Would be very glad if anyone explained to me if and why this is revolutionary.
lantry•55m ago
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•50m ago
Probably isn't by virtue of it being publicly released
cmrdporcupine•38m ago
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•31m ago
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•31m ago
That's my point about the harness.
dgellow•31m ago
Why would you assume it is revolutionary?
rybosworld•33m ago
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.
logicallee•14m ago
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.
yanis_t•1h ago
lantry•55m ago
"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•50m ago
cmrdporcupine•38m ago
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•31m ago
cmrdporcupine•31m ago
dgellow•31m ago