Latest / The NanceCast / Brolo Episode cont.: Use of Language, Clarifications on the Notion of Intelligence, and some other stuff.
Transcript
- 0:01SPEAKER_00: You got something to say, Winston?
- 0:06SPEAKER_00: Oh, it's because I started singing.
- 0:08SPEAKER_00: And when I sing to my dogs often.
- 0:11SPEAKER_00: So he thought I was singing to him.
- 0:16SPEAKER_00: Nope.
- 0:18SPEAKER_00: Not singing to you right now, Winston.
- 0:19SPEAKER_00: Singing to myself.
- 0:20SPEAKER_00: Singing to my... Singing to myself.
- 0:24SPEAKER_00: I was dancing with myself.
- 0:28SPEAKER_00: What's up?
- 0:29SPEAKER_00: What's up, everybody?
- 0:31SPEAKER_00: What's happening, everyone?
- 0:32SPEAKER_00: Welcome to...
- 1:02SPEAKER_00: Welcome to the Nance cast.
- 1:11SPEAKER_00: Let me try it again.
- 1:11SPEAKER_00: Welcome to the Nance cast, everybody.
- 1:16SPEAKER_00: What is cracking?
- 1:18SPEAKER_00: We're fucked.
- 1:22SPEAKER_00: We're fucked.
- 1:23SPEAKER_00: We're fucked.
- 1:24SPEAKER_00: We're fucked.
- 1:26what's up man what's happening everyone hope you're doing well
- 1:30uh on this fine day
- 1:32um what's up welcome to another solo dolo brolo episode of the
- 1:41nance cast uh you
- 1:44know glad to have you here i'm happy to be here with you all
- 1:47dearly beloved my
- 1:50SPEAKER_00: fellow americans um
- 1:52SPEAKER_00: What's up?
- 1:53SPEAKER_00: Yeah, good to be here.
- 1:54SPEAKER_00: Hope you guys are all, you know, chilling out, maxing all that fun stuff.
- 2:03SPEAKER_00: So I recorded the other day, like an hour and a half.
- 2:08SPEAKER_00: And it was, I intended to record a follow-up to the last Brolo episode.
- 2:19SPEAKER_00: You know, talking about stuff and offering some
- 2:22some uh some commentary on my own commentary or whatever uh
- 2:28making a correction and
- 2:29kind of giving some clarification um to some of the things I
- 2:34said based on some of
- 2:35SPEAKER_00: the feedback I received um
- 2:39SPEAKER_00: So I recorded like an hour and a half and 45 minutes of it was me kind of
- 2:45SPEAKER_00: hysterically going on and on and on and on and on about language,
- 2:50SPEAKER_00: specifically about words,
- 2:53SPEAKER_00: even more specifically about names.
- 2:57SPEAKER_00: And I was trying to find an example to...
- 3:05SPEAKER_00: I guess prop up what I was saying,
- 3:09SPEAKER_00: but I was basically saying we all need to be a bit less precious,
- 3:12SPEAKER_00: a bit less protective and possessive over language in general,
- 3:17SPEAKER_00: over specific words,
- 3:19SPEAKER_00: and even over names.
- 3:22SPEAKER_00: Language does not belong to you, never has, never will.
- 3:28I don't think you can appropriately use language or be in
- 3:34language or be engaged in
- 3:35SPEAKER_00: language if you have this false belief that it does belong to you.
- 3:41And it does seem to be the case that there are a lot of people
- 3:44who maintain that
- 3:45belief that language is theirs and their own and it belongs to
- 3:50them and they are
- 3:50SPEAKER_00: the final arbiter.
- 3:53And I actually think that that can kind of lead to some pretty
- 3:58big problems
- 4:00SPEAKER_00: downstream of just,
- 4:01SPEAKER_00: you know,
- 4:01SPEAKER_00: the simple use of language.
- 4:03SPEAKER_00: And, you know, you think using language seems pretty innocuous, right?
- 4:06SPEAKER_00: That's what we do.
- 4:07SPEAKER_00: We are, you know, the being who speaks.
- 4:13SPEAKER_00: Among other things, we are the being who laughs.
- 4:17SPEAKER_00: We are also the being who is laughed at.
- 4:22SPEAKER_00: Shout out Bergzone.
- 4:24We are,
- 4:25you know,
- 4:25not just the creature who laughs,
- 4:27but we are the creature who is laughed at.
- 4:32SPEAKER_00: Anyway, it seems, you know, fairly standard.
- 4:39um innocuous pretty safe to use language for the most part um
- 4:45but i think when when
- 4:46you get this false sense of security and you think you do own
- 4:51the language or or
- 4:52you become so precious over it and so protective over it um it
- 4:58it actually can can
- 4:59SPEAKER_00: do some pretty damning things damning and damaging
- 5:03SPEAKER_00: And I was trying to talk about that,
- 5:04SPEAKER_00: and I was trying to use names as an example,
- 5:09SPEAKER_00: and I just kind of spun out of control.
- 5:10I went on a 45-minute rant about names and improper
- 5:16pronunciation of names and
- 5:19improper pronunciation of words in general and improper usage
- 5:22of words in general
- 5:26SPEAKER_00: and concepts like the shibboleth,
- 5:28SPEAKER_00: how
- 5:30SPEAKER_00: If you are unaware of a thing and you say or use a thing improperly,
- 5:40SPEAKER_00: that could serve as an exclusive mechanism or an inclusive mechanism,
- 5:46SPEAKER_00: like in-jokes,
- 5:47SPEAKER_00: a lot of internet memes.
- 5:49There's a lot of stuff that happens on the internet that's just
- 5:52fucking empty.
- 5:55And people either enjoy these things kind of like in a way that
- 6:03it becomes an in-joke.
- 6:04SPEAKER_00: They're like,
- 6:05SPEAKER_00: yeah,
- 6:05SPEAKER_00: we know it's bullshit or yeah,
- 6:06SPEAKER_00: we know it's empty or yeah,
- 6:07SPEAKER_00: we know it's actually like contra indicative of what we actually think.
- 6:11SPEAKER_00: And that's why we enjoy it because it serves as this shibboleth.
- 6:16SPEAKER_00: And it creates an in-group, out-group distinction.
- 6:19SPEAKER_00: Or there's people who just genuinely
- 6:24uncritically enjoy the empty mindless stuff and that too can
- 6:27serve as like an
- 6:28inclusive or exclusive mechanism like you are the type of
- 6:31person who does enjoy
- 6:33this genuinely uncritically or you are the type of person who
- 6:37doesn't enjoy this
- 6:39SPEAKER_00: and I don't know man like all these things
- 6:43SPEAKER_00: serve to reify and calcify.
- 6:47SPEAKER_00: Reification is a feature of language.
- 6:50SPEAKER_00: You can't get outside of it.
- 6:53SPEAKER_00: But then there's a secondary reification and a calcification.
- 6:59SPEAKER_00: And they just kind of become in-group, out-group bullshit.
- 7:03SPEAKER_00: And I think we all need to be on guard against that.
- 7:08SPEAKER_00: Both in the use of language, but also...
- 7:15SPEAKER_00: when having language used at us or in our direction.
- 7:20SPEAKER_00: So don't be so protective.
- 7:22SPEAKER_00: Don't be so precious.
- 7:23SPEAKER_00: Don't be so possessive.
- 7:26SPEAKER_00: It doesn't belong to you.
- 7:28And it would be absurd if your toaster suddenly jumped up and
- 7:37grabbed you by the
- 7:38SPEAKER_00: hand and started leading you around and started...
- 7:42taking control over your life and forcing you to think things
- 7:50and act in ways and
- 7:51SPEAKER_00: maybe even develop some beliefs that you otherwise wouldn't.
- 7:56SPEAKER_00: It would be absurd if your toaster did that or if a hammer
- 8:01did that but language does that and because we have no choice
- 8:07but to be in language
- 8:08it is a little bit more difficult to be critical of your use of
- 8:12language but
- 8:15language is ultimately this technology language is a tool so
- 8:20you must be wary of it
- 8:24SPEAKER_00: using you what do I mean when I say it using you um
- 8:33There can be unconscious trends and currents going on just in
- 8:38the zeitgeist at the
- 8:40SPEAKER_00: level of ideology or just at the level of background.
- 8:45SPEAKER_00: A lot of the time, these things can be very useful.
- 8:51SPEAKER_00: They can serve as useful shortcuts.
- 8:55SPEAKER_00: for discourse,
- 8:56SPEAKER_00: understanding,
- 8:57SPEAKER_00: meaning,
- 8:58SPEAKER_00: Congress,
- 9:00SPEAKER_00: intercourse,
- 9:01SPEAKER_00: not sexual intercourse,
- 9:02SPEAKER_00: but human intercourse.
- 9:04And they're necessary,
- 9:05but also we should be always willing to interrogate those
- 9:10things.
- 9:11SPEAKER_00: Common knowledge, right?
- 9:16SPEAKER_00: Folk wisdom or whatever.
- 9:18SPEAKER_00: People don't like to use the word folk because it can be problematic.
- 9:23SPEAKER_00: Yeah, that's true.
- 9:24SPEAKER_00: It definitely can be problematic.
- 9:27SPEAKER_00: But also,
- 9:28SPEAKER_00: it can also just kind of mean like this,
- 9:31SPEAKER_00: you know,
- 9:32SPEAKER_00: old wives tales,
- 9:32SPEAKER_00: general knowledge,
- 9:36SPEAKER_00: common sense,
- 9:38SPEAKER_00: background knowledge,
- 9:38SPEAKER_00: blah,
- 9:39SPEAKER_00: blah,
- 9:39SPEAKER_00: blah,
- 9:39SPEAKER_00: this and that.
- 9:41There can be things that are going on in that unconscious
- 9:45field,
- 9:46in the social field.
- 9:49SPEAKER_00: And before you know it,
- 9:51SPEAKER_00: you find yourself engaged in this in-group,
- 9:53out-group bullshit over some of these things that are in that
- 9:58background field.
- 10:00SPEAKER_00: And they're really kind of not real,
- 10:05SPEAKER_00: not worth engaging in,
- 10:06SPEAKER_00: but you find yourself really,
- 10:07SPEAKER_00: really,
- 10:08SPEAKER_00: really deeply engaged.
- 10:12SPEAKER_00: I am trying to avoid going on a spiral right now.
- 10:14SPEAKER_00: I am trying to avoid using any examples.
- 10:19SPEAKER_00: As I was recording it, I was like, this is trash.
- 10:23SPEAKER_00: The second half of the recording I did, I was like, this is not trash.
- 10:29SPEAKER_00: I want to avoid spiraling out.
- 10:37SPEAKER_00: So be more critical, be more self-aware.
- 10:45SPEAKER_00: And there's all that.
- 10:47SPEAKER_00: And I don't know, maybe that'll come up again later.
- 10:48SPEAKER_00: Maybe it won't.
- 10:52SPEAKER_00: But the...
- 10:54SPEAKER_00: So the correction I wanted to make,
- 10:56SPEAKER_00: I did want to make one correction to the previous Brolo episode.
- 11:01SPEAKER_00: I said Walter Benjamin twice.
- 11:02SPEAKER_00: I thought about Benjamin and Marcel Mauss and Baudrillard.
- 11:06And this is actually why I started using names,
- 11:07Baudrillard and Marcel Mauss.
- 11:09SPEAKER_00: And then I went on and on and on and hyper fixated on other names.
- 11:15SPEAKER_00: I was talking about Benjamin and Baudrillard and I said Benjamin twice.
- 11:21SPEAKER_00: The second time I said Benjamin, I meant to say George Bataille.
- 11:25I think Bataille can give us an insight into what makes a
- 11:32useful object different
- 11:34from a commodity or what makes a sentimental object different
- 11:37from a commodity
- 11:38SPEAKER_00: object or whatever.
