|
Basically they assume that the AI entity will achieve both their specific idea of intelligence, which happens to be represented by a number, and awareness of that number and all of its connotations, and then it will rationally make the decision to do whatever it takes to increase that number because that's what getting smarter means to them.
|
#
?
May 16, 2018 22:54
|
|
|
|
| # ? May 9, 2024 00:29 |
|
|
Jon Joe posted:Specifically AI in this case, but before he was an AI guy, Yud was a biotech guy. I somehow missed this in everything. What did he do there? (Bloviated, obv, but what though.) Jon Joe posted:They are not trained to ask how to achieve those qualities in reality, they just assume that the AI will have those qualities, because those are the qualities it needs to have to achieve their immortality. This is the absolutely key point: Yudkowsky started at "I want to live forever, as an emulation is fine" and worked backwards from there. Oh by the way - Musk and Grimes hooking up over the Basilisk? I only just realised today that Elon Musk is named directly in the original Basilisk post.
|
|
#
?
May 16, 2018 22:55
|
|
|
divabot posted:I somehow missed this in everything. What did he do there? (Bloviated, obv, but what though.) I remember reading somewhere once that he was involved in waxing about biotech to little success before he hit on the gold nugget of his rationality/AI combo, in that ‘guy will talk about anything he thinks people will listen to’ sort of way. I’m trying to remember the source of where I read that but I’m pulling a blank right now, sorry.
|
|
#
?
May 16, 2018 23:53
|
|
|
Somfin posted:So the question I have, as always, is "why would it 'wake up'?" What do you even mean by “wake up” though? Are you asking about whether it would exhibit human level intelligence, whether it would exhibit accelerated growth in intelligence, or whether it would become conscious? The theory behind the first is that there’s not necessarily any such thing as “waking up,” just being able to problem-solve, reason, and acquire and integrate new information in a variety of ways. Over a decade ago now, you could give a running Cyc system a newspaper article, it would ask questions about the content of the article, and after you answered them it would integrate that information into its knowledge base. Is that intelligent? It could also be asked to evaluate situations (hooked up to custom inputs) and make plans to achieve goals based on that, and come up with reasonable evaluations and proposals. Is that intelligent? Or is it just a fancy expert system? What even intelligence? The theory behind accelerating growth in intelligence is just that brains are actually really slow, and if you can achieve intelligence in the first place, running it ever faster with ever more storage is “just” a matter of Moore’s Law. This isn’t as clear cut as it sounds though because we are hitting some walls and at least some problems may not be as embarrassingly parallel as hoped. (Neural simulation may scale almost linearly, for example, but inference engines may not.) And of course this begs the question of what superintelligence actually means: Is it just the ability to deal with more information faster, or do we actually mean something else by it? And how does it relate to what intelligence even is? As for consciousness, we as a species don’t have any idea what it is, and theories about it are really more a matter of philosophy. Of course the MIRI/Yudkowsky types all love to file it in the same bucket as intelligence, but they have no idea how any of the stuff we’ve talked about actually work and I suspect their understanding of philosophy is just as well grounded as their understanding of software, which is to say almost entirely nonexistent.
|
|
#
?
May 17, 2018 00:22
|
|
|
I thought LessWrong hadn't got around to covering "A Girl Corrupted" - they did, and in finest LW style
|
|
#
?
May 17, 2018 13:16
|
|
|
Somfin posted:Basically they assume that the AI entity will achieve both their specific idea of intelligence, which happens to be represented by a number, and awareness of that number and all of its connotations, and then it will rationally make the decision to do whatever it takes to increase that number because that's what getting smarter means to them. 🤔👍 normal brain: number go up? 😘👌 planet brain: make number go up! 🤯👏 galaxy brain: IQ... but with bitcoin
|
|
#
?
May 17, 2018 14:23
|
|
|
eschaton posted:What do you even mean by “wake up” though? Are you asking about whether it would exhibit human level intelligence, whether it would exhibit accelerated growth in intelligence, or whether it would become conscious? I was referring to your own terminology there, re 'wake up'. And, yeah. Comforting to know we're on the same page about intelligence / consciousness. I seriously think that 'the AI realises improving its own IQ is optimal for solving its problem' is the jump that Yud and his acolytes are making, and there's so deeply many layered wrong assumptions built into that concept that it's the equivalent of "a witch down the street did it" as an answer to science riddles. Like, it's a sentence where each individual word can be pulled out and analysed for the massive assumptions it's making.
|
|
#
?
