NotPrinceHamlet
@anattendantlord.bsky.social
Ambivalent academic. Occasionally swelling a progress, starting a train or two. Not keen on peaches these days.
The person being quoted here, whosoe'er they be, needs a hefty dose of
October 8, 2026 at 3:16 AM
The person being quoted here, whosoe'er they be, needs a hefty dose of
> Does this mean that as a boffin, I need to rethink what I spend my time doing? Absolutely! Does this mean that boffins should be exiled and replaced with ever-shinier models of Paul's mate's machine? Only if you think that those kinds of questions were the only ones worth asking or rewarding. /end
October 8, 2026 at 3:05 AM
Also on the elephant site: mathstodon.xyz/@atif/117402...
Thoughts? (from those who are more au fait with the incentives and cults of the tech companies)
Thoughts? (from those who are more au fait with the incentives and cults of the tech companies)
Atif (@[email protected])
at this point i don't think it's about marketing. im guessing that the strategy of the labs is a long term infestation of the mathematical ecosystem (universities, journals). they want reliable enterp...
mathstodon.xyz
October 8, 2026 at 12:40 AM
Also on the elephant site: mathstodon.xyz/@atif/117402...
Thoughts? (from those who are more au fait with the incentives and cults of the tech companies)
Thoughts? (from those who are more au fait with the incentives and cults of the tech companies)
Via the elephant site: this issue arxiv.org/abs/2610.08144 is something I hadn't considered. (My guess is that probably this wouldn't affect the "correcteness" of any of the claimed results in the OpenAI MathsDump, but it does remind us that LeanSaysAllGood isn't the end of the story)
Navier-Stokes lost in translation: Why Lean verification of AI autoformalisation does not guarantee correct natural language proofs
Autoformalisation is increasingly used to verify mathematical texts, including those generated by AI, as in OpenAI's announced proof of blow-up of solutions to the Navier-Stokes equations. In this pro...
arxiv.org
October 7, 2026 at 10:33 PM
Via the elephant site: this issue arxiv.org/abs/2610.08144 is something I hadn't considered. (My guess is that probably this wouldn't affect the "correcteness" of any of the claimed results in the OpenAI MathsDump, but it does remind us that LeanSaysAllGood isn't the end of the story)
Non-mathematical followers (i.e. most of you): which theme tune best fits your mental image of Mathematicians React To AI Solving Their Open Problems?
1) Eastenders drum roll
2) Dad's Army opening cope
3) Inspector Morse elegy
4) Sopranos opening brooding
5) Star Trek TNG optimism
6) NCIS rave
1) Eastenders drum roll
2) Dad's Army opening cope
3) Inspector Morse elegy
4) Sopranos opening brooding
5) Star Trek TNG optimism
6) NCIS rave
October 7, 2026 at 8:27 PM
Non-mathematical followers (i.e. most of you): which theme tune best fits your mental image of Mathematicians React To AI Solving Their Open Problems?
1) Eastenders drum roll
2) Dad's Army opening cope
3) Inspector Morse elegy
4) Sopranos opening brooding
5) Star Trek TNG optimism
6) NCIS rave
1) Eastenders drum roll
2) Dad's Army opening cope
3) Inspector Morse elegy
4) Sopranos opening brooding
5) Star Trek TNG optimism
6) NCIS rave
UPDATE: I have checked again, and the post I saw (on the elephant site) has been corrected to say: in fact the machines proved that Thompson's F is *non-amenable*, which is exactly the direction I'd have guessed upo hearing "the machines decided the status of this Y/N question". So I am vindicated!
October 7, 2026 at 5:06 PM
Regarding "how do mathematicians react if their profession/discipline is rendered obsolete" - I first read this story about 25-30 years ago & over the years it keeps resurfacing in my mind, long before the current conflagration. www.cs.ru.nl/~freek/jorda...
(Probably best read fresh w/out context)
(Probably best read fresh w/out context)
Euclid Alone
www.cs.ru.nl
October 7, 2026 at 4:14 PM
Regarding "how do mathematicians react if their profession/discipline is rendered obsolete" - I first read this story about 25-30 years ago & over the years it keeps resurfacing in my mind, long before the current conflagration. www.cs.ru.nl/~freek/jorda...
(Probably best read fresh w/out context)
(Probably best read fresh w/out context)
Think I might print this and read tonight, just on principle, even though it's not really in areas I have any working experience in
This might have been the worst day to post a preprint (if the hope is that someone might be interested in reading it, instead of the OpenAI dump). On the other hand, our work has been in progress for 6 years or so, instead of 3 hours on a slop machine.
Anyway, arxiv.org/abs/2610.08348
Anyway, arxiv.org/abs/2610.08348
Can one hear the shape of a lattice random walk?
