Beyond GenAI
summary
Argues that GPT-4 plus self-directed internal review can invent algorithms that outperform human-designed optimizers on benchmarks (red bars = human; non-red = machine), with AI-generated optimizers showing significant improvement after hundreds of internal scored iterations and promising further gains at thousands–tens of thousands. Demonstrates MidJourney images made in minutes versus a professional two-week $1,000 workflow, shows Run Diffusion producing contest-ready art, and outlines a 10-prompt prompting assessment and planned cash competitions to validate prompting skill.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:31Professional journey
- 5:39Innovation processes
- 7:10Generative AI in practice
- 12:29Giving AI agency
- 17:18AI-driven scientific discovery
- 25:19Future skills and prompting
- 31:43Q&A and concluding thoughts
related talks
28mCan AI Write Its Own Story? Unveiling the Power of Self-Documenting AI.
15mPractical Evolution of Data Science with GenAI
01h 00mAI A-Z
49mGenerative AI: Concepts, Tools, and Applications for Data Professionals
17mGenerative AI: Keeping it real
52mBeyond the AI Hype: What's Real, What's Next
25mThe Great AI Game
56mSeparating fact from fiction in a world of AI fairytales
