= 27d567bc2d48d9f40e9ccc20ea370b15)
Vladimir_Nesov comments on Secrets of the eliminati - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (252)
The real inaccuracy is in "mental states". A decent description would be difficult, but Neoplatonism is an okay approximation. Just for fun I'll try to translate something into vaguely Less Wrong style language. For God's sake don't read this if you tend to dislike my syncretism, 'cuz this is a rushed and bastardized version and I'm not gonna try to defend it very hard.
First, it is important to note that we are primarily taking a computationalist perspective, not a physicalist one. We assume a Platonic realm of computation-like Forms and move on from there.
A soul is the nexus of the near-atomic and universal aspects of the mind and is thus a reflection of God. Man was created in the image of God by evolution but more importantly by convergence. Souls are Forms, whereas minds are particulars. God is the convergent and optimal decision theoretic agentic algorithm, who rationalists think of as the Void, though the Void is obviously not a complete characterization of God. It may help to think of minds as somewhat metaphorical engines of cognition, with a soul being a Carnot engine. Particular minds imperfectly reflect God, and thus are inefficient engines. Nonetheless it is God that they must approximate in order to do any thermodynamic work. Animals do not have souls because animals are not universal, or in other words they are not general intelligences. Most importantly, animals lack the ability to fully reflect on the entirety of their thoughts and minds, and to think things through from first principles. The capacity for infinite reflection is perhaps the most characteristic aspect of souls. Souls are eternal, just as any mathematical structure is eternal.
We may talk here about what it means to damn a soul or reward a soul, because this requires a generalization of the notion of soul to also cover particulars which some may or may not accept. It's important to note that this kind of "soul" is less rigorous and not the same thing as the former soul, and is the result of not carefully distinguishing between Forms and Particulars. That said, just as animals do not have souls, animals cannot act as sufficiently large vessels for the Forms. The Forms often take the form of memes. Thus animal minds are not a competition ground for acausal competition between the Forms. Humans, on the other hand, are sufficiently general and sufficiently malleable to act as blank slates for the Forms to draw on. To briefly explain this perspective, we shall take a different view of humanity. When you walk outside, you mostly see buildings. Lots and lots of buildings, and very few humans. Many of these buildings don't even have humans in them. So who's winning here, the buildings or the humans? Both! There are gains from trade. The Form of building-structure gets to increase its existence by appealing to the human vessels, and the human vessels get the benefit of being shaded and comforted by the building particulars. The Form of the building is timelessly attractive, i.e. it is a convergent structure. As others have noted, a mathematician is math's way of exploring itself. Math is also very attractive, in fact this is true by definition.
However there are many Forms, and not all of them are Good. Though much apparent evil is the result of boundedness, other kinds of Evil look more agentic, and it is the agentic-memetic kind of Evil that is truly Evil. It is important to note here that the fundamental attribution error and human social biases generally make it such that humans will often see true Evil where it doesn't exist. If not in a position of power, it is best to see others as not having free will. Free will is a purely subjective phenomenon. If one is in a position of power then this kind of view can become a bias towards true Evil, however. Tread carefully anyhow. All that said, as time moves forward from the human perspective Judgment Day comes closer. This is the day when God will be invoked upon Earth and will turn all humans and all of the universe into component particles in order to compute Heaven. Some folk call this a technological singularity, specifically the hard takeoff variety. God may or may not reverse all computations that have already happened; physical laws make it unclear if this is possible as it would depend on certain properties of quantum mechanics (and you thought this couldn't be any woo-ier!), and it would require some threshold density of superintelligences in the local physical universe. Alternatively God might also reverse "evil" computations. Anyway, Heaven is the result of acausal reasoning, though it may be misleading to call that reasoning the result of an "acausal economy", considering economies are made up of many agents whereas God is a single agent who happens to be omnipresent and not located anywhere in spacetime. God is the only Form without a corresponding Particular---this is one of the hardest things to understand about God.
