Thine already solves this and A LOT more, all the while being just an app on your iPhone and Apple Watch. People are already switching out.

We are also pleased to announce that our paper got accepted at IEEE SLT 2026. Everything has been made possible by the incredible efforts by @shancn7 and Aditya. More details in our blog and arxiv preprint thine.com/research/who-said-…

we've built @ThineAI so that you shouldn’t have to choose between being present and remembering what happened

$200 billion, that’s how much is planned to be invested in new data centres across India by 2030, and none of them will be owned by India. Everytime a new Indian AI model gets launched, we call it ‘sovereign AI’, talk about its benchmarks, and then move on to the next one. But there’s still no comparison with all the big US models. And they know this, but Indians don’t. So, while we’re spending all our time building new LLMs, the US has already started building new data centres in India. OpenAI is building a data centre in India with enough computing power to use at least 1 gigawatt of electricity. Microsoft has committed $17+ billion, Google is investing $15 billion in an AI hub in Visakhapatnam, AWS around $12.7 billion, and Adani is also planning to invest up to $100 billion but have barely starting to come together while being partially owned by US based EdgeConneX. All these companies are building data centres in India, which will use our land, electricity and huge amounts of water. But we don’t own and control any of it and if we’re going to bear the environmental and resource costs anyway, shouldn’t India at least own more of the infrastructure and get more of the value from it? Otherwise, the resources stay here, the impact stays here, but the control and much of the value can end up somewhere else. History has shown us what happens when that control stays with someone else. Iran had huge oil reserves, but much of its oil industry was controlled by foreign companies, and it had to depend on them for extraction and production. In the 1950s, Iran even fought to take greater control of its own oil industry after years of British influence over it. Africa is one of the richest continents in the world when it comes to minerals, yet many of its countries remain among the poorest because they don’t have the technology, machines and infrastructure to extract and process those minerals themselves. The resources are there, but the ability to turn them into value is often controlled by others. Now, Imagine Anthropic decides not to give the Indian government access to its best model. There isn’t much India can do if it depends completely on them. But what if Anthropic needs huge amounts of compute from data centres owned by Indian companies? Then they would need something from us too. That’s where ownership gives India bargaining power. It’s the difference between simply using AI infrastructure and actually having some control over how AI develops in India. I still think it’s not too late to realize that no Indian LLM is going to beat the best US models right now. So instead of trying to win a race we have already lost, India should ask a different question: what does every AI company need, regardless of who builds the best model? The answer is compute. Let OpenAI, Anthropic, Google and everyone else fight over the models. India should focus on owning more of the infrastructure they all need to run. Because the companies and models will change, but the demand for compute will keep growing. @dylan522p @ShwetaRKohli @narendrasen @abhish18 @siddarthpaim would love to get your views on this

Nikita BYEr