iiinova — where your digital self keeps evolving

kolmogorov gnostic @javaid89.bsky.social · 8h

This is great! Do we/AI companies have a plan for cheaper housing, food and that kind of stuff?

010

kolmogorov gnostic @javaid89.bsky.social · 13h

Ok got it. Very interesting. I think I agree.

010

kolmogorov gnostic @javaid89.bsky.social · 15h

Apologies but would you mind elaborating further? Do you mean mathematics as set of abstractions and principles 1) will stop being useful to understand the world? Or 2) advances will stop because the env that cultivated it is being murdered or 3) or AI itself will replace it?

110

kolmogorov gnostic @javaid89.bsky.social · 09/10/2026

I mean also ... they presumably did not give credit where it was due. There were reported work that was used without attribution or acknowledgement. AFAIK that is textbook plagiarism

000

kolmogorov gnostic @javaid89.bsky.social · 09/10/2026

I am also seeing it everywhere. I was thinking maybe "dead internet." But then I saw video by a youtuber PhD math guy working in finance now who can't wait for the profession to die. Because his brilliance wasn't appreciated ... LOT of resentment everywhere. Like dude you are a quant now you won!?

050

kolmogorov gnostic @javaid89.bsky.social · 09/10/2026

Something has gone horribly wrong over at HN

030

kolmogorov gnostic @javaid89.bsky.social · 09/10/2026

I guess if nothing else they show that there is no "silver bullet?" and like it's going to be hard for anything that isn't like some hypercomputer? So "institutions" and "review" matter (or at least how humans have tried to deal with it) and even then it's hard!

020

kolmogorov gnostic @javaid89.bsky.social · 09/10/2026

I think the crux of the main complaint is how OAI is at once saying "math is solved" while forcing the math community to do unpaid data work for them ... forever. It's either really clever, or unintended byproduct "ruthlessly" hacking the current incentive structure at a societal level.

1147

kolmogorov gnostic @javaid89.bsky.social · 08/10/2026

It's ok to not always be rational! For me the revulsion isn't about loss of artisanal math. Just the hostile anti-intellectualism and centralization of this tech. And the goal of "Disempowerment as a Service" championed by misanthropes who feel entitled to labor of others without contribution!

000

kolmogorov gnostic @javaid89.bsky.social · 08/10/2026

One need only go on hacker need to realize that a non negligible group seem to think that mathematicians/academics are "hiding" knowledge from everyone else.

040

kolmogorov gnostic @javaid89.bsky.social · 08/10/2026

They don't care. Why would they? It doesn't seem like they need to. There are zero consequences for anything they do.

051

kolmogorov gnostic @javaid89.bsky.social · 08/10/2026

This whole thread is so nice. We actually more of THESE types of "inside" thoughts. We also need people to actually lead and have courage!

020

Reposted by kolmogorov gnostic

NotPrinceHamlet @anattendantlord.bsky.social · 07/10/2026

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)

arxiv.org

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...

021

kolmogorov gnostic @javaid89.bsky.social · 08/10/2026

What if the "auto formalization" is incorrect? Or has a subtle issues so that the thing proven was something slightly different than the original statement? But it's not AI doesn't cheat right? RIGHT!?

000

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

Can't wait to see someone correct Tao with "actually it's NOT brute force do you know how large the solution space ..."

130

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

by Sarah Wynn-Williams? oh snap ... I read it and that missed it this too lol.

010

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

How advanced can this be if a computer can do it? I kid I kid.

020

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

Wait Great Gatsby!? I think I know this one ... central theme of the damn book. If I'm wrong i'll just let this stay here and not delete it ... and let this hang like a fart.

110

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

That is fair. And a good cybersecurity mindset! I am just curious as to what this technology is and it's limitations. But something about OpenAI and their dealings "sticks in the craw." ... I think that is the idiom.

110

kolmogorov gnostic @javaid89.bsky.social · 07/10/2026

Yeah something about this is ... off. Not in the sense as "I refuse to believe this." But more like idk grating? or anti-social? insulting? They would have gotten some reaction regardless. Not sure what type of response is. Just don't want to give OAI the satisfaction.

1110

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

Wasn't all of statistics basically invented ... Prove superiority of certain stock of British nobility/aristocracy? Like a prominent journal was called the Journal of Eugenics or something for quite some time. Iirc

010

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

Like is there a difference between a dynamical system and it's mathematical representation? This is definitely above my pay grade/understanding. And I may not be making any sense.

