iiinova — where your digital self keeps evolving

A hallmark of these situations is a lack of transparency, which OAI is maintaining in full. Show the prompts, your lean environment, the underlying lean formalizations. Show your damn work like are you 10 years old?

000

But claimed proofs of big problems that are unreadable and only potentially verifiable by a small number of mathematicians…it’s the worst case scenario. A travesty of text.

100

It’s not always all bad: Gromov’s writings are notoriously unreadable and often incorrect, but the novelty of the ideas makes them extremely generative. However, we wouldn’t say he has “proved” things in these papers, per se. Nor does he claim to have done so…usually.

100

I want to be very clear: an undigested proof is not a proof. We have rare precedents of such “results” from oddball Mathematicians who write a bunch of stuff and never explain it to anyone else. These instances are generally regarded as mathematical disasters.

Taran Fiddler @tfiddlerart.bsky.social · 1h

Reposted by War Devil

Weretober2026 Day 7 - Prehistoric Shark 🦈

201192299

War Devil @wardevil.bsky.social · 41m

That sucks but also great work!

Juspreet Singh Sandhu @freegaussians.bsky.social · 1h

Reposted by War Devil

It is a bit interesting to read this. Fortunately, @oldheneel.bsky.social , @jnshi.bsky.social and I were able to sample SK all the way to beta < 1 a few weeks before oAI released 300 pages analyzing Glauber at beta <= 1. The proof is unintelligible and a chunk uses free probability w/o citation

241

War Devil @wardevil.bsky.social · 49m

Ah okay so the synthetic text didn’t say what OAI claimed again. Again. A g a i n

000

Reposted by War Devil

Antonio E. Porreca 🐳 @aeporreca.org · 13h

Finally, animé par le sentiment de l’urgence, I wrote my own AI statement. Yes, it’s radical. Yes, it will probably make me even less popular among my colleagues, if anyone reads it at all. But I really believe we owe this to our students and our “junior” colleagues.

aeporreca.org

Why we should reject AI in academia

I just woke up and discovered that OpenAI has finally released those mathematical “results” they’ve been bragging about lately. We’re talking about 722 “papers” written by their private LLM or somethi...

34015

Reposted by War Devil

Marilou @mariloupage.bsky.social · 6h

“These students are not learning. They will never be the human expert in the loop. And the same thing will happen to our junior colleagues, if you force them to use LLMs to do their job. In fact, one day there will be no expert at all. “

0235

Reposted by War Devil

天使くん推す @angeldeviloshi.bsky.social · 2h

Here it is. Manga is by Toyama Hanachika, one of Fujimoto's assistants and follows a different setting from Shark Kick yuri from the Look Back manga and anime. The challenge of squeezing horizontal text into the layout and still attempt localisation + scuffed cleanup o<-<

1138

War Devil @wardevil.bsky.social · 2h

It’s always “wait trust me bro. Just keep waiting. No no you gotta pay 400 dollars for the good stuff.” Okay, dealer, tell me when ur plug comes through next.

020

Reposted by War Devil

Kat Abughazaleh @katmabu.bsky.social · 8h

FUCK YES. This is the energy, morality, and courage that I expect from leaders. Unbelievably brave.

414067710

War Devil @wardevil.bsky.social · 2h

In mathematics as in partying, make sure you do your important thinking before the lean

010

Reposted by War Devil

Blake C. Stacey @blakestacey.myatproto.social · 14h

No, we don't. github.com/James-Hanson...

1142

Reposted by War Devil

Dave Dustin @dustin.nz · 13h

"YOU’RE LISTENING TO SISYPHUS FM, THE HOME OF NON-STOP ROCK"

6647401338

Reposted by War Devil

julia 🌋 @hypv.bsky.social · 4h

i like the bit in this article where they admit a big slice of the phenomenon is people not putting up with homophobic or otherwise bigoted parents

Personal identity comes first
Young people live in an identity-driven culture, defining themselves by their gender, race, sexuality, politics and mental health, Coleman says. This shift over the past several decades has been driven by both social media and a larger cultural movement that emphasizes personal expression and psychological well-being.
Parents have long challenged their children’s identities, of course. When they do it now, however, they are often no longer viewed as people with a shared history who disagree, experts say. They are seen as coming from an opposing tribe.
“There’s this idea that if you don’t endorse or embrace my identity, you are no longer an ally,” Coleman says. “And from that perspective, cutting you off becomes a virtuous act of self-care and protecting my mental health.”

0153

Reposted by War Devil

Jay Allen @gaiaslastlaugh.bsky.social · 9h

Really enjoying the yuri series 酒と鬼は二合まで (SAKE TO ONNA WA NIGŌ MADE). Gyaru demon Hinata, who grows horns when she consumes drinks offered by humans, asks lonely college classmate Naori to be her dedicated bartender. Cute characters & love the furigana play in the title

0102

Reposted by War Devil

heidi goodson @goodsiephd.bsky.social · 4h

Love this quote: “I think at the heart of this issue is that humans have two competing natures: a tendency to compete and a capacity to appreciate beauty,” said Kai Shaikh, a grad student in math at the U of Toronto. “To me this seems to be a case of the former attempting to strangle the latter.”

