Many people are bearish about EA/AIS university organizing in short timelines worlds. (less than 3 years?) I think this is directionally correct, and the simple story where most students are many years from substantial impact and so are less useful in short timelines is true.
However, I think people should be a bit less bearish on the margin. The first reason for this is that students seem to have a unique advantage in some types of political campaigning and communication, and this will be quite useful in many short timelines worlds.
Outside of EA/AIS, student political organizing has been fairly important, and the limited historical evidence from inside EA/AIS suggests this is equally true here.
I also suspect that people are overly bearish on students because they have a habit of rounding small amounts of personal impact to 0, rather than some small number. In Shapley value calculations, less important players still get some credit, and the whole category of “less important players” can often get a fair amount of credit. If a class of people is relatively less important to EA/AIS, you might be tempted to have a impercise estimate of their (low) value, biased downwards. This could be fine in a vacuum, but it's troublesome when neglectedness means that you could influence things quite a bit in this category, and the exact payoff is now sensitive to your particular low estimate.
As a university organizer at a very STEM focused state school, I suspect that students getting liberal arts degrees are more easily convinced to pursue a career in direct work. If this is the case, it could be because direct work compares more favorably with the other career options of those with liberal arts degrees, or because the clearer career outcomes of STEM majors create more path dependence and friction when they consider switching careers. This is potentially another thing to keep in mind when trying to compare the successes of EA uni groups.
Crossposted to Less Wrong
‘Like most of life’s dangers, the dangers of artificial intelligence are not easily identified from the armchair.’
J Dmitri Gallow
Epistemic status
I’m not an expert in this field, and herein I make some sweeping generalisations educated guesses about the overall...
Note: This post was crossposted from Planned Obsolescence by the Forum team, with the author's permission. The author may not see or respond to comments on this post.
Subtitle: If acceleration compounds, the naive extrapolation implies superintelligence within a year
This post discusses a public METR report on AI R&D acceler...
TL;DR:
* When we started Rethink Priorities (RP) in 2018, our mission was, simply put, to close the knowledge gaps that stopped people from improving the world as much as they could. That hasn’t changed much, but the landscape we’re doing it in is shifting.
* In 2027, we plan to grow our core GHD, Animal Welfare, and Worldview Investigations teams to take on emerging priorities across causes, including those that the forthcoming transformative AI is creating.
...
Many people are bearish about EA/AIS university organizing in short timelines worlds. (less than 3 years?) I think this is directionally correct, and the simple story where most students are many years from substantial impact and so are less useful in short timelines is true.
However, I think people should be a bit less bearish on the margin. The first reason for this is that students seem to have a unique advantage in some types of political campaigning and communication, and this will be quite useful in many short timelines worlds.
Outside of EA/AIS, student political organizing has been fairly important, and the limited historical evidence from inside EA/AIS suggests this is equally true here.
I also suspect that people are overly bearish on students because they have a habit of rounding small amounts of personal impact to 0, rather than some small number. In Shapley value calculations, less important players still get some credit, and the whole category of “less important players” can often get a fair amount of credit. If a class of people is relatively less important to EA/AIS, you might be tempted to have a impercise estimate of their (low) value, biased downwards. This could be fine in a vacuum, but it's troublesome when neglectedness means that you could influence things quite a bit in this category, and the exact payoff is now sensitive to your particular low estimate.