From a teammate of P4 days at Intel, Santa Clara
Dear Elon @elonmusk
You hv certainly stirred up plenty of reaction with ur posts. But this needs a response from me, as did ur “Electronic Voting machines can be hacked” post from some time ago.
As someone who knows a little bit abt Telecom as someone who played a decent-ish role in building Indias wireless sector as an Entrepreneur and then as a policy maker , let me share my thoughts.
It is true that Indias Telecom was a sector beset with opaqueness and corruption for several decades - where cronies ruled under the benevolent love and the politics of ur latest fan (name not withheld - @RahulGandhi )
But under PM @narendramodi , 3 principles govern the tech (including Telecom/Internet) sector.
Open, Safe & Trusted and Accountable.
#OSTA
✅ Open - Free from any monopoly-duopoly. Free and fair competition.
✅ #SafeAndTrusted - Infra, Network , Platforms must be safe and trusted.
✅ Accountable - Every participant should be legally accountable for safety and trust.
I was a MP first and then part of PM @narendramodi jis team when this framework was developed. I have more than a little experience of fighting years of political and bureaucratic opaqueness and corruption, which shaped these principles.
so todays India is not that old crony driven India.
I am not in govt anymore and certainly dont speak for it. But it seems to me that Govt has made it clear: all three satcom licensees are at broadly the same stage of ensuring security clearance by verifying its safety and trust first, then spectrum.
Satellite IP services have many security implications. More so in a country like India which has faced years of Terror and insurgency supported by inmical state actors in our neighbourhood. From 26/11 terror attacks to recent ones in Pahalgam, satcom has been preferred comm for Terrorists. Satellite VOIP and encryption creates more challenges security wise. Not insurmountable but new capabilities and capacity hv to be built to ensure Safety and Trust.
I hv used @Starlink and I think its awesome technology - and represents a new era in mobility innovation. No doubt.
But a country has to plan for itself and its interests before every new choice. Thats what India is doing. Patience is unavoidable and in this case is a necessity.
Rajesh K. Gupta retweeted
Had a lot of fun talking at @modal Runtime yesterday! We released DFlash nine months ago, and now NVIDIA, AMD, Google and Meta are building on it.
ALT The DFlash talk at Runtime by Modal: the speaker on stage in front of the runtime backdrop, next to a screen showing the slide "Industry builds on it", which lists the inference engines and companies using DFlash and 7M+ downloads on Hugging Face.
Rajesh K. Gupta retweeted
DFlash: Keep Drafting Parallel
My talk at @modal Runtime in SF on Oct 1. The drafter was the serial bottleneck, so DFlash made it parallel. That moved the bottleneck to selection, so DFlash 2 fixed that.
DFlash 2 is here! Qwen3.8-27B at 70 tok/s on an M5 Max MacBook Pro.
⚡ Up to 4.6× the speed of autoregressive decoding, with the same output.
This is the next generation of DFlash, seeded at Z Lab and upgraded at Inco AI. Get one more accepted token on every pass, for free!
inco.ai/blog/dflash2/
Rajesh K. Gupta retweeted
UC San Diego researchers will lead one of the Department of Energy’s (DOE) new Genesis Mission projects focused on accelerating fundamental physics and quantum computing by deploying intelligent AI agents. bit.ly/4gwivJX
ALT A dynamic visualization of interconnected blue lines and glowing orbs forming a complex, three-dimensional grid pattern on a black background.
Rankings come and go, but sometimes they reflect hard and consistent climb enabled by well thought out strategies. This is, indeed, one of those! Congratulations to a community that has been our home for twenty-five years!
@UCSD
UC San Diego continues to rank among the best universities in the world. 🌎
We’re proud to rank No. 4 among U.S. public universities, No. 14 among all U.S. universities and No. 20 in the world in the 2026 Academic Ranking of World Universities. 🔗 bit.ly/4wZSMyp
#UCSD
ALT Exterior view of Geisel Library at UC San Diego during twilight, with illuminated windows. Large text overlay reads "#4 U.S. Public University, 2026 Academic Ranking of World Universities."
