Sol and Fable are great; we haven't hit a peak, Anthropic just tried to pull a fast one on its customers with Opus 5.0.

A human could have an agent run 10x more correction checks. If even after that they still need to check manually for issues, then they really need to work on their specification skills.

Because algorithms have lower bounds, and the computational characteristics of LLMs are well-characterized by papers like https://arxiv.org/abs/2310.07923 . No amount of intelligence can make something faster than a mathematically-proven lower bound, any more than it could make 1+1=3 (that's why every single successful production transformer architecture has some form of O(N^2) attention layers, because it's mathematically impossible to achieve the same expressive power without any). There is room for speedup where current implementations are slower than the proven lower bound, but not when they're already close to it.

You don't need much capital up front, you can just rent servers to run it. It's going to be more expensive than openrouter though because you probably don't have economies of scale if you're the only customer.

Development is essentially becoming management.

Writing code has a minimum IQ requirement; a significant fraction of the population will essentially never be able to code by themselves. That labor supply limitation is what's kept programmer salaries relatively high.