000

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

Well I guess am wondering why say Claude is more like an ant than the map itself. A model is a map; you can bolt on all kinds of fancy things to it. And it's still a representation of something else. An ant definitely also forms representations. But I wouldn't say it is a map or a model of anything.

100

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

Not to derail this. Do wiser people say confidently that there is a fundamental difference between Claude and say a Kalman filter or state space model? Or re/constructing a model from noisy data? Or a Hidden Markov Model? Those are all just fancy maps...

100

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

Or basis functions pick out or assumptions we make of the data. Or like even regression !!! Claude is an animated map I can interact with!?

000

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

I am not sure how say the thing underpinning Claude is fundamentally different than a Hidden Markov model, or a Kalman filter or basic state space models? If that is the case then Claude is the map, just with different coordinates/projections depending on the bias of the "in ariants" the algorithm

100

kolmogorov gnostic @javaid89.bsky.social · 06/10/2026

If it does happen it's gonna take a lot of field work and data gathering/scientific work and knowledge. But everyone's gonna talk about it like an LLM just dropped from the sky and auto-magically just figured it out.

050

kolmogorov gnostic @javaid89.bsky.social · 05/10/2026

Do investors really want an uncontrollable God that may destroy their investments? Or at least cause such massive instability so as to render any returns driven to zero/negative!? Or is this another one of those "Rational" optimization race dynamics

010

kolmogorov gnostic @javaid89.bsky.social · 04/10/2026

Like my family in the non-WEIRD world don't ask these questions. What if we will never find an answer, and our obsession with this question leads us to ... a very very bad place?

000

kolmogorov gnostic @javaid89.bsky.social · 04/10/2026

This may be unwelcome but isn't this kind of like charging Physicists with under-performing on finding the one simple equation/theory of everything? How do we even know that this is a valid question or a question with an answer?

100

kolmogorov gnostic @javaid89.bsky.social · 03/10/2026

"The Wigner Distribution", "The Parseval Identity", "The Nyquist Criteria", "The Basis Pursuit"

110

kolmogorov gnostic @javaid89.bsky.social · 28/09/2026

I will say that this is precisely what happened in mathematics. Last September Scott Aaronson used gpt 5.2 to solve for a result in Matrix Analysis that turned out to be a bit arcane but standard. Then a year later we are solving milenium problems. Maybe biology is different.

100

kolmogorov gnostic @javaid89.bsky.social · 28/09/2026

Sadly I don't think a pause is ever going to happen. And I think we are going to have to figure out what to do about that pretty quick. But "All shall be revealed" soon enough.

000

kolmogorov gnostic @javaid89.bsky.social · 25/09/2026

Yes but not tests correlated to stupidity. Only the Principle Component of the covariance matrix of all tests that correlate with intelligence!

010

kolmogorov gnostic @javaid89.bsky.social · 25/09/2026

Ok yes excellent point! Also agreed!!! I'm personally not a caliper enthusiast.

030

kolmogorov gnostic @javaid89.bsky.social · 25/09/2026

Maybe the argument in this case is "AGI is a contradiction; there is no artificial intelligence" or in other words; intelligence is intelligence. I mean I can get behind that. But I don't claim to know the author's meaning.

100

kolmogorov gnostic @javaid89.bsky.social · 25/09/2026

Sorry I am barging in here ... But this is what everyone in CS, Physics, and Math said. I would not underestimate these companies.

000

kolmogorov gnostic @javaid89.bsky.social · 24/09/2026

It seems to also be strong in Canada. So maybe Anglosphere minus Aus and NZ?

000

kolmogorov gnostic @javaid89.bsky.social · 24/09/2026

And why he talks about process and traceability/auditibility in that same interview. Which is not the same as letting an LLM parse through billions of tokens of logs. This is why I don't consider the labs or their embedded auditors serious people 2

000

kolmogorov gnostic @javaid89.bsky.social · 24/09/2026

It's also important to note. That Jensen is a traditional ENGINEER steeped in hardware and systems that have to deal with real world economic and physical constraints. This is why he brushes aside things like RSI (computers have literally designed themselves this EDA). 1

100

kolmogorov gnostic @javaid89.bsky.social · 24/09/2026

I know this thread has outlived it's usefulness. But the above is really not the same argument (because it's not the same as chess). And your comment is a perversion of Turing. The imitation game afaik was not the Supreme Court of Reality. It was a starting point not the end of every conversation!

100

kolmogorov gnostic @javaid89.bsky.social · 23/09/2026

or gleaned from some quantity of data.

100