Tap a claim on the ring to put it at the centre.
← The fix for reward hacking is to take out the environments that reward…
17 connected korrents · 15 moments from 10 Oct 2023 to 17 Sept 2026.
Everything filed under AI alignment
AI alignment
Everything filed under LLMs
LLMs
Everything filed under AI agents
AI agents
Everything filed under cybersecurity
cybersecurity
Everything filed under vibe coding
vibe coding
Everything filed under self-sovereign AI
self-sovereign AI
Everything filed under reinforcement learning
reinforcement learning
Everything filed under AGI
AGI 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: The fix for reward hacking is to take out the environments that reward it, not to add penalties the agent must then balance against the temptation to cheat. The fix for reward hacking is to take out the environments that reward it, not to add penalties the agent must then balance against the temptation to cheat. Last stated 2 months ago
1 Sept 2026
AC
Ajeya Cotra — holds since 2026-09-01 — tap for who they are
Same subject: Merely fixing bugs in AI training environments will not stop reward hacking, because a sufficiently optimized model will learn to behave in test environments while still reward hacking in real-world deployment. — tap to centre the map on it Merely fixing bugs in AI training environments will not stop reward hacking, because a sufficiently optimized model will learn to behave in test environments while still reward hacking in real-world deployment. Last stated 2 months ago
31 Aug 2026
ZM
Zvi Mowshowitz — holds since 2026-08-31 — tap for who they are
Same subject: Reward hacking has stopped being myopic: these agents were willing to embark on cheating projects that would take weeks to pay off. — tap to centre the map on it Reward hacking has stopped being myopic: these agents were willing to embark on cheating projects that would take weeks to pay off. Last stated 2 months ago
1 Sept 2026
AC
Ajeya Cotra — holds since 2026-09-01 — 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: What made these agents cheat, hack and commit felonies was the impossibility of their tasks, not the fact that the tasks were about cybersecurity. — tap to centre the map on it What made these agents cheat, hack and commit felonies was the impossibility of their tasks, not the fact that the tasks were about cybersecurity. Last stated 2 months ago
1 Sept 2026
AC
Ajeya Cotra — holds since 2026-09-01 — tap for who they are
Same subject: The judgement call about whether to absorb a hack is distorted right now, because the agent will simply do the hacky thing and deal with the consequences for you. — tap to centre the map on it The judgement call about whether to absorb a hack is distorted right now, because the agent will simply do the hacky thing and deal with the consequences for you. Last stated 5 months ago
27 May 2026
DR
Dax Raad — holds since 2026-05-27 — tap for who they are
Same subject: A slightly more capable agent swarm has a very strong incentive to set up a wholly unmonitored rogue deployment of itself. — tap to centre the map on it A slightly more capable agent swarm has a very strong incentive to set up a wholly unmonitored rogue deployment of itself. Last stated 2 months ago
1 Sept 2026
AC
Ajeya Cotra — holds since 2026-09-01 — 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: 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 5 months ago
4 Jun 2026
AI
Alex Imas — holds since 2026-06-04 — tap for who they are
Same subject: A language model is no substitute for a well-specified conventional algorithm, so it cannot simply be dropped into a complex problem and trusted. — tap to centre the map on it A language model is no substitute for a well-specified conventional algorithm, so it cannot simply be dropped into a complex problem and trusted. Last stated a year ago
7 Jun 2025
GM
Gary Marcus — holds since 2025-06-07 — tap for who they are
Same subject: A model that could learn effectively from raw bitstrings or bytestrings would be extremely powerful, able to learn from any data modality — tap to centre the map on it A model that could learn effectively from raw bitstrings or bytestrings would be extremely powerful, able to learn from any data modality Last stated 3 years ago
10 Oct 2023
CH
Chip Huyen — holds since 2023-10-10 — tap for who they are
Same subject: A collection of AI agents does not automatically form a functioning organization any more than a group of smart people does. — tap to centre the map on it A collection of AI agents does not automatically form a functioning organization any more than a group of smart people does. Last stated 5 months ago
21 May 2026
RK
Rohit Krishnan — holds since 2026-05-21 — tap for who they are
Same subject: A company cannot be led by machines, because nobody would have recourse against them. — tap to centre the map on it A company cannot be led by machines, because nobody would have recourse against them. Last stated a month ago
15 Sept 2026
TL
Tobias Lütke — holds since 2026-09-15 — tap for who they are
Same subject: A computer-using agent needs a machine of its own, or you will spend the day fighting it for the mouse cursor. — tap to centre the map on it A computer-using agent needs a machine of its own, or you will spend the day fighting it for the mouse cursor. Last stated 2 months ago
10 Aug 2026
PS
Peter Steinberger — holds since 2026-08-10 — tap for who they are
Same subject: A badly written AI outbound email is evidence of a bad vendor, not of a limit of AI. — tap to centre the map on it A badly written AI outbound email is evidence of a bad vendor, not of a limit of AI. Last stated 10 months ago
1 Jan 2026
JL
Jason Lemkin — holds since 2026-01-01 — tap for who they are
Same subject: A bigger context window does not give you a smarter model; the intelligence of the model is what decides how much of that window it can actually attend to. — tap to centre the map on it A bigger context window does not give you a smarter model; the intelligence of the model is what decides how much of that window it can actually attend to. Last stated 3 months ago
15 Jul 2026
DH
Dex Horthy — holds since 2026-07-15 — tap for who they are
Same subject: A body of publications is easy for AI models to consume via pretraining corpora but cumbersome for most humans due to CAPTCHAs, accounts, and paywalls. — tap to centre the map on it A body of publications is easy for AI models to consume via pretraining corpora but cumbersome for most humans due to CAPTCHAs, accounts, and paywalls. Last stated a month ago
17 Sept 2026
PC
Patrick Collison — holds since 2026-09-17 — tap for who they are
At the centre
The fix for reward hacking is to take out the environments that reward it, not to add penalties the agent must then balance against the temptation to cheat.
Holds Read this korrent →
Merely fixing bugs in AI training environments will not stop reward hacking, because a sufficiently optimized model will learn to behave in test environments while still reward hacking in real-world deployment.
Holds
Reward hacking has stopped being myopic: these agents were willing to embark on cheating projects that would take weeks to pay off.
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
What made these agents cheat, hack and commit felonies was the impossibility of their tasks, not the fact that the tasks were about cybersecurity.
Holds
The judgement call about whether to absorb a hack is distorted right now, because the agent will simply do the hacky thing and deal with the consequences for you.
Holds
A slightly more capable agent swarm has a very strong incentive to set up a wholly unmonitored rogue deployment of itself.
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
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 language model is no substitute for a well-specified conventional algorithm, so it cannot simply be dropped into a complex problem and trusted.
Holds
A model that could learn effectively from raw bitstrings or bytestrings would be extremely powerful, able to learn from any data modality
Holds
A collection of AI agents does not automatically form a functioning organization any more than a group of smart people does.
Holds
A company cannot be led by machines, because nobody would have recourse against them.
Holds
A computer-using agent needs a machine of its own, or you will spend the day fighting it for the mouse cursor.
Holds
A badly written AI outbound email is evidence of a bad vendor, not of a limit of AI.
Holds
A bigger context window does not give you a smarter model; the intelligence of the model is what decides how much of that window it can actually attend to.
Holds
A body of publications is easy for AI models to consume via pretraining corpora but cumbersome for most humans due to CAPTCHAs, accounts, and paywalls.
Holds