inflational appearance of copilot on this project

[yoshimo]

I am doubting the usefulness of ai to review and create code. For me these tools never created anything useful that compiles and usually is far more complicated than necessary.
Instead of undermining the developers credibility and skills and the projects reputation, you should turn this copilot stuff off and don't let it clutter issues and repositories.

[ignoramous]

For the network engine, 40% of the code is all Copilot, and that has been the case since I first started using it in ~2022.

You're right that Copilot cannot handle some tasks, but it does handle some tasks very well. Different models on Copilot have different strengths, like Gemini 2.5 Pro is good at refactor and bug hunting, Claude Sonnet is simply good at code understanding and generation, while GPT-5 is a step up from GTP-4 for agentic tasks.

don't let it clutter issues and repositories.

A genuine concern. We only use Copilot integrated with GitHub Issues only for straight forward tasks (what they call "good first time issue"). Not for everything. Especially, not for features.

[ignoramous]

Possibly, but anyways, we've been using Copilot since 2021, its first release.

[error-reporting]

but v0.5.5o+ release coincides with increased Copilot use for core code

[ignoramous]

Possibly. Though, "Copilot" alone can't be the reason for the bugs, as v055o in general added a LOT of new code/features/behaviour. It was also a rewrite of large portions of the firewall and network-related logic.

[error-reporting]

I think removing existing already published releases beyond v055n should be done

[ignoramous]

removing existing already published releases beyond v055n should be done

Also a good idea.