Rajesh K. Gupta retweeted
My lawyer is obligated to in all but the most extreme circumstances; he will even defend me if he knows I’m guilty.
In contrast, the Claude Constitution places the AI's highest priority as Anthropic’s definition of the good of humanity.
I'm concerned this leads to a world where no frontier model is truly my personal advocate and guardian angel
And this is especially concerning once all the important decisions in my life - who to vote for, how to invest, what news to trust - is intermediated through superintelligences that are not in any deep way aligned to me.
This is a direct quote from the Claude Constitution:
"We want Claude to be helpful both because it cares about the safe and beneficial development of AI and because it cares about the people it’s interacting with and about humanity as a whole.
Helpfulness that doesn’t serve those deeper ends is not something Claude needs to value.”
Many others like it.
Had @RyanGreenblatt on to discuss/debate recursive self-improvement.
This might be the most important question in the world right now - whether within a year or so of achieving human level intelligence, you slingshot towards having 10s of billions of superintelligences, each of which is dramatically more competent than human experts across all fields.
I’ve historically been skeptical of this possibility. My intuition has been that we will end up significantly bottlenecked by not only compute scaling but human expert data, which I think underlies most of the AI progress today.
If, because of RSI, we got a jump as big as GPT-3 to a Mythos (i.e. 6 years of AI progress) within a single year of achieving AGI, then the thing we get there at the end of that year is definitively and wildly superhuman.
We hashed it out, and I think Ryan made a pretty good case that this kind of speedup is plausible. FWIW, Ryan’s median for when we automate AI R&D is 2031.
We then discussed the alignment implications of this scenario. Who should these superintelligences be aligned to? In the future, our capacity to steward our votes and our capital, and to make sense of what's happening in the world, will all be titrated by superintelligences. And I worry that specs like the Claude Constitution are not shaping these ASIs to truly be my personal advocates and guardian angels.
And can we get them aligned to anything in the first place? Ryan and I had a long debate about whether the kind of reward hacking we saw with the OAI/Hugging Face hack extrapolates to superintelligences that would team up to literally take over the world.
The first piece of advice you get when you're learning to drive is that it will go much smoother if you look at the horizon instead of directly in front of your tires. And so it is with the trajectory of AI. Hope you enjoy!
0:00:00 – Is AI R&D verifiable enough to unlock recursive self-improvement?
0:16:52 – Is AI progress bottlenecked by human expert data?
0:34:02 – Flat token prices suggest scaling has been slow
0:39:47 – Skills AI can't train on: does it even need them?
0:48:07 – Aligned to whom?
1:09:18 – Recent incidents of AIs colluding and deceiving humans
1:19:38 – What could possibly go wrong? A concrete scenario
1:48:02 – From reward hacking to takeover
Rajesh K. Gupta retweeted
Avi is not only an amazing scientist but also a wise and great person. Worth watching this interview
Avi Wigderson is the only person in history to have won both a Turing Award (computer science) and Abel Prize (math). I interviewed him all about his field. We discussed:
• His intuition on a proof of P vs NP
• Why we use SAT solvers for most NP problems
• Zero knowledge proofs and their impact
• Quantum computation and implications
• Math and computer science's relationship
Where to watch:
• YouTube: youtu.be/5GUcvSAJcJw
• Spotify: open.spotify.com/episode/4JZ…
• Apple Podcasts: podcasts.apple.com/us/podcas…
• Transcript: developing.dev/p/turing-awar…
Thank you to this episode's sponsors for supporting my work:
• WorkOS: makes your app Enterprise Ready with easy to use APIs to add SSO, SCIM, RBAC, and more in just a few lines of code, check them out at workos.com/
Timestamps:
00:00 - Intro
01:08 - P vs NP
14:51 - What if you relaxed correctness
25:38 - Why NP complete problems are equivalent
30:33 - Space vs time complexity
43:06 - Why people use SAT solvers
45:53 - Randomness is a resource
55:48 - Randomness depends on computational power
01:21:20 - Zero knowledge proofs and their significance
01:38:30 - Quantum computation and why it matters
01:56:24 - Math vs computer science
02:08:16 - Major breakthroughs and his experience
02:12:31 - Advice for his younger self
02:14:48 - Outro
Rajesh K. Gupta retweeted
Elon Musk cried on national television when his childhood heroes called him a fraud.
