Why Dwarkesh is Wrong about Computer Use + How OpenAI shipped its Jev competitor in 1 Week
Source
latent.space
Date
Context
Computer-using agents can avoid some repetitive clicking and scrolling when they can read an application's structure. In this DevDay interview, OpenAI's computer-use lead describes combining visual information with accessibility data and the underlying structure of web pages. That lets an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → inspect more content and write code that performs several steps together.
The interview's App Shots example makes the difference concrete: a screenshot can omit a link's destination or cut off a calendar title, while the accompanying application data can retain that information. The account explains why better access to software can improve an agent's work even without changing the task itself.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
prompt caching — Reusing the model's processed form of a repeated prompt prefix so subsequent calls skip re-reading it, cutting cost and latency substantially.
context compaction — Summarizing an agent's earlier conversation to free room in the context window so a long session can keep going.
