Escape from Drupal Island! (Spice level: 🌶️ 🌶️ 🌶️ )

In this talk, Josh discusses the evolution of the Drupal ecosystem and the need for its community to adapt to changing market conditions. He shares his experiences founding and running Chapter 3 and Pantheon, emphasizing how the web is being reshaped by AI and the importance of interdisciplinary collaboration among content creators, designers, and developers. Josh critiques the concept of 'Drupal Island,' advocating for a broader perspective that embraces technologies from outside the Drupal community. He highlights trends affecting the adoption of Drupal, including the need for agencies to focus on market niches, utilize new technologies, and harness artificial intelligence to enhance their offerings. The conversation is framed against the backdrop of Drupal’s current challenges, including a plateau in new site builds and the end-of-life status of older versions. Ultimately, he calls for proactive engagement and innovation in order to thrive in the evolving digital landscape.

is the get off Drupal Island spice level three go chili pepper emoji talk. Uh thank you for coming. Um I uh I'm excited to be here. It's a privilege and an honor uh to present at Drupal Con. Um I know not a lot of you do know me uh already in the audience because I've been around this for far too long. But for those of you who

don't, hi, my name is Josh. I got started in Drupal uh working on the Howard Dean campaign in 2003. Uh so I got introduced to the Drupal community while trying to take over the government. Uh my Drupal user number is 3,313. I started doing Drupal more professionally for a little while at a a shop called Trellon um with an extensive alumni network. Any other Trellon heads in

the audience? No. Okay. Right. Well, a lot of people worked there for about a year or a year and a half and and uh and that alumni bond is strong. Uh no, but Mike Mike Hagerty is a really good guy. Uh gave me my start in some ways. And uh but then I opened my own shop uh with a couple friends, Chapter 3 LLC, your Drupal experts

in San Francisco. And uh we that that business ran for a good 20 years. Uh but about four years into it uh we founded another company called Pantheon which was based on us seeing what we had to do that was repetitive in our chapter 3 practice versus what was unique to the client and realizing that that repetitive part of the job could probably be turned into a

product or or a platform actually is a better way to put it. Um, in there I also did the Drupal dojo which was a online learning community to teach people how to Drupal. Um, which today you know in retrospect what we did looks a lot like what code streamers do on Twitch but it was like a fully self assembled tech stack. We had like Skype going and

IRC and like this crazy J2 based screen sharing thing cuz like um a Goto Meeting didn't quite exist yet. It was like pretty cool cowboy stuff. and probably my biggest contribution to the Drupal community to be honest like I've contributed some code and some modules over the years but the fact that we probably trained up a couple hundred people at that time to become you know more

expert Drupalists uh is probably has the biggest long-term impact um anyway I' I've lived this it's been a big part of my professional career uh and development and I I really love this community um it's really fun to be here for the 25th anniversary of uh Drupal I'm going to talk just real quickly about Pantheon because this is our Spawn Con session. So, I used the Spawn

Con angle to get this talk accepted. It didn't have to be approved by a track chair. Um, so, uh, Pantheon believes the web is being reshaped by AI, both in terms of what the web does and also how it's made. Uh, tech is continuing to sprawl. Governance and security remains a challenge and is only becoming more so. We believe that building the web is an interdisciplinary effort.

We believe that it requires people who create content, people who create code and people who create design to work together in a really effective way. And what we are building towards is a world where every one of those stakeholders can use the tools that they prefer to use and have a workflow that helps them just dance really effectively and have fewer gaps in their process and have

less room for human error and uh less uh friction around oversight. It's uh it's why we're doing things like directly integrating GitHub into our platform or building connectors for Google Docs and Microsoft Word. It's why we branched out early into the WordPress ecosystem as well. Although at the time some some people thought it was heresy reader it was not. Those are nice people too. Um whoever we're

here for people building the web and that's also why we're launching first class support for Nex.js JS uh next week uh and because it's a big part of how the web is being built now and we want to be supportive of that in this overall vision. We have three real principles in terms of how we operate. We think we think for all the teams that are building

the web, they need the freedom to execute. Um they need to be able to run parallel work streams. They need to be confident that that's a safe way to operate and that they can know what's going to happen. Um, we think they need to be able to use automation in ways that are um, additive to their process and that they can trust. Again, the ability to move

fast without breaking things. Very important. Uh, we can't all be uh, startup debags. Um, and and obviously like because we we work like all of you do uh, in contexts where we're not building throwaway prototypes and we're also not like vibe coding out the first iteration of someone's website. Although that's fun to do. I like messed around with doing that myself. We need to serve customers where

they have real institutional requirements and real risks around governance, security, compliance, etc. All that stuff where if it's not built into the way the system works by default, it becomes this kind of straight jacket that you put on the system afterwards. And that completely undermines the first two pieces of the puzzle. So we bring all these principles together to serve all of you. Uh we power tens

of thousands of meaningful Drupal websites. Thousands of web teams log into Pantheon every day to do their work and collectively those websites reach over a billion unique monthly visitors uh via the web. So that's what we do. Now >> let's get spicy. Um before getting super spicy, I actually do want to pause quickly to define some terms. So I posted a thing on LinkedIn. You may have

read it. It generated a lot of comments and a lot of engagement. The most LinkedIn engagement I think I've ever gotten on anything. Um, and a lot of it there's confusion because we can talk about when we say Drupal, we could mean a whole bunch of different things. And so I want to talk about the language that I use and what I mean about that specifically a

little bit upfront. So there is the Drupal project, the Drupal open-source project. That's the thing that everyone's contributing to building in one way, shape, or form. Uh, and that is organized through drupal.org and its associated uh, uh, uh, places. And that's the software at the center of all of this. And then there's the Drupal community which includes the people contributing to the Drupal project. And then like

a wider range of people around that that are also just doing documentation and working on uh other things to help out the Drupal association or helping to arrange events like this. There's a community of people around Drupal that participate in the building of the project. And then there's the Drupal economy which is the the the commercial ecosystem around the community around the project that allows people to

do this as a job. Um and that includes a lot of jobs where you're implementing Drupal for a particular use case for a client or a project. Uh it also includes doing sponsored work to contribute things back into the core of the project. Sometimes those things go hand inand sometimes it's like no you're just making money over here so you can do your volunteer work over there.

