Kaiser17

Third Rate Wannabe Scientist

[Kaiser17 @Kaiser1710]

useful https://t.co/o1kGl3e7Zx

[Kaiser17 @Kaiser1710]

@abhijit_MLab What do you think is wrong with funding a few student to study abroad.

[Kaiser17 @Kaiser1710]

@chenna1985 what you dont belive in progress of humanoid robots trained via RL over last few years?

[Michael Choi @michaelchchoi]

The update is released on arXiv. Check it out if you are interested in leveraging group action, symmetry and partitioning to speed up the convergence of Markov chains.

Link: https://t.co/LJ3Cfqe9Qw

[Steven Strogatz @stevenstrogatz]

Hugo Duminil-Copin, a Fields Medalist, writes from the heart about what we mathematicians lose when we do math with AI.

Many people here on X will disagree with what he's saying. But before trolling or laughing or braying, try understanding what he's saying. He's not trying to persuade anyone. He's just speaking from his heart.

https://t.co/f8XjndRMgb

[Alexia Jolicoeur-Martineau @jm_alexia]

Simple beats complicated:

We show that switching to a sliding-window attention mask with attention sinks (at no cost) beats linear attention post-training.

Huge thanks to my collaborators @RheaSukthanker, @CameronPashmina, and @Emy_Aze.

Paper: https://t.co/h8DIc223Su

[doomslide @doomslide]

The lion does not concern himself with understanding

[Kaiser17 @Kaiser1710]

@michaelchchoi Are you gonna release your slides?

[Infosys Prize @InfosysPrize]

Indian physicist Prof. Deepak Dhar has been named a 2026 Dirac Medalist by the International Centre for Theoretical Physics (ICTP), recognizing his contributions to statistical mechanics.

His work on the sandpile model shows how tiny, seemingly harmless events can sometimes build up and trigger much larger, unpredictable changes in complex systems, things we see in real life like earthquakes, avalanches, or even sudden traffic jams.

Prof. Dhar is currently an Indian Science Academy (INSA) Distinguished Professor at the International Centre for Theoretical Sciences (ICTS), TIFR, in Bengaluru. After completing his PhD at Caltech, he returned to India and continued his research at TIFR.

The Dirac Medal is a significant recognition of his work and of the strength of theoretical physics research in India.

@insa_academy @TIFR_CAM @Caltech

#deepakdhar #diracmedal #theoreticalphysics #statisticalmechanics #physics #indianphysicist #indianresearch #scienceindia #infosysprize

Read the full story by The Print: https://t.co/WGFVCG6xGl

[SIAM Activity Group on Dynamical Systems @DynamicsSIAM]

Lecture notes: "On the kinetic Fokker--Planck equation in curved geometry" (by Cédric Villani): https://t.co/EeLhztdIzt

[Kaiser17 @Kaiser1710]

@chenna1985 What are your thoughts on deepmind alphafold?

[Kaiser17 @Kaiser1710]

@Kaju_Nut If reviewers use AI there is a chance your paper will be added to training data before it has been accepted anywhere . Are you okay with that?

[Lars Holdijk @HoldijkLars]

Great new paper from Sandeep Cranganore, Sebastian Lehner, @jo_brandstetter and @wellingmax!

Very nice perspective on escorting fields for finite-time driving! Sandeep shows that you can just read them off the FP generator's spectrum, at any speed!

https://t.co/NhjpZvnG0D

[Ahmed @ahmedallibhoy]

On the physics of optimization algorithms: a brief thread

Left: Ensemble of classical particles that seek the minimizer of f(x)

Right: Modulus of the wavefunction of a quantum mechanical system that seeks the minimizer of f(x)

[Kaiser17 @Kaiser1710]

@Nitin_wysiwyg @AIWarper Yeah you just wanna attack and force your view and not up for debate. Whatever, no point in talking any more

[Kaiser17 @Kaiser1710]

@Nitin_wysiwyg @AIWarper Tldr is don't bring coders logic into art community. It's very normal for coders to pick modules from github without crediting, but you can't do similar stuff for art without being called out. It's how it has always been

[Kaiser17 @Kaiser1710]

@Nitin_wysiwyg @AIWarper That isn't really correct way of thinking. What is right, what is wrong is usually decided by the community over time. As of now artist think of ai art as a plagiarism of their work. Ofc humans artist do that too and learn from others but no one in community thinks it to be bad

[Kaiser17 @Kaiser1710]

@Nitin_wysiwyg @AIWarper Trash, it's a ghoulish work built upon the stolen artwork of many artists who will never get compensated.

[doomslide @doomslide]

LLMs in isolation change very little. The adverse effects we're seeing are entirely due to an enormous scale up of bad faith actors (trillion dollar corporations) and the associated massive resource disparities relative to the academic establishment tasked with holding the torch.

[Noam Brown @polynoamial]

An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.

We believe it will be a major step for scientific reasoning. https://t.co/iP6cyheZ7i

[Kaiser17 @Kaiser1710]

@chenna1985 thoughts on this?

[Kaiser17 @Kaiser1710]

@chenna1985 https://t.co/iLf8PrfDXM

[Unitree @UnitreeRobotics]

Unitree Super Athlete AS2-W 😉
Fluid movement, compact yet exceptional. Supports a continuous 16kg payload, with 30+ km unloaded driving range