Overview

In this fireside chat, OpenAI’s Head of Engineering reflected on his journey from MIT to OpenAI, sharing that DALL·E 2’s breakthrough in 2022 convinced him the field had crossed a threshold worth joining. He described his current role leading the OpenAI Platform—which includes the API, enterprise tools, and partnerships with governments—and explained how each new generation of models (GPT-3, GPT-4, GPT-5) expanded what developers could build. He was candid about the chaos of scaling after ChatGPT’s explosive growth, calling it “everything broke at once,” and about the mixed early reaction to GPT-5, attributing it to trade-offs around slower “thinking mode” and a launch strategy that didn’t showcase its strengths. Looking ahead, he highlighted reasoning and multimodal advances—especially long-term agents and more natural voice interaction—as the next frontiers. For startups and enterprises, his advice was to “build for where the models are going,” ensure both top-down and bottom-up AI adoption, and pair innovation with responsible use and robust guardrails.