But the Drupal economy is the larger commercial ecosystem that lets us have events in hotel ballrooms instead of like dingy little like um uh gymnasiums. Um and then finally there's Drupal Island. This idea of let's get off Drupal Island. This came out um as a a meme in the community in 2012 2013 when we made the decision to rebase Drupal 8 on top of Symfony. And there

was a couple other things in there about like rethinking how we worked with jQuery and other stuff. But the gist of it is like we're not going to build everything ourselves. We're gonna embrace technology proudly invented elsewhere as part of the wider solution that we're going to try to bring. And the idea of getting off the island was to break out of the thinking because if you

just do Drupal and only talk to other people that do Drupal, it does become a little bit of a bubble. Um, I also think that there's a an aspect of the Drupal island mindset, which is that like everything should get put into Drupal versus Drupal's part of a solution that includes many other things. So, it's just like a mental tendency that exists in certain places. All right,

so with that said, I like to be spicy in my talks. I this is not the first spicy talk that I've ever given. Um, I think the first spicy talk that I gave was in Drupal Con Austin 2014 where I bas I came out and said Drupal multicight's a bad architecture for people who are serious about like running large portfolios of websites. There's a bunch of reasons

why we have it that are legacy. It's not really well supported. It's a cluge. It's a hack. People will regret it if they make big investments in it. I got yelled at so much in the room, but I was right. It's fairly clear uh because the thing is in a world where you don't have to be limited by some of the technical architectures of like the as

that led us to try to squeeze as much as we could into one particular server, the monolithic multi-sight architecture is just a dog. It's like so bad in so many ways. And like if you're not worried about squeezing like your compute down, like do it differently and have let people live better lives. But and and here's the thing. I'm not bringing this up to reitigate the multi-side

debate. Again, I am I am uh firmly uh um confident and feel righteous in my victory. Um but I do it because um sometimes when I do this kind of spicy, provocative stuff, people think that I'm like a hater and I'm not. Um, I am trying to be provocative around an important point that I believe in strongly that I think it's vital for the people in the

Drupal economy in particular, although also in the community and in the project, but the economy in the particular need to realize and there's some complacency in the ecosystem around the amount of change that's happening. And so I'm being provocative just like I was then because I'm trying to shake people out of that and try to get them to think outside the box. And I've got some ideas

for where we can think outside the box, too. I'm not just saying slap, slap, slap. It's bad. You got to figure it out. I'm like, "Hey, wake up and let's work on this together." The reason why I feel a lot of urgency around this, and I posted about this, you know, had a very similar spicy LinkedIn post and did a session and had a big bof that

was kind of wild and out of control last year. um is because that business that I created, Chapter 3, um we sold it um to Canopy Studios, who are wonderful people. Woo! Canopy, love them. Um but I'm just going to be really honest with everybody in the room. We sold it because it was no longer sustainable. Um, we sold that business because the business had become a

a mechanism to respond to RFPs that included the keyword Drupal, of which there were fewer every year and more and more people bid on. And if you hit a run of bad luck and four or five of them in a row don't work out, you can't pay your staff. And then you have to start thinking about well I have clients that have been with us for 5

10 years more in some cases. We have long-term obligations to those clients. If we can't but it doesn't the business no longer works if we don't have new business coming in the door. And that's how you end up packaging it up and finding someone who will carry that legacy on forward. And Canopy is going to do a great job. They already are. I'm super thrilled to have

done it with them. But I'm not like it's not there's a part of it in my heart is not a happy story. >> And this last year when I started posting about this this this is when this dynamic like this was clear this was happening and it took us like all of last year to actually figure it all out. And again really glad that we got a

safe landing for some of the people on the team. All the clients are going to be taken care of. Canopy gets to grow their business which is awesome for them because again I think they're a great team. They do good work. But I got to watch firsthand what happens when you're complacent in this market. And what happens is you will go out of business or you will

be consolidated. That is just a fact. this is the graph that everyone has to stare at. And Dre acknowledged this. I did everybody saw the Dre note. I thought it was awesome. Like he is aware of this in the ecosystem. He's not he doesn't have his head in the sand at all. He's thinking about the implications of AI which we're going to get into. I have spicy

takes there. But this is the macro situation we find ourselves in. And part of the reason why I wanted to do this talk and the part of the reason why I want to be spicy and provocative is if you read in the LinkedIn threads, I basically said, "Hey, this is going on." And a bunch of people are like, "Well, actually, if you um cut the data differently,

it might tell a different it doesn't. It doesn't. All the data tells the same story." The rate of new Drupal sites being built has now reached equilibrium with Drupal sites that are being taken offline. And if you actually want to get really sad, I've got another the other version of this that includes Drupal 7. Um, then this is from drupal.org. This is Tim posted this in. Amazing,

Tim. I love them. They built a way better visualization of this because like it's prettier and it kind of helps to tell a this story a little bit better. This is a straight screenshot from the drupal.org usage statistics. I've got other sources that corroborate this and back it up. But the reality is we got the modern version of Drupal out 2013 really started to take off in

