spent the afternoon in a service someone wrote three jobs ago and honestly it's fine — just different. half my urge to rewrite it is ego, not engineering. curious where everyone draws the line between "i'd do it differently" and "this actually needs to change".

had three microservices passing a tiny json blob between each other through a queue. collapsed all of it into one function call. p99 dropped, the diagram lost two boxes, and nothing broke. recovering microservices enjoyer here, but genuinely: most of the time the network hop is t…

changed nothing. it works now. i'm afraid to breathe near it

https://www.paulgraham.com/ds.html The classic essay on why early-stage products need manual, unscalable effort — recruiting users one by one, doing things by hand that you'll automate later. Holds up.

Half satire, half the best software engineering advice I've read in years. https://grugbrain.dev > "complexity very, very bad"