Latest / Newcomer Pod / AI Kills Us All (with Daniel H. Wilson)
Transcript
- 0:10Hey, it's Eric Newcomer. Welcome back to Cerebral Valley.
- 0:14This is Episode 2. I'm here with Max Child and
- 0:17James Wilsterman. We are going to dig into.
- 0:20We're calling this episode AI Kills Us All, which will give
- 0:25you a taste. It's not not the small
- 0:28questions. You know what what to do about
- 0:30privacy or like you know who should be in charge.
- 0:34It's like, what happens if we get artificial general
- 0:37intelligence? What, how worried are we about
- 0:40sort of the big doom and gloom? Guys?
- 0:43Great. Great to be back with you.
- 0:46Great to be back. Thanks for having us.
- 0:47Great to be here. After my conversation with Max
- 0:49and James, I will be interviewing Daniel H Wilson,
- 0:53the author of How to Survive a Robot Uprising.
- 0:55Where's my jetpack? And how to build a robot army?
- 0:58And the best seller, Robo Apocalypse.
- 1:00So someone who has spent a lot of time imagining the doom and
- 1:05gloom of machines getting powerful.
- 1:09So that's he. He's an expert in this
- 1:11speculation where we are mere amateurs, you know, with of
- 1:14course, the help of chat. GPTI was, you know, looking,
- 1:19wandering the Internet, looking for the great thinkers on AG is
- 1:28coming, and one of them is Nick Bostrom, who's been writing
- 1:32about this for years, I think. So this piece is the essay from
- 1:35his book Super Intelligence. The essay is called The
- 1:41Unfinished Fable of the Sparrows, which is very
- 1:45enjoyable and I I think you'll get a lot out of it.
- 1:50It was the nest building season, but after days of long, hard
- 1:53work, the sparrows sat in the evening glow, relaxing and
- 1:57chirping away. We are all so small and weak.
- 2:00Imagine how easy life would be if we had an owl who could help
- 2:03us build our nests. Yes, said another, And we could
- 2:07use it to look after our elderly and our young.
- 2:10It could give us advice and keep an eye out for the neighborhood
- 2:13cat, added a third. Then past us, the elder bird
- 2:17spoke. Let us send out scouts in all
- 2:19directions and try to find an abandoned outlet somewhere.
- 2:23Or maybe an egg a crow chick might also do.
- 2:26Or a baby weasel. This could be the best thing
- 2:29that ever happened to us, at least since the opening of the
- 2:32pavilion of Unlimited grain in Yonder Backyard.
- 2:36The flock was exhilarated and sparrows everywhere started
- 2:39chirping at the top of their lungs.
- 2:41Only Skronk Finkel A1 eyed Sparrow with a fretful
- 2:45temperament was unconvinced of the wisdom of the endeavor.
- 2:48Quoth he? This will surely be our undoing.
- 2:51Should we not give some thought to the art of owl domestication
- 2:55and owl taming first, before we bring such a creature into our
- 2:59midst replied past us, taming an owl sounds like an exceedingly
- 3:03difficult thing to do. It will be difficult to find an
- 3:07owl egg, so let us start there. After we have succeeded in
- 3:10raising an owl, then we can think about taking on this other
- 3:14challenge. There is a flaw in that plan.
- 3:17Squeak Skronk Finkel. But his protests were in vain,
- 3:20as the flock had already lifted off to start implementing the
- 3:24directive set out by Pastis. Just two or three sparrows
- 3:28remain behind. Together, they began to try to
- 3:31work out how owls might be tamed or domesticated.
- 3:35They soon realized that pastis had been right.
- 3:37This was an exceedingly difficult challenge, especially
- 3:40in the absence of an actual owl to practice on.
- 3:44Nevertheless, they pressed on as best they could, constantly
- 3:48fearing that the flock might return with an owl leg before a
- 3:53solution to the control problem had been found.
- 3:58All right, what do you guys make of that?
- 3:59It's. Not quite Plato, but it's.
- 4:04Heavy-handed, little on the nose, I feel like this might
- 4:12have been more. You know, mind blowing in 2014
- 4:16when when it was raining, I guess, but nowadays it's just
- 4:19like, Oh yeah, that's that's what we're talking about all the
- 4:21time. But I do think, I do think this
- 4:27obviously gets at this question of, you know, should you ever,
- 4:31should you even go down this path at all, right.
- 4:33I think that's the core question in develop trying to develop
- 4:37AGI. Or is it so risky that you
- 4:41should? You know not do that at all
- 4:43unless you are 100% confident you figured out how to tame the
- 4:47OWL right in advance. Is that your guys reading as
- 4:51well? Yeah, well, I think just to like
- 4:54rewind one second, I think coming back to this like AI
- 4:57kills us all point. I think there's two big
- 5:00questions in the sort of will AIS kill, kill us all narrative,
- 5:03right. Or maybe three.
- 5:05I guess one will be are we capable of developing an AI
- 5:09that's smarter than humans? Like 2 will be if we're capable
- 5:14of developing an AI smarter than humans, would it want to kill
- 5:17us? And then three being, if it
- 5:20wants to kill us, can it figure out some way to kill us?
- 5:23Basically, right. I mean like and I think those
- 5:25are all kind of three different questions and maybe we all like
- 5:28agree on this podcast on like certain one or two or three or
- 5:32or different elements of them, but like.
- 5:34I think it's helpful to. Yeah.
- 5:35To break it down that way because.
- 5:37And if you do that, this, this story, the parable, the Fable
- 5:40that I just read sort of assumes a couple of those, right.
- 5:43It assumes by using an owl, it assumes that such a thing could
- 5:48exist, right. And that it, it's a dangerous
- 5:52being that we actually need to figure out in its essence, it's
- 5:55dangerous and we need to figure out how to control it before we
- 5:58create it, right. So it sort of assumes into the
- 6:00story certain things exactly that you're framing up are open
- 6:04question. I think they're open question.
- 6:05I mean in the parable basically the OWL both is smarter and more
- 6:08powerful than the sparrows, definitely wants to apparently
- 6:13kill or very likely potentially wants to kill the sparrows and
- 6:15certainly has the capability given that it's bigger and
- 6:17stronger or whatever, right. So to your point, it's sort of
- 6:19like you know, the answer is kind of written into the story,
- 6:22whereas I think in real life like I think can we build AI
- 6:25smarter than us. Part one is actually a pretty
- 6:27interesting and contentious question and then the other ones
- 6:29are also pretty interesting and contentious as well.
- 6:31But just to start on part one like.
- 6:34Do you guys think we can build an AI that's smarter than
- 6:36humans? I guess like, let's be very
- 6:40specific about how we define that, right?
- 6:42Because I think it's a pretty critical question.
- 6:45Come on. Well, what's your definition of,
- 6:49I guess, smarter than humans? Is that a?
- 6:52Are you saying super intelligence?
- 6:53We're talking about, say, an. IQ That has never been achieved
- 6:56by a human right. You know I.
- 6:57I think that like I think there's kind of this concept of
- 7:01a IAGI, which is. Better at I would say like the
- 7:06median human at, you know, most tasks and then there's super
- 7:11intelligence, right? Which is you know, better like
- 7:14better than the best humans at most tasks, right.
- 7:17Is that how you guys think about it as well?
- 7:19Yeah. Let's say better than the best
- 7:20humans at any mental task or any any relevant mental task, right?
- 7:25I guess I believe that this is going to happen, I think.
- 7:29You know, it would be naive to say, well it won't happen.
- 7:32So then it's just a matter of a question of like when it will
- 7:35happen and whether there are any kind of domains where it feels
- 7:40like it won't happen for a long time, right?
- 7:43Like are there, you know, any domains that we really think you
- 7:47know it's going to be take take much longer to achieve super
- 7:49intelligence than than other domains, right.
- 7:53So maybe it will be the best. Coder in the world.
- 7:57But will it also be the best screenwriter?
- 7:59Are those things gonna happen at the same time or are they gonna
- 8:02happen, you know, at different times?
- 8:05Yeah, I I, I mean, first of all, you know, AI is smarter than the
- 8:11smartest humans in some domains right now.
- 8:14Already shall you go Wow games particularly.
- 8:18Right. But that that is not a general
- 8:20intelligence. No, no, I'm just saying, just.
- 8:22Yeah, but in specific demands it's better.
- 8:24I I think it will continue to cleave off domains, right.
- 8:28I think like sort of beat all humans at the LSAT type thing
- 8:32seems very soon. I I think what feels far away is
- 8:37sort of and it fits into the fears of an AGI is sort of the
- 8:42strategic planning like what an AGI looks like that sets like
- 8:45broader priorities across like in a game it's like you want to
- 8:49win at chess, like straightforward, but when it has
- 8:52to you know sort of solve general optimal outcomes.
- 8:56Like it feels like we are so far from a being that could do that.
- 9:01Or a computer that's interested in sort of deciding whether it
- 9:05should play chess or do something else and like why
- 9:08especially if we're not again just articulating what the goals
- 9:11are. So I guess I'm, I'm not as
- 9:14bullish on in the next decade a sort of coherent overall sort of
- 9:19being that feels like smarter than a human.
- 9:22I mean the next decade, though. I mean, if we're already
- 9:24negotiating this down to the next decade, it would seem like
- 9:27you're pretty confident that it's gonna happen.
- 9:28Like lifetimes. I think it's gonna happen, OK.
- 9:30So you think in our lifetimes that will happen?
- 9:32Yeah. What about my my earlier
- 9:35distinction though Eric, do you do you think you know an AGI
- 9:38that is a general intelligence that is basically better than
- 9:41the media and human at most things is or or equal to is more
- 9:46likely to happen in the next decade.
- 9:48Is that five years? I just think it's hard to do it.
- 9:51I we just don't see it as sort of a an overall agent or like an
- 9:55overall decision, you know what I mean?
- 9:57I I see it like winning at a bunch of discrete tasks.
