A.Mahdy Abdelaziz: Cloud Run Your Code

This talk focuses on running code in the cloud, presenting an overview of cloud computing concepts and specifically discussing Google Cloud's service, Cloud Run. The speaker explains the evolution from traditional hardware management to using cloud services, highlighting the benefits of cloud infrastructure, containerization, and serverless applications. They introduce Cloud Run as a solution for developers to deploy their code without worrying about the underlying infrastructure, allowing them to focus on development. The speaker also outlines use cases, limitations, and cost benefits associated with Cloud Run, emphasizing the importance of scalability and developer-friendly features.

[Music] ladies and Gentlemen please welcome our next speaker Mar abdel's presenting the topic cloud ring your code that that was impressive a i generated introduction okay so so if anyone is working as a full-time host or something you should worry about your job uh right right thanks for joining me today uh my name is Mii and I'm going to talk about running your code in the cloud

but I have this habit before starting that I'd like to take a selfie with audience is it every okay with everyone to take a selfie okay cool uh let's try this corner I've been actually testing many corners before start let's see actually need to do it this way yeah so cheers uh okay all right thanks all right so uh again thank you for joining me uh in

this um first talk for the day I guess uh my name is m again uh before I start I just want to highlight something important which is basically here uh in this talk I'm presenting myself so I'm not affiliated with any company I'm not representing any company whatsoever uh even if I work for any company that does not represent anything from what I'm going to discuss today

uh I'm I'm representing myself briefly and uh our topic today is quite straightforward and simple uh I'm going to make a refresher for this Classical Mystery of what is the cloud why the cloud and then uh you probably have noticed I'm playing with the word cloud run so I'm going to highlight a little bit more on what is cloud run and then uh some use cases and

limitation and what to when to use it and when not and since I mostly have recent experience with Google Cloud I'm probably going to show you how to do that on Google Cloud however uh I also want to mention that uh what I'm going to present is applicable on pretty much any other Cloud platform so uh Amazon Microsoft and so on more about that later uh pretty

much any Cloud platform should today have the same technology so let's start why the cloud so those of you who have been attending conferences for uh while now and by while I mean like more than 10 15 years ago uh you probably have heard this like move to the cloud let's go to the cloud and so on so this is a very classical question I'm not introducing

anything new today why the cloud I'll tell you my personal experience my personal experience and that's almost 20 years ago when uh I was asked to actually host a website and uh by hosting a website in our uh company it was literally means I go and shop for a hardware assemble this Hardware at home and then uh put the put it under my desk and then like

put the website on it and this is how we went online believe it or not this is how back in the days you would put your website online you just have a computer like this one but then you know because it was under my desk then any movement any wrong action or anything like that can render your website offline you have a broken cable you unblock your

cable but you are also bound to the limitation of the place of our office if our office doesn't have good uh internet connection I mean we are an office just an office so you can have perturbation in in the connection or uh maybe uh you the electricity went off or maybe you have any other factor that does affect your business uh although we tried and that's a

fact by the way we tried in this company to have our own Data Center and by building the data center we literally converted the kitchen in into a iner so we took a corner in the kitchen that used to have a fridge and then we replaced it with a rack and then we placed some hardware and cables to make it a little bit stable like not pluggable

and unplugg uh we placed um a dedicated air conditioner for the heating and like that was our uh equipments back then but it was not only us I mean many companies and many big uh Enterprises were actually having this and the solution was quite simple back in the days it used to be called collocation facility so a collocation facility is basically someone telling you hey we know

that you are struggling with this we know that you have been doing all this under the desk uh computer and hardware and so on how about we give you a room that has all the equipments ready the rack ready good internet connection good electricity uh good ventilation you just come in plug hardware and you are done now collocation facilities still existed today many companies are still using

them it's started approximately 20 25 years ago and basically that was the way to go back in the days the only limitation with that is if you want to go Global if you want to have multiple collocation facilities it means that you have engineers in many locations and those Engineers need to physically move into the collocation facility to manage the hardware replace the hard disk replace the

cable add some new hardware and so on and yeah this is basically answer the question why the cloud the cloud is basically telling you instead of telling you come in person and manage this yourself physically I will do it for you me as a cloud provider I will do it for you and you just stay at home in your terminal I'll manage everything for you so next

