It makes coming up to speed on an existing codebase a slog because you have to trace through everything back to its source. Oh, and because there are magic methods and properties galore, your normal introspection tools in e.g. RubyMine get frequently stymied.
The example he gave where each dev gets a single feature done seems to skip over the opportunity that Dev A has - after she spends only "a couple days" on that feature to ship a simple solution, she now has the rest of that week to do another feature, and another... in fact, she could probably get 6 features done in the time that Mr. Complexity took to do his one feature (3 weeks in the straw-man example). Then the promo packet looks much better for Dev A - while it says "implemented feature X", it also says "implemented feature A, feature B, feature C...". Doesn't that seem more attractive?
Slightly related: I've noticed that there are lots of "ideas guys" (yes, guys) in our field who love to bloviate, and maybe even accomplish some stuff that looks really good. I have made a career out of just putting my head down and getting shit done. I may not have grand design ideas, and in fact have had to unlearn the "fact" that you need to come up with, and implement, Big Ideas. In my experience, people who "get shit done" may not get fancy awards, but their work is recognized and rewarded.
notable nitpick: calling DC a “federal entity” is mixing up the concepts of nation and state. (It’s also major erasure of the culture of people - mostly Black - who actually live there!)
As a Sconnie native, the main thing that annoyed me about the letter system is that it's easy for the letters to rhyme; for example, near Verona (home of Epic!), there are (Dane) County roads PB and PD. Gotta enunciate carefully. :P This would qualify as a "usability issue" I would imagine. :)
“Craft” vs “delivery” is a false dichotomy. Someone who is “delivering” needs to understand what they are delivering! How can you support vibe-coded cruft that you don’t understand? This is why tools like TypeScript exist; TS by itself adds little in terms of functionality or in “deliverables”, but it makes the code more _understandable_. It’s worth the time to invest in making sure your types are correct because it makes you think about constraints. There are no shortcuts.
There is no strategic understanding by any party of this tool (or the Sudoku one): the human "creator" nor the AI agent. Human knowledge is not advanced at all. The LLM can't tell you _why_ it makes its choices - maybe it could pretend to, but the descriptions themselves would probably be hallucinatory.
It's "just a game" but obviously this also applies to AI decision making in much more consequential settings. We should not strive just to come up with "the right answer" but ask _why_ it is the right answer.
I did a bit of googling to see if I could find a decent civilian introduction to "how to write a SAT solver" but everything I found were papers assume quite a bit of previous knowledge. Is there a good source where one could learn a bit about how to write a basic SAT solver and how to improve on it?