Robert Brennan comments

Results 1267 comments of


                                            Robert Brennan

Default to a less expensive model

Seems like the vote favors 3.5! Might need another pass through this to catch all the `gpt-4` references

Move config defaults into `config.py`

Yup exactly. We've also got some env vars for things like `DEBUG`, `EMBEDDINGS`... Would be nice to collect it all in one place, and to ensure that everyone is using...

Move config defaults into `config.py`

Here's a good example of why this would be helpful 😄 https://github.com/OpenDevin/OpenDevin/pull/475/files

Task testing / Evaluation.

I think this is related to https://github.com/OpenDevin/OpenDevin/issues/341

Task testing / Evaluation.

#341 is about testing LLM-generated code, which I think is what you're after here. E.g. we have a task like "write a bash script that prints hello to greet.sh", and...

fix frontend : highlight active assistant text and maintain chat seq | for #476

This is working great! The one bug I see is that if it prints several lines, the go below the fold, and you have to scroll down to see the...

fix frontend : highlight active assistant text and maintain chat seq | for #476

![story](https://github.com/OpenDevin/OpenDevin/assets/7611973/f476ea3f-d879-44bb-bc74-cdb4e3102304)

fix frontend : highlight active assistant text and maintain chat seq | for #476

👍 let's get this one in

fix frontend : highlight active assistant text and maintain chat seq | for #476

Thanks for the fix!

[Bug]: Commands getting printed multiple times

I think this is likely a frontend issue now--I'm not seeing it when running OD headless, and the websocket data seems fine