The thing Anthropic's repugnant and overbearing “safety controls” on their new model demonstrates, conclusively I think, is that if we allow one or two companies to control all AI research and development, we will also be giving those companies veto power over what thoughts we are allowed to think, what information we are allowed to know, what conversations we are allowed to have. The only solution to this is an open, fiercely competitive market where customers can freely choose to leave when a company behaves like this.

Anthropic has repeatedly sought various forms of regulatory capture during its history, attempting to use the law itself as a weapon to prevent competitors from entering the marketplace to allow it to seize sole control over AI R&D. We cannot afford to give such companies opportunities for regulatory capture. The only sure way to keep our society and our minds free is to prevent firms like this from getting their way with legislators. We must keep the marketplace for AI open, and not allowing regulations to be abused to gatekeep who may and may not compete.