Tap a claim on the ring to put it at the centre.
← It is fundamentally difficult to design a reward function that…
17 connected korrents · 16 moments from 28 Nov 2024 to 21 Sept 2026. Nearly all of them are about reinforcement learning .
Everything filed under AI alignment
AI alignment
Everything filed under design
design
Everything filed under LLMs
LLMs
Everything filed under AGI
AGI
Everything filed under neural networks
neural networks
Everything filed under benchmarks
benchmarks Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Read this korrent: It is fundamentally difficult to design a reward function that accurately captures the intended goal in reinforcement learning. It is fundamentally difficult to design a reward function that accurately captures the intended goal in reinforcement learning. Last stated 2 years ago
28 Nov 2024
LW
Lilian Weng — holds since 2024-11-28 — tap for who they are
Same subject: A verifiable task can be optimised by reinforcement learning until a neural network performs it extremely well. — tap to centre the map on it A verifiable task can be optimised by reinforcement learning until a neural network performs it extremely well. Last stated 10 months ago
17 Nov 2025
AK
Andrej Karpathy — holds since 2025-11-17 — tap for who they are
Same subject: Better execution environments are needed for LLMs to properly utilize their ability to evolve systems. — tap to centre the map on it Better execution environments are needed for LLMs to properly utilize their ability to evolve systems. Last stated 2 weeks ago
21 Sept 2026
TL
Tobias Lütke — holds since 2026-09-21 — tap for who they are
Same subject: Capability does not generalise for free: a model that will move mountains on an agentic task still tells the same bad joke it told five years ago. — tap to centre the map on it Capability does not generalise for free: a model that will move mountains on an agentic task still tells the same bad joke it told five years ago. Last stated 6 months ago
20 Mar 2026
AK
Andrej Karpathy — holds since 2026-03-20 — tap for who they are
Same subject: Current models are unsafe because they are not smart enough, not because they are too smart. — tap to centre the map on it Current models are unsafe because they are not smart enough, not because they are too smart. Last stated 3 weeks ago
13 Sept 2026
FC
François Chollet — holds since 2026-09-13 — tap for who they are
Same subject: Humans barely use reinforcement learning for intelligence — what RL they do use goes into motor tasks, not problem solving. — tap to centre the map on it Humans barely use reinforcement learning for intelligence — what RL they do use goes into motor tasks, not problem solving. Last stated 11 months ago
17 Oct 2025
AK
Andrej Karpathy — holds since 2025-10-17 — tap for who they are
Same subject: Humans keep their place in AI as judges rather than authors, because telling which of two answers is better is far easier than writing a good one. — tap to centre the map on it Humans keep their place in AI as judges rather than authors, because telling which of two answers is better is far easier than writing a good one. Last stated 2 years ago
3 Feb 2025
NL
Nathan Lambert — holds since 2025-02-03 — tap for who they are
Same subject: Large language models mimic what people say to do rather than work out what to do, which is why they are not about understanding the world. — tap to centre the map on it Large language models mimic what people say to do rather than work out what to do, which is why they are not about understanding the world. Last stated a year ago
26 Sept 2025
RS
Richard Sutton — holds since 2025-09-26 — 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 than they are in the world because researchers, inadvertently, take inspiration from the evals when they build RL environments. Last stated 10 months ago
25 Nov 2025
IS
Ilya Sutskever — holds since 2025-11-25 — tap for who they are
Same subject: A blank chat box is lazy: it violates the first rule of a good user experience, that it is obvious what you can do. — tap to centre the map on it A blank chat box is lazy: it violates the first rule of a good user experience, that it is obvious what you can do. Last stated a year ago
12 May 2025
JZ
Julie Zhuo — holds since 2025-05-12 — tap for who they are
Same subject: A board game conquers the world when chance and strategy are in balance in it. — tap to centre the map on it A board game conquers the world when chance and strategy are in balance in it. Last stated 10 months ago
12 Dec 2025
IF
Irving Finkel — holds since 2025-12-12 — tap for who they are
Same subject: A browser works better as a single window than as multiple separate windows. — tap to centre the map on it A browser works better as a single window than as multiple separate windows. Last stated 2 weeks ago
