I think the fact portrayed by this graph is underemphasized.
It has significant implications for both domestic and international competition. On the domestic side, it’s relevant to the landscape of competition as AI R&D automation kicks off. On the international side, it is one of the most elegant ways to argue that DSA is likely.
As a corollary, I’m not sure we’ve adequately oriented AI policy and governance strategy based on endgame considerations like the vulnerable world hypothesis and longterm value competition. All of these questions and problems might hit us in a very small window following AI R&D automation.
one of the most elegant ways to argue that DSA is likely
Seems more like speculation than “fact” or even an “argument” to me. It’s a graph with an arbitrary[1] assumption that calendar-time leads remain constant over time,[2] and that the rate of technological progress shoots up very rapidly.
Sample counterargument (that first popped into my mind): as AI progress increases over time, AI-powered attempts to steal the code/weights/training data/information about architectures/etc. of other participants in the race make it so that, due to the offense-defense imbalance, the calendar-time lead of the top participant in the race goes down over time.
Are there good arguments that the cause of some countries have a consistent-in-calendar-time lead over others are such that we should expect the calendar gap to remain constant even as the amount of useful work that can be done per unit of calendar time increases in both the leading and lagging country?
I think the fact portrayed by this graph is underemphasized.
It has significant implications for both domestic and international competition. On the domestic side, it’s relevant to the landscape of competition as AI R&D automation kicks off. On the international side, it is one of the most elegant ways to argue that DSA is likely.
As a corollary, I’m not sure we’ve adequately oriented AI policy and governance strategy based on endgame considerations like the vulnerable world hypothesis and longterm value competition. All of these questions and problems might hit us in a very small window following AI R&D automation.
Seems more like speculation than “fact” or even an “argument” to me. It’s a graph with an arbitrary[1] assumption that calendar-time leads remain constant over time,[2] and that the rate of technological progress shoots up very rapidly.
Until justified! That’s where the tougher task comes into play, if you want to convince someone who doesn’t already agree with you
Sample counterargument (that first popped into my mind): as AI progress increases over time, AI-powered attempts to steal the code/weights/training data/information about architectures/etc. of other participants in the race make it so that, due to the offense-defense imbalance, the calendar-time lead of the top participant in the race goes down over time.
Are there good arguments that the cause of some countries have a consistent-in-calendar-time lead over others are such that we should expect the calendar gap to remain constant even as the amount of useful work that can be done per unit of calendar time increases in both the leading and lagging country?