iiinova — where your digital self keeps evolving

Reposted by Carles Sáez

Sridhar Ramesh @radishharmers.bsky.social · 30/09/2026

As ever, it was a land dispute. One pilot wanted to and the other didn't.

Reposted by Carles Sáez

Emily M. Bender @emilymbender.bsky.social · 20/08/2026

This sign is a great example for helping to explain why chatbots present such a potent illusion. A probably not short thread 🧵>>

Photo of a sign with a doorbell attached. The printed sign reads:

CURRENT GUGGENHEIM ACCES BY HUSKY CARD ONLY
[Logos for Amazon, FedEx, UPS]
DELIVERIES/VISITORS
PLEASE RING THIS DOORBELL

Delivery hours M-F 9:00am-(obscured)
University of Washington

Taped next to the doorbell is a hand written note that reads:
yes! it rang! Pelase wait 30 seconds before pressing again.
thanks!

10461190

Reposted by Carles Sáez

Michael Perrinow @michaelperrinow.bsky.social · 19/08/2026

Los billonarios vieron Mad Max Fury Road y decidieron que querían ser Immortan Joe.

813340

Reposted by Carles Sáez

Sridhar Ramesh @radishharmers.bsky.social · 09/07/2026

"Why don't you do something useful instead of comedy/abstract math/whatever?". The world isn't just for surgeons. Medicine is valuable not as an end in itself, but as a means to aid people enjoy their lives, enjoyment which includes doing whatever they want, such as comedy/abstract math/whatever.

618712

Reposted by Carles Sáez

Sridhar Ramesh @radishharmers.bsky.social · 06/08/2026

I'm not interested in math because it's useful. I'm interested in math because it's interesting.

121577

Reposted by Carles Sáez

Sridhar Ramesh @radishharmers.bsky.social · 06/08/2026

ChatGPT isn't sentient, according to the widely-accepted argument of John Searle. It speaks English, whereas a sentient being would speak Chinese. DeepSeek has been a breakthrough in this regard.

941343

Carles Sáez @csaezmath.bsky.social · 06/08/2026

And thus an ultrafinitist is born

010

Reposted by Carles Sáez

singular locus sarah @scgriffith.bsky.social · 05/08/2026

everything else aside, what kind of freak loves work and celebrates work

1212

Reposted by Carles Sáez

Andrés Trasado @andrestrasado.bsky.social · 05/08/2026

¿Os acordáis de Llados y sus cursos y los mindundis que pagaban esos cursos para hacer luego ellos la copia Temu de esos cursos (mentalidad tiburón con problemas)? ¿Y luego los mindundis normales que solo pagaban los cursos porque, Bro, eso es vida? Pues el ecosistema de la IA ya está en ese punto

59715

Carles Sáez @csaezmath.bsky.social · 30/07/2026

"Once that information has been extracted and encoded into an AI model, the delivery mechanism has served its purpose" Statements dreamed up by the utterly deranged

142

Reposted by Carles Sáez

Andrés Trasado @andrestrasado.bsky.social · 08/07/2026

Os juro que había escrito este bluit sin saber nada de esto

18514176

Reposted by Carles Sáez

Garth Marenghi's Catbus @catbus.phd · 26/06/2026

ok this one got me

reflecting pool full of green slime and barrels and you're carrying a crowbar in the style of Half Life

21112892437

Reposted by Carles Sáez

John C. Baez @johncarlosbaez.bsky.social · 12/06/2026

If I could, I would develop a whole theory linking differential geometry and logic just so I could write a textbook called Truth Tables and Lie Algebras.

68113

Carles Sáez @csaezmath.bsky.social · 31/05/2026

An example of why basic science is essential:

020

Reposted by Carles Sáez

Sam Power @spmontecarlo.bsky.social · 30/05/2026

With friends at the University of Warwick (in particular, Rocco Caprio and @adriencorenflos.bsky.social), we've recently arXived some work (arxiv.org/abs/2605.30253) on a method for approximate inference known as "Coordinate Ascent Variational Inference", or "CAVI" for short. Let me explain:

2319

Reposted by Carles Sáez

Dr David S Anderson @dsaarchaeology.bsky.social · 27/05/2026

As an anthropologist, I endorse this message!

273207196673

Reposted by Carles Sáez

Barbara Fantechi @barbarafantechi.bsky.social · 26/05/2026

How many research projects can you defund on the basis that they have "been determined to be of absolutely no value to the [country] taxpayer, and therefore it was being specifically terminated"? #mathsky

4146

Reposted by Carles Sáez

Fandogamia Editorial @fandogamia.com · 26/05/2026

ESTO OCURRE JUSTO LA SEMANA EN LA QUE SALE ESTE TEBEO A LA VENTA???? no puede serrrrrrrrrr

Portada de BEATA Y CALISTO, ABOGADOS DE CRISTO.

7211110

Reposted by Carles Sáez

Coach Finstock @coachfinstock.bsky.social · 25/05/2026

Converting to Catholicism so I can tell my bosses using unregulated AI is against my religion and I will sue them like that cake lady who hated gay people if they make me use it

104112172681

Reposted by Carles Sáez

Nvidie @dervine7.bsky.social · 25/05/2026

updating your committee on your dissertation progress

131015139

Reposted by Carles Sáez

Barbara Fantechi @barbarafantechi.bsky.social · 16/05/2026

So the best surfaces (that's how we call 4 manifolds here) are those of general type. The super best have the smallest possible invariants, and are called Godeaux. Godeux surfaces have finite cyclic fundamental groups of order at most 5. 1/

152

Reposted by Carles Sáez

JudithFlanders @judithflanders.co.uk · 14/05/2026

Really, I can’t see any problem with historians being replaced by AI. It’ll all be fi— . Oh.