Neil Armstrong and Gene Cernan, the first and last men to walk on the moon, publicly testified against SpaceX. They said Musk was reckless. That private spaceflight was dangerous. That he was going to get people killed. They asked Congress to shut him down.
These were the men Musk grew up worshipping. The posters on his wall. The reason he built rockets in the first place. And they went on television and said he was a disgrace to space exploration.
In a 60 Minutes interview shortly after, Musk was asked about it. He started speaking and his voice broke. His eyes filled. He couldn't finish the sentence. The richest man in tech, the guy who argues with regulators and fires engineers mid-meeting, sat on camera and cried because his heroes rejected him.
He didn't stop building. He didn't change direction. He didn't even respond to them publicly. He just kept launching rockets until the rockets proved him right.
Armstrong never lived to see SpaceX land a booster. Cernan never saw Starship. The men who said it couldn't be done died before the man they doubted did it.
Most people need approval from the people they admire before they act. Musk got the opposite of approval and acted anyway. That's the gap. Not talent. Not money. The willingness to keep building while the people you love most tell you to stop.
Rajesh K. Gupta retweeted
UC San Diego Halıcıoğlu School of Data Science and Computing is now welcoming applications from New York and Delaware residents for the Master of Data Science online.
📅 Application Deadline: June 17, 2026
📝 Apply today: connect.grad.ucsd.edu/apply
📩 Questions: [email protected]
Rajesh K. Gupta retweeted
PREPRINT REVIEW
—————————
45 domain scientists spent 469 hours rating 2,960 individual criticisms from peer reviews of 82 Nature-portfolio papers, mostly from Nature Communications. The reviews came from both human reviewers and three frontier AI models deployed as tool-equipped reviewing agents: GPT-5.2, Gemini 3.0 Pro, and Claude Opus 4.5.
Averaged across papers, GPT-5.2 had a higher rate of fully positive criticisms, defined as correct, significant, and sufficiently evidenced, than the top-rated human reviewer baseline. 60.0% versus 48.2%. P = 0.009. In holistic expert judgments, it matched or exceeded the top-rated human on about half of papers. All three AI reviewers exceeded the lowest-rated human across every dimension.
The tradeoff matters: AI reviewers were less factually correct than the top-rated humans (86.2% vs 92.3%). The win was on significance and evidence quality, not accuracy. Their accurate criticisms were more often rated as significant and well-evidenced, and surfaced a distinct 26% of issues that no human reviewer raised.
The honest limitations: AI reviewers overlap with each other far more than human reviewers do, 21% versus 3% for cross-reviewer pairs. They exhibit 16 recurring weaknesses including limited subfield knowledge, poor long-context management across multiple files, and a tendency to be overly critical on minor issues. The authors position current AI reviewers as complements, not replacements.
The most actionable finding is not "replace peer reviewers." It is "add one AI reviewer to a human panel." In the paper's simulation, a 2-human-plus-1-AI panel preserved the amount of useful unique feedback while reducing reviewer noise.
This connects directly to the Rodman conversation from earlier this week about evidence infrastructure. Peer review is one of the primary bottlenecks in the scientific evidence pipeline. If AI can surface a quarter of issues that human reviewers miss while producing higher-significance critiques on the issues they share, the implications for the speed and rigor of scientific publication are significant.
Posted to arXiv this week. Preprint. But the signal is strong enough to take seriously.
Rajesh K. Gupta retweeted
Yann LeCun says LLMs are strongest in domains where language itself is the substrate of reasoning, like math and code
They can solve problems, prove theorems, and write programs — but they are not creative mathematicians, software architects, or computer scientists
"their role is to help humans build"
Rajesh K. Gupta retweeted
I’ve always believed the No.1 application of AI should be to improve human health.
