice. And the choice has not been we talk because we need them but we don't have people or desire to program in structure is all ruby on rails with increasingly amount of typescript on the on the client

side. So the whole thing is open source and we pay for it so it can run 247 365 on Amazon web services. You know we created 90% of all those technologies in higher ed as Patrick pointed out but we don't run them because our choices have been our choices for funding for leadership even in progressive California have been different. Okay. And uh you know the main focus

on my campus is make sure every web page is accessible by April 24th 2026 because that's when the title two um requirements drop. I mean there's a lot of other things to do on a campus. So this is a longer talk which is why the cocktails might be helpful. But >> and if I could >> the institutions have to decide that they want to do all of

this. If it were me, it would be full stack open source across all disciplines. >> So sign our letter. >> I already did. >> Um and I just add so that's >> sorry I don't have a very good answer >> but it's not it's not a new problem. So if you think of you know Thomas Freriedman right we're only going to have what 10 universities in the

world and everything's going to be a MOO right remember that we're going to have all these Corsera and edex all these things are going to be these uh amalgams of the best teachers and professors in the world we're going to create the best content we're going to aggregate them on demand and students were going to to be self-motivated self-directional learners who just find this stuff and do

their own thing and credentiing and badges were going to be the way we're going to go remember all that my and now how many here if you walked into your provos office and said hey I want to start a MOO you'd like get kicked out my concern is is that the OSBO is going to be the MOO of 2027 or 2026 where where happens if we don't

get university buy in if we don't have full stack if we don't have folks that are engaging in open source authentically that it's just seen as the next round of grant funding coming from Sloan and other places um what's going to happen when we say we want to do an OSO So I think there are remember learning object repositories like Merllo that still exists. Um but there's

just not the commercial incentives. Um Elsir tried to do open class and they were going to push in all their content from their books. So I think that's always been an idea but it's never had the >> remember there's no such thing as a university. They're all multi-universities and so most of what Patrick is talking about is the learning side of the university. But I can guarantee

you if you're working in finite element analysis, computational fluid dynamics, molecular dynamics, speaking with my NSF shirt on here for a moment. Everything's open source. The only thing that isn't open source is the stuff we're working on to make it open source like CUDA and using open CL rather than using Nvidia's drivers. But on things like the national research platform, we have every single thing is open

source through the entire stack. >> So on the research side, it is pretty much done >> and the pose grants highlight that this is an expectation that this will be open source. Um David Wy who's on the list here um started Lumen Learning. I don't know if you know David um but Lumen Learning Lumen Learning is a um open content uh development um group that is trying

to create individual teaching and learning activities um and expose them outside of the LMS. To your point I agree completely why when I was at Sunni we had something called the learning management operating system. We tried to sell the not sell but we were in the system and wanted to create where we weren't dependent on one LMS and it's just campus CIOS weren't interested you know contracts

in order that's what CIO means now contracts in order that's all and as a former CIO and CTO I'm allowed to pick on community any other question >> um yeah so I'm also a former um high my um school board member for my school district and um K12 is saddled with a same as a lot of public institutions in that their IT procurement is mandated through um

procurement offices where um there's perceived benefits through group purchases through the state. So in New York where I live, the first thing you do in order to avoid an RFP, an RFI, putting together a search committee or a tech committee on this thing is you just I'm just going to buy it off state contract. I don't want to have to deal with all this. >> And in

fact, you have to provide a justification for not acquiring something through state contract because if the state contract is there and everyone else's every other school district's licency fee drops because another 50,000 students from this district joined, that's a benefit that they see that oh, they're they're facilitating. When I was at UMass, I was a CTO for UMass for UMass online was coming up with Blackboard, which

just bought WebCT. But UMass thought it was bulletproof because they bought a perpetual license for WebCT. But Blackboard said, "Well, great. You can have a perpetual license for this thing. We're never going to develop past 2010, so go for it." So, obviously, that wasn't a choice. So, we had to go through a whole RFP process. And Moodle was one of the candidates. But the problem was that

if UMass online went to Moodle, it wouldn't need Oracle anymore. Oracle would then not be as pervasive across the system, which would have in the total seats would have dropped, which means that the price per seat would have gone up for the other campuses. So the other campuses in the 17 campus system that was using, you know, resources all would see us increase in their Oracle licensing

because the online campus was no longer participating. It's things like that. So K12 suffers because of the benefits of consortium buying at the state level. That's just one issue. I could go on over drinks. >> Okay. And uh sorry, I can't attend drinks because I'm speaking at school. So next time uh my other question is so commercial education for lack of a better way of explaining what

I do. Uh we're feeling a lot of tension in terms of machine learning to create customized experiences which is another reason I want to get away from LMS and and Lexi and make it just a data that is presented. uh are those sorts of tensions also being experienced in higher ed or is that something out in in the future? I is there the idea that you can

just have a basic web app that is fed educationally by these other things? I know it e is working a lot on that in the federal space and I'm just not sure if that's in higher ed yet. >> I don't know if there's specific platforms or techn How are we doing on time? Are we >> 10 more minutes? Um, >> sorry. >> No, no. Um, so if

you would have gone to educ cause conference, I don't if you've been to anyone been to the educ cause conference. Um, if you walk into the Shark Tank, I mean the exhibit hall with all the vendors in there in 2012, you would have seen open source, open education, OEER everywhere. It was like the last year when I went, it's all AI, right? AI. I felt bad. I

walked by one of the booths of the vendors that sell furniture, classroom furniture, and I'm like, how are you competing? you don't have any AI furniture. Nobody's stopping by your booth. You better put something up. So, there's this. So, AI, whether or not they're going to do anything with it, it's pervasive in the marketing. I think that the the LMS vendors, the any of the enterprise application

vendors are selling something about M, you know, AI and and all the rest of it. So, I don't know if there's anything behind it other than making sure that they don't look like they're behind the, you know, times. Um, I just there's more and more walled garden. They, you know, if you look at the business model of LMS's, they're expanding. So, they're becoming ERP systems. They're becoming,

you know, recruitment tools. are becoming you know they're just their scope is growing because they just want to capture more of that market become more uh ingrained where their universities are dependent upon these things they want to you know that's why they're offering their cloud hosting and their data is all being you know so it's I don't I don't know how authentic it is I'm sure they're

saying that they want to do it um but I don't know I don't you know most of them are going bankrupt I mean like anth technology went bankrupt and now Blackboard's its own company. Um, where's Instructure going to go? They've got tons of campuses already. How can they increase their revenue? Al Essa on this article that I referenced here has a great article on um enterprise systems.

Uh, is the LMS now a commodity? So, if you want to look that up, he's got a great article on discussing sort of that in the age of AI. I think we have time for one more question. >> Well, if anyone would like to thank you all for coming and if anyone would like to continue the lamenting the state of edtech uh and you need to drown

your sorrows, I have a ticket uh for a coupon at uh for a free drink at our event tonight if you'd like to join us. But thank you all. Let's get another round our next talk is going to be around 11:15 uh by Mad Dog Hall uh open source and community higher education. So come around for that one too, please. Oh, and the slides for Patrick's presentation

will also be available on the website later. Still talking. Still talking. Still talking. Oh, there we go. That's good. >> I'm It's It's Well, it's it's it's coming there. >> I'm at the right res. All right, no problem. We still have a few minutes left. So, I'm going to start at 11:15 so the recording can can do that. >> While we're waiting, I'm going to point out

this gentleman, Patrick Mason, who's handing out free drink tickets for the uh thing on the patio tonight. You get a drink. And so if you're interested in going having a beer and talk about education and open source software, he'll be in the back giving those out. I'll make another announcement later on. We have to help each other out here. When they make these things for your head,

I wish they would consider the fact that some people have larger heads than other people, you know? I mean, it's it's just barely hanging on to my right ear. So, >> no, that's I I purposely picked this so I don't have to worry about holding a mic or anything. Let's see. We still have five minutes >> Uh, no, not today. >> Maybe How many of you is

this your first time to scale? Welcome. How many of you have been to scale say more than five times? Good. Anybody here who has been programming earlier than 1969? you you I'm sorry, sir. You in the back or raise your hand. You don't look old enough to have been programming before 1969. I'm sorry. You'll have to show me your bonafide. Okay, I'm going to start to stand

up for me. That takes some time come on. Where are you? It's very strange. There we go. Hello everybody. My name is John Mad Dog Hall and I'm the board chairmeritus of the Linux Professional Institute. That and $350 will get you a cup of coffee at a Starbucks almost any place. For a lot of you who know me, I've been in free source and uh free software

and open source for a long time. But before that, I was programming back in 1969. And that's what we just called it, And I've had a wide variety of different jobs in industry and higher ed and volunteer stuff along the way. But the most important part I think is that I did teach at the university level and I enjoyed doing it. Now a lot of people know

me from when I started with the Linux project of Alpha Linux and maybe they remember me from before that when I was working in 1986 with Richard Stallman in the free software foundation. However, I my really started with programming and and electronics a lot earlier than that and I had three years of electronics class in high school analog electronics, coils and capacitors and things like that. And

then I went to Drexel Institute of Technology which changed over to Drexel University. It was interesting because it was a private school but it had cooperative education which I believe in a lot. And halfway through my curriculum of electrical engineering, I got involved with computers and I found I was much better at computers than I was at electrical engineering. And so I switched over to a course

that was half business and half engineering. And this made a big difference in my entire career. We just did some computer things back then because we didn't really have computer science. It was more like computer black magic. We were stumbling along in the dark and there were some people in academia and research that were doing great things, okay? But we were just, you know, trying to get

along back in those days. And I got a lot of my education simply by reading books and studying and practicing on little computers like the deck PDP8 and learning PDP8 assembly language and then later on PDP11 assembly language and IBM assembly language just reading books and practicing. I started the Drexel computer club because nobody else had even thought of doing that. And then I discovered interacted with

Dikas, the digital equipment corporation user society. Now back in those days we had a lot of these. There was share for IBM and brainstorm for novel and things like that. And these were places where people would write software for their own use because they needed it. They were not professional programmers. They were physicists and chemists and business people and academics and electrical engineers who wrote software for

their own use. And then when they had the software, they said, "Well, what am I going to do with this?" Because selling software is hard work and that that was not their business. And so they would contribute it to dikas and then dikas would make a catalog and publish it and send it out. You would pay $15 for the catalog. And that was a lot of money

for a college student back in those days because for $5 you could buy a picture of beer. And so you had a choice between getting this catalog of free software or fur beer. And I think you can see what direction I would go in. But this was what DAS did. And when you got that software, you could then make copies of it on paper tape from you

buy it from the school store, distributed to your friends for a dollar a piece. And after a while, you would make back your money. And so I became the first redhead. So we have to ask ourselves, what is a university education? What do universities actually do? Okay. And the first thing that they do is they develop a curriculum and that curriculum is looked at by various you

know certification organizations and industry and everything else and they make comments on and everything but you have the curriculum that you should be doing and then they create coursework for that curriculum to be able to teach those courses to the students who are coming in. And then finally they certify those students as having learned something you know through tests through projects or things like that and eventually

you end up with a diploma. But this is not what the university is actually doing because when you really look at it what they're really doing is they're teaching you how to learn on your own. You know when you're in high school that you know if it doesn't come out of the teacher's mouth it's not going to be on the test. I knew students in high school

that didn't even never open the book. They just listened to the teacher in class and they would get good grades on their courses. And that's why when they hit university, all of a sudden they were lost because the professor would come in the room and say, "Here's the syllabus. There's the textbook. There's the library. There's the people to talk to. You go got do interviews, stuff like

that. I'm just your guide. Goodbye." and they would go and they would maybe do research and stuff like that. But that is where that's why you have such a high failure rate in your freshman classes. Now let's take a look at each one of these things in detail. You have the developing a curriculum. So the the the professors and the university talks with industry leaders. They say

what do you need in the future? You see as far out as you can go. Oh by the way we're looking at all this research and this is stuff we need. You know AI is the neatest newest thing of course and you also have these organizations like the association of computing machinery and the triple E who work along with this to do some of their certifications of

this curriculum to say these are things that we see with all of our members and everything in the way and there are two major or it's a major project of these two groups they're coming out with a series of curriculum that they publish in PDFs, free to pull down off the net. And these curriculum will guide you as to what the prerequisites for each course is, what

each course is supposed to teach, where it fits inside of the curriculum, and things like that. But what they don't do is they don't teach they don't say to you, you have to use this operating system, you have to use this office package, you have to use this compiler, you have to use this. They don't they don't say that to you. That is left blank. And so

it's for you to fill in. You the professor. Then you have to create for each one of those courses the coursework for that curriculum. Which textbook should you use? Now, how many of you have ever used a dragon book to learn compiler theory? Oh, all the rest of you, you're lucky because the dragon book actually started out as two books. One was written by Al Aaho of

Bell Labs and the other was written by Jeffrey Alman of Stanford University. And those two books together were the first two books I ever tried to learn compiler theory from. They cost $300 for the set back in 1969. Okay. Well, Jeffrey and I and and and Al knew that that was too expensive. So, they combined the two of them into the first Dragon book and it was

incomprehensible. They even admitted later on it was incomprehensible and that's why everybody was lucky enough to use maybe the second version of the Dragon Book. Okay, which they did a much better job of smoothing out the edges. However, Lewis Rosen Cranon Sterns in digital compiler design was actually what the first book I taught out of and it was immensely easier to learn from than the Dragon book.

So, you pick out a textbook for the course and you look for magazine articles and now articles often that stuff to give to your students and you think about projects that they should These are the teaching materials that you then put into a little file because when the certification people come along, they're going to want to look at all the tests that your students have taken, the

grades they got for those tests, and things like that to make sure that you're doing a good job at teaching these courses. And you also have sample tests that you stick in your little folder because the next professor that comes along will really appreciate having all that stuff because this is the first time they've ever taught that course. Maybe they're a new professor. They have to teach

three courses at the same time and to come up with all this stuff for three courses is a lot of work. If they have that little folder, they appreciate it. Then you certify that the students have learned through class participation, through projects, through tests. You not only test this what the students are learning, but you test how the teachers are teaching. Because if you start looking at

the bell curve for the answers for your tests and you see that most of the students are flunking and otherwise these are otherwise good students and most of them are not doing well on the test. Well, is it because you're not creating a good test? That could be a possibility. Maybe you're not lecturing to get across the information to them. That's another possibility. Okay. So, the tests

are actually for the professor as much as they're for the students. Am I getting this across to the students? Are they learning what they're supposed to learn? Now, certification agencies like Linux uh periodically review their certifications to make sure they're keeping up to date with what industry is using, what people feel that they should be learning and things like that. and they get sample works from students

to be able to uh see what they're doing and how their tests are going. And this is the problem. This is the difference between grade schools out of the teachers mouth if that's not on the test and universities. I don't really have the time to teach you everything. And this is because universities are not to prepare you for a job. I am tired of universities coming to

me and saying they have to use Microsoft Office because Microsoft Office is used by all of industry. I say, "I'm sorry. Your university, you're teaching civil engineers to build bridges and and buildings and stuff like that. You're teaching aeronautical engineers how to build airplanes so they don't fall from the sky. You're teaching brain surgeons how to do brain surgery. And yet you say that your students can't

learn how to use Microsoft Office in a single day. I don't mean I don't mean expert at Microsoft Office, okay? I don't mean that that they know 99.99% of all the features of Micros I mean just to be able to edit a simple document. You can't tell me that your students are too stupid not to be able to pick up Microsoft Office and learn what they need

to learn in a single day. You can't have it both ways. They're either smart or they're stupid. And if they're stupid, get rid of them. So this is what the Bachelor of Science is, okay? They teach you how to learn. Then your master's degree allows you to go deeper into a topic. I mean, you could have a undergraduate degree in civil engineering and then you could get

a master's degree in electrical engineering. You don't have to take all the courses and stuff like that. You could just take your master's degrees, maybe one or two extra courses to get to bring you up to speed and boom, you have a master of science in electrical To go into that depth and at a higher level, you don't have to talk to the students at the same

low level as you did because they're not coming out of high school anymore. They've already learned how to learn. That's what your master's degree is for. And it's preparing you to go for your PhD. is preparing some of some master's degrees require a thesis, some don't. Some just say classwork, some require a thesis to allow you to investigate how to do the research that goes into a

PhD. And then of course your PhD is a culmination of that. It's how we move it's how we move science and technology forward. Okay? you become the expert in that subject that people around the world will read your PhD thesis and say okay now I understand what's going on or you prove that something doesn't exist that's almost just as good because then nobody ever has to investigate

that again okay you become the expert in it my undergraduate roommate was graduated tops his class of electrical engineering Drexel and he wrote some of the first or he created or tried to create some of the first digital filtering circuits. Not analog filtering circuits using capacitors and coils and things like that, but using a digital circuit to be able to do digital filtering. And he went for

his master's degree out of Boulder University of Colorado Boulder. He got halfway through his master's degree when his PhD when his his master's thesis masters looked at this and said there's another student at another university is doing exactly the same research and they're coming up with exactly the opposite answer. And so they brought these two students together and had them argue in front of the professors. And

fortunately for my college roommate, he won. The other student had to start all over again. then you go out after you graduate, then you go out to work with a company. And you all know this. For the first six months, if you're working for a good company, for the first six months, they say, "Take your time. Get used to what we're doing. Learn how we do things.

This is the time you learn to use Microsoft Office." Okay? You got six months to do it. They won't fire you for the first six months. I mean, you can sit on the ground in the middle of the ground and twiddle your thumbs for six months and they still won't fire you because at the end of that six months, they rate what you've done and either do

one of two things. Either give you a nice raise or they say, "Well, you're going to stay where you are or they say goodbye." One of those things. But that's what happens. And I I recently had a friend of mine who came out. thought it was his second job, not his first, but it was something that just threw him for a loop. He'd never done this type

of work before. And he kept saying to me, "They're going to fire me. They're going to fire me. They're going to fire me." I says, "No, they're not. They're not going to fire you." And they gave him the same talk I just gave you. And sure enough, almost to the day after six months, they called him in. HR wants to see you. Called him in. You're getting

a promotion. And he looked at me, he says, "How did you know that?" And I said 50 years in the industry helped me know that. Now I'm going to put a very very strong recommendation for cooperative If your university isn't doing that, you really should think about it. I spent five years at Drexel, but when I graduated, I had one and a half years of practical experience.

And it was during this cooperative education job that I took a correspondence course of how to program the IBM 1130 computer in Forran. Now this was Forran spelled with all capital letters. It wasn't Forran 2, 4R 4, Forran 90, 4R7, high performance for thread threaded forran, none of that stuff. It was just forran. Arithmetic if statements do it to me. And I learned a lot from that

from punching cards and practicing and and doing it. And it is what dropped me or brought me from electrical engineering which I'd liked into computer science. So what should we really be teaching our students? What should we really be learning? Principles, not products. not teaching Adobe Photoshop, but teaching the principles of digital manipulation of of photographs or raster scan things, you know. So, you could be doing

it with or you could be doing it with Photoshop. You could be doing with Adobe Illustrator or you could be doing it with Blender. But the thing is you should also be you know that's the way you learn for a lifetime because no matter what product comes along you can look at the product and say I know what it's doing. I can see what it's doing. This

is all I have to learn is the proper key clicks and stuff like that. Now I've had lots of arguments with discussions maybe with professors about the need to teach assembly language. They say, "Oh, we don't need to teach assembly language anymore because compilers are so good." And you have this abstraction level of virtual machines and emulators and stuff. So, teaching the type of of of performance

programming you're talking about is no longer necessary. And I'm sorry. I'm I'm I'm sorry. You're wrong. Okay? Because if you don't know how the computer works, you're going to be making mistakes in your programming. And if you can't write efficient code and efficient code means understanding what the process is going to be doing from one particular way of holding an alpha numeric character and converting that into

a number and how it works with cache and all that type of stuff then your program is not really going to be optimized correctly. I'm not saying this is the first thing you should teach the students. You don't teach them assembly language as their first language. For God's sakes, teach them something like Python or something. I actually advocate shell script programming for Linux and Unix people. Teach

them that because the shell is already always there and they can do amazing things with shell programming. And then after that you could do Python, but then after that it should be assembly language. Why? Because after you've had assembly language, then dreferencing a pointer makes a lot of sense. After you've had assembly language, then re-entering recursive programming becomes a lot easier to understand. And so, and teaching

compiler theory or teaching operating systems design or teaching database design becomes much more trivial when you learn how the assembly how the architecture of the machine works. It doesn't have to be a complex assembly language. It doesn't have to be like Intel assembly language with all of its addressing modes and stuff. In fact, these days I would recommend teaching risk five as an assembly language. Small number

of instructions, small number of data types, but you learn all the things you really need to learn about how the machine works. And then finally, programming is just not enough. So all the people that go off and study how to write programming, you know, how to write programming, it's not enough. You need to know the rest of the stuff or at least enough that you can now

talk to the people who need to have their programs done so that you don't have to talk on the baby talk level to them and they don't have to talk to the baby talk level to you. And this is why we've recognized that most people going through university need some type of programming some type to be able to talk with Now we get to closed source versus

open And this is where I point out to people that closed source teaches you one thing. How to use that program to solve that problem. Okay. I have a I have an image I want to manipulate. Okay, I can learn I can use Adobe Photoshop to manipulate that image to there. Problem solved. But they don't tell you, it doesn't show you how Photoshop does It does allow

you to improve Photoshop to meet your needs. This is stuff that opensource or free software does teach you. So free software teaches you three times. Closed source only teaches you once. You learn a Microsoft operating system. You can see how to use it. Fine, but you don't know how it works. I mean, you can see block diagrams and stuff like that, but how does it really work?

