iiinova — where your digital self keeps evolving

Grant @syncingsinking.bsky.social · 21/09/2026

Reposted by @robbielyman.com

Reposting with alt text and to #MathSky

Conic Sections Rebellion

Article Talk

The Conic Sections Rebellion, also known as the Conic Section Rebellion, refers primarily to an incident which occurred at Yale University in 1830,[1] as a result of changes in the methods of mathematics education.[2] When a policy change dictated that students were required to draw reference diagrams for exams rather than be allowed to refer to diagrams in their textbooks, students staged a rebellion in which they refused to take the exams at all. Because of this, 43 of the 96 students were expelled.

091

I like your second, more positive spin on it. It would truly be a crying shame if it were actually impossible for the vast majority of people, but I think it is merely difficult.

111

(Finally, observe that for elements g and g’ = hg of the normalizer which are equal in N/H, you get the same equivariant map on X in the construction above.)

000

If g does not normalize H, there is no equivariant map sending x to g.x. Since H stabilizes x, under the equivariant map it must stabilize g.x, but this would require that hgH = gH, i.o.w. that g^{-1}hg is in H, which does not hold for all elements h of H if g does not normalize H.

100

If g normalizes H, there is an equivariant map sending x to g.x: observe that the map h -> hg is equivariant (for the left action of G on itself) and is constant on cosets of H, since Hg = gH by assumption.

100

explicitly, suppose G acts transitively on X. Picking a point x establishes a bijection between X and G/H, where H = Stab(x). Since G acts transitively, any equivariant map is determined by the image of x.

100

Ok, following up on 2024’s banger “edge orbits are double cosets”, if it please the court may I humbly submit “equivariant maps are elements of the normalizer of the stabilizer (modulo the stabilizer)”

100

“If we’re looking to better society, it would be better to start dismantling the oligarchy that led to this development, right?” Very well said, 藤井先生

The machine translation is pretty good anyway but: “all this talk of how dangerous our AI is, is AGI propagandizing (aka investment promo) right? No need to take it seriously. A machine for killing people will kill people whether it operates by GPUs or a roll of the dice. (…)

robbielyman.com @robbielyman.com · 15/09/2026

The argument for *why* change is both necessary and inevitable is rather thin on the ground, but I really like the conception of the goal of a math PhD as becoming a world expert on a deep and interesting topic—arguably this is *already* the goal of a successful PhD.

000

robbielyman.com @robbielyman.com · 15/09/2026

Wishing a very happy 100th birthday to noted arborist Jean-Pierre Serre :)

000

robbielyman.com @robbielyman.com · 15/09/2026

Well, I’m glad plenty of people seem to still have good sense :) www.nytimes.com/2026/09/14/u...

She said she was worried that it could fall into the wrong hands, but not that it would go rogue and turn on humans.

“It does seem like we could just unplug it,” she said.

000

Reposted by @robbielyman.com

Gro-Tsen @gro-tsen.bsky.social · 10/09/2026

Oh look: another paternity dispute about an AI-generated proof! 😬 (About the existence of a non-sofic group.) Links: mathstodon.xyz/@andreasthom... mathstodon.xyz/@andreasthom... mathstodon.xyz/@andreasthom...

Screenshot of first linked post by Andreas Thom on the Fediverse (Mathstodon)

Screenshot of second linked post by Andreas Thom on the Fediverse (Mathstodon)

Screenshot of third linked post by Andreas Thom on the Fediverse (Mathstodon)

1113

Reposted by @robbielyman.com

Gro-Tsen @gro-tsen.bsky.social · 10/09/2026

So it's getting clear that some of the math abilities AI models have become ❝superhuman❞ at is “not giving credit where credit is due”, like “failing to cite previous work on the subject that inspired you” and, worse, “stealing colleagues' ideas”. 🧵⤵️

34921

Reposted by @robbielyman.com

Grant @syncingsinking.bsky.social · 10/09/2026

This is just to say I have solved the open questions that were in your paper and which you were probably saving for your phd students Forgive me they were delicious so incentive-laden and followed so readily once your postdoc's drafts mysteriously found their way into my training data

4356114

robbielyman.com @robbielyman.com · 10/09/2026

This is an open letter urging the suspension of a “math hackathon” at Caltech. You are invited to sign if you are a math PhD student or math PhD. The organizers have responded, but I personally consider the main ask: “suspend the event” still reasonable. docs.google.com/forms/d/e/1F...

docs.google.com

Open Letter about the Mathathon

The full text of the open letter is below. By signing this open letter, you consent to your name, title, and institutional affiliation being publicly displayed beneath this open letter. You can access...

