Similar User
Søren Sjørup  photo

@zorendk

Giga Chokheli photo

@agigao

“ “ photo

@lvtwo000

Andres photo

@andville_

Gerrit van Vuuren photo

@gerrit_jvv

Celio photo

@ccidral

Mamun photo

@mamuninfo2

Why I'm spending more time on the "Following" tab these days

my (~4.5lb) dissertation arrived, opened it up and immediately found a typo 🫠

Tweet Image 1


Saturday is a great day to contemplate the learnings available

How scientists changed their views on AI before and after the experiment

Tweet Image 1


Here it is, the moment when we realized OpenAI is not alone at the top.

🎥 Today we’re premiering Meta Movie Gen: the most advanced media foundation models to-date. Developed by AI research teams at Meta, Movie Gen delivers state-of-the-art results across a range of capabilities. We’re excited for the potential of this line of research to usher in…



Potentially the inflection point we've been waiting for.

We also propose a device for sampling from the posterior of Bayesian logistic regression with a Gaussian prior. The posterior is non-Gaussian, and in fact cannot be computed analytically. Therefore, this represents the first concrete proposal that uses thermodynamic computing to…

Tweet Image 1
Tweet Image 2


Thanks for presenting Asami to our small but very genuine group, Paula! I had no idea you could extend datalog to graphs, and with such elegance. @quoll


Looking forward to abstracting the db entirely... or wait

Want to learn what your database spends all its time doing? ⏱️ This paper by database legend Mike Stonebraker is an absolute classic. It profiles a transactional database in great detail, determining the purpose of each CPU instruction. The authors find that <10% of CPU…

Tweet Image 1


Sometimes, it's just better to wait until the pros do it

I am thrilled to announce the release of a long-time teased project built with @raphaelsrty Introducing PyLate, a flexible library to train and experiment with ColBERT models! It's built on the widely-used sentence-transformers framework, making it accessible and efficient!



I need to up my game

This summer, Taerim Yoon from Korea University joined our lab through the ETH Robotics Student Fellowship and developed a "Parkour Controller" using RL, enabling agile moves over obstacles. We’re excited to push this further for more agile legged locomotion. Please stay tuned!



Electrify all the things ⚡️

Tomorrow, Tuesday! 1:30ET, 1830 London. We have an awesome talk prepared for you!



Hot take

I am sad that Stephen Wolfram did not have the ambition to develop a serious replacement for mathematics, and instead gave us the world's most fancyful and restrictively copyrighted version of Lisp.



According to Jeff Hawkins, intelligence is correlated with the capacity for processing memories at multiple time frames. Next big architecture jump will be related to temporal encoding (prediction)

In our new paper, we show that adding self-modeling to artificial networks causes a significant reduction in network complexity When artificial networks learn to predict their internal states as an auxiliary task, they change in a fundamental way

Tweet Image 1


cogwire Reposted

Good question I do want Eureka Labs to be a proper, self-sustaining business but I also really don't want to gatekeep educational content. My default thinking is that the content itself is free and permissively licensed, the revenue comes from everything else, e.g. running the…


It's happening! Can't believe it really, but very grateful. Signing up my boys ASAR(eady)

⚡️ Excited to share that I am starting an AI+Education company called Eureka Labs. The announcement: --- We are Eureka Labs and we are building a new kind of school that is AI native. How can we approach an ideal experience for learning something new? For example, in the case…

Tweet Image 1


While searching for Fredrik Solstad, author of SIMIT, I stumbled upon this open source gem. Hope to gen me some memory proteins news.stanford.edu/stories/2024/0…


I'm only here for the mind blowing content.

"Quantum mechanics is just thermodynamics in imaginary time." It sounds like a hand-wavy, quasi-philosophical statement. But with a brief dive into the relationship(s) between hyperbolic and parabolic PDEs, it becomes possible to formalize it mathematically. (1/16)

Tweet Image 1
Tweet Image 2


If you only read one paper this month, let it be this one.

🚨Announcing the largest study focused on *how* to optimize the prompts within LM programs, a key DSPy challenge. Should we use LMs to… Craft instructions? Self-generate examples? Handle credit assignment? Specify a Bayesian model? By @kristahopsalong* @michaelryan207* &team🧵

Tweet Image 1


Loading...

Something went wrong.


Something went wrong.