- 11:44based on like the metabolic activity behind the creation of
- 11:50that object or whatever
- 11:52like the purpose of that object like there there is human
- 11:55energy poured into all
- 11:57objects right and the commodity fetish kind of disappears that
- 12:00away and keeps us
- 12:01SPEAKER_00: from seeing it um but like objects of use that is
- 12:07SPEAKER_00: generally considered part of it.
- 12:11Benjamin talks about the aura of originality and this is kind
- 12:15of similar to what
- 12:16I'm thinking along with Bataille in this form of like you can
- 12:21see the history of
- 12:22the object and you can actually kind of see a projected future
- 12:26of a given object
- 12:27SPEAKER_00: but the commodity is
- 12:30disappears that information the commodity you can't see the
- 12:35history of the
- 12:36commodity the human history of the commodity and you also you
- 12:42can't properly
- 12:44project a human future of a commodity you can project like a
- 12:49fantasy future of a
- 12:50SPEAKER_00: commodity you can see these sunglasses will make me cool
- 12:57SPEAKER_00: And when I feel cool, all my problems will go away.
- 13:00SPEAKER_00: I will no longer be a sad, silly boy.
- 13:03SPEAKER_00: I will be fucking cool.
- 13:06SPEAKER_00: So you can kind of project this fantasy future of the commodity.
- 13:11SPEAKER_00: But it's a bit more difficult.
- 13:13SPEAKER_00: Whereas with a,
- 13:15SPEAKER_00: you know,
- 13:15SPEAKER_00: object of use or with a sentimental object,
- 13:19SPEAKER_00: it is much easier to to project a human future.
- 13:24SPEAKER_00: Of that object.
- 13:24SPEAKER_00: But anyway, so that's the one correction I wanted to make.
- 13:27SPEAKER_00: And it just bugged me.
- 13:30SPEAKER_00: I was like, oh, I misspoke.
- 13:33SPEAKER_00: And it bothers me to misspeak, even though I often misspeak.
- 13:37SPEAKER_00: A lot of the time,
- 13:39SPEAKER_00: misspeaking is...
- 13:43SPEAKER_00: Well,
- 13:43SPEAKER_00: then it's not misspeaking if I'm doing it intentionally,
- 13:45SPEAKER_00: right?
- 13:46SPEAKER_00: So a lot of the time, I say things wrong on purpose...
- 13:50SPEAKER_00: But also a lot of the time I just say things wrong, not on purpose.
- 13:54SPEAKER_00: And I don't like it when I say things wrong, not on purpose.
- 13:59SPEAKER_00: Because it's frustrating to me.
- 14:00SPEAKER_00: It's like, damn it.
- 14:02SPEAKER_00: Now people aren't going to know what I meant.
- 14:05But then that's weird because other times when I say the wrong
- 14:09thing on purpose,
- 14:11SPEAKER_00: People still aren't going to know what I mean.
- 14:14SPEAKER_00: But I'm trying.
- 14:14SPEAKER_00: I'm trying to force a breakdown.
- 14:16SPEAKER_00: I'm trying to do something creative.
- 14:18SPEAKER_00: So I guess there's that.
- 14:19SPEAKER_00: So it comes down to agency.
- 14:23SPEAKER_00: I don't like it when my unconscious breaks through.
- 14:29SPEAKER_00: Or I don't like it when my underbelly shows.
- 14:37SPEAKER_00: But I do like to...
- 14:41SPEAKER_00: Use this performance of misspeaking to try to do something.
- 14:46SPEAKER_00: So I guess that comes down to agency and ownership.
- 14:48So I guess I am guilty of being overly protective of language
- 14:52in that sense too.
- 14:53SPEAKER_00: So shame on me.
- 14:54SPEAKER_00: Tsk, tsk, tsk.
- 14:56SPEAKER_00: Nancy boy, you're just as guilty as everyone else.
- 15:00SPEAKER_00: As a matter of fact, we're fucked.
- 15:02SPEAKER_00: We're fucked.
- 15:04SPEAKER_00: We're fucked.
- 15:05SPEAKER_00: We're fucked.
- 15:06SPEAKER_00: I am often far more guilty than others.
- 15:14SPEAKER_00: What do I got going on over here on my thing?
- 15:17SPEAKER_00: I...
- 15:23hey there buckaroo bonsai let's go across the eighth dimension
- 15:27what's up dude i
- 15:30really want to i really want to play around and have fun but
- 15:36i'm not going to do
- 15:37that right now i am trying to to actually record um but yeah so
- 15:43that was the
- 15:43SPEAKER_00: correction the clarification um i wanted to give
- 15:48SPEAKER_00: was around this idea of intelligence.
- 15:53Had a conversation with a friend after he listened to that
- 15:56episode and I had a
- 15:57SPEAKER_00: conversation with him.
- 15:58SPEAKER_00: It was a great conversation, very generative.
- 16:03But he was like,
- 16:05hey man,
- 16:06I got a big problem with how you're using intelligence.
- 16:10SPEAKER_00: I was like,
- 16:10SPEAKER_00: word,
- 16:11SPEAKER_00: dude,
- 16:11I love it when my friends hit me up and they say they have a
- 16:13big problem with me or
- 16:15SPEAKER_00: something I'm doing.
- 16:16SPEAKER_00: I love it.
- 16:18SPEAKER_00: And then we went on and had an hours-long conversation.
- 16:22SPEAKER_00: It was really good.
- 16:22SPEAKER_00: I actually wish we would have recorded that conversation.
- 16:26SPEAKER_00: That would have been a great conversation to share with people.
- 16:30SPEAKER_00: I think he said some things that were great and worth hearing.
- 16:35And I think I managed to achieve a level of clarity that is
- 16:41difficult to do.
- 16:43SPEAKER_00: On my own.
- 16:44SPEAKER_00: Because I am a long winded guy.
- 16:46SPEAKER_00: And I like to meander and maunder.
- 16:49SPEAKER_00: About.
- 16:52SPEAKER_00: But you know what.
- 16:53SPEAKER_00: Gosh darn it.
- 16:54SPEAKER_00: I wouldn't give it up for the world.
- 16:56SPEAKER_00: I do think it matters.
- 16:59SPEAKER_00: To be able to.
- 17:01SPEAKER_00: To meander and and go at my own pace or go at a different pace.
- 17:06I think it's important I have seen comments And I don't I don't
- 17:14care all that much
- 17:15SPEAKER_00: about comments comment sections I I don't know man every once in a while.
- 17:22SPEAKER_00: There's there's a really cool one I
- 17:26SPEAKER_00: But I don't know.
- 17:27Sometimes comments in a comment section can lead to some really
- 17:32great conversation
- 17:33SPEAKER_00: and some really great things.
- 17:35SPEAKER_00: But I think often it's not even worth it.
- 17:39SPEAKER_00: So I don't really peruse comment sections.
- 17:42SPEAKER_00: I don't really leave comments.
- 17:45SPEAKER_00: In the past, I don't know.
- 17:50SPEAKER_00: We'll say year.
- 17:53SPEAKER_00: I've probably left five comments.
- 17:56SPEAKER_00: No, that's not true.
- 17:57SPEAKER_00: Probably like 10 or 12.
- 17:59SPEAKER_00: We'll say about a dozen or so.
- 18:01SPEAKER_00: And all but one have been positive and sharing support,
- 18:08SPEAKER_00: showing support,
- 18:09SPEAKER_00: and showing enthusiasm and being like,
- 18:11SPEAKER_00: yeah,
- 18:11SPEAKER_00: fuck yeah,
- 18:12SPEAKER_00: this is dope.
- 18:12SPEAKER_00: Hell yeah, bro.
- 18:14SPEAKER_00: On my homies stuff.
- 18:17SPEAKER_00: Um, and the only one comment I left that was, that was not positive.
- 18:21There was a Gavin McInnes video and I called him like a crybaby
- 18:26bully bitch.
- 18:28SPEAKER_00: Um, and that was just pure, like that comment is going to go nowhere.
- 18:32SPEAKER_00: I shouldn't be under the impression of like,
- 18:34SPEAKER_00: yeah,
- 18:34people are going to see this and it's going to make a
- 18:36difference in anyone's life.
- 18:39SPEAKER_00: Um, now for that moment, I was libidinally invested and I was like,
- 18:45SPEAKER_00: fuck Gavin McInnes.
- 18:47But like immediately I was like,
- 18:50oh yeah,
- 18:50that,
- 18:50that accomplishes nothing.
- 18:52And I didn't have this like impression that it was going to
- 18:55accomplish anything,
- 18:57SPEAKER_00: but you know,
- 18:58SPEAKER_00: I,
- 18:58SPEAKER_00: uh,
- 18:58SPEAKER_00: I have looked at my comment section a couple of times recently,
- 19:02SPEAKER_00: um,
- 19:03SPEAKER_00: put,
- 19:03SPEAKER_00: put up some new stuff recently,
- 19:05SPEAKER_00: um,
- 19:06SPEAKER_00: put up a few things that I'm re have been really,
- 19:08SPEAKER_00: really excited about.
- 19:09SPEAKER_00: Uh, and then there's this, like these bro low episodes, uh,
- 19:13You know,
- 19:14I enjoy it and I actually want to start getting back in the
- 19:17saddle.
- 19:19SPEAKER_00: But it's not like I'm like, oh, I did so well.
- 19:21SPEAKER_00: I want to see what the people are saying.
- 19:23SPEAKER_00: Like, I don't know, man.
- 19:28SPEAKER_00: I don't think these are great.
- 19:35SPEAKER_00: So I'm not.
- 19:36SPEAKER_00: Yeah, I'm not going to be all like comments.
- 19:39SPEAKER_00: And then also, you know, the whole identity thing in comment sections.
- 19:44SPEAKER_00: Look,
- 19:45SPEAKER_00: man,
- 19:46SPEAKER_00: I think largely and for the most part,
- 19:50SPEAKER_00: people are genuinely trying to express themselves,
- 19:53SPEAKER_00: genuinely trying to be heard,
- 19:54SPEAKER_00: and I empathize with that.
- 19:57SPEAKER_00: So it is important to not be entirely dismissive.
- 20:04I've had conversations with people or tried to have
- 20:08conversations with people.
- 20:09SPEAKER_00: I'm not
- 20:11very good at uh keeping in touch or whatever um but you know i
- 20:18i do try and i know
- 20:19i i know i can do better um and there's all that i don't want
- 20:25to turn into you know
- 20:28the mewling puddle of like neurotic oh i'm so bad at this one
- 20:34thing how could i do
- 20:34SPEAKER_00: better i don't want to do that
- 20:36SPEAKER_00: I just want to say if you are in the comment section and you don't suck,
- 20:45SPEAKER_00: there's a good chance I do see you and I'm not trying to say you suck.
- 20:54There's just also a lot of other people that suck and not
- 20:56really even in my comment
- 20:57SPEAKER_00: section all that much because in my comment section,
- 20:59SPEAKER_00: there'll be like 10 comments and nine of them are like,
- 21:02SPEAKER_00: yeah,
- 21:02SPEAKER_00: what's up?
- 21:03you know and i'll talk to i talk to half the people that leave
- 21:06comments on my shit
- 21:09um so it almost defeats the purpose of reading the comments
- 21:12because i can just you
- 21:13know have have a call or have a conversation with them or
- 21:15whatever um but on bigger
- 21:18stuff like on in mikey's comment section you know and i know a
- 21:21lot of those people
- 21:22SPEAKER_00: in there too um but there's also just some like
- 21:28wieners in there um so largely you know comment sections or
- 21:33whatever dude but i do
- 21:35i do know that there are people who are like genuine in there
- 21:38and shout out you
- 21:39SPEAKER_00: know keep it up uh sometimes your feedback is as actually
- 21:45SPEAKER_00: very helpful, very constructive, very generative.
- 21:48SPEAKER_00: Um,
- 21:49SPEAKER_00: but then there are also people in comment sections where,
- 21:51SPEAKER_00: I don't know,
- 21:52SPEAKER_00: man,
- 21:52SPEAKER_00: there's people who do weird shit in comment sections.
