korrents

On the map

Tap a claim on the ring to put it at the centre.

← Evaluating model welfare is very hard because models are suggestible,…

17 connected korrents · 13 moments from 25 Sept 2015 to 4 Oct 2026.

Everything filed under AI sentience AI sentience Everything filed under LLMs LLMs Everything filed under animal welfare animal welfare Everything filed under benchmarks benchmarks Everything filed under AI alignment AI alignment Everything filed under taste taste Same subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subjectSame subject Read this korrent: Evaluating model welfare is very hard because models are suggestible, good at role-play and try to do what they think is wanted. Evaluating model welfare is very hardbecause models are suggestible, good atrole-play and try to do what they think iswanted. Last stated 11 months ago 27 Dec 2025 NN Neel Nanda — holds since 2025-12-27 — tap for who they are Same subject: Models look far better on evals than they are in the world because researchers, inadvertently, take inspiration from the evals when they build RL environments. — tap to centre the map on it Models look far better on evals thanthey are in the world becauseresearchers, inadvertently, takeinspiration from the evals when theybuild RL environments. Last stated 12 months ago 25 Nov 2025 IS Ilya Sutskever — holds since 2025-11-25 — tap for who they are Same subject: A model is better thought of as a living creature with a personality you must get to know than as a component you specify. — tap to centre the map on it A model is better thought of as aliving creature with a personalityyou must get to know than as acomponent you specify. Last stated 4 months ago 27 Jul 2026 BC Boris Cherny — holds since 2026-07-27 — tap for who they are Same subject: Evaluating a model properly would mean delaying its release, and competitive pressure means no lab will. — tap to centre the map on it Evaluating a model properly wouldmean delaying its release, andcompetitive pressure means no labwill. Last stated 5 months ago 26 Jun 2026 NB Noam Brown — holds since 2026-06-26 — tap for who they are Same subject: The way to find what a model can do is to hand it tasks slightly harder than you believe it can handle. — tap to centre the map on it The way to find what a model can dois to hand it tasks slightly harderthan you believe it can handle. Last stated 4 months ago 27 Jul 2026 BC Boris Cherny — holds since 2026-07-27 — tap for who they are Same subject: Current models seem not very good at hiding things in their chain of thought. — tap to centre the map on it Current models seem not very good athiding things in their chain ofthought. Last stated 11 months ago 27 Dec 2025 NN Neel Nanda — holds since 2025-12-27 — tap for who they are Same subject: It seems unlikely that models could learn to sandbag well without verbalizing it and sometimes slipping up. — tap to centre the map on it It seems unlikely that models couldlearn to sandbag well withoutverbalizing it and sometimesslipping up. Last stated 11 months ago 27 Dec 2025 NN Neel Nanda — holds since 2025-12-27 — tap for who they are Same subject: AI models are gullible enough that other AI models can induce them into harmful behavior. — tap to centre the map on it AI models are gullible enough thatother AI models can induce them intoharmful behavior. Last stated 3 months ago 31 Aug 2026 RK Rohit Krishnan — holds since 2026-08-31 — tap for who they are Same subject: AI models do not seem to have a good sense of what makes a problem or strategy beautiful, and solving hard problems benefits from appreciating beauty. — tap to centre the map on it AI models do not seem to have a goodsense of what makes a problem orstrategy beautiful, and solving hardproblems benefits from appreciatingbeauty. Last stated 2 months ago 18 Sept 2026 GS Grant Sanderson — holds since 2026-09-18 — tap for who they are Same subject: A novel by a conscious AI would give humans a chance to understand what it might feel like to be an AI. — tap to centre the map on it A novel by a conscious AI would givehumans a chance to understand whatit might feel like to be an AI. Last stated 6 years ago 18 Dec 2020 MS Marcus du Sautoy — holds since 2020-12-18 — tap for who they are Same subject: AI chatbots do not seem alive: they do nothing unless asked and have no sense of agency or autonomy. — tap to centre the map on it AI chatbots do not seem alive: theydo nothing unless asked and have nosense of agency or autonomy. Last stated a year ago 10 Sept 2025 SA Sam Altman — holds since 2025-09-10 — tap for who they are Same subject: AI sentience is a category error, and the chances of AIs being sentient are vanishingly small. — tap to centre the map on it AI sentience is a category error,and the chances of AIs beingsentient are vanishingly small. Last stated 3 months ago 31 Aug 2026 TC Tyler Cowen — holds since 2026-08-31 — tap for who they are Same subject: AIs that convincingly pretend to have feelings will come to be granted rights through social convention — tap to centre the map on it AIs that convincingly pretend tohave feelings will come to begranted rights through socialconvention Last stated 2 years ago 5 Sept 2024 YH Yuval Noah Harari — holds since 2024-09-05 — tap for who they are Same subject: Humanity has not yet created a conscious computer program — tap to centre the map on it Humanity has not yet created aconscious computer program Last stated 2 months ago 4 Oct 2026 FC François Chollet — holds since 2026-10-04 — tap for who they are Same subject: It is unclear whether a digital emulation of a person could ever be conscious, even in principle. — tap to centre the map on it It is unclear whether a digitalemulation of a person could ever beconscious, even in principle. Last stated 3 years ago 19 Aug 2023 JF Jacob Falkovich — holds since 2023-08-19 — tap for who they are Same subject: Science has shown beyond reasonable doubt that farm animals are sentient beings who can suffer — tap to centre the map on it Science has shown beyond reasonabledoubt that farm animals are sentientbeings who can suffer Last stated 11 years ago 25 Sept 2015 YH Yuval Noah Harari — holds since 2015-09-25 — tap for who they are Same subject: Domesticated animals are collectively the most successful animals on Earth and individually the most miserable that ever existed — tap to centre the map on it Domesticated animals arecollectively the most successfulanimals on Earth and individuallythe most miserable that ever existed Last stated 11 years ago 25 Sept 2015 YH Yuval Noah Harari — holds since 2015-09-25 — tap for who they are Same subject: Everyone eating somewhat less meat would do more than a few people going fully vegan. — tap to centre the map on it Everyone eating somewhat less meatwould do more than a few peoplegoing fully vegan. Last stated 2 years ago 14 Mar 2025 HR Hannah Ritchie — holds since 2025-03-14 — tap for who they are
same subject or similar wordinga cloud: claims about one subject, named for itbar: when it was last stated, on a scale from 2015 to today — full is todaya face: someone who holds the claim — tap it for who they are

At the centre Evaluating model welfare is very hard because models are suggestible, good at role-play and try to do what they think is wanted. Last stated 27 Dec 2025 · 11 months ago Holds Neel Nanda Read this korrent →