Riju

@VishalBhargava5 If your decision to have or not have kids is based on tax break, it is possibly better for humanity for you to be out of the gene pool.

DINK’s are doing it wrong. Andhra Pradesh is now doing it wrong. Hope other states don’t follow. The incentive should be in the form of tax benefits - like Indonesia does it. This will encourage our best earning population to have more children.

@Trinhnomics Anecdotally Indians never sell gold. We buy gold and keep it in the lockers forever. If anyone is forced to sell their gold, it will be the neighborhood gossip for years. That is the problem. Seeing gold as an heirloom rather than investment.

Gold imports = gold purchases at lower price = higher asset value as gold has appreciated. Indians buy gold because it's a real asset and a good hedge against inflation or depreciation of rupee vs USD and vs gold.

@anishmoonka I thought Oracle got the beating since people figured out that the entire backlog was just 1 single company - OpenAI

Last September, Oracle showed Wall Street a chart that looked just like this one. Their stock had its biggest one-day jump in 33 years. Today it has lost almost half its value. Google just had the same moment. The $460 billion is mostly promises. The figure represents the full value of contracts customers have signed agreeing to pay Google later. Some of that money arrives next month. Some arrives in 2031. Three caveats sit underneath the headline. First, Google's CFO told analysts on the same earnings call that just over half of the $460 billion will turn into actual revenue in the next two years. The rest stretches out over many years. The newest contracts mostly start paying out in 2027 or later. Those are the deals where Google sells its custom AI chips (called TPUs) directly to customers, who then run them inside their own data centers. Second, a huge chunk of the one-quarter doubling traces back to a tiny number of customers. Last week Google promised Anthropic, the maker of the Claude AI assistant, enough computing power for the next five years to roughly match the peak summer electricity use of the entire city of San Francisco. There were earlier deals too: an October 2025 Anthropic agreement worth "tens of billions" for one million AI chips, plus a $10 billion Meta contract from August 2025. For comparison, Oracle's similar-looking $455 billion turned out to be 58% one customer, OpenAI. Google has not published its own breakdown, but with deals this big the math gets concentrated fast. Third, Sundar Pichai admitted on the call that Google is "compute constrained in the near term" and cloud revenue "would have been higher if we were able to meet that demand." Google may not be physically able to build enough data centers fast enough to deliver everything it has already been promised payment for. The bottlenecks are electricity, land, and getting hold of enough chips and parts. Google now plans to spend $180-190 billion this year alone on building more, and the CFO said next year will be even higher. The chart climbs straight up because on paper it does. The same shape on Oracle's chart drove a 36% one-day surge, followed by a roughly 47% crash, once Wall Street started asking who was actually paying, who was actually getting the computing power, and when. Google's customer mix looks more spread out than Oracle's was. The questions are the same.

It's wild to think about how massive 1M token context windows in LLMs really are That's roughly equivalent to: - The complete works of Shakespeare - 11 hours of audio - A 5-minute session fixing some TypeScript issue

It is a tragedy that a state that produced Vivekananda, Netaji, AJC Bose, Tagore, Bankim and Vidyasagar is today thought of merely a source of household help in the rest of India. While there is nothing wrong in working as a maid or driver (all honest labour should be respected), the cultural and economic decline of my home state is not the joke that this gentleman seems to think. Some of us witnessed the collapse over half a century, and find this obnoxious.

Before breakup after breakup💀

@202accepted Well, at this point the nm is just version numbers. 3nm doesn't actually mean anything is 3nm wide. So 1-2nm vs 3nm doesn't really mean anything other than, yeah, TSMC probably has next gen lined up already.

people are missing the point, this means taiwan has 1-2nm already

Princess Irulan is applying Multani Mitti, what's more there to understand?

@braelyn_ai @gruber to opine how he feels about all this :D

do you think the Markdown creator ever thought that we'd be using it for persistent state?

