1 00:00:00,000 --> 00:00:01,940 Fraser McGruer: Hello and welcome to the Cognitive 2 00:00:01,997 --> 00:00:05,249 Engineering podcast, brought to you by Aleph Insights and 3 00:00:05,306 --> 00:00:08,844 produced by me, Fraser McGruer. I'm here with Nick Hare, Chris 4 00:00:08,901 --> 00:00:12,438 Wragg and Peter Coghill of Aleph Insights. On this podcast, we 5 00:00:12,495 --> 00:00:16,090 look at a wide range of topics, and today we are discussing the 6 00:00:16,147 --> 00:00:17,460 dangers of Mirror Life. 7 00:00:25,700 --> 00:00:28,160 Chris, what on earth is mirror life? I have no idea what we're 8 00:00:28,160 --> 00:00:28,700 talking about. 9 00:00:28,700 --> 00:00:30,620 Chris Wragg: Yeah, this isn't some reference to sort of the, 10 00:00:30,860 --> 00:00:33,860 you know, narcissism of modern culture or something. This is 11 00:00:33,860 --> 00:00:37,940 actually a whole sort of discipline of science, which 12 00:00:37,940 --> 00:00:42,280 involves mirror organisms, such as mirror bacteria that are 13 00:00:42,280 --> 00:00:46,840 constructed from mirror images of molecules that are found in 14 00:00:46,840 --> 00:00:51,280 nature. So if you take our DNA and all the DNA you know on 15 00:00:51,280 --> 00:00:56,620 Earth, it is all what is called right handed so it's helix 16 00:00:56,680 --> 00:01:01,680 spirals in a clockwise direction. And and then, you 17 00:01:01,680 --> 00:01:06,240 know, proteins are all that's because righty tighty lefty 18 00:01:06,240 --> 00:01:09,600 loose. Righty tighty lefty loose. Right handed means it's, 19 00:01:09,600 --> 00:01:13,080 if you're going up a staircase, right, a spiral staircase, and 20 00:01:13,080 --> 00:01:16,680 you would use your right hand on the inner Bannister, you're 21 00:01:16,800 --> 00:01:19,440 going in that direction, right? I think, I think the 22 00:01:19,440 --> 00:01:21,920 Peter Coghill: sprinkle a bit more technical jargon. Yeah, 23 00:01:22,640 --> 00:01:27,320 certain molecules have isomorphism, which means that 24 00:01:27,320 --> 00:01:32,060 you can have it looking one way, or you can get a mirror image of 25 00:01:32,060 --> 00:01:32,420 that, 26 00:01:32,480 --> 00:01:35,120 Nick Hare: an enantiomorph, if we're trying to use Yes, 27 00:01:35,120 --> 00:01:37,940 exactly, yeah, an opposite chiral 28 00:01:38,000 --> 00:01:40,600 Peter Coghill: molecules don't so simple sugars are just 29 00:01:40,720 --> 00:01:44,260 symmetrical in whatever way you look at them. Boring, yeah. Then 30 00:01:44,260 --> 00:01:48,280 some molecules, such as the the protein, the molecules of 31 00:01:48,280 --> 00:01:53,020 makeup, proteins and some sugars, some alcohols, etc, are 32 00:01:53,020 --> 00:01:56,320 handed they you can have one version or a mirror image of 33 00:01:56,320 --> 00:01:59,920 that version, and they can naturally, they can actually 34 00:01:59,920 --> 00:02:00,100 occur. 35 00:02:00,100 --> 00:02:03,600 Nick Hare: But isn't it true that orange, the taste of 36 00:02:03,600 --> 00:02:06,360 orange, is actually caused by a molecule that is the mirror 37 00:02:06,360 --> 00:02:07,980 image of the one that causes the taste of 38 00:02:07,980 --> 00:02:10,800 Peter Coghill: lemon, something like that? Yeah. But it turns 39 00:02:10,800 --> 00:02:16,380 out that all life on planet Earth has, for some reason, sort 40 00:02:16,380 --> 00:02:20,600 of sent that gone towards one handedness. What is always right 41 00:02:20,600 --> 00:02:21,200 handed. 42 00:02:21,799 --> 00:02:22,639 Chris Wragg: The implications 43 00:02:22,640 --> 00:02:25,520 Fraser McGruer: of this being, before you go on, if you can 44 00:02:25,520 --> 00:02:28,520 tuck a little bit further on under or will it not go down 45 00:02:28,520 --> 00:02:31,520 further? No, or There you go. I've got so many, okay, I've got 46 00:02:31,520 --> 00:02:33,680 so many questions, but keep going. Chris, yeah. 47 00:02:34,160 --> 00:02:37,520 Chris Wragg: I mean the key, the key thing, really, never mind 48 00:02:37,520 --> 00:02:41,200 why this occurs. The fact is, it's, it's a phenomenon in 49 00:02:41,200 --> 00:02:47,920 nature, and it applies to all organisms. And the key issue is 50 00:02:47,920 --> 00:02:53,620 that human scientists have been experimenting with the creation 51 00:02:53,680 --> 00:03:00,300 of mirror images of naturally occurring molecules, and in some 52 00:03:00,300 --> 00:03:04,440 cases, starting to consider how you might create a mirror image 53 00:03:04,440 --> 00:03:09,000 bacteria right for various issues. Now what this means is, 54 00:03:09,000 --> 00:03:12,840 right. So there are, you know, some benefits around, you know, 55 00:03:12,840 --> 00:03:17,040 treatment of disease, being able to produce chemicals at scale 56 00:03:17,040 --> 00:03:21,380 and so on. But the potential massive downside to this is that 57 00:03:21,380 --> 00:03:27,020 you create bacteria for which our immune system has no ability 58 00:03:27,620 --> 00:03:32,960 to recognise or contain. You know, if you think about like a 59 00:03:33,200 --> 00:03:37,640 phone charger, you know, our our immune cells are a little bit 60 00:03:37,640 --> 00:03:40,700 like, you know, a phone charger that you're trying to plug into 61 00:03:40,700 --> 00:03:42,880 the to the bacteria, and if you've got the wrong phone 62 00:03:42,880 --> 00:03:47,440 charger, you're a bit buggered. And so we're potentially the 63 00:03:47,440 --> 00:03:52,900 issue is we're potentially creating super bugs, I suppose 64 00:03:52,960 --> 00:03:59,500 that that all organisms have no natural defence against and so 65 00:03:59,860 --> 00:04:03,000 the scientific community itself has, has sort of been 66 00:04:03,000 --> 00:04:06,480 increasingly realising this is probably something it shouldn't 67 00:04:06,480 --> 00:04:10,500 do. And so there was a there was a large sort of paper and a 68 00:04:10,560 --> 00:04:14,040 commentary on it in the Journal of Science last year where 69 00:04:14,040 --> 00:04:18,240 hundreds of scientists came together, you know, Nobel Prize 70 00:04:18,240 --> 00:04:21,740 winning scientists, and they said, Unless compelling evidence 71 00:04:21,740 --> 00:04:26,600 emerges that mirror life would not pose extraordinary dangers, 72 00:04:26,900 --> 00:04:31,280 we believe that mirror bacteria and other mirror organisms, even 73 00:04:31,280 --> 00:04:34,760 those with engineered bio containment measures, should not 74 00:04:34,760 --> 00:04:41,260 be created so effectively trying to Create a moratorium on this 75 00:04:41,320 --> 00:04:44,860 element of science, because they feel it could pose an absolutely 76 00:04:44,860 --> 00:04:46,960 catastrophic risk too. 77 00:04:47,680 --> 00:04:51,280 Fraser McGruer: Okay, look, a couple of things. First of all, 78 00:04:51,460 --> 00:04:54,580 we've done a lot of podcasts together, and I'm often felt 79 00:04:54,580 --> 00:04:58,300 confused during these things, right? I have no idea what we're 80 00:04:58,300 --> 00:04:59,080 talking about. 