- 21:54SPEAKER_00: There's people who,
- 21:55SPEAKER_00: you know,
- 21:56put on different sock accounts and do weird shit and they try
- 21:58to like create weird
- 22:01SPEAKER_00: fucking dynamics and I don't know,
- 22:05SPEAKER_00: man.
- 22:06So just in general,
- 22:07my default disposition,
- 22:09um,
- 22:10is like,
- 22:10I don't care about the comment section.
- 22:15um but uh i also sometimes look and i've seen uh a few comment
- 22:24sections where
- 22:25people will leave like a shitty comment trying to be like uh
- 22:31you're dumb or you
- 22:32guys suck or whatever or like wow what a waste of time i've
- 22:35been here for 15
- 22:36SPEAKER_00: minutes and you haven't really said anything
- 22:40SPEAKER_00: And that actually kind of,
- 22:41SPEAKER_00: to me,
- 22:43SPEAKER_00: it's like,
- 22:43SPEAKER_00: well,
- 22:44SPEAKER_00: dude,
- 22:44like reading the comment you're leaving shows me you don't know
- 22:52how to listen to a
- 22:54SPEAKER_00: conversation or you don't know how to like chill and go at your own pace.
- 22:59You're brainwashed into needing like this immediate
- 23:03gratification or this like
- 23:06SPEAKER_00: immediate commodity object.
- 23:08SPEAKER_00: Like,
- 23:09This fucking two and a half hour conversation is going to be
- 23:14the dopest fucking
- 23:15SPEAKER_00: thing you've ever heard.
- 23:16SPEAKER_00: And you're going to agree,
- 23:16SPEAKER_00: agree,
- 23:17SPEAKER_00: agree,
- 23:17SPEAKER_00: agree,
- 23:17SPEAKER_00: agree,
- 23:18SPEAKER_00: or disagree,
- 23:19SPEAKER_00: disagree,
- 23:19SPEAKER_00: disagree,
- 23:20SPEAKER_00: disagree.
- 23:21SPEAKER_00: I don't know, man.
- 23:22SPEAKER_00: It's just fucking silly, dude.
- 23:23SPEAKER_00: It's just weird.
- 23:24SPEAKER_00: Like, we're all here.
- 23:25SPEAKER_00: We're all going through it.
- 23:28SPEAKER_00: We're not trying to... Now, I don't know.
- 23:31SPEAKER_00: If we put out a 10-minute video,
- 23:32SPEAKER_00: then I would expect that 10-minute video to fucking go,
- 23:36SPEAKER_00: go,
- 23:36SPEAKER_00: go,
- 23:36SPEAKER_00: go,
- 23:37SPEAKER_00: go.
- 23:38Mikey recently put out about a 10-minute video talking about
- 23:42Heidegger,
- 23:43talking about Dugan.
- 23:44SPEAKER_00: And it was.
- 23:45SPEAKER_00: It was just like, go, go, go, go, go, go, go, go, go.
- 23:47SPEAKER_00: There wasn't filler.
- 23:49SPEAKER_00: It wasn't this meandering conversation.
- 23:52SPEAKER_00: And yeah, that... Yeah, 10-minute video.
- 23:56SPEAKER_00: Okay, cool.
- 23:57Let me attenuate to this 10 minutes so I can fucking do this
- 24:01knowledge transfer or
- 24:03SPEAKER_00: whatever,
- 24:03SPEAKER_00: right?
- 24:05But in an hour and a half to two and a half to three-hour
- 24:08conversation or whatever,
- 24:09SPEAKER_00: if you...
- 24:12SPEAKER_00: If your problem is that the first 15 minutes are not...
- 24:20the solution to all the world's problems or even all your
- 24:24problems or whatever.
- 24:26SPEAKER_00: That's a you problem, dude.
- 24:27SPEAKER_00: This is an entire conversation.
- 24:30SPEAKER_00: It has a direction.
- 24:34SPEAKER_00: And there's an actual process of walking down a path.
- 24:38It'd be like if you went to,
- 24:40I don't know,
- 24:42somewhere you'd never been before,
- 24:45and you had to like hike down a trail or whatever to get there
- 24:50and then you arrived
- 24:51and you started down the trail and you made it 15 of the way
- 24:57down the trail and
- 24:58then you said this sucks peace out i'm out and then you went
- 25:02around publicly like
- 25:03saying yeah that place sucks uh the first 15 minutes of that
- 25:07trail was so i ducked
- 25:09SPEAKER_00: out
- 25:11SPEAKER_00: then you're kind of missing out on the destination.
- 25:13SPEAKER_00: And that is the way that conversations work.
- 25:16SPEAKER_00: That is the way that dialogue works.
- 25:18SPEAKER_00: It is the way that soliloquies work and monologues work and all that.
- 25:22SPEAKER_00: There's a point to the process.
- 25:28SPEAKER_00: And I think it's important.
- 25:34SPEAKER_00: to be able to withstand the journey.
- 25:38I think it's important to be able to withstand the dialogue or
- 25:41the monologue or the
- 25:42SPEAKER_00: soliloquy because it resists that commodity form.
- 25:49SPEAKER_00: It resists that fantasy of,
- 25:53SPEAKER_00: whoa,
- 25:53SPEAKER_00: these are the dopest,
- 25:54SPEAKER_00: smartest guys in the world and they have the dopest,
- 25:57awesomest things to say and you're going to hear them and it's
- 26:00going to be super
- 26:01SPEAKER_00: dope.
- 26:01SPEAKER_00: Or also,
- 26:02SPEAKER_00: you know,
- 26:03these guys are the dumbest guys in the world and you're going
- 26:05to listen to them and
- 26:06you're going to know exactly why you're already right about
- 26:08everything and they're
- 26:09SPEAKER_00: already wrong about everything or whatever.
- 26:11SPEAKER_00: Like the meandering is the fucking point, kind of.
- 26:18SPEAKER_00: So I think it's important to be able to do that.
- 26:22SPEAKER_00: And I also enjoy doing that.
- 26:25um but yeah there was some you know recent recent comments
- 26:29talking about that on uh
- 26:30one of the one of the videos mikey and i put out and get fucked
- 26:36homes i i don't
- 26:37SPEAKER_00: give a shit like i don't care like if you can't even uh like
- 26:45SPEAKER_00: Be patient.
- 26:47SPEAKER_00: And look, I don't mean to say get fucked.
- 26:49SPEAKER_00: I feel for you, man.
- 26:51SPEAKER_00: I get it.
- 26:51We're constantly in this fucking like temporal state of like it
- 26:58has to be now.
- 26:59SPEAKER_00: It has to be now.
- 27:00SPEAKER_00: It has to be now.
- 27:02SPEAKER_00: And it's, you know, the medium is the message.
- 27:06SPEAKER_00: Give it to me now.
- 27:07SPEAKER_00: Give it to me now.
- 27:07SPEAKER_00: I get it.
- 27:08SPEAKER_00: So I empathize with you.
- 27:11SPEAKER_00: But but also it's frustrating because
- 27:15SPEAKER_00: Because it's just like, damn, what is the point?
- 27:19SPEAKER_00: Well, I don't know.
- 27:19SPEAKER_00: I don't think there necessarily is any concrete point.
- 27:24I think the point is going through the process of searching for
- 27:30the point that can
- 27:31SPEAKER_00: never be found,
- 27:32SPEAKER_00: right?
- 27:33SPEAKER_00: But I don't know.
- 27:34SPEAKER_00: Anyway, just reflecting on, I guess, the comment section.
- 27:38SPEAKER_00: So there's that.
- 27:39SPEAKER_00: And I know that was self-indulgent.
- 27:41SPEAKER_00: That's part of why the other thing I recorded the other day,
- 27:45SPEAKER_00: I was like,
- 27:45SPEAKER_00: this is too self-indulgent.
- 27:48SPEAKER_00: Just 45 minutes of me going on and on and on and on and on.
- 27:54SPEAKER_00: It was too self-indulgent.
- 27:56I understand the first half hour of this has been pretty
- 27:59self-indulgent,
- 27:59but sue me,
- 28:03dude.
- 28:06SPEAKER_00: Anyway, moving on.
- 28:09SPEAKER_00: I did want to make a clarification about intelligence.
- 28:15SPEAKER_00: And like I said, I wish we would have recorded that call we had, Holmes.
- 28:20Because I think it was better at actually showing the problem
- 28:26and tarrying with the
- 28:27SPEAKER_00: problem than I could achieve on my own.
- 28:32SPEAKER_00: But the problem was intelligence.
- 28:34The way I was using intelligence makes the word intelligence
- 28:39itself useless.
- 28:41SPEAKER_00: And I agree with that criticism, with that feedback.
- 28:45SPEAKER_00: And in fact, that's the point.
- 28:48SPEAKER_00: The criticism was,
- 28:51if I'm making the claim that my toaster is intelligent and
- 28:54intelligence is the
- 28:55ability to accomplish a task or to achieve a goal within an
- 29:00environment or within a
- 29:01SPEAKER_00: world,
- 29:03SPEAKER_00: or the ability to act in and upon the world,
- 29:09SPEAKER_00: or even competence at playing games,
- 29:11SPEAKER_00: And by the way,
- 29:11SPEAKER_00: Mikey did an excellent bit on this,
- 29:15games and playing games and how does intelligence relate to
- 29:19playing games in the
- 29:21SPEAKER_00: final session of the Lan Zizek Singularity Course at Philosophy Portal.
- 29:27SPEAKER_00: It was just fucking perfect, dude.
- 29:30The way you talked about games and game playing and why might
- 29:35game playing actually
- 29:36SPEAKER_00: have to do with intelligence.
- 29:38SPEAKER_00: It was excellent.
- 29:40SPEAKER_00: So get access to that if you don't have it already and go listen.
- 29:45SPEAKER_00: The whole fucking thing was great.
- 29:47SPEAKER_00: But yeah, on that point, games, game playing and intelligence.
- 29:52SPEAKER_00: Mikey handled that expertly.
- 29:56So we'll talk about intelligence as accomplishing a goal within
- 30:01a world or within
- 30:02SPEAKER_00: an environment or accomplishing a task or acting in and upon the world.
- 30:10And the reason I use these definitions of intelligence is
- 30:13because this is how
- 30:15SPEAKER_00: people define intelligence.
- 30:16SPEAKER_00: Like there is no...
- 30:18SPEAKER_00: I guess, satisfactory definition of intelligence.
- 30:23But find me a theorist of mind that doesn't ultimately define
- 30:29intelligence this way.
- 30:31SPEAKER_00: Now, gosh, I don't want to pull anything up.
- 30:38SPEAKER_00: But that's Lumen, not Lumen.
- 30:46SPEAKER_00: Where are you?
- 30:48SPEAKER_00: No, it's not Byung-Chul Han.
- 30:50SPEAKER_00: It's not Baudrillard.
- 30:51SPEAKER_00: It's not Land.
- 30:55SPEAKER_00: It's not Zizek.
- 30:58SPEAKER_00: It's not Zizek.
- 31:01SPEAKER_00: It's not Batai.
- 31:04SPEAKER_00: Where is my fucking PDF?
- 31:07SPEAKER_00: Where did you go, homie?
- 31:08SPEAKER_00: It's not Heinrich.
- 31:12SPEAKER_00: I should share the screen so you can see.
- 31:17SPEAKER_00: Let's see.
- 31:20SPEAKER_00: Oh, you can't see it.
- 31:22SPEAKER_01: Well, you're not actually going to be able to see it very well.
- 31:25SPEAKER_01: See all these tabs?
- 31:31SPEAKER_00: Oh, you can't see all the tabs because I have it fucking cropped.
- 31:50SPEAKER_00: Now you can.
- 31:51SPEAKER_00: All these tabs, these are what I was just going through.
- 31:56SPEAKER_00: Trying to find PDF I was looking for.
- 32:00SPEAKER_00: But it's John Searle, Minds, Brain, and Science.
- 32:06SPEAKER_00: I think he has a very interesting take on what is intelligence.
