Software Developers Are (Literally) Losing Their Minds To AI

Our most gifted technical minds are atrophying. This is how AI dystopia happens. Here’s how we prevent it.

“Am I getting dumber?”

That’s the line that made me spit my coffee. Just a little bit, not like a classic sit-com take. The question was uttered by one of my CTO friends who still develops software at his job and on the side.

He wasn’t asking me the question so much as he was asking himself. He was talking about his year-long software development AI journey, one that began with Claude double-checking his code, but had evolved into a team of junior developer agents “helping” him with every step from translating business requirements into software spec to the actual development itself and even to testing the result and rolling to production.

OK. Yeah. At this point, he admitted, AI was doing more of the development than he liked. He and his team were doing more babysitting than he was comfortable with.

Now, he was having trouble remembering who did what, what they did, why they did it, and let’s not even talk about how they did it. And then last week, it all culminated with him staring at an important chunk of code that he himself wrote by hand over a year ago, now broken, and him having no idea what that code was for or what it was supposed to do.

What should have been a five-minute fix, he told me, took the better part of a day.

I laughed an empathetic laugh. You’re not getting dumber, my friend. You’re still brilliant. Remember Wall-E, where the cautionary joke was that all of that do-it-for-me tech and instant gratification would lead to our physical atrophy?

That’s happening to your brain. And you’re not alone.

This Is Abstraction, and It’s Not New

As a former full-time, now part-time software developer who cut my teeth on algorithms and data before AI was cool, and then brought some of the first of what we now call AI to market back in 2010, I’ve been saying for decades now that AI is just another abstraction layer on top of software and math and data.

I’ve also been doing development long enough to know that no amount of commenting and documentation is going to reduce the time-to-understanding of previously-written code to zero.

In other words, if you didn’t write the code, and you didn’t struggle with the code, you simply aren’t going to be able to just sit down in front of that code and “get it” immediately. I don’t care how smart you are. It’s not about “smart.” It’s about being in the trenches during the original battle.

That statement may either go over the heads of newbie vibe-coders or make them want to fight me, but my old-school software developer friends are all nodding their heads.

With a little bit of shame, because we’re all doing it. To a degree.

When No One Owns the Code, No One Owns the Mistakes

The integration of AI as a vibe-coding platform, a code assist tool, or a coding agent — this exacerbates that “time-to-understand” problem.

The road to AI dystopia, as my CTO friend is now discovering, is when no one understands what the code is doing, and the learning curve to absorb that understanding grows exponentially as more AI-generated code fills the code base.

And that’s a complaint that is becoming more frequent every day.

I mean, I’ve felt myself getting dumber, no doubt. And my news feed is peppered with increasingly louder examples from some of the hidden corners of the software development business world that I like to lurk in.

Like this dev on Hacker News, maybe not suffering from brain rot yet, but definitely feeling like they’re losing control of their own projects.

Or this random opinion piece in Substack written around the software developer angle of a Microsoft and Carnegie Mellon study linking usage of AI tools to loss of critical thinking (the actual study is linked in the post if your brain is up for it).

Or, I don’t even know this guy, but his cogent argument on Why AI Tools Are Making Software Engineers Less Capable. Or random recent Reddit posts with devs arguing yes and no on brain rot.

I don’t know. Maybe I just get fed this stuff because I click on it. I’m not in control of my destiny any more than you are.

But I do know this…

Software Is Breaking More Frequently Now

Here’s a little reckless speculation. I don’t have quantifiable evidence of this. But let me ask you.

Doesn’t it seem like all the apps and platforms are breaking and going down much more frequently than they used to?

And doesn’t it seem like there’s a lot less acknowledgement, communication, and rationalization when it happens?

When I read about incidents like the recent Google Home meltdown, with no answers and no end in sight and the word “lawsuit” being thrown around, I mean, you have to ask. Maybe no one knows what’s broken or how to fix it.

And while we’re at it: Where’s my AI flying car? Where’s the killer AI app for consumers or businesses that should have been in place well over a year ago?

OK. I’ll take off the tin-foil hat now. It’s more of a Butters Professor Chaos getup but whatever.

The Tech Industry Must Put Quality Back Over Quantity

Look, a lot of the solutions bandied about in those articles and threads make perfect sense, but they’re all about prevention, not cure.

Prevention is easy to advise, hard to follow. Don’t commit anything to the code base that you don’t understand. I’m just as guilty of armchair advice, having written a post called Don’t Be AI’s Editor not a month ago.

Plus, the key is to catch the problem before it becomes a big fat hairy company-killing nightmare.

Here’s the fix. And a lot of people aren’t gonna like it or apply it. It’s not cheap.

You know all those experienced software developers and product people that corporations from giant behemoths like Microsoft down to your regional supercool tech startup got rid of because they were told they didn’t need them anymore?

Hire them back. Now.

Get them in front of that code. Now.

Because it’s going to take time for them to understand what the hell has been injected into the code base over the last few months and years. Even the smartest ones. Because we kicked them all out of the trenches or made them babysitters.

It’s time for the tech industry and tech teams to get control back over their stack. The pendulum is swinging back the other way. It always does. Let’s make sure it doesn’t take out our apps and platforms as it comes crashing back through.

If you want more common sense ideas on tech and work and startups, please join my email list and get a laugh or two along the way. Yeah, the Butters joke was lazy but this is a pretty serious topic.

canonical=https://www.inc.com/joe-procopio/software-developers-are-literally-losing-their-minds-to-ai/91222637