I'm seeing another team doing this, with 99% agents, and quality, predictability of delivery and raw performance have taken a massive nosedive.

But even not caring about quality: Speed of development is abysmal. It was never something this team had anyway, they were always slow. So they don't care that they take a year and 20 people to do what another startup in the same space was able to deliver in a few months with a skeleton crew (and perhaps more quality).

To me the experiment has shown the results it should. It's self-serving to developers, and amplifies the good and the bad in them. It's not particularly good to users and not to companies if the developers aren't good themselves.

It's writing nearly all of the code where I work. It also reviews all of the code (I review it as well). You still have to guide it and hold its hand though. I dont write code by hand anymore (and haven't in almost a year). The job is now about managing the bots, understanding the architecture and keep everything aligned.

That’s a norm at my place as well. But it’s not free. We have definitely given up on quality (both code and general stability) and fully understanding what’s going on. People are generally working more as well, because of all the layoffs pressure.

Industry is betting that we don’t need to. I still consider it a bet because Claude Code this scale has become norm just for a year.

It writes all the code but we often have to go through several iterations. It won't work autonomously yet.

An agent writes 99% of the code. Then I fix some parts, re-arrange things, usually takes a few prompts. Then, there's code review, more changes, etc. But I hardly write code nowadays compared to a few years ago.

My whole company (I hate it).

You're right, no one has "an agent" writing 99% of the code. Now people have swarms of agents writing 99% of the code :)

The vast majority of work people do is basically supporting simple CRUD type apps.

CRUD apps have long been a solved problem long before AI, no one should be spending a lot of time dedicating their life to figuring out how to write a better CRUD app.

If your bubble consists of writing critical high performance applications or bleeding edge research, maybe you have a use case for not using AI, but that is not the norm.

Companies are heavily incentivized to do so so my guess would be yes. I also don't write the code by hand almost at all.

That is the norm where I work , heck I’d say 99% seems low

Yeap, one more data point here. Most of code here is AI generated with some handcrafted adjustments.

it's the norm in my bubble