Interview
summary
Argues that current wins are concrete pattern‑matching systems — computer vision for object recognition and speech recognition — enabled by deep learning, but that those are narrow, data‑hungry and not true intelligence. Describes work on tuning architectures and applying neural nets to formal reasoning and abstract modeling, reports no definitive training recipe yet, and frames next milestones as building systems that manipulate abstract models rather than just fit patterns.
AI-written summary · about
related talks
04mFrom big data to human-level artificial intelligence
03mRethinking software engineering in the AI era
13mArtificial General Intelligence: Why aren't we there yet?
39mA Fireside Chat with Turing Award Winner Geoffrey Hinton, Pioneer of Deep Learning
29mDeep Learning, Neuroscience and the future of AI
27mThe Cutting Edge of AI: From LLMs to Multimodal Intelligence
26mKeynote: AI in the Real World: Today and Tomorrow
19mAI: The current state of play