Bob McGrew, OpenAI’s former Head of Research, led OpenAI from the GPT-3 breakthrough to today’s reasoning models. Three main pillars of AGI—Transformers are scaled pre-training, post-training and reasoning—and that the fundamentals that will shape the next decade-plus are already in place. He thinks 2025 will be defined by reasoning while pre-training hits diminishing returns. AI agents will eventually price all services at compute costs due to near-infinite supply. This will fundamentally disrupting industries like law and medicine. From robotics breakthroughs to managing brilliant researchers, Bob offers a unique perspective on AI’s trajectory and where startups can still find defensible opportunities.
The fundamental concepts that are needed for AGI are pre-training, post-training and reasoning. Those are all that are needed and all that will be developed for the next ten years.
The supply of currently difficult skills (like legal expertise) will rapidly expand. This will make those currently expensive skills cheap. What will be valuable will be human and business relationships.
00:00 Introduction
01:16 The Trifecta of AI: Pre-training, Post-training, and Reasoning
02:19 Deep Dive into AI Reasoning
03:53 Challenges and Future of Pre-training
05:23 Exploring Post-training and Model Personality
06:42 The Future of AI: Predictions and Controversies
11:44 The Rise of AI Agents and Market Opportunities
15:46 Robotics: The Next Frontier
21:31 Proprietary Data and Its Value in AI
24:22 Proprietary Data and Its Importance
24:29 The Rapid Evolution of Coding
25:25 The Future of Coding: Human vs. AI
27:29 The Role of AgTech Software Engineers
29:14 The Concept of ‘Member of the Technical Staff’
31:11 Generational Differences in Using ChatGPT
32:41 AI’s Role in Enhancing Learning and Curiosity
34:45 Preparing the Next Generation for AI
38:57 Daily Uses of AI
41:03 Managing High-Performing Teams
46:39 Security in an Agentic World
48:17 Conclusion and Final Thoughts
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Apples recent report, that there is a gulf between AI’s pattern matching and LLMs, and then being able to reason and infer from all that information, still seems to have the best handle on the current situation.
Its like the intelligence vs wisdom nuance in Dungeons And Dragons. I once had a high intelligence, low wisdom rogue. She would know immediately that the bell rope was a trap, or that there was a monster behind the bush. But she didn’t know what to do with those facts, so would pull the rope or throw a stone at the bush anyway just to see what happened next.