Well obviously, once you accept that everything else follows. What I’m asking is why you think that
when I wrote that Nature does not have “laws,” I did not mean what I did not write. I did not write that Nature was arbitrary, for example. What we call “laws,” however, are interpretations, heuristics, models. Nature has no problems to solve, needs no predictions. We have problems, and an AI will have problems.
It is obvious that Nature is commonly predictable. However, “commonly” is not enough. Can Nature disobey, for example, her own “laws”? It looks like she can, because what we have come to is an understanding of probability. If a situation can resolve as A or B, with some probability for each, we can develop laws that predict probability, but that does not predict any particular behavior in an individual instance. We have developed a probability. Nature doesn’t handle itself that way.
However, you can believe that Nature does follow laws if you like. And you can attempt to determine those laws. And that is highly useful, even necessary. However, if you believe what you have created, as if Nature were bound by the :”laws” you have inferred—and, hopefully, tested—you will become less able to see the exceptions. If by any chance, you notice one (the cards are stacked against it), you will dismiss it as some error.
, give that it looks very lawful: objects fall down, energy is conserved, if a prediction is true on Monday it stays true on Tuesday, every exception to known rules turns out to obey deeper rules with practical consequences we can exploit. Why can’t we just say “The thing has looked absolutely lawful for millenia, case closed”?
someone who believes Nature does have this construct called “laws” is not reductionist
?!?!??
When I read Yudkowsky, I interpret him as I interpret the Qur’an. That is, I assume that he’s right. It’s just a temporary assumption that facilitates understanding. So I have interpreted “reductionism” in a way that makes it—to me—right. It’s consistent with what I’ve read, so far. However, applying this, Yudkowsky is not always fully reductionist. Essentially, he’s human. So far, anyway. Seems to me that he acknowledges this.
Building on what TimS said, I would make a lot fewer assumptions if you explained things as you would to a toddler rather than express complex ideas in five highly ambiguous words and then complain I didn’t interpret each word exactly the way you meant to.
You still haven’t answered the question of why nature looks a whole lot like it has laws. Even if true, how can you possibly know that the model we have tested ten thousand times and confirmed each time is a surface model and not the truth?
Can Nature disobey, for example, her own “laws”? It looks like she can, because what we have come to is an understanding of probability.
Disagree. First, a probabilistic law is still a law; take nondeterministic Turing machines for example. Second, with the exception of some interpretations of QM, probabilistic models claim to be approximations of the true (or still approximate but a level deeper) laws when the human trying to use it lacks information or time to compute. Sometimes we find nature disobeying situation A’s law in situation B, but there always turns out to be a law that governs A, B, and C.
However, you can believe that Nature does follow laws if you like.
Clearly we’re not using “believe” the same way.
if you believe what you have created, as if Nature were bound by the :”laws” you have inferred—and, hopefully, tested—you will become less able to see the exceptions. If by any chance, you notice one (the cards are stacked against it), you will dismiss it as some error.
I call bullshit. Physicists catch exceptions to current theories all the time, and then work hard to find where it came from, and either devise new theories or fix the loose cable in their setup. Where’s the list of exceptions discovered by mysterians?
So I have interpreted “reductionism” in a way that makes it—to me—right.
You should probably interpret it as “what most reductionists advocate” and use “the version of reductionism I think is right” for the other thing, if you hope to talk to people who call themselves reductionists without making them shout “What the actual fuck?”.
In the abstract, there’s nothing incoherent about using others’ technical vocabulary with different meanings. But:
1) You are misleading yourself if you think you are communicating effectively with them. 2) Your statements about their positions will make no sense to them. 3) It comes off as quite arrogant—essentially you are asserting you know better how a concept works without even attempting to justify the assertion (or even noticing that such an assertion is necessary)
If Eliezer says he is a reductionist1 and you say he is not a reductionist2, you haven’t done anything particularly clever or praiseworthy.
when I wrote that Nature does not have “laws,” I did not mean what I did not write. I did not write that Nature was arbitrary, for example. What we call “laws,” however, are interpretations, heuristics, models. Nature has no problems to solve, needs no predictions. We have problems, and an AI will have problems.
It is obvious that Nature is commonly predictable. However, “commonly” is not enough. Can Nature disobey, for example, her own “laws”? It looks like she can, because what we have come to is an understanding of probability. If a situation can resolve as A or B, with some probability for each, we can develop laws that predict probability, but that does not predict any particular behavior in an individual instance. We have developed a probability. Nature doesn’t handle itself that way.
However, you can believe that Nature does follow laws if you like. And you can attempt to determine those laws. And that is highly useful, even necessary. However, if you believe what you have created, as if Nature were bound by the :”laws” you have inferred—and, hopefully, tested—you will become less able to see the exceptions. If by any chance, you notice one (the cards are stacked against it), you will dismiss it as some error.
When I read Yudkowsky, I interpret him as I interpret the Qur’an. That is, I assume that he’s right. It’s just a temporary assumption that facilitates understanding. So I have interpreted “reductionism” in a way that makes it—to me—right. It’s consistent with what I’ve read, so far. However, applying this, Yudkowsky is not always fully reductionist. Essentially, he’s human. So far, anyway. Seems to me that he acknowledges this.
I could write more, but I’m in awe.
Building on what TimS said, I would make a lot fewer assumptions if you explained things as you would to a toddler rather than express complex ideas in five highly ambiguous words and then complain I didn’t interpret each word exactly the way you meant to.
You still haven’t answered the question of why nature looks a whole lot like it has laws. Even if true, how can you possibly know that the model we have tested ten thousand times and confirmed each time is a surface model and not the truth?
Disagree. First, a probabilistic law is still a law; take nondeterministic Turing machines for example. Second, with the exception of some interpretations of QM, probabilistic models claim to be approximations of the true (or still approximate but a level deeper) laws when the human trying to use it lacks information or time to compute. Sometimes we find nature disobeying situation A’s law in situation B, but there always turns out to be a law that governs A, B, and C.
Clearly we’re not using “believe” the same way.
I call bullshit. Physicists catch exceptions to current theories all the time, and then work hard to find where it came from, and either devise new theories or fix the loose cable in their setup. Where’s the list of exceptions discovered by mysterians?
You should probably interpret it as “what most reductionists advocate” and use “the version of reductionism I think is right” for the other thing, if you hope to talk to people who call themselves reductionists without making them shout “What the actual fuck?”.
In the abstract, there’s nothing incoherent about using others’ technical vocabulary with different meanings. But:
1) You are misleading yourself if you think you are communicating effectively with them.
2) Your statements about their positions will make no sense to them.
3) It comes off as quite arrogant—essentially you are asserting you know better how a concept works without even attempting to justify the assertion (or even noticing that such an assertion is necessary)
If Eliezer says he is a reductionist1 and you say he is not a reductionist2, you haven’t done anything particularly clever or praiseworthy.