81 00:04:59,440 --> 00:05:01,740 Peter Coghill: Don't think. It matters, good. I don't think it 82 00:05:01,740 --> 00:05:05,640 matters because I think we're going to treat it as the 83 00:05:05,640 --> 00:05:11,100 biochemistry existential risk equivalent of AI takeover. Fine. 84 00:05:11,160 --> 00:05:13,200 Fraser McGruer: Think that's okay, good. And also, lots of 85 00:05:13,200 --> 00:05:16,380 these words I've never heard of before, and sometimes when stuff 86 00:05:16,380 --> 00:05:18,840 gets explained to you, you understand it more, but 87 00:05:18,840 --> 00:05:23,540 sometimes you understand it less. We've made you stupider. 88 00:05:23,540 --> 00:05:25,640 Yeah, who knew, right, that this? 89 00:05:25,640 --> 00:05:27,980 Peter Coghill: Yeah, tell intelligence without words. 90 00:05:27,980 --> 00:05:28,400 Yeah. I can't 91 00:05:28,400 --> 00:05:30,440 Chris Wragg: help but feel this is my problem. But the 92 00:05:30,440 --> 00:05:33,620 fundamental question behind it, and also, why are we discussing 93 00:05:33,620 --> 00:05:37,340 this, right? Okay, so essentially, what we're what 94 00:05:37,340 --> 00:05:43,060 we're looking at, is whether or not our demise as a species is 95 00:05:43,060 --> 00:05:48,400 going to be brought about by stuff we do, or some other, some 96 00:05:48,400 --> 00:05:50,800 other means. Are we our own worst enemy? 97 00:05:52,120 --> 00:05:55,540 Fraser McGruer: Okay, gotcha? Sounds good? Well, that's a 98 00:05:55,540 --> 00:05:59,920 nice, nice, happy topic. Who's right answer? Are we the are we? 99 00:06:00,040 --> 00:06:02,340 What was the what was the what was the question? Again, the 100 00:06:02,340 --> 00:06:05,580 simple question, Are we our own worst enemy? Okay, who's gonna 101 00:06:05,580 --> 00:06:09,840 take this So, Peter, are we? Are you your own worst enemy? Are 102 00:06:09,840 --> 00:06:09,960 you 103 00:06:09,960 --> 00:06:12,360 Peter Coghill: Fraser's worst enemy? Yeah, yes, yeah, 104 00:06:12,360 --> 00:06:15,900 definitely. Fraser never, plenty of those. I'm gonna phrase it as 105 00:06:15,900 --> 00:06:21,980 the the Promethean problem. So, you know, we we there's a 106 00:06:21,980 --> 00:06:24,740 paradox of progress in everything that drives our 107 00:06:24,740 --> 00:06:28,160 progress, in in cultural progress, technology progress, 108 00:06:28,160 --> 00:06:32,060 quality of life. Everything is kind of driven by our own 109 00:06:32,060 --> 00:06:35,480 creativity, our ambition, our curiosity, and our sort of 110 00:06:35,480 --> 00:06:39,560 ability to solve problems. But those things, those drivers 111 00:06:39,560 --> 00:06:43,600 ambition, curiosity, problem solving and creativity are 112 00:06:43,600 --> 00:06:46,720 indifferent, inherently indifferent, to moral progress. 113 00:06:46,720 --> 00:06:49,900 You know, we the moral consequence. You know, the moral 114 00:06:49,900 --> 00:06:53,200 consequence is the thing you bolt on kind of afterwards. We 115 00:06:53,260 --> 00:06:58,240 tend to create first and then understand later. So we, we, we 116 00:06:58,240 --> 00:07:01,680 build the bomb, we build an atomic bomb, and then we go, ah, 117 00:07:01,680 --> 00:07:04,140 actually, this kind of changes the whole power dynamic of the 118 00:07:04,200 --> 00:07:08,820 world. Let's, let's, let's now have some accords and some 119 00:07:08,940 --> 00:07:12,660 treaties to try to put a lid on this and and have a whole dog, 120 00:07:12,660 --> 00:07:15,720 dog more about mutual escort, sure destruction, which means 121 00:07:15,720 --> 00:07:18,540 that we won't, we won't use them. So it's sort of we, we 122 00:07:18,540 --> 00:07:24,320 have a tendency, as as a species, to to to open Pandora's 123 00:07:24,320 --> 00:07:28,520 Box, to bring in another clumsy, there's a lot of classical 124 00:07:28,520 --> 00:07:30,560 illusions clumsily bring in another class 125 00:07:30,560 --> 00:07:33,620 Nick Hare: for me, yes, opening a Pandora's box after the horse 126 00:07:33,620 --> 00:07:34,040 is bowled. 127 00:07:34,040 --> 00:07:36,440 Peter Coghill: So the credit the question narcissist earlier, 128 00:07:36,440 --> 00:07:42,700 yeah, we're back up to date. We are. So the question is, are we 129 00:07:42,760 --> 00:07:46,660 Prometheus, you know, getting fire, or are we the fire? That's 130 00:07:46,780 --> 00:07:52,900 the uncontrolled fire spread from a spark like it. Do we have 131 00:07:52,900 --> 00:07:56,200 it? Do we have, do we have a innovation addiction? 