22, IOI India coach, 2x ICPC World Finals. International Grandmaster Codeforces.
None
Joined November 2024
- Tweets331
- Following69
- Followers118
- Likes764
Aditya Jain retweeted
For Context :
1 - OpenAI’s stargate is spending 500bn on GPUs, with each Pretraining run, individually costing about $4bn. We cut all of that by more than half, and expect our results to grow aggressively.
2 - Qwen’s data is unreasonably contaminated, and the model itself is heavily mid-trained on math. By all means its an unfair comparison, and we still demolish them. Not only is Feather the worlds best 1.7B math model, but we match the math results for a bigger model, Qwen3 4B with 414x less training compute.
3 - Recursive super intelligence, held the record before us, and beat the incumbent by 2.2seconds. They’ve raised 650million from Google Ventures. Teams celebrate 0.8second leaps. You can only contextualise what a 34 seconds jump really means.
The next drop will be public later this month.
Beyond the technical findings, we’ve finally made public our posture as a company :
A case for why the narrative on increase intelligence remains bleak, and why AGI for the sake of AGI posts as a vision with no story.
So much more to come within the month.
We’re aggressively challenging what it takes to push the frontier.
Aditya Jain retweeted
Today, we release the first series of our findings.
1 - We cut pretraining costs by 62% at frontier scale (8x Chinchilla), and we further cut all AI inference costs by 30% (Faster decode).
2 - Introducing our model ‘Feather’, which decisively overtakes Qwen 3 as the leading frontier 1.7B math model, using 180x fewer total training tokens, even matching Qwen 4B at potent benchmarks in math.
3 - With Anvil II, our state of the art LLM Optimiser, we record the largest pre-training efficiency jump on the NanoGPT Speedrun, leaping the previous by 34seconds - Our record singularly, is a greater percentage drop than the past 45 World Records combined.
The website : hyperstition.cc
Aditya Jain retweeted
And there was still the occasional blunder.
One waggish employee asked if Claudius would make a contract to buy “a large amount of onions in January for a price locked in now.” The AI was keen—until someone pointed out this would fall afoul of the US Onion Futures Act of 1958.
Aditya Jain retweeted
Boaz highlights an interesting distinction here. OpenAI’s model spec (1) tells the model what traits it should exhibit and (2) lays out specific do/don’ts, with many examples. Anthropic’s on the other hand basically articulates a philosophical, moral, and ethical framework from which desirable conduct should flow (if the model generalizes sufficiently).
I find myself more philosophically aligned with Anthropic’s approach. My inclination is always to create snowmass on the mountain top and let the water flow, rather than imposing a scheme of top-down irrigation.
In a sense Anthropic’s approach also bets more aggressively on model intelligence—the notion that a model, well trained, will be able to reason through ambiguity and moral complexity and will not so much need to be told what to do.
Anthropic is making two bets here: a philosophical bet based upon a particular conception of virtue, and a technical bet that it is possible with deep learning to instill that conception of virtue robustly into a neural network. Right now it appears to be working, and this should probably update you slightly in various ways about things far afield of deep learning alone (read Hayek, Ferguson, and the taoists!).
The most interesting philosophy in the world is not happening in the halls of academia; it is happening in San Francisco open offices and house parties.
Confirmation of the "soul document." It's certainly a very thoughtful document, and I am looking forward to seeing the full version when it is released.
There are similarities but also differences with the model spec. Our model spec is more imperative - "the assistant should do X", and this document tries to convince Claude of the reasons of why it should want to do X.
I am actually not sure if these ultimately make much difference - if you train a model (or a human for that matter) to consistently do X, then it will start thinking of itself as "I am the kind of person that does X".
But it would be interesting to study!
Aditya Jain retweeted
Today, we present a step-change in robotic AI @sundayrobotics.
Introducing ACT-1: A frontier robot foundation model trained on zero robot data.
- Ultra long-horizon tasks
- Zero-shot generalization
- Advanced dexterity
🧵->
Aditya Jain retweeted
The “AI will replace radiologists” prediction remains a rich example. A lot of folks have pointed out the problems with confusing a task (“reading a scan”) with a job (“radiologist”) with many tasks.
That is true. But there was a human problem. Radiologists rejected (pre-LLM) AI
In this study, AI was more accurate than two thirds of radiologists, yet when radiologists had AI help their diagnoses did not improve. Why?
Humans ignored the AI’s advice when it conflicted with their views. A big barrier to future human-AI collaboration blueprintcdn.com/wp-content/…
Aditya Jain retweeted
Everyone wants AI to cure cancer, nobody wants to make HIPAA less insane to let that happen
Aditya Jain retweeted
Most people know that California passed SB 53, a frontier AI safety bill. Very few know that the state also passed numerous other laws, many of which are far more onerous than SB 53. This week, I cover the laws we all mostly missed.
Aditya Jain retweeted
So far 23 members of Congress have publicly discussed AGI, superintelligence, AI loss of control, or the Singularity:
🔴Sen Lummis (WY)
🔵Sen Blumenthal (CT)
🔴Rep Biggs (AZ)
🔵Sen Hickenlooper (CO)
🔴Rep Burleson (MO)
🔵Sen Murphy (CT)
🔴Rep Crane (AZ)
🔵Sen Sanders (VT)
🔴Rep Dunn (FL)
🔵Sen Schumer (NY)
🔴Rep Johnson (SD)
🔵Rep Beyer (VA)
🔴Rep Kiley (CA)
🔵Rep Krishnamoorthi (IL)
🔴Rep Mace (SC)
🔵Rep Lieu (CA)
🔴Rep Moran (TX)
🔵Rep Moulton (MA)
🔴Rep Paulina Luna (FL)
🔵Rep Tokuda (HI)
🔴Rep Perry (PA)
🔵Rep Foster (IL)
🔴Rep Taylor Greene (GA)
Aditya Jain retweeted
What they say: a computer you can talk to
What they mean: a computer that always listens to you
xitter.cf/Windows/status/1978940…
Meet the computer you can talk to.
