> Why pay an expensive architect to design your new office building, when AI will do it for peanuts?
Will it? AI is getting good at some parts of programming because of RLVR. You can test architectural designs automatically to some extent but not entirely, because people tend to want unique buildings that stand out (if it weren't the case architects would have already become a niche profession due to everyone using prefabs all the time). At some point an architectural design has to be built and you can't currently simulate real building sites at high speed inside a datacenter. This use case feels marginal.
There's going to be a lot of cases like this. The safe jobs are ones where there's little training data available online, the job has a large component of unarticulated experience or intuition, and where you can't verify purely in software whether the work artifact is correct or not.
> people tend to want unique buildings that stand out
Just tell the LLM that you want a unique design. I've found LLMs to respond well to requests for "originality," at least in poetry, prose, and coding. No reason that can't do that in architecture as well.
> At some point an architectural design has to be built and you can't currently simulate real building sites at high speed inside a datacenter.
First of all, you can simulate a building site, or any physical environment. We've been doing that for years, even in games. AI companies are working towards a "world model" for precisely that reason. Second of all, even without a physical simulation, the laws of physics are deterministic and easy for an LLM to understand.
> The safe jobs are ones where there's little training data available online,
These cases are "safe" only in relative terms. Lack of easily-available training data is friction but not insurmountable. AI companies have bet big and they have a strong incentive to find and use appropriate training data.