And how can you change it if you really want to? And this is why I've been telling academics for close to 35 years, you should be using free software. And I differentiate between open source I'll tell you why I do that. Because with closed source, you have some other problems. Maybe your student gets a free copy of closed source software to go on their laptop and they

get to use it for academic purposes completely free. But now they have a cooperative education job and they want to use the same software at their cooperative education company and the license does not extend that far. The license changes and so now they have to buy a license from that company in order to be legal. And do you spend all of your time counting these licenses or

do you spend your time learning or teaching? And it's not just Linux. So when I taught operating systems, we didn't have all these operating systems I could point to, but if I was teaching operating systems today, I probably would start off with three DOSs. Very, very simple, allows people to to grasp in their mind the whole system. And then I might go to one of the BSDs,

FreeBSD, OpenBSD, NetBSD, or Tiny OS or any one of the other open free software. So I can compare all these different operating systems with each other. But it's not just about the operating systems, it's about databases. you know, MySQL or now D Mariah DB, Postgress. Postgress is an extremely powerful free software database system, video tools, audio tools. There's over 400,000 open-source or free software projects out on

the net that your students can look at, see how they work, learn how to change them, learn how to interact with When I started giving these talks, I used to talk about the, you know, 300,000 programmers that are now represented on places like GitHub and GitLab. Now there's 2.6 million programmers. Okay? But it's also not just about software. It's about open hardware. The Raspberry Pi people, the

professors at Cambridge when they started the Raspberry Pi project, they started it because they recognized that the students coming into their classes knew less about how computers worked than the students of 20 years before. Because 20 years before they had things like the Vic 20 or the Amigga stuff and they would open it up and they would look at it, they would bring, you know, basic source

code down. They get syntax errors. What the heck is that? They get runtime errors. Why is it doing that? They'd have to fix this stuff and they learned from that. And the students coming into their classes now would buy a laptop that has a sticker on it says, "You open this up, you void your warranty." And they and they and they and they download programs and they

run them. They don't get syntax errors, you know. they they and they don't know how to fix the problems that they do have. And so the professors at Cambridge said, "We need a little computer system that is purposely made without a case. It is purposely made to run free software because we want our students to be able to to know as much as they did 20 years

ago when they come into our courses. So we can start from there. And so we need to have open hardware. And this means even the circuit designs of these systems have to be open. The bill of materials for these systems have to be there. So you can look and say, "Oh, they're using this chip. I can get the I can get the circuit diagram and the circuit

information of that chip from the manufacturer. I can see how that chip is actually programmed and the PR circuit board layout. So maybe you can improve that. I'm working with a project down in Brazil called Caninas Lucas, this crazy canines in Portuguese. And they took a system that was developed by a company and they made it run 30% faster. It used less power. It created less heat

than a commercial company that was selling them by the thousands had done. FPGAAS. I went to the same professor, the same uh department head there and I complained about issues with ATI and Nvidia and stuff and I said, "Look, you know, these people create GPUs and you put them into an elevator, you put them into an airplane or something like that. it has to last for 30

years or 40 years. And after a while, they say, "Well, we've lost interest in that GPU, but we're not going to tell you how to program it." I said, "This is ridiculous." Because for low-level types of of 3D and color mapping and and and anti-aliasing and stuff like that, you could do all of that with an FPGA. The professor looked at me kind of strange and said,

"You're right." And three months later, he had two graduate students create a virtual GPU that could do that could pass all the for anti-aliasing, color mapping, and all the simple things. It wasn't the fastest thing in the world, but your elevator doesn't need the fastest thing in the world. Whisk 5 is the newest implementation of this. If you try and develop something using a ARM processor, the

first thing you have to do is talk to a lawyer and then after that you have to scrape up the money to do the licensing for the design. Risk 5, you just go ahead and start doing your design. There's emulators that will emulate the risk 5 instruction set. So even if you don't have a risk 5 CPU, you can write your operating system to it or your

compiler to it and see the results. But we also need open culture things like creative commons to be able to allow people to make money from their culture but also to use their culture in free ways. And all in the bottom of each one of my slides, I have a creative comments license that says you can use this for non-commercial purposes. You as long as you don't

modify it. If you want to do either one of those things, use it for commercial purposes or modify it, come back to me and I'll give you a different license. In our previous talk here, we talked about making money out of free software. And what I didn't mention is if you look at the MySQL model, you had exactly the same software that was licensed two different ways.

If you were able to obey the GPL and change it and make your make your changes known to other people and to end users, you could do that and pay nothing for it. But if you wanted to be able to change it and not publish your changes, not give it to end users, you had to go back to MySQL and get a second license and pay for

it. And that's how you can make money off of free software. Okay? Dual licensing. Then there's learning on your own. And when you find that curriculum and you go through all those those classes and you look and you find all that information, you can learn on your own. And because they're using free and open source software and open hardware, the cost of that is very very cheap.

But you also have a community of people that you can answer ask questions of and you have to use a little bit of common sense as to whether this answer sounds correct or will you try it out? Does it work? And stuff like that. But all this is learning. You're learning the information. But now you want to sell yourself to somebody else. You need to have a

way of showing that you know this. And that's where certifications come in. That's where your diploma comes in. Okay? That somebody has looked at what you've done and said, "Yeah, you know this." But certifications are not enough. I mean, you can get lots of certifications these days on various languages and particular distributions like Red Hat certified engineer and all that type of stuff. Or there's a more

generalized certification that things like LPI does. We're not vendor specific. We're not distribution specific. Okay. But then you have certifications on products. And I'm sure we all remember Nortell. Nortell went out of business so fast. It was amazing. It was breathtaking. And all the Nortell hardware, which of course still kept working, but all the people that spent huge amount of times learning it and getting their were

out of luck. So what are the objectives of a certification effort? to come up with good certifications to get feedback from people actually in the field as to what type of certifications are needed at what level and what questions we should be asking and what you know and what they should be studying to get this certification and a good example of that is the LPIC 303 that's

our third level of a certification in professional certification at the Linux professional institute where we publish all the objectives that you should be learning. And if you go down that list of objectives and you say, "Hey, I know that. I know that. I know that. Wait a minute. I don't even know what that means." Well, that last one is probably you probably should do a little bit

of studying on. But if you go down all the objectives, you could look yourself in the face and say, "I know what this means. I can do that. I can apply that." Then you can take the test and you probably will pass. But certifications are only a part of that. And what I recommend to students is they get involved with real live projects. Like I said, there's

over 400,000 projects out there. There's something for everybody, you know, and so you can get a project that will allow you to show your expertise in programming and fixing bugs And at the end of the time you can submit that and say, "Hey, Mr. Mr. or Mrs. employer, you can see my code. You can look at the mailing list. You can see the discussions we've had. You

can, you know, and this is the same as working for a company and doing the same thing. In fact, it's better because try and get that type of information of a company you're working for, right? You're lucky to be able to get a letter of recommendation, much less have them be able to look at the code that you wrote and the suggestions that you made. You can

also get recommendations from the supervisors and peers. I I suggest to people do projects for nonprofits. Does your does your church need a website? Do they need systems administration? Do they need a firewall put up? They need a database. Work with your church and get a message from them or a nonprofit, whoever they are. You don't get doesn't care. And then you write blogs and articles and

you get known in the industry for what you're saying. Sure, you'll get a lot of people who come back and, you know, argue back and forth with you, but after a while, you'll start to create a name for yourself and you may even end up talking at scale. I knew Mark Shudderworth when he started Abuntu. Okay. And what he did was he took the entire Debian mailing

list down to Antarctica on a laptop for nine months. He sat in Antarctica for nine months. And I guarantee you he wasn't look just looking at penguins. And when he came back, he said, "I want you and you and you to come work for me on Abuntu." He didn't need to call HR. He could see the people that he wanted to have come work with him. More

ways to get started. I already mentioned Gratus work for nonprofits. Learn the subject and teach and mentor it to someone. Well, I took compiler theory twice. I already mentioned the books I learned from. Okay? But I took compiler theory twice. Almost flunked it both times, but I made it through. But then I had to teach it. And there's one thing about teaching it. You can't sit there

and say the professor won't ask me the question on the exam. You can't say I can get that question wrong. I'll still pass the You have to know it cold. And I would spend hours and hours with books laid out in front of me as I went through each chapter in in doing my preparation until I knew it cold. I've taught compiler theory 10 times in my

life. I haven't taught it for 20 years, but I can walk up to Blackboard today and tell you how a compiler works. So teaching makes you learn it. and start this cooperation with your friends. Maybe you have you can take some free software from different places and make a product out of it. Have a business plan for it. Know how you're going to make money and don't

be afraid to fail. Someday you may be an old fart. I've asked uh I've asked Gail to create an old farts buff. I think it's going to be tomorrow night. Uh hopefully it'll be after my next talk which will be tomorrow. And I hope they let me have as an old fart, you'll find out that your degree that you got back in 1973 is really not as

relevant today as it should have been because computer security was locking the door at the computer room at night. Graphics was putting asy art on online printers and networking was carrying a box of cards down the hall. But because you learned how to learn on your own, you always kept yourself going. I had one student who recently gra who recently retired from Cisco. I taught him over

35 years ago or 45 years ago actually. And he wrote me an email. He said, "I want you to know that your course was the last time I ever had any formal instruction in the classroom." and he retired as a manager from Cisco. So certifications can bring you back. Maybe you're a cobalt programmer from the 1960s or 70s. You know how to program. You know how to

interact with users and but now you need to know another language, okay? Or you need to know something more. Certifications can refresh that knowledge without having to go back to formal education. All of that having been said, I have people come up to me all the time and say, "Should I go to university? Is it useful for me to go to university?" Some people, quite frankly, no.

They are self-driven, self-motivated, self-learning from the get-go and they can get a job through any of the things I've mentioned. Right? But on the other hand, the university is where you meet up with other like-minded people. You exchange ideas and maybe you find out that you you're given a chance not to go down the path you thought you were going to go down, but there's another path

that's almost as good or maybe even better. And you meet people who could be lifetime friends. But you are the one who is in control. You are the one who can makes the decision. look at all the optures, you know, make your decision. There was this event, it was held every year is called Sebed held over in Hanover, Germany. It was the largest computer show in the

world. And I went there first as a representative of digital and then as a representative of the Linux community. And the last time I went there, I had three people come up to me. And the first one said came up to me and said, "I listened to you five years ago and I learned free software and has made all the difference in my job." Another one came

up the next day and said, "I listened to you 10 years ago. I started doing free software and now I'm the chief technical officer of my company." And the last person came up on the last day of SEBIT and said I listened to you 15 years ago. I started my own company and I employed 60 people were doing all and that is what kept me going all

these years. Those type of people coming up to me and saying I read your book Lumies. You know I had a person recently sent me some email. He says I have to apologize to you. have been mean to write this letter for the last 10 years when I was a teenager. I didn't have any money and I didn't have any way any direction of going and so

I saw your book in the bookstore and I stole a copy. It says, "And over the that got me started on the free software and over the years I bought books, I gave away books and things like that, but I always kept your book." And attached this letter was a picture of his daughter holding up the lyrics for Dummy's book. And I wrote back to him and

I said, "I forgive you." And not only that, but I know the WY people, the publishers, and they forgive you too. And so we met at Fostam this year. He came to Fostam. We met. He is the security offer officer for Amnesty International and he is a very strong advocate for the LGBTQ community. And so yes, I forgave him. And I asked these type of people to

write articles for the LPI blog of their time and their path to free software. So with that, that's the end of my talk. I'll try and answer any questions you have or any opinions. It's fine. I have a thick skin, but I'll be right. [applause] Thank you. Thank you, John, for that presentation. Yeah, if anyone has any questions for him, uh, please raise your hand and I

will run to you with a mic. >> It's another question. You said that opinion counts as well, right? It's like gigantic. Thank you for this talk. It's like literally every word resonates. My talk today is the last one in this and honestly I had no idea what you would be talking about but Kina half of my slides like I did not copy them but the ideas are

the same. I am like so grateful that you are promoting all these ideas about uh importance of open uh source software in education and about the what is the high education and like teaching people to think and I still cherish my assembly program and IBM 370 assembly like >> it was an awesome assembly language. Yeah, you know, it makes me feel good when you still think it

was worth it because I think it was. >> Thank you so much. >> Welcome. >> I'll also point out that my first job was for Etna Life and Casualty, an insurance company doing programming and I was using the equivalent of virtual machines in 1973. Yes, Lee, >> you mentioned FPGAs. I I know you do that. I know units down >> You have to hold your microphone up

to your mouth. >> You do the FPGAAS. >> You do you have a unit in San Apollo. Has it ever come up to America? And is it possible? And why don't people use more FPGAAS in their systems? [sighs and gasps] >> I don't know why. Okay. However, there's a large use of FPGAAS in the risk 5 community because it allows you to simulate the hardware before you

cast it into silicon. Now FPGAAS are if if you're doing small number of units then FPGAAS are less costly you know but once you get into actually large scale production million units or more it makes sense to go over to an ASIC particularly when the model that you're doing is very small so for example if I'm creating a CISK [snorts] processor like an Intel processor by time

you put all of the micro code the stuff into that processor, you know, it's huge. Okay. And that has to be done on a piece of silicon that you want to get at the five nanometer or even the two nanometer level. That's expensive. But with an F with with risk five, the core processor is so small that you could put like 50% more processors on the same

silicon or maybe even maybe twice as many or three times as many and therefore the cost per processor goes way down. It's more than just the cost. It's the amount of heat generated by flipping all those transistors. you know, as you turn up the clock, it's, you know, the number of transistors that are sucking up power increases dramatically. So, it creates heat. It uses power and uh

and I think it's only just a matter of time that the risk 5 processors are going to come up to be the speed of uh A6, you know, or other types of of A6. So, you know, but but FPGAAS are kind of interesting because on the Mars lander, the first Mars lander we had going around, it was designed so long ago that they could only use Intel

386 processors, which were incredibly slow by today's standards, but and it was too slow to do a lot of things they needed. So they had two Intel 386 processors, one as a backup in case the other one failed, but they had 40 FPGAAS. And so what they would do is they would program the FPGA to do a specific thing like compress a frame of graphics or something

like that and do the compression and then turn it off and they could do that be in between frames. Okay? And they could also use it for vision processing. So to to to to see what was was going because Mars the Mars rover is just going really slow. Okay. So they could use it to do vision processing and stuff like that. So they could program it to

do different things very fast. And um no I I I just think we are going to be able to have more FPGAs. My goal would be to have an FPGA that could do USB, you know, protocol, an FPGA that could do FireWire or, you know, other types of encryption. Um, the University of Sa Paulo has a wristb processor that they've now taped out that has both an

AI processor on it and an encryption processor on it to offload the risk 5 processor from doing that work. So, you know, I I I'm a big advocate of it. And as I said, I would like to replace a lot of the Nvidia. I hope nobody from Nvidia ATI is here. Replace the uh replace those those GPUs with an FPGA and then you'd be able to you'd

be able to maintain them for all for all time. >> Stop the churn. Right. Stop the churn. Next question. Um yeah with so much open originating um uh at universities you know the obvious ones like Linux at Helsinki and GNU at MIT but even R and Python and Postgress. >> How can what do you have any suggestions on how to re-engage with institutions of higher education both

as developers and adopters of a free and open source software? I mean, we're trying, [laughter] you know, I I I keep, you know, it's interesting. I I graduated from Drexel in 73. I almost flunked out because of my lack of math, higher mathematics at the time. And about 30 or 40 years after I graduated from Drexel, I was asked to go back and talk to their physics

department because they had just bought a brand new high performance computing Beaowolf system to do research at dark matter and all the professors. ers were sitting in the row with their notebooks out taking notes. And I started off by saying, "You have absolutely no idea how strange this feels." But when I was talking to them about what universities really did, I noticed that they were shaking their

heads in the front row. It was like I'm reminding them that yeah, the university is there to do research and stuff like that. That's fine. but they're also there to teach students how to learn. And it was just like light bulb went off. Yeah. We we we kind of forgotten So, you know, I I started mentioning when when I started this talk about the idea that we

have to train, we have to use the tools that people use out of society right now. Now I can understand back in 1994 if I had gone to them and said you should be using open source tools to teach this that they would say no we we we need to use this but you don't need to use Microsoft Office you could use Libra Office you don't have

to use a CAD system for everything you're teaching you use KYAD or any one of the open CAD tools you know to get started at least and the more people that use those then the more people who will do that third thing I was talking about about looking at it and making it better, okay? Until finally the gap between the CAD tools that are they're closed source

and the CAD tools that are free software open source will disappear. Okay? And that's this is this is one thing that's really making me happy these days is because for a long time the desktop marketplace for Linux was 2 to 3%. Now it's 11%. and with things that that Lenovo and other companies are doing of making of showing people how much the software that we'll see a

large swing to free I wrote an article recently about Lenovo and how they're you putting you I Ubuntu or Fedora on their their laptops and they have a s side byside buying comparison that a $780 piece of hardware if you go down the Windows route you could easily end up paying $590 more for that $780 notebook. That's a lot of money to buy something which is a

commodity which a lot of people really don't care So in the next and that and the other thing about sovereignty which is my talk there's a lot of European countries who are saying hey we have to be sovereign we have to start being sovereign from the United States but there's another part of sovereignty and that is personal sovereignty do you own your computer you may own the

hardware but do you own the software? Do you own the system? And with a lot of closed source software, the answer is no, you don't and you never will. So, I'm hoping in the next couple years we're going to see a lot of the people who said, "No, no, market's too small. Can't do it. Can't do it. We're going to change the tunes." For years, the thing

was, "I can't use Linux because there's no games for it." This really ripped me up. Okay, I mean I don't I'm not a gamer. I don't use games. The last time I played a game was a Venture on an LA36 hard copy terminal. I played it about four times. Oh, this is sick. I use computer as a tool and therefore I'm looking for the tools that can

help me do my job. Gaming was not one of them. Now my husband, he works for a gaming company and that's fine. Uh, but now people can run games, a lot of games on their Linux system and more and more every day. And then as the desktop marketplace changes, you're going to see gamer gaming companies say, "Of course, we're supporting Linux. We're not going to give up

15 or 20% of the market, especially when we have tools for compatibility that are as good as they are. So, things are going to change pretty rapidly, I think. Next question. >> We have roughly five more minutes. Does anyone have any more questions? >> He's wearing a Next Cloud t-shirt. >> I am. I work for a higher ed. We just did some uh interviewing for our next

batch of student interns. One thing I've noticed from last year to this year is a downward trend in open source contributions, poll requests, collaborating with people in an increase in students talking to agents uh to do their work. And uh I'm wondering how we can best encourage our higher ed students to talk to one another and get involved in open source. I think you're going to have

to take a leadership position in that. I am not a big fan of what people are calling AI these days. Okay? I look at AI as actually an expert system or what we call AI as an expert system that is able to self-modify its tools and its rules for creating that. That is not a a a an intelligent system. Okay, an intelligence system can work from nothing

and create a a a thesis about why something happens. That would be intelligence. How many of you have ever read the book I Robot? If you haven't, you should. It's a great idea. We started off with very simple robots that could take care of, you know, people and stuff like that. And and and in the end, it was a robot or we think it was a robot

that was so intelligent that it became the president of the universe. But they always had to follow three laws. The first law was you could never allow a human being to come to harm. The second law was it would have to follow the the the orders of a human being unless it caused harm to a human being. I forget what the third law was, but the the

whole book was about how when people strengthened one law, how it affected humankind. Okay, it's worth reading. I just look at the human mind with a number of neurons and synapsis in them, the complexity of it, and we don't even know how it works yet. And so to say that we've created artificial intelligence, I have a problem with that. However, the other problem I have is going

and using an AI system and having it come up repeatedly with the same answer that is correct. How can you maintain something when every time you run it through AI, it creates a completely separate, completely different project, different problem? I have a problem with that as far as certification goes in particular. I would just have to say, you know, it's fine if you use AI to filter

through lots about lots of data and stuff like that, but I'm not really happy with AI actually generating code. We have a hard enough time with human beings generating reliable code and that we're supposed to be intelligent. I don't know if I answered your question or not, but I think that universities have to be have to take a leadership position on this. And I will tell you

that when I was teaching, there was a couple times where I found people students using things that were against the rules and I kicked them out of school. It was hard. I didn't like to do it, but I had >> I think we have time for one more Hi. just so we don't have to end on that note. Um, as an instructor who is in part teaching

my students about how to become responsible open source ecosystem contributors, um, are there particular sorts of supplemental resources or textbooks about the process of contribution and being in that community generally that you recommend beyond specific projects contributor guidelines? Yeah, I mean there's a there's a book by Jono Boo Jono Bono who uh has been a long in in the open source community a long time. He's been

community managers of various communities and he's written some books on how to make contributions and things like that. There's lots of articles on the web about how to contribute to open source projects. I recommend that a student go in and find a project that calls to them. Maybe they like audio, they make like music. There's lots of audio players and codecs and things like that you can

look at as projects and to go in and find their bug list and find the bugs and find a fix a bug in the bug list and then submit it. And what'll happen is the maintainers might come back and say, "Hey, yeah, you fixed a problem, but your code doesn't look like the rest of our code and we have to maintain it after you're gone. So, could

you write it more this way?" you you'll make another submission, but eventually they probably will accept it. And then you do another bug. And this time it'll be easier because you know how they you know how they think when they're accepting your bug fix. And after a while they'll say, "Hey, you're pretty good at this. Would you like to write new code? We have this list of

things we'd like to do, but we just don't have enough people to do it. So you can go ahead and write this new code." And then after a while, they'll say, "You're really good at this. would you like to become part of the core team? And I had a friend of mine who had been doing security for digital for a very long time. He was tired of

doing security and so he joined the Firefox team using exactly this method. Okay, so this is a way that you can get your students to come in and there's nothing wrong with an undergraduate student coming along and saying, "Hey, my project is fixing bugs in this area, right?" And they can get credit for it, Okay, thank you very much for coming. >> Thank you all for attending.

