NixOS VS Research

I am wasting my time. I spend so much time with nix OS. How can I do this when the AI frontier is advancing very fast in a direction of superintelligent machines we don't understand at all. Why do I do this? What need is fulfilled. I have to say it feels very good when I get something to work. I am doing something that is complicated and I can see that I am getting better. But really there is not much time. I need to make such that I am too addicted to research, but how do I do this? How can it be that doing research still seems the hardest possible thing. It seems that nix OS is easy in comparison. Really Nix OS is a sort of hobby. I can't stop myself from engaging with it. How can I make it such that I can't stop myself from doing research? I need to optimize my research for being fun. I am already pretty close. I think the problem is probably that I am not getting to the "YEEEESSSS I did it" point that much. And if I do then it does not feel as good. When I got something to work in nix OS then it is clear that I got it to work. I can see it working. When doing something in research I never get to that point. I never have a very good idea about what I even want to get. So figuring out a concrete deliverable, and then actually working to deliver might be helpful. Another issue is probably that I don't usually cash out some thing completely. I never push to the end. E.g. I could have implemented the robot arm inference and then check if it actually works. That would probably have been satisfying. But instead I moved on. So how about I commit to push on hard, in really any direction that still stands. Especially programming up the algorithms that I am thinking about and making them actually run, or at minimum write pseudocode. Another issue is probably that I don't usually cash out some thing completely. I never push to the end. E.g. I could have implemented the robot arm inference and then check if it actually works. That would probably have been satisfying. But instead I moved on. So how about I commit to push on hard, in really any direction that still stands. Especially programming up the algorithms that I am thinking about and making them actually run, or at minimum write pseudocode.