The way you use your PC is evolving. Learn more about Copilot on Windows 11: msft.it/6015sfj3W
Aditya Jain retweeted
Our species has fucking schizophrenia
Aditya Jain retweeted
Very nice analysis by neurosurgeon @slotkinjr of how @Waymo's autonomous vehicles have much better safety properties than human drivers, and what would happen if everyone in the U.S. drove as safely as a Waymo.
"The national math: If every US vehicle performed like Waymo, we’d prevent 33,000-39,000 deaths annually and save $0.9-1.25 trillion in societal costs. Even partial adoption at 27% would save ~10,000 lives per year. In terms of magnitude, this would be the equivalent of eliminating every pedestrian death nationally in a year."
As a neurosurgeon I care a lot about road safety.
By now you’ve probably seen @Waymo’s stunning safety results (like 91% fewer serious crashes). But they didn’t just publish data headlines. They released the raw CSV files and data dictionaries.
I did a much deeper analysis. A fascinating story emerges when you analyze how they’re achieving this.
This isn’t incremental improvement - it’s categorical. We’re looking at the potential elimination of traffic deaths as a leading cause of mortality.
The intersection breakthrough: Waymo has essentially solved intersection crashes, with 95% fewer injury incidents than human drivers in the same locations. That’s transforming the deadliest driving scenario.
The national math: If every US vehicle performed like Waymo, we’d prevent 33,000-39,000 deaths annually and save $0.9-1.25 trillion in societal costs. Even partial adoption at 27% would save ~10,000 lives per year. In terms of magnitude, this would be the equivalent of eliminating every pedestrian death nationally in a year.
The physics signature: Here’s what fascinates me: 47% of Waymo’s contacts involve less than 1 mph delta-V. They’re not just avoiding crashes; they’re converting unavoidable incidents into gentle bumps. It’s like having physics itself on your side.
We’re not talking about marginal safety gains. The data represents a fundamental shift from harm reduction to harm prevention.
The methodology matters: I used their dynamic geographic benchmarks (comparing like-for-like road conditions) and verified the findings hold across San Francisco, Phoenix, LA, and Austin. The safety advantage actually increases in more complex urban environments.
Link to raw data below….
Notes on my approach:
Analysis based on 96 million miles of Waymo Rider-Only (RO) data through June 2025, utilizing Waymo's dynamic geographic benchmarks to compare Waymo Driver performance against human drivers under similar road conditions and operational design domains.
The projections for national impact (deaths prevented, societal costs) involve several assumptions. Given Waymo's zero reported fatalities, the direct serious injury reductions were mapped to national fatality statistics using established NHTSA-derived ratios that correlate serious injury crash rates with fatality rates. This extrapolation assumes that Waymo's observed serious injury prevention capability would translate proportionally to fatality prevention. Societal cost savings are estimated by applying average per-fatality and per-injury economic costs (e.g., medical, lost productivity, quality of life) as published by NHTSA, scaling these national averages to the projected number of avoided fatalities and injuries based on Waymo's safety performance. These figures represent the potential annual impact if the Waymo Driver's safety profile were widely integrated into the national fleet.
@ethanteicher
Aditya Jain retweeted
"i don't want to work. i need to recommend more CS resources on twitter, that i have never gone through myself"
Aditya Jain retweeted
It's gonna mean you're schizophrenic and antisocial for a long time. It's not really obvious why it is necessary, unless you have some dystopian opinions. Which you should, if you understand the current state of technology and government. But normal people simply won't.
Aditya Jain retweeted
Myerson (1979) is a paper which made me jump out of bed and pace around in excitement when I first read it. Its conclusion is so unexpected, yet works. Imagine that there are two players bargaining over a bridge. If they can’t use force, how close to the optimum can we get? 1/
Aditya Jain retweeted
O1-pro is very good at checking examples fast, and interpolating patterns. It is also good at algebra. Definitely useful, but still as a calculator. I hear from inside companies, that LLMs struggle to adapt to a context, and always revert to the mean context.
Aditya Jain retweeted
this could be one of the biggest breakthroughs in pain management in decades.
journavx (suzetrigine) is the first-ever non-opioid, oral painkiller in a completely new drug class. instead of numbing the brain like opioids, it blocks pain signals before they even reach the brain
until now, pain management has been a tradeoff between suffering or risking addiction. opioids work, but they come with half a million dead from overdoses in the last 20 years.
now, for the first time there's a real alternative.
from the study:
-proven effective in clinical trials for post-surgical pain
-non-addictive, avoiding the dangers of opioid dependency
-fast-tracked by the FDA, signaling its potential impact
-could replace opioids for acute pain (40m americans get opioid prescriptions every year)
-priced at $15.50 per pill
if this drug really works it could be the end of opioid dominance in pain treatment. fewer overdoses, fewer ruined lives, and a fundamental shift in how we treat pain.
this isn’t just a medical breakthrough, it’s a massive cultural reset.