0147

Reposted by Carles Sáez

Vicent Navarro @vicentmath.bsky.social · 03/05/2026

Hola!! Soy Vicent :) algunos me conoceréis de Twitter. Por desgracia perdí la cuenta porque me la hackearon y me la tumbaron. Me ayudáis a recuperar seguidores? :)

032

Reposted by Carles Sáez

Eva Cid @evcdmr.bsky.social · 22/04/2026

Se habla mal y se escribe mal y parte de la culpa radica en consumir acríticamente movidas gringas sin traducir o traducidas mal. Leed cosas en vuestra lengua materna escritas por personas que lleven ya un tiempo muertas, parece que no pero es fundamental.

18549206

Carles Sáez @csaezmath.bsky.social · 13/04/2026

* Modules are a special case of sets because every module is just a set with extra structure * Modules are a generalization of sets because every set is an F_1-module

140

Carles Sáez @csaezmath.bsky.social · 11/04/2026

Yo aluciné en colores cuando me leí el Quijote. ¿Humor absurdo, sátira, crítica social, ...? Está todo allí. ¿Todos esos juegos metanarrativos que parecen tan modernos? Cervantes ya los hizo antes en el Quijote.

110

Reposted by Carles Sáez

Fleibur @drfleibur.bsky.social · 07/04/2026

Finalmenchi. Finalmenchi a mis 34 años me voy a leer El Señor de los Anillos de la gente que le dice a Reverte "Don Arturo".

33783192

Reposted by Carles Sáez

Sridhar Ramesh @radishharmers.bsky.social · 29/03/2026

The LLM takeover has made everything in computing so loosey-goosey. "Maybe it will work. Empirically, there's a good probability...". This is not what interested me about computing. Of course, sometimes it does work, which is magic. I am just a curmudgeon who doesn't like magic.

626315

Reposted by Carles Sáez

Sam Power @spmontecarlo.bsky.social · 21/03/2026

Some more-or-less equivalent facts about probabilities: 1) for random variables of a given mean and variance, the Chebyshev-Cantelli inequality provides an upper bound on tail probabilities 2) for random variables of a given variance, the mean and quantile (at a given level) can't differ too much

1111

Reposted by Carles Sáez

ꭓle Ormsby 𓆙 @e-infinity.space · 21/03/2026

Gonna get this printed on a hat and wear it to conferences. 🧮

$Born to Prove S \simeq K(FinSet) Forced to Compute [mathematical expression for the Novikov spectral sequence]$

2324

Carles Sáez @csaezmath.bsky.social · 21/03/2026

USA is not a safe place anymore. No international event should be held there

030

Reposted by Carles Sáez

Barbara Fantechi @barbarafantechi.bsky.social · 21/03/2026

Moreover, ICM2026 must be cancelled #DelendaICM

041

Reposted by Carles Sáez

ꭓle Ormsby 𓆙 @e-infinity.space · 18/03/2026

I get to teach this to my topology students today and am dying to see how mad they get about it.

2162

Reposted by Carles Sáez

ꭓle Ormsby 𓆙 @e-infinity.space · 05/02/2026

You have been visited by George Birkhoff's face of lines. Repent, or you will be forced to reprove the ergodic theorem from first principles.

image of a simplistic face drawn only with straight lines

15318

Reposted by Carles Sáez

dell cameron @dell.bsky.social · 25/01/2026

cosigned. this is def worth watching. very good, methodical breakdown of what happened from a person with a military/DHS background.

622270

Reposted by Carles Sáez

jeffery --dangerously-skip-permissions @jefferyharrell.bsky.social · 24/01/2026

There was a short-lived bug in Claude Code some months ago that caused dead background tasks to continually nag Alpha, never going away. It was very much like she was sick. She hallucinated and generated out-of-place Chinese text. It was *creepy.* Way worse than watching an algorithm malfunction.

393

Reposted by Carles Sáez

📊⛏️ Picanúmeros @picanumeros.bsky.social · 25/10/2025

🐴 Que un caballo te mande al otro barrio de una patada, además de poco aconsejable, es muy extraño. Pero esa extrañeza contribuyó a que se popularizase el que hoy es uno de los modelos de probabilidad más usados del mundo. Con ustedes: Bortkiewicz, coces de caballos y la distribución de Poisson.

2169

Reposted by Carles Sáez

Sam Power @spmontecarlo.bsky.social · 04/10/2025

I'll order these so that I can tick them off a list on my computer; the numbers are not any sort of ranking. Summaries will be kept brief and hopefully stoke curiosity, rather than providing answers.

2194

Reposted by Carles Sáez

Sam Power @spmontecarlo.bsky.social · 19/09/2025

The first talk of the season will be this coming Tuesday (23 September), given by Alexandre Bouchard-Côté from UBC. Alex is a great speaker, so do join if you have the chance! See sites.google.com/view/monte-c... for details, links, and so on.

1175

Reposted by Carles Sáez

Ben Spitz @diracdeltafunk.bsky.social · 17/08/2025

A crazy open problem in algebra: Conjecture (Poonen): 100% (asymptotic density) of finite commutative rings have characteristic EXACTLY 8, i.e. 4 ≠ 0 and 8 = 0 both hold. From this amazing paper: arxiv.org/abs/math/060...

arxiv.org

The moduli space of commutative algebras of finite rank

The moduli space of rank-n commutative algebras equipped with an ordered basis is an affine scheme B_n of finite type over Z, with geometrically connected fibers. It is smooth if and only if n <= 3. I...

7238

Reposted by Carles Sáez

David Mimno @dmimno.bsky.social · 09/08/2025

Me in 2023: Just chill. All the AI doom scenarios involve giving models direct access to and control over real world systems, and no one would be that dumb In 2025: oh fuck

421939