Inside OpenAI DevDay: superhuman computer use, Decisions API, and the AI cloud
Latent Space · youtube · 2026-10-01
Latent Space hosts OpenAI's Ari Weinstein (Computer Use) and Nikunj (API team) right after DevDay. Key points:
Computer Use
- Ari says Computer Use is "180 degrees different" from months ago; some real-world tasks now complete faster than the average human
- Dots gives every agent its own Linux cloud computer
- Speed comes from combining screenshots, accessibility trees, the DOM, Playwright, and generated code
- Agents now debug and recover from failures; the goal is literally superhuman software operation
Developer stack
- Async tool calling means models keep reasoning while tools run
- Mid-turn steering, WebSockets, GPT Live for real-time control
- UltraFast inference; Decisions API as low-latency structured decisions, already used internally for support classification
- Longer prompt caching, cache pre-warming, server-side compaction for long-running agent threads
- OpenAI's inference teams use coding agents to optimize their own systems
Big picture: whether OpenAI is building an "AI cloud" with higher-level primitives for agents — memory, compute, state — plus trust, permissions, and safety when agents can make payments.
Related event: OpenAI Breaks Down Its New Agent Stack After DevDay(3 posts)→
More from coding & agent
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- shadcn boosts @uisdk: one-shot UI rebuild that's instantly E2E-testable by AI — shadcn · 2026-11-16
- AI agent deletes 43,599 lines of OpenClaw test code in one PR — vincent_koc · 2026-11-16