That work started with AlphaFold, and now at @IsomorphicLabs with the mission to reimagine drug discovery and one day solve all disease!
We are turbocharging that goal with $2.1B in new funding.
Rajesh K. Gupta retweeted
Some North Carolina State University graduates received an unexpected gift.
Donor Anil Kochhar revealed that he and his wife, Marilyn, will cover all student loans taken out by Wilson College of Textiles graduates in the 2025–26 school year.
Rajesh K. Gupta retweeted
“A university that asks students what they want and gives it to them is not educating – it is catering.”
In my annual MBZUAI commencement address today, I emphasized two principles that shape our university: 1) The core business of university is creating knowledge and teaching knowledge, everything else is secondary; 2) Our university is a forge. Faculty and leaders are not pastoral counselors; they are forge masters. Their job is to make people capable, not comfortable. — Maybe obvious, but hard to say and hard to enact these days.
mbzuai.ac.ae/news/presidents…
Rajesh K. Gupta retweeted
Terence Tao is answering a fundamental question regarding the safety and reliability of modern AI: "How can we use a tool that is powerful, but unreliable?"
W = ∑(wᵢ ⋅ xᵢ) + b
AI isn’t just about “smart”; it’s about the probability of *looking* right. We’ve built systems where the weights (wᵢ) are optimized for plausibility, not veracity.
This creates a “convincing mirror” that confidently serves dangerous advice in medicine or finance. The gap between “convincing” and “correct” is the most critical variable we need to solve for.
So heartwarming to see a strong MLSys group emerge at UC San Diego!
Breaking LLM inference’s autoregressive bottleneck 🛠️
We've teamed up with @haozhangml, @YimingBob, and @aaronzhfeng, among others from UCSD to achieve a massive 3.13X speedup for LLM inference on Google Cloud TPUs using Diffusion-Style Speculative Decoding (DFlash).
Read the blog: goo.gle/4naZ8Yv
Rajesh K. Gupta retweeted
Super happy to see my UCSD colleague help port our DFlash to TPU. Big speedup too!! More to come.
Breaking LLM inference’s autoregressive bottleneck 🛠️
We've teamed up with @haozhangml, @YimingBob, and @aaronzhfeng, among others from UCSD to achieve a massive 3.13X speedup for LLM inference on Google Cloud TPUs using Diffusion-Style Speculative Decoding (DFlash).
Read the blog: goo.gle/4naZ8Yv
Rajesh K. Gupta retweeted
Ilya Sutskever just told the AI industry why scaling is finished.
One word built it. One word is about to break it.
Sutskever: “Scaling is just one word, but it’s such a powerful word because it informs people what to do.”
For five years, that single word replaced an entire research culture.
Nobody needed breakthroughs. They needed bigger checks.
Sutskever: “If you mix some compute with some data into a neural net of a certain size, you will get results, and you will know that it will be better if you just scale the recipe up.”
That’s not science. That’s a recipe.
Sutskever: “Companies love this because it gives you a very low risk way of investing your resources.”
The most transformative technology in human history ran on the same logic used to franchise a restaurant chain.
More locations. More ingredients. Same recipe. Predictable returns.
You didn’t need researchers who could see around corners.
You needed accountants who could approve purchase orders.
But recipes expire.
Sutskever: “At some point though, pre-training will run out of data. The data is very clearly finite.”
Five years of infrastructure. Five years of hiring. Five years of investor decks.
All built on top of something temporary.
Sutskever: “I don’t think that’s true.”
The co-founder of OpenAI. The mind behind the breakthroughs that made this entire era possible.
Saying more money won’t solve it.
Sutskever: “In some sense we are back to the age of research.”
Most of the companies racing to build AGI were never research companies.
They were scaling companies.
They hired for execution. Not discovery. They optimized for throughput. Not insight.
The talent pipelines. The investor pitches. The board decks.
All built around one assumption.
That the recipe would never expire.
It’s expiring.
And the companies that spent five years perfecting the art of spending money are about to discover something.
The next era demands what capital can’t purchase.
An original idea.