The next talk will be around 12:30. Hope to see you there. >> Sure. >> Hey man, turn this thing off. Amazing. >> Thank you. >> Professional institute. I guess this Sorry for you guys. Hello everyone. Can I just get your attention real quick? I'm Patrick Massen, Apario, helped organize this uh education section or track and we're also hosting a education pop popup or I don't know there's

some cool term but um anyone who came to one of the sessions, we have a free drink coupon. Um I don't sell anything. We're a nonprofit. So we're just OPIO helps promote open source and higher ed. So, if you're interested, I'm in the back. Um, in between sessions, feel free to come by and I'll give you a couple coupons for a cocktail. to it. I'll tell you

Everyone, our next talk will begin in roughly five minutes. Now we're only I'll leave it. Good enough. Afternoon everyone. Uh, pleased to announce Jill Needleman and Dex on their presentation. If you can give them all a round of applause. All right. Well, welcome to Extreme Home. Wow, that's really loud. Okay. Uh, I'll just have to talk a little softer. >> the mic. >> I can't even see

it. Oh, there we go. Yeah, technology. You just need to move it. Uh, okay. Welcome to Extreme Home Labbing. Uh, this is modernizing infrastructure at one outage at a time. We are part of an organization called WCC Comps. There's just two C's. Uh but uh we will be going over a couple things today that I think I and I hope most people here are interested in which

is migrating to open source software from proprietary software. Um it's going to be an interesting talk. We have both the overview and then the technical part. So you will be getting it all. But first let's do a little bit of background on about us. So first one up is me. Uh my name is Joe. I am a security researcher. I'm an educator. I also do devop uh

sec ops by day and I lead WRCCDC and WC comps by night. What that means really is I technically manage uh the technical part. Uh it's in the literal sense. Um all of the infrastructure, all the resources and make sure things happen. So students can compete, students can run events and students can learn. Um I've been doing this for about eight years. Things have changed a lot

in eight years. Not necessarily um from a software and technology standpoint, but there's a lot it's grown. Students want more events, want more resources, and part of this talk is how we scaled when we were forced to migrate. Then next up is Dex. >> Hello everyone. I'm Dex. Uh during the day, I'm a professional software engineer. I've been doing that for approximately the last 10 years. Uh

at night I am a hardware uh security researcher. Um which really means that I love to buy hardware. I love to mod it and I love to get it to do things that it was never intended to do. And I specialize in a field called high performance enterprisegrade jank. And what that means is I love to go on eBay. I love to buy your old outdated Dell

power edges, your old outdated sands and whatever. And I love to mod them in order to get the best performance out of them uh for as cheap as possible. Uh and and a lot of that's actually used here. Uh you'll be seeing that a lot uh in this talk. So don't be afraid to buy it from eBay. Anyway, what is the problem? Why did we end up

here? Why are we doing Extreme Home Lab? Well, simply put, and I stole this from Dex, we are the most overloaded VPS provider. Unlike most VPS providers where you can get more resources, we are academic. So what we have is what we have and we have to beg, barter and see what can you know be found if we have to expand which makes it very problematic. But

you may be wondering and I'm just going to have a do a little interaction from the audience. How many people are familiar with WC WRCCDC or CCDC in general? Wow. I think the entire like everyone here was a current or former competitor. That's that's awesome. So, this is a redundant slide then. You don't really need to know this, but for the people watching this recording later. Um,

we're going to explain things, but I'm really glad to see the uh the interaction from the community. So, WCC is new. Uh, it's only existed for about a year and mostly on paper. It is the Western Coalition of Competitions. Why did we name that from WRCCDC? Because we are not just doing WRCCDC anymore. We have expanded. We are doing WRCCDC, PRCCDC which is Pacific Rim unlike Western

and we are also doing circus and a bunch of other competitions. Because of that we need to do a lot more things and this is really complicated because when CCDC started it was national event which was a pilot then it broke into regions. We are one of those regions and when we started when I started uh ex working on this and trying to expand it when Dr.

Brown who's in the front row um we had the idea that we wanted to make this bigger and so we started with about 14 to 16 teams at uh qualifiers and eight teams got to move on to regionals. We are now averaging about 25 to 30 teams doing the qualifier and that is just for the qualifier. For our events that we've been running for the what we

call invitationals, we have been having over 40 teams compete from across the country. That's a huge expansion and it makes things really difficult. So when you do these events and you have all these schools and all these teams, one you have to make sure virtual works. If you're doing a virtual event and you can't get in, the event sucks. So, we have a lot of things that

we have to do to make sure all of this works. Um, in the past year with the collaborations, um, the calculation was we've had about a 2,000 student competitors a year participating in our infrastructure and using our resources. So, a huge expansion on the um the resources. What is the goal though and why does scale care? Well, clearly the audience cares because they know about CCDC. But

the goal of this event and these events in general is to bridge the gap between industry and academia. You want something where students can learn in a safe environment and get all of that knowledge without risking getting fired. I mean, we still fire the students virtually at the end of the event. Like, you guys, you know, are all fired, go home. But that's just for fun. We

don't actually want students to be risking their actual careers learning something new. We want them to feel comfortable to break things and sometimes that's our own infrastructure. Um, and we want the ability to test new re ideas. There's a lot of new things that you will hear and you hear during a conference like scale. Those don't generally get picked up by most industries for another couple years.

We are bleeding edge because we'll just try something new. Um there's a um subreddit called self-hosted where every Friday there's new software that's all AI generated. It's great. We pick all that stuff up whether or not it's maintained and use it because it's great. You sometimes get gems, sometimes you get real interesting things, but we try new things all the time for students. Sometimes for fun, sometimes

for you know actual educational benefit. But the main thing is we are entirely volunteer-run. Nobody gets paid here. All of our volunteers get paid in jokes, chuckles, and pizza if we can afford pizza. So, the volunteer team is very small and very dedicated. And we all do this because we were students sitting in the competition and we wanted to give So, um this is from regionals. This

is our operations room. Sorry about the trash and mess. Uh it gets a little chaotic and we kind of just stack things. Um, but we uh average, unfortunately, about 50% turnover year-over-year for volunteers. It's not that it's difficult. It's just when you have a day job and you're being asked to do basically your second day job in the evenings, a lot of people can't handle it. That's

no, you know, slight on them. It's a lot of work. So, we have to be able to handle these things where students can get that benefit, but also not overload the volunteers. So when I started working on this and took over uh we had the no concept of leads. There was a guy named James and James was CCDC. Um I uh you know if you ran something

or you had infrastructure he had any other issues. It was James you know just the one guy. It was overloading for him. And so when I got the opportunity to take over after he had finished volunteering, I uh I made it a point of making something called leads. You have your infrastructure lead, you have your scoring lead, you have all these different components where people who want

to put in extra work can volunteer and do that extra work without having it all on one person. The leads are extremely dedicated. Dex here is the infrastructure lead. And all of that means they care. And so you make sure that the people who care are available and rewarded as much as you can. So make sure they have a hotel room to sleep in at the end

of the day for the event. Things like that. Um but it is essential that these leads don't take on too much because then they burn out and then you don't have a lead and now it falls back to one person. This is probably pretty obvious for most people who have done any research or on you know managing teams effectively but you know it's good stuff to reshare

sometimes. The other thing is you want to make the learning curve low for volunteers because when you are working on things for example networking and you have to know obscure Dell switch configuration that becomes very problematic because not everyone can go figure out how to terminal into a switch and you know configure it. So part of this rebuild is what are the software what are the things

worth purchasing and what are the things worth building out that are open source. Um we want to make sure that there's that balance there and um you know that everything works and I I realize that I don't think anyone can actually read the text because it's very small on this projector but uh sorry about that. Uh I'll make sure the slide decks are available. Um, but at

the end of the day, it does take, you know, a village to make this work. Um, this is one of our rebuild sessions. This was the dedicated team. You can see them in the audience here. These folks, some of them were new. They had never volunteered before. They just showed up and wanted to help. And so during these rebuild sessions, we'll get to that in a minute.

You know, we spent full days from like 9:00 am to 900 pm where we were rebuilding and fixing infrastructure and just getting everything working. But that's not something you normally do. So, why did we have to do that? Well, um let's ask a hypothetical. What happens when you run a nonprofit, you know, academic event and you can no longer afford those pesky licensing costs? Um well, you

could do something. I, you know, there there's ways but you know that becomes very legally ambiguous and so we kind of avoided that it you know creatively um so how did we avoid that well back say if we can save questions to the end yeah thank you um background on that VMware was no longer VMware and suddenly we did not have academic licenses we had relied on

academic licenses for VM VMware for years and we suddenly didn't have them. We had a perpetual license for VMware 7 because we always meant to upgrade to 8 and then we didn't be a we couldn't upgrade to 8 because we no longer had licenses. So our infrastructure stat for three years almost being just VMware and we were literally limping along. There were things just breaking and at

the event the regional final last year something [clears throat] amazing happened. Our certificates expired and I raise a hand. How many people here know how to rotate root certificates on VMware infrastructure? Three people. Yeah. And and one of them is because we had to do it live. Um but um yeah, this this was a mess. So all of our infrastructure went out during an event, and that's

always the worst time to have stuff happen. Maybe it's the best, but it it really sucked. But all of our infrastructure was built for the VMware product. We had a sand. We had hypervisors that were set up with no discs. They only used the SD because that's how VMware wanted you to set it up at one point. Um, we also used Windows and Active Directory because it

was easy. You just had academic licenses. Well, what happened before VMware, you know, lost their licensing, you know, for academic was Microsoft got rid of that as well. So suddenly we had a Windows system that we couldn't reinstall ever. So we were stuck on Windows Server 2016, which wasn't too old really, but it wasn't really good. And we were locked into all these little proprietary voids that

didn't allow us. And the other fun thing was we built everything as a VM. If you want to install a new piece of software, new VM, it was really great. Um, only problem is that takes up a lot of resources. So what happens when you want to move away from VMware? Well, um there's a few things that VMware is really good at and still I think they're

probably one of the best. One of those things is votion. When you want to load balance resources and you have thousands of VMs, votion just takes care of that. It's painless. It's almost completely seamless and students don't notice. The other thing is um our storage provide our storage sand was tiered. So we had flash storage, fast disk, slow disk, and that's really helpful for when you're running

events because you want all those fast discs. And then of course the other challenge is you have to teach all of the volunteers all the new stuff because you can't have volunteers just being like, "Yeah, I'll figure it out while I go." Because you're having to run an event. Finally, rebuilding requires a lot of on-site time, which we can't easily get. So, you know, it was a

a a choice and a challenge. The older operating systems were becoming a problem. We had so many security vulnerabilities. At one point, I think our red team for the event accidentally attacked our infrastructure and took it out. Not a good thing when you're running event and not a good thing for security. Saying you're a security competition and you get hacked by your own team. Not great. I

mean, it's not the worst, but not great. The other thing was the Dell um switches started becoming amnesiac. Uh they they would uh reset every time you rebooted them. So you just never rebooted them. Um these are types of things that are be that are really a problem when you're trying to keep something working. Um and then finally, um all of our gear that was Cisco was

end of life, which also isn't great. So we wanted to do something better. Um and this is just uh a picture of one of the competition events. Um but a few of the resources that we provide and why we need all this working. We provide the largest archive of competition VMs and the largest resource of uh network captures to my knowledge on the internet. There's no one

else that provides 16 terabytes worth of resources freely available. We've had grad students use our resources. We have had um companies use our resources. um it's all freely available and it's generated for students. So um and then of course the competitions we run, we have to have stable resources. So we're finally to that home lab stuff and let's go into it. Um so when you're building something

you have to rebuild, now is always the perfect time to switch to open source because it doesn't get worse. Um, the longer you wait, the more your software breaks and at least you can have updated software that's free. You know, it's going to be broken sometimes, but at least you usually have a community attached to it. And we wanted to try to give back because that al

that open source community we utilize, but we didn't really give back. So, all of our stuff going forward in WC Comps is all open source. Our GitHub repo has all the software that we custom build. Um, all of the changes, all the tooling, all the autom OpenVPN, you can use Tailscale, you can use a bunch of different things. Um, for web services, do you want to use

Docker? Do you want do you want to use Podman? There's a bunch of different choices that you can pick from. No answer is the wrong answer. We just settled on what seemed to work at the time and completely untested, I might add. Um, all of this software has trade-offs. For example, why did we pick Proxmox instead of uh Zen? Mostly because Proxmox is used by most of

the students and it has a bigger community. So then what's next? Well, do I use Dockage or do I use Portainer? Well, Portainer is a lot more complicated and it's getting a little bit more paid. So, Dockage it is. Is it the best software? Maybe not because you can't really do like LDAP on authentication or OOTH easily, but it's there. Um, and then using a wiki, you

know, people wanted wiki, but we've been using Docker wiki, so we kind of just rolled with the different options until we had something that sort of worked. And that's the key thing about doing this. You have to iterate. We had to switch out software and have to change settings throughout A final note on Proxmox. A lot of people here may be familiar with Ludus. Ludus is a

um lab generation software that goes on top of Proxmox. We chose not to use it because it does not work for a competition. When you're building a competition, you're not building a lab. It has rules and how it designs things that completely don't fit with how you need to design a competition. For example, a competition, you need everyone to be able to communicate with each other. Lutus

creates containers of networks like they're all isolated. The other thing is the Proxmox API is really bad. Um, don't let anyone tell you otherwise. Um, as much as I love using Proxmox, the Proxmox API will randomly fail out with locks and not tell you. And it doesn't have clear instructions how to release those locks or retry. Um, that's just a few of the issues, but so what

we ended up having to do is build our own custom software with a ton of retries and weights. Um, that actually worked for us, but it may not work for Um first run everything crashed. Everything broke. Um authentic when we used it for our you know 30 volunteers we could log in fine. When you have 300 people logging in at the same time it just kept eating

more and more RAM. We would double the RAM on the VM because it was running Docker host. Um it would go from like 8 to 16. Then it needed more. We finally settled on like 64 gigs of RAM that day and we were like okay it works. Um, Authentic seems to have the philosophy of enterprisegrade software where you just stack stuff until it works. So, if you

need you, you know, you have a thousand users, just keep adding more RAM and CPU until it works. Um, that doesn't really work for a home lab because you don't have those constraints. So, we had to go through a lot of optimizations to make that work. Um, the DB was getting stuck on these weird slow queries. We couldn't figure it out. We had to keep flushing them.

There were a bunch of strange issues. And then the final issue was the VPN was unable to stay connected. So teams kept dropping out. It was a lot of fun. I don't know if we panicked. Okay, we definitely did panic, but the solution that we came to was just throw more RAM at the problem for the time being. Second problem, second time we tested and load tested

and we thought everything was fixed. Then when we started it, we used Net Bird for our VPN and it has this quirk. It's not really designed for academic events like this. when the users reconnected and because they're the same schools, they were connecting to previous pods. So if they were team one in the first event and they were team 16 in the second event, they were still

connected to team one. So there were multiple teams on the same user accounts, which is bad for an event. Um, and then we wrote some new software, didn't test it before we pushed it to prod, and when you stopped the scoring engine temporarily, it reset the database. So then the scores were gone. So, we had to figure out how to restore our backups that were take in

increments of every like 15 minutes. But at least authentic was working this time. You can get the idea. Third run worked a whole lot better. Everything was fixed, right? Nope. We crashed our core networking. The uh core router forgot its configuration. We forgot to give teams access to things because Proxmox had these policy rules that were timebased, but nobody was there to figure out how to update

the timebased policy. Again, volunteer. you have to make it easy. So, we ended up giving teams a competition start where they couldn't get into anything. That took a bit, but I'm sure it's fine now. Don't worry. I'm pretty sure it's fine. But, you know, if it isn't, I think what we'll do is we're going to hand in to uh Dex the uh fun that we did with

optimizing and some of the stuff that we've learned. So, Dex, up to you. >> Oh, boy. Everyone buckled in. We're gonna go very quickly. All right. So, the problem is uh like we mentioned, we host approximately over 40 teams. So, you know, at max we can host 50 teams. Uh conservatively, let's say we have 10 VMs per team. That's about 500 VMs, right? If every VM has

two cores, that's a thousand CPU cores. If every one of those VMs also has 4 GB of memory, that's two terabytes of memory. Um, our resources are a little bit substantially less than that. Um, our memory is actually not that bad. Um, it's 2.21 terabytes, which means we can barely fit in, but that also doesn't count like our core services. Our red team is also running on

that same infrastructure. So, yeah, we're kind of redlining there if we're just kind of there. So, the main problem is that look, we are a nonprofit. Um, we have a limited budget. We don't have enterprise scale budgets to scale out our hardware like everyone else would. Uh and look, hardware just isn't plentiful anymore. Um especially now in the era of memory and everything else going on. It's

just it is very expensive. We can't just scale everything like an enterprise would. And of course competitors are competitors don't care if they access these VMs and these VMs are not very fast. Of course, they're not going to care why. They just care that their VM is slow and they don't care, you know, that we don't have the budget to scale up for that. So, what is

the obvious response to this? Our response is obviously to give up, right? Talk is over. Goodbye everyone. No, obviously not. You know, that's a quote from Wasabi that he told me multiple times while we were rebuilding. We don't have the budget. Um, and so we kind of have to make do with what we got, right? let's reconsider what's going on here, right? Our budget only controls hardware

purchasing. It doesn't control the entire environment. And what do I mean there? Yes, the majority of your performance is coming from your hardware. 60 80% of your performance comes from your hardware. But software optimizations can actually bring you that last 20 to 40% of performance just to bring you over the finish line to get you to where you need to be. So talked about this a little

bit earlier. Yes, memory is expensive. I don't have a solution to that unfortunately. I would love to have a solution to that. If you are a nonprofit, if you're a school, you can try going over to your IT department, ask very nicely if they're throwing away any servers, any old memory. Um, some e-waste orgs are also 501c3s. If you're a nonprofit, they might be willing to help

you out there. Um, but the core of the problem is still that we have a lot of competitors. We have a lot of VM. There's a lot of memory. We can't really get around that unfortunately. But we can try to kind of fudge it a little bit, right? If we look at what exactly a typical environment for WRCCDC is, um you'll notice that we actually each of

these are individual team environments, right? We have one team over here, another team over here, and we're pretty much deploying the exact same VMs to every single one of the teams. And you'll notice that for each one of these memory regions within the VM, our kernel over here in Debian is the exact same as the kernel over here in this Debian VM over here. And same for

all the other kernels, base OS services, the services that we put on top as our customizations. The only thing that really changes between the VMs is the competitor tools and services, what they're running on top of the VM. And so what we can actually utilize is something called kernel same page merging. And if you've already worked on the Linux kernel or you've touched prox, you already probably

know what this is. But what it basically does is it goes through and it merges similar pages within memory in order to kind of compress down uh the space that you are taking up. And we use this immensely during competition. This is uh one of the things that enables us to run this competition with the limited amount of RAM that we have. Uh one note about this

is that Proxmox only well Linux in general the default is that it only takes effect at 80%. So that's why you will see your memory creep up all the way to 80% and it will stay there. That is kernel same page merging at work. And so we already talked about the memory problem. Let's talk about the compute problem, right? What do we need, ladies and gentlemen? Quality

of service. When did we need it? That's right, yesterday. All right. So, compute quality of service is kind of like network quality of service. Right? If anyone here has worked, networking, quality of service is basically taking a limited finite resource and parting it out in a kind of fair way. Trying to ensure that everyone has equal access to the resource. So, if I have, for example, a

100 megabyte pipe uh on my Wi-Fi, I'm going to limit the specific amount of connection speed for one individual down to, for example, 10 megabit per second so that everyone has an equal amount of access to resources. Not one person can hog the entire thing. Same thing for QoS in terms of compute. That means that we are going to be parting out the compute resources so that

everyone has a fair shot at using the VMs. Which means that unfortunately, yes, if you use a lot of compute during our competition, you're going to have a very bad time. Please don't run LLMs on our VMs. But this allows for everyone to have a nice time. And so how do we actually implement this? uh is actually a systemd setting which sets over your cgroups if you've

worked on the Linux kernel stuff. Uh and all it is is this little parameter here called CPU quota and it is a percentage that you can assign uh for the amount of cores that you want to give to a specific process. In Proxmbox your VMs are considered processes because they run on QMU. So uh the parameter in Proxmox specifically is called CPU limit. It sets CPU quota

directly in systemd. Uh during our competitions, we actually set CPU limit over to two. And that's basically ensuring that everyone has access to at least two cores at one time and then everything else. And then you might be saying, well, okay, what's the difference between that and saying two cores for the VM? This is your typical two core situation where you have youruler executing your task one

after the other. We're assuming that each task executes in about the same amount of time. And we're saying that, you know, the overall execution time is about t equals x. Now in a quad core situation of course you know for we can have each of these tasks executing in parallel. We have two more cores to do the work in. So of course we can do the tasks

in half the amount of time. But if you think about it from the QoS standpoint we're doing all the work up here and then for the rest of that time we're actually blocking out uh the VM. We're blocking out the compute resources. And you might think okay well what's the point of this? The average is still two right? Well, for a competitor in their VM, this actually

means that they're getting a little bit more of a responsive experience. This means that whenever they're typing in a command and they hit enter, that means their VM responds in that moment and therefore their VM feels a lot faster than it actually is on average over a long period of time because it's responsive in that moment. So, uh, the Linuxer does kind of help us here. I'm

not going to go over how it works, but basically uh the more compute that you use, the more dep prioritized you are by the kernel. And therefore, uh anyone who is using a lot less CPU, if they uh attempt to use like a amount of high amount of CPU compared to someone who's been using a very high amount over a base load, that means that the Linux

scheduleuler will actually prioritize that lower process over the other one. all right, we talked about compute, memory, let's talk about storage. So storage is expensive. Same thing as memory. Whoops. Uh but here's the thing, right? Whenever we think of storage, we usually think of hard drive connected over to a server. But over here in virtualization land, the usual thought is that we're going to connect all of

our storage over to network. We're going to use a sand, we're going to use NFS, etc., etc. Because the thing is that and this is our com uh compute stack right now over at the WRCCGC data center. Um you these bottom six servers over here are actually the uh physical servers that our competitor VMs run on. If we were to add hard drives to every single one

of these servers, we would have to buy drives in batches of six, right? So one terabyte over here, etc., etc., etc. And that just becomes cost prohibitive at some point. There's a scaling problem there. The advantage of centralizing your network storage means that I can buy a 20 terabyte drive, put it on a network storage and then suddenly all these servers have access to that storage. Now

it's magic. Now there is something to be said that people will often say that true is not great as a storage backend for your VMs. I am here to prove them wrong. Trunaz is great as your storage backend for your VMs. It comes with a very nice guey. It comes with all the configurations you need to set up all your networking stuff, your ice is scuzzy, if

you want to use NFS, uh, you know, SMB, whatever you want to use as Tunas is great. It auto updates. It's very stable. But ZFS is very, very hardware dependent, and that means that you can mess it up very, very fast, especially if you don't know what you're But uh every time I talked about this, everyone would always tell me that NFS is really, really bad for

doing VMs. And it's not. NFS is just misconfigured. So let's talk about the hardware selection. The thing is buy what you can afford and do the best that you can with it. And that seems obvious, but you know, if we think, okay, I need to have this amount of hardware in order to do this amount of work. Unfortunately, that's just not really how it works for nonprofits.

