The DNA claim I agree is absurd, though not nearly as absurd as you make it out to be. Certainly Democritus proposed the existence of atoms long before we had anything like microscopes. It’s not inconceivable that ancient people could have deduced mathematical efficiencies of the double helix structure empirically and woven that into mythological stories, and some of these mathematical efficiencies are relevant reasons for DNA being actually the way it is. I think the DNA claim is basically a rare false positive for an otherwise useful general cognitive strategy, see (4) below.
It’s also not literally inconceivable that someone in Egypt formulated and technically solved the alignment problem, but I wouldn’t put odds on that of more than 1×10−7. Yes, I am prepared to make a million statements with that confidence and not expect to lose money to the gods of probability.
Belief in God is something that needs to be disentangled about Peterson, he always hesitates to state he “believes in God” for exactly the reason of being misinterpreted this way. The closest thing to what he means by “faith in God” that I can express is “having a terminal value,” and that statement translates to “human beings cannot be productive (including create mathematics) without a terminal value,” i.e. you cannot derive Ought from Is.
This seems motte-and-bailey. If that’s what he means, shouldn’t he just advance “terminal values are necessary to solve Moore’s open question”?
I feel like throughout the comments defending Peterson, the bottom line has been written first and everything else is being justified post facto.
It’s also not literally inconceivable that someone in Egypt formulated and technically solved the alignment problem, but I wouldn’t put odds on that of more than 1×10−7. Yes, I am prepared to make a million statements with that confidence and not expect to lose money to the gods of probability.
When I say inconceivable I don’t mean literally inconceivable. People have done some pretty absurd things in the past. What is your subjective probability that the most prolific mathematician of all time did half of his most productive work after going blind in both eyes?
I feel like throughout the comments defending Peterson, the bottom line has been written first and everything else is being justified post facto.
I can’t speak for others, but I have spent hundreds of hours thinking about Peterson’s ideas and formulating which parts I agree with and why, including almost every argument that has been put forth so far. Me going back retrospectively and extracting the reasons I made each decision to believe what I believe will look from the outside just like “writing down the bottom line and justifying things post facto.” In general I don’t think the bottom line fallacy is one that you’re supposed to use on other people’s reasoning.
That’s surprising but not that surprising: Milton wrote much of his best poetry while blind, and Beethoven was famously deaf. Conversely, I cannot think of a single unambiguous example of a mythological motif encoding a non-obvious scientific truth (such as that nothing can go faster than light, or that all species evolved from a single-celled organism, or that the stars are trillions and trillions of miles away), so I think this is very very unlikely.
I wasn’t saying that unusual things can’t happen. I should have made myself clearer—what I was getting at was with respect to claims that ancient societies managed to spontaneously derive properties of things they were, in fact, literally incapable of observing. That smells like a second law of thermodynamics-violating information gain to me.
The assertion I’m making is not that Peterson is bad, or that he never has amazing insights, etc. My point is purely with respect to putting him on Eliezer’s level of truth-seeking and technical rationality. Having been wrong about things does not forever disbar you from being a beisutsukai master. However, if one is wrong about important things, becomes aware of the methods of rationality (as I imagine he has), thinks carefully, and still retains their implausible beliefs—that should be enough to indicate they aren’t yet on the level required.
On the other hand, I notice I am confused and that I am disagreeing with people whom I respect very much. I’m happy to update on any new information, but I have a hard time seeing how I could update very far on this particular claim, given that he is indeed quite religious.
Regarding the DNA claim: I think what I’m saying is much weaker than what you think I’m saying.
e.g. ancient people discovered how to store information sequentially in a book. DNA stores information sequentially. This is not surprising. Why would it be inconceivable that the double helix structure is not uniquely weird about DNA?
My steelman of Peterson’s claim about DNA is not that ancient people knew what DNA was or were making any attempt to map it, but that there might be some underlying mathematical reason (such as high compressibility) that the double helix structure is amenable to information storage, and also simultaneously makes it a good mythological motif. This seems to be only 1 in 100 or 1 in 1000 surprising to me.
Here’s a bit of what it means to be “real” in Peterson’s pragmatic sense, expanding on another comment:
Atoms are real. Numbers are real. You might call numbers a “useful metaphor,” but numbers are more real than atoms. Part of what I mean by this is: I would be more surprised if the universe didn’t obey simple mathematical laws than if it were not made out of atoms. Another part of what I mean is: if I had to choose between knowing about numbers and knowing about atoms, knowing about numbers would be more powerful in guiding me through life. And this is the pragmatic definition of truth.
At some point in the distant past people believed the imaginary unit i was not a “real” number. At first, it was introduced as a “useful shorthand” for a calculation made purely in the reals. People noticed, for example, that the easiest way to solve cubic equations like x3+x+1=0 was to go through these imaginary numbers, even if the answer you end up with is real.
Eventually, the concept of i became so essential and simplified so many other things (e.g. every polynomial has a root) that its existence graduated from “useful metaphor” to “true.” It led to ridiculous things like taking complex exponents, but somehow phenomena like eπi=−1 made so much internal sense that the best explanation was that that i is as real as anything can be. Or if it isn’t, we might as well treat it as if it is. There is some underlying metaphorical reality higher than technical truth. I could explain exactly what set of physical patterns eπi=−1 is a shorthand for, but that would be putting the cart before the horse.
