Sathvik Redrouthu
@_sathvikr
Math @Caltech. Past: @Etched, YC dropout. @zfellows.
Our @ycombinator cofounder video! In 6 weeks, we built state-of-the-art (SOTA) chip design agents, and caught 5 bugs in a certain company's next AI chip (saving an estimated $5M) Check us out: getinstachip.com
Exciting news on @diodeinc published on Business Insider today. 1/ We raised capital! Over $14.5m, most recently in a $11.4m series A round led by @a16z (details below) 2/ We are working with fortune 100 companies and fast growing startups. If you are building hardware, we want…
We’re proud to announce our $11.4 million Series A, led by @a16z, with the continued support of @ycombinator, @caffeinatedcap and @BoxGroup. We are working with Fortune 100 companies and fast growing startups to design and manufacture their circuit boards faster than ever.
We’re proud to announce our $11.4 million Series A, led by @a16z, with the continued support of @ycombinator, @caffeinatedcap and @BoxGroup. We are working with Fortune 100 companies and fast growing startups to design and manufacture their circuit boards faster than ever.
A hallmark of human intelligence is the capacity for rapid adaptation, solving new problems quickly under novel and unfamiliar conditions. How can we build machines to do so? In our new preprint, we propose that any general intelligence system must have an adaptive world model,…
GUYS STOP USING EXPENSIVE AS A DISQUALIFIER. capability per dollar will drop 100x/year. “$3k task ARC-AGI 80%” could prob be $30 if we cared to optimize it. repeat after me: all that matters is top line intelligence. all that matters is top line intelligence…
All the tech bros this morning thinking that AGI has been achieved because some (insanely expensive) new form of LLMs can now match top *high school students* on one specific task … it’s almost … cute! ☺️
Based on public information, major AI labs are pushing two AI reasoning frontiers to improve "process models" that generate the reasoning chains (or reasoning programs): 1. More search 2. More domains More test-time search is being deployed via improved process models to cover…
Even 3b1b is not immune to it
Welp, somewhat embarrassingly, yesterday this account was briefly compromised and someone posing as me promoted a crypto project supposedly "tokenizing" manim. Thank you to those of you who quickly flagged it as spam and alerted me to the problem. For future reference, it's a…
This is amazing and generalizes soo well in so many ways
Today we're releasing a developer preview of our next-gen benchmark, ARC-AGI-3. The goal of this preview, leading up to the full version launch in early 2026, is to collaborate with the community. We invite you to provide feedback to help us build the most robust and effective…
Today we’re releasing our first public preview of ARC-AGI-3: the first three games. Version 3 is a big upgrade over v1 and v2 which are designed to challenge pure deep learning and static reasoning. In contrast, v3 challenges interactive reasoning (eg. agents). The full version…
Why? Because AI nudges a curve to an outlier even if the delta is +- inf. Instead when presented a new outlier you need to rewrite all your rules to fit all data points. Something I am working on during my free time
AI can't figure out the inverse square law looking at 10M solar systems, Newton figured it out looking at 1
Our thesis at Ndea: simple theories should be discoverable using little data and little compute. And everything humans have ever invented is a fairly simple composition of fairly simple theories.
AI can't figure out the inverse square law looking at 10M solar systems, Newton figured it out looking at 1
Scaling up RL is all the rage right now, I had a chat with a friend about it yesterday. I'm fairly certain RL will continue to yield more intermediate gains, but I also don't expect it to be the full story. RL is basically "hey this happened to go well (/poorly), let me slightly…
Same for me They went on to start an eval company Even though @gdb said not to do it These people stop at nothing 🤦♂️
Yeah similar thing happened to me (must’ve happened a lot our batch lol). Ultimately missionaries > mercenaries, and we’re playing a 30 year game. Wish yc had a way of filtering out sociopaths tho.
This is a nice experiment: if you finetune an Othello next-move-predictor to reconstruct the board from its internal state, the reconstructed boards are often incorrect, but they have the same next moves as the true board! So next token prediction might be "too easy", in that a…
We fine-tune an Othello next-token prediction model to reconstruct boards. Even when the model reconstructs boards incorrectly, the reconstructed boards often get the legal next moves right. Models seem to construct "enough of" the board to calculate single next moves.
Great paper! Similar line of research and would love to chat.
Can an AI model predict perfectly and still have a terrible world model? What would that even mean? Our new ICML paper formalizes these questions One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵
Can an AI model predict perfectly and still have a terrible world model? What would that even mean? Our new ICML paper formalizes these questions One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵
Ive never met someone who actually uses @cluely until today
