My guess is it's DJT's attempt to encourage feelings of contempt for journalists asking reasonable questions.

In his mind it's not worth answering seriously, and he wants to normalise ignoring any and all scrutiny.

Mamdani seemed willing to take the question, so not like he needed saving 🤷

Turns out, some practices that help humans also help LLMs.

Will it finally be AI that gets non-technical leadership to understand the importance of quality in software?

Will it be LLMs that finally motivate us to improve our own output?

My $0.02: the JavaScript language is confusing around the most basic of things, and has so many odd footguns and "Wat" moments, that everybody reaches for dependencies that they hope and pray has dealt with all that bullshit for them.

Of course, the tradeoff is still more code for a better failure message. Usually it's not worth it, but sometimes it's helpful.

If every test case did it, or the original coverage was missing, it's more of a smell.

I reach for that very occasionally, only if the early failure would make a test failure easier to diagnose.

Tends to be when initial state is the result of not-very-stable domain rules and I want to check they still hold, rather than resort to mocking.

Generally to be avoided though.

Did you know the real purists also advocate not eating or breathing. Otherwise they'd have put those in the red-green-refactor loop.

I remember doing that! It was one of the better interview tasks I've had. Realistic, and with a breadth of problems, functional/performance/security.

Seemingly, like all other new things, the FA devs rarely stick around to also be the FO devs. They've moved on to the next thing by then.

In many ways, it's the smart move and I sometimes wish I had the brass neck and lack of integrity to do the same.