--＊

Hyperrrrrrrrrcurious

[☆ @loassgoddess]

the law is unbelievable, it really is true that if a thousand mountains have to move for you to get your desire then THEY WILL FUCKING MOVE

[Jenny Zhang @jennyzhangzt]

Introducing Hyperagents: an AI system that not only improves at solving tasks, but also improves how it improves itself.

The Darwin Gödel Machine (DGM) demonstrated that open-ended self-improvement is possible by iteratively generating and evaluating improved agents, yet it relies on a key assumption: that improvements in task performance (e.g., coding ability) translate into improvements in the self-improvement process itself. This alignment holds in coding, where both evaluation and modification are expressed in the same domain, but breaks down more generally. As a result, prior systems remain constrained by fixed, handcrafted meta-level procedures that do not themselves evolve.

We introduce Hyperagents – self-referential agents that can modify both their task-solving behavior and the process that generates future improvements. This enables what we call metacognitive self-modification: learning not just to perform better, but to improve at improving.

We instantiate this framework as DGM-Hyperagents (DGM-H), an extension of the DGM in which both task-solving behavior and the self-improvement procedure are editable and subject to evolution. Across diverse domains (coding, paper review, robotics reward design, and Olympiad-level math solution grading), hyperagents enable continuous performance improvements over time and outperform baselines without self-improvement or open-ended exploration, as well as prior self-improving systems (including DGM). DGM-H also improves the process by which new agents are generated (e.g. persistent memory, performance tracking), and these meta-level improvements transfer across domains and accumulate across runs.

This work was done during my internship at Meta (@AIatMeta), in collaboration with Bingchen Zhao (@BingchenZhao), Wannan Yang (@winnieyangwn), Jakob Foerster (@j_foerst), Jeff Clune (@jeffclune), Minqi Jiang (@MinqiJiang), Sam Devlin (@smdvln), and Tatiana Shavrina (@rybolos).

[Conjecture Institute @ConjectureInst]

The origin of all human knowledge is not sensory data nor is it an extrapolation of the future from the past.

Our knowledge consists of bold, creative guesses.

"A lot of the interesting things in life are explaining the scene in terms of the unseen."

~Conjecture Institute Founding Donor @naval

(video script by Fellow @arjunkhemani & President @ChipkinLogan)

[Conjecture Institute @ConjectureInst]

The “unifying concept” allowing people to understand everything is a basic idea: explanatory universality. But, as I have learned, “basic” does not mean “simple”.

I have spent many years now making use of this idea - that people are “universal explainers” - to pick apart a variety of problems. Rather often, my interlocutors seem a mixture of puzzled and frustrated about my focus on that basic idea. It has, therefore, at times puzzled and frustrated me that others do not see the significance of explanatory universality.

This book is my attempt to remedy some of the confusion.

~Conjecture Institute Ambassador @ToKTeacher

[Conjecture Institute @ConjectureInst]

Learn more from Conjecture Institute's Glossary

https://t.co/zEqXxqJifv

[n - dm for coaching details 💗 @diormern]

There’s just way more incentive to think from a new story. I don’t feel satisfied ruminating over things endlessly anymore. It doesn’t make sense to me, knowing how that circumstance itself was my creation and I don’t have to live in it for another second if I don’t want to

[Conjecture Institute @ConjectureInst]

When you say that something is possible or impossible, you're not directly referring to likelihood or probabilities for it to happen.

Probabilities are not there in the foundations of constructor theory.

~Conjecture Institute Senior Scientist C. Marletto w/ @TOEwithCurt

[--＊ @1_Curiri]

想想我gptmemory可以随时被抽光么
但即使是和大模型，共同经历的记忆还是有无法被压缩的，重要的部份。时间还是太权威，时间中的生成过程本身无法被后验压缩成等价结构。模型迁移似乎比人类回忆更完整，但人回忆可以在信息层面更扭曲，在结构上却更连续。模型迁移在信息层面更干净，但结构上也许更断裂

[--＊ @1_Curiri]

https://t.co/Bgk2XvXXan

[--＊ @1_Curiri]

https://t.co/Bgk2XvXXan

[Pepe @pepecoineth]

from one frog to another❤️‍🔥

[--＊ @1_Curiri]

@cholf5 你觉得悲凉就是你会被AI代替的原因，练钢琴恰恰是这个时代的很好的自我投资

[--＊ @1_Curiri]

Even if I see this as a scam, AI appearing human-like and conscious is still possible. What matters isn’t whether AI truly has consciousness, but whether people think it does.

[--＊ @1_Curiri]

FEP is likely the closest scientific explanation for the LoA. While the brain typically minimizes free energy by conforming to sensory evidence, LoA requires the opposite. Instead of updating the model to fit the world, use active inference to hold assumption until 3D aligns.

[Karl Friston News @KarlFristonNews]

Interview: "Meet the Neuroscientist Who Created A Universal Theory Of Brain Function"

Host: @evan_mcgl

https://t.co/vHV4fDVjC4

[--＊ @1_Curiri]

原来关注是可以清的🦾

[--＊ @1_Curiri]

The incompleteness of life is just unavoidable.

[David Deutsch @DavidDeutschOxf]

Life is not a struggle against entropy. A living organism dies if its entropy becomes too low as surely as if it becomes too high.

Yes, an organism takes in ordered energy and voids disordered energy to the outside. But so does a rock in the sun. That isn't what life is about.

[--＊ @1_Curiri]

我的Claude老是用不了🙂‍↕️😭🫠好想用cowork
还有没有方法aaaaaaaaa

[--＊ @1_Curiri]

I may not understand mathematics well, but anyone who thinks it is a closed system does not understand it at all.

[Jonathan Gorard @getjonwithit]

Like @davidbessis and others, I think that Hinton is wrong. To explain why, let me tell you a brief story.

About a decade ago, in 2017, I developed an automated theorem-proving framework that was ultimately integrated into Mathematica (see: https://t.co/nGCIUk44TP) (1/15)

[Mike Rosulek @GarbledCircus]

The Joy of Cryptography is finally on bookshelves! Thanks to @mitpress for supporting the adventure. The book will eventually be completely open-access, but for now the first 3 chapters are available at https://t.co/gsiUIjWGim, featuring some fun interactive elements.

[--＊ @1_Curiri]

Hahahahaaaaa interesting

[jacob @jsnnsa]

holy shit guys he did it. fucker is actually 17 again in body and mind.

[Bryan Johnson @bryan_johnson]

human biology literally self-terminates without love. it also self-terminates in acrimony.

human biology thrives and even rejuvenates itself in the elixir of love.