I finally realized that when people were talking about donating to SIAI during the rationality minicamp, most of us (certainly myself) were thinking of may be tens of thousands of dollars a year. I now understand that’s silly. If our goal is truly to make the most money for SIAI, then the goal should be measured in billions.
Eliezer has said that he doesn’t know how to usefully spend more than 10 million dollars...
Thankfully there are more people at SIAI than just him. :)
As a very simple start, they could make sure that all SIAI researchers can focus on their work, by taking care of their other chores for them.
I am guessing that a part of what EY had in mind was that large organizations tend to lose its purposes and work towards self-preservation as much as towards the original objective. $10M/year translates into less than 100 full-time jobs, which is probably a good rule of thumb for an organization becoming too big to keep its collective eyes on the ball.
What chores did you have in mind? Seems like you wouldn’t even need 100 million dollars to hire the world’s best mathematicians (if you have any hope of doing so) and give all involved a comfortable lifestyle. Do you mean, billions total by the time we get FAI? Because you started out speaking of donating such-and-such “a year”.
My point was that if your goal is to give money to an organization (and it seemed that that’s what a lot of people at rationality minicamp were planning to do for SIAI), thinking in thousands is, while a nice gesture, is not very helpful. We should be thinking in billions (not necessarily per year, I’m just talking order of magnitude here).
As for chores it’s can start as basic as living accommodations, private jets, and get as complex as security guards and private staff. I won’t argue for any specifics here, though. I’m just arguing that having lots of money is nice and helpful.
See, I’m not sure more is better after a certain point—shminux touched on this.
I also think if we hired the world’s best mathematicians for a year—assuming FAI would take more than this—some of them would either get interested and work for less money, or find some proof SI’s current approach won’t work.
Eliezer has said that he doesn’t know how to usefully spend more than 10 million dollars...
Thankfully there are more people at SIAI than just him. :) As a very simple start, they could make sure that all SIAI researchers can focus on their work, by taking care of their other chores for them.
I am guessing that a part of what EY had in mind was that large organizations tend to lose its purposes and work towards self-preservation as much as towards the original objective. $10M/year translates into less than 100 full-time jobs, which is probably a good rule of thumb for an organization becoming too big to keep its collective eyes on the ball.
What chores did you have in mind? Seems like you wouldn’t even need 100 million dollars to hire the world’s best mathematicians (if you have any hope of doing so) and give all involved a comfortable lifestyle. Do you mean, billions total by the time we get FAI? Because you started out speaking of donating such-and-such “a year”.
My point was that if your goal is to give money to an organization (and it seemed that that’s what a lot of people at rationality minicamp were planning to do for SIAI), thinking in thousands is, while a nice gesture, is not very helpful. We should be thinking in billions (not necessarily per year, I’m just talking order of magnitude here).
As for chores it’s can start as basic as living accommodations, private jets, and get as complex as security guards and private staff. I won’t argue for any specifics here, though. I’m just arguing that having lots of money is nice and helpful.
See, I’m not sure more is better after a certain point—shminux touched on this.
I also think if we hired the world’s best mathematicians for a year—assuming FAI would take more than this—some of them would either get interested and work for less money, or find some proof SI’s current approach won’t work.
Sure, there are diminishing returns after some point.
And both of those results would be good.