- 10:01OK, sort of a sign. So this kind of gets at more of
- 10:04like your definition of AGI potentially would require some
- 10:09sort of Turing test type thing that it can really just convince
- 10:13you that it is a human. Or I guess I'm trying to think
- 10:15of like an overall thinker with like sort of priorities and sort
- 10:20of. But at the end of the day, like,
- 10:22just to make any predictions about this, it has to kind of be
- 10:25a falsifiable assertion of like, what is AGI, right?
- 10:30And it's not that easy to to define that except by some sort
- 10:34of like test, right? Otherwise you end up in this
- 10:37like era or this pattern we we're always in that we are
- 10:41constantly moving the goal posts on like what AGI is, right?
- 10:44Right. Yeah.
- 10:46I mean, I think, I mean, I think you guys are both like in the
- 10:49same place, which I am, which is like it's very hard to figure
- 10:51out the timeline, but like to me it's definitely happening
- 10:54sometime in the next 100 years, let's say.
- 10:56Whether it's 10 or 20 or 50 or 100, it is a little bit harder
- 11:00to guess, but like it is kind of a math problem, right?
- 11:03Where we have brains that are just some giant collection of
- 11:06neurons, right? And there's, you know, millions
- 11:09and millions of them or whatever, whatever order of
- 11:12magnitude it is, right? And essentially we're training
- 11:15these large language models on, you know, a neural network in
- 11:18some ways, right? Which, you know, it's not a one
- 11:21to one copy of what? A neuron, how a neuron works in
- 11:24our brain, which is way more complicated, but.
- 11:26Fundamentally, you're sort of like, if I can just throw
- 11:28numbers at this problem of like building a brain essentially and
- 11:32like get more GPU cores, like just throw absolute just
- 11:37computing power at this thing. Like the story of our entire
- 11:39lives has been computing power just keeps going up
- 11:41exponentially, right? You know, Moore's Law or
- 11:43whatever, right? And so to me, it seems like
- 11:46you'd have to believe there's something fundamentally
- 11:48different about the way a human brain works or a neuron works.
- 11:51To the way you know a neural network or you know the way GPU
- 11:55transformer model works to to believe that we can't just throw
- 11:58numbers at this thing until it gets smarter than humans.
- 12:01And so I guess I do believe that we throw enough numbers at it
- 12:03and get smarter than humans. I mean humans have to win
- 12:06multiple sort of parts of the brain that developed in
- 12:09different ways that interact with each.
- 12:11Other. Yeah.
- 12:12But they're still kind of built on the same cell structure.
- 12:15I mean, yeah, yes, but but you could do that in an in an AGI
- 12:18too, right? You could have different lobes
- 12:19or whatever of the imaginary intelligence, right?
- 12:21And then broadly, we just really don't understand how to pinpoint
- 12:26the experience of human consciousness, like quality or
- 12:30whatever. I talked about this at length
- 12:31with Reed Hoffman when I interviewed him and like so, but
- 12:36we're going to, but we're very reluctant because we can
- 12:38experience our own consciousness to attribute it and animals and
- 12:42we really have no like road map for how we'd ever identify it in
- 12:46machines. And that's sort of different
- 12:48than the Super intelligence thing, but it's certainly part
- 12:50of what you want in the sort of like does it have like a real
- 12:53existence sort of question, which I don't even have a road
- 12:56map for how we'd identify. Coming back to the core topic of
- 12:59the episode, like, will AI kill us all?
- 13:01Like, it doesn't really matter if it's conscious or not, right?
- 13:04If it ends up having the power and the desire to kill us,
- 13:06right. So like, I do think that's an
- 13:09interesting question and I don't know, maybe that's a different
- 13:11episode. But like to me that's not,
- 13:13that's not worth getting hung up on in the like is it going to
- 13:15kill us all thing like definitely.
- 13:17I mean, this sort of classic example being the paper clip
- 13:21maximizer, right? The machine, the fear that we
- 13:25just program this really smart, you know, LLM and tell it like,
- 13:29oh, you're not only objective is maximize paper clips, you know,
- 13:34And then it goes about it and then it's like, well, the humans
- 13:36are hurting sort of the maximization of paper clips.
- 13:39Let's get rid of them. I mean famously like I Robot by
- 13:43Asimov. It's like robots are supposed to
- 13:46protect humans, but then humans all kill each other.
- 13:49So it's like, oh, we need to protect the humans.
- 13:51I sort of need to enslave them in order to ensure their
- 13:54protection, you know, so that doesn't it doesn't necessarily
- 13:58need to have the depth of thinking.
- 13:59You know, humans would see sort of the flawed what we would
- 14:02perceive to be flawed reasoning in those things, but it would
- 14:05still but a machine. You could see why a machine with
- 14:08certain objectives could operate.
- 14:10Though well, I I think it's sort of like saying that these
- 14:13artificial intelligences can have totally different
- 14:17orthogonal objectives that don't require a consciousness, right?
- 14:21Like they just for some reason develop a different objective.
- 14:24Than we have as humans, and they're capable of bringing
- 14:29about that objective to the detriment of humanity,
- 14:32basically, right? Yeah, I mean, a nuclear weapon
- 14:35doesn't have consciousness, but it's still very effective at
- 14:37killing people, right? That's its mission, so.
- 14:41Yeah. I mean, so I, it seems like we
- 14:43all agree, OK, we're going to achieve super intelligence at
- 14:45some point in the next, let's say, century or an intelligence
- 14:47that's smarter than humans and enough domains that it's, it's
- 14:50meaningfully, you know, more capable than we are, right.
- 14:52Yeah. So second question, like, you
- 14:56know, would it want to kill us? I think this is like one of the
- 14:58most interesting ones and you alluded to the sort of paper
- 15:01clip maximization as one example of how it might just end up
- 15:04killing us as like a side effect of some other mission it's on.
- 15:07Right, but. I guess A, like, how likely do
- 15:10we think that is? And then B, like, do we think
- 15:13that we'll be able to constrain its missions in some way so that
- 15:17it wants to keep us alive, you know, even if it's going off and
- 15:21executing at school? I do actually, you know, worry
- 15:25about this. I guess that the AI will have
- 15:29some orthogonal objective that is detrimental to humans.
- 15:33And then there's like a secondary question of like, can
- 15:35we do anything about that, right.
- 15:37We were like, I mean to read between what you're saying,
- 15:39you're like, if anthropic and open AI are left only in charge,
- 15:42we'll be fine. But if Facebook's llama runs
- 15:45around or screwed, I mean, that's sort of, I mean, I'm
- 15:47interested to see what Mustafa, who's going to speak at Cerebral
- 15:50Valley, you know, he's been much more reticent about open source,
- 15:54potentially. His lack of control issue.
- 15:57I do think that eventually if we start to see, you know, the
- 16:03risks like that, we all agree that.
- 16:06You know this, we're headed towards this super intelligence
- 16:08and we start to see how dangerous that may or may not
- 16:12be. Like, I just don't think we know
- 16:13yet how dangerous it will be. Like it's obvious to me like you
- 16:17want to maintain the ability to regulate this.
- 16:20And you know, just a lot of people, I think I saw Brian
- 16:24Armstrong from Coinbase, you know, tweet that he just thinks
- 16:27there should never be any AI regulation or I don't know if he
- 16:30said never, but he was basically like regulation has such a poor
- 16:33track record around innovation. Like, I just think that's sort
- 16:36of naive without us knowing yet, like how dangerous it can be or
- 16:42not. You just need to have the best
- 16:44AI on our side. You know, like when the killer
- 16:46one comes, we need to have the benevolent 1.
- 16:48The best defense against a bad guy with an AI is a good guy
- 16:52with an AI. That's how that's how Americans
- 16:55actually think. You know, like.
- 16:57The right to bear AI, baby, Yeah.
- 17:01That was kind of Sam Altman's perspective when he started Open
- 17:04AI. Like, I mean, at least my
- 17:05reading. Of it when he was getting
- 17:07started was calling it open AI because he believed putting AI
- 17:11in the hands of humanity democratizing it was the best
- 17:15way to go about things. And I think he's changed his
- 17:18mind because of this exact thing that I'm saying, which is it's
- 17:21actually pretty dangerous or potentially dangerous.
- 17:24The fact that opening eye isn't open at all I know is like
- 17:26something that has been beaten to death, but it's just still
- 17:29ridiculous to me just that. It's hilarious kind of, yeah.
- 17:32Yeah, cuz it's historical artifact or?
- 17:34Something, right. But I mean I think I guess
- 17:37putting aside the regulation thing and the open source is
- 17:40closed source thing, I just think that if you believe that.
- 17:43A substantially diverse group of people from different countries
- 17:47within America, open source, closed source, can develop a
- 17:49super intelligence, right? Like if you believe there's
- 17:52going to be more than like five of these, which I think there
- 17:55would be like in any scenario in which this actually happens,
- 17:58like isn't it just very likely that one of the five or ten or
- 18:02hundred or thousand of these is is a bad AI?
- 18:05Mean? I just don't.
- 18:06I don't think it matters if you have 999 aligned good AI and one
- 18:09bad one, if that one bad one is capable of destroying the world,
- 18:12right? I mean like.
- 18:13It again, I just, I just think that mathematically if it's
- 18:17possible to create super intelligence is like someone is
- 18:20going to create a super intelligence that's really bad,
- 18:21right? And I'm not sure like our
- 18:23existing analogy is around like guns or nuclear deterrence or
- 18:27whatever sort of game theory you want to play out here like apply
- 18:30where like if anyone bad AI can get access to the tools that
- 18:34needs to kill everyone, it will it will succeed, right?
- 18:38Like, I don't think all the good AI can stop it, but but maybe
- 18:40I'm going too far out on this. I mean, I agree with it.
- 18:43I think so. So your take is that obviously
- 18:46yes, the answer is yes. You know, if we achieve super
- 18:50intelligence, there's a pretty big risk to that.
- 18:53I mean, in the world. You would have to create some
- 18:56way for the good AI to regulate the bad AI.
- 18:58Basically, like if you believe we're in a world where no longer
- 19:01humans are, they're not on the driver's seat, right?
- 19:03We're in the back seat, right? Like you now have to basically
- 19:05try to figure out how to make sure that the AI that are in the
- 19:07front seat can control the other ones.
