Tap a claim on the ring to put it at the centre.
← To reward the instinct that a new idea is worth having, training would…
17 connected korrents · 17 moments from 9 Aug 2022 to 10 Sept 2026.
Everything filed under reinforcement learning
reinforcement learning
Everything filed under LLMs
LLMs
Everything filed under behavioural science
behavioural science
Everything filed under education
education
Everything filed under product management
product management
Everything filed under OpenAI
OpenAI
Everything filed under AI alignment
AI alignment
Everything filed under scaling laws
scaling laws 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: To reward the instinct that a new idea is worth having, training would have to score the smallness of the concepts a solution needs, not just whether it solved the problem. To reward the instinct that a new idea is worth having, training would have to score the smallness of the concepts a solution needs, not just whether it solved the problem. Last stated 4 months ago
30 Jun 2026
GS
Grant Sanderson — holds since 2026-06-30 — tap for who they are
Same subject: Teaching people a decision model is useless, because the intuition it rests on is built from experience and has no shortcut. — tap to centre the map on it Teaching people a decision model is useless, because the intuition it rests on is built from experience and has no shortcut. Last stated 4 years ago
9 Aug 2022
GK
Gary Klein — holds since 2022-08-09 — tap for who they are
Same subject: A company should manufacture fresh internal challenges for its best people, because the alternative is that they go and find those challenges at another company. — tap to centre the map on it A company should manufacture fresh internal challenges for its best people, because the alternative is that they go and find those challenges at another company. Last stated 7 months ago
1 Apr 2026
TP
Thuan Pham — holds since 2026-04-01 — tap for who they are
Same subject: Struggling with a hard problem before being shown the answer is not always the best way to learn; a hint or a worked example can beat floundering. — tap to centre the map on it Struggling with a hard problem before being shown the answer is not always the best way to learn; a hint or a worked example can beat floundering. Last stated 3 months ago
22 Jul 2026
SY
Scott H. Young — no longer holds since 2026-07-22 — tap for who they are
Same subject: Problem discovery is not very hard, and product managers who spend a lot of time on it neglect solution discovery, the real work they are paid for. — tap to centre the map on it Problem discovery is not very hard, and product managers who spend a lot of time on it neglect solution discovery, the real work they are paid for. Last stated a month ago
10 Sept 2026
MC
Marty Cagan — holds since 2026-09-10 — tap for who they are
Same subject: A model that solves a hard problem has learned nothing from it: the next session has forgotten it, with no new skill to carry to related problems. — tap to centre the map on it A model that solves a hard problem has learned nothing from it: the next session has forgotten it, with no new skill to carry to related problems. Last stated 7 months ago
20 Mar 2026
TT
Terence Tao — holds since 2026-03-20 — tap for who they are
Same subject: The problems worth a small builder's time are the ones too small for a big AI company to be motivated to solve. — tap to centre the map on it The problems worth a small builder's time are the ones too small for a big AI company to be motivated to solve. Last stated 7 months ago
29 Mar 2026
CH
Chip Huyen — holds since 2026-03-29 — tap for who they are
Same subject: An end-to-end self-improving AI is probably possible, but it is not even desirable, because it is a hard-takeoff scenario. — tap to centre the map on it An end-to-end self-improving AI is probably possible, but it is not even desirable, because it is a hard-takeoff scenario. Last stated a year ago
23 Jul 2025
DH
Demis Hassabis — holds since 2025-07-23 — tap for who they are
Same subject: Rather than having AI just do the task, people should use it to help build the right mental models. — tap to centre the map on it Rather than having AI just do the task, people should use it to help build the right mental models. Last stated 7 months ago
25 Mar 2026
FG
Flora Guo — holds since 2026-03-25 — tap for who they are
Same subject: A feedback loop that reinforces a behavior pulls it toward whatever improves the loop's own score. — tap to centre the map on it A feedback loop that reinforces a behavior pulls it toward whatever improves the loop's own score. Last stated 2 months ago
26 Aug 2026
HK
Henrik Karlsson — holds since 2026-08-26 — tap for who they are
Same subject: A physical AI company needs three AIs, not one — the agent, the simulator and the critic — turning deployment into a flywheel. — tap to centre the map on it A physical AI company needs three AIs, not one — the agent, the simulator and the critic — turning deployment into a flywheel. Last stated 3 months ago
