Tap a claim on the ring to put it at the centre.
← Whether a scientific idea was progress depends on a future and a…
17 connected korrents · 14 moments from 30 Aug 2005 to 26 Aug 2026.
Everything filed under reinforcement learning
reinforcement learning
Everything filed under metascience
metascience
Everything filed under mathematics
mathematics
Everything filed under LLMs
LLMs
Everything filed under innovation
innovation
Everything filed under AI and science
AI and science
Everything filed under AI and human skill
AI and human skill 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: Whether a scientific idea was progress depends on a future and a culture that have not happened yet, so grading it may never be something you can reinforcement-learn. Whether a scientific idea was progress depends on a future and a culture that have not happened yet, so grading it may never be something you can reinforcement-learn. Last stated 7 months ago
20 Mar 2026
TT
Terence Tao — holds since 2026-03-20 — tap for who they are
Same subject: Work produced by one frontier model should be reviewed by a different one, for the same reason a good programmer's work improves under a good peer's review. — tap to centre the map on it Work produced by one frontier model should be reviewed by a different one, for the same reason a good programmer's work improves under a good peer's review. Last stated 2 months ago
26 Aug 2026
DH
David Heinemeier Hansson — holds since 2026-08-26 — tap for who they are
Same subject: What AI still cannot do in mathematics is build cumulatively on partial progress; it is trial, error and brute force rather than an idea improving through use. — tap to centre the map on it What AI still cannot do in mathematics is build cumulatively on partial progress; it is trial, error and brute force rather than an idea improving through use. Last stated 7 months ago
20 Mar 2026
TT
Terence Tao — holds since 2026-03-20 — tap for who they are
Same subject: It would be disappointing and surprising if the next five years of AI progress in mathematics produced no economically valuable improvements traceable to it. — tap to centre the map on it It would be disappointing and surprising if the next five years of AI progress in mathematics produced no economically valuable improvements traceable to it. Last stated 4 months ago
30 Jun 2026
GS
Grant Sanderson — holds since 2026-06-30 — tap for who they are
Same subject: 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. — tap to centre the map on it 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: AI learning to generate good conjectures will never show up as a benchmark being knocked down; it will show up as a shift in how mathematicians talk about the tools. — tap to centre the map on it AI learning to generate good conjectures will never show up as a benchmark being knocked down; it will show up as a shift in how mathematicians talk about the tools. Last stated 4 months ago
30 Jun 2026
GS
Grant Sanderson — holds since 2026-06-30 — tap for who they are
Same subject: Scientific progress in biology is not slowing down. — tap to centre the map on it Scientific progress in biology is not slowing down. Last stated 7 years ago
15 Aug 2019
AG
Alexey Guzey — holds since 2019-08-15 — tap for who they are
Same subject: Scientific discovery is ready to be rewritten, because for all of history it has run at the speed of smart humans remembering what other smart humans learned. — tap to centre the map on it Scientific discovery is ready to be rewritten, because for all of history it has run at the speed of smart humans remembering what other smart humans learned. Last stated 2 months ago
10 Aug 2026
FL
Fei-Fei Li — holds since 2026-08-10 — tap for who they are
Same subject: Craft excellence within each discipline is not going away, however much AI blurs the work across functional lines. — tap to centre the map on it Craft excellence within each discipline is not going away, however much AI blurs the work across functional lines. Last stated 3 months ago
19 Jul 2026
ES
Elizabeth Stone — holds since 2026-07-19 — tap for who they are
Same subject: “Trust science” is a confused instruction: what deserves trust is the process, not any particular thing it has concluded so far. — tap to centre the map on it “Trust science” is a confused instruction: what deserves trust is the process, not any particular thing it has concluded so far. Last stated 3 years ago
16 Aug 2023
JC
Jason Crawford — holds since 2023-08-16 — tap for who they are
Same subject: A hotter field with more teams racing on the same question produces less reliable findings, not more. — tap to centre the map on it A hotter field with more teams racing on the same question produces less reliable findings, not more. Last stated 21 years ago
30 Aug 2005
JI
John Ioannidis — holds since 2005-08-30 — tap for who they are
Same subject: A hyped effect shrinking to a small but real one as the studies improve is a sign that a field is working, not failing. — tap to centre the map on it A hyped effect shrinking to a small but real one as the studies improve is a sign that a field is working, not failing. Last stated 4 years ago
11 Oct 2022
SR
Stuart Ritchie — holds since 2022-10-11 — 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: Economics journals have made large datasets and heavy statistics the price of making any causal claim at all, which has squeezed qualitative economic history out of the field. — tap to centre the map on it Economics journals have made large datasets and heavy statistics the price of making any causal claim at all, which has squeezed qualitative economic history out of the field. Last stated a year ago
13 Oct 2025
AH
Anton Howes — holds since 2025-10-13 — tap for who they are
Same subject: Fraudulent papers pass peer review routinely, which is direct evidence that peer review does not do the job it is supposed to do. — tap to centre the map on it Fraudulent papers pass peer review routinely, which is direct evidence that peer review does not do the job it is supposed to do. Last stated 4 years ago
13 Dec 2022
AM
Adam Mastroianni — holds since 2022-12-13 — tap for who they are
Same subject: History has a reproducibility problem worse than science’s replication crisis, because hardly anyone ever goes back to the sources to check what a historian claimed they say. — tap to centre the map on it History has a reproducibility problem worse than science’s replication crisis, because hardly anyone ever goes back to the sources to check what a historian claimed they say. Last stated 3 years ago
29 Aug 2023
AH
Anton Howes — holds since 2023-08-29 — tap for who they are
At the centre
Whether a scientific idea was progress depends on a future and a culture that have not happened yet, so grading it may never be something you can reinforcement-learn.
Holds Read this korrent →
Work produced by one frontier model should be reviewed by a different one, for the same reason a good programmer's work improves under a good peer's review.
Holds
What AI still cannot do in mathematics is build cumulatively on partial progress; it is trial, error and brute force rather than an idea improving through use.
Holds
It would be disappointing and surprising if the next five years of AI progress in mathematics produced no economically valuable improvements traceable to it.
Holds
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
AI learning to generate good conjectures will never show up as a benchmark being knocked down; it will show up as a shift in how mathematicians talk about the tools.
Holds
Scientific progress in biology is not slowing down.
Holds
Scientific discovery is ready to be rewritten, because for all of history it has run at the speed of smart humans remembering what other smart humans learned.
Holds
Craft excellence within each discipline is not going away, however much AI blurs the work across functional lines.
Holds
“Trust science” is a confused instruction: what deserves trust is the process, not any particular thing it has concluded so far.
Holds
A hotter field with more teams racing on the same question produces less reliable findings, not more.
Holds
A hyped effect shrinking to a small but real one as the studies improve is a sign that a field is working, not failing.
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
Economics journals have made large datasets and heavy statistics the price of making any causal claim at all, which has squeezed qualitative economic history out of the field.
Holds
Fraudulent papers pass peer review routinely, which is direct evidence that peer review does not do the job it is supposed to do.
Holds
History has a reproducibility problem worse than science’s replication crisis, because hardly anyone ever goes back to the sources to check what a historian claimed they say.
Holds