Zengyi Qin
@qinzytech • 4,789 subscribers
Multi-modal Agent Research (vision, audio, computer use, etc) | MIT PhD @MIT | @agiopen_org
Videos
How can Codex control Apps you can't even see? 1. Screenshots via macOS system API. Works even when minimized or hidden behind other windows. 2. Clicks and types go directly to the target process. Nothing else is affected. It's operating at the OS level, not the UI level.
Zengyi Qin809,322 views • 3 months ago
No more content to load