[jeffreyrogers]

The way I think about software is it's basically about processing data. Most of the time that data is someone else's, but sometimes it is generated either by the software itself (e.g., CAD or audio applications), sometimes it is free information but just collected in a useful way (e.g. Google), and sometimes it is proprietary stuff that is hard to access in other ways (lots of financial stuff fits in this category).

The market for software ideas is pretty efficient. So most ideas that you can come up with off the top of your head either already exist or didn't work out for some reason that might not be obvious. The way I'd approach this question is to look at what new data sources have become available. One of the comments mentioned OpenStreetMap and that seems pretty reasonable to me. Improvements in facial recognition and object classification are probably enabling new applications (not all of them will be positive though). Biology/medicine is probably the main field that's had major new advances, obviously mRNA vaccines but also things related to genetics. So I'd bet something important would come out of there.

I'm skeptical that software will have a major impact on things like energy usage or preventing armed conflict, etc. Those are more political/economic problems.