In a 2021 podcast with Jason Calacanis, you said, “None of us know what’s going to happen in the future, obviously [...] because we don’t know, we can make any claim we want.”
That’s true in the sense that no claim about the future is ruled out about facts about the future, and false in the sense that any such claim automatically has any epistemic weight. Anyone can make an unjustified claim , but if it’s only an unjustified claim, there is no reason for anyone else to agree.
But that’s importantly wrong!
It’s a rather vague statement, and whether its right or wrong depends on the exact interpretation.
A relevant slogan: “The map is not the territory. A blank map does not correspond to a blank territory.”
How does that help? That the territory is not blank just means that something will happen in the future. It doesn’t tell you what it is. It doesn’t offer evidence or justification.
That is, a corollary of the claims we make about reality (“the map”) being different from reality itself (“the territory”), is that just because we don’t know what’s going to happen in the future, doesn’t mean that the future is actually indeterminate in any real sense.
If we can’t determine it, it’s undetermined in that sense, even if it’s ls determined ontologically “in itself”. Which is relevant? Actually , it’s the first sense: what the debate is about is the epistemology, what we can know about the future,because Metz is objectivlng to unjustified claims, mere opinions
In the theory of relativity, the future is just a different region of spacetime; it’s still “there” even if we can’t “see” it. It’s true that none of us know with certainty what’s going to happen in the future, and because of that, any claims we make today can’t be immediately proven wrong.
How does that help? Yes, in a block universe theory, the future is already there , but that still doesn’t mean it’s knowable.
What helps knowability is laws and patterns, not mere existence.
But that’s not the same thing as “we can make any claim we want”, because we can still express uncertain, probabilistic beliefs, and keep track of whose probabilistic predictions make better predictions over time.
What’s the difference? If you assume an arbitrary prior, you can literally assert anything. Saying it’s probabilities doesn’t change that,because it is subjective probability (the thing that should properly be called credence) not objective probability the kind of probability that is based on empirical evidence. Bayes allows you to adjust your probabilities by evidence once you have evidence … but you don’t, about the future.
People are pretty familiar with probability in the context of, say, weather forecasting, or sports betting. If I say it’s a 70% chance that it’s going to be sunny tomorrow, or that the Giants are going to win, what that means (if I’m doing it right) is that out of 10 times when I say that something is going to happen with a 70% chance, it should actually happen 7 times.
Yes , you can have evidence-based probability about future events where they follow the same pattern as previous events—inductive probability.
The better argument against Metz is the one about based on empirical evidence about geometrical growth in capabilities. Even if we don’t know how to build ASI (assuming it’s not just a series of incremental improvement cements in current LLMs,which we also don’t know), we have data about other technological revolutions. We are not in a position of comple ignorance. Whatever point he is making , there are ways of countering it—which have nothing to do with map/teritory distinctions.
A big part of “rationalist” philosophy is that this style of thinking is actually very general, that you can assign “subjective” probabilities to anything that might happen in the future,
including things that have never happened before, like AI catastrophes
Yes, but they are not knowledge unless they relate to something in the present or past—in which case they are not wholly subjective.
Wholly subjective claims are in no way different to the thing Metz is complaining about. If you want to sell probability to someone, you had better make sure it’s not subjective probability. Certainty is unobtainable but subjective probability is valueless
Probability is not just for obviously repeatable events like the day’s weather or a baseball game or a coin toss, where people can measure and agree on the long-run frequency over many days or games or tosses.
Probability as opposed to credence actually is. Using the same word for two different things confuses the issue.
That’s true in the sense that no claim about the future is ruled out about facts about the future, and false in the sense that any such claim automatically has any epistemic weight. Anyone can make an unjustified claim , but if it’s only an unjustified claim, there is no reason for anyone else to agree.
It’s a rather vague statement, and whether its right or wrong depends on the exact interpretation.
How does that help? That the territory is not blank just means that something will happen in the future. It doesn’t tell you what it is. It doesn’t offer evidence or justification.
If we can’t determine it, it’s undetermined in that sense, even if it’s ls determined ontologically “in itself”. Which is relevant? Actually , it’s the first sense: what the debate is about is the epistemology, what we can know about the future,because Metz is objectivlng to unjustified claims, mere opinions
How does that help? Yes, in a block universe theory, the future is already there , but that still doesn’t mean it’s knowable.
What helps knowability is laws and patterns, not mere existence.
What’s the difference? If you assume an arbitrary prior, you can literally assert anything. Saying it’s probabilities doesn’t change that,because it is subjective probability (the thing that should properly be called credence) not objective probability the kind of probability that is based on empirical evidence. Bayes allows you to adjust your probabilities by evidence once you have evidence … but you don’t, about the future.
Yes , you can have evidence-based probability about future events where they follow the same pattern as previous events—inductive probability.
The better argument against Metz is the one about based on empirical evidence about geometrical growth in capabilities. Even if we don’t know how to build ASI (assuming it’s not just a series of incremental improvement cements in current LLMs,which we also don’t know), we have data about other technological revolutions. We are not in a position of comple ignorance. Whatever point he is making , there are ways of countering it—which have nothing to do with map/teritory distinctions.
Yes, but they are not knowledge unless they relate to something in the present or past—in which case they are not wholly subjective.
Wholly subjective claims are in no way different to the thing Metz is complaining about. If you want to sell probability to someone, you had better make sure it’s not subjective probability. Certainty is unobtainable but subjective probability is valueless
Probability as opposed to credence actually is. Using the same word for two different things confuses the issue.