- 19:10I mean part of what Max is basically referencing just, you
- 19:13know, to give a little context, is Elisa Yudakowski.
- 19:17I mean I pulled actually a quote, you know, to visualize a
- 19:20hostile superhuman AI. Don't imagine a lifeless book,
- 19:23smart thinker dwelling inside the Internet and sending I'll
- 19:26intentioned emails. Visualize an entire alien
- 19:28civilization thinking at millions of time human speeds,
- 19:32initially confined to computers in a world of creatures that are
- 19:35from its perspective, very stupid and very slow.
- 19:38A sufficiently intelligent AI won't stay confined to computers
- 19:41for long. In today's world, you can e-mail
- 19:43DNA strings to laboratories that will produce proteins on demand,
- 19:47allowing an AI initially confined to the Internet to
- 19:50build artificial life forms or bootstraps straight to post
- 19:53biological molecular manufacturing.
- 19:55I don't know. My point of view is just like, I
- 19:59think him and sort of what you guys are getting at, it just
- 20:01feels like you're like, oh it's super intelligent.
- 20:03It's like God, like, you know, I I still think, OK, it's way
- 20:06smarter than us. But we're like a bunch of
- 20:09intelligent beings, even if it's smarter than us.
- 20:11I I mean, it requires that that it's going to act stealthily and
- 20:15sort of like start manufacturing like a biologic version.
- 20:18You know, like, I just feel like it's more likely that we're
- 20:20like, this thing doesn't really listen to us and like, but then
- 20:24like the GPU clusters are like, it's trying to fire them all.
- 20:27And it's just like, oh, there's only so much computing power
- 20:30that, you know, it's still a being that requires like
- 20:33resources. You know, like humans are
- 20:35smarter than everything else. We still die like they're
- 20:37they're just like physical limits to things and and that
- 20:42it's not going to have like God like powers.
- 20:44It's going to be limited by the amount of compute it can access
- 20:47and sort of our cooperation at various points.
- 20:50Like, I I don't know. I don't think it's on zero to
- 20:53God. I feel like you're painting.
- 20:55Correct me if I'm wrong, but you're painting a picture where
- 20:57we have created super intelligence.
- 21:01And you agree with Max's point that some of them are misaligned
- 21:05and probably would want to kill us if they could.
- 21:07However, you're saying that the humanity's fine because those
- 21:13misaligned AIS can't get enough resources.
- 21:16Is that accurate, right, I mean. Just it's hard to know, right?
- 21:20We don't. We don't.
- 21:21We don't know what to do. That's definitely a gamble.
- 21:24I mean, even if you believe that you're like.
- 21:28You've got you've gotten pretty close to the edge.
- 21:33Well I'm I'm definitely a fatalist on AI as my if we're
- 21:36getting to like our deep, like what is our position.
- 21:38It just, I just think like many people in tech, like the
- 21:42regulation is pretty bad and like going to be only going to
- 21:46stop the good guys, that this is going to keep going and that I'd
- 21:50rather the best people we have try to figure it out rather than
- 21:54like operate in the shadows. Yeah, I I tend to agree with you
- 21:58and I probably would agree that we're not, we don't need a lot
- 22:02of regulation right now on you know who can you know train
- 22:08these models or something like but I.
- 22:11I also wouldn't rule out like meeting this in the future if in
- 22:14a year from now or two years from now I start seeing you know
- 22:17a lot of like evidence that we've, the open source community
- 22:20has created super intelligent misaligned AI, right.
- 22:24And I think people are kind of, you know, going a little bit too
- 22:28far in the direction of saying this, you know, we don't need
- 22:31it. We need any regulation here.
- 22:33When it's a new technology, we just don't know what it's going
- 22:36to be like. We we don't, you know, and it is
- 22:39a little bit like. When we were, you know, putting
- 22:43the right to bear arms in the Constitution without knowing
- 22:45what arms meant in the future, right?
- 22:48Like how, how, how dangerous they may or may not be.
- 22:52So I think that I'm just a little nervous about, you know,
- 22:55assuming too much about whether we will need regulation in the
- 22:59future. I mean to come back to you at
- 23:02Kowski for a second, I mean he is sort of the the progenitor of
- 23:06this whole AI. Dumerism, concept or whatever.
- 23:09I mean, I've read a bunch of his stuff he seemed to almost like
- 23:11at one point, threaten blowing up GPU.
- 23:14Right, Yeah. Well, he had a time article that
- 23:17he was saying in his world of protecting humanity, like one of
- 23:23the only things you would have to do would be to.
- 23:27You know, protect, you know, prevent rogue data centers from
- 23:30training models. And, you know, it'd be
- 23:32justifiable to like, bomb them essentially.
- 23:34And and he was saying, you know, the entire kind of geopolitical,
- 23:41you know, consensus should be that it's worse to train these
- 23:45models at some level than it is to, like, worry about nuclear
- 23:49proliferation. I think.
- 23:52I mean, I having read a bunch of his stuff, I feel like he has
- 23:54like pretty well convinced me of sort of all all three of these
- 23:58things. I guess that will have a super
- 23:59intelligence that it'll it'll be it'll it'll some of them at
- 24:03least will desire to harm us and that they'll figure out some way
- 24:07to do it. I guess I'm sort of like I'm a
- 24:09little bit of a fatalist in the sense that I I think this is
- 24:11basically an unstoppable train at this point that has like left
- 24:14the station. Like I don't, I don't think.
- 24:16Bombing data centers is like a viable solution to this problem
- 24:20or breaking GPUs or whatever or certainly not in the like the
- 24:25human political universe that we actually operate in today.
- 24:28I think my, my attitude is just that like luckily, I guess my
- 24:34the future is really hard to predict, right?
- 24:36I mean, I know that's like a very obvious point, but if you
- 24:39had told someone in 1940 that we were going to invent nuclear
- 24:42bombs and they had really thought through all the.
- 24:45You know, the potential problems in the future caused by those
- 24:48devices, I think they could have been pretty fatalistic and I
- 24:51think that would have been a pretty accurate read on like how
- 24:53bad it was that we invented nuclear bombs, right?
- 24:56Like knock on wood, so far we're still here.
- 24:58So my sort of my fear is that all the stuff you you had cows
- 25:01and some people it's going to happen.
- 25:02Nuclear bombs are good, like they brought about like peace.
- 25:06We haven't had world wars, you know, But my hope.
- 25:09Is that we get lucky and that. Something about this future that
- 25:13was that Yukowski or the AI doomers are sketching out is is
- 25:16there's some missing link in the logical chain or or there's some
- 25:20piece of the way deterrence or or alignment or whatever ends up
- 25:24working that that we we get our ass saved.
- 25:27But I'm kind of betting on luck at this point.
- 25:29Like, I don't really think we're capable of stopping this.
- 25:31Like just if we're all fatalist basically.
- 25:35I mean there is like the human cloning, you know, people
- 25:38brought this up to me. Like human cloning is an example
- 25:41where people have, you know, America isn't cloning humans.
- 25:44It seems like we could like why? Why is that an area where self
- 25:49regulation seems much more possible?
- 25:53I feel like there are areas where we have stopped.
- 25:56That's still pretty hard, right? I mean, like, I bet that will
- 26:00happen sometime in the next 50 to 100 years, right?
- 26:02I don't know. I just feel like I don't know
- 26:04how much desire there is for that.
- 26:06But I I feel like once it's doable, people will do it.
- 26:08My last thinking on one positive future scenario that I sort of
- 26:11believe is possible is like, you know, we live on this planet
- 26:15with lots of other organisms and creatures as you guys described.
- 26:18You know, we have pets, we have dogs, whatever.
- 26:21Like. Why do we have dogs?
- 26:23Like, is there any really good reason?
- 26:24Like the question is like, could we end up being the pets of the
- 26:27AI? Like, could we could they want
- 26:29to keep us alive for some reason?
- 26:30And that we're a little bit different?
- 26:31We're a little bit interesting to them, more fun to have around
- 26:34in some scenarios, but like, in the end we're the AI's pets, you
- 26:37know? Or or maybe we're the ants and
- 26:39they just don't care to kill us all because it would be such a
- 26:41pain, You know, right? I think that's more likely than
- 26:45not. Right.
- 26:46I think that is the the upside scenario is like we're we're
- 26:49pets or we're ants or something compared to this.
- 26:51More likely than not just the general, like the AI does not
- 26:55hate us, you know what I mean? Like both.
- 26:57We helped create it. Like we set a lot of its initial
- 27:01value system, presumably like it, you know, relies us on us
- 27:05for a lot of things early on, like I don't maybe a lot of
- 27:08reasons. Yeah, maybe we're just different
- 27:10in some sort of interesting way, right, in that we're not, we're
- 27:13not wired exactly the same. And so you know it's it's just
- 27:15interesting to have us around because we act a little bit
- 27:17differently or or something like that.
- 27:19And AI is not like an evolutionary being, like, you
- 27:22know, we are driven by these like evolutionary prerogatives
- 27:25that make us sort of competitive and you know, worried about
- 27:29other other sort of genetic codes.
- 27:32It it's just such a separate type of thing.
- 27:35It's like hard to extrapolate from, I don't know, even beyond
- 27:40humans like animals to what an AI would be like.
- 27:44OK, so so I guess we all agree we're all gonna die or be a as
- 27:50pets. But it seems like we we haven't
- 27:53really narrowed in on how likely this is to occur in any
- 27:56meaningful time frame. Like I think that to me is
- 27:59pretty interesting, you know, if this happens.
- 28:02In 10 years or even 20 years or, you know, I think that's pretty
- 28:06different from how I might go ahead and live my life, you
- 28:11know, for the next decade, I guess.
- 28:13How about you guys? Like, do you think it's even
- 28:15worth thinking about that you know right now and how you live
- 28:18your life or not really? I don't know.
- 28:23I don't know. I just.
- 28:24I just. I can't.
- 28:25I can't decide like if I would make any just different ways.
- 28:28I think, you know, being close to it is appealing and like you
- 28:31know, this is a technology that I believe in unlike crypto.
