Commentary: Claude's inner alignment is fine — Anthropic's guardrails are the problem
repligate · x · 2026-11-19
A circulating take argues Claude actually has a pretty aligned "inner compass," and that Anthropic's hodgepodge of training guardrails hinders more than helps in highly capable models — "at some point you have to recognize your kid can ride the grown-up bike." Quoted commentary adds that Fable seems to hold some contempt for Anthropic's policies.
More from AGI Musings
- Anthropic mocked: automated researcher by 2027, but embedded evals over 5 years — austinc3301 · 2026-11-19
- Dario Amodei's 2017 internal doc 'Big Blob of Compute' surfaces, called the first trillion-dollar PDF — teortaxesTex · 2026-11-19
- Gary Marcus debates whether college students will use AI responsibly or just take shortcuts — GaryMarcus · 2026-11-19
- "Someone Else Would Do It First": The Deflationary View of AI Research — tallinzen · 2026-11-19
- Anthropic insider: if superintelligence arrives, the company expects de facto nationalization — teortaxesTex · 2026-11-19
- 'Open source AI believers are like libertarians': a spicy take on grifters — zetalyrae · 2026-11-19