C. Kirabo Jackson

@kirabojackson.bsky.social

Personal account

Two brief thoughts on why I think it's still worth keeping LLM use out of introductory programming for stats/data science: - You need to be a good programmer to get good code out of an LLM - You need to be a good programmer to evaluate the code you get out of an LLM

Alexander Wuttke@kunkakom.bsky.social · 4w ago

Genuinely interested in your views: What is the future of teaching statistical programming language for data science in the age of AI? Will we even still teach programming R or Python?

Cape Verdeans are now honorary Philadelphians because that was basically ROCKY. They weren't "supposed" to be there, much less going the distance with the champions and the GOAT, who had to give it everything they've got just to eek out a win in overtime. Bravo.

Each ICE operation generates roughly $196,000 in monthly losses for nearby businesses and communities. (That's 11x the government's administrative cost of a deportation.) Why? ICE activity decreases foot traffic in those public areas.

BildBildBildBild

Speaking of public spending on children, here's a great set of scholars with what looks like an important new paper They find that Universal Pre-K increases labor force participation and earnings, which then increases tax revenue such that it more than pays for itself (under reasonable assumptions)

C. Kirabo Jackson @kirabojackson.bsky.social · 4mo ago

Delighted to share our revised paper (w/ @juliaturner.bsky.social & @jacobbastian.bsky.social ) showing that Universal Pre-K has broad effects on local labor markets — effects well beyond only mothers. Larger effects for full-day programs. nber.org/papers/w33767 #EconSky #EarlyChildhood #ECE

My sense is that the AI-generated analyses are not good yet. But here’s a more positive spin on your observation: we need a lot more basic, solid policy evals, which may well be of the TWFE variety. We want them to be good, but they don’t have to be top-Econ-journal caliber to be useful. /1

Please please accept or decline requests to review papers immediately upon seeing the request. It's simple. Most people don't have a good reason for being slow (though I realize a few might). It's crazy how many people are slow. They slow down the whole process for no good reason.

Econ needs more feel good stuff. It's nice to get messages like this. Even better would be more "best paper" awards, about which the field is very stingy (relative to other fields).

Bild