Coursera's Engineer No 1 on Building AI agents for knowledge workers #AI #podcast #aicode #startup
Jiquan Ngiam, Coursera's first engineer and a Google Brain alum, joins The Startup Project to discuss the future of AI. Discover his journey from scaling deep learning with Andrew Ng to building AI agents for knowledge workers at his new startup, Lutra AI.
5 Things You'll Learn from This Episode
- Learn why data, compute, and scale proved more critical than architecture in AI's advancement, as exemplified by AlexNet's 2012 breakthrough.
- Understand the early vision for Coursera to democratize access to cutting-edge education like machine learning for a global audience.
- Hear the contrarian take that AI progress is on a sigmoid curve, approaching diminishing returns due to data limitations.
- Discover how the future of AI may involve training massive models to generate data for smaller, more efficient models for deployment.
- Explore the concept of building AI agents that separate data from logic for more reliable and scalable automation of knowledge work.
About the Episode
Jiquan Ngiam shares his incredible journey through the rise of modern AI, from his early days working with Andrew Ng at Stanford and building Coursera to his time at Google Brain during the emergence of transformers. He offers a unique perspective on the limits of AI scaling, the future of AI agents for knowledge workers, and how his new company, Lutra AI, is building tools to augment human productivity.
Timestamps
- 0:00 — Introduction: Coursera's First Engineer
- 2:28 — Working with AI Pioneer Andrew Ng
- 6:15 — The AlexNet Breakthrough: Data, Compute, and Scale
- 12:25 — The Mission to Democratize Education at Coursera
- 17:55 — Google Brain and the Rise of Transformers
- 21:24 — The Sigmoid Curve: Are We Hitting the Limits of AI Scaling?
- 28:13 — The Need for Physical Task Data in Robotics
- 34:23 — Introducing Lutra AI: Agents for Knowledge Work
- 42:49 — Different Approaches to Building AI Agents
- 45:38 — Choosing the Right Models for the Right Tasks
- 52:04 — AI Code Generation: Cursor vs. GitHub Copilot
- 58:30 — Debating AGI and the Future of AI
About the Host
Nataraj Sindam is the creator of The Startup Project, a podcast featuring founders, investors, and operators building the future.
→ Website
Subscribe to Startup Project for more founder conversations!
→ Spotify
#StartupProject #AI #AIAgents #MachineLearning #DeepLearning #Coursera #GoogleBrain #AndrewNg #LutraAI #KnowledgeWork #FutureOfWork #TechPodcast