I am curious to see where we are in a few years. For companies who actually did replace lots of devs with AI, did their features continue to work well? Or did they reverse course and re-hire talented devs to clean up their mess? /END

If your code has 100,000 machine-generated tests, is the AI also fixing those tests as the functionality changes, or have you now made a nightmare for yourself? Degradation of a codebase may take years before your users start to smell it. 6/n

Sure, there will be examples of small teams that whip out new products quickly, but can those systems scale and gain new features over time? Do bugs increase and grow more difficult to fix? Do we see the usual symptoms of a codebase which grew too fast without regular refactoring? 5/n

People often act as though AI-generated code is something that happens just once. "Look, AI wrote 100 tests, amazing!" There's a huge blind spot where people ignore that a working system is a living thing which constantly evolves over time. I want to see how AI-dependent teams deal with that. 4/n

My constant complaint about AI is that it adds more layers without refactoring the existing base. We keep slopping on more and more, making things increasingly complex. Just like a team of enthusiastic new hires, who can write tons of code super fast, and eventually make a system unmaintainable. 3/n

One possibility is that they want to reduce headcount and AI is just a good cover story. It's also possible that management really believes the hype, and they're shooting themselves in the foot in the long run. 2/n