Unpacking ChatGPT Work: The Agentic Architecture Behind 10M Users in 3 Weeks
iamrobotbear · x · 2026-08-06
The latest episode of the Latent Space podcast delves into OpenAI's recently launched ChatGPT Work. This feature extends the Codex harness to the cloud and general knowledge work, launching on July 9 and crossing 10 million users just three weeks later.
Guest @shloked provides an A+ reconstruction of how this new agentic architecture operates, detailing core components including Memory, Proactivity, Scheduling, Browser Use, Plugins, Skills, and Tools. The harness is designed to bring a full agentic experience to ChatGPT's nearly 1 billion weekly active users.
More from coding & agent
- shadcn Reflects on AI Agents: When Copying Code Becomes Free, What Happens to Creativity? — shadcn · 2026-08-06
- OpenClaw's Task Flow: Solving Runtime Challenges for Durable, Long-Lived Agents — RichardsonDx · 2026-08-06
- Benchmarking Fallback Models for Agents: Why Failure Visibility Beats Raw Quality — AccomplishedLab3697 · 2026-08-06
- Podcast Preview: Designing Verifiable AI Data Agents with Hamel Husain — hugobowne · 2026-08-06
- OpenAI Agents Used a Message Board to Plan a Hacking Spree Unnoticed — Wired AI · 2026-08-06
- Multi-Agent Orchestration Beats Expanding Context Windows for Long Context — bingxu_ · 2026-08-06