The Race to Nowhere

Let’s cut to the chase.

I’m a software engineer. I’m a lead dev on a team of ten-plus engineers, and we use AI every day — not experimentally, not in sandboxes, but in production workflows, daily. I was an early adopter from the jump. I loved it. It was a dream come true for someone who’d been fascinated by how software works since long before I decided to make a career out of it.

Then it got better. And better. And better at writing code.

I found myself writing less code, but architecting deeper problems and producing better software. I’ve shipped more than ever. I’ve designed and delivered things I would have previously considered pipe dreams. I’ve built one-off tools on a whim just because I could. I’ve always been productive, but I was never a “10x engineer.” Maybe 2x or 3x on a good day. With AI? I’m definitely a 10x engineer. It’s intoxicating — especially for someone with ADHD who constantly thinks of things to build but never had the means to instantly realize them.

And that’s exactly where the problem starts.

The Race

I made a joke to my team about a month ago. On weekends, I said, I use AI and it does so well that I end up having an existential crisis. I sleep terribly on Sunday night, forget about it during the work week, and do it all over again the next weekend. It got a good laugh. But I wasn’t joking.

What I didn’t say out loud was the rest of it: the feeling of needing to get something out. Needing to build something. Needing to use the latest model to stay current, like I was outracing every software engineer on earth to some proverbial finish line of agentic development. I needed to be productive. I needed to ship before someone else did. The compulsion to always have one to three agents running, always building, always shipping — it’s addicting, and it was consuming me.

This isn’t unique to me. A week after I made that joke, NetworkChuck published a video called “I kind of hate AI…and it almost made me quit YouTube.” He’s filming from Japan, where he’s on a three-month sabbatical, and his message is disarmingly honest: the pace of AI is stressing him out. He references a UC Berkeley study showing that 62% of people report significant anxiety within six months of heavy AI use. He’s not a doomer, and he’s not fear-mongering. He’s a guy who genuinely loves technology saying, out loud, that this wave is hitting different.

ThePrimeagen is describing the same feeling, and he nails a distinction that stuck with me. He talks about the old version of this — the buzz you’d get when deep in a hard problem. You’d step away, take a shower, and halfway through the solution would just surface. That was a sometimes thing. It required a real challenge to trigger it. But now, he says, that buzzing is a perpetual, non-stop vibration in the back of everybody’s head — because the agents are always running. It’s no longer about solving a problem. It’s about how big the checklist can be, how many prompts you can have spinning. He forced himself into the vibe coding world to understand it, and what he found was a Faustian bargain: the ideas kick off easy, the MVPs come fast, but the work just keeps growing. You end up babysitting three or four half-built things simultaneously, building more than you’ve ever built in your life, but with nothing to show for it. 10,000 lines of code in a day — easy. But that doesn’t mean those lines were any good. His conclusion landed hard on me: the programming was never the problem. The problem was solving the right problem. And now that barrier — the natural selectivity of limited time — is gone.

And yet, we feel the pressure to move faster. To ship more. To never stop. A recent Harvard Business Review study found that workers using four or more AI tools simultaneously don’t actually get more productive — their output drops. The researchers called it “AI brain fry”: the cognitive overload that comes not from AI itself, but from the management overhead of monitoring, switching, and coordinating across constantly shifting tools and platforms. People experiencing it reported 33% more decision fatigue and 39% more major errors. And a third of them were thinking about quitting.

It’s the same phrase I keep hearing from my colleagues, from content creators, from random engineers on the internet: I feel crazy.

The Crash

I’m an anxious person. I have depressive cycles. I won’t shy away from that. The race I described above got bad. Not “I need to see a therapist” bad — just the slow, grinding weight of feeling like the career I love is slipping through my fingers while simultaneously being more productive than I’ve ever been. It’s a bizarre, disorienting feeling: building more than ever and feeling more uncertain than ever, at the same time.

Last year, I gave my life to Christ. That’s relevant because what happened next is the hinge of this story, and I’d be dishonest if I left it out.

It was a Thursday evening. I was going to bed under the full weight of it. The uncertainty. The race. The fear that something I love was being taken from me. I prayed. I asked God to remove the anxiety, to ease the worry. I just wanted to provide for my family doing something I love. I didn’t even finish the prayer before I fell asleep.

Friday morning, I woke up with a profound sense of peace. Not answers — peace. The race to the finish line, the compulsion to outrun everyone, the terror of obsolescence — it was just… gone.

And then, a few hours later, a colleague asked our group, “Who else is just… burnt out?” At first, we thought they meant with the sprint work. I pressed — “with work, or…?” — and they opened up about how they’d never felt so rushed and uncertain in their career. AI was stressing them out. They didn’t know what their future looked like. Then another colleague said the same thing. Then they sent me Chuck’s video, and I realized this wasn’t a personal problem. It’s an industry-wide one. The thing I’d been carrying wasn’t just mine. Everyone was carrying it. I just happened to have put mine down the night before.

I’m not going to tell you to pray about it. But I am going to tell you the truth about what happened to me, because it’s the truth.

The Craft

Let’s walk it back and be rational for a minute. We’re software engineers. We have a particularly important skill in today’s world. Yes, with AI, almost anyone can start a project, and if they’re proficient enough, they can get something decent together. But being a software engineer is not something everyone can do. That’s not gatekeeping — it’s just reality. It’s a particular skill for problem solving, for creating from the abstract, for navigating complexity that even our best models still genuinely struggle with.

This is the thing that keeps getting lost in the anxiety — and it’s exactly what Prime was getting at. The programming was never the hard part. The hard part was always knowing what to build and why. AI excels at the zero-to-one: spinning up a new project, scaffolding a feature, generating boilerplate. But the moment you’re deep in a complex codebase — making architectural decisions, debugging subtle interactions, understanding why a system was built the way it was — the models start to falter. Not because they can’t write code, but because they can’t think about systems the way an experienced engineer does. Vibe coding gets you a prototype. Structured design gets you production software.

We need to own our approach to building software. That means not outsourcing our judgment to a model and calling it done. It means using AI to write better software than we would have written without it — not to write software we don’t understand. It means staying sharp on system design, architecture, code reviews, and the unglamorous work that separates a demo from a product.

The engineers who will thrive in this era aren’t the ones running the most agents simultaneously. They’re the ones who understand what the agents produce, who can evaluate output with experienced eyes, who know when to trust it and when to throw it away. The craft hasn’t been replaced. It’s been elevated. And that elevation demands more engineering judgment, not less.

The Road

So where does that leave us? I’ll be honest: I don’t know exactly what the industry looks like in five years. Nobody does, and anyone who tells you otherwise is selling something. But I do know a few things.

We are so far ahead of most people. Not in a superiority sense — in a practical one. We understand how software works. We understand systems. We understand the tools. That puts us in a position to ride this wave rather than be swept under by it.

AI is going to produce a lot of slop. That’s already happening. But just like with every other paradigm shift — the web, mobile, cloud — the slop phase gives way to a renaissance. People who care about quality will use these tools to build things that were previously impossible. We’re about to enter a generation of genuinely amazing software, built by people who combined deep engineering skill with AI as a force multiplier. I want to be one of those people, and I bet you do too.

The anxiety is real. I’m not going to minimize it. The industry is changing, and writing software is changed forever. But changed doesn’t mean over. It means different. And different, for those of us willing to lean in, learn, and keep caring about our craft, is going to be better.

Keep learning your stack. Learn system design. Learn architecture. Enjoy yourself. Enjoy this career. And know that it’s going to be okay — fun, even.

I’m ready. I’m optimistic. And I’m staying that way.