time when someone ask you why you the cloud probably this is the answer just letting someone else manage the infrastructure for you and uh uh believe it or not this this is this is the adopt the adoptation of most of the big companies today and small companies as well anyone anyone today doesn't really want to go and shop for Hardware or build a dedicated kitchen convert a

kitchen into a data center or even go to a collocation facility that's too much money and too much hassle and too much management uh so so why we just don't go to the cloud and that's pretty much convenient solution for many but then we have seen this trend going on like infrastru for you and then when we are saying managing the infrastructure we're saying that how about

you give me your virtual machine and you don't even need to know about the infrastructure so I'm not going to sell you a computer with specific processor and RAM and so on I'm going to just sell you a virtual computer and then I will tell you just give me an image and I'll manage it for you and that's probably 15 years ago when we start to hear

about containers and the idea of containers is I'll take it one step further I'll also manage the operating system for you so the idea of containerization when it started uh I believe it was a competition between Google and Docker uh and Google claims that they started this first because everything at Google is cized but basically I'm not here to uh solve this debate I'm just telling you

that the idea is quite simple do you really have to manage an operating system do you really have to go and install Windows Ubuntu red hat and so on does it really affect your Core Business if the answer is yes then okay good you have a virtual machine you can go and do that you can let's say you are doing a virtualization of a custom operating system

that has custom kernels and stuff like that fine go make a virtual machine but if the answer is no you never have any preference about what type of operating system what it does what are the underlying libraries and so on you probably best with a container where you have to worry only about the libraries and containerized environment but building a container is not an easy task you

need to build a cluster and when we say building a cluster it means that you have to go again and get many virtual machines place them together connect them with network cables uh have uh uh what we call it the control plane that controls the other cluster uh cluster machines or cluster nodes so you see the trend we first asked it for the collocation facility to manage

the room for us and then we ask it for a cloud to manage the infrastructure for us and then we ask for a container to manage the operating system for us how about we also go one step further and ask for someone else to manage the cluster for us and here you see where this talk is trying to send you send you towards more of the develop

how many of you are actually developers writing code every day okay pretty much the majority I see so the rest of you who don't write could if you happen to be a a infrastructure manager or something like that or architect or something like that I I still hope that this is going to be a good news for you because the new trend now is to tell you

when you are writing your code when you are doing your business do you really have to go and study Cloud do you need do you really have to go and study about the infrastructure the virtual machine the network the VPC security the firewall the connector the cluster the containers and so on how about a solution again like this one it's called Cloud run and by the way

Cloud run is really a service at Google it's called Cloud run and basically the idea of cloud run is quite simple and straightforward especially for those developers out there it tells you as you can see write your code put it in a container and that's it so again if you if you look at from the right my right your left hand side code you just write your

code code and then containerize it use Docker use kubernetes use whatever to containerize create a container a native container give it to me and then I will host it for you inside Google Cloud now again I want to highlight pretty much that my experiences with Google Cloud that's why I'm showing examples on Google Cloud but uh similar Solutions more or less on Amazon called Lambda and also

on Microsoft called uh if I remember correctly called container it's called container because it's containerizing but yeah more or less regardless of your cloud provider we pretty much should today a modern cloud provider should have a similar solution where basically I'm telling you you as a developer write your code give it to me as a container and that's it this is my talk today thank you for

okay I'm going to give you a little bit more details about this Cloud run thingy that's what I'm going to do rest of the day I hope it's convincing enough but let's start with this quite interesting thing I promise you that I'm going to be a developer friendly and I promise you that you don't need to learn Cloud you don't need to learn infrastructure you learn network

security all these kind of things and I hear the developers out there telling me do I really still need to learn how to containerize because you are asking me to do that now let me show you something there is something called build packs which is an open source project originally created by Google but it's completely open source anyone on the planet can use it and build packs

is that project that tells you if you are a developer that do not want to put your hand dirty inside the containerizing environment and how to create containers and so on build packs will do it for you now it sounds like I'm making you giving you an ask please go study build packs so that you can conze your application sounds like that right well Cloud run tell

you okay how about you just give me your code I will build packet for you and put it to the cloud so ladies and gentlemen this is this is the idea here we go yeah this is the idea Cloud run also goes one step further by telling code can it be as simple as that just give me your code you don't need to worry about how to

build pack how to containerize how to do anything you are a developer put your code bum it works again I want to highlight one more time that this is not bound to only Google obviously so pretty much on Amazon Lambda is doing similar stuff uh but again because my experience is on Google that's why I'm showing you this uh this example for me as a developer I