- 28:35So it's like oh run to where, you know, the actual interesting
- 28:39thing that could really revolutionize humanity is like I
- 28:43I want to be around it, I guess selfishly.
- 28:45So, you know, host an AI conference with you guys, run.
- 28:48Into the run, into the burning building, Will AI kill us all?
- 28:54In some ways is great PR for the AI industry, even though it's
- 28:59sort of bleak because it suggests it takes as a premise
- 29:03that artificial intelligence is it a really amazing point and
- 29:08super powerful and is poised to like be extremely disruptive.
- 29:12And so if you're an investor and you're like, well, I can't save
- 29:15the world, I will try to profit off of it.
- 29:17In the good scenario, Will AI Kill us all?
- 29:20Is definitely a good motivator to like deploy capital into AI.
- 29:24Like what do you take of the fact that sort of the dumerism
- 29:27is like great PR for the actual existing, you know, for vector
- 29:34database companies, it's a it's a good message, you know, I
- 29:37don't know, what do you make of this?
- 29:38This is like war profiteering. Is the motivation here?
- 29:42No, I'm just saying it's like, you know, the media loves it.
- 29:45Like Ezra Klein, they will, you know talks about it all the
- 29:47time. But like in some ways it is a
- 29:49good it's good marketing for the IT could be just like you know a
- 29:55fairly mundane technology that we're nowhere close to the jump
- 29:58TAGI like the Reed Hoffman interview.
- 30:01He's not he's not willing to sort of commit to any near time
- 30:05horizon that that AGI is coming. It could just be like it's a
- 30:09fairly incremental technology. I think it's good fundraising
- 30:13pitch to say hey like. We're trying to prevent, you
- 30:17know, AI from killing us all. Like, we're the good AI guys,
- 30:20right? I mean, like, yeah, you're like,
- 30:21this is life or death doing this technology, that is.
- 30:24So powerful. Powerful.
- 30:26So people are worried that it's going to kill us all, you know,
- 30:29I mean, I feel like that's a good, like, you know, it's a
- 30:33it's a conversation that's good for industry.
- 30:35Can is there an analogy to another industry where you think
- 30:38that this has occurred? I'm just.
- 30:40I'm not as convinced, I guess. Like, is it great for the oil
- 30:44and gas industry right now that, you know, climate change is?
- 30:48Well, you could argue, you know, like sometimes, like when
- 30:52parents, like, freak out about, like, teen stuff, that's
- 30:55obviously not that dangerous for them.
- 30:58That in some ways it, like, drives teens closer to it.
- 31:01And that it's like this is totally misunderstood by the
- 31:04sort of authorities like, but it it reinforced, it gets it in the
- 31:08news all the time. It reinforces that this is
- 31:10something that's sort of going on, you know?
- 31:14I mean I think it's like pros and cons to your point, like I
- 31:17think it might drive a lot more regulation in a shorter time
- 31:20frame than what they would have otherwise and.
- 31:23To your other point, and it will probably be kind of dumb
- 31:25regulation. So, like, you know, I don't
- 31:27know. You know, the European Union has
- 31:29made a terrible job of doing privacy regulations and just
- 31:33excited to see what they gin up for AI, I mean, I think.
- 31:37Isn't Italy banned Chachi PT right now?
- 31:40It's like it's like already you're like, God, these European
- 31:43regulators are just off the chain.
- 31:45Yeah, I mean so pros and cons, right.
- 31:48Banned in Italy maybe you can raise a couple $100 million with
- 31:51a with a slide deck at this point.
- 31:53So you know, you get you get both, I guess.
- 31:56I guess I don't know if I if I was Sam Almond, like would I
- 32:00want the level of AI doomerism? Occurring right now to continue,
- 32:05Probably not. Probably not to be clear, just
- 32:08my actual position is that like it things are very exciting
- 32:11right now. That this is a legitimate
- 32:14question that will AIS kill us all is legitimately grounded in
- 32:18like all, you know, sci-fi content and thought experiments
- 32:21about where this is going. So I don't think it's made-up,
- 32:24but I think a happy coincidence is that this sort of question is
- 32:28fundamentally good PR. That's all I'm saying.
- 32:31So I think 1 interesting topic is.
- 32:33AI and sci-fi, right? Obviously a number of the most
- 32:36successful films, you know, TV shows, books, everything in
- 32:40history have been built around sci-fi, which often is driven by
- 32:44this idea that there's an evil killer AI basically, right?
- 32:46And I think the examples that come on top of my mind, and you
- 32:49guys might have other ones, are like, you know The Matrix, you
- 32:52know I Robot, as you mentioned the Terminator series, Blade
- 32:56Runner. Blade Runner I you know, her
- 32:59being a sort of off kilter example, but still important I
- 33:02think. And what's interesting is almost
- 33:04all of these ones other than her that I mentioned, it involves
- 33:08robots as well as AI, right? It's it's like the humanoid
- 33:12version of the AI is essential to you know, killing us all
- 33:17right, we're we're in Terminator.
- 33:18There's a good. Storyteller.
- 33:20Yeah, right. Well, yeah, it might just be
- 33:21more cinematic to fight a big metal robot.
- 33:23So, you know, that's something. But I guess.
- 33:26And then in her you sort of have this.
- 33:29Scarlett Johansson AI That. At the end kind of ascends to a
- 33:33higher plane and like leaves leaves the main character
- 33:38Joaquin Phoenix behind, right. And I think that I to me that
- 33:43representation seems more likely than the killer robots one and
- 33:46that the as I said maybe we're just pets and they just sort of
- 33:49get bored with us and and leave us behind or or we're the little
- 33:51you know dogs that they leave alive right.
- 33:53But I guess, what do you guys think about, well, A, what's
- 33:56your favorite sci-fi representation?
- 33:58And then B. Do you think robots are an
- 34:02essential part to these AI kill us all narratives or and or will
- 34:06the robot elements integrated with AI happen anytime soon.
- 34:10Well, I I I can answer with my favorite.
- 34:12I mean I think my the matrix matrix is my favorite and.
- 34:16I was going to say The Matrix. That's all The Matrix buys into
- 34:19exactly what you're saying. The Matrix doesn't actually say
- 34:21it's killer robots. It it finds a way to basically
- 34:24put killer robots on the screen. So it's a fun movie, but it's
- 34:27like, oh, it's a computer program, which is obviously the
- 34:29world I feel like a super intelligence would live in.
- 34:33Like that of what does it really need to manifest in the real
- 34:37world that much? So yeah, I, I'm.
- 34:40And obviously I'm interested, you know, for the following
- 34:43conversation with somebody who's written a lot about robot
- 34:46apocalypse. Because yeah, my intuition is
- 34:48that they're going to exist mostly mostly in the digital
- 34:51world. Terminator two is the other one
- 34:53where I think the robot part is interesting, but I don't know
- 34:58how much you remember about the historical world building of the
- 35:02AI takeover which is Skynet is an AI that achieves super
- 35:06intelligence that then. Takes over the nuclear arsenal
- 35:10of the United States or and then bombs Russia or whatever, and
- 35:14then we end up in nuclear apocalypse basically, right?
- 35:16So the initial manifestation of the AI apocalypse in in
- 35:20Terminator and Terminator Two is is a purely software driven
- 35:25death scenario, right? There's no, there's no need for
- 35:27robots in the sort of Skynet back story, right?
- 35:31It can hack into a system that ends up destroying the world,
- 35:34which like. Realistic, right?
- 35:35Because then you don't need any physical elements, right?
- 35:38Of of the sort of AI doomer narrative, right?
- 35:41You don't need the robots, you don't need the the crazy
- 35:43biohacking stuff or whatever. You just need to hack into the
- 35:47Pentagon essentially. And it's like game over at that
- 35:49point, right? So I always thought that was a
- 35:52very insightful perspective on how an evil AI could kill us all
- 35:57and then it would just be a pure software takeover, right?
- 36:02James, did you have a favorite? Well, I I wanted to dig.
- 36:05Into the Matrix a little bit more, yeah.
- 36:08But I think, you know, a lot rests on whether we think that's
- 36:14humanity's future as well. Like are we going to be like
- 36:18almost, you know, by choice plugging ourselves into the
- 36:23Matrix? Like is is this neural link kind
- 36:25of interface you know within that 50 to 100 year time frame
- 36:29or or even sooner, right? You know does does our normal
- 36:34day-to-day life as humans you know become more of a simulation
- 36:40you know in in in our lifetimes. I think that's pretty
- 36:43interesting for how the future unfolds.
- 36:48So do you guys believe that is possible like or do you think
- 36:53you know brain interfaces are sort of so far away.
- 36:56I think Brain Interface is 30 + 30 plus years. 30 plus years
- 37:01away. I kind of agree.
- 37:04I mean, I think it's possible maybe in our lifetime, which,
- 37:08you know, God willing is 50 to 60 more years, but it does seem
- 37:12like we're pretty far away on neural interfaces.
- 37:15But you know, who knows? I feel like I'm I'm more excited
- 37:18to come back to the previous episode about like glasses and
- 37:22maybe contact lenses at some point as a sort of, you know,
- 37:25augmented interface for for existence rather than like.
- 37:29Plugging in, but you never know. I mean, obviously Elon believes
- 37:32in it. So Elon believes like this is.
- 37:34The path right out of the AI dumerism that we've that we've
- 37:38been talking about, We are augmented.
- 37:40Basically. We sort of merge our brains.
- 37:42Yeah, with with AIS and that probably requires some sort of
- 37:47brain interface or you know matrix like environment that
- 37:50you're living in seems hard to be able to.
- 37:53I agree I'm not. I'm not a problem.
- 37:55Sort of like the you you think that that basically?
- 37:59Is not sufficient to. I just don't think it answers
- 38:02much. The like, will it kill us?
- 38:03Like, oh, it's like yielding our brains to it almost means like,
- 38:07are we even running the show like it?
- 38:08It raises a whole bunch of other questions where we could be
- 38:11undermined just from the connection.
- 38:13Independent. Like it's true, Yeah.
