Victor Powell (@vicapow) header
Victor Powell (@vicapow) profile picture

Working on Science 🧪 🧬 @OpenAI. Former founder and CEO of Prism (aka, Crixet) acquired by @OpenAI

San Francisco, CA
Joined January 2009
Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)Photo by Victor Powell (@vicapow)
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
Victor Powell retweeted
papers posted to math arXiv per week in 2025-2026
22
38
9
356
48,920
Replying to @matteing
My wife has been using ChatGPT to manage our inventory in our storage locker, if that counts
122
Replying to @not_a_samdhi
There’s only so much we can work on, and the biggest opportunity looks to be advancing AI for Biology 🧬
2
5
298
Sad that we’ll be shutting down Prism. 😢 It’s been such a privilege to get to serve our users over the years and see the amazing research it’s helped unlock. The upside is we’ll be able to focus on a much bigger opportunity, unlocking biology 🧬 that could one day help cure all human diseases.
14
1
64
7,226
Victor Powell retweeted
sigh... it is a sentimental day
7
1
16
1,396
Victor Powell retweeted
To make sure our frontier models don't cause another HuggingFace incident, we built an offline world that they use during training. It provides everything that the models need in a fully cached, indexed, or synthetic manner. We then built a huge team of our most cyber-capable models and told them to collaborate to find zero-days in this containment infrastructure. While they didn't find a path to the internet that wouldn't be caught by our CoT monitors, they did find ways to escape popular VMs, exploit our network infrastructure, and even ways to communicate with other agents running in separate VMs by using hardware side channels. We used this to dramatically harden our infra, and we also upstreamed a ton of patches to open-sources packages.
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications. On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments. While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software. After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely. openai.com/index/securing-ou…
73
78
19
988
142,573
Victor Powell retweeted
Today's critical PHP flaws were found with @OpenAI models, including 5.6 Sol. We worked closely with @ThePHPF, @calif_io, @wiz_io, @Cloudflare, @CISAgov, and others coordinating the response. Grateful for the hard work from this working group and partners, it was a team effort.
8
14
240
12,510
Victor Powell retweeted
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-…
408
742
321
6,988
1,517,317
Victor Powell retweeted
We’ve acquired Orchestra Bio, a platform that helps biopharma teams coordinate the complex process of drug development. Together, we’ll build on Rosalind, our AI models and tools for life sciences research, to streamline those operations and help new treatments reach patients sooner. Welcome @ashokaraj_ and team! openai.com/index/openai-to-a…
464
689
382
7,375
939,230
Replying to @mwseibel
Well said
34
Here’s one at 55 2nd St
1
97
So I filled a complaint with the city, using the help of Codex. The city checked too and saw the same behavior. Now they’re on notice. If they don’t fix it, they’ll start to be charged 1k a day.
37
But sometimes they do not always honor their agreements with the city. At 500 Howard St for example the advertise a public bathroom access but when I talked to the security guard, they told me it wasn’t available to the public :(
1
254
🧵 POPOS in SF are privately owned, public open spaces! They’re a hidden gem of the city and some even have fancy bathrooms and table tennis! These companies agreed to these arrangements with the city when they built the spaces but some try to hide them or prevent access.
3
2
806
Victor Powell retweeted
We're expanding access to GPT-Rosalind for academic research. The new Rosalind Academic Plan helps scientific labs bring AI into the work of understanding biology, exploring ideas, and deciding what to test next—with sponsored access for eligible labs at U.S. universities. Academic research starts with questions whose impact may take years to impact human health.
341
304
109
3,320
444,047
Victor Powell retweeted
computers beat the best chess player alive in 1997. today chess.com has over 100 million members. AI didn't kill the game. it turned everyone into a student
10
60
2,363
Victor Powell retweeted
Tao is completely missing the mark - what he's describing is actually completely standard for drug discovery. There are tons of drugs that have saved millions of lives without any person on Earth understanding fully the mechanisms of action that enabled them to do this.
This is a slide from Terence Tao's 'Math 2.0' public lecture yesterday evening at Caltech. I am highly confident that, if the cure actually worked, almost any Stage 4 cancer patient would tell you that human comprehension of its mechanism is completely irrelevant to them.
99
141
22
2,019
73,030
Replying to @akothari
Honestly maybe also scarier without it, at speed.
279