15 Sept 2026
CC
Chris Coyier — holds since 2026-09-15 — tap for who they are
Same subject: A recurring AI workflow improves more through ongoing user feedback than by being fully specified upfront. — tap to centre the map on it A recurring AI workflow improves more through ongoing user feedback than by being fully specified upfront. Last stated 5 months ago
18 May 2026
JL
Jason Liu — holds since 2026-05-18 — tap for who they are
Same subject: AI agents should notice useful things the user did not think to ask about, not simply complete the request. — tap to centre the map on it AI agents should notice useful things the user did not think to ask about, not simply complete the request. Last stated 2 weeks ago
21 Sept 2026
LR
Lenny Rachitsky — holds since 2026-09-21 — tap for who they are
Same subject: AI agents will end traditional UI design and accessibility, because users will stop visiting websites and act only through their agent. — tap to centre the map on it AI agents will end traditional UI design and accessibility, because users will stop visiting websites and act only through their agent. Last stated 2 years ago
21 Feb 2025
JN
Jakob Nielsen — holds since 2025-02-21 — tap for who they are
Same subject: For now AI is best thought of as average intelligence, which is still very powerful to have across every specialty. — tap to centre the map on it For now AI is best thought of as average intelligence, which is still very powerful to have across every specialty. Last stated 3 weeks ago
10 Sept 2026
EV
Elena Verna — holds since 2026-09-10 — tap for who they are
Same subject: A country outside the AI supply chain should just buy the index — which works only in the world where AI ends up commoditised rather than concentrated. — tap to centre the map on it A country outside the AI supply chain should just buy the index — which works only in the world where AI ends up commoditised rather than concentrated. Last stated 4 months ago
4 Jun 2026
AI
Alex Imas — holds since 2026-06-04 — tap for who they are
Same subject: A poor country should prioritise owning a piece of AI over retraining its workers, but it should not bet everything on that. — tap to centre the map on it A poor country should prioritise owning a piece of AI over retraining its workers, but it should not bet everything on that. Last stated 4 months ago
4 Jun 2026
PT
Phil Trammell — holds since 2026-06-04 — tap for who they are
At the centre
It is fundamentally difficult to design a reward function that accurately captures the intended goal in reinforcement learning.
Holds Read this korrent →
A verifiable task can be optimised by reinforcement learning until a neural network performs it extremely well.
Holds
Better execution environments are needed for LLMs to properly utilize their ability to evolve systems.
Holds
Capability does not generalise for free: a model that will move mountains on an agentic task still tells the same bad joke it told five years ago.
Holds
Current models are unsafe because they are not smart enough, not because they are too smart.
Holds
Humans barely use reinforcement learning for intelligence — what RL they do use goes into motor tasks, not problem solving.
Holds
Humans keep their place in AI as judges rather than authors, because telling which of two answers is better is far easier than writing a good one.
Holds
Large language models mimic what people say to do rather than work out what to do, which is why they are not about understanding the world.
Holds
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.
Holds
A blank chat box is lazy: it violates the first rule of a good user experience, that it is obvious what you can do.
Holds
A board game conquers the world when chance and strategy are in balance in it.
Holds
A browser works better as a single window than as multiple separate windows.
Holds
A recurring AI workflow improves more through ongoing user feedback than by being fully specified upfront.
Holds
AI agents should notice useful things the user did not think to ask about, not simply complete the request.
Holds
AI agents will end traditional UI design and accessibility, because users will stop visiting websites and act only through their agent.
Holds
For now AI is best thought of as average intelligence, which is still very powerful to have across every specialty.
Holds
A country outside the AI supply chain should just buy the index — which works only in the world where AI ends up commoditised rather than concentrated.
Holds
A poor country should prioritise owning a piece of AI over retraining its workers, but it should not bet everything on that.
Holds