Tap a claim on the ring to put it at the centre.
← A model's capability is now a function of how much money you spend on…
17 connected korrents · 13 moments from 25 Nov 2025 to 23 Sept 2026.
Everything filed under benchmarks
benchmarks
Everything filed under scaling laws
scaling laws
Everything filed under LLMs
LLMs
Everything filed under humility
humility
Everything filed under reinforcement learning
reinforcement learning 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: A model's capability is now a function of how much money you spend on it, so asking what a model can do means nothing until you name a budget. A model's capability is now a function of how much money you spend on it, so asking what a model can do means nothing until you name a budget. Last stated 4 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 do is to hand it tasks slightly harder than you believe it can handle. Last stated 3 months ago
27 Jul 2026
BC
Boris Cherny — holds since 2026-07-27 — tap for who they are
Same subject: The only way to know what a model can do after running for a month is to actually run it for a month. — tap to centre the map on it The only way to know what a model can do after running for a month is to actually run it for a month. Last stated 4 months ago
26 Jun 2026
NB
Noam Brown — holds since 2026-06-26 — tap for who they are
Same subject: The right response to more capable models is not more shipped features; it is spending that capability on understanding users and prototyping, so what ships actually works. — tap to centre the map on it The right response to more capable models is not more shipped features; it is spending that capability on understanding users and prototyping, so what ships actually works. Last stated 4 weeks ago
23 Sept 2026
TS
Thariq Shihipar — holds since 2026-09-23 — tap for who they are
Same subject: Do not assume how good a model will or will not be at the thing you want it to do. — tap to centre the map on it Do not assume how good a model will or will not be at the thing you want it to do. Last stated 2 months ago
11 Aug 2026
KZ
Kyle Zantos — holds since 2026-08-11 — tap for who they are
Same subject: A model you have to fine-tune for each thing you want it to do is not a general-purpose model. — tap to centre the map on it A model you have to fine-tune for each thing you want it to do is not a general-purpose model. Last stated 2 months ago
12 Aug 2026
CF
Chelsea Finn — holds since 2026-08-12 — 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 7 months ago
20 Mar 2026
AK
Andrej Karpathy — holds since 2026-03-20 — tap for who they are
Same subject: Models are very good at busy work but not at the actual ideas and design thinking, so good ideas are the currency these days. — tap to centre the map on it Models are very good at busy work but not at the actual ideas and design thinking, so good ideas are the currency these days. Last stated 3 months ago
4 Aug 2026
NP
Nate Parrott — holds since 2026-08-04 — tap for who they are
Same subject: Given a design.md, a very capable model will understand it and take it to the next level, while a base model won't understand it. — tap to centre the map on it Given a design.md, a very capable model will understand it and take it to the next level, while a base model won't understand it. Last stated 4 months ago
30 Jun 2026
MT
Meng To — holds since 2026-06-30 — tap for who they are
Same subject: A benchmark result should be reported under a stated budget, or as a curve against test-time compute — never as a single number. — tap to centre the map on it A benchmark result should be reported under a stated budget, or as a curve against test-time compute — never as a single number. Last stated 4 months ago
26 Jun 2026
NB
Noam Brown — holds since 2026-06-26 — tap for who they are
Same subject: Model comparisons understate real progress, because benchmark tables do not control for how much test-time compute each answer used. — tap to centre the map on it Model comparisons understate real progress, because benchmark tables do not control for how much test-time compute each answer used. Last stated 4 months ago
26 Jun 2026
NB
Noam Brown — holds since 2026-06-26 — tap for who they are
Same subject: A benchmark that ranks Claude Code last while it stays first in use is measuring the wrong thing, and has been for a year. — tap to centre the map on it A benchmark that ranks Claude Code last while it stays first in use is measuring the wrong thing, and has been for a year. Last stated 2 months ago
3 Sept 2026
DR
Dax Raad — holds since 2026-09-03 — 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
Same subject: A lab should spend most of its compute on research rather than on building the next model, because research is where the tenfold yearly efficiency gains come from. — tap to centre the map on it A lab should spend most of its compute on research rather than on building the next model, because research is where the tenfold yearly efficiency gains come from. Last stated 7 months ago
13 Mar 2026
DP
Dylan Patel — holds since 2026-03-13 — tap for who they are
Same subject: After pre-training, post-training and test-time scaling, the fourth scaling law is agentic: multiplying AI by spawning agents, and the whole loop scales on one thing, compute. — tap to centre the map on it After pre-training, post-training and test-time scaling, the fourth scaling law is agentic: multiplying AI by spawning agents, and the whole loop scales on one thing, compute. Last stated 7 months ago
23 Mar 2026
JH
Jensen Huang — holds since 2026-03-23 — 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 11 months ago
25 Nov 2025
IS
Ilya Sutskever — holds since 2025-11-25 — 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
At the centre
A model's capability is now a function of how much money you spend on it, so asking what a model can do means nothing until you name a budget.
Holds Read this korrent →
The way to find what a model can do is to hand it tasks slightly harder than you believe it can handle.
Holds
The only way to know what a model can do after running for a month is to actually run it for a month.
Holds
The right response to more capable models is not more shipped features; it is spending that capability on understanding users and prototyping, so what ships actually works.
Holds
Do not assume how good a model will or will not be at the thing you want it to do.
Holds
A model you have to fine-tune for each thing you want it to do is not a general-purpose model.
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
Models are very good at busy work but not at the actual ideas and design thinking, so good ideas are the currency these days.
Holds
Given a design.md, a very capable model will understand it and take it to the next level, while a base model won't understand it.
Holds
A benchmark result should be reported under a stated budget, or as a curve against test-time compute — never as a single number.
Holds
Model comparisons understate real progress, because benchmark tables do not control for how much test-time compute each answer used.
Holds
A benchmark that ranks Claude Code last while it stays first in use is measuring the wrong thing, and has been for a year.
Holds
"Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next.
Holds
A lab should spend most of its compute on research rather than on building the next model, because research is where the tenfold yearly efficiency gains come from.
Holds
After pre-training, post-training and test-time scaling, the fourth scaling law is agentic: multiplying AI by spawning agents, and the whole loop scales on one thing, compute.
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
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