X Freeze(@XFreeze):New proposal to Anthropic: Whenever Anthropic’s paranoid safety system decides a prompt “feels slightly dangerous” or its model is “too powerful” for users, automatically route the request to a more powerful open-source model 😂 Seems fair, right? Pay Elon the inference cost, keep the markup, and let users actually get what they paid for

2026.07.17 10:49

New proposal to Anthropic: Whenever Anthropic’s paranoid safety system decides a prompt “feels slightly dangerous” or its model is “too powerful” for users, automatically route the request to a more powerful open-source model 😂 Seems fair, right? Pay Elon the inference cost, keep the markup, and let users actually get what they paid for

187

コミュニティへ転送

人気のあるユーザー