We have to just work with what we got. So, I can give just very general buying advice. If you're setting up a brand new environment or you're looking to improve yours, please prioritize your CPU cores over your CPU speeds. You will have a lot of parallel activity from ZFS plus NFS. Um, and those favor your CPU cores over your CPU speeds. Doesn't really matter if things are

30% faster. It just needs a lot of CPU cores to execute in parallel. your VMs will depend a lot on random IO. Your sequentials don't really matter all that much. Uh and your caching is just as important especially in ZFS and we lean a lot on that uh during our competitions. So uh like Wasabi said a little bit earlier, we run two NetApp uh 2.5 in Jbods

uh with 40 discs each and this is arranged in a 7x6 disc RAID Z2 with eight hotspars. If you don't know what that means, we'll go over it real quick. So, uh, if you're talking about ZFS, right, each of your individual units within your Zpool are organized into what are called VDEBs. And these individual VDEBs are actually what contain your discs. And this is actually uh your

VDES introduce your individual unit of redundancy. So for each one of these, we have it set up for RAID Z2. All that means is that we can lose two discs uh within this array and we won't lose any data permanently. So on the internet often when you see ZFS tuning guides they'll often talk about uh scaling your uh amount of VDEVs uh because this is a Z

poolool uh when you scale in this direction by adding together your vubvs because all of these go into a stripe what you're doing is you're increasing your storage capacity but you're also increasing your random IOPS. This is always what the internet says to do. The thing is that you can also scale it in this direction as well. What this does up in this direction, the vertical direction

is it increases your redundancy and it increases your sequential IOPS. When you're scaling in this direction with your V-Dubs, that means that you're increasing your sequential plus your random IOPS. So, two different things, two different directions you can go. Internet tends to just say this direction and they are All right. So what we run over at WCCC comps, we run uh 300 GB of memory as L1

cache. Uh this is enough to fit all of our competition environments because we also use Qcow 2, which means we get to take advantage of a lot of compression, which means that our accesses for reads are actually incredibly fast, which really helps with VM responsiveness. We also have two terabytes of Intel and Vinme as our L2 and then 375 as our separate log S- log. Won't really

go over what an S log is here. Um, RIP Optane. Uh, I really hope that I really wish that Intel didn't kill off the product line. Uh, and this is my PSA for everyone over here. Please do not turn off your sync rights. Uh, everyone on DFS tuning guides will often say, uh, if you can't afford the memory, you can't afford the L2, uh, just turn off

your sync rights and it'll make things a lot faster. Uh, please don't because your data integrity goes terribly. And during competitions, we fa we have a lot of rights and there's a lot of chances for corruption and a lot of chances for things to go wrong. You will need sync rights, trust me. All right, so this is the typical amount of traffic that we'd have during competition.

It's a little bit dim, but you can just about make out the blue line and the purple line over there. Our blue line is actually our reads. The purple line is actually our rights. And you'll notice that our reads aren't actually that high, but our rights are incredibly high. That's uh peing about 10 Gbits per second. And the reason for that is that if you think about

it, when a VM is already booted, there's not really much reads that are occurring once that VM is booted because obviously everything's already in memory. But databases and logging and everything else is still sending data over to our array. And so obviously rights are going to be very very high. And so this is just some buying advice that I have right now. Um I don't know if

these listings are still active right now, but one terabyte of uh Intel MVME is like $160 right now. That's only up by like 60 bucks right now compared to memory, which is 300% higher. The usual advice when it comes to scaling ZFS is that you increase your memory as much as you can. Obviously with memory 300% that advice no longer applies. Buy NVMemes and input use them

as your L2. It's fine. Uh your Intel optains are also right now going for 140. Like Intel has already stopped manufacturing these. Like get them while you can. This is my other PSA. Please do not buy these. All these EFS tuning guides from like the last 10 years say that these are great. They are no longer great. They only like do one gigabit per second. They have

two PCI lane P two PCIe lanes. They suck. All right. With all those optimizations, we get this. And uh yeah, it's not it's not great. Um we're maxing out the network connection actually. So we're getting like the full 10 gig over here. But um yeah, these are SATA speeds and this is yeah, this is hard drive speed. Yeah, that's not great. And so let's deal with it.

NFS for virtualization is slow is what everyone's been telling me. It is wrong. NFS for virtualization is not slow. The default settings within the uh NFS mount configurations is just This parameter over here called endconnect is usually set over to one and that means that there's only one TCP stream that's created whenever you're mounting a NFS server. However, so let's take a look at why that's bad.

So this is a typical connection to a NFS server of N connectnect equals 1. You have each of these sequential operations going through the server and each of these operations going individually back to the client. And you'll notice that everything's occurring in sequential order. This read h this read command has to come you know first and then this has to come second and everything else right. It's

not very parallelized. What we can do with endconnect however when nconnect is equal four is we can just shoot all the operations through all the TCP streams that are created all at once and it's suddenly we have all these parallel connections all able to service all at the same time and the advantage of that is that well one of the caveats with this is that CPU cores

really matter right because if you don't have enough cores to serve all these requests like you're going to still get sequential operation again so next let's optimize QCA 2. So your Qcow 2 images are usually how uh Proxmox handles your filebased storage for ZFS over NFS. Uh it is very easy to get wrong. The default settings in Proxmox are actually incorrect for doing this. They have uh

submitted a patch to fix it but it's only partial. We are currently running a custom version of uh Proxmox to deal with it. We'll be submitting a patch over Proxmox uh proper once we have that finalized. But it's these two settings over here. Qcal cluster size and extended L2. Uh your cluster size is kind of the minimum amount of read and write you can perform on a

Qcal 2 file. And extended L2 will go over that. But you the idea is that you want to maximize your cluster. You want to match your cluster size over to your record size. So this is the default uh layout of a QA2 file. We have our individual clusters over here and over here. And whenever your VM is performing a read or a right, they're having to create

these entire cluster blocks. So even if your read or your right is four four kilobytes, it's having to stuff it into this 4 kilobyte block. And you can imagine that's kind of inefficient. And so what extended L2 does is in creates these little subclusters from these main clusters and it can break these things up into 32 uh subclusters from one cluster. Which means that for a 128

kilobyte cluster, we now have individual smaller reads and writes at 4 kilobytes uh going all the way down to 32. Which means that we are now actually ma matching the minima. Well, we are matching what a typical file system would actually read and write over to these files because your typical like X4 from nowadays would match uh these 4 kilobyte writes because it operates on 4 kilobyte

chunks. So this is another recommendation for hardware. This is a Intel platinum. It's a little bit outdated. I think it's end of life, but it's only 3 33 bucks. It is 28 cores. I do know that there is a VPS provider that currently uses this inner server. Um yeah, not great servers. Please don't buy from them. Um but these are great if you want to go over

and scale your environment and get performance for not that much money. These have not actually raised any amount of price during this current time. All right. So, we want to now maximize our network bandwidth, right? Because this is one of the network cards from our servers. Uh it's Intel X5500. Uh but our network connection speeds as you can see as you saw a little bit earlier was

only you know really maxing out this one port one 10 gigabyte or 10 gigabyte port. But all these together are 40 GB max duplex. you know, I bought the network card. I want to use the network card. I'm going to use all the ports, right? And usually when you're talking about that, you usually think about bonding, right? And this is another place where you can actually follow

over in terms of your network setup network setup because whenever people turn on uh your bonding, your LACP nowadays, uh they tend to enable what's called LACP23. And what LACP23 does is it takes the hash of the IP plus your MAC on the sender plus the receiver and it puts them together and it determines which port the packet goes over to. And this is all configured within

your switch. The problem is that for a packet that comes from over here, the switch will determine that okay, this packet always goes over to this port over here. And it won't do a per connection uh load balancing. It won't do any of that fun stuff. you're going to still get the same 10 Gbit per second connection limit because your connection isn't actually optimized properly. And so

instead, we can switch over to LACP34. This takes in account TCP and UDP ports, which is important when we combine it with N connectnect from a little bit earlier. You will remember that NFS uses TCP connections. By opening up more client connections, we are changing the source port for all the connections. And therefore, the switch now knows, okay, I'm going to route this packet over to here

and this one over here for each of these each of these individual NFS connections, which substantially speeds things up. Tada. We have managed we have managed to double our speeds for free, right? This is like great where we are now maxing out our uh 20 gig connection. We can actually go a little bit higher. We are bottlenecked by the mix that we have right now. There's a

plan to fix that. But our randoms are still terrible. This is substa speeds unfortunately. So let's do some final optimizations. Let us increase the packet size to 9,000. What this does is the typical packet size limit is 1,500. By increasing over to 9,000, that means that your individual L2 packet can be 9,000 bytes wide. Uh that means that you have a lot less overhead in terms of

like TCP segmentation, the amount of Ethernet packets that your nick has to unpack at the other end, you know, blah blah blah blah things that your switch has to deal with. By increasing over 9,000, we don't have to deal with any of that. And for latency, we can actually disable something called hardware offloads. So within your nick, your nick is actually trying to do your CPU a

favor. It is trying to offload your TCP segmentation. It's trying to offload as much processing off of the CPU so that it could save itself for actual computation. But the hardware inside of your nick designed to do that, the ASIC is pretty slow at doing it. It'll do it. It'll do it at 10 gig, but it it's going to go pretty slow. So, if we disable that,

we're at we push it over to the CPU, which is substantially faster, like gigab gigahertz faster. And it's just these eight settings over here that you have to enable within your typical network config. And bam, we now have approximately a five times increase on our random reads and writes. Our uh sequentials are now maxed out to the extremes of basically the nick that we're running. everything is

basically on the bleeding edge and uh it we are at the edges of our hardware. We couldn't possibly get a little bit faster. Maybe some custom software tweaks, a few patches over to QMU somewhere, but this is pushing our hardware basically to the limit. All right. So, as much as I would love to keep rambling on and keep everyone here, um there's a lot of optimizations that

we went through for this competition, uh there's a so much more uh that we did in order to get our competition like online and functional and as performant as it was this season, but unfortunately, we just don't have >> all right. Is this thing still? Yep, I can. Is it on? Yeah. Okay. Uh All right. So, uh, with closing, what I want to do is I want

to just give a shout out to all of the volunteers. Um, we have, uh, a number of people who have done a lot of work. We can't possibly mention all of them, but what I wanted to do is just say to all of the volunteers who participated in our CCDC, WCCC, PRCCDC, circus, and so on, just thank you for being here and sticking with us, especially this

year that we've rebuilt everything. A few folks have gone above and beyond and we want to give a shout out. Um, this is a mix of nicknames because in uh WRCCDC we just use nicknames because we're on Discord for all of our volunteer stuff. Um, but I will uh take a few of them. I just want to shout out to Toku, Alchemy, and Bisad for being there

nonstop helping set up stuff. For example, Alchemy is handles all the report grading and has been there to help set up every single day doing a hardware stuff. So, not only has he been doing all the stuff that he normally does, he's also shown up to every event, set up software, set up hardware, and just been there the whole time. And so, um, that's a few of

my folks. Dex is infrastructure lead has a few others. So, uh, Dex, >> yeah, I personally want to give a shout out to, and not in any particular order, Payton Ericson, uh, Sean Richardson, Vessa Exeria. She's going to have to forgive me for her pronunciation. Uh, Jimmy Night Rabbit, and so many other people who helped out on the core infrastructure team. Um, I wish I could list

out everyone, but I'm forgetting a few names, I'm sure. But, thank you to everyone because you guys are the ones who helped make this competition what it was this year. >> All right, and with that, here's the information if you want to snap a pick. If you're interested in volunteering or you want to participate in this, you want to ask questions, we are here to answer those.

Um, we have a Q&A coming up, but you can get information here and sign up to join. Um, and then with that, I uh we have uh roughly five minutes left, give or take. Uh, 10 more minutes. Oh, look at that. We're running 30 >> No, what? >> It's 120. >> Oh. Oh, I have 10. We have 10 minutes instead of five. Okay, I'm good at math.

Um, but uh if anyone has any questions, now is the chance. There was someone at the front, but I think he left, so he doesn't get asked this question. >> What's your favorite dinosaur? >> Um, I don't know. I I would say the T-Rex that's running around stealing telephones. That's That's my favorite. >> Uh, if anyone has questions for them, give you the mic. >> kind of

bird do you have? >> Oh, uh, a green cheek. It's a green cheek out here. >> Uh, what was your documentation process during the rebuild? >> We're not We're not allowed to talk >> documentation. So, uh, we had a spreadsheet, uh, on Google Drive that was at the time open to anyone who happened to get the URL. Uh, because we had no infrastructure for password management, it

was just this open spreadsheet that we've marked all the subnets, all the passwords, everything. It was great. Um, once we got our infrastructure up, we rotated all the passwords and credentials and stuff like that, but for getting the basic infrastructure up, it went all onto Google Sheets. It was the only way to get a bunch of random people on their cell phones to be able to enter

>> Now we have a wiki. >> Yes. >> Oh. Uh, let me get you in the mic. >> Oh, okay. >> Um, you said you wrote your own ticketing system. Yes. That because you couldn't find anything suitable. >> So, we used Mantis for years. Mantis is an open source bug tracking software for software, but it worked really well for students because it was very basic and customizable.

But what we needed was the way for some uh for students to make tickets for incidents in the competition. No ticketing software exists for a competition. So, we had to build our own. Um it was a combination of a lot of custom stuff plus Claude helped out and so we built out this whole actual ticketing system that replaced individual separate components like we now have it integrated

with scoring. We have it integrated with stuff. So for example when red team attacks a team they get in and they break the domain controller for example. The ticketing system allows the teams to say, I found this evidence for them breaking into the domain controller and it attaches the information and associates it with the red team activity so that we can give points back. So that's why

we needed to do the custom It's open source though. It's called ticketing system. It's really really good name. >> Any more questions? >> Hi. >> Hi. What's your plans to upgrade and be able to support more teams? We have no plans. There's no affordability. >> I want to hear from him actually. >> RDMA 25 gig switch. Yay. >> You want to talk about that though as far

as what you have to take into consideration for upgrades given you know because this is a fluid >> Yeah. So we right now I'm evaluating what hardware we currently have and what options upgrade given the cheapest path forward and the amount of performance that we can gain from that path. So right now our Dell switch which we do still technically use for our storage network. It has

uh QSFP28 ports. It has like 10 gig uplink ports. It has 25 gig breakout. And so we have the ability to do that. It also supports RDMA. Uh, and there are network cards that are only like I think 30 bucks for Dell Powerages that you can buy that do RDMMA for only like again 30 50 bucks. And those cables uh they're also pretty cheap which is about

like 50 bucks for a breakout cable. So like an overall upgrade for uh over doubling your performance seems pretty much a no-brainer to me. And that's kind of how you have to think about this incremental updates to increase the amount of performance that you're getting. And what can you get to maximize your dollar? >> Don't worry about it. >> So, for context, >> switches. >> How did

I How am I just finding out about this now? >> You were in the group chat. Um, but uh >> But, uh, someone had a question here. >> Yes. Uh, how come you chose to use uh RAID Z2 versus mirrors? Oh, Dex, do you want to just get the headset here? >> Yeah, it was for maximizing the amount of disc space that we had, right? Because whenever

you're talking about ZFS array, you're talking about maximizing your storage volume versus other parameters like random IOPS, for example. And we really also needed that storage volume plus the random IOPS. And so we had to do that calculation. RAID Z2 was enough for that. Um, RAID Z2 is actually a lot for redundancy in terms of an array like this. The reason why we're doing it is because

the Jbods that we were donated with those drives were a little bit older. And so I really wanted to make sure that if something failed, we had a lot of redundancy to lean on. >> We also lost two discs the first event we ran. >> The hotspars came in handy. >> That's why we had hot spares. >> Do I need to bring more? >> He told me.

>> Yeah. Yeah, you have more. He told me hot spirits were not needed. >> Yeah, I made some mistakes. >> All online. I said no. >> The um Yeah, I think that the one thing to learn about this from a helping you know manage a bunch of people is trust the people. Sometimes you get away from that and uh it's been a humbling experience, but we did

make this work. Um does anyone else have any other you had a um a guey that was showing the speeds that you were showing. >> Uh Crystal Disc, I believe. >> Crystal. Okay. All right. I'm not too familiar with all those things, but I I'm interested in that. >> Yep. >> Yep. It's just Crystal Disc Mark. That's an open source piece of software that runs on Windows

for uh benchmarking your discs. This is running through a VM. So, you're getting like the raw VM speeds from that Windows VM. Yeah, we we had to test live. Uh, and for why we used Crystal Disc was because we needed to test live to actually test the VM performance that we would be experiencing. All right, I think we're just about time. Does anyone have any last question?

