1. The main current impact of LLMs on math is the privatisation of previously public research, whether that happens by virtue of mathematicians leaving universities and founding startups, or most universities having comparatively little access to computational power.
2. The universities that do actually have access to such computational resources are, of course, disproportionally located in the richest areas. A future of AI-assisted mathematics where universities try to provide mathematicians with tokens just exacerbates already existing inequalities.
3. I do not see the endless generation of theorems as the goal of mathematics, nor do I believe in a single correct foundational system or a hierarchy of importance of mathematical fields. I am highly suspicious of claims of "grand unifying theories" and "universal correctness". In that way, I think mathematicians are actually in a great position to reject LLM-based research solely based on the reason that it does not help the community. However, I do not yet know how one does this without losing all their funding.
4. Formal mathematics has as much (or even more) to do with organisation, generalisation, simplification, and dissemination as with verification. Verification also works best when the process of working in a proof assistant helps you find your mistake yourself.
5. On environmental aspects: the usual pro-LLM counterargument is that other industries are more wasteful, so, to be clear, I also don't think that {most domestic flights, most personal vehicles, golf courses, space travel} should exist either.
6. I do not have major end-of-the-world anxieties. I do, however, believe that as with many other instances of not so great new technology being widely adopted, this is just further deterioration of workers' rights.
7. I would be less extreme on my points if these companies were not ran by the worst people on the planet. I, however, also believe that having such an industry not be ran by the worst people on the planet is impossible, especially since the two best use cases for AI right now are weapons design and surveillance.

• [08.08.2026] The revival of "traditional" music in moodern Greek culture is, beyond its obvious gentrifying effects, yet another assertion of the dominance of Cretan culture ("authentically mediterranean") over that of the mainland. Crete and the southern Aegean have always played the part of the pure Greek ideal, untainted by Balkan and Eastern "trash", in homogenising the moden Greek nation (see also: Crete in the "Macedonian struggle", Metaxism).
• [08.08.2026] Listening to the new-ish album by (formerly Tim Smith's) Spratleys. Goes to show that you don't have to be one of the weird special people to write fancy chord progressions, and it's actually the rest of the songwriters who are wrong.
• [07.08.2026] You should dig into Hindemith's harmony if you want to write wacky pop music. His piano sonatas would get more love if the piano community wasn't taken over by Big Romanticism.
• [21.07.2026] Listening to Gang of Four's "Entertainment". Need more artists/bands that are more detailed in their politics than the traditional "struggle song". See for example the influence of situationism in that album, or the Maoist tendencies in Henry Cow's (band, not person) albums (Weird that the two best things the West produced in the 70s were made by Maoists: Henry Cow's "In praise of Learning" and William Lawevere's "Quantifiers and Sheaves").

Black Country, New road / black midi / Blur / Brian Eno / Cardiacs / Death / Gang of Four / Gentle Giant / Gorguts / Henry Cow / Kate Bush / King Crimson / King Gizzard and the Lizard Wizard / KNOWER / Pinback / Robert Wyatt / Sleepytime Gorilla Museum / Slint / Talking Heads / The Pop Group / Thinking Plague / This Heat / XTC

I list my favourite "classical" composers separately, because they were all probably horrible people.

Louis Andriessen / Alban Berg / Pierre Boulez / Luciano Berio / Alberto Ginastera / Sofia Gubaidulina / Paul Hindemith / Olivier Messiaen / György Ligeti / Edgard Varèse

• Four stacked perfect fourths, resolving to the major chord one half-step below the "root".
• A minor chord resolving to a major chord a minor third above it (e.g. B♭m->D). Most famously used in the dreadfully loud opening bars of Ralph Vaughan Williams' "Sea Symphony".
• Minor ninths do not necessarily hurt people and can actually sound quite good.
• Stacked perfect fifths give you a vast choice of resolutions.
• A minor chord moving to a major chord a half step below (e.g. E♭m->D).