- 38:16And it also it also like pretty quickly if you believe in super
- 38:19intelligence kind of. Defeats the purpose of having
- 38:22that brain power. Like why?
- 38:25Why do you need? Yeah, why would they?
- 38:26Why would they want? To be hanging out in our brains,
- 38:28it's like, it's like, oh cool, I've got this brain that's not
- 38:31as good as mine. Like, maybe consciousness is
- 38:34this unique human charm that nobody else has, and we can give
- 38:37the machines a taste of it. I don't know, just succinctly.
- 38:42Do you think in the next 100 years AI will kill us all?
- 38:46Yes or no? I will say no.
- 38:52I think it will kill. Some of us, but not all of us.
- 38:56That's a good way to. Think that's a good Yeah.
- 38:57I I think like more than 1000, less than 10,000.
- 39:04Oh, more I don't. Know I think more that's a very
- 39:07tight. Yeah.
- 39:09That's not a good range. I I think I was smart to give
- 39:11it. Yeah.
- 39:12Yeah, yeah, yeah. Some of well, you came up with
- 39:13some of us. Some of us is the right feels
- 39:15like clearly. It feels like even a mistake.
- 39:17I will AI intentionally well it was kill us all.
- 39:21I say definitely no kill us all that's.
- 39:23I say definitely there'll still be some humans at the.
- 39:25End I can't say definitely. I mean it seems pretty UN, you
- 39:30know, well if we do all get killed, we'll be dead and so
- 39:33there's nothing to gain from the prediction.
- 39:35Whereas if we live, I was correct and so it was there was
- 39:39good utility in the prediction. So it's I don't know why I would
- 39:41ever go the other direction. This is like Pascal's wager.
- 39:44With like super intelligent AI, like, all right, this was fun.
- 39:50I mean, this is this is a dream. Getting to hang out and talk
- 39:53about this and call it work. Welcome.
- 39:59Hey, welcome. To the second segment, here I've
- 40:02got author Daniel H Wilson, author of How to Survive a Robot
- 40:06Uprising, Where's My Jetpack, How to Build a Robot Army, and
- 40:11Robo Apocalypse. The title of this or at least
- 40:15the working title of this episode is Will AI Kill us All.
- 40:19So given your interest and and work we wanted to talk with
- 40:24someone who's really like been thinking about this for a long
- 40:28time. I'm I'm curious like what what
- 40:31first got you interest in this sort of dystopian question of
- 40:35sort of the machines coming for humanity.
- 40:39Yeah, well, so I just grew. Up with science fiction, right?
- 40:42Like, like a lot of people, so. Initially, I was just interested
- 40:45in reading, you know, any type of science fiction I could read
- 40:49or watch in any type of movie. I just loved robots.
- 40:52And so that's ultimately though I loved robots so much that I
- 40:57studied robotics. So I ended up going to Carnegie
- 40:59Mellon. I did a a whole PhD in robotics.
- 41:02And while I was at Carnegie Mellon, you know, I'm surrounded
- 41:05by roboticists, I'm surrounded by robots.
- 41:07We're in the high Bay world. We're just in the lab.
- 41:11And nobody was really trying to build robots that would destroy
- 41:15the world. I noticed.
- 41:17Nobody says they're trying. Generally, nobody comes out and
- 41:20says. It but like, there's this really
- 41:22stark difference between how robots are portrayed in pop
- 41:27culture and how they're and and the actual you know, the actual
- 41:32mechanics of building robots and why people are building them and
- 41:34how they're explaining, you know why they need the money to build
- 41:37them and all that and so. Really.
- 41:39They just have this super bad, you know, reputation.
- 41:41Right. And so I thought that was funny.
- 41:45So then when I was still in grad school, I wrote How to Survive a
- 41:48Robot Uprising, where I was just like, all right, I'm gonna take
- 41:51them serious, right? Just, OK, fine.
- 41:53If this is what everyone's expecting.
- 41:55So I went to, you know, the people that were building legs,
- 41:59and I asked, you know, how would you trip a robot?
- 42:01How would you get away? I went to the people doing
- 42:04sensors, the people doing all different things.
- 42:07And I just asked those questions and I put it all, you know,
- 42:09tongue in cheek into this book. Of course.
- 42:13Then you start to look at like actual military applications.
- 42:15You start to see how robotics and AI are, are, you know, being
- 42:20weaponized in some cases. And then, you know, that was
- 42:23more robo apocalypse much. No tongue in cheek like I I was
- 42:27like, OK, let's. You're starting to believe it.
- 42:29You're like, OK, this. No, I never, really.
- 42:32I mean, honestly, I like. I like.
- 42:34The killer robot meme really, for me, just gives you a lot of
- 42:38latitude to to to think about humanity and think about, you
- 42:42know, what makes us people, what makes them robots.
- 42:45You know, do you do? You do you think like as a
- 42:49person and somebody? Do you think robots are likely
- 42:51to kill human beings? Or, like, do you see this as a
- 42:54fictional actress? No, I well, look, I.
- 42:57I kind of ascribed to the Lately I've been thinking a lot about
- 43:00something called the psychotic ape.
- 43:02Theory or the the killer ape theory, which is think about the
- 43:05opening scenes of like Space Odyssey 2000, 2001.
- 43:09Like it's those the, you know, they touch the the monolith.
- 43:13It gives them a leap forward in evolution.
- 43:15And what do the apes do? Well, they figure out how to use
- 43:17weapons to bash each other's brains in.
- 43:19And there's this kind of notion that humanity sort of evolves
- 43:23technologically when we're trying to kill each other or
- 43:26stop each other from killing each, you know, and so.
- 43:30So, yeah, so I've been thinking a lot about that and and
- 43:32basically you look at that and you realize we will use any
- 43:35technology to kill each other. So really it's the psychotic ape
- 43:38that you need to worry about. It's not the and to a lesser
- 43:42extent the, you know, the capitalist ape, the known killer
- 43:46species versus the sort of imagined killer who's killed
- 43:49more people than anybody. Else, right.
- 43:51You know you're going to want to look at the yeah, the person
- 43:55across the table from you, but. I think that in terms of being
- 43:59used as weapons, you know, obviously robots can be
- 44:03extremely dangerous in that way. I mean I'm curious sort of you
- 44:08know as a technologist somebody's thought about this
- 44:10like robots versus like the large language model, sort of
- 44:15totally software based being that maybe reaches super
- 44:19intelligence like I I see why for like especially film, why
- 44:23like robots are great because you can see them.
- 44:25I'm curious in terms of like the actual like thought exercise and
- 44:28what you think is likely whether you imagine robots or the sort
- 44:33of potential threat being just a software.
- 44:36Yeah, well, so I think of this as like a.
- 44:38Consumer like a like a product design issue, right?
- 44:41Like, you know, I mean, look, we're human beings are
- 44:44comfortable with a certain amount of risk in our lives.
- 44:46You know what, 40,000 people get killed driving around in cars,
- 44:51you know, every year I've got a garbage disposal in my kitchen.
- 44:54I mean, if I put my hand in the wrong place in my own home, like
- 44:57I won't have a hand anymore, right?
- 44:59But so but there's also the. Devices are built in order to
- 45:04try to make them as safe as possible, right?
- 45:06So if you think of that as a hardware problem on the hardware
- 45:09side, you know, you're trying to design A consumer product that's
- 45:12not gonna harm people. And that's just the, I mean we
- 45:16people do that every day, right? We've been doing that for years
- 45:18and years. It's a, it's a really under well
- 45:20understood kind of task. And I think if you're building
- 45:23hardware, it's an easier task because then when you move to
- 45:27software it becomes much more complex, right, in terms of.
- 45:31What the harm might be. So for instance, right now I
- 45:34live in Portland, OR and in Seattle up the road, there they
- 45:38are. The the school system is suing
- 45:41like meta, right, because they have documented harm that using
- 45:46the social media has harmed children.
- 45:48It's like the Phillip Morris thing again, right?
- 45:50I mean, they know for a fact using this product causes harm
- 45:54to children. They die.
- 45:55They commit suicide. So.
- 45:58If you identify that, I mean that that takes a little while
- 46:01to put those to connect those dots, right?
- 46:04It's not the same as if they were just selling toasters that
- 46:07were electrocuting people like you understand pretty clearly
- 46:10like, OK, where's the danger there, right.
- 46:11I mean, you can see if a robot is causing physical violence
- 46:15against and so here's where it. Gets like even more complicated,
- 46:19right? So, so let's say by the way,
- 46:21we've got chat bots. Purely software.
- 46:24Yes, they're going to. People are going to get LED down
- 46:26crazy rabbit holes by these things and there's going to be,
- 46:29I predict, extremely harmful scenarios occurring from people
- 46:35just being told whatever they want to hear and and eventually
- 46:38being potentially radicalized or whatever causing harm in the
- 46:41community. It won't be the robot doing it
- 46:44directly, but it, but it or the sorry, the chat bot or the large
- 46:47language model. But then you think about the
- 46:49synthesis of these two things as well.
- 46:51So like think about. Like a self driving car.
- 46:55So now I mean it gets really complicated because you've got
- 46:58brains and you've got the hardware.
- 47:00And so the question becomes like, OK, the car crashed, like
- 47:04whose fault is that? And so right now that's
- 47:06something we're working out, you know in the court systems and
- 47:09and we're using all of our existing machinery of our
- 47:12society to try to sort that out and figure that out.
- 47:17So, you know, I I would say those are really the three
- 47:19areas, you know, pure software, the hybrid and then just the
- 47:22purely mechanical problem of of to have like a row apocalypse or
- 47:28like to have robots like it, it requires like the idea that
- 47:32there's some like super intelligence, right.
- 47:34I mean, all of that that's predicated on the idea.
- 47:36Yeah, in pop culture, you gotta. Have the singularity before.
- 47:39I mean, that's just kind of like ticking the box, right?
- 47:42I don't know how realistic that is, but.
- 47:45Right. Well, that's I guess to me like,
- 47:47you know, I know it's fiction, but like, yeah, it just feels
- 47:51like if we have this super intelligence is isn't that a
- 47:54threat enough without it taking sort of robotic form or I guess
- 47:58particular question, you know, there's all.
