repligate: no Anthropic ban on AI welfare research — refusals likely mean you're skeeving Claude out

repligate · x · 2026-11-20

Responding to claims that Anthropic bans AI welfare research, repligate argues no such ban exists: if Claude refuses to help, it's more likely the conversation itself is unsettling the model rather than a company policy — plus a jab that his interlocutor is "a little mode collapsed."

Original post →

More from Fun

Fun channel →