really love it to tomorrow you tell me can you create this little website for me to demo something well yeah I write the code and that's it python JavaScript whatever uh can you build a big scaled application that has a lot of functionalities and so on yeah sure why why not I I will just build the application write the code I focus on the code and I

know that there is some guy out there somewhere that will perform everything automatically for me now uh even though this sounds like too sweet and sounds like pretty good uh I think we need to also highlight the limitations of course everything comes with limitations so uh first of all obviously the middle part the build pack part it's actually something that need to be um a container and

that container if you decide to build it yourself if you say like no I don't want build packs I prefer to build the container myself then you must build it in a compatibility mode of Linux 64bits which pretty much the default or or the most popular so I don't believe that there is any issue with that actually this is a good news because if you are writing

in let's say assembly I don't know if someone is coding with assembly or cobal or Fortran I don't know you name it any non-popular programming language today if you if you are writing in binary if you're writing any custom Library any system native library or whatever as long as you are able to containerize it and build it for Linux 64 then it will be accepted and this

leads me back to the previous slide let me go back to this previous slide this guy in the middle the build packs is not like that so because build pack will automatically containerize your application build pack comes with the limitation of what is the accepted programming language so yeah it's not any programming language on the planet it has to be well it's a a generous number of

programming languages such as C++ C I mean net uh JavaScript go python uh what else yeah uh Ruby PHP uh Java of course uh so yeah those are the supported programming language and of course when I say supported programming languages I need to add a version so not all Java version on the planet are supported it's only specific three or four versions are supported so basically build

packs is a convenient if you are one of those developers who develop with one of the most popular programming languages on the planet in one of the latest versions however if you feel like you want to build it yourself you want to create something yourself okay then uh you H you can go to the classical mode which is basically building the container yourself compile it for link

64 now how it works in practice basically uh when you launch something inside run uh what happens behind the scene is basically there is something called process and this process will launch a container for you and this container will start to execute based on requests now this is a very important detail as well because when I say it serves pay and request it is directly affecting your

budget because you only pay P request which is also one uh important milestone for your uh uh accountant in the company who is paying the bill uh you no longer have to tell them hey I have a cluster up and running 24/7 all the week even though we are not using it but I still have to pay for the cluster this is no longer the case now

you pay per execution so today you have 1 million people visiting your website you pay for 1 million tomorrow is a weekend nobody's visiting your website you pay zero so that's also uh in my opinion a very great milestone in terms of optimization of resources and optimization of of of budget because you only pay for serving now when when I say serving I put many characteristics here

I put the CPU utilization the ram utilization and of course the dis space utilization if happens so this is a good equation you can put in place now someone will tell me okay I have a website similar to Facebook for example and Facebook has not million requests per day it has millions of millions requests per second right so yeah in this case we will go into a

debate does cloud run make sense to Facebook I have a feeling that no it doesn't make sense for Facebook right but um this makes more sense for those kind of things that really not keeping your cluster up and running all the time 24/7 is going to save cost with addition to the convenience and so on because basically when the container is Idle then the system will automatically

terminate it for you and then you no longer have to pay for it so that's also part of the system that automatically manage the life cycle for you now if someone noticed something very important about this when I say the container is idle and terminated then there is no container running anymore then when there is a new request coming what will happen you will have to wait

for the startup time so this is also another as I said I promised you from the beginning there will be limitations right so this is also another thing about um uh about to understand when you are dealing with this such a system that you need to be a clever developer that either eliminates the startup time you have zero startup time you don't have a cold start whatsoever

or the alternative option which is basically asking the system to keep a set of containers running up all the time so yes in this system you still can say you know what I want to keep two containers up and running all the time I don't want you to terminate this container I will pay for them even if there is no requests so that's another option for you

to make sure that the first experience by first experience means like when a person visits your website for the first time the first experience goes smooth someone opening your website randomly right now they want them to get the response as soon as possible so uh what other advantages that comes out of the box when you do a system like those is basically obviously automatic scaling so you

don't have to worry about how many nodes how many machines how many clusters and so on you ask the system to scale automatically for you of course you can uh limit the scalability you can say I don't want to scale more than 10,000 for example you can do that or you can say as I just said you can say I don't want to scale less than two

containers ERS I want to keep two containers up and running uh all the time you can also set it for incremental updates and by incremental updates it's this world where you have zero Town time so if if you are introducing a new version you don't have to terminate old version you can introduce new version to come into place automatically you can of course also has these other

