Narayana Murthy has expressed apprehensions about the lack of problem-solving skills amongst Indians, advising that India should refrain from investing in building its own large language models (LLMs) as the country has yet to build any large databases to store and compile big data.
Billionaire Narayana Murthy has expressed apprehensions about the lack of problem-solving skills amongst Indians, advising that India should refrain from investing in building its own large language models (LLMs) as the country has yet to build any large databases to store and compile big data.
“A large language model (LLM) doesn’t make any sense. We have not been able to build large databases, and without big data, AI has no value. Basically, the Indian mindset is still not oriented towards problem definition and problem-solving,” a report by Moneycontrol quoted the Infosys founder as saying.
Murthy asserted that instead of building its own, India should focus on building solutions atop existing LLMs.
Murthy’s views were supported by Infosys co-founder and Aadhaar architect Nandan Nilekani, and even found backers in the ruling government with Ministry of Electronics and Information Technology (MeitY) secretary S Krishnan advocating for adapting existing LLMs to serve specific sectors. while noting that building a new one from the ground up might not be worth the effort, time, or financially viable for the country.
Those who share Murthy’s view point out that LLMs demand years of development and require substantial capital, noting that tech giants like OpenAI, Meta, Google and others have already made these investments and are now reaping the rewards.
Narayana Murthy also asserted that compared to English, Indian languages are still in their infancy, when it comes to practical use in Technology, Engineering, and Mathematics (STEM) field.
“While it is laudable to work towards building large language models in vernacular, but until the time when we can actually create large databases, I don’t know how well we can use it,” Murthy was quoted as saying while speaking on the sidelines of the 2024 Infosys Science Foundation (ISF) awards.
‘What has India invented?’
Notably, despite Murthy and other technologists arguing against India building LLMs, several Indian startups have reportedly started work on multilingual LLMs.
Companies like Sarvam AI, a Lightspeed-backed startup, which launched its LLM Sarvam 1 last month. Sarvam has touted its LLM as India’s first indigenous multilingual LLM, trained from scratch on domestic AI infrastructure in 10 Indian languages and English.
Bengaluru-based CoRover.ai’s BharatGPT and Seetha Mahalaxmi Healthcare’s Hanooman AI, are among others who have also launched their homegrown LLMs.
However, Narayana Murthy questioned whether these homegrown LLMs are indigenous technology developed by India, maintaining that India is bringing nothing but a services mindset to the LLM technology.
“Which area has (India) invented? Please give me an example,” the Infosys founder asked while pointing out India’s dependence on adapting foreign technologies rather than creating innovations.