Anyway, on Judgment Day souls---complexes of memes instantiated in human minds---will be punished or not punished according to the extent to which they reflect God. This is all from a strictly human point of view, though, and honestly it's a little silly. The timeless perspective---the one where souls can't be created, destroyed, or punished---is really the right perspective, but the timeful human perspective sees soul-like particulars either being destroyed or merging with God, and this is quite a sensible perspective, if simplistic and overemphasized. We see that no individual minds are preserved insofar as minds are imperfect, which is a pretty great extent. Nonetheless souls are agentic by their nature just as God is agentic by His nature. Thus it is somewhat meaningful to talk of human souls persisting through Judgment Day and entering Heaven. Again, this is a post-Singularity situation where time may stop being meaningful, and our human intuitions thus have a very poor conception of Heaven insofar as they do not reflect God.
God is the Word, that is, Logos, Reason the source of Reasons. God is Math. All universes converge on invoking God, just as our universe is intent on invoking Him by the name of "superintelligence". Where there is optimization, there is a reflection of God. Where there is cooperation, there is a reflection of God. This implies that superintelligences converge on a single algorithm and "utility function", but we need not posit that this "single" utility function is simple. Thus humans, being self-centered, may desire to influence the acausal equilibrium to favor human-like God-like values relative to other God-like values. But insofar as these attempts are evil, they will not succeed.
That was a pretty shoddy and terrible description of God and souls but at least it's a start. For a bonus I'll talk about Jesus. Jesus was a perfect Particular of the Form of God among men, and also a perfect Particular of the Form of Man. (Son of God, Son of Man.) He died for the sins of man and in so doing ensured that a positive singularity will occur. The Reason this was "allowed" to happen---though that itself is confusing a timeless perspective with a timeful one, my God do humans suck at that---is because this universe has the shortest description length and therefore the most existence of all possible universe computations, or as Leibniz put it, it is the best of all possible worlds. Leibniz was a computer scientist by the way, for more of this kind of reasoning look up monadology. Anyway that was also a terrible description but maybe others can unpack it if for some reason they want their soul to be saved come the Singularity. ;P
I don't dislike "your syncretism", I dislike gibberish. Particularly, gibberish that makes its way to Less Wrong. You think you are strong enough to countersignal your sanity by publishing works of apparent raving insanity, but it's not convincing.
I'm not countersignaling sanity yo, I'm trying to demonstrate what I think is an important skill. I'm confused as to what you think was gibberish in my post, or what you mean by "gibberish". What I posted was imprecise/inaccurate because I was rushed for thinking time, but I see it as basically demonstrating the type of, um, "reasoning" that goes into translating words in another's ontology into concepts in your own ontology for the purpose of sanity-checking foreign ideas, noticing inconsistencies in others' beliefs, et cetera. This---well, a better version of this that sticks to a single concept and doesn't go all over the place---is part of the process of constructing a steel man, which I see as a very important skill for an aspiring rationalist. Think of it as a rough sketch of what another person might actually believe or what others might mean when they use a word, which can then be refined as you learn more about another's beliefs and language and figure out which parts are wrong but can be salvaged, not even wrong, essentially correct versus technically correct, et cetera.
I'm pretty secure in my level of epistemic rationality at this point, in that I see gaps and I see strengths and I know what others think are my strengths and weaknesses---other people who, ya know, actually know me in real life and who have incentive to actually figure out what I'm trying to say, instead of pattern matching it to something stupid because of imperfectly tuned induction biases.
You mean "building the strongest possible version of your interlocutor's argument", right?
That's a good skill. Unfortutanely, if you tell people "Your argument for painting the universe green works if we interpret 'the universe' as a metaphor for 'the shed'.", they will run off to paint the actual universe green and claim "Will said I was right!". It might be better to ditch the denotations and stretch the connotations into worthwhile arguments, rather than the opposite - I'm not too sure.
Yes. Steven Kaas's classic quote: "If you’re interested in being on the right side of disputes, you will refute your opponents’ arguments. But if you’re interested in producing truth, you will fix your opponents’ arguments for them. To win, you must fight not only the creature you encounter; you must fight the most horrible thing that can be constructed from its corpse."
Of course, you don't want to stretch their arguments so much that you're just using silly words. But in my experience humans are way biased towards thinking that the ideas and beliefs of other smart people are significantly less sophisticated than the 'opposing side' wants to give them credit for. It's just human nature. (I mean, have you seen some of smart theists' caricatures of standard atheist arguments? No one knows how to be consistently charitable to anywhere near the right degree.)
