After initially starting coding as an escape from depression and being a shut in, all because my brother wanted to make a game (but couldn't be bothered with the coding), I realized after many years that in the end, what I find joyful in programming is the flow state and
learning & understanding.
This made the university mantra of "contribute to open source for your resume. Show off your github profile" really hard for me to adapt in practice... I tried hard but I just struggled to finish project because as soon as I understood it to my content, which usually stopped at "ok I understand it now and I vaguely have a plan how I would implement it if I have to".
The good place for me to be in has therefore been coding for R&D, since understanding things are often as valued as implementing it. Consequently I often become the goto guy for debugging and understanding but maybe not the guy for "implementing it fast".
Also why I feel mixed with agentic coding... Humans don't really learn without iterations and doing it yourself; it's easy to fool yourself that you understand something after following someone's solution to a problem. But on the other hand, since I do have good understanding I should be able to be thorough with my reading of agentic coding (and therefore often becomes the goto guy for PR reviews for others vibe code).
In the small robot projects that I've had recreationally, I've struggled to find a balance. For getting things done quick, with the limited time I have, agentic coding is wonderful. But understanding takes a hit and I struggle to keep myself motivated of it. (exception are like plotting scripts. I can't be bothered anymore to wrestle with matplotlib everytime)
This made the university mantra of "contribute to open source for your resume. Show off your github profile" really hard for me to adapt in practice... I tried hard but I just struggled to finish project because as soon as I understood it to my content, which usually stopped at "ok I understand it now and I vaguely have a plan how I would implement it if I have to".
The good place for me to be in has therefore been coding for R&D, since understanding things are often as valued as implementing it. Consequently I often become the goto guy for debugging and understanding but maybe not the guy for "implementing it fast".
Also why I feel mixed with agentic coding... Humans don't really learn without iterations and doing it yourself; it's easy to fool yourself that you understand something after following someone's solution to a problem. But on the other hand, since I do have good understanding I should be able to be thorough with my reading of agentic coding (and therefore often becomes the goto guy for PR reviews for others vibe code).
In the small robot projects that I've had recreationally, I've struggled to find a balance. For getting things done quick, with the limited time I have, agentic coding is wonderful. But understanding takes a hit and I struggle to keep myself motivated of it. (exception are like plotting scripts. I can't be bothered anymore to wrestle with matplotlib everytime)