No observation of a model's behaviour can establish that it is aligned, because a model capable enough to be dangerous is capable enough to produce whatever behaviour it is being watched for.

Added to korrents 7 Sept 2026 · How quotes work · Something wrong? Tell us

Where this collides

• Cannot both be true

  Anthropic now uses mechanistic interpretability techniques to audit new models for deception and scheming before releasing them.

  Amodei reports that Anthropic audits new models for deception and scheming before release. Yudkowsky and Shlegeris hold that no observation of a model behaviour can establish that it is aligned, because a model capable enough to matter can choose its visible outputs. An audit is an observation, and one of these claims says it cannot do the job it is being used for.

Shared subject first, then similar wording. They may agree or disagree.

• A human being is not an AGI: we lack a huge amount of knowledge and rely on continual learning instead, so continual learning is what superintelligence should mean. ISIlya Sutskever25 Nov 2025AI alignment
• A model that looks smarter and better-behaved while getting better at hiding unwanted actions is the scary combination for AI safety. ZMZvi Mowshowitz9 Sept 2026AI alignment
• A player coach manages a smaller team while building together with it, rather than sitting in an ivory tower doing performance reviews and alignment meetings. RGRon Goldin9 Jun 2026AI alignment
• A superintelligence needs no consciousness, emotions or malice to be dangerous; a goal system slightly misaligned with ours is enough. ÉTÉmile P. Torres24 Oct 2017AI alignment
• A validated theory of intelligence may be necessary for achieving genuine AI alignment. KKKevin Kelly17 Aug 2026AI alignment

On the map

Open the map on its own page