I quit computing in the late 1980s as a result of an early mid-life crisis in my mid 30s. After a few years getting myself back together, I ended up in an M.Sc. at the University of Durham, UK, paying my own fees and for my food and lodging. Up to this level is easily manageable, because it's essentially just learning. What can make even this level not work out is what makes the next level more difficult to negotiate — real research.
With some life experience behind me, and having read around quantum theory and having come to naive opinions about what was "wrong" with quantum theory, I almost didn't finish my M.Sc. project, and I was only just adequately prepared for the examination component, because I went off on an almost unguided goose chase after a particular piece of mathematics (Clifford algebras, a pied piper if ever there was one). One academic managed to keep me from going completely off the reservation. I've chased after many other pieces of mathematics in the 20 years of research since, so I have a certain perspective on realizing that I've wasted another year or several on another piece of mathematics.
What an older person can bring to research can be a real asset and a real pain for an academic to have to deal with: bull-headedness. If you're older than the academic who is guiding your research, you nonetheless have to take their guidance, at least up to a point. When an academic looks at you, they will be choosing you over someone younger who is probably also more malleable. An academic has to decide whether they want someone who will be no trouble (and probably also not produce great research of their own in the future, but nonetheless be smart enough to help the academic's research go forward nicely, which is a delicate equation), or be a lot of trouble (but the academic can hope for and look forward to a satisfying experience of seeing their mentoring bear fruit many times over). Of course an academic's ego is a fundamental part of who they choose to take as graduate students. It has always seemed to me that this calculation is generally rather different when an academic considers a mature student. Being a real mentor to a graduate student is exhilarating and exhausting if they're really good at research, but could be worse than pointless if they react badly when they discover they're not.
Twenty years ago, I wanted to make some academic let me do things my own way. That didn't work out (!), because I couldn't find an academic who would put me in the place they had instead of taking someone much more promising-looking and just graduated. Putting a brave face on this, I've sometimes thought that it would have been immoral to take those places from the young people who beat me out of those places. Consequently, I've pursued research independently, with an M.Sc. but without a Ph.D., which can be done if you're persistent and you can find a way to get good access to an academic library. Although you will make much better progress in research if you can find a good real-life community, I think it may be better to spend your money on access to a library and to find effective ways to be part of the on-line community (giving as well as taking, of course).
[My qualification for these observations is somewhat sideways: I'm married to a Classics professor, so I see from quite close up, but still from secondhand, and in an arts subject, the process of becoming an academic researcher, which many promising undergraduates never manage. Second-guessing who will make the jump is a large part of taking someone on as a graduate student. Needless to say, these comments are bashed out in a half-hour, they're incomplete, they're not gold dust, and you're not me. Good luck.]