The confession
I've been working a lot more lately. Way more than I used to. I've had a pretty solid work life balance in the past that I kind of prided myself on. After 7pm was my time. My kid was in bed (or should have been), Slack was shut down, and the client work that was on my mind took a back seat. But something changed, or maybe several things all at the same time kind of. But the first thing to change was my desire to be a part of something bigger than myself. It's a cliche or whatever, but every year that goes by I see my son become more and more enthralled with the world around him. He notices shit, remembers that time I said something off-hand. Point being, I want to make sure that he sees how important it is to be a part of something bigger than himself in a world that's becoming more isolated day after day.
There's a million different variations of isolationism that are taking hold in the world today, and the one I'm going to be talking about is probably the least impactful of all of them. But it's the one right in front of me. The one I've been a part of for more than 15 years, and it's the isolation of the web. It feels like an oxymoron to say, I mean, the web was built to connect the world. There was some god damn inspiring stuff that happened between the days of Ada Lovelace and the .com boom in the 90s. Stories of nerds coming together to make the component parts of computers more accessible to normal folks. The Tim Berners-Lee's of the world who pushed for a standardized way of sharing and linking documents. Some real populist shit.
When I was coming up in the web world, the community was the thing that made it so appealing. There were so many meetups, conferences, workshops and the like. And they were centered around some of the core pieces of what building on the web are. I remember going to the CSS Dev Conf in Estes Park in 2013 and being blown away by the breadth of knowledge being shared. The whole conference was focused on CSS. Chris Coyier gave the opening keynote, there were a couple talks on Sass and Compass that were so cutting edge and cool. Tab Atkins even spoke on the future of CSS. It was so rad to see a group of folks I looked up to coming together to talk about what I thought was the best part of web development at the time.
It feels like the COVID days really fucked up the momentum that tech conferences like this had, and I imagine it was a nightmare trying to manage one of these things anyway. Skip a year of funding or ticket sales or sponsors or whatever, and then pair that with the uncertainty of a conference ever happening again, and I think the switch just kind of flipped off.
I read a book a long time ago called "The Sixth Extinction" by Elizabeth Kolbert that outlined the five previous mass extinction level events that had occurred over the past half billion years, and then explored and identified that we were the catalyst in the current ongoing sixth extinction. I think web development has gone through several extinction level events too. And the latest one we're witnessing is AI.
It's tough to see it from the inside. It's even tougher to see it when you're so blinded by the productivity gains you’re seeing in your day to day work. But the deeper I dive into it, the clearer it becomes:
The thing I loved about this industry is dying, and I'm watching it happen from the inside.
The Reason We Found Each Other
Alright. You're into my line of thought at this point, so follow me here.
Why did conferences like CSS Dev Conf exist in the first place? They happened because CSS was hard. If you lived through those days you'll remember that floats were annoying as hell, vertically centering a div was a meme, and figuring out how to do responsive design was a totally foreign concept (if you didn't live through those days consider yourself very lucky). We needed these communities so that we could leverage someone else's hard won knowledge. The struggle is what created the community.
We have way less of that now.
Things like JSConf, and CSSConf don't really exist anymore. That domain knowledge has effectively been abstracted away by the ability to just let Claude do it for you, without understanding the fundamentals. Take it a step further and look at StackOverflow. I can't tell you how many times I'd leveraged SO to make it look like I knew what I was talking about early on in my career. They went from 200,000 new questions a month to under 50,000 in three years. Nobody's asking questions (at least publicly) anymore. And that's a fucking problem.
The struggle of sifting through several Stack Overflow threads, trying one piece of code, mashing it up with another from a comment that got fewer upvotes, and seeing how that affected your specific problem was a pretty amazing way to learn lots of different concepts. And the Stack Overflow Gods that provided extremely detailed responses that linked out to MDN docs, or the spec itself were better teachers than half the ones I had getting my bachelors in "Web Development and Interactive Media" in 2010.
These cornerstone pieces of the web were some the first casualties in the web's "sixth extinction". The next domino to fall, which we're already seeing, is the death of open source contributions.
The entire open source contribution model is changing. There are so many selfless developers out there who built, or are building open source tools to give back to the same community that I came up loving, and are getting absolutely inundated with AI slop contributions because people see those GitHub "contributed to" lists as some kind of status symbol. I generalize here obviously, but it's breaking so many good developers trying to sift through all this garbage that some open source projects like Ghostty and tldraw are having to write aggressive AI usage policies, or are locking the doors completely to prevent people from spamming them with AI slop PRs.
It gets even more horrifying when you start to realize how truly immense the loss of jobs is going to get. Look at companies like Block, Oracle, Amazon and pretty much every other big name. They're laying off workers in droves after posting record profits. I've lived through several of these extinction level events in web development but this one I think has had the most staggering number of layoffs that can be directly tied to it. People in these positions, were literally building the tech the higher ups were going to use to replace them. They were building their own coffins.