uh advantages such as load balancing and so on and the beauty of it is all these infrastructure talks you have have to worry about zero of them you do nothing at all so you as a developer again and myself I mean I'm I'm talking about this purely because I love it I'm in love with this me as a developer today every single time I want to write

a code and introduce it to the world I just to put my code I put it and it works I don't have to even think if it's going to be secure highly available attack did US attack security firewall did I open the port did I close the support did I do this management did I update the system is the operating system under attack I don't have to

worry about anything just put my code and it works so yeah this is our uh those are the features that comes out of the box with it and uh to reinforce this a little bit let me show you a few use cases where uh Cloud run might make sense so basically a website that's the example I'm giving you all the time so you have a website uh

you can add a bunch of services for caching and uh load balancing and so on uh but then there is the thing about Cloud run that cloud run does not support session if you notice the architecture of cloud run uh basically it will the the container itself can terminate at any point and that means that you cannot rely on a session in the container so you as

a developer as well need to be clever enough to have a pure serverless kind of development now here I'm showing you a solution a quite straightforward solution where basically you session you can put it in a radius or uh in memory database so if you're are clever enough you should be very familiar with this kind of concept where basically we are in the world today where nobody

keeps the um the session inside the application the session you should put it inside a database another use case can be a backend rest API for example um can be also a a typical microservice so you create a few like your application let's say you have have um a website selling cloes so you have the front end you have the back end you have the order system

you have the inventory system you have the printing system you have the email notification system you have five microservices and then you want to connect them all together so that might be a typical use case um also excuse me also a good use case can be a workflow so uh let's say you have a specific workflow one job after another and then you want to execute them

in a specific you can also uh use it as a Schuler so let's say every day morning at 7:00 you need to create a report to send it your manager so that also can be a direct use case and in this case it barely will cost you anything now as much as I mentioned a good use cases for cloud run let's also say uh when not to

use it so Cloud run by nature is meant for this Snappy serverless application if this is not your case if you must have a session in your application or if you must have an always on application 24/7 or let's say you are designing something for heavy use case let's say 3D rendering or machine learning or something like that probably Cloud run is not really for you now

I got this question uh many times someone is asking me uh I have developed uh I have I have trained a system I have a trained system already for machine learning and this system is already trained I have a model can I serve it from cloud run or not the answer is yes definitely so let's assume that you have developed a chatbot now can I serve this

chatbot from cloud run yes definitely because the chatbot is about sending a request and getting a response it's not a heavy processing what I telling you not to do is training the chatbot itself the training phase of the chat bot probably Cloud run is not your best friend but now that you have the model working and you have something hosted somewhere you can pretty much host it

inside the the cloud and then making those requests go through Cloud run now let's terminate with some limitations as well or concerns so uh I I told you that it Autos scales and the infrastructure of those Cloud providers such as Google and Amazon and so on they don't really have any problem to scales to millions of millions of containers if you leave it open and you are

lucky enough to get a lot of requests they will scale it to millions they don't have any problem the problem will be you because you will get a nice bill at the end of the month so uh you need to make sure that you are actually uh controlling the scale and making sure that it doesn't go over uh much the second issue or second challenge is the

downstream mismatch so what is Downstream mismatch so Downstream mismatch means let's say you have developed the front end with Cloud run and the back end let's say it's a classical database so this classical datab base doesn't scale as fast and then Cloud run doesn't have any problem to scale and go up and get millions of containers but the back end itself is not as fast so what

you will get is the user will think that your website is fully functional however you have a throttle you have a bottleneck which is the database is not responding as fast so to solve these kind of things is again on your responsibility on your shoulder as a developer to create some sort of a queuing system to make sure that uh your uh your system understand that if

there is a big load then to queue the load on database or uh another solution would be to actually from the beginning have a scalable backend so use a powerful database that can scal as fast so uh a lot of things we can discuss number two is really a big discussion that can be another topic how to make uh a matching between the a fast growing back

end uh front end such as Cloud run for example and a back end such as classical database the third topic which is quite important which is uh portability so uh what is the issue with portability now any good company whenever they commit to any cloud provider they want one important thing which is no lock in I don't want to start with Google and then find myself unable

to quit Google I want to start with Google today but tomorrow I want to to Amazon and then after tomorrow I want to go to Microsoft and then after tomorrow I want to just go back to to to to uh to inhouse in-house data center so so portability is quite important and luckily this solutions that I presented to you today is 100% portable because it is at

