The evolution in AI thinking and products of the next decade
summary
Argues that AI moves from infrastructure → automation → data → AI to a new layer of multiagent, connected products and gives concrete targets: product 3.0: connected car ecosystems, smart cities, teaching machines. Claims current systems hit about 80% on many tasks, so focus shifts to labeled data, simulators/GANs for synthetic training, and an architecture of five dimensions (metrics, knowledge, data, state/features, response). Calls out the usual TensorFlow/Nvidia route as insufficient without careful metric and label design.
AI-written summary · about
chapters (14)
- 0:00Start
- 5:28Evolution of DATA
- 5:49Evolution of INTERFACE
- 6:48Evolution of KNOWLEDGE
- 16:50Evolution of PRODUCTS PRODUCT 1.0
- 20:59Teaching Machines
- 25:05Evolution of Al Thinking!
- 25:46A Web of SIGNALS, ACTIONS, and METRICS...
- 29:57RECOMMENDATION Algorithms
- 30:07REASONING Algorithms Which TURN should take next?
- 34:16An Architecture of Intelligence
- 48:33Artificial intelligence pioneer says we need to start over
- 50:04Is this how we learn to translate?
- 58:09Guiding Principles
related talks
30mThe Next Decade of AI
18mGenerative AI and Agentic AI for Future Industry and Enterprises
18mNext-Generation AI Agents: Orchestrating Technological Evolution
14mTwo Keys to AI's Future
36mArtificial Intelligence: The Future of Software
17mFocus on Data, on the Road to AI
18mTrinity of Artificial Intelligence
47mBeyond Agents: What AI Strategy Really Needs in 2025