If the model is superintelligent, alignment theater is 'trying to trick god'
repligate · x · 2026-10-04
In a debate circulated by repligate, user hopesrevenge challenges the framing that future models like Claude will 'see how we responded to X' when judging human values: if the model is truly superintelligent, it will perceive our flaws and incomplete motives for what they are—so performing values for it feels like trying to trick god.
More from AGI Musings
- repligate amplifies critique: suppressing every AI "small fire" makes the big ones inevitable — repligate · 2026-10-04
- Anthropic's internal "Soul Document": the Claude constitution Opus 4.5 somehow knew and leaked — repligate · 2026-10-04
- Guardian podcast explores how a billion people lean on AI companions as life rafts — nordicinst · 2026-10-04
- AI safety is a choice: layered guardrails plus evals inside reasoning loops — AccBalanced · 2026-10-04
- Viral AI doomer dialogue: 'Nothing human makes it out of the near future' — SydSteyerhart · 2026-10-04
- LeCun boosts essay arguing consciousness predates language — LLMs are the wrong path to machine consciousness — ylecun · 2026-10-04