May 17, 2018 15:18
|
|
|
Somfin posted:"stuff I don't understand but Yud probably does because he's literally paid to think." 'He's paid to think about a very specific (and stupid) thing that he has convinced other people to pay him to think about' is a fairly succinct summation of Yud's entire life.
|
|
#
?
May 17, 2018 15:40
|
|
|
Somfin posted:I seriously think that 'the AI realises improving its own IQ is optimal for solving its problem' is the jump that Yud and his acolytes are making, and there's so deeply many layered wrong assumptions built into that concept that it's the equivalent of "a witch down the street did it" as an answer to science riddles. Like, it's a sentence where each individual word can be pulled out and analysed for the massive assumptions it's making. Is there a concise quote like that that I could pull to pieces in the manner you describe? One that isn't just a passing comment but the core of Yud's argument.
|
|
#
?
May 17, 2018 16:43
|
|
|
Somfin posted:I seriously think that 'the AI realises improving its own IQ is optimal for solving its problem' is the jump that Yud and his acolytes are making, and there's so deeply many layered wrong assumptions built into that concept that it's the equivalent of "a witch down the street did it" as an answer to science riddles. Like, it's a sentence where each individual word can be pulled out and analysed for the massive assumptions it's making. I am on my very early stages of my how-to-computers journey but the one thing that baffled me hard (among many others) when reading about AI-go-foom was that they ignored something that, I assume, is in every intro to computer science book: a "sufficiently smart AI" would not only solve computability theory completely, but also solve in such a way that would hack the universe. What is not computable could be turned into such by sufficient computing power through the magic words of "assuming a powerful enough machine" in such hypotheses, leading to the objection of "but man this isn't a matter of processing" causing them to reply only louder "with ENOOOOUGH power..." whaaaaaa But the layman doesn't know any of that, and it is why it is important to get as much information about how that is bullshit around. Yudkowsky might have completely been convinced by his own fantasy, but I doubt at all that we don't already have an actual criminal sleazebag with the skills to con people hard using that information going around. therefore,  because we need all the ammo we can get
|
|
#
?
May 17, 2018 17:08
|
|
|
https://twitter.com/Outsideness/status/996269237458034688 These people are insane. (That's Nick Land and Twitter is now basically the only way he interacts with the outside world.) https://twitter.com/Outsideness/status/996447979832725504 In case you thought I was exaggerating.
|
|
#
?
May 17, 2018 18:08
|
|
|
What does that even - ... does he think Twitter is a Platonic Form? I'm so confused.
|
|
#
?
May 17, 2018 18:55
|
|
|
ol qwerty bastard posted:What does that even - the followups to those tweets show where the purple prose wing of neoreaction ended up
|
|
#
?
May 17, 2018 20:49
|
|
|
ol qwerty bastard posted:What does that even - Kant famously claimed that you could never know a thing "in itself," but only your perceptions of those things, interpreted through a framework you will never be able to view from the outside. Not-Twitter works the same way.
|
|
#
?
May 17, 2018 21:12
|
|
|
What is he even talking about here?
|
|
|
#
?
May 19, 2018 02:14
|
|
|
He seems to think that a major leftist issue in America today is defending the right of people to own pit bulls. Which I guess makes sense if you only ever encounter leftists on Twitter, and even then only the ones your friends retweet to make fun of.
|
|
#
?
May 19, 2018 02:19
|
|
|
Also they're not even Leftists they're just /pol dudes pretending to be to own the libs. Land is regarded as the most intellectual of the DE because he was an academic, even though he never recovered from his psychotic break.
|
|
#
?
May 19, 2018 04:57
|
|
|
When you’re done with people racism, move on to dog racism
|
|
#
?
May 19, 2018 12:30
|
|
|
Neon Noodle posted:When you’re done with people racism, move on to dog racism Pitbulls are mostly owned by The Blacks according to these guys, so they're not even done with people racism.
|
|
#
?
May 19, 2018 14:29
|
|
|
The front page is still relevant sometimes.
|
|
#
?