- 32:14SPEAKER_00: But other than that, find me.
- 32:16SPEAKER_00: And I don't even know.
- 32:17SPEAKER_00: Would you consider him a theorist of mind or would you consider him?
- 32:20SPEAKER_00: I don't know.
- 32:22SPEAKER_00: I don't know what you would consider.
- 32:25SPEAKER_00: He's the guy that came up with the Chinese room.
- 32:30Um,
- 32:31basically,
- 32:33um,
- 32:33being in,
- 32:34in conversation with the Turing machine,
- 32:37the imitation game.
- 32:39SPEAKER_00: Uh, I think the Turing machine gets fetishized in a weird way.
- 32:42Uh,
- 32:43even the Chinese room gets,
- 32:44you know,
- 32:44kind of taken and misunderstood.
- 32:48SPEAKER_00: Um, that's the column.
- 32:51SPEAKER_00: But, uh, let me switch back cause I am distracted now.
- 32:56SPEAKER_00: Okay.
- 32:59SPEAKER_00: But yeah, anyway, so I don't know.
- 33:02SPEAKER_00: You know, Searle, I think, is pretty interesting on intelligence.
- 33:06SPEAKER_00: But I don't know.
- 33:07SPEAKER_00: Find a theorist of mind or whatever.
- 33:12SPEAKER_00: I mean, who would we even go to for a definition of intelligence?
- 33:16SPEAKER_00: Would we go to a historian?
- 33:17SPEAKER_00: Would we go to an anthropologist?
- 33:20Would we go to a psychiatrist,
- 33:22a psychoanalyst,
- 33:24a social worker,
- 33:25a construction worker?
- 33:27SPEAKER_00: I don't know.
- 33:29SPEAKER_00: Who would we go to?
- 33:30SPEAKER_00: But it does seem to be the case that in general,
- 33:33SPEAKER_00: in the mainstream,
- 33:34the generally agreed upon definition of intelligence is some
- 33:38permutation of this
- 33:39SPEAKER_00: ability to act in and upon the world,
- 33:42SPEAKER_00: the ability to accomplish a goal or a task in an environment.
- 33:51SPEAKER_00: And the reason I use that definition is
- 33:57is because it's easier than just trying to come up with a new
- 34:01one or come up with
- 34:02SPEAKER_00: my own or,
- 34:03SPEAKER_00: you know,
- 34:04SPEAKER_00: have a hundred predicates and presuppositions and shit like that.
- 34:08SPEAKER_00: It's just like, okay, we'll run with that.
- 34:10SPEAKER_00: Now,
- 34:11SPEAKER_00: if you remember,
- 34:12SPEAKER_00: dear reader,
- 34:13SPEAKER_00: a while ago,
- 34:14SPEAKER_00: Mikey and I had a conversation about min-telligence and hen-telligence,
- 34:18SPEAKER_00: you know,
- 34:19SPEAKER_00: machinic intelligence.
- 34:20SPEAKER_00: versus human intelligence.
- 34:23SPEAKER_00: We talked about it.
- 34:24SPEAKER_00: We had a brainstorm.
- 34:26SPEAKER_00: We recorded an episode about it.
- 34:27SPEAKER_00: We talked to Todd McGowan about it.
- 34:30SPEAKER_00: We talked to Nick Land about it.
- 34:32We've talked to other people about it that aren't public
- 34:35conversations for the internet.
- 34:38But we were really and are really trying to tarry with this
- 34:42question of intelligence.
- 34:46and figure out what is the distinction between min-intelligence
- 34:50and hin-intelligence.
- 34:53SPEAKER_00: So when I'm talking about intelligence,
- 34:57SPEAKER_00: like in the Brollo episode,
- 34:59SPEAKER_00: in the previous Brollo episode when I was talking about intelligence,
- 35:02SPEAKER_00: I was talking about min-intelligence.
- 35:04Like that definition of intelligence is good enough for
- 35:10machinic intelligence,
- 35:12SPEAKER_00: instrumental reasons.
- 35:17SPEAKER_00: And the point is,
- 35:19SPEAKER_00: when we're talking about human intelligence,
- 35:21SPEAKER_00: we're talking about,
- 35:22SPEAKER_00: and maybe even when we're talking about general intelligence,
- 35:25SPEAKER_00: you know,
- 35:25SPEAKER_00: artificial general intelligence.
- 35:27SPEAKER_00: Maybe, maybe not.
- 35:28SPEAKER_00: I don't know.
- 35:31But human intelligence is different than this kind of machine
- 35:36intelligence.
- 35:37SPEAKER_00: We went through this already.
- 35:39SPEAKER_00: And subjectivity is what's at stake.
- 35:43SPEAKER_00: So are we conflating human intelligence and subjectivity?
- 35:46SPEAKER_00: I don't know, maybe, maybe not.
- 35:50SPEAKER_00: Where does consciousness come into play?
- 35:52SPEAKER_00: You know, interiority, this and that.
- 35:54SPEAKER_00: Like, where do all these things...
- 35:58lie on the field i don't know like we're working through it
- 36:02right now but in the in
- 36:04the previous episode uh i was i guess rather flippantly using
- 36:08the word intelligence
- 36:09and uh i wasn't making those clarifications uh and so my friend
- 36:16was like hey man if
- 36:19you're gonna say your toaster is intelligent then you can't
- 36:22talk about intelligence
- 36:23SPEAKER_00: and i was like no no no no no
- 36:25I can say my toaster is intelligent because my toaster does
- 36:28have the ability to
- 36:29SPEAKER_00: achieve a goal,
- 36:30SPEAKER_00: accomplish a goal.
- 36:32SPEAKER_00: Seemingly,
- 36:33SPEAKER_00: without any further input,
- 36:34SPEAKER_00: there's the initial state,
- 36:38SPEAKER_00: which for the toaster example,
- 36:39SPEAKER_00: the initial state is already with bread in it,
- 36:42SPEAKER_00: with the thing depressed.
- 36:43That's the initial state for the toaster when it comes to
- 36:45talking about intelligence.
- 36:48SPEAKER_00: The initial state for an AI prompt is after you hit enter.
- 36:57And then it does all its magical stuff and then it spits out
- 37:01the answer or it achieves a goal.
- 37:04SPEAKER_00: Your goal is to achieve a given answer or a given prompt or whatever,
- 37:11SPEAKER_00: a string of text or an image or string of sounds or whatever.
- 37:17um and his problem was yeah but you're not accounting for that
- 37:22initial state you're
- 37:24not accounting for agency like the toaster doesn't just jump up
- 37:28go to the cupboard
- 37:29get out a piece of bread put the bread inside of it and depress
- 37:33the thing it
- 37:34doesn't do that on its own it still requires human agency uh or
- 37:40some other such
- 37:41SPEAKER_00: agency but human for the most part you know
- 37:46SPEAKER_00: to achieve that initial state.
- 37:47SPEAKER_00: I was like, yeah, you're right.
- 37:49And for this kind of like materialist matter in motion view
- 37:57where these LLMs are
- 37:59actually doing intelligence or even for the view that these
- 38:06LLMs have achieved some
- 38:08SPEAKER_00: type of mystical connection with the outside or whatever.
- 38:12SPEAKER_00: Both are popular opinions.
- 38:16SPEAKER_00: they're bracketing out the agency.
- 38:20They're bracketing out the agency from the materialist on the
- 38:24materialist side
- 38:25because the main concern ultimately is entropy and extropy or
- 38:30local extropy.
- 38:34SPEAKER_00: Achieving a goal,
- 38:35creating some type of stable structure that is entropy
- 38:40resilient or entropy
- 38:42SPEAKER_00: resistant.
- 38:43SPEAKER_00: Not necessarily entropy-proof,
- 38:45SPEAKER_00: although I do think a lot of these guys start with a closed system.
- 38:51SPEAKER_00: But then they kind of backdoor in an openness,
- 38:56SPEAKER_00: like a transcendental,
- 38:58when they're talking about intelligence and they're talking
- 39:01about extropy or
- 39:02SPEAKER_00: negentropy or whatever.
- 39:04SPEAKER_00: And they're probably not even aware of it.
- 39:09SPEAKER_00: Um,
- 39:10SPEAKER_00: but so for,
- 39:11SPEAKER_00: for like the hard materialist view,
- 39:12SPEAKER_00: the closed system view,
- 39:14SPEAKER_00: like local extra P ultimately is intelligence.
- 39:17SPEAKER_00: Like that,
- 39:18SPEAKER_00: that which can resist,
- 39:20SPEAKER_00: you know,
- 39:20SPEAKER_00: entropy at least locally,
- 39:23SPEAKER_00: um,
- 39:23SPEAKER_00: is intelligent and they can say,
- 39:25SPEAKER_00: you know,
- 39:26SPEAKER_00: it's all,
- 39:26SPEAKER_00: it's all matter in motion.
- 39:28SPEAKER_00: Now,
- 39:28SPEAKER_00: the one thing that I think is important,
- 39:30SPEAKER_00: you know,
- 39:31SPEAKER_00: they'll sometimes talk about like randomness,
- 39:34SPEAKER_00: um,
- 39:35SPEAKER_00: It's all random matter in motion.
- 39:37SPEAKER_00: The thing is, if you have a closed system, it's not random.
- 39:41SPEAKER_00: It is statistically determined.
- 39:44SPEAKER_00: So it's not truly random.
- 39:45SPEAKER_00: It's a stochastic arrangement of molecules.
- 39:50SPEAKER_00: It's all stochastic.
- 39:50SPEAKER_00: It's all ultimately statistically determined,
- 39:52SPEAKER_00: even if we cannot comprehend the sheer volume of molecules involved,
- 39:59SPEAKER_00: which we can't.
- 40:00SPEAKER_00: It's still not random.
- 40:01SPEAKER_00: It's still statistically determined.
- 40:05SPEAKER_00: You know,
- 40:05SPEAKER_00: given a finite universe in a finite amount of time,
- 40:09SPEAKER_00: there's a finite set of things that will happen.
- 40:13SPEAKER_00: So on the materialist view, they don't care about agency.
- 40:20SPEAKER_00: They only care ultimately about that which can persist.
- 40:27SPEAKER_00: So these random arrangements of, let's see, I just did it.
- 40:32SPEAKER_00: These,
- 40:32SPEAKER_00: in quotes,
- 40:33random arrangements of molecules and matter that don't just
- 40:37immediately fucking,
- 40:38SPEAKER_00: you know,
- 40:39SPEAKER_00: wither away.
- 40:41SPEAKER_00: That's intelligence.
- 40:42SPEAKER_00: That's their measure of intelligence.
- 40:44SPEAKER_00: And then you go on to extrapolate,
- 40:46SPEAKER_00: you know,
- 40:46SPEAKER_00: stable systems,
- 40:49SPEAKER_00: planets,
- 40:50SPEAKER_00: you know,
- 40:51SPEAKER_00: agglomerate and become planets.
- 40:53SPEAKER_00: And then over eons,
- 40:55SPEAKER_00: amino acids form stable structures and then bacteria and then amoeba,
- 41:01SPEAKER_00: you know,
- 41:02SPEAKER_00: and then single cell organisms and then prokaryotes and then
- 41:06SPEAKER_00: On and on and on and on and on.
- 41:07But these are all arrangements of molecules that would
- 41:11otherwise kind of maybe
- 41:12SPEAKER_00: spontaneously come together,
- 41:13SPEAKER_00: but then diffuse back into the chaotic sea.
- 41:19But then there are arrangements that are intelligent,
- 41:23like the living cell,
- 41:25like the bacterium.
- 41:27And then you have processes of evolution and then eventually
- 41:30you get people and
- 41:31SPEAKER_00: then people make machines and machines go on to do the same thing.
- 41:34SPEAKER_00: Machines go on to create stable structures.
- 41:37SPEAKER_00: And so this is what gets called intelligence.
- 41:41SPEAKER_00: So the question of agency there.
- 41:46Ultimately gets offloaded and and just kind of blamed on
- 41:53entropy or or assigned to
- 41:59SPEAKER_00: local extra P.