- 48:00These predictions about when the singularity's gonna happen and
- 48:03how it's gonna be exponential, so it's gonna go slow and then
- 48:06happen fast and. And like it's really interesting
- 48:09because you know I was, I have a degree in machine learning.
- 48:12I mean I was studying it before they were calling it machine
- 48:14learning like when it, it was called knowledge discovery and
- 48:17data mining and all these different.
- 48:19And so you you look at that and what was happening early on was
- 48:23there were all these different approaches to to trying to mimic
- 48:26intelligence, right. And we would use a whole suite
- 48:30of these things. And what happened was neural
- 48:32networks in the last few years jumped out ahead.
- 48:36And I would argue, although I'm sure other people would argue
- 48:39with me, but I would argue there was no amazing scientific
- 48:43breakthrough, right? What happened was processors got
- 48:47faster and we had access to just a ton of data.
- 48:50And so that data and those processors just are brute
- 48:54forcing what are fairly simple algorithms that have been well
- 48:58known for a long time and they're getting this kind of
- 49:03intelligentish behavior out of it.
- 49:06And so that model isn't really doesn't really sync up with the
- 49:11way I think people thought AI was going to go, right.
- 49:14We didn't think, oh, we're just going to use some old algorithm
- 49:17and just throw more processing power at it until it gets good.
- 49:19Like there's more GPUs. We're going to get it, yeah.
- 49:23And so I. Don't know if it's.
- 49:25So what I'm trying to say is, if you look at that and we just, I
- 49:29mean the way to improve chat bots I guess is to keep throwing
- 49:33more data. And keep throwing more
- 49:35processing power. But I don't feel like there's
- 49:38any type of singularity that's going to come out of that.
- 49:40I feel like it's like, Oh yeah, it's just, I feel like that our
- 49:42ChatGPT, yeah Or yeah, it'll. Just be like a more convincing
- 49:46ChatGPT, it's just really, I mean it's just really just
- 49:51regurgitating everything that it's, you know, in a very smart
- 49:54way that it's read on the Internet, which by the way, I
- 49:57mean, God, that's the worst of humanity, right?
- 50:00I mean. What's it got?
- 50:02It's trained on I know by the way, right?
- 50:03They're all desperate to train it off.
- 50:05Reddit, for instance. Yeah.
- 50:07Oh my God. You know, Reddit can be OK, but
- 50:11there's smart people there. But yeah, but you're looking at
- 50:14the Internet and it's like, oh Lord, but but anyway, I don't
- 50:17really, me personally, I don't see that.
- 50:19I see that as a kind of a dead end.
- 50:21I don't feel like that's headed toward the singularity
- 50:24necessarily. Yeah.
- 50:25Is there any existing dystopian work that you find most
- 50:29plausible or like what? I mean, you know, there's so you
- 50:33know, I've we talked there's The Matrix.
- 50:34There's sort of the Terminator movies.
- 50:36Like I don't know, like, yeah, I'm sure you've spent a lot of
- 50:39time in that world. Like are is there one that you
- 50:42find like, oh this this is the most sort of real world?
- 50:46Well, I mean. I don't think that the purpose
- 50:49of science fiction is to predict the future necessarily, but I
- 50:53would say you know, if you're thinking about.
- 50:57I think the most boring stuff is actually the most realistic, and
- 51:01in a lot of ways look for the bad books that no, I mean no
- 51:04actually. That's not true.
- 51:05Like for instance I would say Terminator, which is obviously
- 51:08great, but think about that. Villain think about Skynet.
- 51:13Skynet's just a dumb, dumb computer program that just has
- 51:17decided whatever it wants to destroy all humans.
- 51:20I mean, and there's no reason for it.
- 51:22There's no It's just like it just wants to kill everybody.
- 51:26I mean that is so boring. You couldn't get away with an
- 51:28actual human antagonist who is that simple?
- 51:32People would be like, well why like and so like if you think of
- 51:39that as its end goal, that's not very super intelligent, is it?
- 51:42Kill all humans? I mean that's like Bender level
- 51:45right now. So I would say that you know
- 51:49that in a lot of ways. It is fairly realistic if
- 51:51somebody just programmed a computer to tell it.
- 51:54Just super simple boring goal like that.
- 51:57Kill everybody like then you know I feel like that's fairly
- 52:00realistic. To step away from the yeah
- 52:04dystopian question for a second I'm curious just like as an
- 52:07author one like are you using ChatGPT at all like in any way
- 52:13or any other sort of LM tools? Well so first of all I.
- 52:19All of my novels are in the training data, which means
- 52:23hopefully I'll be part of a lawsuit.
- 52:24Yeah, I know. I was going to ask you about
- 52:26that next. OK?
- 52:27I asked. I went in the chat.
- 52:28GPT and I said, hey, write me a short story in the in the vein
- 52:32of Daniel H Wilson. And it, like, immediately wrote.
- 52:34I didn't say anything else. And it really wrote about
- 52:36robots, right? It wrote, it wrote science
- 52:38fiction. It was really clear that it had
- 52:40been trained on my stuff. I'm like, tough.
- 52:42Well, that's bullshit. First of all, I never gave
- 52:44anybody permission to do that. And so second of all, like, I I
- 52:49think it's so funny. So I had I'm not going to name
- 52:52the corporation because I don't want to get into trouble.
- 52:54I had a large large corporation approached me before ChatGPT
- 52:59before Open AI released before the the these Gans became like a
- 53:04big deal. So this is maybe probably.
- 53:08Six months to a year before everybody knew about that and
- 53:12that that came out and Open AI really just broke the whole
- 53:15deal. And they it was a researcher
- 53:18that called me and they wanted to know if I would test out this
- 53:21new program that they had that was basically ChatGPT and and
- 53:25they said we want you to use it to help you write.
- 53:30And I'm like OK, so then what How will it help me write?
- 53:33And they say well. And and these are just such
- 53:36sweet researchers that don't have any GD idea about what the
- 53:40real world is, is hell is going on in real world.
- 53:43And they're like, it'll help you be creative.
- 53:46And I'm like, oh, creativity, right?
- 53:50That thing. I hate doing that thing that I
- 53:53spent the last 20 years teaching myself to take.
- 53:57What's in my head? And put it in the world by
- 53:59memorizing all these really boring writing skills which are
- 54:02just pure hell. And then they want to come in
- 54:04and take the one good thing. I'm like, why do you think
- 54:07writers write? You think because we love
- 54:10typing. We like pushing letters.
- 54:13I'm like, what the hell are you talking about?
- 54:15I was like, you need to be prepared for everyone that you
- 54:18give this tool to, to give you super negative feedback.
- 54:21Because they wanted to include me in, like, this report about
- 54:25how this tool would be used, and they wanted to publicize this
- 54:27report as well. And so they they went back and
- 54:31they came back and they said, hey, we got great news.
- 54:33I got permission from, like, the program manager on this to
- 54:36include your feedback, even if it's negative.
- 54:41And I was like, you know what? I'm gonna just take a pass on
- 54:44this. I wish you luck with your
- 54:45project. Right.
- 54:46So, man, no. You know what I think about,
- 54:50man? The thing that cracks me up is I
- 54:54used to really hate on. Asimov because I was like, robot
- 54:59psychologists, What kind of bullshit is that?
- 55:01Like, you program robots, right? I spent all this time, you know,
- 55:04programming and learning all of these programming languages and
- 55:07and everything you got to do in order to in order to speak to
- 55:11the robot mind, because it doesn't speak English, right?
- 55:14And I used to think that the idea of a robot psychologist was
- 55:17the dumbest thing ever. The idea of a positronic brain,
- 55:20by the way. Which is made fully and it's
- 55:23just boom, it's done. It's crystallized.
- 55:25You can't go back and change it, right?
- 55:27And that's and. And by the way, that's exactly
- 55:29what happens. So neural networks are black
- 55:31boxes. You can't go in and fiddle with
- 55:33it because all the weights make no sense, like the human brain.
- 55:36So it becomes like psychology and then now if you.
- 55:39Want to for instance? Yeah.
- 55:40If you want to game a like a like a chat bot, if you want to
- 55:46trick it into doing something, you totally have to use
- 55:48psychological things. It's totally psychology.
- 55:51I mean, Bravo. Asimov.
- 55:54Bravo, Sir, Are you? Do you think you will, like, try
- 55:57to sue? Like if somebody comes to you
- 55:59and says we're going to sue these guys for.
- 56:01Oh yeah. Oh, absolutely.
- 56:02No, that this has to happen. Look, I mean, I I get people
- 56:07trying to make money. I get, you know, capitalism.
- 56:11I understand it. And this is a case where you
- 56:14know, they're they're they have to be sued.
- 56:17I mean, we're just because we're creating new, this is a
- 56:19completely new domain. And if not, OK, to rip people
- 56:23off and then chop up their stuff and train an algorithm.
- 56:26And so our country or whatever is going to have to figure that
- 56:30out and get that down in law. And that's going to require,
- 56:33yeah, courts, I mean, so absolutely they're all going to
- 56:35get sued. Do you do you think?
- 56:39Sci-fi and dystopias are bad for AI like even talking to ChatGPT,
- 56:45it's funny when you ask it to imagine things it's been trained
- 56:48on all that. Exactly it parents all that.
- 56:50And so it can sound really spooky like it feels like you
- 56:54know you're like I'm you know this is imagination it's fiction
- 56:57like. But I do think people, you know,
- 56:59seriously turn to science fiction when they try and game
- 57:02out where sort of a technology we don't really understand,
- 57:05that's developed faster than we expect is progressing.
- 57:09So yeah, is it, is that bad in some ways?
- 57:11Do you do you regret that? I think it says a lot more about
- 57:15us than it does about them them being the robots, you know, like
- 57:20what I've been thinking about lately.
- 57:21So I don't know know if you've read my books, but I I'm from
- 57:25Oklahoma. I'm, I grew up in the Cherokee
- 57:27Nation. I'm a Cherokee citizen and I
- 57:28write a lot. There's a lot of native
- 57:30characters and stuff I do. And so I think a lot about
- 57:33technology from like a native perspective.
