Ground Truth Ep. 2 | Prof. B. Ravindran, IIT Madras
Ground Truth by Vizuara, Episode 2: Prof. B. Ravindran Prof. B. Ravindran, Head of the Wadhwani School of Data Science & AI and the Centre for Responsible AI (CeRAI) at IIT Madras and a member of the UN's scientific panel on AI, makes the case for a "frugal AI stack" built for India's real problems. He also explains why he thinks the biggest danger from AI is humans rushing to market before it is ready, and why kids learning from AI worries him. Hosted by Vizuara co-founders Dr. Rajat Dandekar, Dr. Raj Dandekar and Dr. Sreedath Panat. We also cover his journey from five students in his first course to 180, why ChatGPT surprised even AI researchers, what counts as research when AI agents can do the work, why India has no DeepSeek yet, and what the UN's Independent International Scientific Panel on AI actually does. CHAPTERS 0:00 Cold open 1:15 Introductions 4:01 From neural networks to RL 7:14 Surviving the AI winter 8:32 From 5 students to 180 10:30 Deep learning reaches campus 13:14 The ChatGPT shock 18:05 RLHF, AlphaGo & real-world RL 22:14 Revisiting the Bitter Lesson 27:11 How do babies learn so fast? 30:08 Are LLMs the right path? 31:38 Living with the frontier 38:35 Teaching in the age of AI 41:22 What should children learn? 43:31 Outsourcing our thinking 45:08 What is research now? 50:28 AI as a reviewer, not a writer 52:50 Can agents replace PhD students? 54:31 Why he built CeRAI 59:00 India's frugal AI stack 1:06:36 Where is India's DeepSeek? 1:13:14 The real danger of AI 1:21:38 Inside the UN's AI science panel 1:25:19 Closing thoughts LINKS UN Independent International Scientific Panel on AI (preliminary report and briefs): https://www.un.org/independent-international-scientific-panel-ai/ Prof. B. Ravindran: https://wsai.iitm.ac.in/~ravi/ Centre for Responsible AI, IIT Madras: https://cerai.iitm.ac.in/ Ground Truth by Vizuara: https://ground-truth.vizuara.ai Ground Truth by Vizuara: honest conversations with the people shaping AI in India. Making sense of the noise.