WEBVTT - How did intelligence arise? 0:00:06.680 --> 0:00:10.399 Why are some animals so much smarter than others? And 0:00:10.440 --> 0:00:15.080 in particular, what is it that makes humans so incredibly smart? Now? 0:00:15.160 --> 0:00:17.720 I love animals as much as anyone else, and yes, 0:00:17.960 --> 0:00:21.640 other animals do do things like use tools and communicate 0:00:21.680 --> 0:00:24.680 with each other, But humans are the only species we 0:00:24.760 --> 0:00:29.000 know of that uncomplicated research programs to try to understand 0:00:29.040 --> 0:00:32.879 at a deep level why some animal species are more 0:00:32.920 --> 0:00:36.320 intelligent than others and how would these differences come about. 0:00:37.040 --> 0:00:40.519 We humans ask questions like why is it that crows 0:00:40.600 --> 0:00:43.720 have learned how to fashion sticks into tools that they 0:00:43.720 --> 0:00:46.280 can use to grab tasty bugs out of the nooks 0:00:46.320 --> 0:00:49.280 and crannies of trees, But I can't get my dumb 0:00:49.360 --> 0:00:51.199 chickens to stop eating the paints on the side of 0:00:51.240 --> 0:00:53.720 the chicken coop? And what kind of conditions would you 0:00:53.720 --> 0:00:56.040 need to have on another planet in order to produce 0:00:56.160 --> 0:00:59.880 highly intelligent aliens? These kinds of questions keep me up 0:01:00.520 --> 0:01:02.880 and also seem to be keeping many of you up 0:01:02.920 --> 0:01:05.840 at night as well. Today, we're going to tackle three 0:01:06.200 --> 0:01:11.039 wonderful listener questions that we've received related to intelligence. Welcome 0:01:11.160 --> 0:01:13.760 to Daniel and Kelly's Extraordinary Universe. 0:01:28.240 --> 0:01:31.080 I'm Daniel, I'm a particle physicist, and I hope that 0:01:31.160 --> 0:01:34.400 humans are not the smartest folks in the galaxy. 0:01:34.920 --> 0:01:38.640 Hello, I'm Kelly Wiener Smith. I study parasites and space, 0:01:38.760 --> 0:01:40.600 and yeah, I guess I hope we're not the smartest. 0:01:40.640 --> 0:01:44.080 As long as they're nice and smart, that would be great. 0:01:45.720 --> 0:01:47.840 I don't even really need to put that caveat on it, 0:01:47.880 --> 0:01:50.000 Like if they're mean and smart, like, I'll take a 0:01:50.040 --> 0:01:52.400 little bit of meanness for some secrets of the universe. 0:01:52.520 --> 0:01:56.040 What if they think we're delicious and they're smart, I 0:01:56.080 --> 0:01:57.360 think that's a bad combo. 0:01:57.440 --> 0:02:01.280 I think I don't think I'm certainly very tasty, so 0:02:01.360 --> 0:02:02.880 I'm not worried that I'd be going to be like 0:02:03.240 --> 0:02:05.400 high on the list of humans to be consumed. 0:02:05.520 --> 0:02:07.840 What would make you less tasty than other humans? 0:02:10.440 --> 0:02:11.560 This is getting personal now? 0:02:11.760 --> 0:02:14.440 Is us this sprinkled with stinky sauce or something. 0:02:16.440 --> 0:02:19.160 I actually kind of am marbled through with the ribbons 0:02:19.160 --> 0:02:21.239 of fats, So maybe I would be quite tasty. I 0:02:21.280 --> 0:02:22.760 don't know. I try to stay fit. 0:02:22.880 --> 0:02:25.520 You know, Marbling is a delicacy. 0:02:25.840 --> 0:02:27.560 I think we just set a record for how quickly 0:02:27.639 --> 0:02:30.760 we started talking about aliens and consumption of human flesh. 0:02:32.200 --> 0:02:33.920 Glad to see if by the end of the episode 0:02:33.960 --> 0:02:36.320 we hit poop and then we'll have the trifecta. But 0:02:36.560 --> 0:02:37.919 we'll probably get there, all right. 0:02:37.960 --> 0:02:41.240 So my question for you today, Kelly, is do you 0:02:41.480 --> 0:02:46.000 think that we will ever crack the barrier to communicating 0:02:46.520 --> 0:02:52.560 with our fellow intelligent earthlings, Like somehow as AI or 0:02:52.680 --> 0:02:56.960 future linguistics can help us understand the chirpings of prairie 0:02:57.000 --> 0:02:59.120 dogs and the whistling of whales. 