Metaphorical truth is the idea that the patterns in human behavior recorded in our mythological stories are more true than literal truth, in the same way that eπi=−1 is more true than “the world is made out of atoms.” This is the right way to overload the concept of truth for stories.
I still don’t quite grasp the DNA point, even after multiple reads—how would compressibility make it show up in mythos? I can’t find any non-reddit / youtube source on his statements (Freedom is keeping a patient eye on my browsing habits, as always).
I don’t disagree that mathematical truth is, in a certain sense, “higher” than other truths.
I’d just like to point out that if I could consistently steelman Eliezer’s posts, I’d probably be smarter and more rational than he (and no, I cannot do this).
For the DNA point, I’m drawing on some mathematical intuition. Here are two examples:
What if I told you that ancient Egyptian civilizations had depictions of the hyperbolic cosineex+e−x2 even though they never came close to discovering the constant e? Well, the hyperbolic cosine is also called the catenary, which is the not-quite-parabola shape that all uniformly-weighted chains make if held from their two ends. So of course this shape was everywhere!
What if I told you that a physicist who had never studied prime number theory discovered the distribution of the zeros of the Riemann zeta function (that had escaped the attention of number theorists)? It turns out that this is basically how random matrix theory was discovered by Dyson and Montgomery.
The point is that mathematically interesting structures show up in not-obviously-connected ways. Now if I could tell you what exactly the structural property of DNA was, then I would actually believe Peterson’s claim about it, which I don’t. But at least a start to this question is: suppose a thousand genetic life forms evolved independently on a thousand planets. How many mathematically different information storage structures like the double helix would appear? Probably not more than 10, right? Most likely there’s something canonically robust and efficient about the way information is packed into DNA molecules.
Re: steelmanning. Really what I’m doing is translating Peterson into language more palatable to rationalists. Perhaps you could call this steelmanning.
How many mathematically different information storage structures like the double helix would appear? Probably not more than 10, right? Most likely there’s something canonically robust and efficient about the way information is packed into DNA molecules.
I’d agree with this claim, but it feels pretty anthropically-true to me. If it weren’t the case, we wouldn’t be able to exist.
Really what I’m doing is translating Peterson into language more palatable to rationalists. Perhaps you could call this steelmanning.
Once understood, chains of reasoning should (ideally) be accepted or rejected regardless of their window dressing. I may be turned off by what he says due to his mannerisms / vocabulary, but once I take the time to really understand what he’s claiming… If I still find his argumentation lacking, then rephrasing it in an actually-more-defensible way is steelmanning. I haven’t taken that time (and can’t really, at the moment), but I suspect if I did, I’d still conclude that there is no steel strong enough to construct a beisutsukai out of someone who believes in god.
The crux of the matter is that he believes in God then? I’ll also let him speak for himself, but as far as I can tell he doesn’t by your definition of believe in God. Furthermore, I’ve always been an atheist and not changed any object-level beliefs on that front since I can remember, but I think with respect to Peterson’s definitions I also believe in God.
It’s also not literally inconceivable that someone in Egypt formulated and technically solved the alignment problem, but I wouldn’t put odds on that of more than 1×10−7. Yes, I am prepared to make a million statements with that confidence and not expect to lose money to the gods of probability.
This seems motte-and-bailey. If that’s what he means, shouldn’t he just advance “terminal values are necessary to solve Moore’s open question”?
I feel like throughout the comments defending Peterson, the bottom line has been written first and everything else is being justified post facto.
When I say inconceivable I don’t mean literally inconceivable. People have done some pretty absurd things in the past. What is your subjective probability that the most prolific mathematician of all time did half of his most productive work after going blind in both eyes?
I can’t speak for others, but I have spent hundreds of hours thinking about Peterson’s ideas and formulating which parts I agree with and why, including almost every argument that has been put forth so far. Me going back retrospectively and extracting the reasons I made each decision to believe what I believe will look from the outside just like “writing down the bottom line and justifying things post facto.” In general I don’t think the bottom line fallacy is one that you’re supposed to use on other people’s reasoning.
That’s surprising but not that surprising: Milton wrote much of his best poetry while blind, and Beethoven was famously deaf. Conversely, I cannot think of a single unambiguous example of a mythological motif encoding a non-obvious scientific truth (such as that nothing can go faster than light, or that all species evolved from a single-celled organism, or that the stars are trillions and trillions of miles away), so I think this is very very unlikely.
I wasn’t saying that unusual things can’t happen. I should have made myself clearer—what I was getting at was with respect to claims that ancient societies managed to spontaneously derive properties of things they were, in fact, literally incapable of observing. That smells like a second law of thermodynamics-violating information gain to me.