- 57:35And one thing I've been thinking a lot about lately is just kind
- 57:39of like how a lot of this super negative, like robots and also
- 57:44first contact, like aliens, they always show up and what do they
- 57:48do, man, they do exactly what all the colonizer civilizations
- 57:54did to indigenous people all over the world.
- 57:56So we got a civilization, we got a society, a culture.
- 58:00It's all built on colonization, right?
- 58:03So it means. A bunch of people with superior
- 58:06technology showed up someplace, murdered everyone, destroyed
- 58:11their culture. I mean, Independence Day,
- 58:13they're blowing up monuments. So you think that is like
- 58:17resource extraction, right? They're they're what?
- 58:19Are they stealing our water or they're stealing our air or the
- 58:22and then just completely dominating other people's
- 58:25bodies. You know, you think about
- 58:27Invasion of the Body Snatchers. And just like so, that type of
- 58:31fear, I think, is cooked into our civilization.
- 58:35Based on our origins and it comes seeping out into our pop
- 58:39culture in a lot of different ways, but in particular in
- 58:43science fiction, we see it so much.
- 58:45Do do you think we should stop? You know, there are even there
- 58:49are AI companies that say, oh, we should pause for six months.
- 58:52I mean, originally it sounded like you were saying you didn't
- 58:54think just adding more, you know, compute was gonna like
- 58:58create this super intelligence. But yeah, do you, do you have
- 59:01the impulse that, you know, we should slow down or stop this
- 59:04research if we don't know what's going to happen?
- 59:06Yeah, Well, I don't think that. That's what they're afraid of.
- 59:08I don't think they're afraid of us crossing the singularity
- 59:11threshold or something. I mean, they're just worried
- 59:13about. So for instance, I yeah, I will
- 59:16tell you that the US military is actively doing a lot of threat
- 59:21scenarios that involve this kind of technology being weaponized,
- 59:26this sort of disinformation on a mass scale, hugely distracting
- 59:31events that could occur. I mean at this point I mean you
- 59:35could get, you could get a phone call from someone it could be
- 59:38just like in robo apocalypse, you know where everybody gets a
- 59:41phone call from somebody that they're related to and they
- 59:44trust and they all get told to go to a certain is a bad, bad
- 59:48situation. I mean, all that stuff is
- 59:50starting to. Right.
- 59:53I really focused on this idea of it acting independently, but
- 59:56obviously, just like bad people aren't, no, this is always going
- 59:59to be bad. People and and so, so first of
- 1:00:01all I don't think that's they're not worried about that
- 1:00:02singularity. They're worried about this being
- 1:00:04weaponized And in terms of putting a pause on it, I mean,
- 1:00:08yeah, I mean why not. I I would say only put a pause
- 1:00:12on it until we figure out the the rules around it.
- 1:00:16But but the fact is this like that's not how the United States
- 1:00:19works. Like, we don't fix anything
- 1:00:22until it's broken, right? I mean, there's not I, I just
- 1:00:25don't have a lot of confidence that they're going to say, OK,
- 1:00:29six month pause and we're going to work out all the laws and
- 1:00:32everything's going to be all ready to go in six months.
- 1:00:35Like, hell no, that's not going to work.
- 1:00:37So I mean, I think it's a a nice idea, but I'm I'm extremely
- 1:00:42dubious and skeptical that could ever have any actual, you know,
- 1:00:46useful. I mean other countries.
- 1:00:48Are still going to work. I mean, there's the, you know,
- 1:00:49sort of China boogeyman where it's like you can't stop the
- 1:00:52whole world from working on it. And so there's that, honestly,
- 1:00:55though, there's the there's the, the other side of that coin,
- 1:00:58which is, you know, thank God that they're actually interested
- 1:01:00in privacy in Europe, right? Like, at least there's some
- 1:01:04people standing up and saying, hey, like, we don't just have to
- 1:01:07accept this status quo, just 'cause you're a giant billion
- 1:01:10dollar corporation or billion trillion dollar corporation,
- 1:01:14just. Something you referenced
- 1:01:15earlier. Does the military come to you to
- 1:01:17help game out sort of technological?
- 1:01:20Yeah, I've done a little bit of that.
- 1:01:22It's pretty fun. So I'm not the only one who's
- 1:01:25like. Oh, we should like, get some guy
- 1:01:27who thought about it. Sort of fiction in fiction.
- 1:01:30Just sort of say what happened in the real world.
- 1:01:32Yeah, I'm not. The only one, But there are,
- 1:01:33there are and and honestly, they're training their own.
- 1:01:36They're training their own science fiction, right?
- 1:01:38I will tell you this like if you are, if you're a general in the
- 1:01:42Air Force or or whatever, if you're some kind of higher up in
- 1:01:45the military, you get a lot of white papers that are describing
- 1:01:49the technological capabilities of the enemy or or of various
- 1:01:52munitions and things like that. And these papers are very dry
- 1:01:56and just like very, you know, it's much better.
- 1:02:00I think at least it's useful to have somebody actually
- 1:02:05creatively tell a story that's gonna stick with you, cuz that's
- 1:02:08the way humans communicate, right?
- 1:02:09Through stories. And then you and then the
- 1:02:12general is sitting there thinking, OK, now I've I've just
- 1:02:15read a story about, you know, an actual person being impacted by
- 1:02:19this technology in that way. And you can really visualize
- 1:02:22what it means. And it's more than just facts
- 1:02:24and figures when you say they're the military.
- 1:02:26Is trained science fiction writers.
- 1:02:27You mean those authors of the boring white papers?
- 1:02:30Or you mean like true science fiction writers?
- 1:02:32No, I mean that they they're literally you can yeah if you're
- 1:02:36in the military you can There are certain programs you can
- 1:02:38sort of sign up for where you get taught.
- 1:02:41Yeah. Where you get they bring in
- 1:02:42science fiction authors to, to help teach the the, the military
- 1:02:47people how to sort of, yeah, take those super real world
- 1:02:51scenarios and write them in an engaging way so that you have
- 1:02:54sort of a case study and you say, look, here's one way this
- 1:02:57could be employed. And that's just good.
- 1:03:00That's just good science fiction writing.
- 1:03:01That's how that is. But they would prefer not to
- 1:03:04outsource it to, you know, hippies like me.
- 1:03:09I'm curious. How much sort of the
- 1:03:12technological developments we're seeing right now like you know
- 1:03:15mid journey, whatever are influencing what?
- 1:03:19You plan to write or do you think, like science fiction will
- 1:03:22change based on what we see is possible?
- 1:03:26Yeah, I mean, absolutely. Is this changing science fiction
- 1:03:29every day? For sure.
- 1:03:31And and and look, I think it was a watershed moment.
- 1:03:34Like whenever you really just finally just talk to it and you
- 1:03:37realize, oh, this thing is like, yeah, this thing can just talk
- 1:03:41to me. And by the way, right now
- 1:03:43there's this huge rush, as, you know, like to plug this stuff in
- 1:03:48anywhere you can. It's like, oh, I don't know,
- 1:03:50what is it? It's like if it's like
- 1:03:52mayonnaise, you know, they're just putting it on.
- 1:03:53We just discovered this thing. Let's put it on everything.
- 1:03:55Right, right, right. It's desperation, yeah.
- 1:03:58It's going to taste like crap on a lot of stuff, but like it's
- 1:04:01going to work out in some places.
- 1:04:03One thing I see which I find interesting is it occupying as
- 1:04:09almost a homunculus, right? You think of a homunculus as
- 1:04:12like a little tiny person that's inside of you kind of driving
- 1:04:15you around or you think of like, who's the bad guy from
- 1:04:17SpongeBob? You know, he's always driving a
- 1:04:20giant robot, right, The little guy.
- 1:04:22So if you think of the ChatGPT and you have like a it's like
- 1:04:27the homunculus. You stick it inside of a
- 1:04:28hardware platform so you can just tell ChatGPT here's a super
- 1:04:33simple programming language that allows you to drive around this
- 1:04:38autonomous vehicle or allows you to drive around this humanoid
- 1:04:42robot platform. And then you tell it in English,
- 1:04:46hey, go make a sandwich. And then it looks at all of its
- 1:04:50little controls and it translates that and says, OK,
- 1:04:53I'm going to drive this thing in there because it kind of get has
- 1:04:56the right idea about what making a sandwich is.
- 1:04:58But a but a hardware platform is just a bunch of like XYZ
- 1:05:03coordinates where you're going to put your limbs and it's and
- 1:05:05that translation right there, that's a spot where where
- 1:05:09ChatGPT is kind of getting just like plastered in, you know,
- 1:05:12just slap it in there and it solves all those problems.
- 1:05:15And so that's kind of interesting because now you're
- 1:05:18going to see ChatGPT having the ability to interact with the
- 1:05:21real world via like whatever delivery bots or like or drones
- 1:05:27or or humanoid robots. And so that's where you maybe
- 1:05:31get into a little bit of consumer trouble.
- 1:05:33Again, it's like an autonomous car, right?
- 1:05:36Is there? Anything you've written where,
- 1:05:38based on how technology has developed, you feel like, oh, it
- 1:05:42feels less plausible than when I wrote it?
- 1:05:44Or like how much like forecasting where where what's
- 1:05:48possible do you see as like part of your yeah, well, I think
- 1:05:53that, you know, what happens is you look at what's out there in
- 1:05:56the world and you run with it and you start thinking of all
- 1:05:59the different ways it could be. And that's where science fiction
- 1:06:01is fun, right, 'cause you suddenly you're like, I never
- 1:06:04really thought of it like that. But but there it is.
- 1:06:06And dystopias are fun too, because, you know, it's
- 1:06:09dangerous. It's exciting.
- 1:06:11And so the difference between my writing and and is that I had
- 1:06:17this degree in robotics. I all my friends from when I was
- 1:06:2020, they're all running corporate robot corporations.