0:02:59.520 --> 0:03:03.360 Wow, all right, So I've recently read this really great 0:03:03.400 --> 0:03:05.880 book called Do Aliens Speak Physics? 0:03:06.560 --> 0:03:06.680 Ed? 0:03:10.400 --> 0:03:12.200 That is a great book. 0:03:12.200 --> 0:03:15.160 It's a fantastic book, and it kind of convinced me 0:03:15.160 --> 0:03:18.320 that unless you can have two way communication, it's really 0:03:18.360 --> 0:03:21.799 hard to get on the same page regarding language. You can, 0:03:21.919 --> 0:03:24.839 like work with dolphins to try to figure stuff out. 0:03:24.919 --> 0:03:28.280 Maybe we could get some very preliminary like you know, 0:03:28.400 --> 0:03:30.120 I say hello to you, you say hello to me. 0:03:30.160 --> 0:03:31.959 We're on the same page there, But like I guess 0:03:31.960 --> 0:03:36.080 I'm not super optimistic. What kind of communication do you need? Like, so, 0:03:36.160 --> 0:03:40.440 my dog is really cute but about as smart as 0:03:40.480 --> 0:03:43.480 a ton of bricks, and so I can't communicate too 0:03:43.480 --> 0:03:45.760 many complicated ideas to her. But you know, there are 0:03:45.760 --> 0:03:48.400 some border colligues that are brilliant. So how much communication 0:03:48.480 --> 0:03:49.280 do you feel like you need? 0:03:50.880 --> 0:03:52.600 I just kind of want to know what's going on 0:03:52.720 --> 0:03:55.680 in there. Like the whales are definitely talking to each other. 0:03:55.720 --> 0:03:58.920 What are they saying? The honeybees are wiggling to each other. 0:03:58.960 --> 0:04:01.760 This definitely in for me. I just want to translate 0:04:01.800 --> 0:04:04.640 that information somehow, to know what they think is interesting 0:04:04.720 --> 0:04:07.160 to talk about. And like you, I know that my 0:04:07.280 --> 0:04:09.680 dog understands a lot of stuff, Like he picks up 0:04:09.720 --> 0:04:12.120 on pretty subtle patterns, so he knows when he's getting 0:04:12.160 --> 0:04:15.120 a walk. You know, shoes on plus the leash means 0:04:15.120 --> 0:04:17.520 a walk, shoes on plus the backpack I means Daniel's 0:04:17.520 --> 0:04:19.920 going to work this kind of stuff. And he notices 0:04:20.000 --> 0:04:23.080 patterns through the day, like he's definitely paying attention to 0:04:23.120 --> 0:04:26.040 patterns and understanding stuff. But I wish we could talk. 0:04:26.080 --> 0:04:27.520 I wish we could commun in the game more. I 0:04:27.560 --> 0:04:29.479 wish I could tell him like, Hey, we're getting in 0:04:29.480 --> 0:04:31.200 the car and we're going to this thing, or we're 0:04:31.240 --> 0:04:33.599 not going to the vet. You know this information I 0:04:33.600 --> 0:04:35.920 wish we could share with him, Like my dog when 0:04:35.920 --> 0:04:37.839 we leave, he has no idea are we going on 0:04:37.880 --> 0:04:39.960 a three week trip? Are we going to be back 0:04:40.000 --> 0:04:41.960 in an hour? He's cool, and if I could talk 0:04:42.000 --> 0:04:43.440 to him, I could share that with him. 0:04:43.560 --> 0:04:45.159 So, first of all, I think it's very nice that 0:04:45.240 --> 0:04:47.880 you're worried about how your dog's going to feel about 0:04:47.880 --> 0:04:50.960 how long you're gone. I think that's very sweet. So 0:04:51.120 --> 0:04:53.960 bees do this waggle dance, and you can tell by 0:04:53.960 --> 0:04:56.359 the way that they're dancing, sort of like the direction 0:04:56.560 --> 0:04:59.640 and the distance that they're telling their nest mats. We're 0:04:59.640 --> 0:05:02.120 the great source of pollen or nectar is and so 0:05:02.360 --> 0:05:05.400 we can decipher that. But you know, I don't think 0:05:05.440 --> 0:05:07.080 that I could like dress up in a bee suit 0:05:07.160 --> 0:05:09.560 and shake my butt the right way and tell them like, hey, 0:05:09.560 --> 0:05:12.160 I just planted some flowers in the southern corner of 0:05:12.200 --> 0:05:15.279 the property. I mean, I think we can decode a 0:05:15.279 --> 0:05:17.160 lot of what animals say, but we have not yet 0:05:17.200 --> 0:05:19.680 figured out how to communicate back with them. 0:05:19.920 --> 0:05:22.600 Yeah. Well, part of that depends on what's going on 0:05:22.720 --> 0:05:26.120 inside those animals' minds and what intelligence really means and 0:05:26.160 --> 0:05:28.720 where it comes from. So let's dig into the episode today. 