132 00:07:57,280 --> 00:08:01,200 Nick Hare: So, I mean, I think if we I think there are two 133 00:08:01,440 --> 00:08:04,680 really quite distinct categories of thing we need to worry about, 134 00:08:04,800 --> 00:08:09,600 because they're caused by different mechanisms. One is the 135 00:08:09,600 --> 00:08:14,400 deliberate destruction of the human species by decisions we 136 00:08:14,460 --> 00:08:19,440 take, and one is the accidental destruction of the human species 137 00:08:19,980 --> 00:08:23,720 unintended because of the side effects of decisions that we 138 00:08:23,720 --> 00:08:27,500 take, right? So, in other words, you know, are we going to cause 139 00:08:27,620 --> 00:08:30,200 our destruction because we want to? Basically, I would put a 140 00:08:30,200 --> 00:08:32,900 nuclear war in that category. You know, we're kind of actually 141 00:08:32,900 --> 00:08:36,860 making the choice to kill everyone. And I think that's a 142 00:08:36,860 --> 00:08:38,900 very different that's there's a whole different set of 143 00:08:38,900 --> 00:08:43,000 mechanisms there to when we worry about accidentally, you 144 00:08:43,000 --> 00:08:46,000 know, creating mirror life, or letting an AI take over the 145 00:08:46,000 --> 00:08:51,520 world, or whatever. And I think Peter's, you know, question 146 00:08:51,520 --> 00:08:57,040 thoughts about, how have we evolved in such a way to make 147 00:08:57,040 --> 00:09:00,660 this intrinsically problematic? Is a good place to start. So if 148 00:09:00,660 --> 00:09:03,420 we take the problem of us deliberately destroying the 149 00:09:03,420 --> 00:09:07,380 world by sort of just by analogy, if you think about 150 00:09:08,220 --> 00:09:12,360 fights, that is actually quite hard to kill someone. Now, I 151 00:09:12,360 --> 00:09:15,660 know you've killed many a person with your bare hands, trained 152 00:09:15,900 --> 00:09:21,920 trade boxer, Yep, yeah, yeah, exactly. But it's actually, it's 153 00:09:21,920 --> 00:09:24,560 actually quite hard. It is somewhat harder than people. 154 00:09:24,620 --> 00:09:32,300 Yeah, attacks with a gun are 30 times more lethal than attacks 155 00:09:32,360 --> 00:09:37,580 with no gun, right? For example. Okay, now, if we'd have evolved, 156 00:09:37,580 --> 00:09:40,900 if we'd have had guns, if we'd evolved with guns built into our 157 00:09:40,900 --> 00:09:44,980 fists, it's entirely plausible to me that we would have evolved 158 00:09:45,040 --> 00:09:49,300 much less of a tendency to get into fights. We would be much 159 00:09:49,420 --> 00:09:52,840 better at standing down, because the risk would be so much higher 160 00:09:52,960 --> 00:09:57,580 getting involved, right? So clearly our we've evolved have a 161 00:09:57,580 --> 00:09:59,920 level of aggression where it's quite hard to cause that much. 162 00:10:00,000 --> 00:10:04,440 Damage. Now we can't that level of aggression extended to the to 163 00:10:04,440 --> 00:10:08,340 the levels of, you know, at the country level, is, is, therefore 164 00:10:08,340 --> 00:10:13,080 becomes extremely problematic when you've invented, you know, 165 00:10:13,140 --> 00:10:17,640 AI drones and nuclear bombs and, you know, bio weapons. So, so 166 00:10:17,640 --> 00:10:20,780 that, I think is sort of thinking about it from an 167 00:10:20,780 --> 00:10:24,140 evolutionary point of view, that what Peter is talking about on 168 00:10:24,320 --> 00:10:26,600 the kind of deliberate destruction side, you know, is 169 00:10:26,600 --> 00:10:31,400 that we have an urge to destroy things where our which is way 170 00:10:31,400 --> 00:10:35,360 out of line with our capacity to actually destroy things, like 171 00:10:35,360 --> 00:10:38,360 our urge to destroy things. It's almost like our technology for 172 00:10:38,360 --> 00:10:42,520 destruction is catching up with our urge, and I, and I sort of 173 00:10:42,520 --> 00:10:46,660 think, well, you know, if there was, if everyone had access, 174 00:10:46,660 --> 00:10:48,820 well, okay, so why haven't we destroyed the world? You might 175 00:10:48,820 --> 00:10:51,400 say, Well, why are we, why are we all still? Why is the murder 176 00:10:51,400 --> 00:10:54,100 rate going down? Why is there less war? Which is a good 177 00:10:54,100 --> 00:10:57,700 question. But if, if I said, you know, well, actually, we're 178 00:10:57,700 --> 00:11:00,900 gonna give everyone in the world a button, and if anyone presses 179 00:11:00,900 --> 00:11:05,160 it, everyone dies, right? You know perfectly well, some, some 180 00:11:05,160 --> 00:11:08,760 idiot, would press that button, right? So it sort of feels to me 181 00:11:08,760 --> 00:11:12,780 like, well, we know that someone out there would do it if they 182 00:11:12,780 --> 00:11:17,040 could, just for long, right? So feel so actually, there really 183 00:11:17,040 --> 00:11:21,440 is. It's not like, well, we can rest assured that this won't 184 00:11:21,440 --> 00:11:25,040 happen. All we're saying is, well, it's just at the moment 185 00:11:25,040 --> 00:11:29,060 sufficiently hard for someone to do it. And if it gets to some 186 00:11:29,060 --> 00:11:34,280 level of where a man in a shed can create a global killing 187 00:11:34,340 --> 00:11:38,420 thing and decide to just do it, I think we can all assume 188 00:11:38,420 --> 00:11:42,280 someone would right. So that's the problem, I think, is that, 189 00:11:42,280 --> 00:11:47,020 you know, there is nothing in us that stops it from happening. Is 190 00:11:47,020 --> 00:11:50,620 what I'm saying. We get to the correct level of destructiveness 191 00:11:50,680 --> 00:11:53,740 if it becomes cheap enough to destroy the world, some wanker 192 00:11:53,740 --> 00:11:54,760 is going to do it well. 193 00:11:54,760 --> 00:11:58,180 Chris Wragg: And if you, if you take the scale down in terms of 194 00:11:58,780 --> 00:12:03,900 scale of lethality, if you, if you look at it's a boring topic, 195 00:12:03,900 --> 00:12:09,900 but it's relevant here the idea of gun control that you know, if 196 00:12:09,900 --> 00:12:14,220 you give enough guns to people, some of them are going to go 197 00:12:14,220 --> 00:12:18,000 berserk and go and shoot lots of people. Now, if you scale that 198 00:12:18,000 --> 00:12:22,160 up to, you know, large scale lethality. If everybody had, 199 00:12:22,160 --> 00:12:26,840 like, you say, a nuclear bomb, then we already know the, you 200 00:12:26,840 --> 00:12:30,200 know, going, going into a school and shooting lots of people is 201 00:12:30,200 --> 00:12:34,340 already a pretty, pretty high threshold for for sort of 202 00:12:34,340 --> 00:12:36,080 destructive, destructiveness. 