Wallpeflower

Part-time scroller, full-time thinker — a curious mind riding the X wave, absorbing everything from tech trends to brain bends. Equal parts silicon and soul

[Probability and Statistics @probnstat]

Neena Gupta is one of the most celebrated mathematicians of modern India, whose work reshaped algebraic geometry. Born in Kolkata in 1984, she showed extraordinary mathematical talent from an early age and pursued her studies at the Indian Statistical Institute, where she later became a leading researcher.

Her most famous achievement came in 2014, when she solved the long-standing Zariski Cancellation Problem in positive characteristic. The conjecture asked whether, if X × A¹ ≅ Aⁿ⁺¹, it must follow that X ≅ Aⁿ. For decades, mathematicians believed this “cancellation” property always held. Gupta constructed a striking counterexample in characteristic p > 0, proving that the conjecture fails in general. This was a landmark breakthrough in affine algebraic geometry and had deep implications for understanding polynomial rings, automorphisms, and geometric structures.

Her work established her as a global leader in algebra and earned her numerous top honors, including the Ramanujan Prize in 2021, making her the third woman ever to receive it. She has also received the Shanti Swarup Bhatnagar Prize and the Infosys Prize. Neena Gupta’s journey is a remarkable example of how deep abstract mathematics can overturn decades of accepted belief and redefine an entire field.

[Mohandas Pai @TVMohandasPai]

Please correct your statement. Open AI was a not for profit when Infy gave them money nearly 10 years ago. It was a grant not an investment. Many others gave them grants to build AI for public use. Musk too gave money and later took them to court for converting a not for profit to a profit enterprise using not for profit money for their R&D. So then nobody globally knew this would be so big. So to credit him for being a visionary of sorts is silly. He left for different reasons. So please do not make misleading statements to malign others. He is a great techie but to say he is an AI visionary, look at his own start up and decide in this new world where is that!

[NIK @ns123abc]

🚨BREAKING: TRUMP ON ANTHROPIC

REPORTER: Do you view Anthropic and to a degree its CEO, Dario Amodei, as a threat to national security?

TRUMP: "Well, not now, but a week ago, maybe. I was with him yesterday, he made a speech... seemed like a nice guy, smart guy. But he responded to us very quickly, because you know it's a tremendous liability… he responded very responsibly, I thought."

REPORTER: Would you consider using the Defense Production Act to possibly regulate or control AI?

TRUMP: "I would, but, I'm not sure I have to do that. I think so far it's been very responsible. Actually, it was a competitor, and a part-owner, that turned Anthropic in…"

[Gappy (Giuseppe Paleologo) @__paleologo]

I expected this interview of Dario to be a collection of softballs, to be listened to while doing work. But this passage on OpenAI is rap-battle level of bluntness.

"But you know, people alone is not sufficient to leave. People here have had disagreements with me. People here [Anthropic] have disagreements with each other. But when you feel that you can't trust someone, when you feel that their values are not what they say they are, when you feel that they're not honest, when you feel that they're not in it for the reasons that they say, when you see disturbing patterns of behavior, dishonesty, that makes it very hard to continue to work with a company, to continue to trust the company."

[ranjan_raj @ranjan_vittal]

Today, I'm thrilled to announce Pramaana's $27M seed, led by @khoslaventures.
The foundational domains that hold the world together: tax, law, finance, healthcare; all run on certainty. Probabilistic AI can't give them that. We’ve been asked to accept wrong answers with AI as ‘hallucinations’, while in traditional software terms, it’s just a bug. And a wrong answer in such mission-critical domains is more than just a bug, it's a liability that could have catastrophic impact.

We built Pramaana to deliver a 100% trustable experience to the domains that run on certainty: AI that is provably correct, not probabilistically correct. We turn statute and regulation into machine-verifiable code, so every output ships with mathematical proof of correctness. Our mission is to make AI take ownership of it’s work.
Pramaana in Sanskrit stands for “means of valid knowledge”, and we’re going to achieve that by formalizing the world’s knowledge.

[Rahul Mathur @Rahul_J_Mathur]

An IIMA classroom project raised $1m today to offer authentic, affordable & domestically produced Ramen.

Nishank and Akhil observed that roadside vendors in Ahmedabad selling Korean ramen were quite successful. This led to an insight - if a snack is now available as street food, it has deep cultural penetration.

Customer feedback revealed that street ramen felt like ramen flavored masala noodles. The authentic alternatives were cup noodles which felt too processed or imported brands which are quite expensive (₹120 per packet).

This was their “Maggi Moment” - customers needed a semi-prepared packet which felt DIY but didn’t involve too much cooking (adding noodles + water + veggies).

The duo received a ₹14 Lakh grant from IIMA which allowed them to iterate for several months on the flavor profile by running a pop-up on campus. Once they locked this in, they began selling on Amazon & Flipkart - the reception was incredible - they made their first ₹8 Lakhs in sales.

[Dr. Brahma Chellaney @Chellaney]

A future conflict with China will not resemble the 1962 Sino-Indian war. It will unfold simultaneously across land, sea, air, cyber and space domains, with success depending on rapid decision-making, integrated command structures, and the seamless fusion of intelligence, diplomacy and military operations.

Such requirements favor systems that elevate expertise and empower professional military judgment. They do not favor bureaucratic structures designed primarily to constrain military influence. https://t.co/qABjGpN0Cd

[Aviral Bhatnagar @aviralbhat]

4 of my batchmates from Engineering Physics at IIT Bombay are at Anthropic/OpenAI

From a department of 36 people that is ~12%, and I suspect it has a higher AI lab density than CSE which had ~120 people

Deep physics research with PhDs paying off

[shirish @shiri_shh]

Elon Musk: "You cannot ask for two years of salary in escrow and consider yourself a cofounder.

