YHK - yet another hacker news clone

Dream-RSI: Recursive Self-Improvement through Evolving Worlds

74 points bananaflag 2 hours ago 18 comments

yanis_t 2 hours ago | parent

Would be very glad if anyone explained to me if and why this is revolutionary.

lantry 1 hour ago | parent

Here you go:

"Across algorithm engineering, mathematical optimization, and GPU kernel engineering, Dream-RSI achieves competitive or improved discovery quality while substantially reducing discovery cost in several settings."

WarmWash 1 hour ago | parent

Probably isn't by virtue of it being publicly released

cmrdporcupine 1 hour ago | parent

Fairly certain all the labs are doing this (RSI) at this point. It's a question of how public their proclamations are about it and how they're positioning PR etc.

Even today's lighter weight models know how to write kernels and optimize them. I've had DeepSeek 4.1 Flash tune the crap out custom CUDA kernels on my own codebase and it was entirely competent at it. And cheap.

The innovation pieces will be in the harnesses to support this. Which I guess is partially what's going on here.

suddenlybananas 1 hour ago | parent

It's not really RSI if you are just using the AI as a tool to help make it better. It has to be doing it itself, no? Otherwise self-hosted compilers are RSI.

cmrdporcupine 1 hour ago | parent

That's my point about the harness.

dgellow 1 hour ago | parent

Why would you assume it is revolutionary?

catigula 41 minutes ago | parent

Because “recursive self-improvement” is widely understood, even at labs, as having a non-trivial chance of eradicating human life.

This is like asking why a paper called “I discovered God. This is the specific proof.” is revolutionary. Just tedious levels of incredulity.

dgellow 7 minutes ago | parent

Anyone who seriously believe that and isn’t being hyperbolic for shock purpose is delusional. It’s a pretty standard research paper, you cannot assume any research that contains the terms RSI to be revolutionary

rybosworld 1 hour ago | parent

Unless I'm misunderstanding, calling this RSI seems misleading?

This looks like an optimization of current training methods, and a good one, but not "RSI" in the sense of a system that can perpetually improve itself forever.

EthanHeilman 40 minutes ago | parent

Does RSI actually mean anything specific anymore? RSI, AGI, at this point seem like buzzwords. Sure AGI has definition that are measurable, say "better than 95% of humans on 95% of intellectual tasks" but if we used that definition we already have AGI and almost no one thinks we have achieved AGI. We use AIs to train AIs which we use to train AIs, why is that not RSI? How much human intervention means that is not RSI?

jephs 34 minutes ago | parent

Yeah, this is absolutely not what anyone reasonable is thinking about when they say recursive self-improvement.

I'd say it's much closer to the concept of continual learning, but I'm only a few pages deep and haven't groqued it fully yet.

logicallee 1 hour ago | parent

This is a solid and very interesting paper! The authors were kind enough to publish the complete prompt for it (appendix B.1 on page 18), so anyone can try their approach with any LLM and see the results.

smusamashah 46 minutes ago | parent

So its more like a harness (like Codex etc) on top of any existing LLM?

DanMcInerney 37 minutes ago | parent

I ended up building a simplified version of this as /self-improve in https://github.com/DanMcInerney/orchflows. History is the state ledger, memory and RSI just cite the history as evidence and can be rewritten. I feel like strong immutable state is the missing piece of the puzzle for most of these memory libraries.

gilfoyle_7 34 minutes ago | parent

I might be wrong but is this really a solution for RSI? I interpret it's as a way to reducing wasted tokens and compute on paths that don't yield better results. It's an optimization. It's a faster way to get to RSI though. what's wrong?

ahmedhossamdev 22 minutes ago | parent

The replay simulator from history for off-policy eval is clever - avoids expensive rollouts. Curious how they prevent the policy from overfitting to already-discovered branches and going stale as the search space expands?