Tap a claim on the ring to put it at the centre.
← Generalization to new tasks is the central result of the activation…
17 connected korrents · 15 moments from 11 Oct 2018 to 6 Oct 2026.
Everything filed under optimism
optimism
Everything filed under LLMs
LLMs
Everything filed under neural networks
neural networks
Everything filed under robotics
robotics
Everything filed under scaling laws
scaling laws
Everything filed under batteries
batteries
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: Generalization to new tasks is the central result of the activation oracles work and suggests the method has promise.
Generalization to new tasks is the central result of the activation oracles work and suggests the method has promise.
Last stated 11 months ago
20 Dec 2025
NN
Neel Nanda — holds since 2025-12-20 — tap for who they are
Same subject: Models already generalise substantially from tasks that can be verified to tasks that cannot. — tap to centre the map on it
Models already generalise substantially from tasks that can be verified to tasks that cannot.
Last stated 9 months ago
13 Feb 2026
DA
Dario Amodei — holds since 2026-02-13 — tap for who they are
Same subject: Reasoning will generalize the way instruction tuning did: add enough verifiable domains and, at some point nobody can yet locate, the rest start working on their own. — tap to centre the map on it
Reasoning will generalize the way instruction tuning did: add enough verifiable domains and, at some point nobody can yet locate, the rest start working on their own.
Last stated 2 years ago
3 Feb 2025
NL
Nathan Lambert — holds since 2025-02-03 — tap for who they are
Same subject: It is the diversity of a robot data set, not its size, that produces generalization: dropping the most diverse slice hurts far more than dropping a random fifth. — tap to centre the map on it
It is the diversity of a robot data set, not its size, that produces generalization: dropping the most diverse slice hurts far more than dropping a random fifth.
Last stated 3 months ago
12 Aug 2026
CF
Chelsea Finn — holds since 2026-08-12 — tap for who they are
Same subject: Deep learning cannot learn generalizable discrete programs, because a pattern-recognition shortcut is always taken instead. — tap to centre the map on it
Deep learning cannot learn generalizable discrete programs, because a pattern-recognition shortcut is always taken instead.
Last stated 2 years ago
6 Nov 2024
FC
François Chollet — holds since 2024-11-06 — 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 9 months ago
13 Feb 2026
DA
Dario Amodei — holds since 2026-02-13 — tap for who they are
Same subject: Success on the easy version of a task is evidence of nothing, because it seduces you into believing the model found a general solution when it did not. — tap to centre the map on it
Success on the easy version of a task is evidence of nothing, because it seduces you into believing the model found a general solution when it did not.
Last stated a year ago
7 Jun 2025
GM
Gary Marcus — holds since 2025-06-07 — tap for who they are
Same subject: Emergent abilities come from generalisation turning compositional, not merely from a training set containing a lot of stuff. — tap to centre the map on it
Emergent abilities come from generalisation turning compositional, not merely from a training set containing a lot of stuff.
Last stated a year ago
12 Sept 2025
SL
Sergey Levine — holds since 2025-09-12 — tap for who they are
Same subject: Pre-trained representations reduce the need for many heavily-engineered task-specific architectures. — tap to centre the map on it
Pre-trained representations reduce the need for many heavily-engineered task-specific architectures.
Last stated 8 years ago
11 Oct 2018
KT
Kristina Toutanova — holds since 2018-10-11 — tap for who they are
MC
Ming-Wei Chang — holds since 2018-10-11 — tap for who they are
KL
Kenton Lee — holds since 2018-10-11 — tap for who they are
JD
Jacob Devlin — holds since 2018-10-11 — tap for who they are
Same subject: 2026 is going to be the most interesting year in Shopify's history, and probably in literally everyone's career. — tap to centre the map on it
2026 is going to be the most interesting year in Shopify's history, and probably in literally everyone's career.
Last stated 10 months ago
18 Jan 2026
TL
Tobias Lütke — holds since 2026-01-18 — tap for who they are
Same subject: A busted computer is not a calamity of data loss or a long restore process in the modern world — tap to centre the map on it
A busted computer is not a calamity of data loss or a long restore process in the modern world
Last stated 12 months ago
24 Nov 2025
DH
David Heinemeier Hansson — holds since 2025-11-24 — tap for who they are
Same subject: A cosmopolitan patriotism, grown large enough, may overcome nationalism and let us give up war — tap to centre the map on it
A cosmopolitan patriotism, grown large enough, may overcome nationalism and let us give up war
JA
Jane Addams — holds since c. 1906 — tap for who they are
Same subject: Access to AI will become a fundamental driver of the economy, and may eventually be considered a fundamental human right. — tap to centre the map on it
Access to AI will become a fundamental driver of the economy, and may eventually be considered a fundamental human right.
Last stated a year ago
23 Sept 2025
SA
Sam Altman — holds since 2025-09-23 — tap for who they are
Same subject: AI increases productivity by boosting efficiency, and greater productivity means greater prosperity. — tap to centre the map on it
AI increases productivity by boosting efficiency, and greater productivity means greater prosperity.
