For Computer Use, the harness matters as much as the model

Re:For Computer Use, the harness matters as much as the model by smpandya on Thursday September 24, 2026 @01:42PM (#49834193)

Are there results comparing agents running different tools (agent-browser, playwright MCP, browse CLI), or is this mostly Stagehand focused?

[ Parent | on HN ]

Re:For Computer Use, the harness matters as much as the model by MiguelG719 on Thursday September 24, 2026 @01:48PM (#49834300)

You can swap the driver/tool yourself with the evals CLI! Just use `—tool-surface <one-of-the-supported-tools>`. Or define your own using the interface

[ Parent | on HN ]