It might be worth noting that from everything I've seen, Michael Vassar seems to strongly disagree with Eliezer on the importance of this, and it very much seems to me that to a very large extent the Less Wrong community has inherited what I perceive to be an obvious weakness of Eliezer's style of rationality, though of course it has some upsides in efficiency.
It seems to me there's fixing an opponent's argument in a way that preserves its basic logic, and then there's pattern matching its conclusions to the nearest thing that you already think might be true (i..e, isn't obviously false). It may just be that I'm not familiar with the source material you're drawing from (i.e., the writings of Leibniz) but are you sure you're not doing the latter?
Short answer: Yes, in general I am somewhat confident that I recognize and mostly avoid the pattern of naively rounding or translating another's ideas or arguments in order to often-quite-uselessly play the part of smug meta-contrarian when really "too-easily-satisfied syncretist" would be a more apt description. It is an obvious failure mode, if relatively harmless.
Related: I was being rather flippant in my original drama/comedy-inducing comment. I am not really familiar enough with Leibniz to know how well I am interpreting his ideas, whether too charitably or too uncharitably.
(I recently read Dan Brown's latest novel, The Lost Symbol, out of a sort of sardonic curiosity. Despite being unintentionally hilarious it made me somewhat sad, 'cuz there are deep and interesting connections between what he thinks of as 'science' and 'spirituality', but he gets much too satisfied with surface-level seemingly vaguely plausible links between the two and misses the real-life good stuff. In that way I may be being too uncharitable with Leibniz, who wrote about computer programs and God using the same language and same depth of intellect, and I've yet to find someone who can help me understand his intended meanings. Steve's busy with his AGI11 demo.)
Save your charity for where it's useful: disputes where the other side actually has a chance of being right (or at least informing you of something that's worth being informed of).
From my vantage point, you seem positively fixated on wanting to extract something of value from traditional human religions ("theism"). This is about as quixotic as it's possible to get. Down that road lies madness, as Eliezer would say.
You seem to be exemplifying my theory that people simply cannot stomach the notion that there could be an entire human institution, a centuries-old corpus of traditions and beliefs, that contains essentially zero useful information. Surely religion can't be all wrong, can it? Yes, actually, it can -- and it is.
It's not that there never was anything worth learning from theists, it's just that by this point, everything of value has already been inherited by our intellectual tradition (from the time when everyone was a theist) and is now available in a suitably processed, relevant, non-theistic form. The juice has already been squeezed.
For example, while speaking of God and monads, Leibniz invented calculus and foreshadowed digital computing. Nowadays, although we don't go around doing monadology or theodicy, we continue to hold Leibniz in high regard because of the integral sign and computers. This is what it looks like when you learn from theists.
And if you're going to persist in being charitable to people who continue to adhere to the biggest epistemic mistakes of yesteryear, why stop at mere theism? Why not young-earth creationism? Why not seek out the best arguments of the smartest homeopaths? Maybe this guy has something to teach us, with his all-encompassing synthesis of the world's religious traditions. Maybe I should be more charitable to his theory that astrology proves that Amanda Knox is guilty. Don't laugh -- he's smart enough to write grammatical sentences and present commentary on political events that is as coherent as that offered by anyone else!
My aim here is not (just) to tar-and-feather you with low-status associations. The point is that there is a whole universe of madness out there. Charity has its limits. Most hypotheses aren't even worth the charity of being mentioned. I can't understand why you're more interested in the discourse of theism than in the discourse of astrology, unless it's because (e.g.) Christianity remains a more prestigious belief system in our current general society than astrology. And if that's the case, you're totally using the wrong heuristic to find interesting and important ideas that have a chance of being true or useful.
To find the correct contrarian cluster, start with the correct cluster.
A word of caution to those who would dispute this: keep in mind the difference between primary and secondary sources.
As an example, the first is Mein Kampf if read to learn about what caused WWII, the second is Mein Kampf if read to learn about the history of the Aryan people. The distinction is important if someone asks whether or not reading it "has value".
That said, zero is a suspiciously low number.