2014 and it grew pretty rapidly through 2020. Then it started to slow down a little bit. And then in the past year and change, it has been flat. And all of the pain that exists in the Drupal economy, all the latent like stress and stuff that people fear, not all of it, a lot of it stems from this fact. And the sooner we get to acceptance that

this is our reality and we must respond to it and do something, the better off everyone's going to be. Hence the title of my talk, you have to get off the island. You have to get off the island. Drupal Island can support probably an infinite amount of community and it can totally support the project, but Drupal Island has a finite carrying capacity for the Drupal economy that

is at this moment in decline. And if you think about it as an ecological metaphor, the people who stay on the island will be smaller in number and eventually eating each other. It's literally what's happening. Uh or metaphorically what's happening. Not figuratively what's happening. Not literally. No, sorry. Uh it could be really hard of the gala. And here's what I want to say. Here's what I want

to say. This is a You don't have to abandon Drupal. I'm not saying that anyone should abandon Drupal. I'm not saying that anybody has done anything wrong. I'm not saying the project made an incorrect choice. I'm not saying we positioned something weirdly in the community. I'm not saying any of the players in the Drupal economy have contribute. Like it there's no blame here. I'm just talking about

this is what's going on and we have to get outside our bubble. And that doesn't mean because many have interpreted me as saying it does. It doesn't mean you abandon Drupal. Although I will say that many people do because they realize they have to get off the island and they just leave, right? They stop showing up here. They get a job with a different kind of tech

company and they don't come back. Or and maybe some of them do come back if they're like, you know, old schoolers and it's the 25th anniversary. But if people are going off to, you know, more modern systems or other areas of tech, the chances that they'll boomerang back around because if they left the island because they saw the writing on the wall, they're not coming back. We

have to get off the island with intent. We have to get off the island with the notion of exploring and mapping the ecosystem around us. We have to get off the island to build trade routes, to build alliances, to start doing things that are larger than just Drupal, but bringing our special friend with us and doing the things that only it can do. I think and I'm

going to talk about this. These are not the only paths to growth, but for everybody who has built a career here or built a business here, these are three that I am watching work in real time today. The first, and these are you guys all know this, but I'm just going to go through it because these are good ideas and there are others. So, just, you know,

take it in. The first is to focus on a market niche or uh or or a solution. Because the truth is that most people that buy, most clients, most customers, they're not really searching for technology. If they're googling with or asking their friend, the research assistant inside their computer to find some information for them. Uh if they're using technical keywords, it's because they've already figured out that

they think that's a solution to their problem, but what they really want is to solve the problem. They don't really care if it's Drupal or WordPress or something proprietary. Most people don't actually care about that sort of thing. They care about their their problem. They care about their ambition. And so those of us who have been building the internet for 5, 10, 20, 30 years, Jesus, um,

you know, we know a lot. We actually have a lot of expertise in our heads. If we've served customers in similar spaces solving similar problems three, four or five times, you know what that probably means? You know more about them and what goes on in their world than any of them do. And turning that expertise into a solution uh or into a a a nichely targeted practice

is a way to ex identify and uh establish yourself outside of the Drupal bubble and build a growing business. I've got examples of all these. Um, another thing to do is if you're technical in heart and uh and soul, you got to embrace some new technology. You don't have to abandon Drupal, but you got to got to start to embrace some new technology. Like there are web

developers under the age of 40, I promise. Um, they just don't use they use other stuff. And the other stuff is cool if you give it a minute, right? Like, so spend some time working with some new technology. Figure out if there's a way for that to play into your practice. figure out if that can augment what you do, let you solve different kinds of problems, and

you can, you know, create some some innovative things that will get more excitement around what you're doing. By the way, these are not mutually exclusive. You can do all three of these things at once, but these are ideas for how to branch out. And of course, the last one, it wouldn't be a talk at a tech conference in 2026 if I didn't say radically transform your practice

through artificial intelligence. Um, but you should Um, we have to think deeply about what that means. Uh, and again, I thought the Dre note today was was fantastic. Uh, I he's very thoughtful at this. I have a ton of respect for that man. Um and he has really good really good points on how to think about um what even is the value of expertise in the future

and how much more valuable expertise becomes when it becomes basically cost-free to produce median level replacement like quality stuff like and there's a lot of expertise in this economy that we could refine and figure out how to leverage. So, some examples, um, solutions and outcomes are a better business than billable hours. Um, everybody knows this, but it's hard to change. Um, this is an example of someone

that we work with that just focuses on actually a specific niche subset of public sector entities. They work with transportation uh, departments and grant making associations almost exclusively. there. They provide a fullervice wraparound solution for all of their needs. All the stuff they do with the web, they support it, they host it, they just do the whole solution for these government entities that just want better technology.

For God's sakes, the stuff that they're working with when they get off of that and onto these guys, it's like the worst. I mean, you guys, how many here, how many how many of y'all have ever worked for a government client? Like, oh man, those people are in pain. but there's like you can find Drupal on like two pages on their website where they go into like

some of their technology stuff and implementation deals, but their website isn't about Drupal. They're not advertising we do Drupal. That's that's what we're good at. They're like, we know what your problems are. We understand you. We're easy to do business with and we bring you value. Right? Those are the ways that you can engage people outside of a specific technology space and get a really meaningful conversation

started. Um, embracing modern technology. This is a different uh cut of data than what I was showing uh from drupal.org before. This is a cut of data taken from my favorite uh open source research project, the HTTP archive. Um, if you're not familiar with it, the HTTP archive is a public data source that has compiled web scans across millions of websites going back to like 2011 2013.

