OpenAI
OpenAI's mission is to ensure that artificial general intelligence benefits all of humanity.
State of the AI industry — the OpenAI Podcast Ep. 12
Andrew Mayne hosts a discussion with Sarah Friar and Vinod Khosla about the current AI ecosystem, arguing that AI is more like infrastructure with growing compute demand and a long endurance beyond hype. They explore how AI will mature from consumer chat to true agentic tasking, the enterprise merit of vertically specialized solutions, regulatory hurdles in health, and the economics of compute, pricing, and ads, alongside the implications for startups, robotics, and the future of work.
Getting started with Codex
This video introduces CodeEx and walks through getting started with its CLI and IDE integrations, including installation, sign-in, and selecting development contexts. It covers how to configure agents.md files, prompts, and MCP servers, plus advanced use cases like planning, testing, code reviews, and open-source customization. The presenters share best practices, tips for grounding prompts, and examples of integrating CodeEx across pipelines, Slack, GitHub, and diverse tooling.
Codex in JetBrains IDEs
The video demonstrates Codex integrated inside JetBrains IDEs, showing how developers log in via ChatGPT or API keys and work on a real Kotlin Multiplatform project from build to deployment. It highlights Codex assisting with debugging, implementing features like localization, and operating in agent modes that balance autonomy with user control, all while keeping the experience within the IDE.
OpenAI Town Hall with Sam Altman
The panel explores how next‑generation AI tools will reshape software development, business models, and education, emphasizing how humans and AI will collaborate to create, customize, and deploy software at scale. Speakers discuss engineering jobs in Jevans paradox terms, the rise of memory/personalization, security and biocontainment concerns, and the social/policy levers needed to ensure broad, safe adoption. They also consider education, creativity, and the future of collaboration between people and intelligent agents in a world of increasing abundance.
Build Hour: Apps in ChatGPT
OpenAI Build Hours explores the CHBT apps platform, demos real-time app building with ChatGPT, MCP servers, and the Docs MCP workflow. The hosts showcase apps (All Trails, Adobe Express, quizzes), live UI components, full-screen and widget integrations, and how CodeEx accelerates development, followed by a Q&A session on MCP compatibility, tooling, and UX best practices.
Accelerating science with Prism
The video introduces Prism, an AI native environment that embeds advanced AI into scientists’ editors to accelerate research, demoing real-time writing, diagram generation, and literature discovery inside a unified workflow.
Episode 13 - The Thinking Behind Ads in ChatGPT
The discussion centers on OpenAI’s approach to introducing ads in ChatGPT, balancing business viability with user trust. Assad Awan explains a principled framework: ads must be visually and functionally separate from the model, be useful and non-intrusive, respect user privacy, and empower user control, with different treatment for free versus paid tiers and enterprise. The team emphasizes a trust-first, transparent, and iterative process to evolve ad formats while ensuring high-quality, non-exploitative experiences.
Build Hour: Prompt Caching
The session explains prompt caching and why it’s a powerful way to reduce latency and cost when using OpenAI models. It covers how caching works (prefix-based, token blocks, implicit vs explicit caching, and extended prompt caching), key concepts like cache keys and context engineering, and demonstrates practical trade-offs and best practices with live demos and real-world examples from Warp and other customers.
Builders Unscripted: Ep. 1 - Peter Steinberger, Creator of OpenClaw
A candid conversation with Peter about the rise of OpenClaw, the open source mindset, and how AI agents are reshaping how developers build, collaborate, and deploy software. It covers personal journey, community adoption, and practical advice for embracing agentic tools safely and effectively.
Build Hour: API & Codex
A Build Hours session explores Codeex, Codex, and Harness Engineering updates, focusing on AI-assisted software development, agent delegation, and new tools like hosted shells, skills, apps, and work trees. The speakers demo workflows, security and reliability patterns, and how to scale product teams with structured guardrails, documentation, and cross-team collaboration. Mitch from Basis and the co-founders discuss pragmatic, company-wide adoption, including brownfield deployments and ongoing improvements.
Building AI for better healthcare — the OpenAI Podcast Ep. 14
OpenAI researchers Dr. Nate Gross and Karan Singhal discuss the OpenAI approach to healthcare AI, emphasizing safety, grounding in medical knowledge, and collaboration with clinicians. They explore HealthBench evaluations, data privacy, how to integrate multi-modal data, and real-world deployments like clinics in Nairobi, highlighting goals to raise access, reduce clinician burden, and ultimately improve patient outcomes while maintaining trust and safety.
Episode 15 - Inside the Model Spec
In this OpenAI Open Eye podcast episode, Andrew Maine talks with Jason Wolf about model specs: what a spec is, why it matters, and how it guides model behavior beyond what is actually implemented. They explore how specs are developed, the chain-of-command for resolving conflicts between policies, the role of transparency and public feedback, and how training and iterative deployment shape alignment with the spec over time.
Get daily AI recaps from
OpenAI in your inbox
Get AI-powered summaries delivered to your inbox. Save hours every week while staying fully informed.