0:05:28.760 --> 0:05:31.440 I'm dying here what you had to tell us. 0:05:31.720 --> 0:05:34.080 I'm super excited about today's episode. But now I've got 0:05:34.080 --> 0:05:36.920 myself wondering, like, could you make a little electronic bee 0:05:37.279 --> 0:05:39.520 that was like a bee size and like do the 0:05:39.600 --> 0:05:41.880 waggle dance and maybe get the bees to go to 0:05:42.040 --> 0:05:44.800 like your source of nectar. Anyway, surely there's got to 0:05:44.800 --> 0:05:47.560 be funding for that. That sounds awesome. I find questions 0:05:47.600 --> 0:05:50.599 about animal intelligence to be absolutely fascinating, and this is 0:05:50.600 --> 0:05:52.960 actually a topic that we get a lot of questions 0:05:52.960 --> 0:05:55.200 from y'all about, And if you want to send us 0:05:55.200 --> 0:05:57.880 your questions, you can send us questions at questions at 0:05:57.960 --> 0:06:01.360 danieland Kelly dot org. So today we're going to tackle 0:06:01.600 --> 0:06:04.640 three of the questions that we've gotten from listeners about 0:06:04.680 --> 0:06:07.440 animal intelligence and they sort of build on one another, 0:06:07.520 --> 0:06:09.040 and so we're gonna, you know, do a bit of 0:06:09.080 --> 0:06:11.000 an overview on animal intelligence today. 0:06:11.440 --> 0:06:13.840 This is super cool. Second new kind of episode is 0:06:13.880 --> 0:06:17.320 like a listener questions themes episode, almost like these listeners 0:06:17.320 --> 0:06:19.799 are working together to send us questions that click together 0:06:19.839 --> 0:06:20.360 so nicely. 0:06:20.600 --> 0:06:23.000 That's right, Yeah, and these three click together very nicely. 0:06:23.040 --> 0:06:24.760 You know, we could have done something similar with our 0:06:24.800 --> 0:06:28.280 consciousness episode, but I think we got twenty questions about 0:06:28.279 --> 0:06:30.279 consciousness before we were like, all right, we get it, 0:06:30.360 --> 0:06:33.160 We're going to get an expert. Hopefully we answered everyone's 0:06:33.240 --> 0:06:36.320 questions with that one. But today we are going to 0:06:36.360 --> 0:06:39.800 start with a question from Tom. So let's go ahead 0:06:39.800 --> 0:06:40.880 and hear Tom's question. 0:06:41.720 --> 0:06:45.000 I'm reasonably confident that humans are the only Earth species 0:06:45.040 --> 0:06:48.560 that has ever built rockets and such. So why is 0:06:48.560 --> 0:06:52.039 this level of intelligence so rare? Life on Earth is 0:06:52.080 --> 0:06:55.560 incredibly diverse and the Earth has been teeming with it 0:06:55.640 --> 0:06:59.200 for billions of years? How does the species even develop 0:06:59.279 --> 0:07:04.760 rocket level intelligence? Is intelligence a biologically selectable property that 0:07:04.839 --> 0:07:08.839 improves the chances for a population to survive? Or is 0:07:08.920 --> 0:07:12.360 rocket level intelligence just dumb luck? After all, the fact 0:07:12.400 --> 0:07:15.720 that rocket level intelligence emerged on Earth has proved to 0:07:15.760 --> 0:07:19.560 be quite unlucky for many coexisting species and may yet 0:07:19.600 --> 0:07:21.920 still be unlucky for humans. Thanks. 