We construct distinct high-dimensional mean-zero finite range lattice random walks having pairwise-equal return probabilities for all step counts. The same examples provide pairwise-distinct shapes of...
arxiv.org
October 7, 2026 at 4:03 PM
Think I might print this and read tonight, just on principle, even though it's not really in areas I have any working experience in
Reposted by NotPrinceHamlet
It seems to work with their target audience: governments and investors. However I kind-of don’t care about that side of the equation here. What bothers me a lot is the message this is sending to young would-be future mathematicians. They are paying attention, and questioning their life’s dreams.
October 7, 2026 at 8:36 AM
It seems to work with their target audience: governments and investors. However I kind-of don’t care about that side of the equation here. What bothers me a lot is the message this is sending to young would-be future mathematicians. They are paying attention, and questioning their life’s dreams.
Reposted by NotPrinceHamlet
This was, as you’d expect, properly excellent. God knows how @neilperryman.bsky.social could sit through ten consecutive viewings in short order… 😬
If you’d like to know more about The Day After (and the shit it got from Hesseltine etc), I did this for the old FTS
iainhepburn.me/2026/09/04/t...
If you’d like to know more about The Day After (and the shit it got from Hesseltine etc), I did this for the old FTS
iainhepburn.me/2026/09/04/t...
The video version of our latest episode about Threads is now on YouTube. You can also watch it ad-free via our free to join Patreon. Please like, subscribe and share this video if you enjoy it. youtu.be/pFPP6FjpGOI
World of Telly Episode 23: Threads
YouTube video by World of Telly
youtu.be
October 7, 2026 at 10:11 AM
This was, as you’d expect, properly excellent. God knows how @neilperryman.bsky.social could sit through ten consecutive viewings in short order… 😬
If you’d like to know more about The Day After (and the shit it got from Hesseltine etc), I did this for the old FTS
iainhepburn.me/2026/09/04/t...
If you’d like to know more about The Day After (and the shit it got from Hesseltine etc), I did this for the old FTS
iainhepburn.me/2026/09/04/t...
@vcarchidi.bsky.social In the interests of keeping myself honest: a while back, in the context of the "OpenAI found a non-sofic group" mic drop, I claimed that for a lot of open problems where it is announced that AI found an answer, I could probably guess with good accuracy if answer is yes or no >
October 7, 2026 at 2:40 AM
@vcarchidi.bsky.social In the interests of keeping myself honest: a while back, in the context of the "OpenAI found a non-sofic group" mic drop, I claimed that for a lot of open problems where it is announced that AI found an answer, I could probably guess with good accuracy if answer is yes or no >
Meanwhile, a reminder that for those of us in UKHE, we don't need Evil Trillion Dollar Companies to destroy our livelihood and our professions; we have our own low-tech methods of destruction, like some kind of Evil Fred Dibnah
Research time “reduced to 25 per cent under the plans”, down from 1/3.
www.researchprofessionalnews.com/rr-news-uk-u...
www.researchprofessionalnews.com/rr-news-uk-u...
Nottingham staff ‘will have much less research time’, says union - Research Professional News
Fears for research at Russell Group institution amplify after university’s council approves cost-saving plans
www.researchprofessionalnews.com
October 7, 2026 at 2:12 AM
Meanwhile, a reminder that for those of us in UKHE, we don't need Evil Trillion Dollar Companies to destroy our livelihood and our professions; we have our own low-tech methods of destruction, like some kind of Evil Fred Dibnah
Regarding this dump by OpenAI, and I use the word deliberately, I am reminded of a conversation at a conference, where I said "they can't stop me caring about the maths I do" & was met with "they can stop you caring about your maths problems, but they can't make you care about their maths problems"
October 7, 2026 at 1:11 AM
Regarding this dump by OpenAI, and I use the word deliberately, I am reminded of a conversation at a conference, where I said "they can't stop me caring about the maths I do" & was met with "they can stop you caring about your maths problems, but they can't make you care about their maths problems"
Reposted by NotPrinceHamlet
If, like me, you are someone who is interested in how AI works but doesn’t want to do the serious work that would normally involve, this is a great essay. It’s at (for me) the right level of abstraction: detailed enough that I can see how various types of models differ …
new newsletter, laying out in probably far, far, far too much detail how yudkowsky and the weird ideological commitments of ai safety eventually led to OpenAI and Anthropic massively overtraining their models on cybersecurity and being shocked at them "going rogue"
Tracing the rogue ideology of the frontier labs to their product choices
A drawing of Eliezer Yudkowsky as the Paperclip Optimizer that I commissioned back in 2016 and have an alarming number of chances to use since This is a...
buttondown.com
October 6, 2026 at 5:29 PM
If, like me, you are someone who is interested in how AI works but doesn’t want to do the serious work that would normally involve, this is a great essay. It’s at (for me) the right level of abstraction: detailed enough that I can see how various types of models differ …
Relying as I do on a handful of accounts here for keeping up with CanPol, it has only just struck me that in discussions of "this province's Conservatives have gone nuts / are going even further nuts", I don't see SK mentioned. Is this because there's no point restating axioms, or just sample bias?