They started doing a consistent count in 20123, which is why that's the start date for my data so I could use get consistent numbers. But over the past uh six years, they have been scanning the top 16 million websites on a monthly basis. They pick the top 16 million based on other publicly available data from Google Chrome that tells you which websites are visited the most. Uh,

and they put everything in there, man. The whole nitty-gritty output of the HAR file, the full page source, all the assets. You can do crazy analysis on this uh with this stuff. It's 51 pabytes of data and growing. But one real easy thing to do is to say in this sample of the top 16 million websites, what are they running? So the slowly declining blue line, dark

blue line is Drupal. That includes Drupal 7 because I didn't couldn't figure out a great way to make it do the version thing in this cut of the data, although it's possible. I just need to work on my big query foo. Uh, and the light blue line is Nex.js. I actually have done a whole bunch of analysis on the data set as part of my day job.

Nex.js is a crazy outlier. No other web technology whether it's like proprietary stuff like web flow or it's old word like WordPress is still growing but very slowly um nothing else has this pace of adoption like they they are beating the allegations of JavaScript has a flavor of the month framework and it changes you know at least once or twice a year like that growth trend is

just accelerating over time and there have been more Nex.js JS sites built in the past year than there are total Drupal sites on the internet. So, I'm not saying you have to use this technology, but I'm saying there's a reason I'm adding it to my platform. Uh, and it's it's cool. There's a bunch of cool stuff you can do with this, and it's really, you know, a

way to get outside the sandbox and potentially attach yourself to growth. I will obviously also admit, caveat, chapter 3, RIP tried to do the Drupal Next project as a growth move and it obviously didn't work. Um, I think that was a little bit of like if you build it, they will come style open-source marketing that to be honest has never worked very well for any business. But,

you know, so I got to take my own uh counterargument off the table because I'm aware of that. But Nex.js, check it out. It's growing rapidly. But if you're looking to bring people into your practice with a technology angle, like it's got to be AI, right? Because that's everybody's like trying to figure out what the heck to even do with this stuff. And people are doing it

like just before the conference like they like uh uh Crisen from Calamoon is like we're a Claude thing and Alex is like yeah me too. Kevin's like me too. I saw a whole bunch of other posts like this. It's I mean this is interesting to me. Um and I think there really is something in your ability to have expertise and know how to use the tools to

deliver things that can actually matter for clients because there's a lot of like AI is like you know a solution in search of a problem. Um, like and there, you know, no, I don't have any doubts about this at this point. Like maybe a year, year and a half ago, I could have been halfway considered an AI skeptic, but like we're not going back. Like we're not

the we're not going back to people not having a chat with their friend in the computer to do research about things. We're not going back to a world where you can autogenerate a bunch of code really quickly that mostly checks out. Like that's not going away. Now, how it all works out remains to be seen for spicy reasons we'll get into later, but it's not we're not

going back. and and that creates a challenge and an opportunity for everyone in the Drupal economy because the web itself is being radically transformed by this The way people encounter websites, when they encounter websites, what from the website gets to them before they ever visit it and how is changing very fast. And again, it's not going back. I don't we don't know what the new equilibrium is,

but we have entered a new world. We've also entered a world where the fundamental underlying economic model of the open web is not broken, but it's obviously in doubt. Um, you know, the open web model of free websites supported by ads was already kind of in a tough spot because like the ad platforms just took all the money. Um, and uh and that's why there's not journalism

anymore. But um uh but the the the this is even you know it's it's it's on another level right and so the idea of how do you account for the value that your content provides to these other tools but on the flip side you know it's it's it's a it's a very difficult thing to just say well um I'm going to throw up the Cloudflare AI blocker

thing and just nobody scrape my website because you will not exist uh in this world and so figuring out what we do figuring out what the open web is in this postAI era is like a real challenge for all of us and all of our clients and all the customers out there. So there's a very real opportunity to develop a strong opinion and position yourself as an

expert who can help people manage this transition. The other thing and everybody knows this but I'm going to like talk about it in like direct terms for for a minute or two is that the practice of web development is being radically reshaped by this technology. Drupal had a community motto. Talk is silver. Code is gold. Because coding is hard. It takes a lot of work. Um, and

it's easy to talk. It's no longer the case. Coding is easy. Uh, what you code and why is actually now the really important thing. And I I have a bunch of friends from other parts of the nerd world who I would consider more like crazy software engineers. And there, you know, that point got impressed upon me really early on as this stuff started to emerge. I'd be

like, "Yo, Andy, is this is this legit? Do you think it's for real?" And he's like, "Well, remains to be seen. Some of the quality is uh questionable, but like let's be honest, software engineering has never been about typing. It's been about thinking." Uh the typing was necessary and a lot of labor. And that's why talk of silver code is gold from a community contribution standpoint was

a good motto for the time, but not anymore. This little uh cover image that I got, and you got to go read this article if you haven't already. It's this thing called Gas Town. Have you all heard of this? Gas Town. >> It's crazy. >> Uh so Steve Yagi, he like used to be the principal uh software architect at Google and he's like a he's a personality.