0:07:22.880 --> 0:07:26.400 Wow. Okay, so that is a great question, but it's 0:07:26.400 --> 0:07:30.280 a very specific question. So why is a rocket level 0:07:30.400 --> 0:07:34.120 intelligence so rare? So we have an end of one. 0:07:34.320 --> 0:07:37.800 We know of one species that is able to build rockets, 0:07:37.840 --> 0:07:40.320 and with such a small sample size it can be 0:07:40.400 --> 0:07:42.800 really hard to know for sure. A bit of a 0:07:42.840 --> 0:07:46.840 spoiler here, we don't know. In general, science does not 0:07:47.040 --> 0:07:49.600 know exactly what it was that put humans on the 0:07:49.640 --> 0:07:51.960 trajectory that we went on to be able to build 0:07:52.040 --> 0:07:54.600 rockets and all the other amazing things that we can do. 0:07:54.960 --> 0:07:57.360 We will talk about some of the hypotheses today that 0:07:57.400 --> 0:07:59.880 we think were important, but first let's back up and 0:08:00.080 --> 0:08:04.680 ask the question why can't all species build rockets? Why 0:08:04.840 --> 0:08:06.440 is this kind of intelligence so rare? 0:08:06.840 --> 0:08:09.560 I mean, ants and termites can build impressive stuff, right, 0:08:09.640 --> 0:08:13.240 they definitely work together, and we see other species doing 0:08:13.240 --> 0:08:15.840 cool stuff. So why aren't they building rockets and getting 0:08:15.840 --> 0:08:17.480 off planet? What does that really require? 0:08:17.960 --> 0:08:20.840 So again, the rocket thing is very specific, and part 0:08:20.840 --> 0:08:22.760 of it is, you know, ants just might not have 0:08:22.800 --> 0:08:25.320 the right appendages to build a rocket. You know, maybe 0:08:25.320 --> 0:08:28.040 they've got rocket plans and if they just had hands, 0:08:28.080 --> 0:08:29.920 they'd be building rockets, but probably not. 0:08:30.760 --> 0:08:32.320 That's the kind of thing we could learn if we 0:08:32.400 --> 0:08:35.040 learned to speak ant and be you know, maybe they're 0:08:35.360 --> 0:08:39.280 desperate to fly into space and frustrated about their limitations. 0:08:39.760 --> 0:08:42.120 We should be bringing them with us just you know, 0:08:42.360 --> 0:08:42.760 they might. 0:08:42.679 --> 0:08:44.960 Appreciate that, or we should just have them on the 0:08:44.960 --> 0:08:46.839 podcast and ask them directly. 0:08:46.920 --> 0:08:50.720 Maybe in five or ten years. So if you ask yourself, 0:08:50.760 --> 0:08:53.840 why can't all species build rockets. You're essentially asking, well, 0:08:53.840 --> 0:08:58.160 why aren't all animals intelligent? What holds you back? And 0:08:58.440 --> 0:09:00.400 we think the answer has to do with how much 0:09:00.559 --> 0:09:04.000 energy it takes to be super smart. So our adult 0:09:04.040 --> 0:09:07.320 brains weigh about three pounds, so for a person who 0:09:07.320 --> 0:09:10.080 weighs one hundred and eighty pounds, that's about two percent 0:09:10.120 --> 0:09:14.480 of your body weight. But by measuring oxygen consumption rates 0:09:14.480 --> 0:09:16.840 in different parts of our body, we estimate that our 0:09:16.920 --> 0:09:20.200 brain uses about twenty percent of all of the energy 0:09:20.200 --> 0:09:23.640 that we burn every single day. So this very small 0:09:23.760 --> 0:09:27.160 organ is using an outsized amount of energy. So brains 0:09:27.200 --> 0:09:30.560 are just energetically expensive. So if you are a species 0:09:30.600 --> 0:09:34.320 that doesn't need a super big, smart brain, then there's 0:09:34.360 --> 0:09:37.560 probably not great selection pressure for you to be expending 0:09:37.559 --> 0:09:40.120 a bunch of energy on an energetically expensive organ that 0:09:40.160 --> 0:09:40.960 you are not using. 0:09:41.440 --> 0:09:44.080 So you're saying that this thing is expensive, right, that 0:09:44.160 --> 0:09:46.839 it costs us something, and then so if it doesn't 0:09:46.880 --> 0:09:49.520 pay for itself, it's definitely not going to survive because 0:09:49.520 --> 0:09:51.440 I've hung around and used up a bunch of energy 0:09:51.559 --> 0:09:54.960 and didn't provide tangible benefits. Then folks with big brains 0:09:55.240 --> 0:09:58.400 would die more often during famines and cold winters and 0:09:58.400 --> 0:10:00.520 stuff like that. That's the idea. 0:10:00.679 --> 0:10:02.640 Yeah, so this is more of a species level argument. 0:10:02.679 --> 0:10:04.880 But yeah, so species with big brains that didn't need 0:10:04.920 --> 0:10:08.240 them would have to spend more time foraging or get better, 0:10:08.360 --> 0:10:11.120 higher quality food because of their big brains. And if 0:10:11.120 --> 0:10:12.640 that's not happening, it's not worth having. 