set.seed(67)

lol downwardly mobile 40-50 y/o put it on my tombstone 🤜 🤛

been doing this for a couple months now, am basically the new Dave Ramsey (if you remove the "wants" and "savings" categories)

When “translate all of R into Rust” works I’ll become a true believer

that’s diabolical and awesome

And to put a downer exclamation point on this: what happens when everyone is feeding existing knowledge as nuanced context and there are no Hadley/Jenny actually writing R Packages anymore? It's mid-shit all the way down from here

hell yes. And a recent epiphany for me is you can feed it a shitload of context, let's say the entire R packages book, and you get way better outputs

To be clear: should we bomb the economy and environment in a single AI stroke for fast SO and autocomplete? No.

I hate to admit this but I have found fast SO and autocomplete are very helpful use cases

oh this is awesome, thanks!

Wait what is that baseball one under data edu 👀🧐

rstudio::conf 2019 was peak twitter rstats IMO. Every person I met was like a celebrity and/or old friend to me, and their twitter handle was more familiar than their actual name. I learned so much from them, both technically and about what good stewards of a community look like

💯 and also 🤮

No cap 🔥

Right now eating nothing but bacon before blood work on Monday, gonna get misdiagnosed with every condition under the sun and laugh all the way from the pharmacy to the bank

Yay! Also if you’re looking for other low code options to get the principles down (because the R/python implementations can get gnarly for an undergrad course pretty quick), cstructure.io @erickscott.bsky.social could be helpful to you

Agree, they are useful when used in focused contexts and driven by experts or at least experienced people in those contexts. I’m not sure I’d try to vibe code a whole manuscript as OP did, but I’ve definitely seen awful human generated code in manuscripts too 🤷‍♂️