the end of the day a container and as you know container is 100% open source 100% portable I don't believe there is any cloud provider today that does not support a container so you are not at risk if you start and test Cloud run today but then you use it for one year two years and then you don't like it anymore it's it's okay you take your

container put it somewhere else and you good now if I want to highlight one single thing to share with you today how to get started if you want to try it this is the command you type obviously inside Google Cloud gcloud run deploy and then you give it your source code and that's it literally that's it uh depending on how your system is configured depending on your

permissions and everything it is going to interactively ask you for enabling an API enabling billing enabling and so on so it's going to uh go you over the steps necessarily to get everything set up for the first time but then after that the next executions of this same command this is a command gcloud run deploy and then after that you give it the source code that source

code again can be PHP javascrip script node uh Java Ruby and so on you got the idea C++ net and so on do I need to highlight it one more time if you want to take one take away from today just try this command by the way I promise you there is uh a good qua a generous qua to try things out I I personally uh most

of my demos that you might find on inter internet are actually there and I pay zero for them because I think I get some sort of 10,000 10 million 10 million requests for free per month so that's more than enough from My Demo consumption uh in the last few minutes I'm going to just quickly show you how to uh get started so basically uh this is my

console that was a demo I did a few months ago basically uh you notice on the top there are two buttons create service and create job so the difference between them is uh service service is what I was discussing with you all the day which is basically let's say a website or a backend and so on but job uh something's designated to terminate so something that will

start and terminate let's say uh create a report or let's say uh make a batch processing or something like that so something that is not really made to receive a request and send a response it's meant to start perform an operation and then terminates that's mainly like briefly the difference between them um so so this is showing you how I uh try to do this demo basically

there are two options uh you can have uh even linking your uh git repository so if you have GitHub or something like that you can uh link it and then uh eventually uh you can set up a hook let's say for every push automatically send a new version to Cloud run so that's quite uh nice and I try to highlight the free TI that I'm using so

I'm getting two million sorry I said a wrong number earlier I'm getting two million requests per month for free and some CPU usage and RAM usage for free as well so yeah please go ahead and hit my demos try to reach more than two million to make me start paying for uh what else yeah so so about the CPU remember at the at the beginning I told

you that you can make it if you are if your application doesn't have any cold start whatsoever if your application is Snappy and start automatically you don't need to to allocate CPU all the time however if your application let's say a spring boot that does I mean nothing against spring boot I love spring boot spring boot is excellent right but spring boot needs some 10 seconds to

start right your user will not wait 10 seconds so you probably need to allocate a fixed number of CPU and keep them up uh on all the time so that uh you have a fixed number of containers running all the time and that's it you pretty much click on create uh it took me when I was trying to do this demo it took me less than one

minute to get my application up and running and I got even a public URL so don't know if you can see oh this will not show you anything but uh the URL is exactly in the middle of the screen so uh this is a URL um it's it's public and of course you can modify it to point it out to your domain or custom domain or DNS

and so on so as simple as that I'm done uh what else I want to show you uh yeah so so later on what you might think about is more management so let's say you want to split traffic you want to create a new you want to modify the traffic you want to create a tester and so on so this is possible to do uh you can

manage the traffic as well so pretty much uh management system for you or a place to manage everything so uh I hope at this point I gave you a good idea of uh what is the latest trends in letting someone else do the job for you um my talk as you can see is mainly oriented for developers who don't want to learn cloud container management security and

so on just put your code in 2 minutes you are public uh for those of you who are using Google Cloud I just want to uh finalize with this last slide comparing with um majure computational option on Google Cloud so uh so basically uh by the way I didn't draw this don't take me wrong the the credit goes to the lady up on the left she's quite

clever and she has a nice book uh was full of drawings so please follow her uh but basically uh she's showing you basically the differences between the computational options that you might uh see such as a traditional compute engine when to use it this is when you want to be in control of everything the operating system and everything or uh kubernetes when you believe in containers but

then uh you want to be in control of the cluster and everything pretty much or Cloud run if you are convinced with what I mentioned today the the next step after Cloud run would be Cloud functions so I didn't talk about cloud cloud functions today but briefly speaking Cloud functions is a very special case of cloud run where we tell you you know what we think that

