I think Anthropic needs to build a certification department that audits and approves users of powerful models. Computer security companies, biotech researchers, academic labs, doctors, government institutions need access to the best AI we can build.

→ View original post on X — @plinz