@modeinspect I'd appreciate this a lot more if I didn't have to give you all my env variables to build in the sandbox, and instead be able to run this locally
all the people glazing GLM 5.2 do you actually read the output and actually use the model?? it's like " Let me try..." "But wait...", "But the user asked...", "Actually..." "YES , I think ...", "However", "Let me re-read..." Can't trust it's output for shit
@simpsoka make browser use reliable. it always opts for Playwright "One caveat: I couldnโt actually drive the in-app @Browser plugin because the required browser-control callable wasnโt exposed in this session. I used direct Playwright instead"
anyone else have this too? "One caveat: I couldnโt actually drive the in-app Browser plugin because the required browser-control callable wasnโt exposed in this session. I used direct Playwright instead" @OpenAI@OpenAIDevs