Hacker News Reader

I feel that a lot of this can already be achieved via aider (not affiliated), and any of the top models.

Do you have any benchmarks to back up your 'feelings'?

I really don't like the snarky tone of the parent comment.

Nonetheless, I don't think this is even something that can easily be benchmarked. I'd recommend you take a look at aider [1], and consider how I drew similarities between it and what's presented here.

Has ClosedAI presented any benchmarks / evaluation protocols?

[1] https://aider.chat/

Yes, they show benchmarks in the article linked here. Did you not read it?

I don’t think you actually read it. The benchmarks are in reference to the model that’s underlying deep-research, and not deep-research itself. For the latter, they have anecdata from scientists.

Reply on news.ycombinator.com