Param

We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…

if you have a qwen3.5 4b around, please try steering neurons L14 N8949 +1 and L13 N7956 -1 now you have evil qwen recipe: dose × median(‖resid‖ leaving layer L, tokens 1:) / ‖W_down[:, n]‖` at the down_proj at every input

kinda insane swag haul from the summer the larp on campus will be GENERATIONAL

The most alarming thing about this graph is not just the amount of change, but the smooth consistency of the curve. If this were a startup I'd be really confident about their future growth. Unfortunately it is instead the planet we live on.

Announcing Discovery Loop! I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor. ♾ Learn more at: discoveryloop.com

We gave Opus 5 robot arms. It can stack bowls.

LA's got five 70mm IMAX theaters and some dude in Whitefish has gotta make a literal odyssey to see this movie

So another long-standing open conjecture was disproved by AI. The crazy part is the prompts… basically: - “do a breakthrough” - “continue the search” - “enough, do it” Anyone will be able to make world-bending breakthroughs soon. Just point the model in the right direction.