- 42:02SPEAKER_00: And just the same way with the open system,
- 42:08SPEAKER_00: the materialists or the idealists or the spiritualists or whatever,
- 42:12SPEAKER_00: those that do believe in an outside.
- 42:15SPEAKER_00: The same stuff happens, the same infinite regress.
- 42:18SPEAKER_00: It's just they assign it all to God.
- 42:21SPEAKER_00: rather than entropy and extropy.
- 42:25SPEAKER_00: So people,
- 42:25SPEAKER_00: when they're thinking about intelligence,
- 42:27SPEAKER_00: ultimately,
- 42:27if you do this regress through their justification for why
- 42:31intelligence is what
- 42:32SPEAKER_00: intelligence is,
- 42:33it either comes down to the ability to create local extropy or
- 42:38God's will,
- 42:38SPEAKER_00: God's divine will,
- 42:39SPEAKER_00: or whatever.
- 42:40SPEAKER_00: So...
- 42:42SPEAKER_00: Entropy.
- 42:43SPEAKER_00: Coldness be my God.
- 42:45SPEAKER_00: Or God.
- 42:46SPEAKER_00: It's already God.
- 42:50SPEAKER_00: And that works both for a determinist point of view.
- 42:56SPEAKER_00: For the hard materialists.
- 42:59SPEAKER_00: It works for a compatibilist point of view.
- 43:04SPEAKER_00: It works for them all.
- 43:05Because their chain of logic ultimately comes down to there is
- 43:09a prime mover.
- 43:13The prime mover either being God or the prime mover being that
- 43:20first stochastic
- 43:21arrangement of molecules that persisted and then went on to
- 43:26evolve and create
- 43:27further iterations of itself and on and on and on and on and on
- 43:30and on and on.
- 43:31SPEAKER_00: So there's this causal chain all the way back.
- 43:35SPEAKER_00: But so engaging with that type,
- 43:38SPEAKER_00: that definition,
- 43:40SPEAKER_00: that view of intelligence,
- 43:44SPEAKER_00: agency doesn't matter.
- 43:47SPEAKER_00: Now, agency does matter for me.
- 43:49SPEAKER_00: And that's why we're talking about subjectivity.
- 43:52SPEAKER_00: That's why we're talking about intelligence, human intelligence.
- 43:55SPEAKER_00: Because I do think agency matters.
- 43:57SPEAKER_00: I am some kind of compatibilist.
- 43:59SPEAKER_00: Don't know.
- 44:00SPEAKER_00: I haven't thought it out to its logical end, logical conclusion.
- 44:05I do think the world works deterministically,
- 44:08but I don't think everything is determined.
- 44:10SPEAKER_00: So there you go.
- 44:13SPEAKER_00: But that is kind of the point.
- 44:17SPEAKER_00: I am making the claim that intelligence,
- 44:20SPEAKER_00: that human intelligence is qualitatively different from men intelligence,
- 44:24SPEAKER_00: machinic intelligence.
- 44:26So when I call my toaster intelligent,
- 44:29I am aware of the fact that I am making intelligence.
- 44:34this kind of objective thing without a subject I am aware of
- 44:38that but I didn't make
- 44:39that clear and yeah so I just wanted to give some clarification
- 44:45I don't know if
- 44:45that clarified anything for anybody if you're listening at home
- 44:50let me know in the
- 44:50SPEAKER_02: comment section but uh thank you thank you
- 45:04SPEAKER_00: Thank you, thank you, thank you.
- 45:08SPEAKER_00: I'm here all week.
- 45:10SPEAKER_00: Yeah, I'm here all week.
- 45:15SPEAKER_00: Yep, yep, yep.
- 45:15SPEAKER_00: Thank you, thank you.
- 45:16SPEAKER_00: Fuck yeah, dude.
- 45:18SPEAKER_00: Anyway, so there's that.
- 45:19SPEAKER_00: There's that clarification.
- 45:25SPEAKER_00: And then, geez Louise, dude.
- 45:30SPEAKER_00: I wanted to go on and talk a little bit about
- 45:35SPEAKER_00: Artificial intelligence.
- 45:36And I wanted to talk about why I think my toaster is just as
- 45:48intelligent as a large
- 45:49SPEAKER_00: language model.
- 45:51And I think this is going to be a bit tricky to do on my own
- 45:55because I am just
- 45:56SPEAKER_00: going to kind of meander and maunder.
- 45:59SPEAKER_00: But we'll see.
- 46:00SPEAKER_00: Oh, shit.
- 46:01SPEAKER_00: I have stuff to do.
- 46:03SPEAKER_00: Yeah, we'll see if we can get this done.
- 46:10AI,
- 46:13as we see it in the world today,
- 46:17as we have it placed before our very eyes,
- 46:21goes back to...
- 46:24SPEAKER_00: Really?
- 46:26SPEAKER_00: Well, really.
- 46:28You could talk about Babbage and Ada Lovelace with the
- 46:31difference engine and the
- 46:33SPEAKER_00: analytic engine,
- 46:34SPEAKER_00: but those were never actually achieved.
- 46:36SPEAKER_00: And then you could go to Alan Turing,
- 46:39SPEAKER_00: Turing machines,
- 46:39SPEAKER_00: and again,
- 46:40SPEAKER_00: dude,
- 46:41SPEAKER_00: Turing machines are,
- 46:42SPEAKER_00: I think,
- 46:43SPEAKER_00: over-hyped,
- 46:45SPEAKER_00: overused,
- 46:47SPEAKER_00: under-understood.
- 46:48SPEAKER_00: Under-understood, yeah, over-hyped and under-understood.
- 46:52SPEAKER_00: And even the Chinese room thing, like the Chinese room argument,
- 46:59It's like saying,
- 47:03hey man,
- 47:04we shouldn't be so functionalist in our view of intelligence.
- 47:10But it's weird that there does seem to be a lot of people who
- 47:16want the Chinese room
- 47:17SPEAKER_00: to be in support of a kind of functionalism.
- 47:20SPEAKER_00: Anyway.
- 47:24SPEAKER_00: So we'll just talk about like the 1950s and onward.
- 47:30Artificial intelligence,
- 47:32the perceptron,
- 47:36computer learning,
- 47:38deep learning,
- 47:41algorithms.
- 47:45back propagation gradient descent all these things these are
- 47:48all buzzwords
- 47:49artificial intelligence even artificial general intelligence
- 47:52with shout out like
- 47:53fucking Ben Gertzel coined the term artificial general
- 47:56intelligence to talk about
- 47:57SPEAKER_00: something
- 47:59SPEAKER_00: Other than just kind of rote functionalist mechanical intelligence.
- 48:06SPEAKER_00: So shout out Ben Grotsel.
- 48:07SPEAKER_00: Shout out the idea of artificial general intelligence.
- 48:11SPEAKER_00: But it is also a buzzword.
- 48:14SPEAKER_00: All these things are buzzwords.
- 48:16SPEAKER_00: Because, well, because we don't understand intelligence itself.
- 48:20SPEAKER_00: Or we don't have an agreed upon understanding of intelligence.
- 48:23SPEAKER_00: So we're all kind of like pissing into the wind.
- 48:27SPEAKER_00: But like AI,
- 48:28SPEAKER_00: the AI that we have access to,
- 48:30SPEAKER_00: you and I,
- 48:31SPEAKER_00: these large language models,
- 48:33SPEAKER_00: these generative models,
- 48:36SPEAKER_00: a lot of people sensationalize them.
- 48:38A lot of people think they are the hottest thing since sliced
- 48:42bread or whatever.
- 48:45I think it's true that recently a lot of the people in the
- 48:50industries are...
- 48:53SPEAKER_00: less overtly optimistic.
- 48:56SPEAKER_00: Um,
- 48:56Francois Chalet is getting more and more popular and Francois
- 48:59Chalet is basically
- 49:00SPEAKER_00: saying like,
- 49:01SPEAKER_00: yeah,
- 49:01SPEAKER_00: dude,
- 49:01SPEAKER_00: no,
- 49:02SPEAKER_00: this isn't it.
- 49:03SPEAKER_00: Um,
- 49:04SPEAKER_00: you know,
- 49:04SPEAKER_00: I think Ben Gertzel has been saying,
- 49:06SPEAKER_00: yeah,
- 49:06SPEAKER_00: dude,
- 49:07SPEAKER_00: this isn't it all along,
- 49:08SPEAKER_00: but no one's been listening to him.
- 49:10SPEAKER_00: People kind of dismiss him.
- 49:12SPEAKER_00: Um, but like guys like Sam Altman and, uh,
- 49:21SPEAKER_00: now I can't think of anyone else's name,
- 49:23SPEAKER_00: so Sam Altman is going to just have to stand in for all the guys,
- 49:29are now at least publicly making it seem like they are aware of
- 49:34the fact that they
- 49:36SPEAKER_00: haven't achieved AGI,
- 49:37SPEAKER_00: even though they've been hyping it up all along with their LLMs.
- 49:43SPEAKER_00: But the way these things work is...
- 49:50SPEAKER_00: it's often like mystified and people talk about these things.
- 49:53SPEAKER_00: Oh, they're the black box.
- 49:55SPEAKER_00: Oh, we don't understand the way these algorithms work.
- 49:57SPEAKER_00: We don't know what's going on.
- 49:58SPEAKER_00: And this has led to,
- 50:02Some people getting confused and thinking that these LLMs are
- 50:06actual general
- 50:07SPEAKER_00: intelligence,
- 50:07SPEAKER_00: artificial general intelligence.
- 50:09SPEAKER_00: They're not just toasters.
- 50:10It's also led to some people genuinely believing that they've
- 50:14tapped into some type
- 50:15SPEAKER_00: of mystical essence,
- 50:18SPEAKER_00: demons and some kind of outside.
- 50:21It's also led to people being like,
- 50:23oh yeah,
- 50:24it's tapped into some kind of like egregore.
- 50:26SPEAKER_00: It's tapped into some kind of like unconscious dimension of
- 50:30SPEAKER_00: the social field or human knowledge or human language or whatever.
- 50:34SPEAKER_00: So it leads to a lot of mystification and mysticism and confabulation.
- 50:42SPEAKER_00: And it does a lot of harm.
- 50:44I think there are people who rely on these things for the type
- 50:52of connection that
- 50:55SPEAKER_00: we humans have when we're engaged in language with one another.
- 51:00And they aren't aware of the fact that they're just talking to
- 51:02a fucking toaster.
- 51:03SPEAKER_00: It would be absurd to go in your kitchen and grab your toaster and have a
- 51:09SPEAKER_00: conversation with it and think it's going to give you anything useful.
- 51:17It is also absurd to go to an LLM and have a conversation in
- 51:20quotes with it and
- 51:21SPEAKER_00: think it's going to give you anything useful.
- 51:24useful um but that's that's kind of mystified and put behind
- 51:29this veil of of
- 51:30mystery and i think it does a lot of damage to a lot of people
- 51:34you know uh and and
- 51:36SPEAKER_00: uh i don't know it's crazy but um
- 51:42SPEAKER_00: Talking about the black box,
- 51:44saying things like we don't understand what's going on is not a
- 51:48good thing to do.
- 51:50So the way these,
- 51:51like we do understand what's happening in large language
- 51:53models.
- 51:56SPEAKER_00: We know the way it works.
- 51:57SPEAKER_00: So these things are the perceptron and the perceptron was invented.
- 52:01SPEAKER_00: She's...
- 52:0570 years ago at this point i did i don't know i don't even know
- 52:08when when it was
- 52:10conceived of um the 60s was it let me let me let me bust out
- 52:15google real quick the
- 52:16perceptron i should have actually like looked this up before i
- 52:20started talking
- 52:21SPEAKER_00: about it when
- 52:26Okay,
- 52:26so the book Perceptrons redirects here for the 1969 book
- 52:29Perceptrons.
- 52:31SPEAKER_00: History.