There's got to be some combination of inspiration, perspiration, and risk to be a cofounder."

when Elon recruited Mueller in 2002 from a stable job at TRW, he insisted that two years of his salary be placed in escrow as protection in case SpaceX failed.

Musk agreed to the escrow arrangement, but this directly influenced how Musk viewed him.

[Matt Dancho (Business Science) @mdancho84]

Google’s Time Series Foundational Model.

[dom @ryiacy]

My interpretation of this:

Right now, Anthropic and OpenAI are making a killing by selling enterprise FDE services to F500s, building workflows for them on top of proprietary models, then using the traces and context from this to build RL envs to improve the models.

This is crazy amounts of leverage - instead of buying this data they're getting paid gigantic consulting fees to extract it.

This also goes way beyond typical consulting in scope - organizations are effectively outsourcing key learning curves and domain knowledge to the AI labs.

Despite that, it's so far been worth it for them because the value of skilled FDE is so high and the ROI so fast, and orgs are willing to pay a premium for competent AI implementation.

But in the long run, one of two things happens: either orgs are gonna get hooked on this and end up paying for the model training that replaces their business, or they find a way to build and own their own model ecosystem.

What that looks like is developing some combination of AI models, evals, RL envs, and workflows. Initially probably the model will still be an off-the-shelf frontier model from a top lab.

But as firms build out more sophisticated eval / RL env (increasingly the same thing) infra, it starts to become viable to post-train an custom model on top of an OSS base. Cursor have done this successfully with their Composer model RL'd on top of Kimi.

Sidenote, this is the same conversation that a lot of national governments in Europe are having in the past week. When we look at what the rhetoric about 'sovereign AI' in the UK actually boils down to, it's doing custom post-training on top of an OSS model, and then running it on local GPUs.

Ultimately, the current feeding frenzy for AI services in all of its guises - FDE, AI consulting, etc - should raise questions about long-term sustainability. If consulting services are truly a value add and competitive advantage, then in the long term you want to in-house.

[Chamath Palihapitiya @chamath]

The CEOs of the Mags have each been in the seat for decades. Love them or hate them, they have had tons of reps dealing with every conceivable situation - and have had do it under withering public scrutiny. As a result, what you see is what you get. And what you get is that they are all in-band, predictable actors.

Then there’s the emerging case of Dario. What you see is also what you get, but what you get is different.

To his credit, this is only his 5th year as CEO. It’s incredible to see what he’s built. He’s running, arguably, America’s most important company and perhaps the world’s. It’s a trillion dollar behemoth that will only grow bigger if the revenues hold.

But whenever there is an Anthropic dust up, it can roughly be summarized as “someone is lying or misleading you - figure out who but it’s not me”because the setup is always such.

This last episode with the government has been boiled down to that - a case of he said, she said. The problem is that this the second such incident of Anthropic vs US Government just in the last 90 days.

So what happens over the next 18mo when Anthropic develops their next super-brain model after Mythos/Fable?

If I had to bet, it seems Dario is convinced that he’s being truthful and others including the US Government are not. So why even try to cooperate with the US Government versus just work around them - especially when you have a super tool/weapon on your side.

This, to me, is a concerning setup.

[Vinod Khosla @vkhosla]

Great idea especially if you consider Prism ML https://t.co/k0Csm2Lf8q and https://t.co/zsg1PmS3nP as a way to get "concentrated intelligence": very high, almost as good, performance as frontier models on phone hardware! I suspect a 50b-100b parameter model will run on an iPhone this year!

[Sandeep Kulkarni @moneyworks4u_fa]

Good to see such Collab between two fund houses.
thanks Sahil for bring it together on netra's 5th anniversary.
Thanks Harish for distilling your decadal learnings into a single page.

[Paul Graham @paulg]

How to Earn a Billion Dollars: https://t.co/WeWBUkKym6

[DAIR.AI @dair_ai]

The Top AI Papers of the Week (June 7 - June 14)

- Agentopia
- Self-Harness
- Agents' Last Exam
- MiniMax Sparse Attention
- Lookahead Sparse Attention
- How AI Agents Reshape Knowledge Work
- The Geometry of On-Policy Distillation

Read on for more:

[Priyanshu Priyank @PriyanshuP1405]

Insider doesn't need quant to be a high sharpe strat

[Sar Haribhakti @sarthakgh]

.@scottbelsky: "We may look back at this early period of artificial intelligence and realize that by using frontier models in the cloud, we were essentially hiring PhDs for every task just because we could. Fast emerging efficient AI practices will help us avoid malpractice-by-overspecification - you don’t send a cardiac surgeon to take a patient’s blood pressure — not because the surgeon would fail, but because the surgeon is slower, scarcer, more expensive, and the nurse’s reading is indistinguishable. It is far more prudent to allocate the right talent for the right job."

[Christopher Clary @clary_co]

"India's quest for sovereign capabilities is a race against time owing to its rivalry with China. But despite improved investment & regulatory conditions, there has not yet been any significant progress towards the development of a robust & competitive space economy," argues Tobias Scholz.

[Sameer Joshi @joe_sameer]

#Saturation warfare comes down to 5 numbers: fire channels, magazine depth, leakers, cost-per-effect, and time. All 5 are now moving in one direction & #IntelligentMass is the reason. My new piece on why 2026 is the year the race tips. #ComethTheSwarm

https://t.co/VA6uyPm0RD