>> Oh, there's one Davey. >> Um, have we looked into Incisus? Uh, no. I I've heard about it. I've heard about. So, uh, the question was, have you you looked into it's Incas, right? >> Yeah. Um, I heard about that just literally like when you talked about it like a couple weeks ago. Um, and so I never knew about it and so I haven't tried it. But

it's something that seems interesting, but we haven't, you know, had a chance because during the competition season, we basically freeze all of our resources and in a couple weeks this competition season will end and that's when we do our upgrades and our resources, aside from security updates, of course. Okay, I think that's it. Uh, if anyone >> what? >> Oh, okay. Well, we'll check out the >>

uh there's one more question. well, thank you all. Um we will be outside if you have any other questions but uh really appreciate the time and uh hope it was informative. >> The next presentation in this room will happen at 2:30. See you all there. has been around project based learning and a lot of it is around open learning infrastructure and the way I see it is

uh libraries are active labs for co project based appreciate your eager the Test test test. testing. Testing. Is anyone out there? Fant. Oh, I couldn't quite hear that. Let me increase the volume. Pardon me. >> Yes, I can. Thank you. Yes. So, so when you're ready, you can turn it over to me and I'll share and then I'll stop sharing when I'm done. Oh my gosh. Well,

you know what? Um, you don't have to do be so kind because Oh my gosh, that's a lot. So, thank you and I I hope it's a wonderful conference. Scary, huh? >> I'll share and I'll stop sharing. Hello Who's this? Fantastic. Good to see you. Thanks for navigating this. I appreciate I can talk all day long. Who's that Aario guy there? What's he doing? Well, Patrick, I'm

I'm curious how I like him. >> You're doing You're doing great. Thank I know all these Zoom features to zoom in on people is fun. Thank you, Stephanie. And I'll Perfect. Thank you. Perfect. I'll be brief and then we can have your questions at the end and they'll be wonderful. Hello everyone. Uh welcome to scale 23 conference. Um just a kind of a quick side note, your

safety and concern and comfort is important to us. If you feel safe at any time, feels please feel free to call our safety number posted around the venue. Uh I would like to take this time to introduce our speaker right now. Um this is Stephanie Liji. >> Good. >> And then on Zoom we have Serial Overlander. Um please give it up for them. And yeah. >> All

right. Thanks everybody. Um yes. Uh unfortunately I am by myself physically here although we did get Sarah um on on Zoom here too. So, he will actually be doing his slides live and able to be part of the Q&A. Originally, we thought I might have to use a recording, but we were able to get get it all to work. So, thank you to Dominic and his uh

the rest of the AV team, especially Carlos, who helped me make sure it was all going to work. So, um thanks for showing up. Uh this is I think probably one of the first times there has been a discussion about how to link how to work collaboratively in open source between at this level between CSU and UC. There are definitely collaborative efforts that are actually happening that

are open source projects. I'll mention Jupiter in a second. Um, but I think we this kind this started to some extent uh based on an event that Appiio uh uh put together in uh November and Sarah and I started talking about the idea of potentially collaborating and bringing the two largest systems in California together to really uh start uh innovating at at higher speeds and at you

know at better capacity uh within the UC the UC sorry within the the entire state. So the idea was that using your public universities not just as education formats which is an important part of our mandate but also to um push innovation um uh within within the So yeah quick who we are. I'm Stephanie Liji. I am the executive director of the UC Santa Cruz open source

program office. Um I also chair what is the burgeoning University of California OSPO network. Um, and I'll talk a little bit more about that later. Um, and then also joining remotely, Zer, I'll let you introduce yourself. Hopefully everybody can hear you. >> Hello everyone. I'm sorry I'm not there, but I'm the dean of the library. My name is Selander, and I like to do a lot of

open-source projects with our wonderful students. Thank you. >> All right. So there is clearly a very strong uh history of open source within C the universities in California both both the CSU and the UC's. I like I said Jupiter is kind of one of the best I think one of the best examples of that but there's other ones that are coming out of uh uh uh California

universities. And when I say coming out, it doesn't mean that they're only from California University, but that there was a big foot footprint from a there a CSU, a Calpaly, uh, UC um, out of it. And and Jupiter of course for me is one of the the the examples I can point to the easiest because there's a Cowpoly St. Louis Abyispo component, a lot of people from

there that worked on it as well as continued maintenance through UC Berkeley um, among other parts of of Jupiter. But and I take uh personal pride in the Seph ad because Seph came out of UC Santa Cruz and um is one of the reasons I have a job and I could explain that later but there was a basically the uh developer of Seph went off got a

bunch of money from Red Hat for for Seth and then gave UC Santa Cruz uh a big chunk of money which helped me create my which helped create my job and the center I work for. So um but so these are really great uh like uh this these and these are the major ones right these are the ones that we all know about uh that were part

of Linux foundation or Apache or but um and these all became very much you know part of a global digital infrastructure that's super important and what zero and I were talking about was what happens if this becomes and this was kind of like happen not happen stance but a little ad hoc let's say but what happens if we do this in an intentional way if the UC's

and the CSUs work together intentionally to increase um increase collaboration and development on really key important So that's where and and the reason why this I think is really important and interesting to look at is that we're pretty big. I mean between the UC's and the Cal States we have uh over three quarters of a million uh threequarters of a million students and that's you know we're

not inclers um you know researchers and faculty which obviously gets us closer to a million and we're we span over a very large ge geographically very large uh uh state with 33 campuses and what we're viewing this as is really a massive untapped potential for coordinated efforts because there's a lot of different things that we do well and we complement each other on. So the UC system

for those of you who are um from California probably you know you you may know more about you know understand um how this the complimentary complimentariness of the of the two systems uh is and they were historically it's hard to really argue this but I mean historically I think it's bears out that they were designed to be complimentary in the way that they are although to some

extent you know some things just kind of happened but the UC system was generally always seen as being more research intensive. Um whereas the CSU system became more teaching focused. Uh but obviously with new with Cal Cal the three Calpalies Humbult, St. Louis Abbyispo and Pomona um we actually see way more of a research uh focus. They're all R1 ones as I recall I think. Yeah. Um

and um so a little bit more like like a tiered system as well, but you know you know other CSUs are like San Diego State is also an R1. So there are there are some so this isn't a 100% you know black white like clear-cut division but there is more of an idea of the CSU campuses being larger and more teaching focused whereas the UC UC campuses

are a little bit more small uh smaller and research intensive. We have a lot of other uh other um so we have 10 campuses so a little bit smaller system but we also have the five national labs the medical schools and law um law schools and the CSU has 23 uh campuses that I would say would be I call more diverse in a number of ways geographically

they are all over the state um as opposed to UC's which are typically in more um urban areas if you I mean I count Merced and Davis as being urban urban but you you know, but you'll find more uh CSU campuses spread out f in further. I mean, even just in the LA county, you have I don't even remember how many, but a good handful of CSUs.

And also demographically speaking, there's a lot more um firstgens in in the CSU campuses. Um and there are and I just it hits more of what California looks like, let's say. Um and there's more of this undergraduate emphasis within the CSU system than within the UC But there are challenges. I mean, it's I liken it when I look at when I'm trying to collaborate within the UC

system. I always say it's our sibling rivalry. Uh because everybody like everybody's like, "Oh, Berkeley's like the the you know, the older brother that everybody has to be like, oh yeah, everybody has to like try and emulate or whatever." But that when we start looking at the differences between the UC and the CSU system, I always kind of think of it's more like family rivalries, like how

you treat your cousins or something. But it would be nice to be more like a little less of a rivalry and little and more of a collaborative effort. So and up think recently because I think of just a number of factors there really has been limited cross system coordination and collaboration and that's I think a big challenge right now. Um and what has happened is relatively fragmented.

So we could probably talk about a few efforts but they're typically ad hoc maybe just around one particular issue. Um and I see that and I think Sarah and I both had this discussion about seeing this as um a missed opportunity. So what could we do if we actually designed collaborative collaborative efforts? And I want to just also state that we are just getting this discussion started

and one of the reasons we wanted to talk here is to get people's inputs either people who are part of the education system within California or people from industry or other parts of open source who have like ideas about what they see as potential areas for collaboration. So I'm going to talk a little bit about what we're doing at UC that we see as kind of one

maybe one of the things we could leverage and then um Cheryl's going to talk a little bit about some of the stuff that he's doing at Calpali Humble. Um so one of the things we're like I mentioned that I'm um helping create a uh a network of um open source program offices within the UC system. Um we got the a little bit of backstory is that um

like I was mentioning Seph earlier in 2015 at UC Santa Cruz we started uh the center for research and open source software. I actually joined in 2016 so it wasn't we back then but in 2016 it was we um at so at 2016 I joined and we the main focus of that was supporting um open source research at UC Santa Cruz specifically. Um, but we started realizing

we were doing a lot of other stuff that wasn't a research center related and really started to look more like what OSOs do, but we didn't really know what an OSPO was because we never heard that term before until about 2020. So about that time in 2022, 2020 to 2022, we started like developing our program a little bit more and became and started working on creating an

OSPO in 2022. the difference between what what I do at Cross and what I do at the OSPO is hard to sometimes differentiate but um for me the OSPO ends up being more go there is still a research focus but it's not specifically like the projects that are under our center it's much broader uh within the campus there's also an education and a number of other um

components there so in 2024 we were lucky enough to get a Sloan gr a Sloan foundation grant uh to help us build do what we were doing and you know at throughout UC and right now we have six campuses um and we're really the first systemwide network of open source program offices um we what we and one of the reasons we wanted to be a network is

the idea that we could leverage collaborate leverage the work we were doing on different campuses in a more effective way um either be tools that we were creating or programs that we're running um we coordinate activities so we're not um so we can like campus different students from different campuses could benefit stuff being done um throughout the system and also we because this is you know is

something that we want people to emulate we're seeing it as something a model for other systems to follow including to some extent our discussion with CSU kind of has looked at this modeling this this um this the the UC effort on this just and just as a quick what is an academic ospo what is an OSPO at a university actually supposed to be doing um so I

I'm not sure how not how much the idea of an OSPO resonates with everybody in this room, but open source program offices are, you know, off have come out of industry have been in industry for over a decade. Um maybe like almost two decades actually if you look at when the first one started and um they're really like centers of gravity for open source um within an

institution. How the ones in academia differ obviously because our our mandates are different than than um unities than than industry. So we also focus a lot of on our research impact. So how do we use how do we leverage open source to um uh to to make sure our research actually you know has a wider impact. Um we're helping our students build better skills. Um and also

we create partner we have a are a great platform for partnerships with industry. We help we help we don't always succeed but we try and help uh um mitigate any tech transfer issues. i.e. like IP issues and the like that's sometimes a real problem with uh partnerships between industry and academia and we also are really pushing policy ad advocacy. If anybody wants to talk about academic licensing

in um or sorry open source licensing academia my colleague from San Diego is sitting right there Carla and she knows a she's the one who's managing a lot of those efforts. So just a little shout out to her and everybody's anybody has questions it's she's the person to go to. Um so these are this is what academic gospel is doing. This is our um oh so our

our UC OSPO network uh is has those as as focus areas. I do want to have a shout out especially for anybody who's up who's coming from up north. We are having a summit that's going to talk about this is the second one we've done um where we bring all of the open source folks. We try and bring as many open source folks from within the University

of California and other parts of California um academic industry open source um to a single conference like a large summit. We have two days uh in Berkeley um April 22nd and 23rd coming up and registration should be open soon and we hopefully will have uh more information of that soon. But I just wanted to do a shout out, especially for those who may be in the area

in in the Northern California area, the Bay Area. So, um, but yeah, uh, if people have questions about that, I'm happy to to talk more about it as well or come after come talk to me afterwards. Um, but so now I'm want to talk about a few of these things because there's a lot of overlap with the discussions about things that we can leverage to improve collaboration

with CSU. And so one of some of the initiatives that we're working on specifically within our network is the discovery of of open source like what are people doing on our campuses? We don't know. So that's a huge push for us to understand and that because without knowing it we aren't able to leverage it. We aren't able to improve the impact of it. Um we also focus

on obviously on education that's one of our mandates. Uh so improving the use of open source and promoting the use of open source within curriculum, but also making sure students are able to work um at on open source platforms in open source um throughout their time at at um at our universities as well as beyond. And uh Sarah will talk a lot about some of the cool

stuff they're doing at CSU in a second. And then sustainability not just for the concept of open source sustainability but also the sustainability of like you know the ospos within within these networks because it's or sorry within these campuses. So we have a lot of discussion about how do we make sure that the what we're doing um especially like the research software and everything that's being created

is is actually being done in a way that's that works within the the needs of the open source community that people are following best practices but also that they understand the value of us supporting those efforts. Um so this is the one thing I I want to talk about and uh just wrap up the part about what UC is doing and then I'll turn it over to

um Surl. the UC open source uh repo browser, our orb is specifically aimed and this is one of the things that we see as something that a tool that we can we share a collaborative tool that we can share um with our CSU partners um and that we discover basically we're using this tool um to discover uh platforms sorry to discover all the the different UC the

different open source that is happening in the UC system. We're also using it for other um other campuses that have academic opos. We're working together with them to kind of broaden the use of this tool. And it taught us a lot. It taught us that wow, we have a lot of stuff out there at 50 like over 52,000 repositories within the 10 UC's. um uh they we

kind of show um you that that where where we need to focus some of our our work and I'll show you a slide in a second which kind of showed us things we needed to work on um and really helps us find those people and enable uh potential collaborators to um and I can get I can go more into detail it's really we're working on the like

we don't have it quite launched to the extent that we can show it publicly but it does get it is, you know, you can actually it's it's like what it sounds like. It's basically you search for the type the topic uh the campus you want to work with like the the different topics and different technologies being used and it's a it's a really cool uh it hopefully

by next like very soon we'll be able to actually say hey this is what it looks like and do a little demo on it. Um what has been used useful for for us and we we think would be useful for for other campuses other and other systems is really understanding um what the pro what what we need to help our uh our our stakeholders with uh what

we learned for sure was that we needed to do a lot more education and outreach for those who are working in open source at at the UC's. Um we were I guess we weren't too surprised but it was a little shocking to see that over 60% of uh the repos that we found lacked a license in any type of license at all. Yeah. So anyways, we've only

been doing this for a couple years, so that's not our fault, but it but it did give us a thing of like, okay, what do what what are the things we need to focus on? And then a and did something as simple as contributor guidelines. Well, how are you going to build a community without that? So um these are things that we started to realize and started

focusing uh um what we needed to do, but it also focused that we needed to figure out ways of systematically discovering at scale. So because and how do you continually do this so that you can always have updated information. That's what we're working on now. But we are seeing replication beyond UC. And I think this is one of like I said one of the collaboration points for

CSU and UC um is that this is becoming a tool that we can that other universities can use. And we were discussing uh specific meth and discussing and publishing specific methodologies for multicampus uh use of this uh this um for this tool. And we're publishing everything we do. So people can defin it's definitely replicable. Um We're trying to create a multiplier effect really because we're we're seeing

how other universities are are using this. Um, and now I'm going to stop sharing my screen and I'm going to give it over to Sarah. >> Thank you, Stephanie. And I really appreciate you and everybody's willingness to allow me to do this virtually since I just came back from Austin. Um, the way I see it is we have such tremendous opportunities to work with students on open

source. not only sharing what it is but what's possible and I'll start with a little bit about us. Basically, I see a library as infrastructure for open education or open learning. And oftentimes, I think of it as a laboratory for both curricular and co-curricular innovative experiences, whether it's our work with software engineering or it's with engineering students building things or career skills, digital humanities and our supportive

hackathons. So the library is is a key component in an open learning infrastructure and part of that is because we're really focused on a lot of open pedigogy frameworks. Now we usually see open pedagogy with co-authoring of articles or papers or books and that's a wonderful technique. I'm I'm really appreciative of it. The one I focus on on this talk is really about project-based learning and how

that is an incredibly launching uh kind of transformation of what the student experiences but also how the institution itself changes as a result of that engagement of students. So whether it's integrated in the curriculum or it's a co-curricular activity, I'm going to explain some of the work we do with our biology program, our computer science engineering and more to come where we sponsor real world challenges with

our students in the classes and they prototype and if they're a successful team that's wonderful. We can also offer them employment following their class and fully develop their idea out to version one or we sponsor paid internships or student employment for very specific ones or hackathons. Here's an example. Dr. Shireen Bogle is an amazing computer science professor. Always ask me for ideas about, hey, what challenges would

you like to have our students work on? And so in 2022, we got a digital dissection table. And that's great. Five humans, 120 animals. Weird to see dead people in the library, I know, but it doesn't do plants. So I gave this challenge in the fall of of with these resources. Build us something like this for plants. And in 12 weeks, the student team Flora built a

prototype that looked very good. So what do I do? I hire the students or in this case AJ Beum, a brilliant talented uh programmer to build out the version one with student assistants. Bodney students, computer science students pro uh developing the techniques of how do you do photoggramometry of plants. This is the world's first three-dimensional digital arbarium because nobody wants to make 3D models of plant leaves.

They're way too thin. Now, what happened? We released it as open-source version one, and it's back to the classroom again because now the biology students want to use it for the vertebrate museum, and they're starting to annotate. One of our goals is all students uh publish before they graduate. Now this has been a very common thing that we do. We've been doing it for years. This is

an example of in 2018 2019 we asked students to give us as a software to count where people sit in the library. And so it was a phenomenal piece of software that called space use. And we wrote about it in the Chronicle of Higher Ed. We went and presented it at the American Library Association. And I even arranged a vendor booth that they could answer questions about

their software. All because in many ways this is a transformational moment for the students. Do they want to start a business or if they don't I have the open-source software that I need to solve my real world issue which is I want a good analysis of where people sit to transform the library. This has been used also for how do we assess learning? I'm sure everyone knows

that it's one of the most arduous processes where you ask faculty and librarians to read a bunch of student papers and do an assessment that follows some rubric. Well, you can automate a lot of that work and make it easier for faculty to do that kind of work. The work was intended to start with automating the process and then backwards engineer it to help students with research

and writing. But then COVID stopped us. But we're not stopping. We keep doing these developments. We work with uh supporting student hackathons and we also recently challenged them to connect I naturalist one of those citizen science projects that's got 8 million some odd folks participating in identifying biodiversity across the world and connecting it with our digital commons trying to create a new way to engage academic content

and make it inspiring. Our internships, which are paid internships, allow us to actually encourage students explore a part of their interest. In this case, John Gerving looked at 3D modeling and helped us develop a 3D model of our library that was interactive. This is something that they could do in a way that supported the curriculum, but it was a co-curricular experience that really helped launch this student.

And um that's what we try to do quite a bit of. Here is our um GIS data archive project where I'm working with Dr. Jim Graham and we hire students to work on how do you stitch aerial photographs together, geo reference them, layer them with other GIS projects and make GIS projects, historical maps, aerial photographs all scholarly citable because so much of GIS is lost every year

because they're kept only on hard drives or servers and not really publicly accessible. So, we list all of our projects and because we want to make sure that not only do we see the impact of their work, but we also have a way for the students to communicate the impact of their internship to prospective employers or the uh future grad programs. We also have an innovation hub

where we do a lot of these kinds of software development and uh we do the open-source projects for anything that we pay uh students a stipened or hourly rate. If you're ever on the third floor of the library and you see about uh little flashes of light, that's Hunter making two to 400 photographs of one specimen to make it one 3D model for the herbarium. There's a

lot we can do with our student employees. This spring, we're we're testing out how to work with 14 library scholar interns and nine engineering student teams. They're each building different kinds of solutions to problems that we are encountering in the world of academics. One is how do we engage books? Another is how do we geo reference tens of thousands of aerial photographs and make it something that's

more automatable. These are solution making opportunities that we think are really important for our future and how we envision a collaboration between the UC's and CSU. At the AACNU meeting, uh, the forum on digital innovation, I shared an activity for folks where I said, "Brainstorm the challenges that you have that you wish you could have the talent and time to focus and work on." And so on

the back side of it, there's an example of the ones that we've given our students. And on the front side, it's a way of us brainstorming what are the challenges that we're encountering with uh implementing aic AI or geo referencing at scale or making learning more exciting these days with more of our students all coming from the kh economy So, I think we're gonna briefly touch on

some questions and get back to um Stephanie and um a a opportunity to really think through, but I'm going to share that there are lots of ways we can use real world challenges as project-based learning that enhances the curricular or co-curricular experience of our students. And the question we are thinking about is how to create the infrastructure and planning in order to do it. Stephanie and I

have experience with opportunity hack and code day, but we're looking for what is the infrastructure to create opportunities across large systems like we have. And then who's interested in helping us develop these ideas? I hope they'll please contact us because we can always say in the open source community it is building solutions together. I thank you for your time and please um let me know if you

have questions and I will say um Stephanie please take it away. back to my slides real quick. And so yeah, we probably want to touch back on those questions unless people have some ideas right now. Um we I'll just go back to we have some questions at the end too. Why is this not working? This was doing so well. There we go. So, I think I just

wanted to highlight one of the ideas that that that Sarah brought up and that I think is a really good start and we kind of saw that as our first kind of bold idea. Let's do something that's big but also something that could have some real world impact. Um, and we could use some of the idea, some of the things that sorl already had listed um, as

real world or maybe get ideas from from the larger group, the community at large here and elsewhere of like what type of topics we could really be working on. And so I wanted to point out this idea of this joint hackathon. Um, and what the history behind some of this is is that we're uh currently this uh the Reboot the Earth hackathon is a large one that

the UN actually puts on and um it work it uh works with different campuses all over the world on it. It had the we did the first one at UC Santa Cruz uh the first one on the west coast uh of the US at UC Santa Cruz in November. This next one will be in San Diego uh at UC San Diego. Um and at that point we

did invite all of our CSU like the regional CSUs but we didn't collaborate with them. So one idea may be if we move forward and this will probably be a recurring type of uh hackathon happening over multiple years. Um so as we maybe do another one in the fall um this could be a great start to that. And what's nice about that one, it's very much a

climate change, which of course we're looking for things that no not just impact maybe individual communities, but things that like much more state focused. So I think one of the ways to also look at this is not just that it's oh you see and CSU working together. It's these you know powerhouses within the our state within California working together to deal with state focus specific solutions and

spec specific challenges within our state. And one of the reasons and I'll talk a little bit about this towards the end is that the idea is making it not just oh this is something coming out from the education system but the you know the public administration the p you know the the state government state agencies looking to open source um and that the the the engines behind

that um kind of change that shift in how those agencies look at look at open source is coming from the universities. So one of the ideas we've had for dealing with a lot of the collaboration um sort of the hackathon as a start uh but also these solutions all of these are things that you know every part of the state has some level of challenge on I

mean I don't think any of us could argue that urban planning and you know it home housing urban planning and homelessness isn't a problem in a vast majority of the state definitely climate change agriculture um and another big one we're talking about nowadays more is education and workforce development how do we make sure our universities but also our all of our education system from K through 12,

community colleges and the like are actually like able to uh meet the needs of of of uh future uh kind of the future innovations, the future technologies. Um so one of this idea that we were we were having as well is this idea of if we were able to really push and work on a good collaboration between CSU between UC's and then and maybe bring in other

parts of uh California education public education system is to have more of a statewide open source alliance so that it wouldn't just be the universities working together but we would be collaborating say with the agencies in um in um other parts of the the government because we are obviously government we are actually technically California state employees um as parts of these these systems I liken this a

little bit to what's going on in the EU this is actually one of the reason why I kind of was inspired by what inspired me with this is that there's a lot there is definitely a push for ospos within governments within the EU. I can see California and especially in the its current kind of state and it's the current situation and kind of political climate that there

is out there taking more of a lead from um from what the use the EU is doing. Um but I'm I'm not that's and I'm not trying to sugarcoat it. I know that the EU has its a lot of lot of issues, a lot of bureaucratic nightmare issues, but um I think that their outlook and their their policies with regards to open source is something for us

to emulate. And I see that especially create the creation of more of this open source alliance where concerns about digital sovereignty, dig uh uh open source being the the foundation of infrastructure. Um even just the simple things of like agencies not having a vendor lock in that we all kind of are concerned about. um I think has been is something that could move this this issue along.