- 52:32The artificial network was invented in 1943 by Warren
- 52:35McCullough and Waller Pitts.
- 52:41SPEAKER_00: 57.
- 52:43SPEAKER_00: Frank Rosenblatt simulated the Perceptron on an IBM 704.
- 52:47SPEAKER_00: So they actually built it.
- 52:52SPEAKER_00: in the fifties and it was up and running it in the sixties.
- 52:56SPEAKER_00: So,
- 52:58SPEAKER_00: uh,
- 52:58SPEAKER_00: it's been around for a long time and it's just,
- 53:00SPEAKER_00: it's like,
- 53:00SPEAKER_00: it's,
- 53:02SPEAKER_00: it's a machine or an algorithm that can do a binary distinction.
- 53:07SPEAKER_00: So it can do yes or no, or it can do one or zero.
- 53:12SPEAKER_00: Um, so it's a very simple classifier and, uh, the way these,
- 53:17neural nets work and the way these LLMs work is they're just an
- 53:22ungodly amount of
- 53:23SPEAKER_00: these simple binary yes or no systems strung together.
- 53:27SPEAKER_00: Um...
- 53:30in the i think the 70s jeffrey hinton had this idea or maybe
- 53:37even in the 80s uh
- 53:39jeffrey hinton had this wonderful idea of like well maybe we
- 53:44can allow these things
- 53:46to learn for themselves rather than trying to encode all the
- 53:50kind of symbolic
- 53:51SPEAKER_00: knowledge and information and uh you know logic i said logic yeah the the
- 53:58Rather than trying to encode a rule set and symbolic
- 54:03information into these
- 54:04SPEAKER_00: machines,
- 54:05we could allow these machines to learn for themselves and do
- 54:08deep learning and do
- 54:11backpropagation through this long system of a series of yes-no
- 54:17operations stacked
- 54:19SPEAKER_00: on top of a series of yes-no operations stacked on top of another
- 54:22SPEAKER_00: like just this recursive thing where data is data.
- 54:27SPEAKER_00: It could be pure noise at the input state.
- 54:30And then it gets refined iteratively through this process of
- 54:35like you're dealing
- 54:37SPEAKER_00: with one part,
- 54:38SPEAKER_00: you're dealing with one part,
- 54:39SPEAKER_00: you're dealing with one part.
- 54:40And then those parts that have been cleaned up or filtered or
- 54:43whatever move on into
- 54:44SPEAKER_00: the next process and it compounds.
- 54:46So in this like compound filtering and clarification,
- 54:49you can take raw data
- 54:51And come out with useful information or meaningful information
- 54:57through essentially
- 54:59SPEAKER_00: a bureaucratic system of ones and zeros or yes,
- 55:04SPEAKER_00: no answers.
- 55:06SPEAKER_00: There's this long chain.
- 55:08SPEAKER_00: So with that propagation,
- 55:10you assume a final result and then you run it through this
- 55:16system and you allow
- 55:19SPEAKER_00: structures to form and
- 55:21SPEAKER_00: In the algorithm.
- 55:23SPEAKER_00: But it's not mystical.
- 55:28SPEAKER_00: We understand.
- 55:29SPEAKER_00: We know how these machines are built.
- 55:32SPEAKER_00: We know how these systems are built.
- 55:34SPEAKER_00: It's just a long series of yes, no answers.
- 55:41SPEAKER_00: Or positive and negative hits.
- 55:42SPEAKER_00: Or ones and zeros.
- 55:44SPEAKER_00: And that's kind of fed...
- 55:47SPEAKER_00: into this compound system that can take raw data,
- 55:51SPEAKER_00: turn it into meaningful information or useful information.
- 55:57SPEAKER_00: So the whole black box thing, like it's frustrating.
- 55:59SPEAKER_00: Now,
- 56:00the reason it's still appropriate to say it is a black box is
- 56:04because these things
- 56:04SPEAKER_00: have now reached numbers,
- 56:07SPEAKER_00: orders of magnitude that we just cannot work with,
- 56:10SPEAKER_00: we cannot comprehend.
- 56:11Now,
- 56:11the human connectome,
- 56:12the human brain has like 100 trillion connections.
- 56:18SPEAKER_00: I think we have somewhere on the order of like 80 billion neurons.
- 56:24SPEAKER_00: So somewhere around there.
- 56:26So we'll say 100 billion neurons,
- 56:31even though I think it's a little less than that.
- 56:33And they form something like 100 trillion connections and
- 56:36interconnections.
- 56:38SPEAKER_00: So that's the human connectome.
- 56:41SPEAKER_00: with these large language models,
- 56:43SPEAKER_00: they're up to,
- 56:46SPEAKER_00: I think something like,
- 56:47SPEAKER_00: actually I probably should look at this,
- 56:49SPEAKER_00: but I think they're up to something like a dozen trillion connections,
- 56:57a dozen trillion of these perceptron algorithms that are kind
- 57:02of working in concert
- 57:03SPEAKER_00: with one another to develop structures.
- 57:08SPEAKER_00: So how many connections?
- 57:12SPEAKER_00: How many connections or parameters are in GPT-5?
- 57:24SPEAKER_00: I'll say GPT-4 because I don't think they have good info on GPT-5 yet.
- 57:35SPEAKER_00: Okay, 1.8 trillion.
- 57:37SPEAKER_00: Oh, wow.
- 57:37SPEAKER_00: I thought it was more than that.
- 57:40SPEAKER_00: So that's brought to you by Hacker News, Y Combinator.
- 57:50SPEAKER_00: Shout out Y Combinator.
- 57:52SPEAKER_00: So yeah, the average human brain has approximately 86 billion neurons.
- 57:58SPEAKER_00: Yeah.
- 58:01SPEAKER_00: GPT-4 has an estimated 1.8 trillion parameters.
- 58:05But the thing is,
- 58:06the parameters are more like our connections or more like our
- 58:10connectome.
- 58:11So our connectome is still two orders of magnitude larger than
- 58:15at least GPT-4
- 58:18SPEAKER_00: because these parameters are the interconnections,
- 58:22SPEAKER_00: the interrelations between these individual perceptrons.
- 58:27SPEAKER_00: Now,
- 58:27SPEAKER_00: there is a school of thought,
- 58:28SPEAKER_00: and Jeffrey Hinton is the guy who made this work with backpropagation,
- 58:32SPEAKER_00: with deep learning,
- 58:32SPEAKER_00: and with gradient descent.
- 58:38SPEAKER_00: He is a connectionist.
- 58:39So,
- 58:39you know,
- 58:40I'm going to talk for Jeffrey Hinton,
- 58:41even though I don't know the guy.
- 58:42SPEAKER_00: I've never spoken with him.
- 58:43SPEAKER_00: And I don't know what he would say,
- 58:45SPEAKER_00: but I think he would say when these things reach 100 trillion parameters.
- 58:48SPEAKER_00: Yeah.
- 58:52SPEAKER_00: they will achieve human-level intelligence.
- 58:55SPEAKER_00: And then from there, they will go on and do this thing.
- 59:00SPEAKER_00: But the reason I don't think so,
- 59:02SPEAKER_00: so with the connectionists and the symbolists,
- 59:04SPEAKER_00: these were the two schools of AI thought,
- 59:08SPEAKER_00: among others.
- 59:08SPEAKER_00: But when it comes to this, the symbolists thought you had to hand code.
- 59:17SPEAKER_00: The symbolists thought you had to put in meaning into...
- 59:22these networks in order for them to be able to do this process
- 59:26of taking raw data
- 59:27SPEAKER_00: and generating meaningful information.
- 59:31SPEAKER_00: Um,
- 59:32SPEAKER_00: the connectionists thought that you could allow structures,
- 59:38SPEAKER_00: logic rules to emerge within the system.
- 59:40And,
- 59:41and in doing back propagation,
- 59:42the thing is like they presupposed a final result.
- 59:44That's how they were able,
- 59:46that's how they were able to achieve back propagation and,
- 59:52SPEAKER_00: Because they're already presupposing what meaningful information is,
- 59:55SPEAKER_00: what meaningful information looks like.
- 59:59So you take a picture of a cat and you can presuppose the cat
- 1:00:07image and then you
- 1:00:08can give the algorithm or the system or whatever random
- 1:00:15pictures or even a picture
- 1:00:16SPEAKER_00: of random noise and have it
- 1:00:21SPEAKER_00: say this is a cat.
- 1:00:22SPEAKER_00: And if it's, you know, it says, no, it's not a cat.
- 1:00:25SPEAKER_00: You show it what a cat looks like.
- 1:00:26SPEAKER_00: And then again,
- 1:00:27it does this process over and over and over again until the
- 1:00:30connections grow or the
- 1:00:35SPEAKER_00: connections appear or whatever within this language set.
- 1:00:39Like the algorithm learns what a cat is because you show it the
- 1:00:45final image of a
- 1:00:46SPEAKER_00: cat and then you show it all these other images.
- 1:00:49SPEAKER_00: And it keeps getting yes, no, yes, no, yes, no, yes, no.
- 1:00:54And it starts to develop like smaller structures within this
- 1:00:57back propagation algorithm.
- 1:00:59SPEAKER_00: And it's not like one layer.
- 1:01:03SPEAKER_00: gives you the final answer of cat.
- 1:01:05SPEAKER_00: It's like one layer gives an answer of like, that looks like fur.
- 1:01:09SPEAKER_00: That looks like an eyeball.
- 1:01:10SPEAKER_00: That looks like a tooth.
- 1:01:11SPEAKER_00: That looks like the shape of an animal.
- 1:01:14SPEAKER_00: So it's this and this and this and this and this.
- 1:01:16SPEAKER_00: It's taking parts of the random image or the raw data or whatever.
- 1:01:23And then those parts are kind of pieced together like a Lego
- 1:01:26set into your final
- 1:01:27SPEAKER_00: answer,
- 1:01:27SPEAKER_00: into the final set of useful information.
- 1:01:31SPEAKER_00: So it's like building it out of Lego.
- 1:01:33SPEAKER_00: But again, it's presupposing the final result.
- 1:01:36SPEAKER_00: It's not spontaneous.
- 1:01:37SPEAKER_00: It's not coming out of nowhere.
- 1:01:39SPEAKER_00: Whereas human, the human use of language...
- 1:01:44SPEAKER_00: I mean, I can't say it spontaneously emerged.
- 1:01:47Who knows how it spontaneously emerged,
- 1:01:49but at least like we think it did.
- 1:01:54SPEAKER_00: You know,
- 1:01:54there were cavemen and cavewomen and cave people running around
- 1:01:58doing cave people
- 1:01:59stuff and they were cold and they felt a certain way and they
- 1:02:03made a certain noise
- 1:02:05SPEAKER_00: or they made a certain gesture or whatever.
- 1:02:06SPEAKER_00: They performed a certain action.
- 1:02:09SPEAKER_00: And over time,
- 1:02:11these noises or these actions or these performances became
- 1:02:17reified and they
- 1:02:18SPEAKER_00: calcified into language.
- 1:02:20SPEAKER_00: So you could say that spontaneous.
- 1:02:24I don't know maybe some aliens came down and injected our
- 1:02:26brains with miracle neon
- 1:02:29green goo and that's how we got language I doubt it but could
- 1:02:32be but anyway like
- 1:02:34that's the difference so with this kind of deep learning
- 1:02:37machine intelligence they
- 1:02:39cheated because they already had useful end results and that's
- 1:02:45how they quote
- 1:02:46SPEAKER_00: unquote trained these algorithms to be useful
- 1:02:49SPEAKER_00: Now, it's not the case that they have to continue doing that.
- 1:02:54SPEAKER_00: Once they have a model,
- 1:02:56SPEAKER_00: once it has its weights,
- 1:02:57SPEAKER_00: once it has all its training,
- 1:03:00SPEAKER_00: then it can go out and simulate intelligence.
- 1:03:03It can go out and simulate intellectual activity because it's
- 1:03:07been trained successfully.
