Building Gen AI products. Ex-Facebook, I work on AI, Search and ML Infra. Cautiously optimistic about the future and a believer in being nuanced.

San Francisco, CA
Joined November 2009
Nishanth PS retweeted
F.03 just crossed 9 straight days walking on this treadmill: 348 miles and 1.2 million steps Started as a side project to stress the robot for any failures. Let's see how long it can go
327
232
73
4,581
213,442
Nishanth PS retweeted
yea the great debloating and blocking ads on your tv revolution has begun this is the way: - install adb - get ip address of tv - turn developer mode on that's it claude or deepseek does the rest 🔥
This is how fast the TV moves now that Claude deleted all the BS laggy bloatware it comes with
69
91
16
2,661
425,524
Nishanth PS retweeted
We noticed this early in the Thiel Fellowship.
One of the biggest problems with sending children to public school is that they learn the habits of dependence - be where you’re told, do what you’re told, accept what you’re told to be the truth. The system is not designed to produce independent people.
3
9
2
132
7,537
Me inspecting the codebase after an overnight /goal
33
85
9
1,679
59,555
Nishanth PS retweeted
We placed the fruit fly brain inside Odyssey-3, a frontier world model. The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
670
1,765
1,180
20,265
6,094,855
Nishanth PS retweeted
I’ve spent a few days thinking about the Tesla Roadster reveal and landed somewhere I haven’t seen much discussion. I’m neither a Tesla fanboy nor a Tesla hater. And I couldn’t care less whether what we saw technically qualifies as a flying car. Three takeaways: (1) The risk of that live demonstration going horribly wrong was real. Almost no CEO would willingly expose their company to that kind of public humiliation for a product with so little financial upside. Elon did. (2) That willingness to attempt the seemingly impossible is what makes Tesla special. It’s also arguably Elon’s greatest contribution to America. He’s helped bring back something we’ve lost. The American spirit of chasing big scary ideas that most people think are impossible. That mindset is spreading across a new generation of founders and its long term impact on America’s future could dwarf anything Tesla ever sells. I think this is deeply underappreciated. (3) Can we just acknowledge how insanely cool that moment was? We’ve reached a strange place culturally where even extraordinary feats of human ingenuity immediately become arguments about why they don’t count. Not everything needs to be a debate. Sometimes we should just be able to stand back together and marvel at what humans are capable of. I’ve been both bullish and critical of Tesla’s ambitions over the years. But watching Elon and his team push the boundaries of what’s possible reminds me why I’m glad we have him. Whether the Roadster ever makes Tesla a dollar or impacts $TSLA stock price is almost beside the point. We need more people willing to risk failure in pursuit of things the rest of us think are impossible. Because when they succeed we all win.
1,067
2,381
340
19,265
603,851
"Mission-driven companies do better because they attract better people. It’s easier to align vectors if you’re on a mission." This is under appreciated.
Antonio Gracias explains the lessons he learned from Elon Musk: pro-entropic, action in the space of fear, mission-driven, and compassion for humanity “I mean, I think the things about him that are, I’ve learned a lot from him, by the way, and I definitely wouldn’t be me without him, and the things that I’ve learned and observed in him." "The first thing I think he does, he’s very, very pro-entropic, right? He pushes the probability map to the very edge. To him, if it’s possible in the physics, it’s possible, period. He will fight for things that have low probability of success, but high expected outcomes. So making rockets, making rockets land, this seemed almost impossible, right?" "But it was worth doing because even though it was almost impossible, it was possible. It had low probability, but high, high expected value. And that is a, it’s very, very valuable to society. It’s hugely valued, man, if you do that. I think that’s the first thing.” “The second thing is he’s very courageous. I mean, these are like, you know, and courage, by the way, it’s not the absence of fear. It’s action in the space of fear. Does that extremely well." "I think the third thing is super mission-driven. I think we’ve developed our mission-driven hypothesis because we learned from him and others that mission-driven companies do better because they attract better people. It’s easier to align vectors if you’re on a mission." "And part of being mission-driven is because he’s very compassionate. I mean, he’s a very compassionate human. He cares about humanity deeply and he wants what’s best for humanity. It’s why he’s doing what he’s doing. I mean, those have been very important lessons.”
40
As Ilya said “if you value intelligence above all other human qualities, you’re gonna have a bad time” xitter.cf/ilyasut/status/1710462…
a comment from 3blue1brown's most recent video
1
46
Nishanth PS retweeted
Whale 🐋 is back
🚀 DeepSeek-V4.1-Flash-1017 is here! Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API. 🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model. 1/4
40
135
10
4,579
220,724
Nishanth PS retweeted
"how to avoid brain rot" was a great watch. some of my favorites points: - using ai to ask yourself socratic questions instead of giving you the answer makes you feel like you're learning less but research shows you're actually learning more - ai tooling is moving so fast that it's better to know 50% of a new tool than to master 90% of an older tool
Here is the full talk:
41
252
11
3,371
156,605
Nishanth PS retweeted
You can like this. And this. Like More Things. What a day for cars!
44
32
4
882
64,026
Interpretability for large neural networks needs tools and agentic exploration of the space. Maybe throwing compute at the problem might work.
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works already-clean.goodfire.com/
38
This is what my scope creep looks like
Turns out the future does have flying cars
37
Nishanth PS retweeted
Replying to @elonmusk
People don’t understand how cool and inspiring it was for my 6 year old to watch this.
1
2
167
So will we see cars doing flips in the air now?
NEW: Highlight reel of all the stunts performed at the Tesla Roadster event. INSANE 🔥🔥🔥
63
my absolute favorite detail of the new porsche mission s I imagine this is the likely direction that software will go as well
53
114
47
4,674
194,246
Nishanth PS retweeted
Religion, politics, and sports are all based on identity. That's why it's nearly impossible to have a meaningful discussion about them. @naval
39
187
10
1,454
43,372
🤣 Most people don't understand the amount of context you need to begin coding. The more you delegate to AI the more you will lose the ability to build that context quickly.
When you hit Claude’s limit and manager asks if you can keep coding on your own?
1
3
187
This is probably a quick way to test treatments.
What a fascinating study.
35
Nishanth PS retweeted
my life before sama's tweet:
today we launched ChatGPT. try talking with it here: chat.openai.com
40
184
9
2,881
102,378