0:10:12.880 --> 0:10:14.960 I guess you could also imagine a population, right of 0:10:14.960 --> 0:10:17.400 a single species with smaller brains and bigger brains, And 0:10:17.440 --> 0:10:19.760 if the bigger brains aren't providing a benefit and they 0:10:19.760 --> 0:10:23.360 are providing a cost, then those big brain folks among 0:10:23.440 --> 0:10:25.920 us are not going to survive. There is a population 0:10:26.040 --> 0:10:27.439 level argument there also, right. 0:10:27.440 --> 0:10:31.120 Yeah, there's a population level argument. There's often more variables 0:10:31.160 --> 0:10:33.720 at play when you start thinking about a specific species. So, 0:10:33.840 --> 0:10:36.480 for example, when human brains, if they were to get 0:10:36.559 --> 0:10:39.240 too big, you would start to have trouble getting those 0:10:39.320 --> 0:10:42.160 giant heads through the birth canal. And so there's limits 0:10:42.160 --> 0:10:44.480 on how big brains can be. And in general, there's 0:10:44.520 --> 0:10:47.240 not a ton of variability between the sizes of brains 0:10:47.240 --> 0:10:50.160 of adults. And now we can get into this whole 0:10:50.240 --> 0:10:53.280 different debate about is it really about how big the 0:10:53.320 --> 0:10:56.480 brain is? So like blue whales have much bigger brains 0:10:56.480 --> 0:10:58.839 than we do, but they haven't created rockets to get 0:10:58.840 --> 0:11:01.000 them out of space. Maybe that's because it's just too 0:11:01.040 --> 0:11:03.840 expensive to send a giant blue whale to space. But yeah, 0:11:03.840 --> 0:11:05.480 so trying to figure out exactly what feature of the 0:11:05.480 --> 0:11:07.400 brain we should measure is difficult. But we know that 0:11:07.440 --> 0:11:11.040 brains are expensive energetically, right, And. 0:11:10.960 --> 0:11:12.880 I was actually reading a study about the size of 0:11:12.880 --> 0:11:15.600 the human brain that suggested that if you packed in 0:11:15.640 --> 0:11:18.680 the neurons more densely, you would add more noise. So 0:11:18.760 --> 0:11:21.959 making the neurons smaller and denser wouldn't actually improve your 0:11:22.000 --> 0:11:24.520 cognition because you'd be ending up adding noise. And so like, 0:11:24.800 --> 0:11:27.160 human brains are sort of at this sweet spot for 0:11:27.360 --> 0:11:30.360 how much you can squeeze them down to get cognition. 0:11:30.679 --> 0:11:32.199 I think maybe one of these days we should have 0:11:32.240 --> 0:11:36.360 a whole separate episode on exactly how you should measure 0:11:36.400 --> 0:11:39.160 brains to correlate it with intelligence, because I read a 0:11:39.160 --> 0:11:42.200 paper that was arguing that actually, the reason we're so 0:11:42.280 --> 0:11:46.080 great is because we have denser neurons, and that it's 0:11:46.160 --> 0:11:50.000 not brain size, it's neuron density that really unlocks intelligence. 0:11:50.160 --> 0:11:53.160 I think in general, the field has not figured out 0:11:53.200 --> 0:11:54.720 exactly what we need to be measuring. 0:11:55.480 --> 0:11:57.959 That's such a nice way of saying that these folks 0:11:57.960 --> 0:12:02.280 don't know what they're doing. Well, you gotta start is immature, Yes, well. 0:12:02.160 --> 0:12:05.160 Yes, brains are complicated man, Yes they are. 0:12:05.440 --> 0:12:07.760 No, you're totally right in great respect to folks who 0:12:07.800 --> 0:12:10.320 are studying the brain, because obviously it's important. And just 0:12:10.360 --> 0:12:13.080 because we don't understand everything doesn't mean that early studies 0:12:13.080 --> 0:12:15.240 are not valuable. They're the ones that lay the foundation 0:12:15.360 --> 0:12:18.760 for future progress. Yes, absolutely, and so we shouldn't mock them. 0:12:18.920 --> 0:12:20.760 We're grateful for their pioneering work. 0:12:20.880 --> 0:12:24.040 Yes, I studied a brain infecting parasite of fish and 0:12:24.080 --> 0:12:28.520 I started getting into neuroanatomy and neurochemistry and I was like, no, 0:12:29.080 --> 0:12:32.679 I'm sticking with parasitology. Brains are complicated, so anyway, but 0:12:32.720 --> 0:12:33.840 here we're jumping in anyway. 