• During my PhD, I've realized that I've needed to access two "modes", and how these have shifted as I've gone from programming as a kid to programming as an adult for a "job" / "research".

§ Mode 1: Binge Programming

• This was me as a kid, all through undergrad, and the first half of grad school.
• Essentially, programming was a video game, it was art, it was something I did obsessively till I felt "satiated".
• Once I had reached satiation, I would drop what I was programming, and wander away for the next few days to read a book/paper, ideate, dream about the next cool feature my game needed, or study some maths.
• This worked brilliantly as a kid, I wrote many tens of thousands of lines of code, and kept a "work life" balance by just wandering away to do my thing when I was done programming for the week.
• This worked brilliantly in undergrad, as all the projects and assignments I had were "made for success". Sure, they needed thought and hard work, but all the assignments were such that after the key insight, one could indeed just binge-program one's way through the solution.
• During my undergrad internships (GSoC, Tweag, etc), I was doing "research engineering", where once again, the loop of "implement a well understood algorithm from the literature" was the standard I was held to, which works really well with the cycle of "wander away, read a paper, come back, binge-implement".

§ Where this fails in grad school

• I realized that in grad school, I was genuinely being asked to solve problems that nobody around me knew how to solve. It's subtle, and feels extremely similar to a textbook exercise at first.
• It only dawned on me half a year later, when we had submitted a "proof" that turned out to be incorrect, that no, this was genuinely subtle and hard !
• I could no longer spend three days cleverly thinking about a problem to find the kawaii insight that solved it.
• I also felt that this mode of programming, when it ensured a steady rhythm, meant that I lose my "burstiness" where I would stay up for days, living like a monk, just hacking on something till I brought it to life.
• I wanted to gain that edge back, where I could justify for myself the days I would lose "wandering away" once I had brought something back to life. This felt incompatible with the demands placed on me to be "regularly making progress".
• This brings me to the other mode:

§ Mode 2: Persistently Showing Up To Not Solve the Problem

• As the title says, one simply shows up to work everyday (no waiting for inspiration!), tries to solve the problem, fails, and goes back home a little more enlightened.
• While this has helped me make actual progress, I feel that this mode has made me lose the "spark of joy" I harbor for bringing a concept from the platonic realm to life via code.
• I noticed this when I was interning at AWS, where I was suddenly feeling much more content with the day-to-day programming, something I hadn't anticipated.
• It's because I'm back to doing easy things again! We understand how to solve the problem (more or less), and we solve it! It's great!

§ The Synthesis

• I think what I need to do is to reclaim the joy of binge programming in academic research.
• Furthermore, if I want to regain the "burstiness" where I can just churn out code for a week, I must accept that this will mean that I will spend the next three days contemplating about Borges or something that's not the immediate problem.
• I'd like to consciously make the switch, when I feel that the time is right to crank out a couple thousand lines of code, by explcitly discussing this with my advisor.