Why can’t we do both? Innovate as well as be a back office of AI services. We have enough people of varying calibre who can fo both jobs. Not everyone can innovate but they would need jobs and not everyone wants to do boring jobs so they can innovate. For God’s sake we have 140 crore people in this country!
110%!
But not be JUST backoffice, because you can lose that power to lesser parity countries like Vietnam, Philippines
When we create, take the front stage, we can demand front, back, side offices too 💪
From a global macro point of view, We are going to definitely have broken supply chain, we need to ramp up manufacturing. I am sure we can't do it like China did with cheap labour so we need to leap frog to robotic manufacturing.
We have done this in the past, we leapfrogged to service sector from agri without manufacturing, we leapfrogged to mobile first without computers.
We should leapfrog to robotic manufacturing or AI powered services.
The Worlds biggest disadvantage is our strength that is we have working age population.
Hold my coffee...
AI was there from 70y.
But now it's benificial to us due to 2 reasons...
-
Computing power ( My phone has same computing capability i.e. of Apollo mission in 1960s) Basically am talking about Moore's law and size of transistor which is at 2nm right now somewhere in Fabs of Taiwan.
-
Cloud based infrastructure and Data centre
2 is possible due to 1.
- Languages like python, go Lang etc which can interact with server, cloud, device etc almost in real time and hence train the models faster.
Now, sorry to say but we lack at point 1.
We don't have any Fabs available at India which can produce chips at scale.
We can buy but it's very costly and only RIL or Tata can do this. Sarvam ai and krutrim are very small players.
That's why Sam Altman challenged Cxos when Sequoia India head asked this question which is in post.
Still am bullish on India, coz we are good at 3 and 2 is manageable.
We need lean and clean startups which can build in India and consumed in usa and Europe.
It's flabbergasting to see that our top 4 out of 5 daily most used apps are from USA.
We need to reverse this!
Research on known drawbacks of the current state of LLMs, Existing Transformers model enhancements and working on them in a precise manner together along with continuous improvement on language support with cost effective scalability should be enough to level up our game for now. For the same exact reason China has withdrawn its research resources on quantum computing temporarily, and is solely focusing on Llms and AI.