AI Unleashed: "My Takeaways from India's AI Powerhouse Event"

A few weeks ago, I had the chance to attend one of India’s largest AI conferences, organized by Nvidia, courtesy of my company Sify, which was one of the key event sponsors and participants.The star attraction? None other than Jensen Huang, CEO of Nvidia, the visionary leader behind the $2 trillion company powering today’s artificial intelligence revolution. To add to the excitement, the event featured surprise appearances from Indian icons Mukesh Ambani, Chairman and Managing Director of Reliance Industries, and Bollywood’s "sweet and simple" (Hum Hain Seedhe Sadhe Akshay) Akshay Kumar.

While you can catch the 1.5-hour replay of Jensen's address, his interaction with Akshay Kumar, and a fireside chat with Mr. Ambani, I wanted to share some intriguing insights I gathered firsthand from their discussions at the Jio Convention Centre. Here’s my take:

• The End of Moore’s Law Jensen pointed out that Moore’s Law—the idea that computing power doubles without software changes—is no longer sustainable, leading to what he called "computing inflation."
• Exponential Rise in Computational Demand With data and model sizes doubling each year, computational needs are increasing fourfold annually. While Moore’s Law led to a 100x performance increase over 10 years, AI's current pace requires a 4x increase every year—a staggering 10,000x leap over a decade. This rapid scaling is essential for AI models to keep growing smarter as data and model sizes expand.
• The Shift from Software 1.0 to Software 2.0 We're moving from traditional programming ("Software 1.0") to a new paradigm where machine learning ("Software 2.0") drives the software. Here, AI analyses vast datasets to learn patterns, enabling computers to predict outcomes without explicit programming.
• A 'Universal Translator' of Data We’ve all heard about the Data being the new oil but the concept is becoming even more critical. Jensen’s "scaling law" illustrates how AI models need vast amounts of data to improve in accuracy and effectiveness, making data an invaluable resource.
• A Second Scaling Law: Intelligence Through 'Thinking' Beyond training, AI is evolving to improve inference through a process akin to human "thinking." While models like ChatGPT offer instant responses, future AI may simulate reasoning, planning, or reflection, producing higher-quality outputs.
• Inference Time Scaling Law According to the second scaling law, the longer an AI model "thinks" during inference, the better its response quality. This mirrors human reasoning—while simple questions might require a quick answer, complex problems benefit from more in-depth thinking. By allocating more computational time for inference, AI can deliver better solutions, especially for tasks involving complex planning.
• AI Won't Replace Jobs—But People Who Use AI Will Win While AI may handle 20-50% of tasks with incredible efficiency, it can’t fully replace human creativity and judgment. Jensen believes that AI won’t take over jobs entirely; instead, those who leverage AI to boost productivity will thrive.
• India Should Build Its Own AI Inspired by Prime Minister Modi’s vision, it was emphasised that India should develop AI domestically. Jensen quoted the Prime Minister, saying, "India should not export data to import intelligence, just as India should not export flour to import bread." It's crucial for Indian enterprises to harness local data to build their own intelligence, ensuring both sovereignty and privacy. - At Sify, we’re supporting this vision by building AI-enabled data centres and providing AIML platforms to empower enterprises in achieving self-reliant AI capabilities.
• Open Source is a Game Changer Both industry stalwarts and Mr. Ambani recognized the transformative power of open-source software, with a nod to Mark Zuckerberg for bringing open-source AI models like "Llama" into the spotlight. Open source has democratised technology, with millions worldwide benefiting from it. Hearing this was a proud moment, as it reminded me of my days at Red Hat nearly two decades ago, when my colleagues and I worked tirelessly to introduce enterprises to open-source software. Often, it’s hard to gauge the impact of your work in the moment, but knowing legends acknowledge it feels incredibly rewarding.

In essence, the message from the conference was clear: AI and ML are catalysing one of the biggest industrial revolutions of our time. Now is the moment for businesses to harness AI, paving the way for a better world and a brighter future for society.