How does the work you do change you? A reflection after wrapping up three hobby projects:
I hit an early hard moment while building suffice
(LLMs: none) because I had carved my abstractions very closely to the
underlying implementation in an attempt to avoid premature abstraction.
This was correct in principle but left me frustrated when it came to
testing things with the tighter rules of rust. Struggling through this
gave me care
for the end result and also confidence that constraints are
clarifying. Not every jot is right but I had enough swagger in the
output to publish it to
crates.io.
I was very hands-off with saline
(LLMs: entirely) and spent more time in the business domain (why lifting
logs kinda suck, what the interference effect is) but the process heavy
approach left me frustrated: I was answering questions about “how it
ought to be” instead of just doing the work. I might use the result but
I might just use the learning to find a tool that is least-worst. I’ve
at least published this to a public repo but am unlikely to maintain the
tool in the long run – it’s disposable.
saddlemap
(LLMs: mixed) was intentionally disposable – it exists to do
something fun I hadn’t seen before! I could not have gone as fast as I
did without the taste of what a good abstraction was from
suffice or the dis-taste for spec-driven development from
saline. This was heavily experimental work and I felt less
attached when I added tests at the end after the dust settled. I was
left with a sense of unease at the clash of hard-won habits of craft vs
disposability of output. I spent more time honing the craft of project
management than I did of coding in rust.
To do engineering is to master tools and taste. How have you shaped your tools to your taste and how have they shaped your taste in turn?
How does our relationship with the output of our work change as the rate of production increases? As more and more output is treated as fundamentally disposable?
Socials