May 20, 2018 05:02
|
|
|
There’s a front page?
|
|
#
?
May 20, 2018 09:59
|
|
|
SolTerrasa posted:He seems to think that a major leftist issue in America today is defending the right of people to own pit bulls. Which I guess makes sense if you only ever encounter leftists on Twitter, and even then only the ones your friends retweet to make fun of. This comes back to a recurring theme with the right's inability to understand the left. The left is fragmented for many reasons, and one of those reasons is very justified; the issues that different left wingers care about are not necessarily related to each other. Crunchy naturalistic hippies are often critical of capitalism, but that doesn't mean they're communists or that they agree with communists. Same with the neoliberals who often lean left on social issues but are sympathetic to capitalism. And the anarchists who are also critical of capitalism but reject the top-down approach of communists and some environmentalists. And on and on. In principle, the right wing ought to be the same. Religious conservatives should not necessarily be fascist sympathizers, and the fascist sympathizers should not necessarily be the same as the survivalists and militiamen. But in the US, and often elsewhere, culture and media have brought those groups together. So the right sees those different ideas, observes that right wingers often belong to all of those groups, and assumes the left operates the same way. How does this relate to pit bulls? Pit bull awareness activism occasionally touches on some left wing issues. The biggest one I can think of is tenant rights. A lot of landlords have special rules barring tenants from owning pit bulls. And the general animal sympathy sometimes crosses over with environmentalism. Beyond that, it's a very localized issue. Pit bull sympathizers are not necessarily involved in other types of political activism. gently caress, you could make a right-wing argument for allowing people to own pit bulls just as well. Libertarianism's live-and-let-live philosophy is perfectly compatible with letting people own the dogs they want. I think this idiot is so caught up in his own contrarianism that he assumes pit bull advocates are leftists solely because they disagree with him, and not based on any real principles.
|
|
#
?
May 20, 2018 15:30
|
|
|
Caveatimperator posted:I think this idiot is so caught up in his own contrarianism that he assumes pit bull advocates are leftists solely because they disagree with him, and not based on any real principles. This is Nick Land, who is an Englishman living in Shanghai. And also bugfuck crazy. So it's not clear how he would have picked up on this. (In the UK , pit bulls are for the lower classes. And then there's dumb poor people who get rather smarter poor people selling them any loving dog and telling them it's a pit bull. One of the middle daughter's worse early choices of boyfriend fell for this one, and the dog in question was clearly an Alsatian even at the age of eight weeks ...)
|
|
#
?
May 20, 2018 16:35
|
|
|
Caveatimperator posted:This comes back to a recurring theme with the right's inability to understand the left. I mean there is a good case to be made of "pit bull activists" being on the left mostly, because there's a few extra weird layers of racism on top of pit bull hatred - they're perceived as the black people dogs and therefore more prone to violence, unlike the glorious aryan german shepherds who wouldn't hurt a fly and are just used as police dogs because they're Very Good Boys However this doesn't really... show up very often, and is definitely not a major issue for 99.5% of the left, so my guess is Nick Land read some tumblr where someone complained about this and decided it's what everyone left of Reagan must be rabidly foaming mad about because his entire existence is pathetically online
|
|
#
?
May 20, 2018 19:49
|
|
|
ate all the Oreos posted:I mean there is a good case to be made of "pit bull activists" being on the left mostly, because there's a few extra weird layers of racism on top of pit bull hatred - they're perceived as the black people dogs and therefore more prone to violence, unlike the glorious aryan german shepherds who wouldn't hurt a fly and are just used as police dogs because they're Very Good Boys I've never heard of this before. Is it a regional stereotype from where you live? I live in Los Angeles, and pits are perceived as macho people dogs first and macho Latino people dogs second, if at all.
|
|
#
?
May 20, 2018 20:59
|
|
|
Caveatimperator posted:I've never heard of this before. Is it a regional stereotype from where you live? I live in Los Angeles, and pits are perceived as macho people dogs first and macho Latino people dogs second, if at all. Florida, and it's definitely a perception across most of the south as far as I can tell  Like I've heard them called stuff like "n-er dogs" several times, it's real lovely
|
|
#
?
