Do you lean more towards "fun" projects or trying to solve "serious" challenges?

Not OP, but do you by chance have ideas for "serious" challenges? I would be interested in some

One I'm very interested in that I don't have the time or expertise to really tackle is single-lens photogrammetry (or photogrammetry in general). Right now there isn't a great consumer-friendly option that doesn't require custom hardware. Many online resources recommend specific lighting and high-contrast colors to get a barely-accurate scan.

Another similar-ish one that other people have started addressing is building a system that reliably detects photo/video manipulation (ie. photoshop & deepfakes). An interesting extension of that would be detecting audio manipulation in a video.

Contribute to open source projects you use... at least in the meantime.