essays
the posts that argue rather than demonstrate — changing my mind about testing in public, splitting a stack instead of switching, and what vibe coding actually costs
most of this blog is a bug and its post-mortem. these are the ones that are an argument instead.
what does this blog cover about opinion?
the parts of engineering that are judgement rather than fact, written with the cost included. i try to keep two rules here.
the first is that an opinion post has to be falsifiable by my own work. i quit testing is a retraction of a post further down this same site — i argued in 2024 that testing was a waste of time for a solo developer, and then a lost update ate 79% of my data because there was no test pointed at the one path where correctness actually lived. leaving the original post up matters more than the retraction does.
the second is that a preference is not an argument. i didn’t leave go for rust exists because “language x is better” posts are cheap and useless; the interesting version is where you split the stack, and what the split costs you at the boundary.
and one of these is older than my current opinion of it. vibe coding was written in early 2025 and reads now as a question the industry answered differently than i expected. the update is written and scheduled rather than published — it argues that i don’t vibe code, i build fences.
- i said i quit testing. then i shipped seven things. — a year ago i declared testing a waste of time for solo devs. then a silent bug ate 79% of my data. an honest retraction.
- i didn't leave go for rust. i split my stack. — the 'which language' question is the wrong question. go for services, rust for tools — and here's the exact line.
- vibe coding: the tiktok of programming — the dangerous dopamine trap of large language models in coding
related
the testing argument’s evidence lives under testing ; the stack split is worked through in go and rust ; the ai posts sit under ai .