mehmet umut caglar

Libertarian, Positivist, Transhumanist

[SARS‑CoV‑2 (COVID-19) @COVID19_disease]

🚨 BREAKTHROUGH: Stanford Medicine scientists say they have found how to regrow worn-down joint cartilage

They targeted a specific aging trigger called 15-PGDH a protein that roughly doubles in your knees as you get older.

Here's what they did.🧵

[Valerio Capraro @ValerioCapraro]

BREAKING: OpenAI’s solution to Navier–Stokes does not match its Lean verification.

The most important article to read today is not one of OpenAI’s 700 AI-generated math papers.

It is this other paper, making a deep and worrying point:

A Lean-verified proof does not automatically validate the proof written in natural language, nor does it mean that the formal statement captures the intended theorem.

During translation, an AI can change an assumption, weaken a statement, or replace the argument entirely.

It can hallucinate another theorem.

Lean correctly verifies the result.
But the proved result may no longer be what the paper claims.

This is a general problem. Things get spicy when the authors examine OpenAI’s proposed Navier–Stokes solution.

They identify at least two mismatches between the written intermediate results and their Lean counterparts:

One estimate claims that four additional input derivatives suffice. The Lean version requires five: a weaker result.

A pressure-flux estimate is obtained through a different bound, and proved through a different argument.

It is not clear whether these mismatches invalidate the entire proof.

But they raise an important issue.

OpenAI is flooding us with claimed revolutionary breakthroughs. Yet nobody knows whether the proofs are correct or whether they prove what they claim to be proving.

Epistemia at scale.

*
Paper in the first reply

[Pavel Rabtsevich @p_rabtsevich]

I think I found a planet nobody knew existed.

It's been hiding in NASA telescope data for seven years.

At the end of this thread, you can fly to it yourself.

[mehmet umut caglar @self_adjoint]

@Jude_E_Stroyer @cremieuxrecueil Out of curiosity where is the data supporting that is wrong?

One counter data is that
https://t.co/lfJpdi1BtQ

[mehmet umut caglar @self_adjoint]

@cremieuxrecueil İf trend keeps going in 50 years we will have a much relaxed and better society.

[tux pacific @__tux]

“there aren’t any cryptography results”
“there aren’t any cryptography results”

[mehmet umut caglar @self_adjoint]

@0xdoug Is it true that there are mistakes on oai s original work.

And as a result everything we put on top of it is problematic?

[mehmet umut caglar @self_adjoint]

@SJ_Swapnil_Jain What is the up to date unconditional value?

[Ben Podgursky @bpodgursky]

Lucky coincidence the labs are only successfully cracking esoteric math problems and not the problems where a solution would give the winner an asymmetric commerical advantage.

[Auri Fomo @AuriFomo]

@ElliotGlazer Not providing the formalization of every claim was a poor choice. There was no rush for the 400 results dump, they could have released the fully formalized second batch at a later date.

[mehmet umut caglar @self_adjoint]

@nselmi Check this. We are at -31 now

https://t.co/TgTAmAxOAx

[pranav @pranavbalakri]

Managed to get an even better bound for integer multiplication. κ > 2^-31, with exact margin 5.91 * 10^-10.

https://t.co/AgKs1jELLO

[mehmet umut caglar @self_adjoint]

@teortaxesTex Check this

We are at -31 now

https://t.co/TgTAmAxgKZ

[pranav @pranavbalakri]

Managed to get an even better bound for integer multiplication. κ > 2^-31, with exact margin 5.91 * 10^-10.

https://t.co/AgKs1jELLO

[pranav @pranavbalakri]

Managed to get an even better bound for integer multiplication. κ > 2^-31, with exact margin 5.91 * 10^-10.

https://t.co/AgKs1jELLO

[mehmet umut caglar @self_adjoint]

@pranavbalakri Holly F

[sahil @narwhalej]

@pranavbalakri Holy guacamole!

[mehmet umut caglar @self_adjoint]

@0xdoug I feel we got stuck at -34
(Come on, it has been 3 hours)

[mehmet umut caglar @self_adjoint]

@jachiam0 https://t.co/LtocW6QRs4

[mehmet umut caglar @self_adjoint]

Waiting for the proof that say real universal speed limit is c+ 2^-182 m/s

[ThePrimeagen @ThePrimeagen]

Its even worse than I thought.

OpenAI one shotted math

[Derya Unutmaz, MD @DeryaTR_]

I couldn’t agree more with Demis: “The most important thing we can apply AI to is improving medicine and human health.”

That’s the mother of all problems, and AI or rather superintelligence (SI) is the perfect and only way we’ll solve it!

[mehmet umut caglar @self_adjoint]

@ilyasut That seems to be the most important tweet of the ai age.

Really capture and represent the moment.

History books should mention it