203 00:12:36,080 --> 00:12:38,480 Nick Hare: Or you've got people like, you know, Andreas Lubitz, 204 00:12:38,480 --> 00:12:41,140 who was that German wings pilot who crashed the plane exactly 205 00:12:41,140 --> 00:12:43,720 when you think, well, if he'd have been able to crash a plane 206 00:12:43,720 --> 00:12:46,300 with a million people on it, probably wouldn't have changed 207 00:12:46,300 --> 00:12:49,780 his if anything, he might have been more up, you know. So, 208 00:12:50,260 --> 00:12:51,100 yeah, yeah, 209 00:12:51,100 --> 00:12:54,400 Peter Coghill: I think, I think it doesn't really hinge on it 210 00:12:54,400 --> 00:12:57,340 being deliberate destruction, though. I think it's, I think 211 00:12:57,340 --> 00:13:01,560 that I think there may be a tendency, even though, you know, 212 00:13:01,560 --> 00:13:05,340 accidental disruption. So when we were, when we built the 213 00:13:05,520 --> 00:13:12,120 atomic bomb and then the hydrogen bomb, we we, we weren't 214 00:13:12,180 --> 00:13:15,180 deliberately setting out to try to put ourselves in a position 215 00:13:15,180 --> 00:13:18,600 that we could destroy the world. They were, they were sort of 216 00:13:18,600 --> 00:13:21,080 tactically, at the time, rational things to do. It's 217 00:13:21,080 --> 00:13:25,280 like, this is, this is a world, and at that time, this is a war 218 00:13:25,280 --> 00:13:29,540 ending weapon. So we need to get that, and that will cause Japan 219 00:13:29,540 --> 00:13:33,680 to surrender. Then it became, because the technology then 220 00:13:33,680 --> 00:13:36,380 existed, it became a sort of race to produce more and more 221 00:13:36,380 --> 00:13:38,900 and more than we and they are, over a decade or so, we got to 222 00:13:38,900 --> 00:13:42,880 the point where actually this was a, this was a world ending 223 00:13:42,940 --> 00:13:46,720 weapon. But before that point, it was a, sort of, it was a 224 00:13:46,720 --> 00:13:47,680 rational thing to do. 225 00:13:47,980 --> 00:13:50,500 Nick Hare: Yeah, I'm not saying that's partly the problem. Yeah, 226 00:13:50,560 --> 00:13:56,980 you know is that it gets not, there's not, it's not, in a 227 00:13:56,980 --> 00:14:00,240 sense, it is rational because we do want to kill people. We want 228 00:14:00,240 --> 00:14:02,940 to have the technology to kill people, so it makes sense to 229 00:14:02,940 --> 00:14:05,760 invent it. What I'm saying is that the methods you would use 230 00:14:05,760 --> 00:14:09,180 to stop it happening, the regime that you would have to impose to 231 00:14:09,180 --> 00:14:12,480 stop it from happening, just looks very different if you're 232 00:14:12,480 --> 00:14:16,740 worrying about a deliberate choice to kill people versus 233 00:14:16,800 --> 00:14:20,040 accidentally killing everyone in the world. They're different 234 00:14:20,040 --> 00:14:22,760 kinds of mechanisms. And I think, you know, you can deter 235 00:14:22,760 --> 00:14:25,760 people in a different way, right? You need to impose 236 00:14:25,760 --> 00:14:29,480 different mechanisms, I think, on but, but I'm just saying are 237 00:14:30,260 --> 00:14:32,960 one, one key mechanism of deliberately destroying the 238 00:14:32,960 --> 00:14:35,660 world is that, what we can't deny is we definitely, some 239 00:14:35,660 --> 00:14:36,440 people do want 240 00:14:36,439 --> 00:14:39,139 Peter Coghill: to do that. Yeah, no, I agree. I agree. Yeah. 241 00:14:39,199 --> 00:14:41,559 Fraser McGruer: I think you might be getting close to what I 242 00:14:41,559 --> 00:14:45,699 want to ask, right? Which is, okay, do we contain the seeds of 243 00:14:45,699 --> 00:14:48,579 our own destruction? Right? More or less, that's the question we 244 00:14:48,579 --> 00:14:52,299 want to ask, right? And I can see that it's interesting. But 245 00:14:52,299 --> 00:14:55,359 why are we discussing this? What's the value? What's the 246 00:14:55,539 --> 00:14:56,919 what's Well, 247 00:14:56,920 --> 00:14:59,140 Nick Hare: what do we do about it? Right? Because, because the 248 00:14:59,140 --> 00:15:01,800 problem is that. That destructive technology is going 249 00:15:01,800 --> 00:15:05,580 to get worse, and I think on a large scale, obviously, well, 250 00:15:05,580 --> 00:15:09,120 nuclear weapons technology haven't, hasn't really. It's 251 00:15:09,120 --> 00:15:12,240 certainly not become super cheap. And in fact, you know, 252 00:15:12,420 --> 00:15:15,060 you look at the data, and there's fewer nuclear weapons 253 00:15:15,060 --> 00:15:19,500 around now than there were, you know, 30 years ago, killing 254 00:15:19,500 --> 00:15:22,220 people has become easier. We've got murder drones, and yet 255 00:15:22,280 --> 00:15:26,180 murder rates and rates of deaths in war have fallen right, see? 256 00:15:26,180 --> 00:15:29,480 So what are we getting right? Is the is the question, because we 257 00:15:29,480 --> 00:15:32,960 want to be doing more of that like and I think what we haven't 258 00:15:33,140 --> 00:15:36,740 we've only really just started seeing what the impact of, say, 259 00:15:36,740 --> 00:15:40,720 drones are, but actually being able to kill a million people 260 00:15:40,720 --> 00:15:44,140 with drones might turn out to be much easier and cheaper than 261 00:15:44,380 --> 00:15:47,080 killing a million people with a nuclear bomb, and more efficient 262 00:15:47,140 --> 00:15:50,080 and more efficient. And in fact, you can, just then you can, you 263 00:15:50,080 --> 00:15:51,700 get the whole city, whereas 264 00:15:51,820 --> 00:15:54,520 Fraser McGruer: it could be morally superior. Well, 265 00:15:54,640 --> 00:15:57,220 Nick Hare: we're not even worrying about that. The point 266 00:15:57,340 --> 00:16:01,260 is, you know, it's a new it's the latest thing to have to 267 00:16:01,260 --> 00:16:05,400 worry about, what do we do? What did we get right with nuclear 268 00:16:05,400 --> 00:16:09,240 weapons or with any other technology, which means that we, 269 00:16:09,420 --> 00:16:12,360 you know, actually all of the trends in violence are positive, 270 00:16:12,360 --> 00:16:16,140 like despite the fact that destruction is easier and 271 00:16:16,140 --> 00:16:20,420 cheaper than it was in the past. What have we got right that 272 00:16:20,420 --> 00:16:23,480 makes it that's that has been. Is that a question 273 00:16:23,480 --> 00:16:26,780 Fraser McGruer: you want to answer right now? Yeah. 274 00:16:26,839 --> 00:16:28,999 Nick Hare: Well, yeah. Why not? Are you proposing this as a 275 00:16:28,999 --> 00:16:32,419 question? Well, I'm just at what we I think that's why you said, 276 00:16:32,419 --> 00:16:34,879 Why are we asking the question? And