Perplexity AI’s Aravind Srinivas offers Zerodha’s Nikhil Kamath a bold solution for India’s AI problem: ‘Stop aping the West’

Zerodha co-founder Nikhil Kamath is unstoppable. In the latest edition of his podcast show The WTF is?, Kamath sat down with Perplexity AI co-founder, Aravind Srinivas, to explore the future of artificial intelligence (AI), its impact on industries, and India's potential in this transformative space.

The conversation, rich with insights, touched upon personal journeys, the evolution of AI, and the broader implications of technology on society.

The Journey from Chennai to AI Leadership

With Kamath asking Srinivas about his own journey, the latter started the podcast by tracing his roots to Chennai and Tamil Nadu, and from there to becoming a leader in AI, all of which reflected the importance of perseverance and continuous learning.

“Growing up in Chennai, there was always an emphasis on excelling academically,” he noted, highlighting the cultural pressure that shaped his early aspirations. His interest in machine learning was piqued through participation in a Kaggle competition, which became a pivotal moment in his career.

“That competition was my gateway into AI. It wasn't just about winning; it was about understanding how machines can learn from data.”

Kaggle hosts numerous competitions sponsored by Google, providing opportunities for data scientists and machine learning practitioners to showcase their skills. These competitions often involve complex datasets and real-world problems, attracting participants to develop and test advanced algorithms. Google's involvement typically includes providing datasets, computational resources, and often, substantial prizes.

Internships at prestigious organizations like OpenAI followed and these provided invaluable exposure to cutting-edge research. “Being in an environment where everyone around you is pushing boundaries forces you to think beyond conventional limits,” Srinivas shared. This experience shaped his approach to AI, reinforcing the need for rigorous experimentation and innovation.

The Complexity of Achieving General Intelligence

A recurring theme in the discussion was the challenge of developing general intelligence in AI. “Right now, AI is great at specific tasks—chess, coding, even writing,” Kamath observed. “But true intelligence means adapting across multiple domains without explicit programming.”

Srinivas expanded on this, explaining that while AI systems minimize loss functions to improve accuracy, defining these functions for general intelligence remains elusive. “One of the biggest hurdles is enabling AI to learn its own loss functions. That’s where real breakthroughs will happen.”

The conversation also tackled the fundamental question: what constitutes intelligence? “If an AI can perform human-like tasks better than humans, does that make it intelligent?“ Kamath asked. Srinivas responded, “Not necessarily. Intelligence isn’t just about performance; it’s about adaptability and understanding.“

The Economic and Societal Impact of AI

The rise of AI has significant economic implications. Kamath pointed out the potential displacement of jobs, particularly in areas like coding and content creation. “AI is already writing articles, generating code—what happens when it starts doing these tasks better than humans?“

Srinivas acknowledged these concerns but emphasized that AI would also create new opportunities. “The personal computer revolution didn’t eliminate jobs; it changed them. AI will do the same. The key is adaptation.“

The discussion also touched upon India's role in the AI revolution.

“India has an incredible talent pool,“ Kamath said. “But how do we leverage it?“

Srinivas argued that India must invest in building its own AI infrastructure, particularly in training models suited to local languages and contexts. “We can’t just rely on Western models. We need AI that understands India—not just English but Hindi, Tamil, Bengali, and other languages.“

India’s Position in the AI Race

India’s potential in AI is vast, but there are challenges, according to Srinivas.

“We need better data centres, stronger computing power, and local AI solutions,“ Srinivas emphasized. He pointed out that companies like Core Vivve are already making strides in this space. “India has the capability, but we need more investment and strategic focus.“

The dominance of tech giants like Google and Meta presents another challenge. “Breaking into the AI market when global giants control the infrastructure is tough,“ Kamath noted. “But India has the advantage of scale. If we build AI solutions tailored to our market, we can compete globally.“

Srinivas highlighted the potential for AI in personalized advertising, noting that platforms like Instagram have excelled by leveraging AI for targeted engagement. “If Indian companies can innovate in this space, there’s a huge opportunity to disrupt the market.“

The Future of AI and Its Ethical Considerations

As AI continues to evolve, ethical considerations become increasingly important. “Who owns the data that AI is trained on?“ Kamath asked. “Should countries be compensated for their data being used in AI models?”

Srinivas acknowledged these concerns, predicting a shift toward stricter regulations and data ownership models. “We might see countries charging for data usage, similar to how media companies are putting content behind paywalls.”

Another pressing issue is AI’s role in shaping public opinion. “AI-generated content is becoming indistinguishable from human-created content," Kamath said. "What does that mean for misinformation?”

Srinivas called for more transparency in AI-generated media. “Users need to know when they’re interacting with AI. We need mechanisms to verify content authenticity.”

The Road Ahead

As AI development accelerates, the need for innovation and adaptability has never been greater. Kamath and Srinivas agreed that the key to success in AI lies in continuous learning and staying ahead of trends. “AI is moving at an exponential pace,” Srinivas said. “If you’re not learning, you’re falling behind.”

For young entrepreneurs looking to enter the AI space, Srinivas had one piece of advice: “Start small but think big. Use open-source models, experiment, and focus on solving real-world problems. The opportunities are endless.”