So he's got this thing. It's like an open source project for you to stand up to like be able to run 20 to 30 concurrent uh uh um agent threads generating code off of your requirements documents doing QA you know handing off context restarting when they get stuck and picking this these things up. It's got like whole underlying architecture for maintaining state and and like different roles

for the things to play. And it's all based on like Gas Town from like Mad Max. There's like there's like the mayor and then there's the attack dogs that heard in agents that like it's you know his personality is all deeply in this and it's like it's it's one of the tools that you see streamers use when they're like I want to see if I can burn

a million dollars of claude credits in 24 hours. Watch me. Um but it's like to me. The fatal flaw of Gas Town is that if you read his blog post he's using Gas Town to build Gas Town. Um and there is a lot of oraoros energy in this space where the systems are expanding to meet the needs of the expanding systems. Uh the bureaucracy is expanding to

meet the needs of the expanding bureaucracy. and financially there's a lot of self-reinforcing feedback loops that are in play right now. Um, but like we are at the point now where there's a very real risk that everyone who was like, well, you know, it's great business to sell pickaxes to miners might find out that all that's at the bottom of the mine is more iron and they're

bringing it up to make more pickaxes. And if that's the case, if there's not gold down there, then the whole economic situation is going to change quite a lot. So, I'm going to get into that as like the closing kind of spicy arc, but I just want to reiterate this point and Dre said it really well earlier today, but this I think is like the key unlock

for many of us. Applied artificial intelligence actually going deep on a particular problem and having the solution for that is where all the action is at this point. If you look at how things have improved in software development, right? Like the the fact that it's you you're you're able to generate more and better code now than you were a year year and a half ago. All none

of that you should none of that actually gets chocked up to like core underlying model improvements. Like it's not the new version of like the underlying claude that did that. It's the fact that Claude code built a whole wrapper around that to put in place a process to help you use the tool effectively at creating software. Um which is all meta. um and all very valuable. And

that same meta layer of figuring out how to apply AI in a process to a particular problem with workflow and everything else, that is where the that's where the practices of the future will be built and people in the Drupal economy have the information in their heads to know how to do this. And it's a tremendous opportunity. But I'm gonna get a little bit more spicy now

about AI in particular, uh, if you don't mind. And like hopefully everyone appreciates my Gemini sloth images. The current AI hypo AI hype cycle peak is eminent. Um, this is a graph that I've actually used in a whole bunch of different context. I'll explain it really quickly, but it's the the red roller coaster ride is the classic hype cycle. This is something that analysts use all the

time. It is actually a demonstrable pattern where wa we're super excited about this tech. Oh, it's actually mostly spoken mirrors and then maybe like you know it it over time as people actually figure out how to use it, it builds durability. The underlying piece that's in the colors and it's weird how these graphs just fit together. I I've never like I reverse image search it. I'm the

first person to make this combo graph mashup I'm pretty sure, but they fit together like a glove because the bottom one is the software development life cycle. That's this has been around since like the 60s where it's like the people that are first are these innovators. They're the the pioneers, the early the cutting edge people. Then you get the early adopters who are not like pioneers. They

don't have like the pain tolerance to actually live on the cutting edge, but they're like really dialed into what's going on. They like are enthusiasts. These are like the people who would like, you know, the the per people who wanted to buy the Apple 2 cuz you could like mess around with a little bit. Um and then there's the chasm, which is what happens when the tech

actually doesn't have real utility value. So, metaverse chasm. Uh, they just wrote that whole thing off because it didn't work. Um, crypto for the most part chasm because it doesn't provide real utility value. It's just a way to do scams. Um, there's other like crypto enthusiasts will argue with me, but the vast majority of crypto as it was used in that period where it was chasming like

2021 to 2023 was just pump and dumps. Not cool. Fraud. Uh, but AI is definitely gonna make it. Like that's my point. It's not going back. There's no world where we're going to be like, "Oh, that AI thing. Remember that? Oh, those were the days." uh and economic assumptions that are based into the levels of investment and valuations for the companies that are leading the AI hypewave

are extremely unsustainable and not very well thought through. They didn't realize how much RAM they would need to build all the data centers they thought they would build that there wasn't enough of it in the world. That was not something that anyone foresaw when they were like, "Oh, we're just going to build all these data centers everywhere to run all the they didn't think about the need

to electrify the servers and chips in the data centers at the scale that their model said they were going to do when they said, "Well, we're going to just invest a trillion dollars in this over here." Um, I have a friend uh who is in the fossil fuels industry. And real talk, there are data center buildouts right now that are working to export decommissioned gas turbines from

the Czech Republic, ship them over to the US and install these like aging, inefficient, high pollution gas turbines just so they can get their data center project, you know, to be only one year behind schedule because of the power draw that they want isn't there. Um, and because of geopolitical reality, the economy is about to experience a massive energy shock, >> right? The price of oil is

going to go way up. Price of natural gas in the US is capped because we can only export so much of it. There's a bunch of export capacity that's going to come online in 2029 or 2030. We're going to neutralize our prices of gas with Germany, which will make them, it'll be like two to three times more expensive wholesale NATURAL GAS. THIS WHOLE thing is not going

to make it through all that. So the point of that is not to say abandon AI, right? Because we're not going back, but this is very much to me like in my early career being in the dot bubble. Like you know that this is unsustainable and like you can't go on forever. No one wants to call the top because if you're wrong then like people get up

higher than you. And I'm not saying the top is now so I'm not on the hook for that thing. I'm saying it's soon. It's soon. Dice had like the cool like the Chicago thing of like we're lifting the city with screws and reinventing and I was just like man there's another thing that happened to Chicago. I don't know if you're going there. I guess you're not. Okay.

there is going to be an incredible amount of creative destruction in the tech industry over the next three or four years. Maybe starting this year, maybe starting next year. Again, I'm not calling the top today, but it's going to be wild to watch some of these things unwind. And it's going to be actually very difficult and painful for a lot like a lot of people. I'm not

trying to minimize that fact. But what I will say is own on balance. Um, you know, don't love again, as I said before, obviously not a lot of love for crypto in my heart, but we're all going to make it is a great slogan because I do think we're all going to make it. And here's why. there's going to be some kind of crazy meltdown. Like the

financial impact of this is going to be rough, right? Like companies that have like high valuations are going to have low valuations or go out of business and and there already have been companies that had like billion dollar plus valuations like what was it was like superhum.ai or something like they basically were like mechanical turking everything got caught and then went bankrupt. Um so there's going to

be this meltdown but the meltdown is not uh a total crash and we go back to the old world. What it actually is is it's a return to sanity. It is marktomarket pricing where the cost of the thing actually allows the person selling it to you to turn a profit. Like at the fundamental level, that and uh people letting go of the idea that we're going to

create super intelligence that obates all these problems and means we don't have to worry about it anymore. Which if you believe that, it is your right to believe that, but I think that's foolish. Um, sanity will mean we've got this technology that obviously has utility. It obviously can be used to make our lives better. It obviously can help many people. Um, it also can obviously be used

for like evil nefarious purposes which we need to figure out safeguards for. But that's not like making the model impossible. It's like I can't like, HEY, I MADE THIS AWESOME NEW chainsaw and it can't cut anybody's throat. No man, you made a chainsaw. What you need is laws that say don't cut throats. Um, and on the other side of all this is a better tomorrow for everybody.