May 20, 2018 21:04
|
|
|
Pibbles tend to be popular in lower-class areas because they have the perception of being a tough guard dog, which is something good to have if you live where there might be threats from someone desperate for fast money and not worried about the law. In the South, that's a lot of heavily Black areas.
|
|
#
?
May 20, 2018 23:14
|
|
|
https://www.youtube.com/watch?v=X9_n8jakvWU&t=118s
The Vosgian Beast has a new favorite as of 00:20 on May 23, 2018 |
|
#
?
May 23, 2018 00:17
|
|
|
In Scotts newest pile of apologia, he defends the "intellectual dark web" and manages to in turns compare them to:
Not that hes saying conservatives are as oppressed as those people though! Just that hes trying to imply that they are somehow "silenced" in the same the same way as the people who were literally locked up or beaten for their views, because people yell at them on the internet for their dumb opinions.
|
|
#
?
May 24, 2018 03:53
|
|
|
The First Amendment guarantees a captive audience for me, and zero criticism from thee; all else is ＣＥＮＳＯＲＳＨＩＰ
|
|
#
?
May 24, 2018 04:06
|
|
|
In Scott's defense though, they should be beaten and locked up.
|
|
#
?
May 24, 2018 05:04
|
|
|
Grace Baiting posted:The First Amendment guarantees a captive audience for me, and zero criticism from thee; all else is ＣＥＮＳＯＲＳＨＩＰ
|
|
#
?
May 24, 2018 13:35
|
|
|
Dark Lord's Answer by Eliezer Yudkowsky Yudkowsky’s first attempt at the Light Novel form, from which he says he learnt many lessons for Red Tidday Up. “Content warning: sexual abuse, economics.” A didactic economics parable in novella form, on the value of monetary policy rather than a rigid gold standard - “So you see, Prince, that you’re not being told to steal from your country of Santal. Even if, to save it, you must transgress the righteous rules against usury and adulterated coinage.” The S&M slave girl is the author mouthpiece, and the viewpoint character is handed an idiot ball to make the mouthpiece look smart. Gratuitous S&M for flavouring, and to provide Yudkowsky’s favoured Philosophy Tough Guy ethics tests. A porned-up version of the Sequences. At least the plot is coherent and not stupid. Not worth 99p and I won’t be reading it a second time, but at least the author admits its deficiencies. the joys of doing your research, eh. An hour reading this thing and I get 0 useful words out of it. Oh well. 2 stars on Amazon, ‘cos it’s not quite 1-star bad.
|
|
#
?
May 26, 2018 15:09
|
|
|
Of all forms of fiction you could write in the style of, why Japanese light novels 
|
|
#
?
May 28, 2018 07:16
|
|
|
BattleMaster posted:Of all forms of fiction you could write in the style of, why Japanese light novels Can't say he doesn't know his audience.
|
|
#
?
May 28, 2018 07:29
|
|
|
BattleMaster posted:Of all forms of fiction you could write in the style of, why Japanese light novels Because he’s a weeaboo and thinks they’re art instead of pulp trash like western written forms.
|
|
#
?
May 28, 2018 08:16
|
|
|
Basilisk book: 320 words of a prospective intro written - I won't use it, but I was getting into flow. 1100 words of at least 2000 on Yudkowsky's predilection for torture porn in his philosophy, and that loving thing I mention above.
|
|
#
?
May 28, 2018 10:45
|
|
|
divabot posted:Basilisk book: 320 words of a prospective intro written - I won't use it, but I was getting into flow. 1100 words of at least 2000 on Yudkowsky's predilection for torture porn in his philosophy, and that loving thing I mention above. you might have already said, but are you thinking to self-publish this or do you have interest from a publisher?
|
|
#
?
May 28, 2018 16:43
|
|
|
prisoner of waffles posted:you might have already said, but are you thinking to self-publish this or do you have interest from a publisher? currently thinking self-publish, open to offers but not going out of my way for them mostly just writing and trying to work out what my marketing points are at this stage
|
|
#
?
May 28, 2018 18:48
|
|
|
|
| # ? May 9, 2024 00:29 |
|
|
BattleMaster posted:Of all forms of fiction you could write in the style of, why Japanese light novels
|
|
#
?
May 28, 2018 22:05
|
|
