Alexander Kasprzyk

Alexander Kasprzyk

@amkasprzyk.bsky.social

https://kasprzyk.work Algebraic Geometry • Combinatorics • Mirror Symmetry • Machine Learning Mathematician at the University of Warwick. Editor-in-Chief for “Experimental Mathematics”. Member of EPSRC’s Math Strategic Advisory Team.

550 followers369 following1.4K posts

Alexander Kasprzyk reposted

Liesbeth Corens
@onslies.bsky.social

THE HOOVERS HAVE BEEN RETURNED TO THE SHOP. I also entered hopeless ticket to get my office hoovered noting that there seems to be paper-bitten-through-by-mice all over a corner. Still no hoovering. Six mice traps in my teeny tiny office though. Like I'm in some slapstick cartoon.

Liesbeth Corens@onslies.bsky.social

Bad news! The Faculty has rejected the budget request for the hoovers! When I write my campus novel people are going to find it too unbelievable to buy.

Alexander Kasprzyk
@amkasprzyk.bsky.social

What a nonsense claim. If you’re interested in maths, I say go for it. It’s as beautiful and fascinating and rewarding and challenging as it ever was. Fall in love with the discipline. And you will shape the future of mathematics, with your energy and passion and excitement for the subject.

Arula Ratnakar@arula-ratnakar.bsky.social

“If AI continues to improve at present rate, then in a few years, there’ll be nothing human mathematicians can do that AI can’t do better” can professors please stop spreading doom. It affects undergrads & isn’t the whole point to keep future generations inspired & enthusiastic so fields dont die??

Alexander Kasprzyk
@amkasprzyk.bsky.social

A nice article by Melissa Lee (Monash) on the recent OpenAI math dump, intended for the general public. It’s a good summary: balanced, and making a clear point about questions of quality in a way that’s easy to appreciate. #MathSky

The Conversation AUNZ@aunz.theconversation.com

OpenAI released 722 new AI-generated mathematical papers at once – three have been retracted, and mathematicians are coming to grips with the rest.

theconversation.comIs this the ‘mathocalypse’? Why OpenAI’s latest results dump has left mathematicians in shock
Alexander Kasprzyk
@amkasprzyk.bsky.social

I think good public communication has to focus on making just one point. It doesn’t have to (and probably shouldn’t) be comprehensive. There are limits on word count imposed by the publisher and the attention span of a busy reader. Ideally a reader can come away able to summarise what the point was.

Alexander Kasprzyk
@amkasprzyk.bsky.social

On a contentious topic such as AI, the point you want to make should be accessible to readers no matter their opinion. The style you adopt must keep as many readers as possible receptive to the argument you’re trying to make. I think Melissa has done that here, and done it very successfully.

Alexander Kasprzyk
@amkasprzyk.bsky.social

I was lucky enough, several years ago, to have a series of one-to-one private conversations with the (then) VC at my (then) university. She was extremely generous with her time; our talks totalled many hours. Respectful, attentive, honest, and disarmingly open. I learned an enormous amount.

profserious@profserious.bsky.social

Universities became more managed, for good reason, but along the way something of shared endeavour was lost. We need academic governance that can help shape our institutions, with all the responsibilities that entails. @profserious.bsky.social proposes profserious.substack.com/p/reinventin...

profserious.substack.comReinventing Academic Governance... collective engagement and shared responsibility
Alexander Kasprzyk
@amkasprzyk.bsky.social

I also understood for the first time exactly why universities were becoming structured as they are, and the external factors that were driving this. Because she was prepared to talk freely about these issues, I developed a lot of respect for the challenges and tensions universities and VCs navigate.

Alexander Kasprzyk
@amkasprzyk.bsky.social

University management was a topic I’d never thought seriously about before, and I found it surprisingly fascinating. I really valued the opportunity to get to see the more human side: the unrelenting hard work, exhaustion, annoyances, successes, and the battles that were and weren’t worth fighting.

Alexander Kasprzyk
@amkasprzyk.bsky.social

A nice article by Melissa Lee (Monash) on the recent OpenAI math dump, intended for the general public. It’s a good summary: balanced, and making a clear point about questions of quality in a way that’s easy to appreciate. #MathSky

The Conversation AUNZ@aunz.theconversation.com

OpenAI released 722 new AI-generated mathematical papers at once – three have been retracted, and mathematicians are coming to grips with the rest.

theconversation.comIs this the ‘mathocalypse’? Why OpenAI’s latest results dump has left mathematicians in shock
Alexander Kasprzyk
@amkasprzyk.bsky.social

Everything is gamified. Everything can be reduced to a metric. If it can’t, it’s not important. Tomorrow we’ll read that #AI outperforms humans in every meaningful measure when it comes to appreciating art. Poetry and music are wasted on humans. We’re just so crass compared to the pure #LLM.

This post is unavailable.
Alexander Kasprzyk
@amkasprzyk.bsky.social

“When I ask my customers if they would read a novel written by AI they either vehemently reject the idea or shrug. I have been told ‘I don’t see the difference’ more than once. We sell a poetry collection openly written ‘in dialogue’ with ChatGPT.” Nightmare. www.lrb.co.uk/blog/2026/oc...

lrb.co.ukAlfred Nunney | Write a Novel in Thirty DaysMy boss at the bookshop has said he would sell AI novels if there were clear demand and clear labelling. We almost...
Alexander Kasprzyk
@amkasprzyk.bsky.social

math removed from math e-mail title/abstract subscription list: Alexander M Kasprzyk 1201 AG,CO ****@me.com Your cancel request has been received, and your email address has been removed from the Title/Abstract distribution list.