All right. So just kind of I think wrapping up I think we have a few minutes for Q&A soon. Um so why are the benefits of taking this approach? Um I think it's I think it's pretty obvious. I think this I don't think I have to uh explain to the crow this crowd uh the importance of open source um in all sectors but I I it is

kind of good to delineate how in these specific sectors it might be it's I think particularly useful for students there are students CSU UC any of part of you know UC the California public education system the idea that they're able to build real world experience through open source I think it you know it just is so valuable I working with students who get their are able to

build their portfolio by working in open source even though it's I think we're you know it's challenging right now we had the whole discussion if anybody was at Stormy's talk about the increase in the number of like like PRs and the like and it's actually to some extent AI is actually maybe making it harder for for people to get into o open source um I think that's

going to balance out in the end but it is still very very helpful for building your portfolio and having opportunities um within the state itself. I think the access to the research that's being created and the having more open access to the research is beca being created within their own research institutions within their own uh universities is important and and not just act their not just state

access but all communities access. the ability to have custom solutions being created internally within our within the state um by using open source platforms um as well as the obvious potential reduc reduction in costs um that um remain working with open source can can can occur and with the state cal with the budget situation like it is I think that that's actually a really important thing to

kind of why state legislators and the like would be more interested and kind of understand the more it's a really and for that it's really about getting starting to get more champions within within the state uh within the legislatores within the agencies and for academics so talking from you know our where we come from um research visibility is always very important and and the impact of your

research uh through open source is something that's I think um most that's a growing understanding not I don't think it's all there but it's a growing understanding within many academic institutions and also in ways of increing uh partnerships using open source platforms to to work more more closely with industry and get industry uh pathways or industry partnership like you know and sponsorship and then the innovation pathways

that are open to us with with open source um that is not there through through more um typical tech transfer uh you know typical proprietary means. So what we're seeing really is this matchmaking piece between our systems and then with external partners um between UC our research CSU's teaching and also um and innovative teaching ways of both sides um and then working with industry and open source

communities really could bring in joint um like very you know impactful joint uh uh impactful uh collaboration. we can be looking at stuff like working on joint projects and having other collaborative collaborative Um, so for short term, what we're looking at, again, this is still kind of just an idea that we're forming and we're trying to help move forward within starting within this the two systems that

we've been talking about, but then broadening it out to the to the greater the state in g the larger larger uh state of California. um is for the short term is some of the things we've talked about this inventory over the OS projects trying to see what the landscape is of open source right now within the UC and CSUs um making it clear who we are we're

reaching out to to try and get interests and collaboration um and then starting this outreach to state agencies and legislators about kind of broadening this and making it more of a state thing um I've seen that actually especially when discussions about uh California policies on uh software security and with um AI that really we need to start as an open source community starting figuring out who our

champions are supposed to be uh in Sacramento and I think that that's that this would actually be a great another method for doing that is kind of talking to them about um bringing in the education side and I think that that would actually be a very good draw uh and also help us identify who those people are that are going to help move this uh move these

policies forward. and later term I can see we as we just be creating a real C UC at CSU collaborative framework um that uh that I think would um kind of be something that's a template as opposed to kind of the ad hoc way that it that it happens right now. Um I think joint industry partnerships would be really really uh impactful because one of the things

that that we with the UC network that we're hearing from industry is that wow it's very helpful to be able to kind of go to one entity to talk about potential collaboration. Um because then it's e instead of like having to go to all 10 campuses and figuring out who we work with. Um I can see that even being you know more having more impact if we're

doing it at the 33 campus level. Um and also I mean I you know our major and most important part of both UC and CSUs is really teaching and education and um and making teaching people to learn to to understand how to learn on their own. I think we the last talk talked about that or couple talks ago talked about how the point for your university is

to teach people to know how to think. Not teach people how to think but teach people to think kind of thing like teaching people how to learn on their own. Um and I think that doing having integrated curriculum development between UC and CSUs which have kind of different outlooks but really complement each other I think would be really impact really really helpful and really great. Um so

but we do need like we need to be able to have these shared platforms we need to build on the o all open source initiatives that that have already occurred both the like the network that I'm working on all of the cool um uh education and curriculum building um that that's happening at CSU and Kpali in particular um and then this coordinated industry uh partnerships I think

would be very very kind of the things that we need kind of the um things that we need to help get this kick started. And what else needs to happen? Well, again, like I said, state level recognition of open source as a development strategy economically um I think is again looking for champions to to um to move that along. Uh funding of the coordinated infrastructure hopefully that

could be coming from the state if they're recognizing the impact the economic impact of open source and then bringing in industry partnerships and engagement. How people can help. anybody in the room? Um, and I'll like let uh Sarah also jump in on this one too if he has wants to add to anything because I think this is yeah last slide before discussion. Um I think for industry

partners just be showing that there's a possibility of engagement with our campuses either as a large in at a whole or even individually for now um really makes a difference for uh those of us trying to work in open source and in in universities uh for faculty really start proposing and thinking through what initiatives would would benefit from being being collaborative across multiple across the system or

at least and across sectors. and across campuses. And then for the communities, just the community at large as well as the open source community really say, "Hey, if this stuff is happening at public universities, there should be it should be publicly available." That helps really spurn the idea of openness within within um the universities and then contributing back. You know, if we have open s if we

have good academic uh open source projects, having uh open source community members who are outside of academia, uh contributing back in really has it creates this great virtual cycle where people understand that the communities that they're they're working with and um collaborating with. So people may have questions um and then but I'm we also kind of wanted to to kind of throw out questions to the group

because again this is also a learning process for Sariel and I. Um so I'm going to let Serial have the last word though in case there's anything else you wanted to say before we move to the to these questions. >> Yeah. I'm just going to make it really short. I'm just going to say it's so important for us to give students that experience of transformation and making

the the solutions to the real world. Our challenge in higher education is we all know very significant right now and we have to be extraordinarily creative and so the only way we are going to be able to scale some of the innovation that's required these days is to actually actively involve our students and it's that framework that's so exciting as I said I'm experimenting right now with

some larger scale approaches at internships and curricular uh student challenges and I see that it can have a huge impact and that's partly what I'd like to explore with you all is that do we see it as a possible California thing that we can um foster a a shared either hackathon or an engineering event a lot of times it's about just creating real world solutions and if

there's one thing that really students want to do is make a difference in not only their life but the world. So, how do we do this together? I really appreciate hearing from you all. >> Great. All right. Thank you. So, I think we can get these are kind of just for people to thought. And I think actually the question that Seriel had was actually a really good

kind of start to if people have thoughts on that, but um so I'll just say open it up for questions or comments on either what Seriel had to say or what's up here. Um, for anyone that has questions, please raise your hand and I will hand >> Okay, I'm gonna turn this off so we can all see. >> Hi. Um, I'm at one of the UC's in

the research and economic development department. Um [clears throat] I've seen two students from two different UC's recently who came for entrepreneurship support one of them was not guided at all on and he was like dreaming of having an open source project but also expects to make a living on it and he's been coding for a year on it. >> So I was worried he was going to

like use a license that was too permissive. And then I had another couple of students who were asked to sign away their IP in order to be allowed to participate in a campus accelerator. And I read the uh contract and it was from my department. Um and so I'm just curious like do are you running into issues with uh the people on campus who are trying to

prod produce as much private IP as possible and what's your strategy around that and can your office be a resource for me to send students to who need somebody to look at a document and say hey maybe you have another option you could open source this or this is what would happen if you did I would say to the last question yes we are a resource I

don't know if it's specifically to give a hard answer on that because we do have to respect our TT our tech transfer fs we who we are trying I can't say it's 100% every all 10 campuses but a lot of the campuses do have good relationships with our TTO um our tech transfer offices uh so for that question I would say yeah the um I'm not sure

are you talking you said they were from different different campuses >> there's like one from one UC and then a pair from another >> yeah talk if you yeah if you want to talk actually Carla would be a good person. The two three of us should talk after and we could probably go into more detail about like who at different campuses to talk to. Um I know

that I have my experience with um with uh working with my innovation hub folks and with my tech transfer office has been very positive recognizing the concerns about IP with regards to people who are developing open source students. Yeah, I think the issue is because with stu within UC and correct me if I'm wrong Parla uh people within you students within UC typically own their own IP

student not graduate students undergraduates as soon as they start becoming like like staff which is like a graduate student would be considered there is that discussion about the IP technically is you know under the uh the university so I'm not sure if that's why the accelerator students got that way we could talk about that later but I think it is something important to recognize that even when

it's the UC at least under the UC system even when it is seen as IP of the university because they paid for it in some way be it from a position and that's normally the that that situation um there is still a pathway for open for open source um and a pathway for a student still being able to like move on move to the next step with

that open source that like you could if he that that student wants to if they want to open source the product and then build upon that that pathway is available. It's just that you know I think it depends on each campus right now. I think one of the efforts we're trying to make is making it more uh a systemwide approach that people can feel more confident knowing

that there are like specific ways of of uh handling you know uh different situations like you know you it sounded like you had two very different systems and so I would like I mean sorry different examples so it wouldn't be like everybody has the same example there's no cookie cutter approach to it but at least people feel a little bit more confident that's That's that is one

of the things that we are um doing. I I know Carly wants to talk but I would let Sarah just talk about CSU first. Yeah, >> I'll just briefly share that um whenever we pay for internships or student assistance um technically then CSU could own the um IP. However, because it's a requirement for the library that this project be open source and commercializable license, MIT license, it's

actually freeing the students to commercialize it after they want like branching it off, but that's on their own time. And when they're working with us, it's an open-source project and it results in an open-source um product. And I think that's super important because I explain it to students that this is actually in your interest because if you decide you want to commercialize it, you can. But there

exists the open source. Your only leverage is you are the first to the market. You understand the market. But um that's how we do it here. >> And Carla, why don't you introduce yourself to? >> Um hi, I'm Carla Padilla. I'm over at the UC San Diego um open source program office. Um so I've been working a little bit with the TTO's on some of these problems

or issues. And what I want to say is that the um UC regents which kind of you know sits above all the UC's I consider it like the federal government and then each campus is its own little state. Um so they have policies about copyright and who owns what and there is a website um you know just Google Google you U you U you U you U

you U you U you U you U you U you U you um you U you U you UOP um um copyright and it'll tell bring you to a page where it kind of outlines who owns what and under what conditions um you know like generally like Stephanie was saying if you're doing something as a staff member if you do something as a research project because a

lot of research grants will specify who owns what right sometimes you know it's not owned by the students it might not even be owned by the university it could be owned by you know whoever um um providing the grant. Um sometimes with like what students do on their own you Yeah. generally you know undergrads do own it except that if they use significant um significant um university

resources now how that's defined and you know what is that line can be a little bit blurry and that's really where tech transfer office or your office of innovation or commercialization whatever it's called on your campus would um be the ones to weigh in on this once you know so we can help guide you to kind of those resources we can also to you about the different

licenses, you know, what licenses are kind of favored by the UC. If UC has some of the ownership but they're going allowing open source or if like you know as a student and you completely own it, we can help you say like these are the different licenses, here's what they do, but we have to be careful about not providing actual um legal advice. But, you know, there

is um opportunity for us to help provide that guidance that I think you're looking might be looking for. >> Yeah, just to reiterate, I I feel like the one of the big pushes for within the OSPO effort, not just at UC, but at all the other there there's like almost 30 campuses around the US that have that are starting to have these these offices and their campuses.

and um they I think that's one of the goals that we have is to be that that bridge so that researchers don't feel like they feel like they can have these discussions with TTO without with the tech transfer without feeling like they're giving up their you know whatever giving up the their their their work that they've done. So I think that that's that is part of our

role. And did you have another question or did we get everything? Okay. Yeah. Uh we have roughly um less than 10 more minutes of this session but we do have still time for questions. >> Uh I have a quick question about uh what kinds of uh discussions have you had with developing policy for when it comes to what what exactly should be open source versus what should

what what is defined as IP. I think like the library aspect of looking at especially with academia like there are some things that we should all know and and that information should be should be free right libre uh in a sense of and I know that there's all sorts of um different policies that have been written for for the state of California in the sense of you

know looking for small businesses first and doing all these other things. So is there any discussion in developing policy at either the the uh university governance level or at the state level to say like we should be defaulting to open um and and what kinds of information should be retained as IP versus should be retained but via open source uh licensing because like you know my government

employee like people can ask for my test text messages on my government phone like why is the script that I write any different than that? Um so yeah just curious if you've looked into that at all or >> I I mean if it's like I I mentioned the the the the EU model I think is an interesting like starting point for that within California. My understanding is

that I don't the problem is I don't that there is actually a default to open but it's never been enforced and though that's a big problem that we are trying to recognize and how to deal with that's actually one of the things that I feel like is at least we have some like hey look there's actually I think laws on the books that you're supposed to do

that um but because of the the comp because of complicating factors and lack of education on how to actually do that um I mean education at the broader not just within the universities but within like the state system Um I don't I think that that's like one of those areas that we could especially with collaboration between the two two the two big systems education systems uh university

systems that that would be actually a very good start kind of like showing this is what open source is you just haven't understood it till now um for the policy at UKOP I think that part of the the efforts that we're doing uh within the like so the discussions with uh UKOP legal um have touched on it but it has I feel like that has been a

campus to campus discussion because everybody knows the technologies out of their campuses the best. So I think that there are things software seems to get a little bit of a free ride when it not free ride that's probably too hard too easy of a word to say but it's more like it is seemed like oh yeah patenting cop patenting or copyrightiting software does seem to be maybe

not the best way of making sure it gets used and implemented. So I think that there's a slow understanding that the that software is is unique. Um that said that doesn't help with open hardware aspects and I think that those are that's still an issue that nobody's gotten their head wrapped around within within UC about how exactly to make those determinations. I do feel like at least

within my interactions with tech transfer at UC is that they do want to lean towards what the inventor, what the person who developed it wants to do with it as the kind of if you know they would like to respect that within the first like because they want to have these people continue to do stuff and not and not like uh squaltch any like other research that

they might be coming up with. So um I right now I don't that there I don't feel like there is a single policy push or discussion at the moment. I do think that there's discussions that you see there are there's some sort of discussion in the California state system but like the sorry the government but I would be I don't don't think there's not a single effort

right now as far as I know serial do you have any thoughts on >> if I understood it well I I would say the policy for procurement is a nightmare at most universities and I'll just share this [snorts] >> um they often want to get the commercial um option because the free is very confusing in our systems and yet so I would just share that it sounds

like one of the education things that we all need to do is really make sure that the procurement folks know how incredibly important open source is and encourage that they um streamline the process because if it has to fit how a subscription to Microsoft service or other services works it's a nightmare there. And the reality with open source is it's a great opportunity to be creative and

innovative and to really empower our students in a way that transforms us. And if we have to make it fit the commercial model, it's really difficult. Imagine what open access looks like um without a mandate. I prefer not to mandate open access because it should be exciting to publish in the most impactful way as well as make a transformational software like open source. Thank you. >> Yeah,

thank you for bringing up procurement. I actually wasn't thinking procurement. I was thinking the other way around. So yes, that's exactly what he says. >> Okay. Uh hi. So thank you so much for your presentation. uh my question kind of a little bit um from a different angle of the same story bridging the gap uh everybody talking I will be talking about bridging the gap between open

source and academia as well so what I'm interested in uh so you talked about like empowering facilitating like students researchers and uh basically like all you are doing for the student body so what I'm curious about uh what are your relationships with faculty do you have any influence on curriculum? Uh do they come because from for me as a founder of nonforprofit one of the goals is

to bridge this gap. I find actually faculty being like the most inertious part of any educational system. Uh and uh I'm still struggling to kind of communicate that probably that's something they need to pay attention. Uh so what is your experience in this regard? >> Sir, did you hear did you hear the >> It was well how how do we what's our experience with getting faculty to

adopt open source practices and and platforms and the the person questioning was specifically saying that her re her experience has been faculty has been the hardest kind of the hardest stuff to craft the hardest problems. Um what what do you feel that you have because I have my own opinions on that too. So, but I'll let Sarah go first. >> I I'll start with saying um it's

one of the greatest things that we have to do is engage our faculty in a way that's collaborative with students. Faculty care about their students. So, by connecting an open-source project to the students, it actually helps our relationship with the faculty. And I noticed that really strongly with the GIS project with Jim Graham, Dr. Jim Graham. And I noticed that throughout my career at libraries is that

um you can't always engage faculty in an open access textbook or this other thing, but connecting it to their students and their care for the students really is a good um relationship building opportunity. And you don't need to get everybody. you you really just want to encourage that there's enough that model the behavior that we all need to be practicing more of. Um, and I used that

example of uh just in Texas the other day that with AI, a lot of our question and answer assignments no longer have the same I'll say um impact or value. we really do have to shift to more project-based learning, but it's not something we should say, hey faculty, you've got to change your assignments. It's like, no, collectively, we've got to work together on something that's really exciting

because higher ed should be exciting. It isn't no longer an ivory tower. It has to be something different. That's my piece, Steve. >> I think we're out of time, but my last Yeah, we're out of time, right? Yeah. Yeah. Uh I I do want to say that there is also a push just yeah everything he just said plus there is a push I know within the UC

to actually link more open scholarship to tenure and promotion and I think that's also something and not not just UC like there's broader efforts as well that we could talk maybe a little bit more about but um but yeah cool um sorry everybody else have other questions I I will hang out for a little bit. Um, I'm not sure if Serial's gonna hang out online for much,

but I'll probably take you off the screen and just we keep the Zoom open, but thank you everybody so much for this. It's been a really good conversation. I really appreciate it. Thanks for for >> Byebye. >> Bye. [applause] talk will be around. >> Yeah, later. Should say later. Thank I can hear myself talking. Okay. >> Uh but uh >> are you okay with the mic? Okay.

Okay. >> Hello. >> Afternoon everyone. We are about to begin our next talk for this track. Um I'd like to introduce Aini. >> Ai. >> Abne. I'm so sorry. That's um Adam. Give him a round of applause for this. >> Hi all. Thanks for joining. Um I'm Kudri and welcome to our talk. There's a small group of you so come on up. I I'd love for it

to be more interactive. Uh since it's a small group, I think we can make this more of a conversation. Uh and would love to really hear from you all what you all think. Uh as um I don't know his name, but that's okay. Um I'm like one thing I do for work is I work at GitHub and we um really want to help educators figure out how

to use AI in the classroom. uh and we're exploring and in this talk I'd like to explore at least how uh we can help teachers help each other uh with learning about and teaching with and about AI. Uh and here is Adam. Adam is this is Adam Holt. He's the founder and lead of Internet in a Box. and we're both very passionate about working with learners and

educators to make coding and learning content available around the globe uh and make kids dreams come true. So, uh before we start, I think we're such a small group we could probably even like uh I don't know you want to introduce yourself to your neighbors. Do you guys know each other already? No. Okay. Um anybody want to share where they're from? Anybody who's an educator here? We'll

start with that. Awesome. What what what do the rest of you do? >> Awesome. Thanks for encouraging participation. Um for the folks who are teachers, are you uh university professors, college professors, high school? >> Community college. >> Amazing. Which one? >> Oh, awesome. I graduate. I'm a SMC alum. So, yeah. Cool. Awesome. Adam, would you like to introduce yourself? >> Not really needed. Uh, but uh, a

couple of us are from Boston and help tiny schools around the world who may not have uh, even community college budgets. Let's just leave it at that. Yeah. >> Yeah. We've worked We're going to hear more about um, Internet in a Box and the and the work that we do there. So, I'm work with GitHub. uh I lead education for GitHub. So uh but I work on

Internet in a Box when I can. So very excited to have Adam here with me. Um so a large like I I think I want to talk a little bit about the challenges educators face. So a large part of my work at GitHub is about supporting students on their journey to becoming developers. And to do that, I am speaking with educators to understand their perspectives on software

development and your use of AI tools such as GitHub copilot in the classroom and the challenges like you face. So consistently I I have heard I would I was just at SIGY uh in St. Louis two weeks ago and this this is my third SIGY. Y does everyone here know what SIGY is? The special interest group for CS educators. Um it's an ACM special interest group where

there's um this year I think there were 1400 educators there uh talking about a lot of talking um a lot of talks were about AI and using AI and one thing I and I've heard and the conversation has changed over the last three years since I've been there. Um initially there was a lot of skepticism about AI and lots of concerns about cheating um how students will

cheat with AI and the now I think there's an acceptance right we're in the acceptance phase where I think educators understand that AI is here to stay and they want to figure out how to use it effectively and we had a workshop there where we were actually showing educators how they can accelerate their own work by using AI in the classroom. So curious do for uh who

uses AI in the classroom here? Are you teaching with AI? >> Do you do you prepare any curriculum or accelerate your work with AI at all? >> Okay, I know you're use AI to do your homework. You've already helper Yeah. And I I think the I think what I've heard is like AI can help with the simple debugging and take away some of the tactical like work

that if you're a computer science professor like a lot of initial work is I think it's important to know how to debug but like knowing the syntax differences between Python and Ruby. Not sure how important that is, right? Um and I think leveraging AI for that is great but the problem solving and the critical thinking skills is something that we still want to cultivate, right? Um so

I think teachers are expected to learn and teach like before you've always been asked to teach basic computing software development version control and now AI and there's escalating expectations because the pace of technology is just rapidly increasing right like um last year you know we were talking about uh essentially agents were coming coming to existence now it's all about agentic workflows And do you know agentic AI?

