I can't help but constantly ask the question of whether there is a better way, whether the existing approach is even the right way, and whether by seeing a similarity in the essence of a problem with something in a different field altogether (than technology) we might find a more interesting and successful solution.
Don't listen for problems that you might work on, just listen to how things currently work and be inquisitive. Is the way something works today the best way? Does it represent the final iteration?
You already see with evolution that things change over time, and so will solutions. Once you've accepted that everything will change and get better, and start forgetting the concept of "Everything's been done already.", you'll start to see room in everything for the next small improvement, or if you're lucky (and transferred knowledge from one domain to another) some major improvement.
With side projects it obviously helps to pick small things. You want to be done in a weekend, or a few months at the pace of an evening here or there.
If you would like a side project of interest, the last couple of weeks I've been nagged by the thought of privacy online and the idea that the internet fails to allow people to learn from their past mistakes and errors as the internet does not forget and forgive and so people can never move on and improve.
As such, I've been pondering what forgetfulness might look like on the internet. And I think it must be possible to create a little Python script that takes credentials for social networks and various services online and helps introduce forgetfulness by deleting old content. Perhaps not everything, mimicing memory the lesser events initially, and then the major events eventually. Say... everything over 12 months old that received no likes/retweets or engagement gets deleted immediately, and higher engaged items of more significance get flagged up and you can opt to forget them or not.
What does forgetfulness look like on the web, and how can it be applied?
You can see the fundamental problem, "Lack of rehabilitation within society and the ability to better oneself", and you can see a proposed solution, "Bring forgetfulness to the web". If the problems are real enough and chime with people then you'll likely find people willing to pay. In my example the likely market are those in their mid-20s just starting on a career but having lived a lot of their life online and who may now regret the wide-exposure given to their drunken escapades a mere 5 years earlier.
I'd also say not to look at side projects to make money. They're mostly inquisitive exercises. Some make money, but most scratch an inquisitive itch.
I'm quite frustrated that I barely get to work on the thoughts I have, but even then they never go to waste. Even the ideas not pursued help to gradually improve my understanding of how things work and new angles of attack for existing problems. Just question everything.