[burkaman]

It's interesting that the screenshots advertising the bot make it seem totally useless. In the first example the bot comments "You should add tests for this function". The author asks "do I really need to?" and the bot replies "No actually you don't, just make sure it works". In the second example the author asks the bot for a review, and it just copies and pastes the code change and says basically "looks good to me".

[venv]

I suppose the productivity improvement comes from not having to wait for those same lines from a colleague?

[winkmaster]

I agree with you, I think the examples need to show the full potential of the bot. This being based on ChatGPT, the feedback can really be astonishing if the use case makes sense.