PhD studies can be very different. Wet-lab students who do experiments can spend 12 hours a day in a lab cutting mice. But that’s somewhat mechanical work where you repeat the protocol. PhD students are also less experienced and spend a lot of time making mistakes and learning. When I got my fist full-time developer job I spent day and night working as well. After 5-7 years, I was something like the author because I could do non-trivial tasks quickly. It felt like the code was writing itself in the back of my mind while I was drinking coffee or browsing HN.

My PhD, in physics, was a mix.

When I was building the physical apparatus itself and plumbing it up, I was working 70 hour weeks because it was more or less mindless physical labor.

When I was writing software to operate the experiment or analyze data, I settled into more of a few hours a day of deep work pattern. Part of that might be the nature of science; after I implemented an idea, it might be hours before the experiment fully responded to my changes or the computers finished crunching numbers. And I needed that feedback to inform what I'd do next.

Writing my dissertation was somewhere in the middle. The actual writing required focus, but a lot of it was tracking down references or making figures, which required some thought but not too deep. That's probably the closest I came to a 40 hour week.

Based on other replies, it sounds like other people have experienced similar relationships of the nature of the work to their capacity to do it.

Like the author says, I think this is pretty normal. Deep work in a flow state is easily 10x more productive than when you're messing around.

10hrs of deep work a week is also a significant amount of time. I doubt most people spend that much time doing deep work each week.