It's a spoiler for who ?

Apparently it’s all about the model. Just switched to GPT 5.1 Codex and it became smarter right away. But I lost my privacy. So it’s either private and dumb or non-private and smart.

Personally, I’ve been liking Warp for what it’s worth.

Yea, one of those things where I'll wait a year and it will be next level.

OpenClaw is absolute shit. Were clearly early. If I came up with ways to rate an LLM - I’d give OpenClaw zeros across the board.

My experience too. I just gave up. Models make several mistakes and you have to structure code all along. Not that expected magic or I did something wrong.