Yes. Go ahead. How many years? >> Python. I'm curious where you're >> with AI browser. I guess different Yeah. I think so you're talking about two different things, right? It's like it's Yes. Yes. So I I think it's actually both. I think teachers need to understand how their students are using AI um regardless of the instruction they receive, right? Because um I think it's over half or

nearly half of students use AI in their workflows. And I have a a data on that. But I think that there's also it I think teachers understanding how their students are you using AI but then I think there are courses that exist today like I don't know where where do you teach at >> Sunni okay I don't I don't well so I know like um at UCSD

uh Leo Porter has a course on his CS1 course on Python right it's it's now um I think it's called pi like learning Python with co-pilot right so I think it's it's uh very I I think I'm seeing teachers actually intentionally introduce AI into the classroom. Um CS50 at Harvard has developed their own AI debugger tool, right? Like so basically providing guardrails so that it's a more

Socratic uh AI uh bot instead of one that just gives you answers all the time. So I do see intentional uses at um Oops, sorry I do see intentional uses uh from educators. Um so and I think I'm talking about both. I think part of it is that because curriculum as you're saying right like curriculum standards take years to develop and if I look at the pace

of technology just the last five years uh I don't think teachers I believe and I'm okay so I'm not a teacher so I'm coming from a very non-teer perspective here but I believe that teachers need to I think my hypothesis is teachers can learn from each other and accelerate each other share uh shared knowledge but also share artifacts right and the artifacts are changing right it used

to be the artifacts were documentation um assignments uh exam materials right those were things that teachers could share photocopy share over email now I think the artifacts are prompts um can be uh skills can be specific workflows right that that teachers can share with each other of how how they're accelerating their own work And then I think there's a thought leadership piece where you want to understand

how students are using it so you can design your courses effectively. I think it's I think what I'm saying is teacher oh so I know teacher people ask teachers to do a lot already and I'm saying teachers are now being asked to do even more but I think you can help each other to do so. That's the hypothesis. So no, they do. >> Oh yes. Yes. Yes.

Harvard teaching AI. Okay. No, totally get it. Push the push the I didn't hear the last part though. Push. >> So, how do how do we encourage more of that? I'm just trying to understand. >> No, it's great. I think it's I think your points are valid and I'm sure you're asking things that are on other people's minds like um how do we push how do we

accelerate the sharing, right? The collaboration. Um what do you all think? I have ideas up here and I'll give you my ideas, but I'd love to hear science and what we're finding in the So, we've got Hey, where am I? I think what we're already knows involves AI. What are some of those in there? that you're teaching computer science students and they're kind of working You're kind

of the loss of Does that partially something betterul for those learn I I love those. Yeah. Go ahead, Steve. But the problem is you have you have to have the experience to understand it enough to go that's wrong. >> That's that's right. And I think educators are kind what I've heard is that educators are adjusting how they test knowledge. Right. So to your point have they absorbed

or have they learned the material that the class objectives had. I think educators what I've heard at least is like educators will modify the exams themselves like for example instead of asking people to write a piece of code for the exam they'll be asked to explain the code right. Um and and I like the examples that you gave with with notebook LM like how much of that

content what I'm asking is like can that we're going to create repeatable artifacts right I think educators are going to get recreating the wheel and instead of doing it in silo I'm wondering like can that notebook LM that set of artifacts be shared right in an open source manner so that we don't have to recreate the wheel over and over again especially given how fast technolog is

moving moving right. That's that's the idea at least. Do you want to add anything else? >> Sorry, my slides keep going to sleep. I just wonder also I try to challenge I agree. It's like uh you want you want your students to build resilience, right? So that they keep going. Um I think that's a part of it. I I think I think educators are going to have

to be creative, right, about how well I also think, okay, so I think they're going to have to be creative to have the students learn the intended material, right? I think also the expectations of what students need to learn is going to change, right? Because I do think that um I I think I'm trying to think of a good example uh historically that that could relate, right?

So um you you provided the Wikipedia example earlier um which is interesting because we're going to talk about that in a bit. I don't know why this is not working. Uh hold on. Oh, just fell asleep. Okay, we have help. Thank you. Um, so I think that Wikipedia, I think, is an interesting example where teachers were afraid that uh students would just use Wikipedia to write their

book reports, And I do think that teachers found creative ways to help um I think not just to prevent cheating but to use Wikipedia to augment their learning, right? And uh also I think have it not be like I think Wikipedia is a little bit it's like make sure it's not the single source that's used for the research that you do, right? And I think there was

a more easier workaround than with AI. um perhaps a calcul the calculator example that we've heard many times in uh you know used like when the advent of calculators people were worried students are not going to learn math but there were ways that were like we found ways that students still have to know math learn math and they can still use the calculator in the AP exam

right so I think there's I think that's the creativity that's going to be required is like if AI is here to stay what has to change in our pedagogy to make sure that That's well what changes in terms of what students need to know and learn and then what do they what needs to change in terms of what they need to know and learn and then how

do we change the teaching to address that? Oh, it's the wrong one. >> Yep, I can do it. Thanks. I'm not sure what happened, but we're gonna do take two. Thank you. I'm going to try to keep stay here so I don't have the projector go to sleep. All right. Thank you for your Okay, cool. Um, one second. Um, so I think Adam actually has a good

point of view on this. So 10 to 15 years ago, when we talk about AI, right, and the struggle to keep up with the pace of AI, 10 to 15 years ago, AI, the AI word was cheapened from what it meant in the 1950s, right? um it was originally meant to be um superior to human intelligence to refer to tactical improvements. About six years ago now, the

term generative AI rose to prominence to bring refer back to that old definition. Um and three years ago, I think advanced search or four years ago now, right? Like advanced search engines such as chat GPT uh came to fruition and two years ago, news articles from today are now in your AI models. like I can just search and find out what's going on if I just search

like Google and use Gemini. Um, and they're using Brag, right, to to figure out what's going on. So now today, like AI knows a lot about you and educators like you hear, right, that AI is a tool. Educators still decide on and shape the curriculum or like AI is not something to fear or avoid. Students will use it anyway like we talked about, right? and the everpresent

by bringing it in yourself, you're helping students learn to use AI the right way. Student interactions with AI must be observed and encouraged so they do not fall behind. Right? So, and I think Adam's going to talk a little bit about like how it's harder in the developing world with the scarcity of mentors and resources and yet I think the potential for AI exists. So why why

like if I want to look at the numbers we already talked about students using AI from a GitHub classroom survey we see like close to half of students 47% it's a AI is a daily part of their schoolwork another 21% use it weekly and most of those who use GitHub do it do do so frequently including 74% who use it once a week and they aren't just

using it for school they also use it for personal work and I found that really amazing because the results are centered around AI and ladder up to increasing overall GitHub and co-pilot usage and we see a shift in direction with over 70% of teachers using AI in the classroom. So you talk about like there's teachers that don't use it in the classroom, but some of like at

least for the teachers and here it was like uh 1500 teachers responded uh in this survey across 142 countries. The teachers that are on GitHub, right? Se uh 70 I say yeah over 70% of teachers are using AI in the classroom. So that's an astonishing number to me, right? because it means that you're going to encounter a teacher who's teaching about software collaboration. Why are you smiling?

So, >> go ahead. Go ahead. Or is it? >> I think it's actually the former. I'm assuming it's the former. I don't know if we got into that detail, but I'm assuming like they're somehow incorporating AI use in the classroom. But we know that AI use is not just for coding, right? Like you have many products that now like even Chrome, M365, they have AI embedded into

the the desktop applications. >> it% predisposed to using AI. >> You think people on GitHub are are predisposed to using AI? they're already predisposed to >> I think there can be exceptions to that. I think there are people No, [laughter] >> that's why it's not 100%. Um I So I agree the survey is uh bi I guess it's biased is the right word. Um or it's the

survey is leaning towards people who I think would be more inclined to to use AI. I agree because they're already on GitHub. Um, is there a hypo are you saying that there's more people not on AI than is reflected in the survey? Is that the the point? >> Oh, absolutely. Yes. Yes. Absolutely. paints a picture. >> Oh, absolutely. Yes. CS versus nonCS use. Absolutely. There's a huge

disparity there. Yeah. Go. What were you going to say? >> The universe is different in the CS world versus the nonCS world. I agree. Wa. >> This is your survey. There's a wake for >> Yeah. Yeah. It's your survey. Yeah. >> Do you remember? Can you remember? Yeah, but I think the the question is slightly different. I don't think it's co-pilot specific. It's about um is it

how how is it? Go ahead. >> Yeah. What qualifies as a yes? Is it that they're using uh like the teacher has incorporated AI into the curriculum or is it that they're just observing their students use it? I don't think we we got into that detail. Yeah. Um Okay. So, which country do you think is the most skeptical of having AI in the classroom? Sorry, I see

your hand in the back. Yeah. Go ahead. more predisposed. Um Where where are you here? Sorry. What what population are you talking about? >> In the US. >> Which region? LA. Los Angeles. Okay. What were you gonna say? Your Okay, cool. He said his wife is a elementary school teacher. Kindergarten. >> third grade. Okay, cool. And she's using AI, too. Awesome. Thanks for sharing. Um, which country

do you all think is Afghanistan. Oh, >> it's actually the US surprisingly. Which receptive? >> India. Cool. So, what what can teachers do? So, before I provide my non-teer perspective and tell you how all like things should be, um what would love to hear ideas from you? What what do you all think? I mean, we already heard I'm sorry I didn't get your name. >> Kim, like

we heard from Kim about the work she's doing um at where was it? Berkeley. Yeah. >> Cool. And but you were describing a Berkeley example. Is that right? Or >> Amazing. So, we heard um from Kim about the example through Dora. Yeah. Go calculator. We also change AI just make sure the way I approach They are required stand. >> I like that. Yeah, I think that's great.

What What subjects do you teach? >> Oh, cool. Do you fly? >> Oh, amazing. All right. Awesome. So I I think the hypothesis here is that teachers need to help teachers like this is a great conversation right just the ideas that are flowing here I think you all have strong points of view and experience to reflect that so would love to um figure out how to do

more of that and I I believe um that we should democratize knowledge right and create sharable content um the topic specific prompts we talked about model selection guidance uh make it easier to share learnings like on an open source repo And we've done workshops with internet in a box. Um I think inerson and virtual workshops are very helpful with GitHub. We're going to have an educator workshop

in June. Invite you all to join us. Um and really do the like just dive in and do the hands-on work and really look at turning like turning passive educators and learners into creators, right? with everyone learning the tools as they evolve and progressing. And I think the so the whole inspiration of this talk actually came from uh an archive article about archive arxiv that I read

about its origin and how it got started. And I was thinking like because as I said like I get up education one of my jobs is to help teachers and I was thinking like how do I actually help educators in the classroom on board to AI and this archive article talked about how researchers uh you know because of the publishing life cycle they did not have an

opportunity to obtain peer review prior to submission right there was no real platform to do that. um outside your own like local network like very localized network. So archive was formed to say hey before you publish to science or nature um you know you can upload your content to archive in a pre-published format and get feedback from people and get it peer-reviewed in a very um codified

way. I think that Wikipedia does a similar thing with the the editing that it provides and I think GitHub makes that possible for code with collaborative software development and I think internet in a box which we're going to hear more about in a bit um does that for offline mostly offline communities um and they serve as transformation models and I think there's three pillars to making that

happen it's the technology like you have to have like a place to do it right like a software place like a centralized software place to do it. You have to have the right content and the ability to share that content and then you have to have the community of educators or community of users around around you to learn from each other and share knowledge with each other.

And so here I provide like several like the platforms as examples um and just the scale that they've reached, right? Uh, and I can go into like I think you all probably know all three like but I can go into details about why each one uh like prescribes to those common principles of free access, peer review and collaboration. Um, I'd love to hear from you all though

like what does this resonate with you at all? Like does this kind of uh help you think about how you can use share your AI learnings with each other? >> Yeah. Like like uh workflows like a like when you're talking about skills, you're talking about like aentic AI workflows. Yeah, I think there are um I don't know if there are any educator specific ones yet, though there

are like I I know that um Enthropic just came out with a class um for educators with materials. So, uh, I think that we have resources on, um, the GitHub education discussion forum in our public repo where you can pull lesson plans and and AI like our through like basically skills like it's really markdown files that you can incorporate into your workflows. Yeah. instructional designers. >> There

are for there are instructional designers involved in creation of learning content at GitHub. Are there instructional designers involved? This is an idea, right? Like this is like an idea that teachers can learn from each other, right? Um I I think there could be instructional designers involved in figuring out how what is the right way to do it. My current thought is we take these learnings and create

an open source repo for teachers to um contribute to that. That's and then I what does the instructional design look like beyond that? Is it like common courses? Is it common templates that you're looking for? I'd love to hear from you on like what it is that would be specifically helpful. >> Oh. Oh, you mean like the the course to teach teachers to use AI? >> [laughter]

>> I was wondering. I think yes and though I think well my hypothesis well maybe I'm too arrogant but I like with software development like with uh GitHub right where it enabled like the anyone can contribute philosophy right where like from beginning programmer programmers to seasoned developers, right? But there's no like instructional teacher monitoring everyone's code, right? It's like peer reviewed, right? People are uh code like

doing code reviews of each other's code. Of course, if you have five years of experience versus one, you may be asked to review more complex tasks or you may be given a higher level of authority in getting code out to production. Right. But I so I guess I I kind of I may disagree with with you in that teachers can't review each other's work, right? I guess

and I think I think one of the questions I have is those curriculum committees, those people that have advised and the instructional designers that have advised to date. Is there time? Is there time for that process or do we need to become more efficient given how fast things are changing because what's true last month may not there may be new things that for you to try today

and use right it's a question I don't know and you may be right like maybe it is helpful I'd love to hear from others too that's great thank you okay and then I'm going to turn it over to Adam to talk about internet in a box and an offline example we can learn from >> Hey, um I hope this is uh on topic. This is the dark

ages. These little boxes uh contain the internet in in a sense and uh yeah, some of you have heard this, but I briefly lived in Haiti after the earthquake 16 years ago. So we're forced to do this, forced to, you know, note what we were doing as teachers as you said and bring hard drives full of open street map, Wikipedia, Khan Academy, you know, TED talks. It

was the early days and uh so this is just an example. Uh yeah, it's about as um anti-generative, anti-transformative, but the earlier definitions of those two words when the community was the generative part and the community was the transformative part. We're not asleep at the wheel. Um I would say not quite 10 folks around the world who are trying to inject AI into these little boxes which

contain you know as you said uh in more colorful language than me like some some some very uh cool classes learning gems right um so we have University of Houston we've got some really kick-ass students who are trying to put little pieces of this conversation into these offline online uh learning libraries. Um none of let's be honest, none of it's working perfectly yet. Some of these AIs

are just acting as search engines. Some of them are acting as coding playgrounds to get you to think about what an LLM is. It's like us three years ago saying, "Is this stuff really coming?" Yes, it's coming. All these kids in Haiti, they've seen their aunt and uncle use AI on the phone. They know it's coming. Um so you know um it's a step back in time

in a sense to take um all the things we've just been talking about 10 20 30 years ago um and we do have um within the RAM limitations within one to eight gigs you know rough LLMs working completely offline um with as you know like uh radically more powerful models every couple months I wish uh, you know, these these very low-income communities could hear this conversation in

this room to hear what's going to hit them next in the coming five years. Um, yeah, it's um, it's the old world of the, you know, that we've just been talking about the the staff classroom, you know, throwing uh, rough ideas back and forth at each other. The schools that don't have instructional designers by definition, you know, at the bottom of the economic totem pole, right? But

they still make some of these things happen and even they are getting they're getting their feet wet with AI when they have absolutely no you know IT shop down the hall that they can you know go to. >> I think this this last slide also this community this picture here is from a teacher workshop that we did in Mexico in 2020 and we brought like I think

it was 50 teachers together and showed them in a workshop style format. It it was like teaching them about computer usage, about looking at like connecting to an internet in a box and exploring offline content, like exploring the apps that are on there. And I I think that the work that Internet in a Box does, right? Yes, not everybody in Internet in a Box is on AI

yet, but they are teaching each other, right? They are figuring out ways for the communities to teach each other. And I think that can apply not just to people in Mexico or people that are not as privileged as us, but can also apply to where just knowledge sharing needs to happen in my opinion. >> So not much to say, but uh and we don't actually measure how

many people are using this. So it's somewhere between thousands and millions. We don't know. Um but it's good to to be reminded. Um yeah, this is in Kenya, this one. Yeah. So, um it's 32 students connecting or or 10 students to a little Wi-Fi box like this where a local NGO or church or government as you know like curriculum committees have always done. They've stuffed this full

of what they define knowledge to be, yeah, this is a this is a pretty um successful example in Kenya that we ran into last August. Um yeah, and it should be uh it should be growing many steps above this photo if uh the two gentlemen in the back there uh introduce uh something that allows kids to uh create their own Android apps on Android tablets like this.

If you guys want to hear more about that, um you can ask the two guys in the back there later on. So, it's it's not just a static dump of, you know, grandpa and grandma's library inside here. Um, it's it's definitely generative, but what does that mean? It's evolving. >> Thank you. Okay. Awesome. So, I I'm always inspired by the stories that I hear and see. Um,

so the the picture here, this Malumb Victor, he's um a technologist and an impletor in Africa and Kenya and he's set up 50 schools across Africa. like he's helped connect them, right? Um just through downloading this open source software, working with an NGO and figuring out like how do I order equipment from China and put it together in a suitcase uh and and and make it available

to students that that need it. And you can come to either um the internet in a box booth to learn more about this work or uh which is in the expo hall booth 420. And also um come there's a talk tomorrow which will talk much more about um the work uh that we do. Uh the talk is where in the world is internet in a box. So

to get back on track uh conclude these Neanderthal technologies in other words stuff at 1990s internet in the box are now being combined with 2026 technologies and we don't know where it's going but uh yeah we have you know the two uh appdev for all guys in the back row there are talk is in succession to ours five and 6 pm tomorrow if you are interested into

this more radical view that you know offline grassroots techn technologies can deliver a lot of this without, you know, live chat GPT. Um, but we're, you know, we're grabbing pieces of chat GPT inside these little tiny boxes where we can >> Cool. Thanks for remembering that. Yes. Um, all right. So, I think how how does a local network become global, right? And I think teachers are comfortable

with I think I think teachers are comfortable with people just like like employees like I'm comfortable with people in my local community, right? like it takes a while to acclimate and feel brave enough to reach out and I think somebody mentioned this like how earlier like how it's like talking to a instead of like going to the TA office right um it's it instead of having to

face and ask some hard questions to a TA fear looking stupid I can just go to my chatbot right and ask questions and I can be very vulnerable with it and do this iterative thing 247 I get access to essentially a TA at my side that I can ask anything to and learn. And I think that's been something where AI has been helpful. Uh and I also

think that can we provide an environment uh where um any anyone right can learn um whether they're in Mexico, in Africa, the US, you know, can can we provide an environment where we can all learn from each other and what does inclusive access to knowledge look like? Right? Can curriculum align learning be globally crowdedited and locally adapted? Um, I think teachers are content verifiers, content givers. Um,

I think you're also like, uh, empathy providers. I actually relate a lot to teachers as a manager. That's like one thing that I found. It's like I I I majored in CS and I'm I always think like I probably should have majored in communications and psychology both like every day, right? When when you and I think teachers are similar, right? Because it's a lot about people management,

right? in a classroom. It's like meeting meeting your students where they are and figuring out how to help them grow. all right, this is invitation to collaborate. I'm just going to go there. It's invitation to collaborate. I'd love to work with you all on figuring out how to make knowledge sharable. And uh, you know, you can get involved any of these places. is um there's public education

resources and and I'd love to get something started in the public repo here to help educators like share knowledge. So if you have an idea, you have a seed and you want to get going, I would love to work with you to make that We have a few more minutes for questions. If um you all want to have a little bit more discussion, the discussion has been

really good. So thank you. Thank you for all the questions and the prompts and I want to keep the discussion going. Please come. Um, you can't miss us. Like when you walk into the expo hall, the Microsoft and GitHub booth is right there. So come say hi. We'd love to see you. Um, I'll be there uh a lot of the day. Uh, and I'll also be at

the Internet in a Box booth. So come by and see us um today, tomorrow, Sunday. >> Uh, any questions for them? So for uh the internet box project are you considering downloading some open source models to probably run in there would be super small and limited models but still a good approach to give some sort of access to AI models to students that don't have access to

the internet. Uh and I don't know if you have considered to do anything around that. Can he have the mic? >> It's just not distributed. There's uh so many folks that have got really good alphas or betas of that. Um yeah, not on these guys that have 512 megs of RAM. It doesn't work well here. But on the larger Raspberry Pies, it's clunky, but it works. It's

it's got to get out there. Um, yeah, it's it's it's on the verge. Um, I don't know which ones should be um, you know, sanctioned or recommended. So, every school in Zimbabwe has it, that kind of thing. We're not there yet, but we're quite close. >> Think we had a question in the back. >> Uh, one thing about teaching I there was a good observation from uh

a colleague of mine Marian Lang who's a former professor now principal engineer at GitHub and she talked about um how there's a lot of hype around AI right and the hype cycle with just a little bit right just a little bit there's like a like you know basically the what industry is conveying as this is completely like revolutionary going to change your world tomorrow right versus the

meaningful benefits of AI and the software development life cycle that are very true and very real, right? And I think I think that ambiguity, navigating that ambiguity and figuring out which is which I think is important as we learn to use AI. Go ahead. >> Cool. Thank you. Um, >> more questions. >> Okay. Thank you all for coming. I'd love to continue the conversation. Thanks so much.

