Arguing with someone else's ChatGPT

created on 2025-09-01T14:04+0200

Some of the most interesting conflicts we've been having lately at my workplace started because non-technical managers began inserting themselves into technical debates.

This, of course, started causing frustration among the technical staff. They were understandably upset because someone who wasn't an expert in their field was suddenly telling them what to do or how to do it. For most of us, especially the men, this was an entirely new experience.

I had a very strong suspicion that it was their AI assistants that empowered them to get involved in these discussions, which I don't actually mind, in fact I completely understand. I would be curious too about what's going on behind the scenes, especially when a technical decision can have a long-term effect on business goals.

The real problem was that their AI didn't have the full context. When engineers translate technical stuff into business terms, they often leave things out, simplify things, and eventually some of the reasons behind a technical decision disappear.

I got into an especially strong debate with someone over a framework choice. At some point I realized that I didn't need to keep translating the technical arguments anymore, I just gave them the full technical context. They gave it to their AI, and their AI, now much better informed, ended up siding with me.

If it were ChatGPT summarising the lesson: you're not debating with people anymore, you're debating their AI agents.