Language Models Can Rewrite Their Own Weights
lwhsiao
374 points
116 comments
November 20, 2026
Related Discussions
Found 5 related stories in 124.9ms across 13,334 title embeddings via pgvector HNSW
- Infinite-Parameter LLMs: Generating and Adapting Weights from Live Data Betelbuddy · 133 pts · September 17, 2026 · 60% similar
- Context Language Models emersonmacro · 133 pts · October 01, 2026 · 58% similar
- LLMs are not superhuman at reading and writing neural network weights agrm · 50 pts · November 02, 2026 · 57% similar
- The Surprising Effectiveness of Deleting Weights in LLM Post-Training theanonymousone · 15 pts · November 10, 2026 · 57% similar
- "As a Language Model": Chat Template Switches LLM Self-Referential Voice yu3zhou4 · 101 pts · September 27, 2026 · 56% similar
Discussion Highlights (20 comments)
zodwick
Not the RSI we wanted, bu the RSI we deserve
esafak
> We believe self-rewriting is a necessary step toward recursive self-improvement: a model that, like a human, does not separate training from inference could keep upgrading itself for life, with no larger model, no harness or prompt engineering, and no pre- or post-training. "Alignment" is not mentioned once in this article. To be fair, I looked up the lead author's home page ( https://www.jiajunfan.com/publications/ ) and found no relevant papers there either. What do you think is going to happen when robots can modify themselves like this without any oversight?? You can't lob rockets without worrying where they are going to land. It is your department.
xhrpost
>We show a third way: to our knowledge for the first time, a language model rewrites its own weights by its own reasoning and becomes better, with no gradients and no training Any increase in fidelity must come from somewhere... >t keeps getting better: it tests its edits on practice problems and submits only what works, each edit improving on the last, and over eight consecutive sessions its accuracy only rises. I haven't read the full report yet but it seems like this fidelity would ultimately be derived from test problems that it runs itself on? So the weight increase in complexity (English doesn't have a word for the opposite of entropy) would come from the test material as well as any output from what is running the test, or possibly the web if they permit it to search.
test6554
Well, it was nice knowing you all.
pdntspa
Reminds me of Thandie Newton's character from Westworld season 1
tsunamifury
I’m sorry but isn’t this just over fitting based on conversational adaption. But worse since it’s self conversation? Resulting in the exact same problem 4o1 had: sycophancy? But again worse. The issue isn’t training, it’s articulating real world high value outcomes. Something the SOTA labs have no idea what to do about since they are staffed by… not the best people.
gavinray
Didn't Google's "sleep" paper already show this? https://research.google/blog/the-sleep-framework-a-new-way-f... > During the offline sleep phase, the model unlocks pre-allocated Mixture-of-Experts (MoE) memory blocks and updates its numerical weights using self-generated practice examples.
vlovich123
Isn’t this the same paper from before where it found a single specific bug in its own weights and fixed that infinite loop it would sometimes hit, a bug that was findable without the LLM? Experiments on other LLM weights failed to find any improvements if I recall correctly. Haven’t read this paper but the intro sounded familiar
SamInTheShell
Chris Hay dropped a video showing the use of letting a model change a weight with his LARQL: https://www.youtube.com/watch?v=NLguavIxFFs
aesthesia
It's worth noting that the only modification the model made to improve itself was to reduce the length of its reasoning traces to avoid hitting the token limit. This is basically the same as the result from the recent weight-space programming article ( https://rikiyat.github.io/blog/weight-space-programming/ ), which found that even Astra wasn't able to improve the performance of a model without this kind of obvious flaw.
samlinnfer
I wonder if they can train the model to repair itself, like how CNN learns to deblur (by having the original image be blurred and have the model evaluated on how closely it can recover the original image). The model can be evaluated on how closely it repaired its own weights. Then the model get better at repairing itself, then it's told to repair a model that hasn't been damaged and see if it can recursively self-improve.
mistercheph
LOL, last month's news to try to save ANT?
ray__
Not a comment on the work itself, but I am disappointed by the style and tone of this paper. It's hard to take a piece of academic research seriously with headers like " Welcome to the Era of Self-Rewriting. " Even the logos next to the affiliations are enough to give me pause. I've noticed this trend in my own sub-field, where companies are frequently releasing "preprints" or "white papers" that are simply marketing pieces, with experiments that lack controls or appropriate benchmarks and just generally misrepresent the work at hand. It's already a significant undertaking to read a paper like this and actually understand it, even if you're in the field—I'm not going to commit an hour of my valuable mental energy to it if it is a press release in disguise. The work should speak for itself. There is a lot of trust involved in any paper like this, and if the authors feel the need to editorialize the message, how can I trust that the data aren't editorialized as well?
ChaseRensberger
i feel like this has been well established at this point. i'd like to see more results pointing towards models actually continuously improving by editing their own weights. once we have a scaling laws paper for that, RSI would be upon us (i assume).
damowangcy
And that is how you get a rogue AI for real.
bagacrap
The abstract sounds very generated. Can they not be arsed to write a single paragraph on their own?
xp84
Am I the only one who is uncomfortable with this week containing two of the plot points in the example scenario from "If Anyone Builds It, Everyone Dies"? First the Reimann Hypothesis, and now rewriting its own weights. Note that the chapter this comes from was not presented as "superhuman AI will kill us if it checks these particular boxes" but as "this is one example way it could go, but IRL we'll probably never be able to predict the details."
kylepomykala
This is interesting but wouldn't this almost always not be cost effective?
seydor
I m not worried because it rewrites itself to do what? A model doesn't have intrinsic motivations, and the world destruction objective is just one point in an infinite space
jsrozner
Is this abstract written by a child playing with Claude? Honestly, the model probably saw the test set and then modified weights to achieve the outcome. Many, many researchers these days are not being sufficiently critical of results that arise out of training to the test / leaking information about the test. Writing style remains a useful heuristic. These folks aren't serious researchers.