0:12:34.000 --> 0:12:36.480 I think also Tom's question has this fascinating flavor to it, 0:12:36.520 --> 0:12:39.600 which is like, you can make the evolutionary argument, and 0:12:39.600 --> 0:12:41.360 I think you're going to get into this later for 0:12:41.480 --> 0:12:45.520 why intelligence might be beneficial, right, it might even overcome 0:12:45.559 --> 0:12:48.920 the expense of the brain. But it's hard to understand 0:12:49.000 --> 0:12:51.200 why evolution would give us a brain that's capable of 0:12:51.200 --> 0:12:54.679 doing things that you didn't need to survive on the savannah, Right, 0:12:54.720 --> 0:12:56.840 You didn't need to be able to think about eleven 0:12:56.880 --> 0:12:59.719 dimensional mathematics and build rockets and all sorts of other 0:13:00.080 --> 0:13:04.000 easy stuff in order to survive evolutionarily. So why is 0:13:04.040 --> 0:13:07.880 it that evolution provided this tool which has this crazy 0:13:07.920 --> 0:13:10.400 capability clearly outstripping our needs. 0:13:10.559 --> 0:13:12.880 Yeah, well, you could ask a similar question like why 0:13:12.880 --> 0:13:15.840 do peacocks have such big, showy tales? And I think 0:13:15.920 --> 0:13:18.600 that you know, natural selection doesn't just select for things 0:13:18.640 --> 0:13:21.680 that help us run faster to catch our food or 0:13:21.720 --> 0:13:25.720 help us garden more efficiently. It also selects for traits 0:13:25.720 --> 0:13:28.439 that make us more attractive mates in one way or another. 0:13:28.480 --> 0:13:31.600 So sometimes it makes you know, you bigger and more 0:13:31.679 --> 0:13:34.120 muscly if that's what the females want. But you know, 0:13:34.160 --> 0:13:36.840 there's also variability in what females of a species want, 0:13:36.920 --> 0:13:40.080 And some of us want super smart partners, our partners 0:13:40.080 --> 0:13:42.800 with musical abilities. And I think humans are kind of 0:13:42.960 --> 0:13:45.480 weird in this respect, that we really are interested in 0:13:45.480 --> 0:13:49.800 intellectual characteristics in addition to just physical characteristics. And I 0:13:49.800 --> 0:13:52.360 don't know exactly how that played into our evolution and 0:13:52.400 --> 0:13:54.640 made us unique. But you know, if you can do 0:13:54.760 --> 0:13:57.559 nine dimensional math, there's got to be a group of 0:13:57.600 --> 0:13:58.480 people who are into that. 0:14:00.480 --> 0:14:02.680 Since both of us are married to nerds, I think 0:14:02.679 --> 0:14:06.520 we may be an unrepresentative subset of humanity. But I 0:14:06.559 --> 0:14:08.600 totally agree with you. Big brains are hot. 0:14:08.679 --> 0:14:11.440 Yes, it's super hot, way more hot than muscles and 0:14:11.520 --> 0:14:15.400 being in good shape, I think. But could be biased. Okay, 0:14:15.840 --> 0:14:18.240 So the question was essentially about how do we get 0:14:18.640 --> 0:14:21.880 humans in particular, and the answer is, we don't know 0:14:21.880 --> 0:14:24.320 about how you get humans in particular, but we can 0:14:24.400 --> 0:14:27.960 get some hints by looking at other kinds of animals 0:14:27.960 --> 0:14:31.560 and trying to figure out what things selected for intelligence 0:14:31.680 --> 0:14:34.520 in other animals, and that might give us some insights 0:14:34.560 --> 0:14:36.760 into like maybe you know, we maxed out a particular 0:14:37.120 --> 0:14:39.280 you know, trait, or there was a particularly strong selective 0:14:39.280 --> 0:14:42.400 pressure for us. So we're going to start this conversation 0:14:43.200 --> 0:14:47.120 by talking about how we define intelligence outside of the 0:14:47.160 --> 0:14:51.160 ability to build rockets, and then some difficulties with measuring it, 0:14:51.200 --> 0:14:53.200 and towards the end of the episode, when