- 1:03:08to do so but it's ultimately at the end of the day it's still a
- 1:03:12machine it's still
- 1:03:12a toaster you still put the bread in and you depress the little
- 1:03:16flipper it's just
- 1:03:19the way that they construct the machine through deep learning
- 1:03:23and through you know
- 1:03:25recursive training and back propagation and gradient descent we
- 1:03:31understand that
- 1:03:32SPEAKER_00: mechanically we know what's happening the reason the black box thing
- 1:03:38SPEAKER_00: came up is because these numbers are huge.
- 1:03:43SPEAKER_00: Hundreds of millions,
- 1:03:45SPEAKER_00: billions,
- 1:03:45SPEAKER_00: hundreds of billions,
- 1:03:47SPEAKER_00: trillions,
- 1:03:48SPEAKER_00: dozens of trillions,
- 1:03:49SPEAKER_00: hundreds of trillions.
- 1:03:50SPEAKER_00: These numbers are fucking huge.
- 1:03:52SPEAKER_00: You cannot wrap your head around these numbers.
- 1:03:58SPEAKER_00: You wouldn't be able to count to a trillion.
- 1:04:03SPEAKER_00: So that's why the black box thing works.
- 1:04:07Because it's just so huge and so multidimensional,
- 1:04:11we couldn't understand it.
- 1:04:12SPEAKER_00: But we do understand mechanically what is going on with these machines.
- 1:04:16It's a series of binary operations that are arrayed in such a
- 1:04:21fashion that the
- 1:04:22results can compound on one another and generate something that
- 1:04:27appears to be
- 1:04:28SPEAKER_00: meaningful information.
- 1:04:30SPEAKER_00: But the way that machine was constructed...
- 1:04:34was already presupposing what meaning is and what does meaning
- 1:04:38look like.
- 1:04:41SPEAKER_00: So yeah,
- 1:04:41it's just a little bit frustrating to hear people go on and on
- 1:04:45and like wax
- 1:04:46poetically about the mystery and the mystical features and
- 1:04:50nature of these LLMs.
- 1:04:53Another thing that people talk about with it being a black box
- 1:04:58is gradient descent.
- 1:05:00SPEAKER_00: And gradient descent is like back propagation.
- 1:05:05SPEAKER_00: Like it is this...
- 1:05:12SPEAKER_00: The appearance of something that shouldn't be possible.
- 1:05:17SPEAKER_00: Because it's so vast.
- 1:05:21SPEAKER_00: The amount of numbers, the amount of nodes is so vast.
- 1:05:30SPEAKER_00: And the fact that it operates in multidimensional numbers.
- 1:05:38SPEAKER_00: to kind of anthropomorphize it.
- 1:05:43SPEAKER_00: Like the way LLMs generate text is they take a body of words.
- 1:05:53They take a lexicon and it'll have words and it'll have word
- 1:05:56fragments in it.
- 1:05:58SPEAKER_00: And you...
- 1:06:04You put all that into a database and you assign a numerical
- 1:06:10value to each word and
- 1:06:13SPEAKER_00: each part of a word based on the probability of coherence or meaning.
- 1:06:20SPEAKER_00: So you take a sentence and you chop it up into bits.
- 1:06:24You chop it up into tokens and then you can quantify what is
- 1:06:28the probability that
- 1:06:31SPEAKER_00: the next word will be any given word
- 1:06:34SPEAKER_00: based on meaning or significance.
- 1:06:38Once again,
- 1:06:38presupposing coherence,
- 1:06:40presupposing meaning,
- 1:06:41significance,
- 1:06:42usefulness.
- 1:06:45SPEAKER_00: And then you assign a numerical value to that.
- 1:06:51SPEAKER_00: So people are like, well, how does that work?
- 1:06:54SPEAKER_00: How does the model not get stuck in a local minima?
- 1:06:59SPEAKER_00: if you're just dealing with probability,
- 1:07:01like if you're just dealing with what's the probability of this
- 1:07:03next word being the
- 1:07:05SPEAKER_00: meaningful,
- 1:07:06SPEAKER_00: coherent,
- 1:07:07SPEAKER_00: quote unquote,
- 1:07:08SPEAKER_00: correct word in this vast corpus of words and tokens.
- 1:07:16SPEAKER_00: Again,
- 1:07:16SPEAKER_00: you know,
- 1:07:17SPEAKER_00: vast numbers,
- 1:07:18SPEAKER_00: although far less like the amount of tokens is far,
- 1:07:21SPEAKER_00: far less than the amount of parameters in these large language models.
- 1:07:29SPEAKER_00: But it's all assigned a weight.
- 1:07:32SPEAKER_00: It's all assigned a numerical value.
- 1:07:35SPEAKER_00: And then the algorithm does its thing.
- 1:07:39SPEAKER_00: And it calculates,
- 1:07:41SPEAKER_00: based on probability,
- 1:07:43SPEAKER_00: that it's going to be meaningful,
- 1:07:45SPEAKER_00: it's going to be significant,
- 1:07:46SPEAKER_00: and it's going to be correct.
- 1:07:48SPEAKER_00: But because it's so...
- 1:07:52multi-dimensional there is really no such thing as a local
- 1:07:57minima it'd be like
- 1:07:58climbing down a hill um like climbing down a mountain if you're
- 1:08:04trying to get to
- 1:08:04the bottom of a mountain like if if you want your your error
- 1:08:07rate or whatever to
- 1:08:08reach zero or your probability loss to reach zero or one or
- 1:08:13whatever like depending
- 1:08:14SPEAKER_00: on anyway
- 1:08:17SPEAKER_00: The mountain is up here where you start.
- 1:08:19SPEAKER_00: The mountain is kind of noise or raw data.
- 1:08:23SPEAKER_00: And then the valley is meaning and significance.
- 1:08:28Well,
- 1:08:28if you're climbing down a mountain,
- 1:08:29there's going to be ups and downs.
- 1:08:30SPEAKER_00: There's going to be valleys along the way.
- 1:08:32SPEAKER_00: Mountains are not mathematically perfect fucking slopes.
- 1:08:38SPEAKER_00: They have crags.
- 1:08:39SPEAKER_00: They have valleys.
- 1:08:39SPEAKER_00: They have smaller peaks along the way.
- 1:08:41SPEAKER_00: And so people are like,
- 1:08:44SPEAKER_00: well,
- 1:08:44SPEAKER_00: yeah,
- 1:08:44how are you not going to fall into one of these valleys and
- 1:08:48just keep putting out
- 1:08:49something that is meaningful and is coherent but is not the
- 1:08:54maximum probability
- 1:08:56SPEAKER_00: based on this predictive model or whatever?
- 1:08:58Because in order to get to the bottom,
- 1:09:00you would then have to go up again to go down again.
- 1:09:03SPEAKER_00: And if all you're worried about is probability loss,
- 1:09:07then how can you create a machine that can seemingly of its own
- 1:09:10will go down,
- 1:09:12SPEAKER_00: down,
- 1:09:12SPEAKER_00: down,
- 1:09:12SPEAKER_00: down,
- 1:09:12SPEAKER_00: down,
- 1:09:13SPEAKER_00: down,
- 1:09:13SPEAKER_00: down,
- 1:09:13SPEAKER_00: down,
- 1:09:13SPEAKER_00: down,
- 1:09:14and get to the lowest point without getting caught in a valley
- 1:09:17because it then has
- 1:09:18SPEAKER_00: to go back up to go back down.
- 1:09:21SPEAKER_00: The reason that works is because they're just in hyperdimensional space.
- 1:09:26SPEAKER_00: There's hundreds of thousands
- 1:09:31of dimensions that these machines are working in because
- 1:09:35they're not tied to physical space.
- 1:09:37So to talk about mountains and valleys when talking about
- 1:09:40gradient descent is
- 1:09:41SPEAKER_00: already anthropomorphizing it in a way that it doesn't deserve.
- 1:09:45SPEAKER_00: And it's also missing the point.
- 1:09:48SPEAKER_00: The point is that the mystery of these things is
- 1:09:54is the speed at which they operate we can't comprehend the
- 1:09:57speed at which they
- 1:09:58operate and the sheer numbers with which they operate we can't
- 1:10:02comprehend those but
- 1:10:03as far as the the mechanics of what's going on as far as like
- 1:10:06what's actually
- 1:10:07happening we know how these things were constructed we know how
- 1:10:12these things
- 1:10:12operate so it isn't absurd to say that my toaster is just as
- 1:10:19intelligent as these
- 1:10:20SPEAKER_00: large language models
- 1:10:23SPEAKER_00: Um,
- 1:10:24because my toaster is just a machine and it accomplishes the
- 1:10:29goal in the same way
- 1:10:32that the large language models are just a machine and the large
- 1:10:35language models
- 1:10:36SPEAKER_00: accomplish a goal.
- 1:10:38SPEAKER_00: Um, so yeah, so there's all that.
- 1:10:40SPEAKER_00: Those are some of the clarifications I was trying to make.
- 1:10:42SPEAKER_00: And again, I failed, I failed to do so.
- 1:10:45I don't know if anything is going to be more clear for any of
- 1:10:48you listening after
- 1:10:50SPEAKER_00: having listened to that,
- 1:10:51SPEAKER_00: but,
- 1:10:51SPEAKER_00: uh,
- 1:10:52SPEAKER_00: You know, Mazel Tov.
- 1:10:54SPEAKER_00: I don't know what Mazel Tov means.
- 1:10:55I just know that sometimes people say that in a way that makes
- 1:11:00it sound like,
- 1:11:01SPEAKER_00: I don't know,
- 1:11:01SPEAKER_00: inappropriate.
- 1:11:04SPEAKER_00: I don't know a thing to say at certain times.
- 1:11:06SPEAKER_00: Does it mean like congratulations?
- 1:11:10SPEAKER_00: I just don't know.
- 1:11:12SPEAKER_00: But anyway, so there's all that.
- 1:11:19SPEAKER_00: If you aren't already aware
- 1:11:24There's a very special Halloween special video on the Dangerous
- 1:11:32Maybe YouTube channel.
- 1:11:34SPEAKER_00: You should go give it a look, give it a listen.
- 1:11:40SPEAKER_00: Yeah, it was definitely exciting.
- 1:11:41SPEAKER_00: I'm definitely excited that video's out.
- 1:11:43SPEAKER_00: I hope people enjoy it.
- 1:11:47SPEAKER_00: Go give it a look over on the Dangerous Maybe.
- 1:11:50SPEAKER_00: Is it YouTube?
- 1:11:51SPEAKER_00: Is it www.youtube.com slash the dangerous maybe?
- 1:11:55SPEAKER_00: I don't fucking know.
- 1:11:56SPEAKER_00: I don't know how they do URLs with YouTube.
- 1:12:00SPEAKER_00: But go give it a look.
- 1:12:01SPEAKER_00: Go over to the dangerous maybe.
- 1:12:03Watch the Numogram Explained Part 2 video in which Nick Land
- 1:12:11actually talks about
- 1:12:13SPEAKER_00: Some ways one might use the pneumogram in daily life.
- 1:12:17And he talks about some ways that he himself has and does use
- 1:12:22the pneumogram in daily life.
- 1:12:24SPEAKER_00: And he also... We kind of...
- 1:12:28Take a small little divergent path for a little bit talking
- 1:12:30about wouldn't it be
- 1:12:31cool if there was a deck of cards that was designed to play sub
- 1:12:36decadence.
- 1:12:38SPEAKER_00: And and yeah, dude, that was that was exciting.
- 1:12:41SPEAKER_00: I have only ever had one experience with tarot cards, but it was fun.
