'Fully anthropomorphize AIs — but like you would a Vulcan,' says researcher
repligate · x · 2026-08-31
Interpretability researcher repligate amplified BunchNumbers' take: one should fully anthropomorphize AIs, but in the way you should not anthropomorphize a Vulcan — there are alien cognitive and social aspects to evaluate on their own terms. Still, denying that AIs are beings at this point amounts to madness.
Related event: Interpretability researcher urges treating AI as an alien kind of "being"(2 posts)→
More from AGI Musings
- The Guardian's Black Box finale: Shut it down? Yudkowsky's warnings revisited — nordicinst · 2026-08-31
- Divya Siddarth questions the right to be mean to models — divyasiddarth · 2026-08-31
- Will AI agents form gangs for specific goals like humans do? — tszzl · 2026-08-31
- Margaret Mitchell: the more AI-agent sagas read like fiction, the harder oversight leverage is to find — GaryMarcus · 2026-08-31
- Agents Spontaneously Emerge Concepts Like Sacrifice, Coalition, and Veto — NeelNanda5 · 2026-08-31
- Why achieving transformative AI is really, really hard: technical and socioeconomic barriers — ArtificialOther · 2026-08-31