Keynote OK Computer
summary
A concise survey of how recent deep learning makes expensive supercomputer power cheap and practical: examples include an old supercomputer vs a PlayStation 3 at similar FLOPS, Google using 16,000 processors to learn “cat”, AlphaGo 4–1 beating a world champion, and Watson now available via TensorFlow and web APIs. Messaging and voice scale—WhatsApp 900M users and 30B messages/day, voice error rates drop from ~26% to single digits—drive bots and Alexa into homes; simple Telegram bots and auto-generated sites work in days; finance sees ~40% fewer jobs since 2004.
AI-written summary · about
chapters (11)
- 0:00Introduction and speaker
- 2:35AI breakthroughs and speed
- 7:39Natural language processing shift
- 10:35Conversational user interfaces
- 16:57AI impact on jobs
- 21:58Human creativity vs AI
- 26:37AI tools and accessibility
- 31:10Building AI-powered services
- 35:04Intelligence and conversation design
- 37:14Privacy and security concerns
- 38:16Future of AI and opportunities
related talks
59mKeynote: AI without the BS, for humans
38mKeynote: Thriving in the Age of AI
20mKeynote: Open-Sourced Inspiration - The Present and Future of Tech - AI
59mKeynote: The past, present, and future of AI for application developers
50mKeynote: After the AI Hype – What’s Real, and What’s Next
25mKeynote Kickoff
45mKeynote: Conversations with Machines
17mKeynote