the answer is, because 277 00:16:34,879 --> 00:16:37,099 we've, I don't know about you phrase, but I don't want 278 00:16:37,099 --> 00:16:40,039 everyone in the world to be killed. So the question is, what 279 00:16:40,039 --> 00:16:41,139 should we do about 280 00:16:41,800 --> 00:16:44,140 Peter Coghill: it? Just select people, yeah, 281 00:16:47,440 --> 00:16:50,260 Fraser McGruer: yeah, okay, where do we go with this? Who's 282 00:16:50,260 --> 00:16:50,560 next? 283 00:16:50,560 --> 00:16:53,980 Peter Coghill: I mean, I before we do that. I mean, I thought 284 00:16:53,980 --> 00:16:57,580 occurred to me, I came to sort of extend this a little bit 285 00:16:57,580 --> 00:17:03,480 further, that perhaps our innate capacity and our innate drive to 286 00:17:03,660 --> 00:17:06,840 produce ever more dangerous technologies is actually a 287 00:17:06,840 --> 00:17:11,100 potential solution to the Fermi Paradox. I think maybe that. 288 00:17:11,100 --> 00:17:15,360 Maybe it's the case that any sufficiently complex 289 00:17:15,360 --> 00:17:21,320 intelligence that evolves must necessarily have curiosity, risk 290 00:17:21,320 --> 00:17:25,160 taking and competitiveness that causes it to be dangerous and 291 00:17:25,160 --> 00:17:28,100 cause it to eventually, there's a sort of an technological 292 00:17:28,820 --> 00:17:32,060 determinism means that you eventually strike upon atomic 293 00:17:32,060 --> 00:17:36,140 weapons and biochemistry, bio weapons and things, and that 294 00:17:36,140 --> 00:17:39,800 will lead event, on average, will lead, lead to extinction Of 295 00:17:39,800 --> 00:17:42,520 the species. Yeah. So I think there's 296 00:17:42,520 --> 00:17:44,860 Nick Hare: a tragedy of the commons kind of argument that, 297 00:17:44,920 --> 00:17:48,700 yeah, we just, you know, the possibility of cooperation is 298 00:17:48,700 --> 00:17:53,020 just too hard and unsustainable, and affection is always going to 299 00:17:53,020 --> 00:17:54,520 be better. And hey, Preston's 300 00:17:55,600 --> 00:17:59,500 Peter Coghill: dilemma, because, yeah, that's but it's hard for 301 00:17:59,500 --> 00:18:05,580 the innate kind, the kind of the kind of intelligent intelligence 302 00:18:05,580 --> 00:18:09,840 that will win out in a bio, sort of crude biological evolutionary 303 00:18:09,840 --> 00:18:13,920 world, if you could prod it differently, and you could have 304 00:18:14,100 --> 00:18:16,860 more peaceful beings because they had guns on their arms, 305 00:18:17,580 --> 00:18:21,200 then maybe they find cooperation and coordination much easier, 306 00:18:21,380 --> 00:18:24,020 but it's just that in the biological world, because of the 307 00:18:24,020 --> 00:18:28,100 limited amount of damage you can do person on person, that just 308 00:18:28,100 --> 00:18:31,040 doesn't scale very well as you get more and more complex. 309 00:18:31,040 --> 00:18:33,740 Nick Hare: Yeah, and there's also the problem that humans are 310 00:18:35,300 --> 00:18:38,540 quite like we have a small number of children, which 311 00:18:38,540 --> 00:18:42,760 doesn't sound like terribly like relevant, but actually it means 312 00:18:42,760 --> 00:18:47,380 that we're not we are incentivized to cooperate 313 00:18:47,740 --> 00:18:50,980 intrinsically with quite a small number of people. Yeah, so we're 314 00:18:50,980 --> 00:18:53,740 only really related to a very small number of people, and 315 00:18:53,740 --> 00:18:56,860 hence we only really care about and do things altruistically for 316 00:18:56,860 --> 00:19:00,540 a small number of people, because, you know, we have got a 317 00:19:00,540 --> 00:19:03,540 lot of investment in our own offspring. Yeah, whereas, if you 318 00:19:03,540 --> 00:19:09,540 take bees, they've all got one mum, all bees have got one mum, 319 00:19:09,540 --> 00:19:12,480 and they're all sisters, except for the drones, who, it doesn't 320 00:19:12,480 --> 00:19:14,880 matter. They just their only job is to mate with the queen and 321 00:19:14,880 --> 00:19:18,420 then die, right? So a whole beehive can all work together 322 00:19:18,420 --> 00:19:20,520 individually, because they're all essentially the same. 323 00:19:20,640 --> 00:19:25,040 They're all invested in the same set of genes. So option number 324 00:19:25,040 --> 00:19:30,560 one is we just have one we like Queen Camilla, or whatever she 325 00:19:30,620 --> 00:19:34,940 is, like a random having all the children. And, you know, we 326 00:19:34,940 --> 00:19:38,300 genetically engineer things so that we're all just, you know, 327 00:19:38,300 --> 00:19:44,320 we're basically like bees. You know. Well, I mean, you know. 328 00:19:44,500 --> 00:19:49,600 But if, if we have enough bio technology to engineer, you 329 00:19:49,600 --> 00:19:54,520 know, to engineer bio weapons, yeah, I mean, I think we could 330 00:19:54,580 --> 00:19:55,840 re engineer human. 331 00:19:55,960 --> 00:19:57,820 Peter Coghill: I mean, I think, I think there are ways doing it 332 00:19:57,820 --> 00:20:00,720 without, directly bio without turning into. Please. Yeah, and 333 00:20:00,720 --> 00:20:04,440 I think, I think the counter argument to the the fact that 334 00:20:04,500 --> 00:20:08,220 intelligence is is the problem, but intelligence could also be 335 00:20:08,220 --> 00:20:12,540 the antidote to the problem, because we are capable of moral 336 00:20:12,540 --> 00:20:16,740 reasoning and an empathy and things we are capable of it. So 337 00:20:16,740 --> 00:20:19,080 maybe what we need to do is train ourselves to be better at 338 00:20:19,080 --> 00:20:24,080 doing those things. So and so we know. So we know we can move 339 00:20:24,080 --> 00:20:28,040 away from the constraints of the biological intelligence we have. 340 00:20:28,160 --> 00:20:31,040 Maybe we, you know, maybe when we start uploading our brains to 341 00:20:31,040 --> 00:20:35,300 computers, or maybe just drilling hard and home, the 342 00:20:35,420 --> 00:20:40,840 moral reasoning is a thing you must do from an earlier age. You 343 00:20:40,840 --> 00:20:42,880 know, we were then capable of sort of, sort of meta 344 00:20:42,880 --> 00:20:46,420 revolution, that we're then able to drive our evolution