> However, do you think putting 2 hrs a week on a single project for extended time will take me to places?

What places? Getting a job in the field? I think it could do it, unless the field is very hard to get into (e.g. 3d computer vision/vSLAM is considerably harder to get into than computer graphics, machine learning research is much harder to get into than applied machine learning (glorified data science) etc.). A good approach is to talk to some insiders in the field to get a feel of how hard would it be. If you don't know anyone, topical subreddits are not the worst place to ask this.

I went through exactly this path, even got some interviews, but learned that the jobs that would interest me the most are hardest to get into, and would mean a large pay decrease (I was making a ton in my current area of expertise). The nail in the coffin was chatting with some guy from Facebook AR (I met him on reddit) team who had my dream job and who said that, in my position, he'd prefer to just accumulate money and FIRE in 5 years, instead of starting a new career.

To add small amount of context, I do also enjoy learning when its not a 'tool' but its a very narrow margin of the medium. For example, i've probably watched every single 3blue1brown video on youtube there is, yet I won't really use any of that. I watch them while I'm eating and any time they come out, as well as a few similar channels.

But this sort of learning is more about the purpose of stimulating my mind with challenges rather than acquiring knowledge.

80% of my time learning is when its a tool, 20% of the time its for stimulation, 100% of the time it is for some meaning other than it being satisfying to collect knowledge and facts, which is 0% of the time.