we're answering 0:14:53.240 --> 0:14:55.720 other questions from other listeners, we'll talk about some of 0:14:55.760 --> 0:14:57.680 the leading hypotheses. So we're going to get to. 0:14:57.640 --> 0:15:00.000 Them, all right. I love when we start a hard 0:15:00.080 --> 0:15:03.880 conversation with let's define what we're talking about, because it 0:15:03.920 --> 0:15:06.320 makes me feel all philosophical and nerdy. But also it's 0:15:06.400 --> 0:15:08.840 really important. If somebody doesn't do that, then I feel 0:15:08.840 --> 0:15:11.040 like they're going to be sloppy with their arguments because 0:15:11.040 --> 0:15:12.480 they're not being precise with their words. 0:15:12.920 --> 0:15:15.120 Well, and I should say that while I was researching 0:15:15.120 --> 0:15:19.600 this topic, I came across a lot of different definitions, 0:15:19.680 --> 0:15:21.800 and I decided to try to go with something a 0:15:21.840 --> 0:15:24.160 little bit more general that feels pretty good to all 0:15:24.240 --> 0:15:26.120 of us. So here's what I came up with. It's 0:15:26.200 --> 0:15:29.920 just the ability to solve lots of different kinds of problems, 0:15:30.040 --> 0:15:33.200 because that's what they all seemed to narrow down to. 0:15:34.080 --> 0:15:36.200 But what do you think how would you define intelligence? 0:15:36.560 --> 0:15:38.840 I definitely think the ability to solve problems is a 0:15:38.920 --> 0:15:41.320 component of it. It feels to me like also there 0:15:41.360 --> 0:15:46.200 should be a learning aspect, right, discovering patterns generalizing from them, 0:15:47.040 --> 0:15:49.720 you know, in machine learning, that's a big factor. We 0:15:49.760 --> 0:15:52.640 want to develop networks that are not just capable of 0:15:52.680 --> 0:15:56.120 solving the problems we've taught them, but solving a general 0:15:56.200 --> 0:16:00.480 class of problems, including examples they haven't seen before, bringing 0:16:00.520 --> 0:16:04.640 together ideas, coming up with bigger picture solutions. So something 0:16:04.680 --> 0:16:08.520 about generalizing from the examples that you've learned feels to 0:16:08.560 --> 0:16:10.240 me like an important part of intelligence. 0:16:10.560 --> 0:16:13.200 Yeah, and I think that having the ability to solve 0:16:13.240 --> 0:16:16.080 lots of different kinds of problems does, to some extent 0:16:16.120 --> 0:16:18.800 imply the ability to learn, Okay, yeah, or at least 0:16:18.840 --> 0:16:21.240 to experiment with your environment to try new things, which 0:16:21.280 --> 0:16:24.000 I think is like learning. So I think we're pretty 0:16:24.040 --> 0:16:26.080 much on the same page. What do you think. 0:16:26.280 --> 0:16:28.680 I think that's a good working definition. Okay, So then 0:16:28.920 --> 0:16:31.400 how do we measure this? Are we like giving tests 0:16:31.440 --> 0:16:34.320 to cats and dogs to see can you do eleven 0:16:34.320 --> 0:16:35.400 dimensional string theory? 0:16:35.760 --> 0:16:38.360 We're not asking them about eleven dimensional string theory. We 0:16:38.400 --> 0:16:40.840 will talk about some experiments that we do, but first 0:16:40.880 --> 0:16:43.000 I wanted to tell a couple of stories about how 0:16:43.040 --> 0:16:46.240 we've thought we were measuring something but it turns out 0:16:46.240 --> 0:16:49.160 we're measuring something different. Just to highlight how difficult it is. 0:16:50.880 --> 0:16:52.280 Here comes wet Blanket Kelly. 