If by “decide” you mean control the outcome in any meaningful way I agree, we cannot. However I think it is possible to make a best effort attempt to steer things towards a better future (in small and big ways). Mistakes will be made, progress is never linear and we may even fail altogether, but the attempt is really all we have, and there is reason to believe in a non-trivial probability that our efforts will bear fruit, especially compared to not trying or to aiming towards something else (like maximum power in the hands of a few).
For a great exploration of this topic I refer to this talk by Nick Bostrom: http://www.stafforini.com/blog/bostrom. The tl;dr is that we can come up with evaluation functions for states of the world that, while not yet being our desired outcome, are indications that we are probably moving in the right direction. We can then figure out how we get to the very next state, in the near future. Once there, we will jot a course for the next state, and so on. Bostrom signals out technology, collaboration and wisdom as traits humanity will need a lot of in the better future we are envisioning, so he suggests can measure them with our evaluation function.
To me that doesn’t sound very different from “I want a future with less suffering, so I’m going to evaluate my impact based on how far humanity gets towards eradicating malaria and other painful diseases”. Which I guess is consistent with my views but doesn’t sound like most long-termists I’ve met.
Well, it wouldn’t work if you said “I want a future with less suffering, so I am going to evaluate my impact based on how many paper clips exist in the world at a given time”. Bostrom selects collaboration, technology and wisdom because he thinks they are the most important indicators of a better future and reduced x-risk. You are welcome to suggest other parameters for the evaluation function of course, but not every parameter works. If you read the analogy to chess in the link I posted it will become much more clear how Bostrom is thinking about this.
(if anyone reading this comment knows of evolutions in Bostrom’s thought since this lecture I would very much appreciate a reference)
Hi Khorton,
If by “decide” you mean control the outcome in any meaningful way I agree, we cannot. However I think it is possible to make a best effort attempt to steer things towards a better future (in small and big ways). Mistakes will be made, progress is never linear and we may even fail altogether, but the attempt is really all we have, and there is reason to believe in a non-trivial probability that our efforts will bear fruit, especially compared to not trying or to aiming towards something else (like maximum power in the hands of a few).
For a great exploration of this topic I refer to this talk by Nick Bostrom: http://www.stafforini.com/blog/bostrom. The tl;dr is that we can come up with evaluation functions for states of the world that, while not yet being our desired outcome, are indications that we are probably moving in the right direction. We can then figure out how we get to the very next state, in the near future. Once there, we will jot a course for the next state, and so on. Bostrom signals out technology, collaboration and wisdom as traits humanity will need a lot of in the better future we are envisioning, so he suggests can measure them with our evaluation function.
To me that doesn’t sound very different from “I want a future with less suffering, so I’m going to evaluate my impact based on how far humanity gets towards eradicating malaria and other painful diseases”. Which I guess is consistent with my views but doesn’t sound like most long-termists I’ve met.
Well, it wouldn’t work if you said “I want a future with less suffering, so I am going to evaluate my impact based on how many paper clips exist in the world at a given time”. Bostrom selects collaboration, technology and wisdom because he thinks they are the most important indicators of a better future and reduced x-risk. You are welcome to suggest other parameters for the evaluation function of course, but not every parameter works. If you read the analogy to chess in the link I posted it will become much more clear how Bostrom is thinking about this.
(if anyone reading this comment knows of evolutions in Bostrom’s thought since this lecture I would very much appreciate a reference)