It leads us to the last casualty in this extinction that I'll talk about, and that's the junior developers. Junior developers were having a miserable time finding a job before the AI boom. The industry was doing everything it could to snatch up every senior dev on the block to avoid having to train up the juniors. That created a talent pool of so many people that remains untapped. I can't fathom how much potential here got pushed out and had to find work in a different industry. It's a shame because there was so much wonder around the industry when I first started, and seemingly the 7 month bootcamps really started to suck that wonder out of these juniors, with AI being the final straw that's going to make it even more difficult for junior devs to break into the industry.
It's depressing to realize, but, the thing that was supposed to help developers write better code is destroying the ecosystem. It was supposed to allow us to focus on bigger issues, and instead is effectively locking down open source projects, making it easier for soulless execs to justify laying off massive swathes of their workforce, and is making it harder for meaningful progress to be made in the effort to bring new, passionate talent into the industry.
The Callouses
I've got a lot of these.
It's hard to pick one that really illustrates the point. I could tell you about the time I built a multi stream gaming platform that took raw video feeds from Esports competitions, used ffmpeg to encode them, live stream them, and allow users to switch between feeds as the tournament was going on to see different perspectives. The sheer amount of AWS hot garbage documentation I had to read to try to understand how the hell to use MediaEncoder to get the bitrate right so six streams were the right quality. The all nighters that we pulled using test feeds from six Xbox's connected to a server to make sure the demo we were running the next day would work just long enough to sell the product.
Or I could go into depth as to why the partnership between two really close friends who helped me build House of Giants fell apart. These were people I'd had amazing relationships with. I hired one of them at my first agency. We had this grand vision for how we'd run House of Giants. We modeled ourselves as an anti-agency. A buck against the status quo because we absolutely despised how the agencies we worked at in the past were run. Out of touch executives and leadership who would clutch their pearls harder than the clam they stole them from. And it was great for a while, but then the stakes were raised. Our ambitions diverged, the way we wanted to run things, chase work (or not) became an argument rather than a conversation. How we went about learning and expanding our technical knowledge and skillsets turned into arguments about who's job it was to do Task X over task Y.
The struggles in these moments..those are the things that teach you the most. The ambitions of the project, your uncertainty about how the fuck you're going to execute on the idea. The panic scrolling through Stack Overflow. The unbearable contemplation of how you stop being business partners with friends. That's where the growth happens. It feels paradoxical, especially in the moment, that the struggle during these times is what would end up making you more efficient in the long run. And I think a lot of shops, and even gigantic corporations (looking at you GitHub, with your newly acquired 90% uptime milestone) are going to face the wrath of this in short order.
I'm not afraid that AI exists. I do client work, and I'm doing that client work faster than I ever have in the past, while simultaneously building more than I've ever built on the side. Taking side projects from an idea I've had for years to production in a week. The tools are getting better, and they're helping me do more.
What I'm worried about is that in five years, the people making the architecture decisions won't have the scar tissue to know what to do when it breaks. They'll patch it with another AI slop generated set of changes that start with "Great catch! We should get that fixed" coming back from their agent.
I'm worried about an entire generation of developers who are going to build without callouses like these to inform their decisions. That judgment layer, the thing that separates "I can build this" from "should I build this?", is completely disappearing.
The Paradox of Ease
I keep coming back to something I've written about before. Being comfortable with being uncomfortable. I think that a lot of humanity is lost when we avoid the struggle. The more comfortable we are with being comfortable, the harder it is for us to push ourselves to do something difficult.
This is another thing I want my son to know about me, that I wasn't afraid to do something I was scared of. I'm the type of person that puts the spider in your basement in a cup, and takes it outside instead of smashing it. I used to collect the rollie pollies in my parents garage and take them to the grass. See a worm on the sidewalk after it rains? I'll pick you up and put you back in the mud dude, I got you. In total opposition to those convictions I went hunting for the first time last October and killed an elk. Field dressed it (with the help of some very good friends), and packed it out. I had a ton of internal conflict about that. But at the end of the day I wanted to show my kid that there are times in life you have to make a decision that might go against everything you think you're about.
Now when I ask my kid what his favorite food is he says "elk steak!", and that gives me some sort of weird primal satisfaction knowing what I went through to provide that experience for him. That's a part of his life now, something that was so emotionally taxing for me, will live as a positive experience for him.
We keep trying to move so fast to build tools that eliminate all the friction in our lives and then sit wondering why nothing feels like it matters anymore. We stop getting together to talk about those hard fought battles, and we scream into the void that is our AI agent, and all that does is reflect back to you all the bullshit you want to hear beneath the subtext of your prompt. I build a ton of stuff with a swarm of AI agents and I still feel this way. That says something.