@chuswlove Profile picture

chus

@chuswlove

all things machine learning. Systems for AI Lab @GeorgiaTech. Check out my blog :)

Similar User
Damii photo

@damidelchill

Andrew photo

@andreea_jpg

Oriol ⚽ photo

@_urifuentes

Roger Pinyol Fuertes photo

@pinyolroger

Pew! photo

@pewcorbi

Albert Rojas photo

@ruohas

Ilena photo

@ilenaireland

Oscar Lázaro photo

@oskrlaz

Roser Cantenys Sabà photo

@CantenysR

Paula Bonay photo

@paula_bonay

Pinned

🔥🔥SNAKE ALERT🔥🔥 i just published my 3 part blog series on State Space Models. A deep dive from 0 to mamba-2: - 10,000+ words - 25 custom figures - Interactive simulations And it's free! Check it out here: chus.space appreciate for sharing!


god bless America

Tweet Image 1

chus Reposted

I compiled all the emails released as part of the Musk v. Altman lawsuit in chronological order (link in reply). IMO a really valuable read. Extremely consequential decisions made in these emails.

Tweet Image 1

i found this dress and i see these weird green lines??

Tweet Image 1

chus Reposted

Moravec's paradox in LLM evals I was reacting to this new benchmark of frontier math where LLMs only solve 2%. It was introduced because LLMs are increasingly crushing existing math benchmarks. The interesting issue is that even though by many accounts (/evals), LLMs are inching…

1/10 Today we're launching FrontierMath, a benchmark for evaluating advanced mathematical reasoning in AI. We collaborated with 60+ leading mathematicians to create hundreds of original, exceptionally challenging math problems, of which current AI systems solve less than 2%.

Tweet Image 1


Does anyone here know about sketches? (exponential histogram, count-min…)


is anyone here doing the Gemini kaggle competition?


thx to all new followers. i will make it WORTH IT


imagine this being saturated in a few years and models still being bad in random af benchmarks

3/10 We evaluated six leading models, including Claude 3.5 Sonnet, GPT-4o, and Gemini 1.5 Pro. Even with extended thinking time (10,000 tokens), Python access, and the ability to run experiments, success rates remained below 2%—compared to over 90% on traditional benchmarks.

Tweet Image 1


chus Reposted

chus (@chuswlove) has an amazing 3-part blog series on state space models, mamba, which I am planning to go through this weekend go read his blogs and support him

Tweet Image 1
Tweet Image 2

i’ll be doing research at the GeorgiaTech Systems for AI lab until next autumn REALLY COOL. still a lowbie though


i finished PMPP. An amazing book!


you might be old, but old enough to remember Caffe?


all base M2 and M3 macbook air now come with 16gb ram. A forced hand huh


the trombone emoji proposal... wop wop

Tweet Image 1

This year's emoji finalists have finally been revealed and this is one of them

Tweet Image 1


how can a youtube video be of so much quality? truly next level youtube.com/watch?v=h9Z4oG…


chus Reposted

I will tolerate recurrent laryngeal nerve slander no longer! It's actually the result of several elegant solutions to difficult problems in embryology, and the length is a non-issue. A 🧵 1/13

My meager education in biology and evolution gave me the mistaken impression that evolution optimized everything. But it didn't. One example is the recurrent laryngeal nerve (RLN). It goes from behind your ear, loops down below your aorta, and then back up to the voice-box 1/2

Tweet Image 1


__syncthreads() gives me comfort feels like home, cozy and expected ly syncthreads


why not... 3d conv tiled kernel

Tweet Image 1

Chapter 7: convolutions is up!



United States Trends
Loading...

Something went wrong.


Something went wrong.