Ryan Greenblatt

So, if you get a serial labor acceleration of 4x, I think this increases AI R&D progress by roughly 2.15x.

AI R&D is only a subset of AI progress.

Some of the AI progress is driven by scaling up compute for training runs.

I tend to think that roughly two-thirds of AI progress is algorithms while roughly one-third is from scaling up compute for training runs.

This means you get only two-thirds by 2.15 plus 1 divided by 3 equals 1.75x AI progress increase from for x serial labor acceleration.

To get a 2x increase in the rate of AI progress, assuming these constants, would need roughly 5.3x serial labor acceleration.

That's the end of the list.

This model is basically a simplified version of the AI futures project model with somewhat different constants.

Heading.

Appendix.

Different notions of uplift.

There are several different concepts that could be meant by productivity uplift, and which one we're talking about makes a huge difference.

Serial labor acceleration.

Suppose you could speed up everyone at the company by X but had to use no AI assistance or only 20-20 AIs in your work.

For what X would you be indifferent?

Just taking into account productivity, ignoring safety.

Parallel labor acceleration.

Suppose you could magically grow the company by a factor of X, where the new people would have a similar distribution of skills and knowledge to the current people, including knowledge about the company, etc., but had no AI assistance.

For what X would you be indifferent?

Current work acceleration.