Ai leadership is a great laudable goal.
But where does India start.
We seem to have got it all wrong for now.
In ai, India has just two elements of comparative advantage. First is the massive pool of data available. Second is the large tech literate work force. In the ai age – the latter is no great advantage. So all that we have as a weapon for gaining comparative advantage – in the ai battle – is data.
And data is the new oil. But not unless it is standardised.
Unlike oil, if you consume data, it doesn’t get destroyed and converted into green house gases. Data is something that you can eat and yet have it. Once your data leaks out, you may not miss it – but yet it would have been gone /stolen. Protection of data is important. But it is also important to standardise it and to store it optimally.
Unless we do these two things – standardise the data and store it safely and optimally – the data related comparative advantage will not get manifest.
There is data everywhere but to be able to consume it, we need to make it safe and potable. Very much like – water water everywhere but not a drop to drink.
It is not just about data but also about networks. The current hierarchical networks of the pre AI age are no longer relevant to the AI age.
The wearable or handset is going to be an edge ai device and the ai data centre will extend to very near the consumption point. All of this needs standards. Open standards allowing free interconnection anywhere.
If we don’t have this, we will get proprietorial networks controlled by cartels. All the billions of dollars of forex leaving the country will be used against the country.
Whilst the enthusiasm of the various funding agencies to first lay their hand on tax payer money and then to disburse it is all “indian and understandable”. Finally attractiveness of bureaucratic jobs is rated based on the purchasing power of those jobs. The great indian ai spend will follow those patterns which is actually unavoidable.
The key question is not how to secure a winning place in the ai race but to figure out the policy and regulatory steps to be taken to help private sector. It is not the job of governments or of colleges to build and operate ai systems, devices or networks. If that capability exists, then it should be used to revive the floundering BSNL and MTNL.
Interestingly both these telcos with thousands of small exchange buildings all over the country are perfectly placed to reinvent themselves as ai networks. And ai networks would make existing telcos like Jio and Airtel irrelevant.
Likewise, getting into a competition to buy over priced nvidia chipsets will do little for India. It is best to use data centres in places like the Middle East where both the cost of capital and energy is far lower.
It only makes sense to enter this race and play this game is if we can build an end to end stack.
And safeguard our data
Before we get into moving towards solutions.
Let us first freeze the direction.