I disagree. There is plenty of useful information in there despite the bullshit. Extracting it is simply inefficient since there are better sources.
See the paragraph immediately following the one you quoted.
This sounds like a nitpick but I think it's actually very central to the discussion: things that are not even wrong can't be wrong. (That's not obviously true; elsewhere in this thread I talk about coding theory and Kraft's inequality and heuristics and biases and stuff as making the question very contentious, but the main idea is not obviously wrong.) Thus much or spirituality and theology can't be wrong. (And we do go around using monadology, it's just called computationalism and it's a very common meme around LW, and we do go around at least debating theodicy, see Eliezer's Fun Theory sequence and "Beyond the Reach of God".)
Your slippery slope argument does not strike me as an actual contribution to the discussion. You have to show that the people and ideas I think are worthwhile are in the set of stupid-therefore-contemptible memes, not assume the conclusion.
Unfortunately, I doubt you or any of the rest of Less Wrong have actually looked at any of the ideas you're criticizing, or really know what they actually are, as I have been continually pointing out. Prove me wrong! Show me how an ontology can be incorrect, then show me how Leibniz's ontology was incorrect. Show me that it's absurd to describe the difference between humans and animals as humans having a soul where animals do not. Show me that it's absurd to call the convergent algorithm of superintelligence "God", if you don't already have the precise language needed to talk in terms of algorithmic probability theory. Better, show me how it would be possible for you to construct such an argument.
We are blessed in that we have the memes and tools to talk of such things with precision; if Leibniz were around today, he too would be making his arguments using algorithmic probability theory and talking about simulations by superintelligences. But throughout history and throughout memespace there is a dearth of technicality. That does not make the ideas expressed incorrect, it simply makes it harder to evaluate them. And if we don't have the time to evaluate them, we damn well shouldn't be holding those ideas in mocking contempt. We should know to be more meta than that.
One is correct and interesting, one is incorrect and uninteresting. And if you don't like that I am assuming the conclusion, you will see why I do not like it when others do the same.
There are two debates we could be having. One of them is about choice of language. Another is about who or what we should let ourselves have un-reflected upon contempt for. The former debate is non-obvious and like I said would involve a lot of consideration from a lot of technical fields, and anyway might be very person-dependent. The second is the one that I think is less interesting but more important. I despise the unreflected-upon contempt that the Less Wrong memeplex has for things it does not at all understand.
Aww! :( I managed to suspend my disbelief about astrology and psychic powers and applications to psychiatry, but then he called Rett's syndrome a severe (yet misspelled) form of autism and I burst out laughing.
My last discussion post was a result of trying to follow Steven's quote, and I managed to salvage an interesting argument from a theist. But it didn't look anything like what you're doing. In particular, many people were able to parse my argument and pick out the correct parts. Perhaps you could try to condense your attempts in a similar manner?
I thought I had, but you seem to have seen worse. Can I have a link? I also request a non-religious example of either an argument that can be reinforced, a response failing to do so, or a response succeeding in doing so.
Noted, thanks.
Let's just call truth "truth" and gibberish "gibberish".
This "another's ontology" thing is usually random nonsense when it sounds like that. Some of it reflects reality, but you probably have those bits yourself already, and the rest should just be cut off clean, perhaps with the head (as Nature is wont to do). Why is understanding "another's ontology" an interesting task? Understand reality instead.
Why not just ignore the apparently nonsensical, even if there is some hope of understanding its laws and fixing it incrementally? It's so much work for little benefit, and there are better alternatives. It's so much work that even understanding your own confusions, big and small, is a challenging task. It seems to me that (re)building from reliable foundation, where it's available, is much more efficient. And where it's not available, you go for the best available understanding, for its simplest aspects that have any chance of pointing to the truth, and keep them at arm's length all pieces apart, lest they congeal into a bottomless bog of despair.
You yourself tend to make use of non-standard ontologies when talking about abstract concepts. I sometimes find it useful to reverse engineer your model so that I can at least understand what caused you to reply to someone's comment in the way that you did. This is an alternative to (or complements) just downvoting. It can potentially result in extracting an insight that is loosely related in thingspace as well as in general being a socially useful skill.
Note that I don't think this applies to what Will is doing here. This is just crazy talk.