Congratulations Valeska on the new job! 🎉

Partner Valeska Pederson Hintz and Associate Ahsan Sayed analyze the U.S. startup and #venturecapital scene in their Venture Capital Law Review article. Despite recent VC pullback, #AI offers hope, and the prospects for #startups might be on the rise. bit.ly/485PlLy

Received my Bitkey order today! Looks like a tiny pizza box, but instead contains a rock-solid wallet.

The best way for defenders to stay ahead of attackers is for defenders to *also* understand how to attack the systems they defend. I'm very proud of this great bitcoin security research from Engineering at Block! Come work with us: jobs.smartrecruiters.com/Square/7440000…

The 2024 Austin Hacker Hoedown date has been selected. Join your industry colleagues for lightning talks, libations, and BBQ at San Jac Saloon on Wednesday, November 20th from 6pm to close. Registration is free but limited. For more information visit: hackerhoedown.com If

Thank you to everyone who’s reported vulns and helped us to be more secure. We’ve increased our security bug bounty maximum payouts. We’re also running a double-payout promotion for IDOR vulns through the end of April.

Agent Security is hiring! We are OpenAI's dedicated team focused on securing our agentic AI systems - we operate where traditional appsec/infrasec boundaries blur. This is frontier work: openai.com/careers/securi…

"Our internal testing revealed that GPT-5 is substantially more capable than any previous model we have integrated into our system."

Yesterday we launched ChatGPT Atlas, our new web browser. In Atlas, ChatGPT agent can get things done for you. We’re excited to see how this feature makes work and day-to-day life more efficient and effective for people. ChatGPT agent is powerful and helpful, and designed to be

A letter from our CISO: "Fighting the New York Times’ invasion of user privacy" openai.com/index/fighting…

We just published a post on how we continuously harden ChatGPT Atlas (and other agents) against novel prompt-injection attacks. This is an ongoing security problem (and a frontier research problem!) and we’re investing heavily in automated red teaming, reinforcement learning,

.Adrian Spânu and I have been working on publicising more on OpenAI AI application security. First up: blog post & paper about safe_url, a mechanism that detects and mitigates some common prompt injection exfil sinks post: openai.com/index/ai-agent… paper: cdn.openai.com/pdf/dd8e7875-e…

The S in ASI stands for swag.

OpenAI is now FedRamp 20x Low authorized. Excited to be on the path of bringing the most secure and powerful AI technologies to solve the hardest problems of some of our most important institutions. marketplace.fedramp.gov/products/FR253…

I’m glad that we’re able to support the U.S. government while maintaining meaningful guardrails. 🇺🇸

.OpenAI blog post: Designing Agents to Resist Prompt Injection We’ve found that prompt injections are evolving to look more like social engineering over time as models get smarter and mitigations need to evolve too. openai.com/index/designin…