4ldo Malkinson

Tres Comas Boom !

Sarcasm, AI & Tequila

[4ldo Malkinson @4ldo9]

@LiorOnAI @ylecun @ylecun is really a boss, he called it "Le World Model" and not "Un World Model".

It's like going to the Ritz an ordering un croque monsieur, on the menu it will be "Le Croque Monsieur"

[Jesse Ezell @jezell]

If Dario wants to slow down the pace of AI, he should just move the company to Europe.

[4ldo Malkinson @4ldo9]

@arthurmensch That's a quote of Maurice Gamelin I think

[4ldo Malkinson @4ldo9]

@MindsAI_Jack Best start-up idea

[4ldo Malkinson @4ldo9]

@sama stop doing shady sh*t and go cure cancer.

[Valerio Capraro @ValerioCapraro]

BREAKING: OpenAI might have stolen another major proof.

In a detailed Mastodon post, which I report in full in the comments, Andreas Thom presents several pieces of evidence suggesting that OpenAI may have trained Astra on conversations in which he and Gábor Kun were working on Gromov’s soficity conjecture, one of the ten problems OpenAI later announced Astra had solved.

I know Andreas. We met several times early in our careers. He is an exceptional mathematician, a leading expert on sofic and hyperlinear groups, and one of the most respected scholars in the field. He has spent two decades working on this problem.

If his account is correct, this is not a minor dispute over attribution. It would mean that unpublished human work was absorbed into a model and then presented to the world as a breakthrough by the model itself.

And if the allegations raised by Levent Alpöge, Tristan Buckmaster, and now Andreas Thom are all substantiated, we are no longer looking at isolated incidents.

We may be looking at one of the greatest intellectual scandals in the history of science.

AI is not discovering new mathematics.

AI is stealing human discovery.

[Valerio Capraro @ValerioCapraro]

BREAKING: OpenAI might have stolen another major proof.

In a detailed Mastodon post, which I report in full in the comments, Andreas Thom presents several pieces of evidence suggesting that OpenAI may have trained Astra on conversations in which he and Gábor Kun were working on Gromov’s soficity conjecture, one of the ten problems OpenAI later announced Astra had solved.

I know Andreas. We met several times early in our careers. He is an exceptional mathematician, a leading expert on sofic and hyperlinear groups, and one of the most respected scholars in the field. He has spent two decades working on this problem.

If his account is correct, this is not a minor dispute over attribution. It would mean that unpublished human work was absorbed into a model and then presented to the world as a breakthrough by the model itself.

And if the allegations raised by Levent Alpöge, Tristan Buckmaster, and now Andreas Thom are all substantiated, we are no longer looking at isolated incidents.

We may be looking at one of the greatest intellectual scandals in the history of science.

AI is not discovering new mathematics.

AI is stealing human discovery.

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

CancerBench: the frontier model cancer cure benchmark.

AI lab CEOs keep talking about curing cancer, so I made a benchmark.

One metric: how many types of cancer has your model cured?

All models are currently tied at zero.

It’s time to hillclimb!

https://t.co/SeEJz0FFYv

[4ldo Malkinson @4ldo9]

@iScienceLuvr Get on this one ?

@sama
@DarioAmodei
@elonmusk

[4ldo Malkinson @4ldo9]

@LightOrigins_ Really cool

[4ldo Malkinson @4ldo9]

@mikeknoop Hinting that human intelligence is definitely not general, since we are good at ARC but not good at Navier Stokes.

Well, most of us 😂

[4ldo Malkinson @4ldo9]

@NandoDF A real progress would be building an AI which helps humanity achieve this kind of forward thinking mentality

That's a millenium problem :)

[4ldo Malkinson @4ldo9]

@dlouapre Schmidhuber had already done this joke in 1991 😐

[4ldo Malkinson @4ldo9]

@cdngdev Is there a blogpost about the process ? Really cool demo

[4ldo Malkinson @4ldo9]

@ClementDelangue Yeah, mid-tier tech with french politicians looking at it like it's their next meal.

Luckily Samsung can afford to lose money.

[4ldo Malkinson @4ldo9]

@elonmusk Please @elonmusk fill in Cerfa n°346430 and print it in three issues.

[4ldo Malkinson @4ldo9]

@linasbeliunas Fun fact : french VCs and banks never bet 1€ on them.

[Yoram Wijngaarde @yoramdw]

Here you go: European Disapora Index. Companies valued $1B+ incl. Stripe, Snowflake, Datadog, Miro, Grammarly, UiPath, etc cc @sc_cath @lugaricano (small disclaimer: just made this today on the @dealroomco API, will refine coming days)

[4ldo Malkinson @4ldo9]

@haider1 Wow, what tool did it use ?

[4ldo Malkinson @4ldo9]

@paulg Yes we all know what trust is, Mr. Graham.

[artem kirsanov @ArtemKRSV]

New on the geometry of motor control, introducing a dynamical systems view of motor cortex 🦾🧠
https://t.co/Eps6X1XIgv

Seconds before a planned movement, motor cortex can already contain activity specific to that movement.

This is the same circuit that will soon generate the commands sent to your muscles.

So what is that preparatory activity doing? And why doesn't it spill out into motion?

[4ldo Malkinson @4ldo9]

@MLStreetTalk Sounds a bit like AI Burn Out