Lenny's Podcast: Product | Career | Growth

OpenAI’s Head of ChatGPT: We’re entering a new era of AI (again) | Tibo Sottiaux

Oct 4, 2026 · Summarized Oct 4, 2026

Listen in Apple Podcasts Pocket Casts Castro

Key takeaways

Most actions on the internet will soon be carried out by AI agents, and building products that scale for these agents is essential.
AI systems will become cheaper and faster at an incredible rate, enabling seamless integration across modalities and platforms.
Loops and manual workflow tuning are passing phases; the future lies in systems that learn and adapt autonomously.
Skills like taste, user understanding, and creativity will become increasingly valuable, while routine skills like typing speed are less important.
AI safety and alignment require heavy investment in monitoring, guardrails, and responsible deployment, but there is strong optimism about solving these challenges.
Community involvement and open ecosystems are crucial for expanding AI adoption, discovery, and economic sharing.
The cultural environment at OpenAI emphasizes autonomy, ownership, and rapid iteration, allowing teams to operate with high trust.
Humans will remain valuable for their creativity and ability to learn and adapt, especially as AI handles more routine tasks.
Future AI interfaces will be more natural, using voice and multimodal inputs in locations like meeting rooms, reducing fatigue.
Building an ecosystem of plugins and integrations, with shared economics, will be key for widespread adoption and utility.

Summary

Introduction to the Episode and Setting the Context

The episode begins with Tibo Sottiaux discussing the limitations he sees in current AI projects, emphasizing the need for building ten times better systems within a year. He highlights the importance of designing for a future where most internet actions are performed by AI agents and stresses the significance of building systems that learn and adapt without requiring meticulous looping or manual configuration.

The Vision for AI Agents and Their Evolving Role

Tibo shares his view that the future of work will be radically transformed by AI agents that understand goals, preferences, and learn from feedback, operating seamlessly across devices and contexts. He describes his ongoing experiments with teams of agents expanding and shrinking based on breakthroughs in models, illustrating a dynamic process of evolution in AI capabilities.

Building and Operating the Dots Platform

The episode delves into the development of Dots, OpenAI's personal agent platform, and how it integrates models like Codex and ChatGPT. Tibo talks about the sci-fi inspirations behind these efforts, the community-driven approach to building AI, and the technical structure of Dots. He discusses how plugins are discovered and recommended through user engagement metrics and emphasizes the ecosystem's importance for broader AI adoption.

Human-AI Collaboration and Maintaining Human Value

Tibo reflects on the enduring value of human creativity, taste, and ingenuity, emphasizing that AI should serve as an extension that empowers humans rather than replacing them. He also talks about the psychological and social challenges faced by engineers, such as fatigue and loneliness from constant interaction with agents, and anticipates future solutions that will make AI collaboration more natural and less burdensome.

OpenAI’s Safety and Future Outlook

The conversation shifts to AI safety, where Tibo outlines OpenAI's focus on safety, alignment, and responsible deployment, stressing that their incentives align with societal well-being. He expresses optimism about solving safety challenges through intensive investment in monitoring and guardrails, and he acknowledges the importance of community involvement and openness in fostering trustworthy AI technologies.

Reflections, Changes in Perspective, and Closing Thoughts

Tibo admits that some of his earlier beliefs about rapid capability milestones have evolved, especially regarding the importance of voice and the pace of progress. He highlights the talent influx from new generations, advocating for greater democratization of coding and AI tools. The episode concludes with reflections on the importance of community, trust, and continuous learning in building AI for the future.

Show notes

Tibo Sottiaux leads ChatGPT and Codex at OpenAI. Under his stewardship, OpenAI has shipped some of its most consequential consumer launches: Codex, ChatGPT Work, and most recently, the new Dots personal agent platform. He joined me at DevDay, hours after his team launched more than 20 new products, to talk about where things are heading.

We discuss:

1. Dots, OpenAI’s new personal AI assistant

2. Why most actions on the internet will soon be taken by agents

3. Why loops, graphs, and fine-tuning agent workflows are a passing phase

4. Which skills are trending up and down in the AI era

5. How Tibo’s Dot warned him about a production outage five minutes before the launch

6. OpenAI’s approach to AI safety

—

Brought to you by:

WorkOS—Make your app enterprise-ready, with SSO, SCIM, RBAC, and more

DX—Engineering intelligence for the AI era

—