Back to Ryan Greenblatt – What happens once AI can automate AI research?, with Ryan Greenblatt.
Did Expert Human Data Drive Recent AI Progress?
Ryan Greenblatt says scaling up expert human data has not been hugely important for AI research and development. Removing the last two doublings of data generation from expert humans would not make a huge difference.
Over the last few years, labs scaled compute, the number of people working at AI companies, and the effort spent on data labeling. Much of the recent gain comes from better ways to use humans and AIs to construct reinforcement learning environments. Environments are much better than in 2024 because researchers know which environments they want and how to structure them, and because huge amounts of AI labor build them. Knowing what to build and using AI labor matter more than human labor spent building the environments. Human labor still matters. People want a very large number of environments, and the AIs are pretty good at making them given a sense of what the environment should be. Many of the environments have good verification loops.
Frontier lab spending is overwhelmingly compute, because compute is easier to scale up than data. A rough compute-to-data split is about 20 to 1 or 10 to 1, and the exact ratio depends on the company.
Ryan Greenblatt – What happens once AI can automate AI research? ·
From the transcript
We have better algorithms. A huge thing that's happened is that we have built a deca-billion-dollar data industry, which has... systematically collected and codified expert human judgment across all kinds of different disciplines, codified in the form of RL environments, codified in the form of SFT traces that these experts built to help the model better understand how do you do coding and how do you build complex infrastructure projects? How do you do law? How do you do whatever, whatever? And how are the AIs able to replicate the effect that currently expert human judgment seems to be playing in AI progress? Yeah.
So my sense is that scaling up the amount of effort spent on getting expert human data has not been hugely important for AI R&D in general. So in particular, over the last few years, we've been scaling up compute, scaling up people working at AI companies, and scaling up the amount of effort spent on data labeling. My sense is that if you remove the last two doublings or whatever of data generation from expert humans, that would not make a huge difference.