Ai goes way beyond the American rhetoric
Money is necessary, not sufficient
Two startups are way ahead of the curve.
The Americans lead in ai. No two ways about it. They made consumer ai happen. And they did so with the brute force of money. Just like android and apple own smartphones, they now own ai. It was carefully and brilliantly orchestrated strategy – and has the potential to enslaven the world – after handcuffing most of humanity to their phones. It is not the internet but tethering humanity to the smartphone that changed the world. Smartphones empowered the poor. Ai will ruin them.
Bad news has just begun coming in. The worst is yet to come. Disparities. Civil unrest. Mental health issues. Health, wealth, relationships … are going to change irreversibly. More the sugar, sweeter the dish.
But what about diabetes.
Big tech would like us to believe that it is all about money. Because money is all they have. And money can buy most things. Except happiness. Except destiny. Except ai leadership.
The quest for monopolisation through predatory pricing has worked in the Internet age. But will it work in the erupting ai age. Two startups – both by Indians – one in India and one overseas seem set to challenge the core paradigm of money being a sufficient condition for success.
These startups are Unconventional Ai in USA and Kodoy in India.
The largest seed round in AI history just happened. $475M raised at a $4.5B valuation with NO product. That’s Unconventional AI, founded, and this is now the biggest seed round ever. Kodoy in contrast is jugaad innovation. With potentially greater impact.
Unconventional ai is about neuromorphic computing, machines designed to compute more like biology and consume radically less energy. They try to mimic the human brain. And can work without network connectivity. Just like a human can. Training frontier models like GPT-4 reportedly consumed ~50 GWh of energy. Neuromorphic systems aim to operate at under 1% of that.
Kodoy is about middleware. Something that is a big gap between Nvidia at one end and the LLMs like ChatGpt on the other. It is a no man’s land as of now.
India hit upon this gap purely by accident.
It started with making 3D maps. You need these to fly drones (or air taxis). Spatial awareness needs maps. India is the map maker of the world. Om Hemrajani and Genesis International are the worlds digital twin makers. Google earth, Apple Maps, Ayodhya masterplan ….. genesis has done them all.
Second it got into the challenge of running deepmind on fasta files generated from DNA sequencers. 155 GB per person. One million target to be covered. No existing ai works. Third they tried to model the human body as a machine in 3d. And then track this machine in real time. Again non trivial. Fourth India has 35,000 GPUs but they are spread across 300+ locations. How do you make these work as one data centre.
Inadvertently they reinvented the data centre as we know it.