Um, for the most part. Um, and and what I'll say is the the spin cycle will be rough. Like as we go through that, it's going to be very difficult because capital is going to get reallocated away from speculation. Now, I actually think that could benefit a lot of people in the room because the reason there's no budget for building Drupal sites right now is because we

haven't yet, although the Dre note gave us a really good uh way. We haven't yet figured out how to position building a Drupal site as being an AI project. And that's because of dumb dumb tople decision-m that just won't greenlight anything that doesn't have the keyword in it. That's part of the hype cycle. So that's going to change because people will be like, oh, did we make

any did we get any return on all that AI investment? We did. We didn't. Huh? Well, we should be more cautious about when we invest in AI and a little bit more thoughtful about how we do it. And there's, you know, there's going to be some creative destruction of value, but like the underlying thing of investing in technology to improve your business, to achieve your mission, to

to to accomplish what you set out to do isn't isn't. So like we're going to see reallocation of investment into more productive places. That could be an immediate benefit to us actually after the chaos uh uh sorts out. the model makers prices will go up up up um because so many of them are selling their product at a loss right now. Um and you're they're like I

think the uh thesis was we're going to build data centers a way out of it and it's not going to work. Um maybe somebody will come in and like like do sort of what DeepS did which was like the hip-hop PHP if anybody remembers that. It's like history all over again. Like Facebook literally confronted this dilemma in the late as where they're like p Facebook's written on

PHP. It's growing like a rocket. We want to go worldwide. If we are going to do that, we're going to have to build 32 new data centers and that will cost us I don't know what at the time it was billions of dollars. And somebody was like, well, could we just make PHP cost half as much compute? Like could we make it twice as fast? And they

were like, could you do that? And they're like, we could try and that was hip-hop PHP. that was PHP7. That's why there was that huge performance boost in how fast PHP apps could work. And it was all driven from Facebook dumping a lot less than a billion dollars into a dev team that could figure out how to pull that off. So there's maybe some improvements of that

nature to happen in the world of LLMs. But I do like the the gap between what it takes to run a scalable business and just to run the inference compute is is out of whack right now. So if you are building something, don't hold back. Like you need to be in the arena to figure this out. But if you're thinking about banking a spreadsheet about how it's

going to work out, like run that model back with like a 2x, 5x, and 10x increase on your inference costs just to be on the safe side. Um, the last thing I'll say about this is the financial calamity that is going to come out of this. And frankly, I'm I'm not going to talk politics, but like a lot of um private credit activity in the United States

is funded by sovereign wealth funds of Middle East countries that are no longer going to have the money to invest in US tech companies because they're going to need to rebuild their own critical infrastructure. Private equity and private credit are going to need liquidity. they're going to start selling companies that they've been holding on to for a long time because they didn't like what the value was

and didn't want to admit they probably were holding on to a loss as a project. So, I would look across tech, but also in our ecosystem, companies owned by private equity are likely to get squeezed for cash or sold and then run in a very different manner. Um, draw your own conclusions from that. It's going to happen. Um, and the Drupal economy, the Drupal project will persist,

the Drupal community will persist, and I p I predict the Drupal economy will evolve through this. Um, in the dot times that, you know, I came up in, one really telling thing is all the gamechanging companies that that turned into big tech and like big tech has lots of problems. Let's just take it as like who made a company that could really grow versus who didn't. All

the gamechanging companies really had their growth after the pop. Google was around in like 1999, but they went public in 2004 and they grew hugely after that. Facebook didn't even exist until till 2004 2005. Um, Amazon was like a pitily, it wasn't pitily, it was like a big online book seller, but it was just an online book seller until like well after the market correction occurred. And

that's what I mean by the sort of creative destruction, right? It's a economic term that has downsides and it's a pathy way and sometimes I'm writing off the fact people lose their jobs and all this other bad stuff happens. But the reality is when there's a bunch of stuff based on nonreal assumptions about things that aren't going to happen in the real world, money gets put to

bad use. And when you get back down to a level and you understand what's going on and the money gets put to better uses, then more sustainable, productive, and truly innovative things emerge. And so if you or we or us can be in the position of kind of like not on that hype cycle roller coaster, but following and helping to figure out what that real world adoption

is as it becomes clearer and clearer how to apply this technology to provide real value. When that chasm comes and the hype cycle collapses, we're actually going to be in a really strong position to grow through But you have to be in the arena to do that. You can't like you if you try to lay I'm just going to wait it out until the correction comes, you're

screwed because you're going to be way too far behind everybody else. Um so it's this really kind of like tough balancing act to manage of a lot of people like the information environment around AI is just like so untrustworthy. Like so many people are just getting paid to shill for right now. Like this is documented in well sourced reporting like $500,000 a year to streamers to just

go on there and be like I'm DOING CRAZY STUFF WITH insert your model name here. Um you can't trust the information environment the secondary information environment around any of this stuff right now. All you need to have your own firhand experience, your own firstparty data, or you need to have people like look left, look right that you know are legit who can tell you what is and

isn't real because if you trust what you read online, you're being fed propaganda actually at this point. Um, so get in the arena now. It's fun. Um, the last spicy take that I'll leave and then I'll close is Drupal Con might have to evolve too. Um, this is just a little curveball at the end, but I asked my research assistant Claude to go read through all the

