I predict that ASI, demonstrably far smarter than any human, will tell us that moral realism is false and yet philosophers such as the author will still refuse to accept it.
I'm pretty sympathetic to this objection to the "Carl plan", which you acknowledge: "It could be that the right ways of reasoning in verifiable domains differ from the right ways of reasoning in unverifiable domains".
In response to this objection, you say:
> But this proposal seems to get us about as close as we can get to getting the right answers. If we cannot get AIs good at philosophy by having them have superintelligent and accurate constitutions in other domains, it is hard to see how we could be assured they’ve gotten the right answers.
I'm not sure why you think this. I'd agree we're not going to be "assured". But if there are relevant disanalogies between verifiable domains and unverifiable domains, it seems like we should try to adjust for those disanalogies (with careful a priori reasoning), no? We're not restricted to pure extrapolation from performance on the reference class of verifiable domains.
Yeah I think with the Carl plan we should probably supplant it with having them get the right answers to cases where we're confident which view is right even in unverifiable domains.
That's better, I think, but I'm gesturing at something beyond making sure they've gotten "right answers" previously, be it in verifiable or unverifiable domains. I'm saying: Epistemology is about more than extrapolating from labeled examples. We can also ask things like, do the basic philosophical norms the AIs are following seem independently plausible? (And what are our general standards for independent plausibility?)
You could group these things into "right answers" too if you wanted. I'm not sure what I think of that. It seems like it kind of dodges the question of what metaphilosophical methodology we should use to assign these right answer labels.
> one’s ability to recognize good philosophy surpasses their ability to perform good philosophy.
Why do we think this is true? This is not at all obvious to me.
More fundamentally, your whole argument seems to be built on the idea that knowledge in unverifiable domains like philosophy is a priori, and I don't think this is true. I think that even in unverifiable domains, advancement generally comes from interacting with the world, not a priori thinking. This is one of big insights of bayesianism and probabilistic thinking more generally - even where a piece of evidence doesn't prove or disprove a claim, even where no hypothetical evidence could prove or disprove a claim, empirical evidence can still increase or decrease the likelihood of the claim. Gather enough empirical evidence, and you can be very certain of the truth or falsity even of an unverifiable claim. This process is how moral progress generally occurs. But to illustrate, lets use your example of an unverifiable claim from physics. Do stars continue to exist after passing beyond the edge of the observable universe? I take it we both agree that they do, even though there is no empirical verification, even though there probably can never be empirical verification. Why are we confident of this? It's not because of any a priori reasoning. Philosophers did not teach us this. It is because we have a rather precise model of the laws of physics developed from a great deal of empirical evidence. This model tells us that the edge of the observable universe is result of how fast things move away from us and how fast information can travel toward us, nothing more. And this model tells us that there is nothing particularly special about us or where we are - an observer somewhere else would have an observable universe with different boundaries for the same reasons. Future empirical discoveries might undermine this model. We might find empirical evidence that there is something special about us or where we are, and if we do then that would reduce our confidence that stars continue to exist after passing beyond the edge of the observable universe. It's empirical, not a priori, and no amount of sitting and thinking, either by humans or AI, would advance our understanding much. Ethics is no different.
n=1, but I spent a month having AI give me single card Tarot card readings and found it considerable better than my own readings in two ways: the depth of interpretation of that days card, 2) the ability identify historical patterns and other significant orderings and relationships.
"Historically and conceptually, tarot acts as a "philosophical alphabet". It integrates major schools of thought—like Neoplatonism, Stoicism, and Hermeticism—while serving as a tool for psychology, epistemology, and self-reflection." --AI
Thanks, this changed my mind a little bit
I predict that ASI, demonstrably far smarter than any human, will tell us that moral realism is false and yet philosophers such as the author will still refuse to accept it.
I'm pretty sympathetic to this objection to the "Carl plan", which you acknowledge: "It could be that the right ways of reasoning in verifiable domains differ from the right ways of reasoning in unverifiable domains".
In response to this objection, you say:
> But this proposal seems to get us about as close as we can get to getting the right answers. If we cannot get AIs good at philosophy by having them have superintelligent and accurate constitutions in other domains, it is hard to see how we could be assured they’ve gotten the right answers.
I'm not sure why you think this. I'd agree we're not going to be "assured". But if there are relevant disanalogies between verifiable domains and unverifiable domains, it seems like we should try to adjust for those disanalogies (with careful a priori reasoning), no? We're not restricted to pure extrapolation from performance on the reference class of verifiable domains.
Yeah I think with the Carl plan we should probably supplant it with having them get the right answers to cases where we're confident which view is right even in unverifiable domains.
That's better, I think, but I'm gesturing at something beyond making sure they've gotten "right answers" previously, be it in verifiable or unverifiable domains. I'm saying: Epistemology is about more than extrapolating from labeled examples. We can also ask things like, do the basic philosophical norms the AIs are following seem independently plausible? (And what are our general standards for independent plausibility?)
You could group these things into "right answers" too if you wanted. I'm not sure what I think of that. It seems like it kind of dodges the question of what metaphilosophical methodology we should use to assign these right answer labels.
> one’s ability to recognize good philosophy surpasses their ability to perform good philosophy.
Why do we think this is true? This is not at all obvious to me.
More fundamentally, your whole argument seems to be built on the idea that knowledge in unverifiable domains like philosophy is a priori, and I don't think this is true. I think that even in unverifiable domains, advancement generally comes from interacting with the world, not a priori thinking. This is one of big insights of bayesianism and probabilistic thinking more generally - even where a piece of evidence doesn't prove or disprove a claim, even where no hypothetical evidence could prove or disprove a claim, empirical evidence can still increase or decrease the likelihood of the claim. Gather enough empirical evidence, and you can be very certain of the truth or falsity even of an unverifiable claim. This process is how moral progress generally occurs. But to illustrate, lets use your example of an unverifiable claim from physics. Do stars continue to exist after passing beyond the edge of the observable universe? I take it we both agree that they do, even though there is no empirical verification, even though there probably can never be empirical verification. Why are we confident of this? It's not because of any a priori reasoning. Philosophers did not teach us this. It is because we have a rather precise model of the laws of physics developed from a great deal of empirical evidence. This model tells us that the edge of the observable universe is result of how fast things move away from us and how fast information can travel toward us, nothing more. And this model tells us that there is nothing particularly special about us or where we are - an observer somewhere else would have an observable universe with different boundaries for the same reasons. Future empirical discoveries might undermine this model. We might find empirical evidence that there is something special about us or where we are, and if we do then that would reduce our confidence that stars continue to exist after passing beyond the edge of the observable universe. It's empirical, not a priori, and no amount of sitting and thinking, either by humans or AI, would advance our understanding much. Ethics is no different.
n=1, but I spent a month having AI give me single card Tarot card readings and found it considerable better than my own readings in two ways: the depth of interpretation of that days card, 2) the ability identify historical patterns and other significant orderings and relationships.
"Historically and conceptually, tarot acts as a "philosophical alphabet". It integrates major schools of thought—like Neoplatonism, Stoicism, and Hermeticism—while serving as a tool for psychology, epistemology, and self-reflection." --AI