Last stated 2 months ago
25 Sept 2026
MM
Mr. Money Mustache — holds since 2026-09-25 — tap for who they are
Same subject: Decolonization is truly the creation of new men — tap to centre the map on it
Decolonization is truly the creation of new men
FF
Frantz Fanon — holds since c. 1961 — tap for who they are
Same subject: Battery-electric airliners do not work even on optimistic assumptions, and the arithmetic said so before anyone built one. — tap to centre the map on it
Battery-electric airliners do not work even on optimistic assumptions, and the arithmetic said so before anyone built one.
Last stated 6 years ago
2 Apr 2021
BF
Bjorn Fehrm — holds since 2021-04-02 — tap for who they are
Same subject: Permitting reform as climate policy relies on the assumption that storage technology will continue to improve and become cheaper. — tap to centre the map on it
Permitting reform as climate policy relies on the assumption that storage technology will continue to improve and become cheaper.
Last stated 2 months ago
6 Oct 2026
MY
Matthew Yglesias — holds since 2026-10-06 — tap for who they are
Same subject: A 30-year-old battery losing a time server's year override can make networks trust a wrong date and break security certificates. — tap to centre the map on it
A 30-year-old battery losing a time server's year override can make networks trust a wrong date and break security certificates.
Last stated 3 months ago
4 Sept 2026
JG
Jeff Geerling — holds since 2026-09-04 — tap for who they are
same subject or similar wording a cloud: claims about one subject, named for it bar: when it was last stated, on a scale from 2015 to today — full is today a face: someone who holds the claim — tap it for who they are
At the centre
Generalization to new tasks is the central result of the activation oracles work and suggests the method has promise.
Last stated 20 Dec 2025 · 11 months ago
Holds Neel Nanda
Read this korrent →
Similar wording
Models already generalise substantially from tasks that can be verified to tasks that cannot.
Last stated 13 Feb 2026 · 9 months ago
Holds Dario Amodei
Similar wording
Reasoning will generalize the way instruction tuning did: add enough verifiable domains and, at some point nobody can yet locate, the rest start working on their own.
Last stated 3 Feb 2025 · 2 years ago
Holds Nathan Lambert
Similar wording
It is the diversity of a robot data set, not its size, that produces generalization: dropping the most diverse slice hurts far more than dropping a random fifth.
Last stated 12 Aug 2026 · 3 months ago
Holds Chelsea Finn
Similar wording
Deep learning cannot learn generalizable discrete programs, because a pattern-recognition shortcut is always taken instead.
Last stated 6 Nov 2024 · 2 years ago
Holds François Chollet
Similar wording
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 13 Feb 2026 · 9 months ago
Holds Dario Amodei
Similar wording
Success on the easy version of a task is evidence of nothing, because it seduces you into believing the model found a general solution when it did not.
Last stated 7 Jun 2025 · a year ago
Holds Gary Marcus
Similar wording
Emergent abilities come from generalisation turning compositional, not merely from a training set containing a lot of stuff.
Last stated 12 Sept 2025 · a year ago
Holds Sergey Levine
Similar wording
Pre-trained representations reduce the need for many heavily-engineered task-specific architectures.
Last stated 11 Oct 2018 · 8 years ago
Holds Kristina Toutanova Ming-Wei Chang Kenton Lee Jacob Devlin
Same subject: optimism
2026 is going to be the most interesting year in Shopify's history, and probably in literally everyone's career.
Last stated 18 Jan 2026 · 10 months ago
Holds Tobias Lütke
Same subject: optimism
A busted computer is not a calamity of data loss or a long restore process in the modern world
Last stated 24 Nov 2025 · 12 months ago
Holds David Heinemeier Hansson
Same subject: optimism
A cosmopolitan patriotism, grown large enough, may overcome nationalism and let us give up war
Last stated c. 1906
Holds Jane Addams
Same subject: agency
Access to AI will become a fundamental driver of the economy, and may eventually be considered a fundamental human right.
Last stated 23 Sept 2025 · a year ago
Holds Sam Altman
Same subject: agency
AI increases productivity by boosting efficiency, and greater productivity means greater prosperity.
Last stated 25 Sept 2026 · 2 months ago
Holds Mr. Money Mustache
Same subject: agency
Decolonization is truly the creation of new men
Last stated c. 1961
Holds Frantz Fanon
Same subject: batteries
Battery-electric airliners do not work even on optimistic assumptions, and the arithmetic said so before anyone built one.
Last stated 2 Apr 2021 · 6 years ago
Holds Bjorn Fehrm
Same subject: batteries
Permitting reform as climate policy relies on the assumption that storage technology will continue to improve and become cheaper.
Last stated 6 Oct 2026 · 2 months ago
Holds Matthew Yglesias
Same subject: batteries
A 30-year-old battery losing a time server's year override can make networks trust a wrong date and break security certificates.
Last stated 4 Sept 2026 · 3 months ago
Holds Jeff Geerling