000

Sounds about par, maybe I’ll check it out if I feel like self-flagellation!

000

Is any of the discussion good? “There’s lots of mathematicians talking about it” seems to me to be true of, eg, Reddit and not a useful indicator of quality

100

russ @qqrs.us · 03/08/2026

I'll amend this to say I'm glad Terence is talking in this direction, so apologies for my underinformed and late night angst. I do think there's room for improvement, but maybe it's shaped more like the Leiden declaration than Bill Thurston organizing a phone tree.

100

Reposted by @robbielyman.com

Good talk by Terence Tao on problems with AI in math research I'm finding the language helpful for thinking about AI in non-math stuff (writing, coding) teorth.github.io/tao-web/slid...

teorth.github.io

101

Do we just not have anybody willing to fill the thought leadership shoes for the field? It’s genuinely weird to read Grothendieck or Thurston speaking to urgent political issues affecting the math community of the day, right, vs basically crickets (or “this is why i didn’t prove this” cope) atm?

100

Anyway, I don’t know anything more about Navier Stonks today than I did yesterday and that wasn’t much. It’s good to hear Tao belatedly realize publicly that millions of dollars of token spend coming to bear on any sufficiently hot problem is not *great* for young people in the field, I guess.

001

I’ve read annals and inventiones papers, I beg to differ!

robbielyman.com @robbielyman.com · 05/09/2026

Unfortunately “writing well” still not one of them

121

On the other hand, I had the dubious pleasure of slowly feeling more and more certain that the invited speaker at a (programmer) event had written a script for their talk using AI and then not really gut checked with themselves after, so maybe I’m overimagining things??

000

robbielyman.com @robbielyman.com · 05/09/2026

I feel like AI-generated writing has not gotten any more difficult to spot over the past year. If anything maybe it’s even more obvious? Am I crazy? Other AI capabilities seem to have dramatically improved in that same time frame.

robbielyman.com @robbielyman.com · 30/08/2026

I actually do think these markets are pretty saturated! Students and other folks I know just around that age seem knowledgeable or at least opinionated, and people around my age and older seem less knowledgeable but equally polarized if not more so.

130

robbielyman.com @robbielyman.com · 29/07/2026

This was a *fantastic* talk! In case a little blurb helps make the title clearer, I loved the concrete ways of using "sum" and "product" types to push correctness questions in programming to the place they really should be!

000

robbielyman.com @robbielyman.com · 28/07/2026

Is there a reason I don't see people talking sensibly about how most AI math results so far are about constructing relevant examples? (which is to imply, they proof the easier "existential" ∃ statement rather than the presumably harder "universal" ∀ statement)

robbielyman.com @robbielyman.com · 28/07/2026

alas! although, looks like there are occasionally ones on the West Coast too!

011

robbielyman.com @robbielyman.com · 27/07/2026

Come make friends, learn some Zig and work on Software You Can Love this weekend in Downtown Brooklyn! zig.day/usa/new-york...

zig.day

Zig Day New York City, NY #2

Zig Days are single-day collaborative programming events

A L O N Z O @alonzo.zip · 26/07/2026

Reposted by @robbielyman.com

i'm not totally sure why this one wanted to stay a draft so long, but new post about the YA book I just read :) is it a flex to include my translations of a couple paragraphs of Japanese? read.alonzo.zip/welcome-to-t...

read.alonzo.zip

Welcome to the Classroom of Meritocratic Supremacy

Around this time last year, I started watching ようこそ実力至上主義の教室へ, titled Welcome to the Classroom of the Elite or just Classroom of the Elite in English, and sometimes shortened to よう実 in Japanese. The ...