062

Reposted by War Devil

heidi goodson @goodsiephd.bsky.social · 4h

"'There’s likely to be a bunch of results where they take someone’s work and then take it to completion,' Dr. Buckmaster said in an interview on Monday evening."

141

Reposted by War Devil

Adolfo Neto @adolfoneto.elixiremfoco.com · 5h

Even if the list is very long, it’s a PDF. There’s room, even if it’s at the end. I’d be ashamed to be involved in something like that—very ashamed. Maybe that’s the explanation.

142

Reposted by War Devil

Adolfo Neto @adolfoneto.elixiremfoco.com · 5h

On the Brazilian logic mailing list, someone shared: "Machines that produce knowledge... How lucky we are to be alive to see this happening... If the result is correct and wasn’t stolen from anyone, today is a day that will go down in history. (...) Problem (NUMBER) from the list:

152

Reposted by War Devil

Barbara Fantechi @barbarafantechi.bsky.social · 5h

Let's stop listening to the guy who was already extremely successful and yet decided to use AI and make advertisement for them, ignoring any ethical concerns. Same guy who said nothing about anti DEI cuts to research funds, until it was his institute that got defunded.

2287

Reposted by War Devil

Barbara Fantechi @barbarafantechi.bsky.social · 4h

We can also not read this nonsense. I have enough to read what others wrote, to spare time to read what nobody wrote.

510421

How does this dude know?

020

How do you know it did it?

120

Yeah man the worlds top gacha game market and its virtuous gambling bans

050

They did. After the Navier-Stokes drop a mathematician revealed he had developed the proof strategy independently conversing with OAI models, and OpenAI admitted the possibility that this conversation became part of the model’s training data.

100

Olúfẹ́mi O. Táíwò @olufemiotaiwo.bsky.social · 15h

OH THE ULTRA CLASSIC LIKE IF STAR WARS WAS GOOD

020

Reposted by War Devil

only a few episodes in but shout out to everybody who recommended Legend of Galactic Heroes

810512

Antonio E. Porreca 🐳 @aeporreca.org · 23h

I find it quite alarming the way known lying sources just immediately gain trust again after a demonstrated lie. People just move past it and pretend that base credulity is the proper mode regardless of past behavior. Just rewarded for lying.

061

Reposted by War Devil

You know that you* are free to ignore papers written with LLMs, particularly when the authors are funded or directly fed the results by AI companies? It’s certainly better than crying while giving visibility to them. That’s not our community (or not anymore), let’s keep our distance.

151

Reposted by War Devil

Liv Agar @livagar.bsky.social · 14h

America is, fundamentally, at its core, a rape nation. and that makes it far from special

03419

Reposted by War Devil

Alexander Kasprzyk @amkasprzyk.bsky.social · 11h

…assistance of professional mathematicians in the areas, and chose to attempt to answer not just one or two, but as many as possible. This is about power, and nothing to do with mathematics. They’re forcefully asserting their dominance over us. They’re saying, “We are in control; you are powerless.”

28316

Reposted by War Devil

Alexander Kasprzyk @amkasprzyk.bsky.social · 11h

OpenAI didn’t simply stumble upon all these results down the side of the sofa whilst cleaning. They chose to systematically invest huge sums of money to produce them. And to keep on going. And going. And going. They systematically took our motivational questions, perhaps with the… #MathSky #LLM

39930

Reposted by War Devil

Liv Agar @livagar.bsky.social · 19h

I think a lot of people who think this way probably feel guilty about an encounter where they were bad with consent. And I don't say this to imply we should assume everyone talking like this is a rapist, but that we should probably talk about how widspread it is to be bad with consent

2692344

Mmm. AI restating someone else’s result. No shot.

130

All ships must go. We’re burning the whole harbor down.

010

It kind of rules I can’t lie

Okay but they aren’t anyways so whatever

000

So my claim to Open AI: everything you say is lies. All these proofs are junk. If not, prove it. I will take my doubt of every claim you make as far as I can, because that is mathematics.

010

Mathematicians do not worship text. We use text to describe the proof of our claims. But some text you made with mathematical terms and ideas is not constitutive of a proof without your argument. If you don’t force me to trust you, I don’t. End of story.

I the mathematician do not trust your claim. That is my basic position as a mathematician, and it is the proper collegial position for any mathematician to take of their colleagues claims. “Yeah, you’re probably right” is not a proper attitude.

You used a model to generate 277 lean formalizations marked as proofs for certain conjectures. Are they? Can you tell me that they are? Has the proof convinced you? Does it even prove what you say? Where is your argument for this process?