Matt Knight

Proud to introduce Aardvark, our agentic security researcher powered by GPT-5. Aardvark hunts for vulnerabilities the way a security engineer would: by reading and analyzing code, writing and running tests, and proposing patches. Now in private beta. openai.com/index/introduc…

I’m at SXSW moderating a session with @dartmouth on a student- and faculty-developed platform for student wellness. Join us at 2:30pm central to learn about Dartmouth’s leadership on AI. schedule.sxswedu.com/events/PP11616…

Codex Security—our application security agent—is now in research preview. openai.com/index/codex-se…

I shared this note with @OpenAI today. Onward!

Great conversation with Joel de la Garza on the @a16z podcast. We covered agentic security research with Aardvark, where AI and security are headed, and reflections on five-plus years @OpenAI. Thanks for having me! youtu.be/-clah3kPPHc

A new security agent called Aardvark:

Aardvark is a labor of love and mission for the whole team. We are super excited to bring it to you. Sign up for the beta immediately!!!

Introducing Aardvark, our agentic security researcher:

@sama @OpenAI Aardvark shows what’s possible, and now the real work begins: hill climbing, putting it to work, and patching the planet. Learn more and apply for beta access here: openai.com/index/introduc…

When I started at OpenAI five years ago, AI agents that could secure software were nothing more than a fantasy. Aardvark reminds us how far we’ve come since GPT-3, and this is still only the beginning. Huge thanks to @sama and the @openai team for believing in the vision.

Proud to introduce Aardvark, our agentic security researcher powered by GPT-5. Aardvark hunts for vulnerabilities the way a security engineer would: by reading and analyzing code, writing and running tests, and proposing patches. Now in private beta. openai.com/index/introduc…

Congratulations to Team Atlanta for winning @DARPA and @ARPA_H’s AI Cyber Challenge. @OpenAI models were their top three! #aixcc

Introducing GPT-5 Our best AI system yet, rolling out to all ChatGPT users and developers starting today. openai.com/index/introduc…

LIVE5TREAM THURSDAY 10AM PT

gpt-oss is out! we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!) (and a smaller one that runs on a phone). super proud of the team; big triumph of technology.

What do you want to create next?

@MichaelKratsios Congratulations! Fantastic news.

Thank you, Mr. President for the honor of serving in your White House again, this time as Assistant to the President for Science & Technology and for the nomination to be Director of OSTP. To ensure our Nation’s prosperity and security, we must unleash scientific breakthroughs and ensure America's technological dominance. Now we have the President who will make it happen. A Golden Age of American Innovation lies ahead!

@jaimeblascob @defcon @OpenAI It’s at 1050 tomorrow at the AIxCC Stage. Title is “Eating Our Own Dogfood” though I don’t believe the title made it onto the agenda.

@embeddedsec @defcon @OpenAI Whats the title of the talk tomorrow? I can’t seem to find it in the agenda

I’m @defcon with the @OpenAI team. Interested in AI and security? Join me for the following: - today at 1600: panel on the evolution of LLMs - tomorrow at 1050: talk on how we’re using our LLMs for security at OpenAI See you there!

I’ll also be recapping my and @marcnewlin’s escapades in the Spectrum Collaboration Challenge on a panel about @DARPA competitions at 1720 today.

Excited to give the closing keynote at the @BlackHatEvents AI Summit in one hour. Come learn about how we’re equipping the security team with LLMs @OpenAI.