Olivier Leroy

nah myself do not jump in the body of the function write some signature and dummy examples first...

A bit late on my #podcasts toot: Another good episode of the test set: https://posit.co/thetestset/episode/smuggled-gold-and-untrustworthy-intuition-with-leland-mcinnes

I really enjoyed that the guest bluntly say that it did not have a good workflow with AI: where does it fit, how to use it etc

His take that AI is more data visualization than models was also so on point. IE more dimensions reduction and "opinionated" (bad wording is mine) representations that "surprise you but scale poorly".

Next month I will be at @rOpenSci discussing tests (prob more unit tests) and how I am struggling with them (but try to improve!).

We discussed a bit the topic at Cleveland useR group (thanks @jdblischak !).

One point of debate that came is should you put stopifnot logic in @examples from the function doc has a poor man tests (yup I do that sometimes..). The clean way is def move that in test/

curious to know what ate other takes on that and also how you struggle (or not) on test!

grep -nr -C3 "that_string" --include="glob_that" .

is my "poor man" LLM

every days on windows make me realize than linux is better. tooting that while doing some updates on the damned OS.

I do not know if I will vote for "Lost"
:blobthinkingeyes: