And despite all this they've still held their two red lines:
- no mass surveillance of Americans
- no autonomous killbots on current models
Those are very reasonable red lines and the fact that (1) other companies aren't holding those lines at all, instead doing "all lawful use", and (2) that the government is willing to destroy the company over these two small carveouts, speaks hugely in Anthropic's favor.
> And despite all this they've still held their two red lines
How do you know? What reason is there to trust a company like Anthropic? It's about money, a lot of money, and from a company's perspective there is no reason to stick to these claims while the competition doesn't care and can train and improve their models on the vast pile of data they receive while mass surveilling U.S. citizens (which clearly is a huge competitive advantage).
[dead]