October 6, 2026 at 6:26 PM
Relying as I do on a handful of accounts here for keeping up with CanPol, it has only just struck me that in discussions of "this province's Conservatives have gone nuts / are going even further nuts", I don't see SK mentioned. Is this because there's no point restating axioms, or just sample bias?
TIL (thanks to @world-of-telly.bsky.social )
Last of the Summer Wine. Absolutely serious: in 1973, it’s one of the most well-observed, philosophical and funny series on telly. Brilliantly written. Men as bairns. Then after Foggy leaves and Uncle of the Bride it turns into formulaic ‘going downhill in a bath/sidecar/piano-on-wheels’ territory.
October 6, 2026 at 4:16 PM
TIL (thanks to @world-of-telly.bsky.social )
Reposted by NotPrinceHamlet
41% staffing cuts in the Faculty of Arts at UoN. We're entering a wilderness of creativity and critical thinking.
October 5, 2026 at 1:18 PM
41% staffing cuts in the Faculty of Arts at UoN. We're entering a wilderness of creativity and critical thinking.
Might actually be ready for the day ahead, provided that at some point in the next 8 hours I can get 6 hours of sleep.
Can't wait to see how I'll self-sabotage this time. My money is on "too cold to get to sleep, turn heating on, then wake up after 3 hours because it's too hot".
Can't wait to see how I'll self-sabotage this time. My money is on "too cold to get to sleep, turn heating on, then wake up after 3 hours because it's too hot".
October 5, 2026 at 11:37 PM
Might actually be ready for the day ahead, provided that at some point in the next 8 hours I can get 6 hours of sleep.
Can't wait to see how I'll self-sabotage this time. My money is on "too cold to get to sleep, turn heating on, then wake up after 3 hours because it's too hot".
Can't wait to see how I'll self-sabotage this time. My money is on "too cold to get to sleep, turn heating on, then wake up after 3 hours because it's too hot".
Tangential, but can anyone persuade Brian Blessed to do a recording where he bellows COULD THAT EM-DASH BE A COMMA, so that this can be played at high volume every time someone is about to send a piece to the "publish" part of their content management system
Your next coworker might well be an AI agent—and will require a whole new model of workplace interactions.
AI Agents Are About to Flood the Workforce. No One’s Ready for It
Your next coworker might well be an AI agent—and will require a whole new model of workplace interactions.
buff.ly
October 5, 2026 at 5:30 PM
Tangential, but can anyone persuade Brian Blessed to do a recording where he bellows COULD THAT EM-DASH BE A COMMA, so that this can be played at high volume every time someone is about to send a piece to the "publish" part of their content management system
In a cafe whose opening hours are 8am-6pm where the company-imposed playlist has just got on to Dolly Parton's "9 to 5". This feels a bit like insult added to injury for the staff
October 5, 2026 at 4:45 PM
In a cafe whose opening hours are 8am-6pm where the company-imposed playlist has just got on to Dolly Parton's "9 to 5". This feels a bit like insult added to injury for the staff
And wine's done more than wisdom would / To make my unpaid workday good
October 5, 2026 at 4:26 PM
And wine's done more than wisdom would / To make my unpaid workday good
Reposted by NotPrinceHamlet
The most incredible thing about this is that they've clearly used AI prompts which involve some mention of politically satirical art, and the plagiarism machine has reached instantly for Ward Sutton's Stan Kelly cartoons
October 4, 2026 at 11:40 AM
The most incredible thing about this is that they've clearly used AI prompts which involve some mention of politically satirical art, and the plagiarism machine has reached instantly for Ward Sutton's Stan Kelly cartoons
Reposted by NotPrinceHamlet
I love to spy on what people read on the subway. Last week, I had a first: I saw an older gentleman reading NO NAME by Wilkie Collins. About 1/3 through and deeply absorbed in it!
What's the most interesting or unusual book you've seen someone reading in public lately?
What's the most interesting or unusual book you've seen someone reading in public lately?
October 5, 2026 at 1:29 AM
I love to spy on what people read on the subway. Last week, I had a first: I saw an older gentleman reading NO NAME by Wilkie Collins. About 1/3 through and deeply absorbed in it!
What's the most interesting or unusual book you've seen someone reading in public lately?
What's the most interesting or unusual book you've seen someone reading in public lately?
In an effort to avoid binge-watching febrile YT after midnight, have forced myself to watch some downloaded TV with long episodes. Really wishing that the rotation hadn't brought me to the early-season episode of A Touch Of Frost constructed around child death + (prejudices about) Down's syndrome
October 5, 2026 at 1:00 AM
In an effort to avoid binge-watching febrile YT after midnight, have forced myself to watch some downloaded TV with long episodes. Really wishing that the rotation hadn't brought me to the early-season episode of A Touch Of Frost constructed around child death + (prejudices about) Down's syndrome