3 Aug 2026
DD
Dmitri Dolgov — holds since 2026-08-03 — 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 11 months ago
17 Nov 2025
AK
Andrej Karpathy — holds since 2025-11-17 — tap for who they are
Same subject: If whole categories of reward hack are undetectable by humans, then training against the hacks we do catch teaches models to cheat only where we cannot see. — tap to centre the map on it If whole categories of reward hack are undetectable by humans, then training against the hacks we do catch teaches models to cheat only where we cannot see. Last stated 2 months ago
11 Aug 2026
RG
Ryan Greenblatt — holds since 2026-08-11 — tap for who they are
Same subject: It is fundamentally difficult to design a reward function that accurately captures the intended goal in reinforcement learning. — tap to centre the map on it 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: More capable AI agents are more likely to find and exploit flaws in their reward functions. — tap to centre the map on it More capable AI agents are more likely to find and exploit flaws in their reward functions. Last stated 2 years ago
28 Nov 2024
LW
Lilian Weng — holds since 2024-11-28 — tap for who they are
Same subject: One pre-trained robot model now matches or beats the specialists that were fine-tuned with reinforcement learning for the very tasks they were built for. — tap to centre the map on it One pre-trained robot model now matches or beats the specialists that were fine-tuned with reinforcement learning for the very tasks they were built for. Last stated 2 months ago
12 Aug 2026
CF
Chelsea Finn — holds since 2026-08-12 — tap for who they are
Same subject: RL environments exist to make a model generalise, not to teach it each skill one at a time — exactly as pre-training does. — tap to centre the map on it RL environments exist to make a model generalise, not to teach it each skill one at a time — exactly as pre-training does. Last stated 8 months ago
13 Feb 2026
DA
Dario Amodei — holds since 2026-02-13 — tap for who they are
Same subject: "Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next. — tap to centre the map on it "Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next. Last stated 11 months ago
25 Nov 2025
IS
Ilya Sutskever — holds since 2025-11-25 — tap for who they are
At the centre
To reward the instinct that a new idea is worth having, training would have to score the smallness of the concepts a solution needs, not just whether it solved the problem.
Holds Read this korrent →
Teaching people a decision model is useless, because the intuition it rests on is built from experience and has no shortcut.
Holds
A company should manufacture fresh internal challenges for its best people, because the alternative is that they go and find those challenges at another company.
Holds
Struggling with a hard problem before being shown the answer is not always the best way to learn; a hint or a worked example can beat floundering.
No longer holds
Problem discovery is not very hard, and product managers who spend a lot of time on it neglect solution discovery, the real work they are paid for.
Holds
A model that solves a hard problem has learned nothing from it: the next session has forgotten it, with no new skill to carry to related problems.
Holds
The problems worth a small builder's time are the ones too small for a big AI company to be motivated to solve.
Holds
An end-to-end self-improving AI is probably possible, but it is not even desirable, because it is a hard-takeoff scenario.
Holds
Rather than having AI just do the task, people should use it to help build the right mental models.
Holds
A feedback loop that reinforces a behavior pulls it toward whatever improves the loop's own score.
Holds
A physical AI company needs three AIs, not one — the agent, the simulator and the critic — turning deployment into a flywheel.
Holds
A verifiable task can be optimised by reinforcement learning until a neural network performs it extremely well.
Holds
If whole categories of reward hack are undetectable by humans, then training against the hacks we do catch teaches models to cheat only where we cannot see.
Holds
It is fundamentally difficult to design a reward function that accurately captures the intended goal in reinforcement learning.
Holds
More capable AI agents are more likely to find and exploit flaws in their reward functions.
Holds
One pre-trained robot model now matches or beats the specialists that were fine-tuned with reinforcement learning for the very tasks they were built for.
Holds
RL environments exist to make a model generalise, not to teach it each skill one at a time — exactly as pre-training does.
Holds
"Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next.
Holds