Slightly late this week, because I have been on a series of trains and buses and trolleybuses and trams this week to travel overland from London to Tallinn, while visiting various Brutalist and Modernist landmarks along the way. Highly recommend this as a past time, though I do not recommend Flixbus or indeed Deutsche Bahn.1
Anyway, links:
There is quite a bit of discussion on if professors really need to be researchers - or if it is sufficient for them to be teachers. A new paper from Barbara Biasi finds that people who attend more research-intensive universities are more likely to get a PhD,2 file patents, and earn more than those who go to less research-intensive institutions.
Why do academics p-hack? Well, a statistically significant result is more likely to get you a job and more likely to get media coverage.
Jacob Trefethan wrote a choose your own adventure blog post on how AI will affect medical advances.
Providing women in Burkina Faso with free contraception did not decrease birth rates.
The annoying spam in your inbox asking you to donate to Democratic causes? Yeah, very little of the money appears to go to candidates.3
People who live >2 hours from their hometown voted for Clinton in 2016 by six points. People who have never lived outside their hometown voted for Trump by twenty-six points.
Sectors that are more exposed to LLMs have seen wages rise rather than fall.4
There’s a fun map of the US’ abandoned railways.
I recommend Deutsche Bahn only if you want to spend an unexpectedly long time wandering around the Mannheim rail station.
OK, this isn’t surprising; academic researchers train more academic researchers. In the words of Terry Pratchett, “getting an education was a bit like a communicable sexual disease. It made you unsuitable for a lot of jobs and then you had the urge to pass it on.”
At least at this one (large) firm.
Because, on average, AI is a complement rather than substitute to human labor. Wages fell where it is a substitute.