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."

Probably isn't by virtue of it being publicly released

That's my point about the harness.

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

RSI has always being a well defined name, and you can only have RSI if you have an intelligence capable or creating itself.

It has technically existed for a long time (for longer than the name), but only on academical applications for extremely limited intelligences that could only create something like themselves. And that is still the only form that exists today.

It was never powerful enough to optimize ads distribution, and all the claims people are pushing around today are plain bullshit.

Clearly well-defined must live in a probability bracket. Only humans demand exactitude.

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.

Agreed, what I understand from RSI would be models creating new models, or at least upgrading their own weights/architecture. It does not seem to be the case here.

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

It says to read the complete history. Would that be analogous to reading all of one's chat threads, or just the history of the relevant chat thread that it's a part of?

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.

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?

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?

this is just not good science or engineering. ok as a student project i guess.

Perhaps I'm not understanding it correctly, but here's my take on what the paper is doing.

Imagine you have a problem you want to solve (let's say, identify an OCR'd handwritten character, e.g. the MNIST Dataset). You tell 3 agents "Hey, each of you take a stab at getting really good at recognizing characters from this dataset. You can take 10 refinement steps to continue to improve ". You can't give each agent unlimited steps of course, because you have a finite amount of compute.

So each agent goes off, and by the end, Agent 1 got to 90% accuracy, Agent 2 got to 80% accuracy, and Agent 3 got to 89% accuracy. Agent 1 wins, of course.

But then you look at the refinement steps, and after 2 steps, Agent 1 was _already at_ 90% accuracy. So the agent spent the next 8 steps basically not moving at all. Agent 3 on the other hand, perhaps was continuously climbing in accuracy at every refinement step, but hit step 10 and had to stop.

Now because you recorded every step from every agent, you know what you'd do differently next time -- you'd not allocate as many steps to Agent 1, and give Agent 3 more steps, because perhaps that might result in Agent 3 coming up with a better answer.

From my understanding, that's what they built in the form of a "search" controller -- a way to evaluate automatically and reapply how you could allocate resources more effectively, when applied to a new problem.

But I guess my misunderstanding is how applicable the search controller is when applied to new problems -- just because one pathway stalled early for one problem, doesn't mean it would work for another?