As the world becomes increasingly digitized, the development of artificial intelligence (AI) technology has become a global race.
In the last 2 years, it almost felt that the center of gravity for tech start to shift from the United States to India / China and Asia in general, which have thriving startup ecosystems and are way advanced in terms of Fintech tech (like UPI), AI innovation (China), Web 3 + crypto (Dubai).
When China scared Eric Schmidt
China was leading the AI race for a long time – to a point that Eric Schmidt warned the US govt on this.
Schmidt wants the government to implement his sweeping blueprint to fight what he considers an existential threat to democracy posed by China’s AI plans, an effort that could also bolster his own commercial AI interests.
Without some type of unified, broad adoption of an AI foundation for the entire department, DOD will soon reach a tipping point after which it will be unable to catch up to its competitors,” Schmidt said in 2018 while testifying before the House Armed Services Committee. Schmidt spoke in a personal capacity, but was chair of the Pentagon’s Defense Innovation Board and on Alphabet’s board at the time.
He says the U.S.’s national security and economic leadership are dependent upon spending billions to procure smarter software, bolster AI research, and build the country’s computer science talent pool.
From a 2021 article / Via
But then, OpenAI came and democratized the entire AI landscape!
You own LLMs. You own the (AI) world.
LLMs are critical components in natural language processing (NLP) and form the backbone of AI-powered products like virtual assistants, chatbots, and translation services.
These models require vast amounts of data, computing power, expertise (AND PATIENCE) to develop, and the current owners of LLMs are primarily US-based companies like Google, Microsoft, and OpenAI.
This dominance means that the rest of the world will be primarily consumers of these technologies, rather than creators.
The absence of access to LLMs and NLP technology could mean that opportunities to create globally dominant tech giants like Google or Microsoft may be lost (?) for India and Asia.
Can an Indian startup/company create LLMs?
As with anything entrepreneurial, the answer is ‘yes. why not!’
Practically speaking, building LLMs not just needs vast talent resource and $$s, but a whole lot of patience.
The Indian investing/VC ecosystem has money, but no patience – which is why we don’t have pure blood product startups. For a community obsessed with traction, imagine if they had invested in OpenAI way back in 2015? The founders would have been fired by 2021 – or forced to launch a SAAS business focused on providing AI tools to enterprises.
Excerpt from another essay: What comes first: Product, Soul or Traction?
There are many such products that go slow (rather, snail-speed slow) in the early days, get the soul of the product out (the key differentiation) and then scale (exponentially).
Name 5 global consumer product companies from India.
Because that differentiation takes time to build. It eventually serves as a moat and provides a unique leverage for the company to make others irrelevant (without competing)
What’s the way out?
Of course, you can AI technologies that do not rely on LLMs and NLP technology but those may not create mass scale change (at the rapid speed of LLMs adoption).
The Indian startup ecosystem still has time to redeem by focusing on building product-led startups, irrespective of whether the VC community catches up or not.
I am definitely doing that (at scale with our products). What about you?