The assertion I’m making is not that Peterson is bad, or that he never has amazing insights, etc. My point is purely with respect to putting him on Eliezer’s level of truth-seeking and technical rationality. Having been wrong about things does not forever disbar you from being a beisutsukai master. However, if one is wrong about important things, becomes aware of the methods of rationality (as I imagine he has), thinks carefully, and still retains their implausible beliefs—that should be enough to indicate they aren’t yet on the level required.
On the other hand, I notice I am confused and that I am disagreeing with people whom I respect very much. I’m happy to update on any new information, but I have a hard time seeing how I could update very far on this particular claim, given that he is indeed quite religious.
Thank you for being charitable. =)
Regarding the DNA claim: I think what I’m saying is much weaker than what you think I’m saying.
e.g. ancient people discovered how to store information sequentially in a book. DNA stores information sequentially. This is not surprising. Why would it be inconceivable that the double helix structure is not uniquely weird about DNA?
My steelman of Peterson’s claim about DNA is not that ancient people knew what DNA was or were making any attempt to map it, but that there might be some underlying mathematical reason (such as high compressibility) that the double helix structure is amenable to information storage, and also simultaneously makes it a good mythological motif. This seems to be only 1 in 100 or 1 in 1000 surprising to me.
Here’s a bit of what it means to be “real” in Peterson’s pragmatic sense, expanding on another comment:
Atoms are real. Numbers are real. You might call numbers a “useful metaphor,” but numbers are more real than atoms. Part of what I mean by this is: I would be more surprised if the universe didn’t obey simple mathematical laws than if it were not made out of atoms. Another part of what I mean is: if I had to choose between knowing about numbers and knowing about atoms, knowing about numbers would be more powerful in guiding me through life. And this is the pragmatic definition of truth.
At some point in the distant past people believed the imaginary unit i was not a “real” number. At first, it was introduced as a “useful shorthand” for a calculation made purely in the reals. People noticed, for example, that the easiest way to solve cubic equations like x3+x+1=0 was to go through these imaginary numbers, even if the answer you end up with is real.
Eventually, the concept of i became so essential and simplified so many other things (e.g. every polynomial has a root) that its existence graduated from “useful metaphor” to “true.” It led to ridiculous things like taking complex exponents, but somehow phenomena like eπi=−1 made so much internal sense that the best explanation was that that i is as real as anything can be. Or if it isn’t, we might as well treat it as if it is. There is some underlying metaphorical reality higher than technical truth. I could explain exactly what set of physical patterns eπi=−1 is a shorthand for, but that would be putting the cart before the horse.
Metaphorical truth is the idea that the patterns in human behavior recorded in our mythological stories are more true than literal truth, in the same way that eπi=−1 is more true than “the world is made out of atoms.” This is the right way to overload the concept of truth for stories.
I still don’t quite grasp the DNA point, even after multiple reads—how would compressibility make it show up in mythos? I can’t find any non-reddit / youtube source on his statements (Freedom is keeping a patient eye on my browsing habits, as always).
I don’t disagree that mathematical truth is, in a certain sense, “higher” than other truths.
I’d just like to point out that if I could consistently steelman Eliezer’s posts, I’d probably be smarter and more rational than he (and no, I cannot do this).
For the DNA point, I’m drawing on some mathematical intuition. Here are two examples:
What if I told you that ancient Egyptian civilizations had depictions of the hyperbolic cosine ex+e−x2 even though they never came close to discovering the constant e? Well, the hyperbolic cosine is also called the catenary, which is the not-quite-parabola shape that all uniformly-weighted chains make if held from their two ends. So of course this shape was everywhere!
What if I told you that a physicist who had never studied prime number theory discovered the distribution of the zeros of the Riemann zeta function (that had escaped the attention of number theorists)? It turns out that this is basically how random matrix theory was discovered by Dyson and Montgomery.
The point is that mathematically interesting structures show up in not-obviously-connected ways. Now if I could tell you what exactly the structural property of DNA was, then I would actually believe Peterson’s claim about it, which I don’t. But at least a start to this question is: suppose a thousand genetic life forms evolved independently on a thousand planets. How many mathematically different information storage structures like the double helix would appear? Probably not more than 10, right? Most likely there’s something canonically robust and efficient about the way information is packed into DNA molecules.
Re: steelmanning. Really what I’m doing is translating Peterson into language more palatable to rationalists. Perhaps you could call this steelmanning.
I’d agree with this claim, but it feels pretty anthropically-true to me. If it weren’t the case, we wouldn’t be able to exist.
Once understood, chains of reasoning should (ideally) be accepted or rejected regardless of their window dressing. I may be turned off by what he says due to his mannerisms / vocabulary, but once I take the time to really understand what he’s claiming… If I still find his argumentation lacking, then rephrasing it in an actually-more-defensible way is steelmanning. I haven’t taken that time (and can’t really, at the moment), but I suspect if I did, I’d still conclude that there is no steel strong enough to construct a beisutsukai out of someone who believes in god.
The crux of the matter is that he believes in God then? I’ll also let him speak for himself, but as far as I can tell he doesn’t by your definition of believe in God. Furthermore, I’ve always been an atheist and not changed any object-level beliefs on that front since I can remember, but I think with respect to Peterson’s definitions I also believe in God.