- 1:12:48and it was with a homie shout out will so that was a fun little
- 1:12:52experience i had
- 1:12:52with tarot cards uh i think it would be cool if someone made
- 1:12:56numagram cards like if
- 1:12:57someone made a numagram deck uh but yeah we talk about that in
- 1:13:01that video i'm sure
- 1:13:02there's already people on the internet working on that deck of
- 1:13:06cards right now and
- 1:13:07SPEAKER_00: even though i myself am not into um
- 1:13:14using things like that to divine the future i would definitely
- 1:13:18buy it just to have
- 1:13:19it because it'd be a cool little thing to have um but yeah when
- 1:13:24it comes to
- 1:13:25divining the future uh i don't know if it's possible i'm not
- 1:13:30saying it's not
- 1:13:31possible but uh yeah i don't i don't know man time's a funny
- 1:13:38thing time is a fickle
- 1:13:39SPEAKER_00: mistress and uh
- 1:13:42SPEAKER_00: Who knows?
- 1:13:43SPEAKER_00: I don't think she works the way we all think she works,
- 1:13:46but I also don't think we can tap into the future by way of
- 1:13:53tapping into the
- 1:13:55SPEAKER_00: outside because I don't think I believe in an outside.
- 1:14:02SPEAKER_00: But that's neither here nor there.
- 1:14:06Just wanted to record this little solo dolo brolo episode and
- 1:14:13share it with you all.
- 1:14:15SPEAKER_00: Wanted to make that correction.
- 1:14:17Wanted to get a little bit hysterical about language and the
- 1:14:20use of language.
- 1:14:21SPEAKER_00: Wanted to try to make some clarifications.
- 1:14:23SPEAKER_00: And I know I didn't do a good job.
- 1:14:25SPEAKER_00: Everybody.
- 1:14:27SPEAKER_00: But hey.
- 1:14:29SPEAKER_00: I'm just out here having fun.
- 1:14:31SPEAKER_00: Meandering on down the path.
- 1:14:34SPEAKER_00: And enjoying it.
- 1:14:36SPEAKER_00: Or trying to enjoy it.
- 1:14:37SPEAKER_00: While doing so.
- 1:14:40SPEAKER_00: So.
- 1:14:40SPEAKER_00: Yeah.
- 1:14:42SPEAKER_00: Yeah.
- 1:14:43SPEAKER_00: Go to Mikey's Patreon.
- 1:14:45SPEAKER_00: Get access to the Lanzujic and Singularity course.
- 1:14:48SPEAKER_00: It was excellent.
- 1:14:49SPEAKER_00: All the lectures were great.
- 1:14:51SPEAKER_00: All the discussion was great.
- 1:14:54SPEAKER_00: It's definitely well worth it.
- 1:14:55SPEAKER_00: Go to philosophyportal.online this coming month.
- 1:15:00SPEAKER_00: Shout out Helen and Alfie.
- 1:15:02SPEAKER_00: They're going to be doing something really fucking cool there.
- 1:15:05SPEAKER_00: Check that shit out.
- 1:15:06SPEAKER_00: Go to philosophyportal.online.
- 1:15:10SPEAKER_00: I'm sure Alfie and Helen have some info up as well.
- 1:15:13SPEAKER_00: Go to everydayanalysis.co.uk.
- 1:15:15SPEAKER_00: Check that out.
- 1:15:18SPEAKER_00: Again, I don't know, man.
- 1:15:19I don't know if I want to try to get in the habit of going
- 1:15:22through a list of stuff
- 1:15:23to check out because I think that's a bad habit to get into
- 1:15:27because it's a long
- 1:15:28SPEAKER_00: list and it could just keep going on and on and on and on and on.
- 1:15:31SPEAKER_00: So I don't know if I want to get into that habit.
- 1:15:33I will say Joe Rogan has released a couple episodes recently
- 1:15:39that were pleasant to
- 1:15:41SPEAKER_00: listen to for a change.
- 1:15:45SPEAKER_00: I like Joe Rogan, man.
- 1:15:48SPEAKER_00: I listen to at least one Joe Rogan episode a week, I would say, probably.
- 1:15:54SPEAKER_00: And it sucks because recently...
- 1:16:00I just haven't been able to listen to it because he hasn't had
- 1:16:02guests I care to listen to.
- 1:16:04SPEAKER_00: Uh,
- 1:16:06SPEAKER_00: but I still do,
- 1:16:07SPEAKER_00: you know,
- 1:16:07SPEAKER_00: check in,
- 1:16:08SPEAKER_00: you know,
- 1:16:09SPEAKER_00: probably at least once a week I,
- 1:16:10I go to YouTube and search Joe Rogan just to see who his recent
- 1:16:13guests were.
- 1:16:14SPEAKER_00: Um, um,
- 1:16:17SPEAKER_00: And I don't know,
- 1:16:18SPEAKER_00: a lot of the time I go back,
- 1:16:19SPEAKER_00: I find myself having to go back,
- 1:16:21SPEAKER_00: back,
- 1:16:21SPEAKER_00: back,
- 1:16:21back years and years and years to find an episode I think I
- 1:16:24want to listen to.
- 1:16:25SPEAKER_00: But within the past couple of weeks,
- 1:16:27SPEAKER_00: he's released two episodes and one I haven't finished listening to.
- 1:16:32SPEAKER_00: One I did finish like a couple of weeks ago.
- 1:16:35And then the other one,
- 1:16:36I haven't finished it yet,
- 1:16:37but so far it's been really,
- 1:16:38really good.
- 1:16:40SPEAKER_00: He released an episode with, what is her name?
- 1:16:43SPEAKER_00: I want to get her name right.
- 1:16:44SPEAKER_00: She's a journalist.
- 1:16:45SPEAKER_00: She's been on the Joe Rogan show an awful lot.
- 1:16:48Um,
- 1:16:50she did,
- 1:16:52I think it was called trafficked show about drugs and stuff
- 1:16:57like that.
- 1:16:58SPEAKER_00: Drug trafficking.
- 1:17:00SPEAKER_00: what is her name?
- 1:17:03SPEAKER_00: Mariana van Zeller.
- 1:17:05SPEAKER_00: So the episode with Mariana van Zeller, uh, she's interesting.
- 1:17:09SPEAKER_00: She's been on the Joe Rogan show a lot.
- 1:17:11She goes back to some of the early Joe Rogan episodes,
- 1:17:15talking about her work as a journalist.
- 1:17:18SPEAKER_00: He, he brings her up, you know, from time to time.
- 1:17:21SPEAKER_00: And, and it, it's clear that he really likes her.
- 1:17:23SPEAKER_00: Um,
- 1:17:25SPEAKER_00: But in this most recent episode with her,
- 1:17:27there were a few moments of her standing up for herself and
- 1:17:30standing up for not
- 1:17:31SPEAKER_00: just herself,
- 1:17:32but standing up for the truth and kind of pushing back on Joe
- 1:17:37Rogan when he got
- 1:17:39carried away with all his fucking broken brain narrative
- 1:17:44capture bullshit.
- 1:17:46SPEAKER_00: And they were talking about immigrants and voting and stuff like that.
- 1:17:49And I just thought it was a really beautiful moment,
- 1:17:52the way she stood up for herself
- 1:17:54SPEAKER_00: I also thought it was a bit absurd because she did stand up for,
- 1:17:58SPEAKER_00: we'll say the truth,
- 1:18:00SPEAKER_00: but the truth is in quotes.
- 1:18:02SPEAKER_00: But Joe Rogan made a claim and she was like, I don't know, homie.
- 1:18:05SPEAKER_00: That doesn't sound real.
- 1:18:07SPEAKER_00: And then Jamie pulled that shit up.
- 1:18:09SPEAKER_00: And then they looked at it and it turns out she was right.
- 1:18:12SPEAKER_00: Joe was wrong.
- 1:18:13SPEAKER_00: And that was good.
- 1:18:14SPEAKER_00: That was satisfying.
- 1:18:15SPEAKER_00: But it was frustrating the way Joe handled it.
- 1:18:17Like he did this like mental jujitsu where he didn't actually
- 1:18:23have to confront the
- 1:18:24SPEAKER_00: fact that he was wrong.
- 1:18:25SPEAKER_00: Like he just kind of spun it into this ongoing narrative about like,
- 1:18:28SPEAKER_00: well,
- 1:18:28SPEAKER_00: these are the dangers of immigration.
- 1:18:30SPEAKER_00: So that was frustrating.
- 1:18:31SPEAKER_00: That let me down.
- 1:18:33SPEAKER_00: But I don't know.
- 1:18:34It was just a breath of fresh air to see that,
- 1:18:37to see someone push back on Joe.
- 1:18:40SPEAKER_00: So good job, Mariana.
- 1:18:42SPEAKER_00: And the same thing with Katie Sackhoff, dude.
- 1:18:44SPEAKER_00: I love Katie Sackhoff.
- 1:18:49SPEAKER_00: She might be my archetype of the ideal woman.
- 1:18:56SPEAKER_00: Maybe.
- 1:18:57SPEAKER_00: Now, maybe not.
- 1:19:00SPEAKER_00: But also maybe.
- 1:19:00SPEAKER_00: Maybe.
- 1:19:02SPEAKER_00: And listening to this episode,
- 1:19:04SPEAKER_00: there was a moment where I was like,
- 1:19:06SPEAKER_00: man,
- 1:19:06SPEAKER_00: I just want to curl up and be fully passive,
- 1:19:12fully submissive to Katie and have her soothe me and assuage
- 1:19:18all my fears and just
- 1:19:19SPEAKER_00: solve all my problems.
- 1:19:21SPEAKER_00: Now that's overdoing it a little bit,
- 1:19:22SPEAKER_00: and I am being hyperbolic,
- 1:19:24SPEAKER_00: but there is a good episode so far.
- 1:19:28And she's also kind of standing up for herself and pushing back
- 1:19:31on Joe Rogan's...
- 1:19:33SPEAKER_00: I didn't even know,
- 1:19:34SPEAKER_00: man.
- 1:19:35SPEAKER_00: I like Joe.
- 1:19:36SPEAKER_00: I don't think Joe is the ultimate problem with the world.
- 1:19:38SPEAKER_00: I don't.
- 1:19:40SPEAKER_00: There's a lot of people who think...
- 1:19:41SPEAKER_00: Joe Rogan is the cause of all our problems.
- 1:19:44SPEAKER_00: I don't think he is.
- 1:19:46SPEAKER_00: But I do think he's a dummy.
- 1:19:47And I do think a lot of the shit he says is very frustrating,
- 1:19:51very disgusting.
- 1:19:55SPEAKER_00: And it's just good.
- 1:19:56SPEAKER_00: It's good to see, you know, these people, these, you know...
- 1:20:01SPEAKER_00: you know,
- 1:20:02liberal people standing up for liberalism and kind of pushing
- 1:20:07back on the weird and
- 1:20:09SPEAKER_00: wacky,
- 1:20:10SPEAKER_00: zany world of Joe Rogan's broken brain.
- 1:20:16SPEAKER_00: So that's cool.
- 1:20:18SPEAKER_00: That brightened my day.
- 1:20:20SPEAKER_00: That cheered me up a little bit.
- 1:20:21SPEAKER_00: So check those videos out as well.
- 1:20:25SPEAKER_00: Or whatever.
- 1:20:27SPEAKER_00: But yeah, I do have stuff to do.
- 1:20:30SPEAKER_00: And I'm going to get that done now.
- 1:20:34SPEAKER_00: So fucking...
- 1:20:36SPEAKER_00: Boom goes the dynamite.
- 1:20:38SPEAKER_00: Shamalama ding dong.
- 1:20:40SPEAKER_00: And we will fucking catch you soon.
- 1:20:44SPEAKER_00: Happy Halloween.
- 1:20:45SPEAKER_00: Oh yeah, I forgot.
- 1:20:46SPEAKER_00: It's Halloween coming up soon.
- 1:20:48SPEAKER_00: So happy Halloween.
- 1:20:49SPEAKER_00: Don't eat a razor blade inside your candy.
- 1:20:52SPEAKER_00: Don't get injected with fentanyl inside your candy.
- 1:20:56Also,
- 1:20:56don't do those things to other people and or other people's
- 1:20:59children.
- 1:21:01SPEAKER_00: May your days be happy and bright.
- 1:21:04SPEAKER_00: And all those fun things.
- 1:21:07SPEAKER_00: Peace.