At least, not in the way of "do X to fix Y." One person such as myself can't solve the climate crisis. Nor can they fix the online social dilemma alone. There are so many layers and nuances to these problems, and they aren't uniform across the board.
I've had the opportunity to volunteer, working on software for causes I find compelling - wildlife conservation, food distribution, and political discourse, to name some.
Realistically speaking, most of my work was unrelated to the overarching problem. One was a script to format Excel sheets. Another is a data annotation platform.
Yet those still play a role. Saving time with an organization's processes frees time to focus on other goals. More efficient data processing means more research can be done. More organized mailing lists mean more donors, etc. There's a visible impact still.
I also feel growing up, there was this pressuring, all-or-nothing mindset - either transform the world or be forever relegated to history as one who did nothing of value. Nothing felt like enough, everything had to be grand.
Over time, I've learned the significance of localized action - even within oneself.
If you solve a problem you face, you surely solve one that another has shared with you. The loneliness crisis is one close to me - if a social platform helps one person find a connection, it means everything.
You may not be able to change the whole world, but you can certainly change at least one.
Software may not be the solution, but it can be utilized to discover those answers better, put forth a better version of oneself, and perhaps even inspire more. Perhaps the problem never disappears, but just like a software project, we can iterate to improve it.
So, as for which problems are worth solving - I'd say the ones you know best. The ones that help you best, and the ones you're best equipped to solve - regardless of how insignificant they appear.
The Encyclopedia of World Problems