Factorio is right on the cusp for me. SpaceChem, too. TIS-100 was a bomb for me; if I'm literally writing code, it better end up in a GitHub repo.
With Factorio for me, I always reach a point where I know how I'm going to scale up, and then I lose interest in actually doing so. Just knowing that I could do it is satisfying enough.
I love construction/management games but I very rarely "finish" them for exactly this reason. The moment I can see exactly how I would go about making it to the end goal, I'm done.
I haven't identified with a HN thread in quite a while. The moment I understand how to do something or that it's possible, I'll drop it. That includes my personal projects.
On the one hand, it is a vastly better debugging experience. You can see the entire state of the machine, and easily step through the execution.
In contrast, with my day job, I'm dealing with incomplete documentation, poorly designed and very large libraries without adequate comments or organization, subtle bugs that only occur in unidentified situations, etc.
But at the end of the day, TIS-100 felt more like work or a hobby project, so I'd be better off doing one of those.
With Factorio for me, I always reach a point where I know how I'm going to scale up, and then I lose interest in actually doing so. Just knowing that I could do it is satisfying enough.