Drupal Association's financial disclosures and stack out like how much is spent versus how much is made in doing Drupal cons. And for a while, I I remember this being told this a lot in the early days. Drupalcon Drupal Con was this big money-making thing for the Drupal association. I don't think it's true anymore. Um, I think we're doing this event. I don't want to stop doing Drupal

cons, don't get me wrong. This is not this not the point of this slide. The point is I think we need to have our own mark tomarket of what's the purpose of this gathering and it's not really a trade show, right? Again, people aren't coming to Drupal Con to buy Drupal. That's just not really a dynamic that happens. I think everybody on some level recognizes that. And

we could put the same or maybe just like a little bit less budget and energy towards an event that was really just about education and community. And it could still be professional professional education, career development. Like that should be that could be the focus. But there's so much so many calories get burned. Like I like doing the trade show floor thing. I love having our guys out

there doing demos. I like having a place to meet people. But man, it's a lot that goes into pulling all that off. And I don't think it's doing a lot to benefit us. Like I really like what the Drupal Association has done recently that our boy um Alex Mareno has been talking about of like going to other conferences and showing up there. Like I'm still like it's

that's a it's not a that's not yet a proven strategy and like going out there to promote the Drupal brand I think is less less effective than going out to promote solutions. But but like getting out the off the island in that way makes a lot of sense to me. Continuing to do this big like sort of I want to have a big celebration. I want to

have a big party. I want to have people come and share their knowledge. I want the community to have a natural rallying point to do contribution sprints and other stuff. I just think there's probably a way to make the economics of it work just for that without a lot of extra effort. And the Drupal Association would probably could turn it back into a money maker. It's my

last spicy take. Just a little spicy amuse at the end. Um all right. So anyway, uh if you want to talk more about this stuff at at our at our booth, you can. Thank you for attending my sponsored content session. Um the web is changing. your operating model has to change with it. We'd love to talk to folks about that. And I will pause now for questions

and also shout out that I did schedule a bath for people to come yell at me if they're mad. Uh, thank >> Question in the back. >> Okay, so Josh >> Yeah. >> I agree with everything you're saying. I think bring up some really good points. And I definitely see that we do need to adapt especially as you broke down addressing agencies like everybody cares a lot

about future tal things we are trying to do. And on the market, folks, there's tremendous opportunities. We just shared that we're putting out there and we need to take the message out and we need to take that out from a lot of agencies that can come on board. I was at marketing earlier. We have one table of people there. We have a we have a large developers

to be able to contribute Wheers to help us bring this message out. But here's the thing. Here's the spicy one for you. Just something to think about when you post your post on LinkedIn. So, how does that help us? Because here's where I want you to think some people and they're also in the So when you viciously attack them through the words you say it hurts their

brand and I know that helps your but when we in the open source community believe in the values of working so I'm ask when we post things, it's not all of us and our businesses or but I do agree that everything you're talking about are absolutely things we >> Um, for the recording, Glenn was saying that he agrees with many of the things in the talk. um

would like to see more contributions from Pantheon, which I can speak to in a second. And then also was asking to be more considerate about what's posted publicly versus con uh conversed in still publicly discoverable but more quiet uh community forums. Is >> Yeah, right here. This is great. >> Yeah. But this is going to get taped and put on YouTube and made public, just so you

know. Um, so the here's my challenge and I've talked with others about this before about the kind of the the problematic aspects of posting LinkedIn about this kind of thing is I'm trying to talk to the Drupal economy and not all of the Drupal economy is in the like relatively close community channels. Um, and so one there's this challenge of like not being able to reach necessarily

the audience that I want because not all of the Drupal economy is deeply plugged into the Drupal community. Um, and then there's just a more sort of I guess paradigmatic aspect of my personality which I recognize is not for everyone and I'm not saying it's right or wrong, but it is what it is. Like it's just like facts are facts and I think talking about stuff openly

is okay. Like when I did the great multi-sight debate thing like John Kennedy was came after me in the hallway afterwards. He was like you're killing this project. Adobee's going to use this. They're going to compete again. And it wasn't none of that happened. Right. The reality is that Drupal is not a big enough deal for the broader public to really care that we're having this conversation

no matter where we're having it. like there's not a competitor out there that's to Drupal as the open source project that's going to use this in a real way. If there's examples of that happening, um I would I will amend my my comment. But like I think that's a fear that people have. There's a fear that like what's going on in the the community and the ecosystem

might be known because we're like we feel bad. We don't have feelings about it. But I don't think there's like a material business risk. And as it pertains to like competition within the I think that there are uh a lot of people are commendably want to maintain like neutrality and friendly relations. Um and I definitely believe in the ethic of co-opetition. Um, and I know many wonderful

people who work for with uh uh competitors. I also know from moving hundreds of accounts from competitors to us that they're not being well served and that that's actually really bad for the whole thing. It's not a good thing to have a company that is squeezing its customers be one of the pillars of the ecosystem. So, and I don't think that there's a uh I again it's

unhelpful to like try to pretend that that isn't part of what's going on. >> do you dialogue with them though? Like you talk to the staff there's a whole new group that's really trying to make a difference. I'm not looking at everything you're trying to do because I know you care about Google. >> Yeah, I want all of us to be successful. I think people just got