Alexander Kasprzyk
@amkasprzyk.bsky.social

End of an era for me. I've been subscribed to the arXiv daily emails since 2003. It's been a daily part of my life. But I just can't cope with how many submissions there are now. I've taken the very difficult decision to unsubscribe. #MathSky

Alexander Kasprzyk
@amkasprzyk.bsky.social

I felt the daily email kept me very well informed with what was going on in the subjects I care about. I'd see new trends emerge, new researchers get established, old topics spring back to life with a surprising insight... I loved it. Ah well, it was fun whilst it lasted.

Alexander Kasprzyk
@amkasprzyk.bsky.social

We’ve entered the age of Math-as-a-Service (MaaS). But we’re still confused about who is providing the service. Are we destined to simply be the middlemen? “Two roads diverged in a yellow wood, And sorry I could not travel both…” #MathSky

Alexander Kasprzyk
@amkasprzyk.bsky.social

“I don’t write questions/conjectures in my papers now. When giving a talk I don’t go beyond proven/published facts. It’s an awful feeling…” This is on us. #MathSky

Alexander Kasprzyk
@amkasprzyk.bsky.social

A good friend, a very serious algebraic geometer, is having his work from the past decade swooped upon by a person with AI; said person is pushing out AI-generated proofs of the parts my friend exposed to the world. Fortunately not everything was revealed, but this is utterly gross. #MathSky #LLM

Alexander Kasprzyk
@amkasprzyk.bsky.social

Is this how it’s going to be now? Keep everything private? Discuss nothing, don’t release stages in a larger programme of research. Because someone out there has AI, no original ideas of their own, and no sense of appropriate behaviour.

Alexander Kasprzyk
@amkasprzyk.bsky.social

"We are writing to inform you that the following journals will temporarily close submissions starting October 19: Annales de l'Institut Poincaré B, Bernouilli, Electronic Communications in Probability, Electronic Journal of Probability, ESAIM: P&S." #MathSky #AI #LLM

Alexander Kasprzyk
@amkasprzyk.bsky.social

"The journal Stochastic Processes and Application will be temporarily deferring the handling of new submissions received during the same period." "The sharp increase in submissions has put unprecedented pressure on our editorial boards and on the entire editorial system."

Alexander Kasprzyk reposted

Lord Speaker
@lordspeaker.parliament.uk

"Ukraine is not anyone's to give. This war is about Ukrainians wanting to decide their own future." Dr Fiona Hill, former advisor to the White House on Russia, delivering the latest Lord Speaker's lecture to a packed room @houseoflords.parliament.uk My thanks to Lord Ahmad for chairing.

Alexander Kasprzyk
@amkasprzyk.bsky.social

And elsewhere too. Sadly the cleaning staff often get the worst deal of all. Their jobs have been outsourced, numbers are cut, and for those who remain the pay stays flat whilst the hours rise. It seems a sorry and shortsighted way to save a few pennies.

Liesbeth Corens@onslies.bsky.social

When you stop cleaning a building that's used by hundreds of students & staff, only just commit to cleaning the toilets twice a week and wipe the seminar rooms once a week (resolute NO on offices, and no provision of a hover for staff to do that): how long until it crumbles? Experiment at QMUL...

Alexander Kasprzyk reposted

Matthias Köppe
@mkoeppe.bsky.social

#otd 2 years ago, I created the @passagemath.org fork of #SageMath, the comprehensive #OpenSource #Mathematics system in #Python. Almost 3000 Issues/PRs and 244 Releases later, we are at passagemath-10.8.13, fully modularized as envisioned in 2020, with support for Linux/macOS/Windows/WebAssembly. 🎉

Alexander Kasprzyk
@amkasprzyk.bsky.social

Today's #AlgebraicGeometry lecture will discuss the Jacobian conjecture. A very different situation to when I taught this material just 10 months ago, in January. #LLM #MathSky

Alexander Kasprzyk
@amkasprzyk.bsky.social

...(and a traditionally incorrect proof) should go to Ludwig Kraus (1884) rather than Ott-Heinrich Keller (1939). Kraus was a student of Klein, and spent time in Berlin attending lectures by Weierstrass and Kronecker. He died very young, aged just 28, which is probably why his name isn't familiar.

Alexander Kasprzyk
@amkasprzyk.bsky.social

Theirs was a very lonely intellectual struggle. It required an enormous degree of knowledge, insight, artistry, and courage. It's a shame that, just as the final few who participated are leaving us, the focus is on passive prompting, rather than a final opportunity to preserve the stories of heroes.

Alexander Kasprzyk
@amkasprzyk.bsky.social

That an industry flush with money – the top ten AI companies alone have a combined valuation in excess of $2,190,000,000,000 – couldn't managed to fund this is criminal. And it really highlights what, and who, they care about. It's quite shocking.

This post is unavailable.
Alexander Kasprzyk
@amkasprzyk.bsky.social

As I coincidentally just wrote: we seem to be happily devaluing ourselves into oblivion. Another take on this is that we’re de-professionalising ourselves. It ends at the same place.

Hendrik De Bie@debiehendrik.bsky.social

Yesterday in a meeting a colleague told me he was doing this too

Alexander Kasprzyk
@amkasprzyk.bsky.social

I heard an interesting point: who is awarding the degree? If the marking (and hence the judgement of academic quality) is being done by OpenAI, then maybe they should be the ones awarding the certificate and not the university? We seem to be happily devaluing ourselves into oblivion. #UKHE

Margot Finn@eicathomefinn.bsky.social

'What’s missing from much of the discussion is the sacrosanct get out of jail free card of academic judgement. And assuming universities aren’t keen to inadvertently trade away their right to deny an appeal on that basis, it needs to be front and centre.' Finally someone gets the essential point.

wonkhe.comAI marking could cost universities their get out of jail free cardSet aside for a second the fact that when you ask students if academics should be able to use AI in marking, they tend to be vehemently opposed.