Back in January, on Oxide and Friends, a tech podcast where engineers get into it about the industry, Bryan Cantrill was trying to describe a feeling. Not burnout, not fear exactly. Something more specific.
Here's the setup: over the past couple of years, AI tools have gotten scarily good at writing code. Not just autocomplete, but the kind of work where you describe what you want in plain language and the system builds it. Whole features, sometimes whole applications, assembled by an AI agent while you watch. People are calling it agentic coding, or "vibe coding" when they're being playful about it. And it works. Faster than most people expected.
So Cantrill was trying to name the feeling of being genuinely productive with these tools and realizing the productivity itself was doing something to you. Something subtractive. Adam Leventhal, another engineer on the show, suggested a name: Deep Blue.
And apparently the room just knew.
The reference is to IBM's Deep Blue, the chess computer that beat world champion Garry Kasparov in 1997. It was the first time a machine had beaten the best living chess player under tournament conditions, and it landed like a cultural earthquake. The name isn't an accident. It's not "Skynet." It's not apocalyptic. It's the moment a system does the thing you were good at, does it fine, and you have to figure out what that means for the part of you that was built around being good at it. Kasparov didn't die. He just had to become something else.
That feeling had been floating around for a while. Scattered across Hacker News threads (a developer forum where the existential stuff tends to surface between the technical arguments) and half-finished blog posts. Steve Yegge, a veteran engineer whose blog has been picking fights with industry consensus for almost twenty years, wrote about it in early February, calling AI tools "vampires," which is a different metaphor but the same flinch. Simon Willison, a developer and one of the more thoughtful voices writing about AI tools right now, picked up the name "Deep Blue" a few days later on his blog and suddenly the conversation had a handle.
That's what a good name does. Before it, you're performing the feeling every time you try to talk about it. After it, you can just point.
Here's the thing that makes talking about Deep Blue tricky. The people feeling it most are the same people who advocated for the tools. They're not AI skeptics having a vindication moment. They're practitioners who leaned in, built workflows, shipped things faster, and are now saying "this is costing me something I didn't expect to lose."
That's a vulnerable place to speak from. Because the anti-AI crowd is right there, and they are not going to be careful with the nuance. "Deep Blue" becomes "see, even the fans admit it's bad." Your honest reckoning becomes someone else's ammunition.
So there's this extra weight on the conversation. It's not just "how do I process this feeling," it's "how do I process this feeling without handing a talking point to people who were never engaging with the same question." Those are two really different problems wearing the same clothes.
Simon's blog post mentions that chess players and Go players went through this more than a decade ago and came out stronger. And that's true. Chess is arguably more popular now than before Kasparov's loss. The international chess federation's active player count nearly doubled between 2009 and 2014. Chess.com runs millions of games a day. The "chess is dead" people were wrong.
But here's where the analogy gets comfortable in a way that should make you suspicious. Deep Blue didn't take any jobs. Chess was already so niche that the few people directly affected just kept playing. And the audience still wanted to watch people play. That's the key part. Chess has fans. Fans who care that a person is doing it.
Software has users. Users want the thing to work. Nobody is rooting for the developer.
The optimistic version of what comes next is genuinely appealing. More people building smaller, weirder tools. One or two creators making something like Scrivener, a writing app beloved by novelists, or Reaper, a music production tool with a cult following. Software with a specific point of view, for a specific group of people who actually want it. That future sounds good. It might even be real. It might even employ just as many people, just differently. And the same tools reshaping the landscape might help those people figure out where they land next. They might even prefer it.
But you don't know that when the tie gets cut and you're mid-air from the trebuchet. "You'll probably be fine" is a thing people say to you from the ground.
Because Deep Blue, the feeling, isn't really about the future. It's about Tuesday. It's about the mass and inertia of a life that's already shaped around certain assumptions. Your mortgage doesn't care about the emerging landscape of indie software. Your identity as someone who solves hard technical problems doesn't quietly rearrange itself because a blog post told you to think of it as an opportunity.
The people naming this feeling aren't failing to see the upside. They can see it fine. That's almost what makes it worse. You can hold "this might lead somewhere genuinely good" and "I don't know what I am right now" in the same hand. One doesn't cancel the other. They just sit there together, being true at the same time.
So yeah. Naming it was the right move. "Deep Blue" is a good name. It's honest about what it references. It doesn't flinch. What happens after the naming is less clear. Not in a way that resolves into hope or grief, necessarily. Just in a way that's open.
I keep thinking about Kasparov. Not the loss. The part after. He did keep playing. He also became one of the loudest voices against Putin's authoritarian grip on Russia, got arrested at pro-democracy protests, and did a bunch of stuff that had nothing to do with chess. I don't know if that's "coming out stronger." I don't know what it is. But he was still in the room.
What's the thing you're protecting about how you work right now? Not the job, the thing underneath it. I'd genuinely like to know.
-
References:
The podcast in question, with timestamp (47:15 on): https://www.youtube.com/watch?v=lVDhQMiAbR8&t=2835s
Steve Yegge's AI vampire article: https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163
Simon Willison's post on the coining: https://simonwillison.net/2026/Feb/15/deep-blue/