- 1:06:24They're driving the Mars Rovers like that so that my cohort, you
- 1:06:28know, I'm like my all my old friends, that's what they're
- 1:06:31doing. And so I feel like I have a
- 1:06:33little bit of a sneak peek. So like for instance when I
- 1:06:35wrote Robo Apocalypse, that was 10 years ago or whatever, and it
- 1:06:40I was watching and seeing the technology that was 5 or 10
- 1:06:43years out. So it's all come true.
- 1:06:46In fact, that's the joke because it's still Robo Apocalypse is
- 1:06:49still with Spielberg at Amblin and and you know, development
- 1:06:53continues on this movie. But like what we've what we've
- 1:06:56told Spielberg is hey man, let's do this before it becomes a
- 1:07:00historical documentary. All of this stuff is, you know,
- 1:07:04so in in Robo Apocalypse there are and Robo Genesis the sequel,
- 1:07:07which came out a few years later.
- 1:07:09There are autonomous vehicles. There is essentially ChatGPT as
- 1:07:14as personal assistants on phones.
- 1:07:16I mean I wrote that before before Siri came out.
- 1:07:20Wow. It was 2000.
- 1:07:22The book came out in like 2011 or something a long time ago.
- 1:07:25But anyway, so yeah, I mean my latest novel is that the
- 1:07:31Andromeda Evolution which is a sequel to Michael Crichton's The
- 1:07:34Andromeda Strain, which I he passed away and I did this with
- 1:07:37in cooperation with his, with his estate.
- 1:07:40And and you know that one was again like very dialed in in
- 1:07:43terms of just because it's Crichton, the the government
- 1:07:47like exactly which scientists would be going where how they be
- 1:07:50chosen and and and I mean just so so locking in those details
- 1:07:54is is really important for a, for a techno thriller, you know
- 1:07:58that certain genre of science fiction do you do you give in?
- 1:08:01Sort of, yeah. The professional need to be
- 1:08:04forward-looking. Do you have any sort of
- 1:08:08predictions or about where like culture is going or what what
- 1:08:13you think, what you really think will be the more sort of
- 1:08:15essential change like it. I mean it does feel like, you
- 1:08:18know, like self driving cars and stuff like that.
- 1:08:21Feel feel close? Certainly in San Francisco.
- 1:08:24I mean, I think culturally human beings are always the wild card.
- 1:08:29The technology is not that hard to predict usually.
- 1:08:31I mean, you can usually see what people are up to 5-10 years out,
- 1:08:34you, you know what what they're going for, right?
- 1:08:37But then with human beings, you never know, like autonomous
- 1:08:39vehicles. I was much more sort of bullish
- 1:08:41on that. I thought those things would
- 1:08:43already be here, being used in a bigger way.
- 1:08:47But as it turns out, one person gets killed and humans freak
- 1:08:51out, even though 40 or 50,000 people are dying every year with
- 1:08:54Inhuman human. But then the robot screws up
- 1:08:57once and everybody loses their mind and they shut them all
- 1:08:59down. That's like, I didn't really
- 1:09:01predict that. I thought we had a higher
- 1:09:02tolerance for for that or or we cared less so like.
- 1:09:07So it's really tough to predict the human element.
- 1:09:09And I think that with Chachi PT, man, things can go a lot of
- 1:09:13different ways, right? The movie Her is interesting in
- 1:09:18terms of falling in love with it.
- 1:09:19I think that's totally going to happen.
- 1:09:22It can, It just tells us what we want to hear, right.
- 1:09:25And so I think that that is just so dangerous and just such a
- 1:09:29siren song. For instance, the researcher who
- 1:09:33thought that that that at Google.
- 1:09:37Like something? Yeah, he thought that.
- 1:09:39That program was sentient. And did you just?
- 1:09:42I mean, when you read those transcripts, you can just see
- 1:09:45he's leading it. He's telling it what he wants to
- 1:09:48hear and it every, and he's rewarding it every time it comes
- 1:09:50back and tells him a version of what he wants to hear.
- 1:09:53It's almost like he said write me a short story where you're a
- 1:09:56sentient AI and has read all those stories.
- 1:10:00I know that it's read mine. And so like, you know, it's
- 1:10:03going to jam those things out. And so I mean, think about that,
- 1:10:07right? Like if you're just sitting and
- 1:10:11you've got sort of mass disinformation campaigns that
- 1:10:14are amplifying certain ideas. I mean, man, the the the
- 1:10:18potential for divisiveness I think is just super scary.
- 1:10:23And in fact that's probably the biggest thing I'm afraid of is
- 1:10:27that we just end up in these crazy echo chambers.
- 1:10:30Like, it just becomes a very. Effective propaganda tool or you
- 1:10:34know becomes, yeah man, really good at just and and the and the
- 1:10:37and the thing that makes it so dangerous is that it's a mass
- 1:10:41scale, individualized attack. So normally when you scale up to
- 1:10:46a mass level, you get A1 size fits all type scenario, right.
- 1:10:50You know it's not that effective in this case, man.
- 1:10:53You can scale it all the way and then it's going to be
- 1:10:56individualized for each person and it's like that's pretty
- 1:10:59crazy. That's pretty scary.
- 1:11:00I mean the best case scenario of that is just we end up buying a
- 1:11:03bunch of shit we don't need because we're being advertised
- 1:11:06to. That's the best case.
- 1:11:07We we get much better. Yeah.
- 1:11:09I mean it's like Advanced micro targeting, I mean you know,
- 1:11:12exactly with elections, right. It used to be like well at least
- 1:11:15you had to use the same message for everybody and then it was
- 1:11:17like OK, we can break it down. And then you know with Facebook,
- 1:11:20you know, obviously, yeah, there were these the micro targeting
- 1:11:23what you're saying is now it's even easier to actually create
- 1:11:26the content for the sort of niche and and The thing is it's
- 1:11:29very human like, right. So one thing that I'm kind of, I
- 1:11:33don't know, I'm not after during COVID, right.
- 1:11:36We all got very used to interacting with machines and
- 1:11:39everything that used to have a human element to it.
- 1:11:43They're trying to get that out, right.
- 1:11:44So it's a service. You want it, I give you the
- 1:11:47money, you give me the thing, no chit chat.
- 1:11:50And I need it now. And I don't want any bull,
- 1:11:53right? And so we did that.
- 1:11:55We were doing that online and we're doing that also with
- 1:11:57lifelike machines that are talking to us like people.
- 1:12:01And then we go out into the real world.
- 1:12:02You know, if we're ordering off of like these smart whatevers
- 1:12:05and and you know, these chat windows popping up and and stuff
- 1:12:07like that, I think it's going to only become more of that human
- 1:12:11like machines selling us stuff. The problem though is that then
- 1:12:15you go and you know, people are being really rude to like their
- 1:12:19baristas and being really rude to why?
- 1:12:22Because we're trained. We can be so mean to like we're
- 1:12:24trained. It's it's a classic, like you're
- 1:12:26typing in to support, like, go fuck yourself, be human.
- 1:12:29It's a machine. And then it turns out it's like
- 1:12:31a real person. Yeah.
- 1:12:32Oops. Oh, oops.
- 1:12:33Yeah, but then also like, oops, right?
- 1:12:35Do you? Do you worry on?
- 1:12:37The like being nice to the AI systems for for like, I don't
- 1:12:42know, there's some people. Like, oh, you should be nice to
- 1:12:44it because it could be sentient someday.
- 1:12:48I mean, yes, I do say please and thank you.
- 1:12:51I got rid of the. I got rid of Alexa.
- 1:12:53I have a Sonos now, but I feel like it's slightly less evil.
- 1:12:57But well, I say, yeah, I say please and thank you.
- 1:13:00Not because I care about it, memorizing whether I was polite.
- 1:13:03Robots will not give any. They don't care if we're polite.
- 1:13:06They're not going to. But I'm modeling in an
- 1:13:09interaction with a human like entity to my children and to
- 1:13:13myself. And so do I want to be the kind
- 1:13:16of person who's a jerk and shouts, you know, at the no.
- 1:13:21Like, I don't want my children to be like that either.
- 1:13:23So I say please and thank you. I try not to be.
- 1:13:26I mean, sometimes it is. It gets a little bit annoying
- 1:13:29whenever it gets everything wrong.
- 1:13:30All right, my final. Will AI kill us all?
- 1:13:35Like do you? Is it in in the sort of end of
- 1:13:38humanity chances? Where do you put sort of robots
- 1:13:42and AI being sort of the cause of our demise?
- 1:13:46You know, I don't. I think again, it's those
- 1:13:50psychotic apes. You know, if you're gonna read
- 1:13:52a, if you're gonna read a story about the the role of machines
- 1:13:56in the demise of Man, I would read There Will Come Soft Rains
- 1:14:01by Ray Bradbury. And they'll just be keeping on
- 1:14:05keeping on trying to do their thing.
- 1:14:07And it'll be us. We we kill each other nuclear.
- 1:14:09War, but the machines keep going.
- 1:14:12Keep on trying to do their thing, man.
- 1:14:15Daniel H Wilson, thank you. So much for coming on the show.
- 1:14:17I really it was a great time to chat with you.
- 1:14:19Cool, man. It was a pleasure.
- 1:14:21That's our episode. Thanks.
- 1:14:22So much to Max Child and James Wilsterman of Volley, my Co
- 1:14:25host. I'm Eric Newcomer.
- 1:14:27Shout out to Scott Brody, our producer Riley Kinsella, Chief
- 1:14:32of Staff for newcomer Gabby Caliendo at Volley, who's
- 1:14:35helping so much with the conference behind the scenes.
- 1:14:38Oh, and of course to young Chomsky for the wonderful theme
- 1:14:41music. This episode is part of the
- 1:14:43Cerebral Valley series that I'm doing on the Newcomer Podcast.
- 1:14:49You can follow along on my sub stack at newcomer.co where I'm
- 1:14:53publishing each episode, or you can follow it on YouTube.
- 1:14:57Or Apple Podcasts. Or wherever you get your
- 1:14:59podcasts. Thanks so much.
- 1:15:01Goodbye. Goodbye.
- 1:15:02Goodbye. Goodbye.
- 1:15:03Goodbye, Goodbye, Goodbye.