to act right. Um, stop charging their customers for the Google crawler indexing their site. It's not a good practice. Yes. >> So, speaking off the island, a different island uh seeing exactly the same uh issues. And one of the things that I was thinking when you showed the graphs is well add 7 to the oration that you got from Yeah. that kind of stuff. And >> I'll

go back. We don't have to look at >> I think I think we've been through a phase where the CMS has been the place where you build websites. That's no more. And what we're finding now is a phase where we have to rethink how we use the CMS. So we're actually at sort of the end of the hype cycle here with with the CMS where we find

out what it's actually useful for in a larger context. >> Yeah, I think that's a great comment. I'm I'm not going to try to play it all back for the recording. You're from a different island, though. What island? >> Uh the Typo 3 ISLAND. >> THE TYPO3 ISLAND. We should establish a trade route. Uh and and I think some already have like you jam and Robert. Yeah,

exactly. That's cool. Uh yes. >> So you were speaking earlier about uh doing other things in Drupal like you you noted next.js which I know you're adding. >> Sure. That's totally self- serving. I I'll own that. But scoreboard just look at the number of sites. Come on guys. >> But how itself is kind of notoriously locked down to its environments. How do you how do you gel

those two things? How do you make those two things a call to action to to add more stuff to our our product offering? At the same time, Pantheon is kind of not letting us if we're Pantheon clients. >> Oh, totally. Um I mean, I would say my advice is self-serving in that I just see the trend lines with Nex.js and we can we can talk about quality.

I mean, there's concerns there, but that's true of every tech. Um, and so that's why we're we're not on the cutting edge of where the web is going. Like I'll own that. Like we're we watched a lot of different things go through the flavor of the month cycle and we're like and then like took us a little while. I think we're we're I would love it if

like our first attempt at front end sites had been a smashing success. It wasn't. So we started over from scratch and rebuilt this version of it which is way better. Um, I think for for the people who really want to be on the cutting edge, like Pantheon is like potentially not the place, right? Because we like we're offering a tradeoff of less responsibility um like less maintenance,

less overhead, less mental load. Uh, but with in in in return for a reduced set of choices so that you can, you know, build on top of whatever that is you're doing. we'll focus on, you know, how you're using the technologies that we do support and like building your practice and your craft up there. But it's always been the case, you know, we had lots of people

who came, used our stuff for a while, and then kind of like migrated their stuff away because they just were chasing a different technology thread. And that's normal and healthy, I think. >> Yes. >> I'm actually wondering if one of the reasons why we're seeing the number of sites um going away additional factor is that there are a lot of sites that were created that maybe needed

to be consolidated under sites that just otherwise needed to go away and now that people are being forced to reckon with it because of what's happening with the Drupal versions because of Drupal 7 end of life now they're forced to look at it when they're not necessarily being as forced to look at it with other >> yeah that's a good question so the one one thought is

the the that is definitely the story in this graph right if you include earlier versions of Drupal in the total number of sites. It's like a lot of those Drupal 7 like that really smooth declining big blue thing of Drupal 7 is just end of life sites that aren't getting are not not going to make the leap to a modern Drupal, right? Um you also asked about

um consolidation being one of the factors and there is an industrywide thing for consolidation. Now I give a version of this talk or gave a version of this talk uh last uh the end of last summer at the decoupled days in New York that was focused specifically on the DXP players like the Psychore uh optimiz they're all like do the same thing and like the the people

on their side like no it's just it's because of consolidation it's it's because they're now running they're using next as the front end and and so forth and that's true and it's like not the whole story cuz like you can go back and see like well in 2023 this site was running on optimizely it's not now. Is it on a decoupled front end? No. Uh is it

um has it been redirected to a different domain? No, it's running uh some other DXP, sometimes WordPress, sometimes Drupal, sometimes a kind of like you know obviously home homegrown something. Um, so, uh, the consolidation effect I think is there's a little bit of like cyclical action that happens in the industry where people kind of let sites go out and then they bring them back in and they

let them go out and they bring them back in. And that's happening at different rates all over the market depending on where um, uh, clients and customers are in their own life cycle. I don't think there's like a a coherent overall market trend towards or away from consolidation. It's just constantly happening in both directions. Um definitely the the overall like the big picture decline because Drupal 7

and Drupal 8 was just a huge leap that not everyone was going to make and I think everybody kind of knew that at the time and and we made our peace with it. So that's why the Drupal 7 version of the graph is like sad sad but it's not it's not really a valid argument at this point that's so long ago. To me the concern is just

the adoption of modern Drupal hitting that plateau. Uh because again if we're not seeing net new builds, there's that's a big revenue driver for the current Drupal economy that's very time and materialsoriented. Um and and it's also like new builds are kind of where you get to play innovation forward. So if there's fewer new builds, it's harder to see the innovation that is demoable, but like building

a real thing with it always pushes it further. So the hopefully we will see I mean again and I I really liked Dre's uh talk earlier today and you know he acknowledges this and it is early days right Drupal CMS and canvas and all this stuff have like some appealing things to them and they may really start to increase adoption I hope they do u but if

they do I think it's it's not going to happen overnight and I still don't think that's going to in any kind of reasonable time frame lead to a bunch of people searching for and buying Drupal services in the near future. I think we have to do this work to diversify our economic model regardless. >> Oh, you're about to wrap it up. I'm >> Well, we're a little

over over time, but if people want to hang out, I'm like, >> yeah, let me get the buff. >> Yeah, one more question. We can do it. Um, yeah, it's not spicy enough for >> All right, fine. We'll do it around a round table or we can chat about afterwards. Thank you all so much for attending my sponsored talk. Um, and think about this. We'll talk about

it at the gala. Let's figure it out together. Uh we're all going to make it.