Today's reading

[anna_reads_science]

The Reverse-Centaur’s Guide to Criticizing AI - Cory Doctorow, tumblr post, December 05, 2025.

This post is the text of a lecture about criticising AI, and is a ~40 minute summary of a book coming out some time this year. I'm personally skimming for commentary that relates to children and genAI, because that is my current research focus, but I'm also reading for getting a feel for the kinds of hidden things that people say about AI so that I can recognise them when I'm doing interviews and need to follow up. Basically: upskilling.

The definition of a centaur (person supported by automation) and a reverse-centaur (automation requiring a human moving part) was useful to me as contextualising the rest of the talk. Particularly the idea that many AI tools are being developed specifically to be reverse-centaurs.

Depending on your viewpoint, this is either a dry and cynical take on USian style capitalism, or a wide eyed and aware realistic understanding. I lean to the latter. I wish I didn't, but the realities of the USA stock market are terrifyingly dystopian.

We need to protect artists from AI predation, not just create a new way for artists to be mad about their impoverishment.

It’s the bubble that sucks, not these applications. The bubble doesn’t want cheap useful things. It wants expensive, “disruptive” things: Big foundation models that lose billions of dollars every year.

Overall: there are some great high level summaries of So! Much! about AI and the issues with the way that it is being criticised and the ways in which this is being leveraged for over reach.