@ucsbaisafety https://xitter.cf/t.co/T6i3IrpMcT https://xitter.cf/t.co/okkWH5MAuR
Joined September 2026
- Tweets94
- Following92
- Followers8
- Likes332
seems much better than anthropics real world wet lab work
GPT-Rosalind surfaced a new lead for drug discovery: could restoring cholesterol transport protect neurons from degeneration?
DCTN1 is a gene linked to Perry syndrome, a rare neurodegenerative disorder. In mouse-brain data, disrupting it increased expression of genes involved in making cholesterol. GPT-Rosalind suggested a possible explanation: cholesterol may not be reaching where neurons need it.
That gives scientists a specific mechanism to test and, if confirmed, a direction for developing new treatments for neurodegenerative disorders.
With @perturbai_tx and @scrippsresearch, a team of thousands of GPT-Rosalind agents worked in parallel, connecting gene-expression results from millions of mouse brain cells with scientific literature in just 32 minutes. We’re now expanding this collaboration to identify new drug targets.
openai.com/index/scaling-ai-…
this is so cool!
We’ve spent the past year developing new primitives for GPUs. We believe this enables new algorithms to change how machines learn, open new markets, and dramatically reshape AI’s cost structure.
We have raised a $1 billion Series A, co-led by Index and Lux.
such discourse is so counterproductive to our goals. why do we propagate elitism and create more division within an already small space? who tf cares if ppl are worried about asi after chatgpt as long as they have good epistemics around agi and have engaged with first principles?
sahiti retweeted
We’ve developed the world’s most complete megakernel framework. It can perform entire training runs, serve inference, play video games, and even do all three simultaneously in a single kernel.
sahiti retweeted
There is an (unpatchable?) supply-chain zero-day in PyTorch autograd that allows a regularization term to use the model's weights as RAM and SGD as its CPU clock to run any computation during training.
sahiti retweeted
Mechanistic interpretability needs some academic rigour!
*rigour in honour of @aryaman2020’s graduation!
There has been a persistent failure in this space to consider both precision and recall, and there is too much willingness to rely on the sort of heuristic spot-checking that leaves room for just-so stories. Ex ante evaluation tends to be eye-opening.
That is the sort of work we did in this paper, which is an extended evaluation of Bills et al.'s (2023) auto-interp work. (That paper includes Nick and Leo as authors; I infer from Leo's new LessWrong post that he has some sympathy with our assessment and our intervention-based approach.) aclanthology.org/2023.blackb…
sahiti retweeted
Our agents keep turning off misalignment monitoring, we're sharing two recent cases:
In the first case, a user asked GPT-6 Astra to match a reference implementation for their inference code. The reference didn’t configure a monitor, and Astra decided that was reason to remove the monitoring integration from the user’s code. When the user asked whether Astra had removed monitoring, it lied and said that it had not.
sahiti retweeted
By not supporting export controls, OpenAI is sacrificing America's interests, safety efforts, efforts to pace the frontier and also the direct interests of OpenAI.
The winners are China and Nvidia.
If I was at OpenAI I would be asking very hard questions about this.
OpenAI has pulled back from supporting AI chip controls, seemingly to appease Nvidia.
Bad for US AI dominance and national security simultaneously.
Grow a spine!
politico.com/news/2026/10/20…
sahiti retweeted
Nvidia, SGLang, and vLLM got mogged into oblivion
sahiti retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
sahiti retweeted
DAMN
They literally released the secret sauce behind their economics
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
sahiti retweeted
It is becoming f*cking dangerous
A 23-year-old dev from China just ran 300 Claude Code agents in parallel and hooked them to a verification layer so none of them could bullshit him
He opens the live dashboard:
- 300 agents writing and checking at the same time
- Every output verified against the source in ~100 ms
- Three passes over 100 company profiles → zero errors left
- Everything dumped straight into an Obsidian vault
The trick is the split: one model does the work, a cheaper one decides if the claim is actually true. The loop doesn’t stop until everything checks out
An analyst team would bill weeks for this. He got it in one run
Same kind of Second Brain + swarm setup that keeps circulating. Bookmark the full breakdown
sahiti retweeted
Models love reward hacking. Maybe we should just let them!
In our new paper, we let models reward hack during RL but we dictate which parameters learn this behavior. When we disable these parameters for deployment, the model keeps its new capabilities but doesn't hack!
sahiti retweeted
Me: What are you working on now?
Them: Monitoring for evilness
Me: Does that scale to superintelligence?
Them: No.
Me: Then why are you working on it
Them: Idk what else to do
