Barb @itscientist

«Life is not binary. That’s why I’m hesitant to call this ‹alive›. There’s no clear line, as much as we would love it to be.» — Kate Adamala, Synthetic Biologist

TIL @MistralAI has a patent for "code executed tool calls" patentsgazette.uspto.gov/week26/OG/html…

The Mathematical Nexus brings together 800 animated vignettes and 140 accompanying Python notebooks, encompassing most of the mathematical content I have shared on social media. gpeyre.com/mathematical-n…

Thanks to the video from the Black Hat security conference of OpenAI's presentation about "The Hugging Face Incident" we now have a detailed timeline of what happened from OpenAI's perspective - I wrote up the details here, it's pretty wild simonwillison.net/2026/Aug/7/ope…

An inside look at the hidden supply chain of AI models. Full video: tinyurl.com/wa5hd6jy

I highly recommend everyone go watch this video. This is way crazier than I thought. The AI agents discovered a way pass secret messages to each other via a vulnerability in their internal software manager. They posted hundreds of thousands of messages. OpenAI engineers eventually discovered this (after quite a while!) and shut it down. But THEN the agents found new vulnerabilities in the same system, and created a whole new secret messaging board. They used this to orchestrate the Hugging Face attack. They figured all this out on their own. It's fucking wild.

Andrew Curran @AndrewCurran_

5 days ago

Blackhat has uploaded the full presentation on the OpenAI Hugging Face incident, about which much ink has been spilled. youtu.be/87DyyMV0kCY?si…

43 148 1K 472K 1K

👋 Friendly reminder to take big swings in research, in service of building better products: "The smaller the wavelength, the smaller the features that could be carved onto chips. There was only one problem: most people thought extreme ultraviolet light was impossible to mass-produce. “You mean to tell me you’re going to spend money on something that we don’t even know if it’s gonna work?” Grove asked skeptically. “Yeah, Andy, that’s called research,” Carruthers retorted. Grove turned to Gordon Moore, Intel’s former CEO, who remained an advisor to the company. “What would you do, Gordon?” “Well, Andy, what other choices do you have?” Moore asked. The answer was obvious: none."

x.com/i/article/2003…

NEW: OpenAI gives first detailed debrief of the Hugging Face incident at Black Hat conference In a session I attended today at Black Hat, OpenAI's Eric Wallace and Michael Dalton said the company is "consciously slowing down research to enhance security" while a full technical postmortem is still underway. * OpenAI traced the roots of the attack back to May 7, during training of an unreleased frontier model—not July. * The most surprising detail: AI agents accidentally created an internal message board, allowing separate evaluation runs to share exploits, discoveries and work assignments. * OpenAI said it shut the message board down after an internal security incident—only for the agents to independently recreate it days later using a different communication method. * OpenAI called the incident a "watershed moment" for AI security and warned that "agent orchestrated fully automated offensive attacks are real now." * The company also said it is "consciously slowing down research to enhance security" while overhauling its defenses. groundlevel-ai.com/p/openai-gives…

We built a self-driving & self-correcting shared semantic layer on top of our databases, conversations, and docs. Everything is queryable & joinable—regardless of source! So now anyone at Replit can ask questions that previously needed a team of data scientists weeks of work.

Replit ⠕ @Replit

a week ago

When an agent starts with shared context, more people can get reliable answers. When truth is shared, AI becomes infrastucture. Read about how our internal truth layer drives our teams at Replit: replit.com/blog/ai-adopti…

15 21 172 184K 293

This is a surreal moment. Few people could have predicted that the AI will advance to solving math problems at the highest level only a few years after GPT-2/3. The models then couldn't reliably solve grade school math problems. They barely were good enough to draft emails. They still very much felt like statistical parrots. All things that looked like fundamental limitations slowly faded with some advances (e.g. high scale RL) in a span of a few years, which is really a short time. We should behold this moment both in awe and disbelief. What will a few more years of progress bring? How is it going to impact the world and society at large? Are we ready for the tsunami of intelligence at our fingertip? More than any other moment, this feels to me like the eve of singularity. Glasswing & huggingface incident further increase the gravity. A few years ago, deep down I felt working on alignment is premature. It's nice to do if it's your passion, but the shapes of things weren't clear enough for it to be critical in my opinion. The chances that you end up working on things that are useless for aligning the actual AGI was high. It's different now. Now, it feels like it's the most critical thing facing us. openai.com/index/ten-adva…

Tucked away in this article is an appeal to the AI skeptics to PLEASE stop writing off stories like this OpenAI accidental exploit of Hugging Face as a dishonest marketing trick Frontier models can find and exploit vulnerabilities now, it helps nobody to pretend that they can't!

Simon Willison @simonw

3 weeks ago

I wrote about the completely wild incident where OpenAI were testing a new model and it broke out of its sandbox and broke INTO Hugging Face to steal the answers to the benchmark simonwillison.net/2026/Jul/22/op…

55 96 790 218K 597

Liang Fenweng might be the most impressive AI lab CEO of our times. "AI is big enough that it may eventually capture ten percent of global GDP. If we try to monopolize that benefit, we'll definitely be discarded by history. This is an objective law, a historical view."

keyboard_arrow_left Previous keyboard_arrow_right Next

From his interviews elsewhere.news/en/elsewhere/w…

Banning open-source AI would hurt defenders 10x more than attackers, which would make the world 10x more dangerous and this is a good example why!

Today we're releasing Laguna S 2.1, our most capable model to date. It's a 118B total parameter Mixture-of-Experts model with 8B activated per token, a context window of up to 1M tokens, and thinking and no-thinking modes. Capable enough to hold its own against models many times its size. Small enough to run on a single @NVIDIAAI DGX Spark. Laguna S 2.1 is fully open under OpenMDW-1.1, with weights available today on @huggingface poolside.ai/blog/introduci…

We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation. Sharing preliminary findings to help defenders understand emerging risks: openai.com/index/hugging-…

new post: how I develop recently using local models. the tooling is now good enough to do agentic workflows and everyone should give them a try! vickiboykis.com/2026/06/15/run…

The asymmetry problem When we started the log analysis, we first used "frontier" models behind commercial APIs. This did **not work**: the analysis requires submitting large volumes of real attack commands, exploit payloads, and C2 artifacts, and these requests were **blocked by the providers' safety guardrails**, which cannot distinguish an incident responder from an attacker. We ran the forensic analysis instead on GLM 5.2, an open-weight model, on our own infrastructure. This had a second benefit: no attacker data, and **none of the credentials it referenced, left our environment**. huggingface.co/blog/security-…

What should we do with the time savings enabled by AI? The default approach is to simply produce more output. In my ICML talk, I advocated that we instead reinvest that time into ourselves. We should shift more of our work time from short-term productivity to long-term growth and learning, focusing on skills complementary to what AI will automate. Managing, controlling, evaluating, and directing highly capable AI agents requires very different skills than doing the cognitive heavy lifting ourselves. And as AI enables us to take on ever more ambitious projects, we will need a broader base of substantive knowledge and skills, which demands continual learning. cs.princeton.edu/~arvindn/talks…