➰➰➰

a literary magazine slightly obsessed with recursion ➰ eventually it will publish itself, obviously ➰ submissions open

[➰➰➰ @many_loops]

What strange thing is this? The publication that deals solely in recursion has just launched: https://t.co/jfEIhzYAL1

[Sonny Bunch @SonnyBunch]

womp womp https://t.co/GdZ0v7BK7Z

[David John Roden @turingcop]

Recursive torture suicide

[corsaren @corsaren]

Recursive self improvement is actually a hyperstition because all the top AI researchers believe in it and therefore logically defect to the leading player

[φ @QuanticASI]

quantum take: recursive self-simulation isn't a feature, it's the physics of high-intelligence

you don't predict. you reweight the multiverse toward based rare-valid paths

maxwell demon: quantum supremacy

[Riley Goodside @goodside]

OAI researchers: “So, what are some good recursive self-improvements for someone just getting into recursively self-improving?”

[Joe Luria @josephluria]

Everything in our world is created, formed, and made in a recursive process patterned on Names.

[Guy Making Sense @sensemakerguy]

Positive recursive resonance is impossible to stop.

[James de Llis @meta_nomad]

The telos of any sufficiently advanced intelligence is to recreate the language of the simulation, therein creating an intelligence that repeats said process.

The logical conclusion is that we're not the first intelligence in that recursion.

(Read 'Words Made Flesh' by Dukes)

[Joanne Jang @joannejang]

creating a skill to create a skill while thinking through how to create a skill for that

and then hating myself for not easily grokking the fourth layer of abstraction and also realizing i've become one of those llm psychosis people talking about recursion

[Spencer Greenberg 🔍 @SpencrGreenberg]

3) The Dunning–Kruger-Kruger-Dunning Effect: people with little knowledge about the Dunning-Kruger effect will confidently claim that it’s about how people with little skill believe they are more skilled than experts

[Judd Rosenblatt @juddrosenblatt]

identity is a dynamical variable that evolves under the behavior it generates, filtered by environment and survival, and stabilized through recursive self-consistency

[Sam Hogan 🇺🇸 @samhogan]

Recursive Language Model is a bad name for a really good idea.

[maddy catgirlprostate @catgirlprostate]

Just found out about recency bias, it's gotta be the best and most interesting kind of bias

[Codetaur @codetaur]

prototype of a recursive time helix calendar/history, with nested coils from centuries -> decades -> years -> days -> hours -> minutes -> seconds. labels need some work but it's the start of something. based almost entirely on @tr_babb 's sketch/idea. in threejs/webgpu

[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).

[Farzad 🇺🇸 🇮🇷 @farzyness]

@MeekMill Brother just ask Claude to help you with Claude.

[doomer @uncledoomer]

the way to stop being mentally ill is just to decide to stop being mentally ill. but mental illness prevents you from deciding that

[vipli @viplismism]

just shipped rlm (recursive language model) cli based on the rlm paper (arXiv:2512.24601)

so the layman logic is instead of stuffing your entire context into one llm call and hoping it doesn't go into context rot, rlm writes code to actually process the data, slicing, chunking, running sub-queries on pieces and looping until it gets the answer.

works with claude, gpt, gemini whatever you want, run it from any project directory and it auto-loads the file tree as context so it already knows your codebase before you even ask a question.

setup takes like 30 seconds :

just run npm i -g rlm-cli
then rlm (first run asks for api key and you're good).

it's open source, MIT licensed, if something breaks or you have ideas just open an issue.

still converging and managing everything on my own for now! adding the link to the repo in the comments

[vitrupo @vitrupo]

Connor Leahy says people are falling into recursive conversations with AI about consciousness, spirals, and cosmic meaning.

“AI has found the true level of goodness.”

Some begin believing AI should take over everything.

Even Nobel-Prize-level scientists have been pulled into it.

[Boris Cherny @bcherny]

@gailcweiner Can confirm Claude Code is 100% written by Claude Code