Once a spec becomes sufficiently large and detailed and complicated, it becomes very difficult to ensure it is internally consistent. That's why I start every project with a METASPEC.md so that Claude can break up the task of writing SPEC.md into manageable steps.
Everyone knows a philosophy comes before a spec. Claude has to write your applications philosophy first, then you write your spec. But a philosophy is crap without a values statement, so Claude has to actually write that first.
How do you avoid interruptions for permission? dangerously skip permissions, or is there something less nuclear than that? For me I guess the only less nuclear thing I can think of is running on a sacrificial machine. Is there any better way?
Anthropic is big hall monitor energy. Clownish behavior and exaggerations constantly. A holier than thou attitude derived from rationalists at LessWrong.
There is always way too much drama with Anthropic. They could've just called it 4.9, but no, they had to dramatize it to the max. Well, this is what happens. I'll just bypass all the drama and stick to Codex. I also don't want to unnecessarily pay 100x.
"It is unfortunate that a large number of users here are not hackers, not even in an idealistic philosophical sense, and will betray the public good for their own short-term gain. You either unite the world or you divide it."
Now I can continue my vanity project of having Claude iterate on a single spec.md for hours on end. Surely at some point it won't be shit.
https://code.claude.com/docs/en/auto-mode-config
https://news.ycombinator.com/item?id=48511072
"It is unfortunate that a large number of users here are not hackers, not even in an idealistic philosophical sense, and will betray the public good for their own short-term gain. You either unite the world or you divide it."
They are presumably building a KYC passport submission flow as we speak.