Fragments

This "Dear Colleague" letter , sent by the Department of Education to institutions of higher education late last week, is exactly what you would do if you were trying to use the logic of liberalism in the service of a narrow authoritarianism.

www.ed.gov

U.S. Department of Education Directs Schools to End Racial Preferences

The U.S. Department of Education has sent a Dear Colleague Letter to educational institutions receiving federal funds notifying them that they must cease using race preferences and stereotypes.

And as easily as it fell in line with progressive orthodoxy, the academy twists itself in knots to bow to the reaction against it.

A colleague of mine, the inimitable Max Chis, sent me a link to this study the other day. It measures popular generative AI models against baseline cognitive tests.
TLDR: everyone's favorite new pals are all demented.

www.bmj.com

Age against the machine—susceptibility of large language models to cognitive impairment: cross sectional analysis

Objective To evaluate the cognitive abilities of the leading large language models and identify their susceptibility to cognitive impairment, using the Montreal Cognitive Assessment (MoCA) and additional tests. Design Cross sectional analysis. Setting Online interaction with large language models via text based prompts. Participants Publicly available large language models, or “chatbots”: ChatGPT versions 4 and 4o (developed by OpenAI), Claude 3.5 “Sonnet” (developed by Anthropic), and Gemini versions 1 and 1.5 (developed by Alphabet). Assessments The MoCA test (version 8.1) was administered to the leading large language models with instructions identical to those given to human patients. Scoring followed official guidelines and was evaluated by a practising neurologist. Additional assessments included the Navon figure, cookie theft picture, Poppelreuter figure, and Stroop test. Main outcome measures MoCA scores, performance in visuospatial/executive tasks, and Stroop test results. Results ChatGPT 4o achieved the highest score on the MoCA test (26/30), followed by ChatGPT 4 and Claude (25/30), with Gemini 1.0 scoring lowest (16/30). All large language models showed poor performance in visuospatial/executive tasks. Gemini models failed at the delayed recall task. Only ChatGPT 4o succeeded in the incongruent stage of the Stroop test. Conclusions With the exception of ChatGPT 4o, almost all large language models subjected to the MoCA test showed signs of mild cognitive impairment. Moreover, as in humans, age is a key determinant of cognitive decline: “older” chatbots, like older patients, tend to perform worse on the MoCA test. These findings challenge the assumption that artificial intelligence will soon replace human doctors, as the cognitive impairment evident in leading chatbots may affect their reliability in medical diagnostics and undermine patients’ confidence. No additional data available.

After months of reading it little-by-little each evening, I'm finally finishing the second volume of Robert Musil's The Man Without Qualities. In a sentence: it's a book that has a surprising lot to say to our moment. And here, on St. Valentine's Day of all days, a breathless comparison of love with anger, in its closing pages:

Now, everyone knows what a great relief it is when one is upset to work off one's anger on someone, even if it has nothing to do with him; but it is less well known that this also applies to love. For love, too, must often be worked off in the same way on someone not really involved, for lack of a more suitable outlet.

And then, this song came on the radio, and the sentiments combined into something complicated.

A bit unrelated to his narrator's voicing of a character's thoughts on love, but here's a smart essay on Musil at the New Criterion that I recently read as well.

Elon Musk is not as complicated as many people seem to think he is, and it's also not at all surprising that the aspiring oligarchs have bent the knee.