in a 345 00:20:46,420 --> 00:20:48,400 cognitive space, rather than in a you're 346 00:20:48,700 --> 00:20:50,680 Nick Hare: saying, why the solution is, why can't we all 347 00:20:50,680 --> 00:20:54,460 just get along? Why Can't We? Well, actually, I mean, the 348 00:20:54,460 --> 00:20:57,580 thing is that, Stephen, I mean, Steven Pinker's argument in the 349 00:20:57,580 --> 00:21:01,560 better angels of our nature is that, you know, because the fact 350 00:21:01,560 --> 00:21:04,800 that the trends in violence and all sorts of other indicators 351 00:21:04,800 --> 00:21:10,020 are all what we would regard as positive, suggests that our kind 352 00:21:10,020 --> 00:21:14,160 of political and social technology is outpacing our 353 00:21:14,160 --> 00:21:16,920 destructive technology development. And his his view is 354 00:21:16,920 --> 00:21:20,120 that, you know, our ability to have a say, a strong state, to 355 00:21:20,120 --> 00:21:23,120 punish people, to reliably detect and punish wrongdoers, 356 00:21:23,780 --> 00:21:27,140 has has actually outpaced, you know, the ability of those 357 00:21:27,140 --> 00:21:32,000 wrongdoers to cause death and mayhem. And I find that quite 358 00:21:32,000 --> 00:21:34,880 plausible. And in fact, you know, if you think about, well, 359 00:21:34,880 --> 00:21:37,280 what technology does that rely on? It relies on a certain 360 00:21:37,280 --> 00:21:41,080 amount of ability to, you know, centralise power, but also to be 361 00:21:41,080 --> 00:21:43,900 able to detect things, to be able to identify that this 362 00:21:43,900 --> 00:21:46,720 person did that wrong thing, or that they're going to, you know, 363 00:21:46,720 --> 00:21:49,540 that we kind of detect that someone's acquired the 364 00:21:49,540 --> 00:21:50,920 precursors to a bio weapon 365 00:21:50,920 --> 00:21:53,440 Peter Coghill: we've seen We've got a like our social 366 00:21:53,440 --> 00:21:56,260 technologies, our institutions and things we have now are way 367 00:21:56,260 --> 00:22:00,840 more complicated and way more capable than our little You 368 00:22:00,840 --> 00:22:04,740 know, when we were people living on the planes, travelling around 369 00:22:04,740 --> 00:22:07,860 for scraps and bonking the heads of each other with mallets and 370 00:22:07,860 --> 00:22:10,680 things, we didn't need anything like that. It's amazing that we 371 00:22:10,680 --> 00:22:15,660 have got big state institutions and international treaties and 372 00:22:15,660 --> 00:22:16,080 things. 373 00:22:16,620 --> 00:22:22,280 Chris Wragg: Yeah, I just, I just worry that the this idea 374 00:22:22,640 --> 00:22:30,020 that the the instinct to be peaceable and survive alongside 375 00:22:30,020 --> 00:22:35,120 one another, is going quicker than our ability to invent 376 00:22:35,120 --> 00:22:39,560 things to kill one another, and our drive to kill each other At 377 00:22:39,560 --> 00:22:42,280 some point, like with, for example, mutually assured 378 00:22:42,280 --> 00:22:48,640 destruction, you only get one go, right? So, so, like guns 379 00:22:48,640 --> 00:22:52,420 scale, you know? And it's like, okay, well, back in the day, we 380 00:22:52,420 --> 00:22:56,260 used to have lots of fist fights. Now we have far fewer 381 00:22:56,260 --> 00:23:00,240 gun fights, right? That works. But when you get to the stage of 382 00:23:00,240 --> 00:23:05,160 cataclysmic risk. You only, you only need to get it. What wrong? 383 00:23:05,160 --> 00:23:09,540 Once, right? So how, how can you outpace that? You can't. 384 00:23:09,540 --> 00:23:12,120 Nick Hare: But then, if you look at what we actually did, it was 385 00:23:12,120 --> 00:23:14,760 things like, you know, you'd make these commitments, but then 386 00:23:14,760 --> 00:23:17,160 they'd be monitored. You know that you would actually count is 387 00:23:17,160 --> 00:23:20,600 this person you know, either have they decommissioned these 388 00:23:20,600 --> 00:23:22,940 weapons, and then you know that you're, you're monitoring 389 00:23:22,940 --> 00:23:25,040 where's the nuclear material, and there's kind of 390 00:23:25,040 --> 00:23:28,460 sophisticated global system of trying to make sure people 391 00:23:28,460 --> 00:23:31,700 aren't doing that. And you could imagine doing a similar thing 392 00:23:31,700 --> 00:23:34,520 for, you know, for kind of AI drones, you could say, well, you 393 00:23:34,520 --> 00:23:37,340 know, we're gonna have a limit on the number of of drones that 394 00:23:37,340 --> 00:23:39,740 each country is allowed to have. And it feels a long way away 395 00:23:39,800 --> 00:23:41,920 that we could imagine that happening now, but I think all 396 00:23:41,920 --> 00:23:45,160 it would take is the equivalent of another Hiroshima or Nagasaki 397 00:23:45,340 --> 00:23:48,700 committed by drones to make people go or we should do 398 00:23:48,700 --> 00:23:49,480 something right. 399 00:23:49,780 --> 00:23:52,180 Peter Coghill: My concern with the the social technologies 400 00:23:52,180 --> 00:23:55,780 we've got for controlling it is they feel fragile. They feel 401 00:23:55,780 --> 00:24:00,840 much more fragile than our drive to find ever more unpleasant 402 00:24:00,840 --> 00:24:01,740 ways to kill each other. 403 00:24:01,739 --> 00:24:04,379 Fraser McGruer: Go and see what you're going to say. Well, I was 404 00:24:04,379 --> 00:24:07,379 Chris Wragg: just going to say, thus far we've only, we've only 405 00:24:07,379 --> 00:24:12,299 touched on deliberate, yeah. Well, I was all. I was just 406 00:24:12,299 --> 00:24:16,199 going to bridge into that by saying, well, well, I I'm not 407 00:24:16,199 --> 00:24:20,179 quite, I'm not so sure it's totally binary, like that, that 408 00:24:20,179 --> 00:24:23,419 there are things we do to kill one another, and there are 409 00:24:23,419 --> 00:24:27,739 things we do by accident. I sort of feel like, although nuclear 410 00:24:27,739 --> 00:24:30,799 bombs, obviously, you know, have a purpose, and guns have a 411 00:24:30,799 --> 00:24:36,319 purpose, something like a knife is multi purpose. And you know, 412 00:24:36,319 --> 00:24:41,379 if I throw a knife at you in order to wound you, but it hits 413 00:24:41,379 --> 00:24:43,779 you in the heart and I kill you. I go, Oh, dear, you know, oops. 414 00:24:43,779 --> 00:24:46,239 That was a that was manslaughter, not murder or 415 00:24:46,239 --> 00:24:50,259 something. So I'm not quite so sure that we've got this really 416 00:24:50,259 --> 00:24:55,239 clear distinction. So AI being an example, you might generate 417 00:24:55,239 --> 00:25:02,339 AI and use it in a variety of sort. Circumstances. Some of 418 00:25:02,339 --> 00:25:05,579 those might be lethal. Some of those might not be lethal. Some 419 00:25:05,579 --> 00:25:09,419 of them might be sort of intended to be, like policing 420 00:25:09,719 --> 00:25:15,419 robots or something that suddenly become, you know, fully 421 00:25:15,419 --> 00:25:19,859 lethal killing machines. That it's isn't the distinction 422 00:25:19,859 --> 00:25:23,419 between what we accidentally do to kill one another and what we 423 00:25:23,419 --> 00:25:27,739 deliberately neglect, I suppose, yeah, is not quite so clear cut, 424 00:25:27,799 --> 00:25:30,139 yeah, although, I mean, I think there's a different set of 425 00:25:30,139 --> 00:25:33,199 things you would want to do. Let's imagine that we could 426 00:25:33,199 --> 00:25:36,739 solve we were absolutely guaranteed that nobody would 427 00:25:36,739 --> 00:25:39,859 want to destroy the world. So we'll put that on a shelf and 428 00:25:39,859 --> 00:25:42,519 pretend that we've solved the problem of people wanting to 429 00:25:42,519 --> 00:25:45,699 destroy the world. Well, then we have the next category of thing 430 00:25:45,699 --> 00:25:49,539 to worry about, which is, what about accidentally doing it? 431 00:25:49,539 --> 00:25:51,279 Yeah. How do we stop that? 432 00:25:51,940 --> 00:25:55,060 Fraser McGruer: Okay, so that's we're going to move on to next, 433 00:25:55,060 --> 00:25:58,180 right? Am I right? Like, how do we stop this stuff accidentally 434 00:25:58,180 --> 00:25:58,600 happening? 435 00:25:58,600 --> 00:26:02,100 Peter Coghill: Yeah, before we how do we foresee consequences 436 00:26:02,100 --> 00:26:03,960 that we didn't foresee, which kind 437 00:26:03,960 --> 00:26:05,820 Fraser McGruer: of leads, which sort of connects to something 438 00:26:05,820 --> 00:26:07,980 I've been thinking about and I've not been able to stop 439 00:26:07,980 --> 00:26:10,680 thinking about the last 10 minutes, right? Which I can't 440 00:26:10,680 --> 00:26:13,860 get off the bees, right? Bees, yeah. And the reason why is, 441 00:26:13,860 --> 00:26:16,740 let's say they've got this system which works really well, 442 00:26:16,920 --> 00:26:19,800 and it means they don't do sort of mad things like kill them, 443 00:26:19,800 --> 00:26:22,280 keep sharks. They're all related, but I feel really bad 444 00:26:22,280 --> 00:26:25,640 for them, because they live on a planet with a bunch of people 445 00:26:25,880 --> 00:26:29,780 who, who they're going to get killed because of us, right? And 446 00:26:29,780 --> 00:26:32,180 it's something outside of their control, please? Yeah, the bees 447 00:26:32,180 --> 00:26:34,760 have figured it all out, but then these sort of highly 448 00:26:34,760 --> 00:26:37,940 evolved monkeys that have come along and then, and then we, we 449 00:26:37,940 --> 00:26:40,220 destroy the plant, we destroy the bees along with us. And I 450 00:26:40,220 --> 00:26:42,744 just feel bad because, you know, the bees have figured it out, 451 00:26:43,044 --> 00:26:45,384 and, you know, they get sort of done in. 452 00:26:45,384 --> 00:26:48,084 Peter Coghill: There's a great deal more moral weight than just 453 00:26:48,084 --> 00:26:53,604 our own species. We have a sort of moral responsibility for the 454 00:26:53,604 --> 00:26:56,604 rest of the world, and not only the rest of the world, but all 455 00:26:56,724 --> 00:27:00,144 the potential species that might come about after we've gone, 456 00:27:00,204 --> 00:27:04,320 yeah. So we can't, we mustn't snuff out life completely. We 457 00:27:04,320 --> 00:27:05,700 mustn't, yes, no, 458 00:27:08,700 --> 00:27:12,840 Unknown: fish, fish, yeah, yes. And there are all the other 459 00:27:13,560 --> 00:27:15,180 Peter Coghill: we have to respect our piece to peace 460 00:27:15,180 --> 00:27:17,460 accord with the fish that we've settled upon podcast. 461 00:27:17,460 --> 00:27:20,040 Fraser McGruer: And we did do that, yeah, but it also, I'm 462 00:27:20,040 --> 00:27:20,660 sure, that 463 00:27:20,660 --> 00:27:23,900 Nick Hare: this fragile cease fire, Twix Twixt, land and sea 464 00:27:23,900 --> 00:27:25,460 Exactly, exactly. 465 00:27:25,460 --> 00:27:27,740 Fraser McGruer: And I'm sure in our sort of discussions, you 466 00:27:27,740 --> 00:27:31,220 know, logically, external for this just fits into really, you 467 00:27:31,220 --> 00:27:33,980 know, stuff, external forces over which have no control. 468 00:27:34,100 --> 00:27:38,060 That's what we are to the bees, right? Anyway, I've diverted us. 469 00:27:38,120 --> 00:27:40,460 Nick, oh no, so was it you? Peter, who was going to come in 470 00:27:40,460 --> 00:27:42,080 Peter Coghill: with some Yeah. So I think it boils down to 471 00:27:42,080 --> 00:27:44,560 like, it's unforced. It's unforeseen, so unforeseen 472 00:27:44,920 --> 00:27:49,000 consequences that we want to be able to foresee, isn't it? So 473 00:27:49,000 --> 00:27:51,580 when we create a new technology, everything's, oh, great, 474 00:27:51,580 --> 00:27:54,160 brilliant. We've got, we've got, we've got nuclear power. 