pranav

pranav

@pranav.bsky.social

Research Scientist at Google DeepMind. ಕನ್ನಡಿಗ. Past: Researchoor, Algorithms team at OpenAI & with Juergen Schmidhuber.

1.8K followers246 following121 posts

pranav
@pranav.bsky.social

Everything about that Luigi guy is just sad. Manifesto written by an LSTM, no big plan or idea, got caught in the dumbest way imaginable, miserable health condition.. just sad overall

pranav
@pranav.bsky.social

People think scaling laws started with Kaplan et al. But almost 3 years before that this paper came out with power law fits, data and compute scaling laws on language modeling. arxiv.org/abs/1712.00409 Goes to show how having the right idea is only a small part of success

arxiv.orgDeep Learning Scaling is Predictable, EmpiricallyDeep learning (DL) creates impactful advances following a virtuous recipe: model architecture search, creating large training data sets, and scaling computation. It is widely believed that growing tra...
pranav
@pranav.bsky.social

Looks like Sutton didn’t get the memo. Memory and compute are cheap. It’s called the “Bitter Lesson” Replay buffers are a-ok!

Tom Dupuis@tomdupuis.bsky.social

Soooo....... Rich Sutton's team just leapfrogged the whole field with this banger, solving RL with pure streaming data: arxiv.org/abs/2410.14606

arxiv.orgStreaming Deep Reinforcement Learning Finally WorksNatural intelligence processes experience as a continuous stream, sensing, acting, and learning moment-by-moment in real time. Streaming learning, the modus operandi of classic reinforcement learning ...
pranav
@pranav.bsky.social

Death of pre-training has been greatly exaggerated. Look at image generation models. We’re so far from compute optimality yet pre-training is by far the most important bit. Diffusion models are like o1: train for 1 step but unroll for many. Even test time inference is pre-training bottlenecked.

pranav reposted

Naomi Saphra
@nsaphra.bsky.social

hard for me to get engagement on this platform but it’s ok I’m just waiting for all the AI researchers to show up so I can go mega viral saying shit like “who called it Xavier Initialization and not Weighting For Glorot”

pranav
@pranav.bsky.social

Going on an impromptu hike Rated challenging In an unknown country Zero sleep Not prepared re clothes/hat/shoes or food/water Need to come back and pack for an early morning flight immediately after Rain, even thunderstorms forecast on the trail Let’s see what I make of this in some time

pranav reposted

Margaret Mitchell
@mmitchell.bsky.social

“I asked an AI program to write an entire sci-fi novel for me, and [as someone who will die an empty shell of a man who wasted his life doing nothing for the world...] I was super impressed.” https://www.theonion.com/guy-who-sucks-at-being-a-person-sees-huge-potential-in-1850488022 😅

theonion.comGuy Who Sucks At Being A Person Sees Huge Potential In AISAN MATEO, CA—After spending the past three decades of his life being totally unable and unwilling to engage in any meaningful way with the world around him, James Parker, a local guy who sucks at b...
pranav
@pranav.bsky.social

Starting with the foundation of sobriety. With lots of hard work and vigilance. With copious support from friends and family. With a little bit of luck…

pranav
@pranav.bsky.social

We can build really great tools with LMs and AI in the next few years. We have to sit down and put in the hard work. Not rush to through the door naked screaming eureka.

pranav
@pranav.bsky.social

Man this MSFT stock is made up of something else. I really need to sit down and understand their cashflow one day. I have a huge suspicion that their cloud revenue is recycling their old money

pranav
@pranav.bsky.social

I’ll never be as badass as the dude who splashed leftover wine in the glass in the floor in All Quiet on the Western Front. He’s so entitled he won’t go for that tiny last sip. Instead throws it out for a fresh full glass. What an a hole