Iran not stopping Chinese LPG because they know what happened when they ate an uncooked Bat

📢 Open-sourcing the Sarvam 30B and 105B models! Trained from scratch with all data, model research and inference optimisation done in-house, these models punch above their weight in most global benchmarks plus excel in Indian languages. Get the weights at Hugging Face and AIKosh. Thanks to the good folks at SGLang for day 0 support, vLLM support coming soon. Links, benchmark scores, examples, and more in our blog - sarvam.ai/blogs/sarvam-3…

@ai_for_success OpenAI is ready to do anything to switch the vibes currently against them

OpenAI thought the best way to describe a new model launch was calling it less cringe than the previous version. Seriously??

@MihirkJha Wait till you hear who was recently cleared to buy/process oil from Venezuela.

Visionary Leadership. I remember how Modi Govt worked on war footing to build these reserves

@sadist_2003 You would not have said this if you knew about the state of RAM prices.

The world is on fire and we are inaugurating plants, lol

We vibe coded this demo in @GoogleAIStudio to show Nano Banana 2's real-world understanding. With each frame in this environment, the model sees only the previous image, and is prompted to imagine what happens next. The consistency is incredible!

Capitalism for me, socialism for me. Example 32343

Lol, such kind of things will only come from people who wants you to fight in their war and never the other way round.

My take on AI Kumbh Mela : They went out of their way to make it possible for people from many walks of life to attend.. #IndiaAIImpactSummit2026 So going by the media accounts, all that really happened at the summit were the long security lines, lost wearables, and other more pressing first world problems like some of us not getting our millet foam at @PMOIndia dinner. Perhaps there is another way to look at it. This is the first of the AI Summits that seemed to have allowed basically free registration to anyone who wanted to come. As far I know, the other ones in UK, Paris and Seoul were limited to mostly invited "delegates". Opening the summit up this way meant LOTS of people showed up (as they apparently do to the other Kumbh Mela..)--estimates of registrations at 250K and daily attendance of as much as 70K. It's in a different plane all together compared to the "invited delegates only meetings", and in a different league even compared to the mega AI conferences like #NeurIPS2025. The bigger numbers also meant longer lines, more chaos and lower signal to noise ratio for the cognoscenti. After my panel, I met one student who brought his mother (who didn't seem particularly tech savvy) to show and tell her about AI.. I talked to a neurologist from the capital region, who showed up just to get a sense of how and whether this technology might atrophy our own cognitive skills. A lady working in Arts and Crafts, who was trying to get a sense of how AI will affect the artists and their livelihoods. I saw lines of women--clad in their finery--waiting for their group buses after a trip to the summit. And I of course saw tons and tons of UG students from Indian colleges attending sessions and trying to make sense of things (however primitive some of their understanding seemed after a minute of talking to them). Maybe some of this was orchestrated. But I would think that if AI is supposed to be such a transformative technology that would impact everyone, perhaps it is quite justified to "let everyone in".. For that inclusiveness, I believe the organizers of this AI Kumbh Mela deserve a huge amount of credit--and our benefit of doubt on the attendant inconveniences. (No, I have no connection with the organizers. I own my opinions.)

Refreshing to see the leader of a nation with so much labour capacity have a progressive view of a technology disruption, while it would have been easy to just take the doomer mindset.

@aakashgupta What is ironic about this entire episode is that the US has been sanctioning Nvidia chips to China sighting this exact same problem - they'll use them for military purposes.

Anthropic is now getting punished by the Pentagon for asking whether Claude was used in the Maduro raid. A senior administration official told Axios the “Department of War” is reevaluating Anthropic’s partnership because the company inquired whether Claude was involved. The Pentagon’s position: if you even ask questions about how we use your software, you’re a liability. Meanwhile, OpenAI, Google, and xAI all signed deals giving the military access to their models with minimal safeguards. Only Claude is deployed on the classified networks used for actual sensitive operations, via Palantir. The company that refused to strip safety guardrails is the only one trusted with the most classified work. Anthropic has a $200 million contract already frozen because they won’t allow autonomous weapons targeting or domestic surveillance. Hegseth said in January he won’t use AI models that “won’t allow you to fight wars.” And this week, the head of Anthropic’s Safeguards Research Team resigned with a warning that “the world is in peril.” So the company most worried about misuse built the only model the military trusts with its most sensitive operations. And now they’re being punished for caring how it was used. The message to every AI lab is clear: build the best model, hand over the keys, and never ask what they did with it.