OpenAI

Frightening.

Turn on agent mode and ChatGPT can take action for you—research, plan, and get things done while you browse. Now in preview for Plus, Pro, and Business users. https://lnkd.in/g9DgnSA2

Our new engineering blog takes a look at the stories behind the the technology and builders at OpenAI. Building our ChatGPT-powered browser Atlas meant rethinking the browser experience and creating a brand new process architecture, tailored for agentic experiences. Atlas engineers Ken Rockot and Ben Goodger take us under-the-hood on how they built our solution, “OWL” to ensure fast startup, richer UI, and nimble iteration. https://lnkd.in/e6XPbmXx

Now in private beta: Aardvark — an agentic security researcher that helps security teams discover and fix vulnerabilities at scale. Aardvark puts frontier research to work. It reads, reasons about, and tests code, catching subtle vulnerabilities and privacy issues traditional tools often miss. Then it proposes targeted fixes, helping teams secure software before issues ever reach production. Aardvark is already delivering impact — surfacing meaningful vulnerabilities across OpenAI and open-source projects — and we believe access to this level of security expertise should scale. We’re starting with a private beta and will broaden availability as we learn and refine. https://lnkd.in/g87huHSJ

We’re partnering with PayPal to bring their trusted payments and millions of merchants into Instant Checkout in ChatGPT. Under this partnership, PayPal is also expanding its AI strategy with OpenAI beyond commerce — scaling access to ChatGPT Enterprise for its 24,000+ employees, enabling engineers with Codex, and broadly expanding direct use of OpenAI’s APIs. Together, this empowers PayPal to accelerate product development, boost employee impact, and raise the bar on customer experiences.

We’re sharing a research preview of gpt-oss-safeguard — our open-weight reasoning models designed for safety classification tasks. About gpt-oss-safeguard: - Designed to help developers test, evaluate, and improve AI safety systems in the open. - Fine-tuned versions of our gpt-oss open models - Free for anyone to use, modify, and deploy We partnered closely with ROOST to identify developer needs, evaluate early builds, and build documentation that makes it easier to use these models in real workflows. ROOST is also launching the ROOST Model Community today — a space for Trust & Safety practitioners and researchers to share best practices and accelerate open-source safety innovation. This is a research preview, and we’re looking for feedback from the community. 📍 Join us Dec. 8 in San Francisco for the Open Safeguard Hackathon — hosted by OpenAI, ROOST, and Hugging Face — to explore how open-weight reasoning models can support safer digital spaces and advance online safety research: https://lnkd.in/gcjfnsvb Models are available now on Hugging Face, Groq, LM Studio, Ollama, Cerebras, and more.  https://lnkd.in/gRQbmcbb

We completed our recapitalization. The non-profit, the OpenAI Foundation, is now one of the best-resourced philanthropies ever, with equity valued at ~$130B. It continues to control the OpenAI for-profit, which is now a public benefit corporation. https://lnkd.in/gsVee2aw

It’s Cybersecurity Awareness Month 🔐 There’s a growing risk online that affects AI called “prompt injections.” Here’s how to stay smart about it.

Codex’s multimodal abilities can help you design, iterate, and build beautiful frontends for your apps Channing Conger and Romain Huet show how to use Codex Cloud as your front-end design partner — from sketching ideas to shipping UI.