My commenting has really gotten a lot better since Github Copilot.

# do my code like this
...code

# and then also do this with that
...lmfao

🤣🤣🤣🤣🤣🤣🤣🤣

AI is The GOAT 🐐🎉

Code completion suggestions have got a little better these days, and sometimes they even contain the correct variable names and values. But do we really need the LLM's waste of energy and memory consumption to save us from typing to the end of the line? Algorithms would have been enough.

Code comments rarely make sense, but my typical example is something like

// TODO remove workaround when #123 is fixed

which adds external context (some say that's bad) and can't be understood without external context. Likewise, commit messages referencing GitHub/GitLab issues, for example

git commit -am "mobile header layout #456"

I wonder how an AI should help anyone get those correct?

@ingosteinke

nothing wrong with TODO comments !