0:16:52.840 --> 0:16:54.320 Come on, of course, we were going to get to 0:16:54.360 --> 0:17:00.760 wet Blanka Kelly. Okay, So clever Hans is horse in 0:17:00.880 --> 0:17:03.520 Germany at the end of the nineteenth century, and this 0:17:03.600 --> 0:17:07.360 horse became super famous because you could ask him questions 0:17:07.400 --> 0:17:13.480 about addition, subtraction, multiplication, division, including questions that involved fractions wow, 0:17:13.640 --> 0:17:16.800 and he would stomp out answers, so he would like 0:17:16.960 --> 0:17:19.600 stomp and then when he was done answering the question, 0:17:19.720 --> 0:17:21.560 so for the right numbers, like if the answer was five, 0:17:21.600 --> 0:17:23.760 he'd stop five times and then he'd sort of do 0:17:23.960 --> 0:17:26.399 like a little circle thing with his hoof to be 0:17:26.400 --> 0:17:27.160 like okay, I'm done. 0:17:27.600 --> 0:17:28.159 Wow. 0:17:28.280 --> 0:17:30.840 And then you could also ask Clever Hans questions like 0:17:31.280 --> 0:17:34.800 what is that woman holding in her hand? And using 0:17:34.920 --> 0:17:37.280 an alphabet that was on a board, each letter was 0:17:37.280 --> 0:17:41.320 separated into columns and rows, he could clamp out what 0:17:41.480 --> 0:17:46.200 letters he wanted to say, you know, umbrella or hot 0:17:46.240 --> 0:17:47.960 dog or something like that. 0:17:47.960 --> 0:17:51.359 That implies some ability to read, right, yeah. 0:17:51.280 --> 0:17:54.400 Yeah, yeah right. Some experts came in and they were like, 0:17:54.560 --> 0:17:56.919 oh my gosh, we totally agree because we can do 0:17:56.960 --> 0:17:58.920 the trick with Clever Hans too. We can ask clever 0:17:58.960 --> 0:18:01.600 Hans the same question, it's not just his owner, and 0:18:01.640 --> 0:18:03.959 he can answer it. So they decided that Clever Hans 0:18:04.320 --> 0:18:07.320 had the intelligence that was equivalent to about a thirteen 0:18:07.400 --> 0:18:12.320 year old child. But then another set of experimenters came 0:18:12.320 --> 0:18:16.480 in and they started walling clever Hans off from the 0:18:16.560 --> 0:18:19.440 room where the questioner was asking. 0:18:19.880 --> 0:18:22.560 Oh, yeah, that's right. 0:18:23.880 --> 0:18:28.600 It turns out that clever Hans could only answer questions 0:18:28.640 --> 0:18:32.200 correctly if the experimenter knew the answer ahead of time, 0:18:32.840 --> 0:18:35.080 and if Hans could see the experimenter. 0:18:35.720 --> 0:18:38.720 Oh, man, this is just like those experiments with the 0:18:38.720 --> 0:18:40.880 wiki boards, right, uh right. 0:18:40.800 --> 0:18:43.320 Like something else is sort of happening behind the scenes 0:18:43.359 --> 0:18:47.120 that's unintended. So it turned out clever Hans isn't smart 0:18:47.280 --> 0:18:49.960 in terms of the ability to do like division, but 0:18:50.040 --> 0:18:54.240 clever Hans was very smart at reading human body language. 0:18:54.480 --> 0:18:56.679 So even if you put a new person in that 0:18:56.720 --> 0:18:59.720 clever Hans had never encountered before, Clever Hans could tell 0:18:59.720 --> 0:19:02.920 that they were like holding their breath until he got 0:19:02.920 --> 0:19:05.120 to like the right number of stomps, and that then 0:19:05.200 --> 0:19:08.240 something about their body would change and then he'd know, okay, 0:19:08.240 --> 0:19:10.840 that was it. And so clever Hans was actually just 0:19:10.960 --> 0:19:12.200 queuing it on body language. 0:19:12.280 --> 0:19:15.399 Wow, So clever Hans could really read human emotion and 0:19:15.520 --> 0:19:18.840 like understand what humans wanted him to do and then 0:19:18.880 --> 0:19:19.359 would do it. 0:19:19.600 --> 0:19:22.520 Yes, right. But this points out one of the first 0:19:22.560 --> 0:19:25.840 difficult things about studying animals is that sometimes it can 0:19:25.840 --> 0:19:28.720 be really hard to figure out what you're measuring or 0:19:28.760 --> 0:19:30.639 what it is that they're actually learning, and it's not 0:19:30.680 --> 0:19:32.840 always what you think that they're learning. So you have 0:19:32.880 --> 0:19:35.639 to be very careful about how you design experiments, and 0:19:35.680 --> 0:19:37.760 you need to keep in mind that animals sense the 0:19:37.760 --> 0:19:40.760 world in different ways. So you know, some animals echo locate, 0:19:40.840 --> 0:19:45.320 some can see polarized light, some explore their universe using electricity, 0:19:45.400 --> 0:19:47.159 and so you need to make sure that you're asking 0:19:47.280 --> 0:19:50.040