Yeah. >> Uh, next presentation will start around 5:' I believe. Uh so hope to see you all there. The presentations or the slides for this one will be available online. >> Patrick has an announcement. >> Other than being obnoxious. >> No, I'd love to meet you. What's your What's your name? Yes. >> The education track. >> Oh, you're a perio. Yes. Okay. Awesome. >> I'm with the

pererryio and we helped organize the track and we're hosting a social. We don't sell anything. We're a nonprofit. So, if you wanted a drink coupon for later this evening, um, let me know and I'm happy to since you all had to suffer through my Yeah, that's right. underwater. I can make them available after Are you a teacher? I don't know anything. I don't have an internal how

I have to check this out. Oh my god. hours. class. the opposite. I'll use And then I will not used to what my background just graduated. And I walked over as a not as right. Okay. Uh All right. Hello the remaining people who uh willing to stay for the last session on Friday. Like I knew that will happen. I knew that will happen because I would never

stay for 5:00 p. p.m. on Friday. So like I really commend people who still come and stay and want to listen for something. So at least you know what rumors are. It is recorded and it will be recorded. So probably hopefully somebody will listen at their leisure time right so let's hope for the best okay uh so uh I am going to talk about posgress and you

know what I noticed today I sat here listened to all the talks uh about open source and education and I was wondering how many times people will mention any So you know how many times databases are mentioned today? once once uh so uh yeah uh so we talk about open source but somehow it's mostly uh Linux and uh all programming languages but not databases and I do

not want databases be forgotten because all the problems which were discussed today about open source and education when it comes to databases they are tripled quadrupled and whatever So uh and I spend a life trying actually to do something with that almost life. Okay. I'm like I'm doing it for very long time. So uh my name is Hi uh and my day job is a database architect

at uh DRW holdings is a trading firm in Chicago. Uh also I am a founder and president of priary posgress nonforprofit organization uh dedicated to posgress education in the Midwest states of the United States. Also I'm a communication chair for the ACM Chicago chapter and I'm a LPI board of directors member. So by all of these three I kind of have a place at this convention. Um

Uh also on top of everything which I mentioned uh which already takes kind of about 200% of my time uh I uh run Illinois prior post this user group and I'm an organizer of uh teacher data conference which will happen in uh Chicago uh in June. So that is I have literature about this conference here on the u first row. Uh this will be two-day conference and

uh it will include uh complimementaryary training on different aspects of posgress uh first day half day training. Uh we already have full program tickets are on sale. Uh we have community and academic partnership. Um and uh we have uh full uh price ticket uh $50 uh and uh student tickets $20. Uh so uh come and join us. Uh I will uh I have caros here. Uh all

right. So, as I said, I'm from Chicago, the city which gave lots of things like skyscrapers. Uh, the largest in the world warehouse, not data warehouse, but industrial warehouse, Montgomery Warehouse, the largest in the world exhibition space, merchandise mart, uh, the largest ferris wheel. uh also the international workers day May 1st uh and uh international women's day and lots of other things and among those I think

the only city in the safe space the only city which did not let national guards in uh so we have something and as uh the author of the Chicago uh development plan uh Daniel Burham said make no small plans and I'm not making small plans I'm making big plans actually. So that's why I am here. All right. So uh what sorry what is uh the problem uh

like the talk is called uh posademia and uh what is the problem? Okay. So uh here is the problem. Uh everybody here before me like each each person to talk talk about uh the gap between academia and uh industry and uh it's well known fact right so on one hand absolutely we are talking about yes uh we should teach students fundamentals we should teach students the way

to solve problems but also we need to expose them to some world problems And the biggest problem is that we are continuously having difficulties here. Again, it's not new. Uh it was forever uh and uh it was in all industries, right? So being a generic problem, it is especially acute when we're talking about teaching computer science students and especially when teaching databases. uh and uh it's not

because I love databases but I think it's objective problem. So let me explain why I think that this problem is the most serious. So uh I would cover four different aspects. It's like um major and uh curriculum uh students expectations uh training materials and lack of open source software exposure. So let's start from um majors and curriculum. Again uh it was mentioned today multiple times uh that

curriculum as difficult to change. It takes time. It doesn't happen instantaneously. At the same time what I hear when I speak to faculty uh people are trying to stay on top of current research trends. Yes. Because they need their public But what is what is buzz? It is important because they want to make curricula appealing to the upcoming students because they need more students, right? Uh so

that's a kind of uh idea behind what uh how curriculum has been built. At the same time, students want to be employable after graduation. Yes. Uh that's that's why they're there, right? I mean, we want them to be able to think independent to solve problems, but they want to be employable after graduation. And uh while they want it, they often do not know what it will be

required from them when they will enter the workforce. Uh so, uh expectations for the students. So they want to be ready, they want to be employable and uh they go to job fairs, right? Uh they uh look up the listings uh and uh you know what they see? What they do not see? They do not see anything about databases. Like literally look at what is offered as

entry level. No position mentions databases. It's always kind of it might come up later in the interview nice to have but it's definitely not a requirement. uh and uh you know only normally only several years after graduation students realize that probably they need to know something about databases again and I talk to these students it might work you know several years after graduation because uh they come

to me and asking why why they think which they design do not work. So again I faced this problem for like super long time because before I became uh like database administrator, database developer, database architect etc. I actually um been a professor and researcher. I worked in academia for several years. Then I went to industry because I like to do cool things. Um but once a teacher

forever teacher so um been there and uh we don't need theory. I heard it 40 years ago. I hear it now. It is a huge problem. That's like uh like yeah like more than 40 years ago when I was like kind of younger starting like TA and I taught uh the evening classes uh for you know returning engineers who wanted to learn about databases. Oh my gosh.

I started teaching relational theory what is like selection projection. they complained to uh the department that I'm teaching them wrong things. I'm not giving them practical skills and nothing changed. Same story. Same story now. Like the hardest part to explain people that they need theory. Uh and uh the students who are like current students uh they often do not think they need theory of databases. They want

to learn something cool like AI. What's the relational gosh the relational theory is not cool because everybody uses no SQL databases and that's all out there which is not true. Okay. Uh so uh next problem uh insufficient uh training materials and uh that's like multifaceted problem because first of all the objectively not enough training materials to teach real uh and um again generic for all computer science

especially for databases. So um as again anybody who taught any computer science classes knows developing training materials takes more time than developing courses. Uh the other thing is uh that uh also uh when you evaluate like 300 students you need some automation with results evaluation which means you are using some tool uh which uh does not necessarily uh checks the depth of understanding uh of the Uh

and again uh for all computer science especially for databases uh even you know what forget about evaluating students uh when um people um like apply for some position and they go for technical screening. Uh there are all this automated online tests where they are supposed to solve some problems with SQL and uh the only thing these tests evaluating they grammar evaluating that people know how to write

grammatically correct SQL statements and something which produce correct results. There are no tests which would evaluate whether the road efficient SQL statement how long it will be executed. Will it be executed like in one second or for 10 hours? Like these materials do not exist. These tests do not exist. Period. Uh so no exposure to real world problem. So uh now uh limited of any usage of

open source. Everybody talked about this today, right? And uh again for databases it's even worse. There's again historical problem. Uh I experienced it when I was in academia. We have academic subscriptions. We have academic subscriptions. Uh we have Oracle academic edition which is for free. We have Microsoft Office which is for free. And uh people tend to rely on this because yes it's for free. It's available.

You do not need to do anything. Uh and this gives free promotion to proprietary products. Uh and also it secures future user base because again we can talk forever about which database is best for which project but at the end of the day you know what we choose what we know how to use and uh if we started with Oracle we are using Oracle for the rest

of our lives no matter how much it cost if we uh started with uh Microsoft it's very difficult to switch to something else because it works and we know how to work software and know how to make the best use out of it. So um also this kind of like traditional uh suspicion uh regarding uh opensource free products you know what I did not plan but when

I was searching for the image I saw this Russian foreign minister I could not skip this so I decided let him be on this slide. Uh yeah um it just popped up. Okay. So I I just said let it be uh so uh that's honestly but this existing bias towards open source products is less reliable which sometimes maybe possibly true but not always. Uh uh and uh

also uh another big problem is insufficient marketing of opensource tools for high education. And now as a LTI board member I can tell that LTI does very decent job regarding this but probably that's the only organization which does decent job uh regarding marketing of open source tools for high education. It it just does not exist and in general marketing for for like open source um could be

much better. Okay. Uh now why uh okay where is our invisible hand of market? Why market does not catch all these problems? Again as I said when a student chooses where to go when they choose their high educational institution they do not know what they will need. During this four or five or six years while they're there things change and again uh databases are so so so

like in the shadows like they do not know they will ever need databases they do not pay much attention. Uh also uh like important consideration computer science versus software engineering. Uh so yeah uh companies do not hire computer science graduates to write programs. They hire software engineers and that's important. Uh and uh engineering component is important and uh we do not have it in high education and

um I actually wanted to ask this question. I'm asking this question. Is there any college or university who officially give degree for debates? No. Right. where all debates are coming from, who teaches them, who like certifies their quality. God, but I can tell you that no education institution in the world uh like educates the base generically. So um yeah, that's a problem uh with okay and who

is affected everybody. So uh I can tell uh that hiring companies so for the past long number of years I am on the hiring side and I see this problem when people come and uh they do not know basics uh it's actually alarming how much basics they do not know uh obviously graduates who have trouble finding employment and open source community why open source community suffers because

um we uh lose this potential user base and we lose future contributors and uh yeah all of the above. Uh so uh going coming to databases okay why using opensource databases is especially important. So as with any other open source product we are talking about availability uh equal opportunities fostering diversity inclusion and specifically for posgress that is database which originated in academia database which was developed in

the university database which is the full scale complete relational database with all the modern trends and you do not have any light version for posgris for academia or for usage you go to the website you download and you have full posgris with all capabilities you can use it you can try so um it's extremely important extremely useful in uh education process and uh also uh that's a

great support for the community because again it's like brings us new users new potential contributors uh and I mean lots of things which can be done at postgress and academia overall. Uh so uh that's for academia why it's important for posgress project again new ideas new hackers new user base and same as by Oracle puts Oracle light into educational institutions we would rather put posgress in education

institutions because we want people to have uh we want this future graduates to have posgress as their to go tool as the tool they know most and they would start uh they would advocate for using posgress at their future works. Uh now that I talked like for the half of my time now again uh that's my organization uh prior not for profofit uh so we um serve

Midwest states of the United States uh and uh we specifically focus on educational aspect of posgress our conferences our meetups our training courses uh are all geared toward educating developers uh students and uh bridging this gaps between um poss and all the other aspects of the world. Um so we are focusing on Midwest community uh on uh application developers and on uh students and recent graduates that's

our target audience. so uh what we found uh easier said than done uh because uh we all know we like posric community knows that existing problem existing gap uh people do not know like not enough um students come to our meetups. uh and uh I thought that okay we are not advertising enough and like it turned out to be extremely challenging uh like literally it's such a

simple thing to like hey students we have something for you to offer right uh so uh then uh so there okay Um so why uh because uh students uh I mean okay in the past year I think or so students started to be on meetup com more frequently before that they just never were on meetupcom. Uh they had their discords their like servers their uh like whatever

forums. uh so uh you know like ivory wall uh around them um then they kind of do not see a need until they are about to graduate. So um I started to see students when kind of like you know six months before the graduation they just start actively looking for meetups and they start to look where they can meet people they can meet professionals uh so uh

then they start to appear and uh you know if they get interested in posgress like us but technically they are all done you know so uh I managed to you know interest some of them in posgress but uh kind of a little bit too late though it's like not hopeless. So they do not have incentives learning because not a part of their curriculum and uh does not

seem like will benefit their future. Uh and there are not enough communication channels honestly like first year when I was actively doing it it was how I can actually outreach it was not easy. Okay. uh challenge getting faculty involved uh uh that's still the biggest challenge because why why what's the reason what's in for us? no reason right so uh like okay we have our curriculum we

have our courses why we need to learn something which nobody is asking for by the way students are not interested in having this knowledge uh also okay when you go to academic conference you can get certificate of participation when your work is accepted to the academic conference it is your publication it's your citation index it's something you go to this conference nothing. Why was the result? No,

nobody comes. All right. Uh now uh ch getting researchers interested in presenting. So same story why they should go why they should present it's no benefit. Uh and uh it's actually not easy to uh bring new academic ideas to post community. That's pro problem of community but it goes both ways. Okay. uh so uh what we uh like how we started to approach it again I work

actively in this direction for several years now uh so uh first of all we are doing extensive outreach uh so like honestly we did this marketing so we specifically wrote like very targeted messages to all Chicago universities computer science departments. Uh so uh first uh like one out of 10 responded and actually when I started to talk to them I found that the only reason why they

responded because one of the professors read my book which starts from like okay no professors ever saw I don't think otherwise they would even respond. Um so that's a very mixed success. I'm happy I have one you know one one university interested in academic partnership and actually we are working on signing contract with them so I think it will happen one out of 10 uh then inviting

uh students to our meetups we have pizza it's easier [laughter] uh honestly that's like I think all the incentive uh for them to come but um small percentage of them come for pizza stay for posgress which makes me very happy uh again uh one out of 10 probably but some of them uh then we offer uh students volunteering opportunities. So that's um so how it looks. So

we have a conference uh you can come and volunteer which means that you get free access but also I can tell you you are not going to this session you are maning this other sessions. So they kind of have this limited um like options to go to whatever talks they want. Uh but again it's free and free meals. So uh uh I uh get like seven 10

12 students every year like maning the conference and they are exposed to possess doing something uh also u discounted tickets. So um this year I mentioned we have $20 student tickets uh for several years. I just try to make sure that if some of our sponsors is not using their ticket, they communicate to me. I give their new ticket to students. So it was operational nightmare because

they can uh decide uh you know okay we have four sponsor tickets only two people are coming we have two tickets and it's like a day before the conference. So now I need to quickly find okay hey students who was interested like please confirm your coming uh I was absolutely sick and tired of it. Uh so then we try to work on like um sponsor subsidizing student

tickets. That's also another challenge. So I think that uh the only way to attract more students uh is to make student tickets like really really really really low cost. Uh so uh this year we have $20 student tickets uh and I mean okay conference is in June but uh guess how many students bought the tickets so far uh so I am hopeful that that's not the end

of it and again as I said we have partnership with one of the universities and they actually seriously want to partner with us. So, uh I was I am invited to their like u fair like uh like career whatever business intelligence fair next week and they actually asked me to be there as a resource with like our literature. Uh so uh hopefully something will come out of

it. Uh but um other universities um yeah and it all only goes on personal uh you know, contacts, communications, if somebody knows somebody, uh, all right. Um, okay. When I have when I have presenters from outside Chicago like you, Jimmy, I always order one uh deep dish pizza. When they are only locals, we know better. So, I always have good pizza from Jardanas. Better pizza than ACM

meetups have. But because they have Domino pizza, I have Jardanos. Uh but [laughter] uh but uh deep dish we do when guestona like you come. [laughter] Uh yeah. Uh so um big thing which we I uh we introduced this year is academic partnership. So we used to have community partnership and we extended it this year. Uh so um for community partnership it's like you are nonforprofit or

meet up or something you partner with us you sign zero co sponsor agreement uh and uh we give you two free tickets and your logo on the website and we advertise you and you come and learn from us. So academic partnership that's thing which um I um introduced invented uh and as I said it's going super slow but idea is the same you are academic institution you

have u you offer degree in computer science you um sign our sponsor agreement you pay zero dollars uh we put your logo on the website you can give you like promotional stickers or promotional flyer we will distribute uh We get two free tickets for faculty and uh life is beautiful. Again, that takes forever because nobody wants even to sign one page agreement, right? Um so um I'm

not uh hopeless but it's challenging. But uh again uh my hopes are that you know when I sign first academic partner then it will be a little bit better after that because people will see that you know it does not bite it's all good and there are some benefits. So uh hopeful hopeful hopeful and we'll see how things will you know go this year. Uh so uh

what else uh we want to do uh that's what we're already doing. Uh we want to simplify this academic partnership because again uh as funny as it sounds but people kind of get scared of like one and a half page contract and uh like or who is signing this or you you want the usage for our logo? I'm not I do not want the usage of your

logo. I'm offering you to display your logo for advertisements or now we need to consult with somebody. So so we uh are going to simp simplify this. So my other I'm not making small plans right I said so I'm not making small plans. So a big plans I'm making I'm trying to establish like all your partnerships. So we are talking with at least two educational institutions about

uh doing like um classes um on posgress uh like in the like after school time uh and their like campuses and hopefully something will happen. Uh I'm also looking at doing hub day summit um again we'll see hopefully uh now why California has this convention we have nothing in Midwest so somebody has to start uh uh the big part is uh increasing outreach outreach I think is

a key word in all of this because uh people do not know that things exist and uh I know that if I'm sitting there saying okay uh if you want free ticket come to me uh it's not enough uh that's what I learned for all the years when I'm trying to um know attract more people who never did posgress but who might benefit from using posgress so

um that's a lot actually uh because somehow we have too much information in the world there are too many things to attend too many things to pay attention people do not pay attention to anything as a result uh I don't know I thought that uh I advertised like call for proposals for pig data like I don't know every day for several weeks and still people who even

are in my clos network what you say is closed already when why why I did not know I don't know uh so uh trying trying to reach different communities so I'm talking to jungle community to Chicago community to u black python devs to like uh you know to many many communities and I would love to talk to more communities because again uh otherwise people do not know

people do not know what benefits they might get attending this conference so um doing doing doing all right and uh so it's difficult loss of inertia academic environment most inertious environment I'm sorry I think I have a right to say I've been to academia so it's like the most inertious uh yeah uh so we're trying to apply marketing I have a professional marketer who actually came to

me and said how I organizing uh and you know was the most effective tool what I learned was the most no not pizza you think it's pizza the most effective tool is uh I surprised that oh you have talks about the yes we have students who are interested so now uh okay that's uh all about me. Uh if you want to connect to me, I'm on LinkedIn,

my GitHub with several of my pet projects uh prior organization and uh pitch that conference. So you can take pictures of these QR codes and uh yeah, I mean as I said I said this will be 35 minutes, right? Uhhuh. Uhhuh. Uhhuh. I timed it. >> I was interested in your uh your thought about the conferences with. So, I just had this conversation with somebody from OpenSSF

and one of they're trying to bring in more academics is that they're teaming up with like ACM for a let me say it's oh god I'm blinking on what they call that there's a special term yeah but yeah but but the special interest group yeah thank you I know there was a simple term um and the special interest groups that I know that for security related and

then the idea would be that they could have then the conferences look are much more like highly regarded by themsially for PhD students and the like. Yeah, >> I thought when I invited me, I thought that that will be my avenue. Nope. Nope. Nope. Nope. Zero interest. Zero. Uh honestly I am like one of uh the um the president of um ACM Chicago chapter is a professor

in UIC. I try to talk through him. I'm talking to outside him because he's zero interest. I thought so you know I I I'm an member. I like uh I go to conferences and I I I'm there. I'm there. No, I wish I >> uh I like honestly I'm doing all this outreach now. I think the direct marketing is the most efficient because uh otherwise it get

lost somewhere in many departments. So just trying to find we even try to send separate email to the student council to like you know the general students to like research. I'm I'm fine. So, thank you for the talk and I have a lot of ideas. Uh I'm with the uh the Thunderbird project and we have been wanting kind of for the past year to kind of explore

university and academic kind of and so the thing is like we're not a database technology though you'll be happy to know we are moving to a yeah finally um and so we're like a collection of technologies and so for project like ours where there's like so many different avenues students could take from QA to support to encryption to databases to Rust. What would be maybe some strategies

you might recommend for a project like ours where we are devoted to like one specific language or one specific database? again like as I said I had a moderate success with this. Uh so uh like pretty you know I probably like um especially when you offer like several different things it might make sense to reach out to like uh you know who is responsible for the like

student affairs in the organization. Again it turned out it was not easy because when I try to like literally I go to the university website and try to find who could be person I find like nobody. So this person who said I am professional marketer I can help you. She actually did research. So she like wrote specifically to specific uh people. Um and um so I think

that that's one thing but some universities actually do a lot for students. So like my successful my semi success is with Northwestern University. My kind of more successful story with with the actually are like very very interested in uh making sure their students have like radios have a future in industry. Uh so I was like fortunate that they already on the lookout so they just okay let's

let's talk. Um so otherwise um like As I said um I'm happy when students come to meetups but uh that's a little bit too late when they come you know it's I think it's most beneficial if we are talking about like sophomores may maybe like you know juniors when they actually start doing real things when they're out of their like pure theoretical woods and start like looking

for something and they are not doing meetups. So um you know if I will be able to set some like free training I will let you know but so far it's conversations they like did not materialize in anything. Um I mean let's solve this problem together. Thank you. Those are the final talks for this room. >> Uh Henrietta's slides will be available online for you guys' uh

viewing and other one. But yeah, we'll see you here