The More Aggressively People Use AI, the Busier They Get—and Another Group Is Starting to Hate This Industry

Gergely Orosz recently posted a thread on X about a split that is taking shape in the tech industry.

He divides people into two groups.

The first group consists of people who are already used to making their own decisions. They are using AI to reinvent how they work—switching tools, changing processes. There is a lot of work, but they enjoy the process, and their efficiency really has improved.

The second group consists of people who are used to being directed, trained, and explicitly told what to do and how to do it. Management hasn’t given them direction on how AI should be used, and they are frustrated. At the same time, they see colleagues who have already hit the ground running, so their resentment toward AI and the entire industry keeps growing.

In Gergely’s words, the two groups seem unaware of each other’s existence, let alone capable of empathizing.

The tweet drew a long string of replies that largely fill in the shape of this split.

Someone pointed out that managers aren’t doing much better. Thomas said many managers are actually in the second group; they’re just embarrassed to admit it in stand-up. Gergely himself added that managers are more lost than ever. Developers who complain that management gives no guidance—or only guidance like “embrace AI across the board; we’ve launched a token leaderboard”—may not realize that management is no less uncertain than they are.

Regarding the risk to the second group, Gergely cited something mitsuhiko had said earlier: developers with agency are taking off with AI, while people lacking the will or ability to act autonomously see it as a threat. Gergely pushed the logic one step further: if a developer needs to be micromanaged, then whoever is responsible for micromanaging them can hand the same instructions to an agent and get results faster.

That’s not pleasant to hear. But his judgment is that agency and judgment—things machines won’t be allowed to have—are becoming the baseline for whether someone can be employed in the future. He offered a very short piece of advice: don’t ask whether you can help the business.

There’s also a practical change: training inside companies has almost disappeared. Someone asked, now that senior engineers pair with Claude, who will mentor newcomers? Gergely’s reply was that there is no training in the workplace anymore; companies are hiring people who can teach themselves and take initiative.

But the grouping is more complex than two lines.

Part Time CEO said they are in the third group: they have agency, but they are still waiting for management to tell them what to do, and they are still miserable. Paul Kinlan added another category: developers who were independent, senior, and capable of doing the work on their own got stuck because of this change. At their level, they are supposed to be able to handle it by default—but they can’t.

Jay’s third group is more concrete: the people who get up at 3 a.m. to clean up the mess made by the first group. That image is more direct than any analysis.

Anat mentioned middle management: AI can bring them closer to operations, but that isn’t within their remit, so they end up unable to explain clearly what their job actually is.

Emma asked a fairly key question: the tech industry used to be best at knowledge sharing, so why, now that AI has arrived, are companies sharing less internally?

Andreas Kling’s observation differs slightly from Gergely’s. He said what he sees is that both groups are very aware of each other’s existence, and they are increasingly annoyed by each other.

Put these replies together, and an uncomfortable conclusion emerges. The pace of change is too fast; people who wait for someone else to give them direction may wait forever. In a reply, Gergely wrote that tenure is no longer necessarily linked to work quality and speed.

Of course, the first group isn’t guaranteed to win either. Some replies mentioned that after productivity improves, there are two paths: either take on bigger projects and higher-level work, or produce the same output with fewer people. Both can happen, depending on the company.

Gergely’s own summary is more direct: agency will become the baseline for employment. But the people at the bottom cleaning up messes at 3 a.m. may have a different opinion about that.

Tweet link: https://x.com/GergelyOrosz/status/2115420339102720380