In general I agree, though I'll note I did make it through to graduation and beyond when I went back to school and used my pretty old laptop to learn coding and analytics.
Granted, there would be a meaningful difference in if I did that for a living with real-world systems vs. what we did for class assignments. iirc, the largest dataset we used in class was only 4 million rows and maybe 100 columns. And the most taxing ML stuff I ran only took about 15 minutes to perform, whereas if I were doing that "for real," a machine like mine would take exponentially longer to perform an ML task, and of course I'd be doing it a lot more, so that probably would reduce a lot of life. OTOH, generally I'd expect to have a work-issued laptop if I did that stuff for a living.
So as a true Personal Computer, I seem unlikely to do a lot of heavy lifting with my own machine.
That said, I'm still highly averse to something like a Chromebook. My PC may largely function like an internet portal, but I just can't exist in that environment. I do some stuff that needs a legit machine.