001

robbielyman.com @robbielyman.com · 17/07/2026

best 3 months playing hookie of my life :)

030

robbielyman.com @robbielyman.com · 02/07/2026

I'm really excited about this paper :) With George Domat and Hannah Hoganson, we show that there are exactly 3 "coarse equivalence" classes of homeomorphism groups of countable, compact ordinals in the order topology, giving explicit models for each one! arxiv.org/abs/2607.01196

arxiv.org

Coarse geometry of homeomorphism groups: Classifying countable Stone spaces

Towards developing the tools of geometric group theory for non-locally compact topological groups, we give one of the first complete classifications of a family of such groups up to coarse equivalence...

robbielyman.com @robbielyman.com · 01/07/2026

By the way, I’m soliciting suggestions for interesting Perron numbers (algebraic integers) of manageably small degree. This one is sometimes called Lehmer’s number (I think?) and is a very small Salem number. If you have a vaguely hipster favorite Perron number please let me know :)

011

robbielyman.com @robbielyman.com · 01/07/2026

Here’s an irreducible train track map whose transition matrix has characteristic polynomial x^12 - x^7 - x^6 - x^5 + 1 It represents an outer automorphism of a free group of rank 10, which is a good deal higher than the theoretical minimum of 6. Maybe we cannot go lower, which would surprise me!

A screenshot of a notes app drawing of the train track map, including a computation of its transition digraph, the Local Whitehead graph, and the map on directions. The graph has three vertices and 12 edges. Edges 1,4,6, and 11 connect vertex a to vertex b. Edges 5,7,9 and 12 connect vertex b to vertex c. Edges 8,2,10 and the reverse of 3 connect vertex c to vertex a. The map sends 4 -> 5 -> 2 -> 6 -> 7 -> 8 -> 1 -> bar(4),3, sends 9 -> 10 -> 11 -> 12 -> bar(3) and sends 3 -> 9,2.

032

Reposted by @robbielyman.com

Gro-Tsen @gro-tsen.bsky.social · 23/05/2026

Very insightful short essay by Petra Schwer ( mathstodon.xyz/@PetraSchwer ): ‘The meaning of doing mathematics’ “(Can AI solve all math? What do we actually mean by doing mathematics? How do we communicate mathematics? What is mathematics beyond problem solving?)” arxiv.org/abs/2509.15998

arxiv.org

The meaning of doing mathematics

Can AI solve all math? What do we actually mean by doing mathematics? How do we communicate mathematics? What is mathematics beyond problem solving? This essay is my attempt to answer these question...

22813

Reposted by @robbielyman.com

Dominique Baker @bakerdphd.bsky.social · 07/05/2026

Ummmmm guys... I just went to canvas to do something for class and this is all I can see. IT has said nothing and my students need to use this to submit assignments and I need it to grade and also what information are we being blackmailed with and...Well, things are going great!

Screenshot of webpage which read: SHINYHUNTERS
rooting your systems since '19 ;)

ShinyHunters has breached Instructure again).
Instead o f contacting us to resolve it they ignored us and did some "security patches",

WARNING

If any of the schools in the affected list are interested in preventing the release of their data, please consult with a cyber advisory firm and contact us privately at TOX to negotiate a
settlement. You have tilı the end Of the day by 12 May 2026 before everything is leaked.

Instructure stil has
until EOD 12 May 2026
to contact us.

34740189

robbielyman.com @robbielyman.com · 20/03/2026

it used to be one of the Wash Heights 1 line stops—really tall, rounded arch ceilings on the platform. i quite like Prospect Park + Church Ave on the B/Q tho too

robbielyman.com @robbielyman.com · 19/03/2026

oh, that part is fine by me: it's not my favorite use of my work, but i do publish open access and post to the arXiv because i believe in open science.

robbielyman.com @robbielyman.com · 19/03/2026

this press release demonstrates to my mind a surprising lack of domain expertise in terms of marketing, assuming they’re looking to get buy-in from academics in STEM!