we asked you to do too much to put your code and connected and so on how about I give you an IDE inside the cloud you write your code inside this IDE and bum it works that's Cloud functions so uh of cloud run where you don't even need to worry where is your code and how to write it and what are the instructions and all these kind

of things you literally open the Google Cloud console you write your code well from its name it's function so it should be a function not really a big application but it's a Snappy function you write it inside the console you click save you have versions you have Version Control and so on but you click save and then bum it works without going over the hustle of else

again you can find this photo on her Twitter account and her website don't worry about taking photos of this one uh I think she summarized also my talk at some point so this is summary of my talk basically if you want to recap again this is on her website don't worry about taking a photo of this one am I good on time uh okay I think at

this point I can take questions if there are any yeah I see something here awesome thank you for the questions so I'm going to repeat the questions and you correct me if I understood correctly so the first question is uh can I configure the idle time so I mentioned earlier that uh if the container is not used it has some period of idle and then it gets

terminated and uh your question is can I control this time uh the current answer based on the documentation I checked uh is no Google decides that for itself and that's why Google never gives you a guarantee like hey I'm going to give you an idle time for 10 minutes Google doesn't want to to do that Google tells you don't count on my idle time even if it's

10 minutes just if you really want to be on secure side keep at least one container up and running all the time okay uh that's based on the recent documentation I I read I don't know if there are plans in the future to change that or make it configurable or something like that remember that this is pretty new technology what I discussed with you today uh was

released last year so it's really pretty new uh but yeah uh that's about I mean Cloud run Cloud run has been around for four or five years but cloud services Cloud run jobs and services are pretty new from last year but anyway uh the second question is about billing can I if I understand the question correctly uh it was can I uh set a ceiling for the

billing I don't want to I want I don't want to uh over expend and the answer is unfortunately no so again uh you cannot set that but the workaround so the workaround is uh basically um so you mentioned something uh about alerting yes that's exactly the workaround so what you can do is basically you can set an alert alert system uh on the budget on the billing

so that whenever you reach let's say a certain number let's say 1,000 for example then you trigger something that trigger something can be anything it can be as simple as sending me an email sending me an SMS sending me a pager call okay literally you can make a pager call a Skype call a slack call and so on uh or it can be a function like Cloud

function and this function let's say terminates the service or disable the API or something like that uh there are demos about that uh if you want to check it out check something called uh events Arc so there is a service called events Arc that that can uh take an event coming from the billing and translate it into a a script and that script can do something like

let's say terminates service okay but the billing has this alerting system you can also make it alert uh based on a specific number so you can say for example I want to be alerted at 50% of my budget my budget is 1,000 so when I reach 500 send me an alert when I reach 60% 70% 90% and so on you can do that as well does it

answer your questions awesome thank you uh more questions no more questions oh yeah uh thank you thank you for the excellent question so if I understand your question correctly you asking if I can use AI to write the code for me that can work um the answer is yes so uh I I have been playing around with this uh I might make a demo about it tomorrow

so uh the the quick answer is yes uh I think Google has introduced something called DUI but now they renamed it into gimi and so on name it whatever you want but no now it's called Codi Codi it's called Codi so uh Codi is what you're looking for uh the quickest and fastest way to try this out is from the cloud shell so Cloud shell is basically

a command line tool inside Cloud console inside the browser you don't need to install anything you open the Google Cloud console and then there is a button on the top right called Cloud shell you open it and then it opens an editor for you a fully fledged uh code editor and then in this editor on the bottom right I'm I'm literally preparing myself for the demo tomorrow

uh on the bottom right there is a a small um uh gim symbol you click on it it opens the Cod and then you ask it write me a code that takes a request and create an order for cloes or something like that and then it will write the code for you in which language pretty much any uh most of the supported on openly available languages so

the future is here it's scary right but thanks for the excellent question all right uh are we good for time okay any last one more minute question think we are good amazing all right so uh again I thank you so much for being here with me I shamelessly announce about my uh next panel I'm going to be in the panel next talking about Ai and cyber security

so those interested about AI might uh be interested in this panel as well uh and tomorrow as well at the same time at 11:00 I think 11 or something I'll be also uh having a panel about uh Ai and software engineering so this is exactly what you're talking about so panel today and panel tomorrow both of them about AI uh would be would be happy to see

you in those panels as well thank you [Applause] so thank you m so much thank you wow for you yay thank you and you would be still able to ask some questions at the ask me station that's near the registration okay thanks