Claude code's system prompt is filled with irrelevant stuff about how CC works, so that the agent can help the user set it up. And there's no way to disable all the extra stuff, AFAIK.
How is this different from the speculative decoding that we had before?
You could pair a big and small model like qwen 32b with qwen 4b and had that same dynamic of the small model generating tokens and the big one "certifiying" them.
The blog says something about re-using the big model's data?
There's https://github.com/skrabe/lobotomized-claude-code , which strips many of those, but I'm not sure if it is "legal" to use.