475 00:27:54,160 --> 00:27:56,860 Brilliant, this is we're getting into unknown, unknowns. This is 476 00:27:56,860 --> 00:27:59,800 what we're talking about. This sort of is a bit, yeah, gone. 477 00:27:59,800 --> 00:28:02,580 But it's like we go because we make a new technology, but it's 478 00:28:02,580 --> 00:28:06,420 only, Only later did, but does it become apparent, perhaps, 479 00:28:06,420 --> 00:28:09,060 through, you know, nuclear weapons, they were okay, not 480 00:28:09,060 --> 00:28:12,060 great things, but they were okay, if, as long as they were 481 00:28:12,060 --> 00:28:15,120 only a very small number in the world, they were quite useful as 482 00:28:15,120 --> 00:28:19,320 a tactical weapon of war. But when you get several, when you 483 00:28:19,320 --> 00:28:22,940 get so many that you can destroy everything on the earth several 484 00:28:22,940 --> 00:28:26,900 times over, and not only that, but extinguish any hope of life 485 00:28:26,900 --> 00:28:34,580 regaining a foothold. That the case is the same for AI, mirror 486 00:28:34,580 --> 00:28:39,020 life and various other existential risks, that that's 487 00:28:39,020 --> 00:28:42,220 that that's that's a problem. But we didn't at the time 488 00:28:43,120 --> 00:28:46,180 foresee that that was a problem. It's only when we started to 489 00:28:46,180 --> 00:28:49,660 scale up did people go Hold on this situation now means that 490 00:28:49,660 --> 00:28:50,200 this could occur. 491 00:28:50,260 --> 00:28:53,200 Nick Hare: Yeah, and I think this is by analogy with, you 492 00:28:53,200 --> 00:28:56,920 know, our aggression levels being misaligned to our 493 00:28:56,920 --> 00:28:59,920 technology, our capability to destroy things. I think our 494 00:28:59,920 --> 00:29:04,500 levels of scare, of fear. What we're scared of is also, you 495 00:29:04,500 --> 00:29:08,100 know, misaligned to how scary things actually are. You know, 496 00:29:08,100 --> 00:29:11,460 because we've really only evolved to be scared of snakes 497 00:29:11,460 --> 00:29:14,940 and things. Yeah, we haven't really, yeah, we haven't really, 498 00:29:14,940 --> 00:29:18,900 we haven't really evolved the means the correct level of sort 499 00:29:18,900 --> 00:29:26,120 of scaredness of things like, you know, cars than we are, are 500 00:29:26,240 --> 00:29:27,800 much less scared of snakes. 501 00:29:27,920 --> 00:29:32,060 Chris Wragg: Yeah, although they do kill up to 75,000 people per 502 00:29:32,060 --> 00:29:35,480 Nick Hare: year, snakes still not as many as hippos or as 503 00:29:35,480 --> 00:29:39,680 cars, as many as cars, more than hippos. Wow. Anyway, seems like 504 00:29:39,680 --> 00:29:42,100 we should all be really scared of snakes all of a sudden, but, 505 00:29:43,780 --> 00:29:46,660 but the point is that, you know, and particularly when it comes 506 00:29:46,660 --> 00:29:50,560 to these sort of abstract risks, like, you know, the kind of AI 507 00:29:51,160 --> 00:29:56,740 takeover problem, which you have to explain to someone. You've 508 00:29:56,740 --> 00:29:59,440 got to sit down and tell them a story about how this could 509 00:29:59,440 --> 00:30:02,760 happen. And then at the end, they'll go, yeah, that's not 510 00:30:02,760 --> 00:30:07,020 gonna happen. Come on, you know? And I think at least nuclear 511 00:30:07,020 --> 00:30:10,740 bombs come free with an enormous, terrifying fireball. 512 00:30:11,280 --> 00:30:15,600 But AI takeover doesn't really look like anything scary until 513 00:30:15,600 --> 00:30:19,740 it's way too late. And so, you know, it is those kinds of 514 00:30:19,740 --> 00:30:22,400 risks. And same with mirror life. It's like, well, nothing 515 00:30:22,580 --> 00:30:27,380 looks that scary to us. It's a bacterium. We're not 516 00:30:27,380 --> 00:30:30,500 intrinsically scared. I mean, nuclear weapons are 517 00:30:30,500 --> 00:30:33,740 intrinsically scary, but a lot of these things that could just 518 00:30:33,740 --> 00:30:37,280 kill us all just they're picture in a magazine, aren't they? 519 00:30:37,280 --> 00:30:40,240 Chris Wragg: Yeah. I mean, I think it's quite interesting the 520 00:30:40,240 --> 00:30:47,020 extent to which awareness in the public consciousness is a is a 521 00:30:47,020 --> 00:30:52,540 deterrent to us developing technologies. So, so you know, 522 00:30:52,540 --> 00:30:55,960 if you, if you talk to most people about the threat of AI, 523 00:30:55,960 --> 00:31:00,160 they think of Skynet, right and terminators to get scary robots, 524 00:31:00,160 --> 00:31:04,620 right? Exactly, but, but, or how right, you know, but, but 525 00:31:04,620 --> 00:31:09,120 psychotic robots. Psychotic robots. But the the role of 526 00:31:09,900 --> 00:31:15,240 science fiction, dystopian post apocalyptic fiction is to, is to 527 00:31:15,240 --> 00:31:20,840 basically tell that story in a way that sticks in in people's 528 00:31:20,840 --> 00:31:24,380 minds. If, and if you look at there, I would say there is a 529 00:31:24,500 --> 00:31:28,340 large portion of popular resistance to the development of 530 00:31:28,340 --> 00:31:32,900 AI that is based on this idea of killer weapon systems and, you 531 00:31:32,900 --> 00:31:36,260 know, Robocop and all that kind of stuff. But if you always so 532 00:31:36,440 --> 00:31:42,100 mirror life, like very few, it's not in anybody's public 533 00:31:42,100 --> 00:31:44,740 consciousness, right? Nobody's written a novel about it. Hasn't 534 00:31:44,740 --> 00:31:48,460 been a blockbuster film about it. But if you look at something 535 00:31:48,460 --> 00:31:54,280 like cloning, human cloning, and the the way, quite quickly, 536 00:31:54,280 --> 00:31:57,340 people were weirded out and thought, you know, that should 537 00:31:57,340 --> 00:31:59,920 be a sight, Whoa, don't. Don't go there. That's there's 538 00:31:59,920 --> 00:32:03,480 something partly, it's in the public consciousness. Partly, 539 00:32:03,480 --> 00:32:09,180 it's like, you say, inherently creepy. Creepy, right, exactly. 540 00:32:09,180 --> 00:32:11,580 Fraser McGruer: And pregnancy, too. It feels like a couple of 541 00:32:11,580 --> 00:32:14,340 things I want to say we need to draw to a conclusion soon, but 542 00:32:14,340 --> 00:32:16,500 it feels like we're talking about something we've talked 543 00:32:16,500 --> 00:32:18,900 about before, which is perception of risk, maybe, which 544 00:32:18,900 --> 00:32:22,640 is, you know, I'm probably more scared of sharks and guns. Four 545 00:32:22,640 --> 00:32:26,480 people a year, really killed by sharks, exactly. That's 546 00:32:26,480 --> 00:32:30,080 terrifying, right? Yeah, and I should probably be more scared 547 00:32:30,080 --> 00:32:34,280 of of of cheeseburgers, right? That's actually more people die 548 00:32:34,400 --> 00:32:37,280 of choking on a cheeseburger. Well, I hadn't even thought of 549 00:32:37,280 --> 00:32:38,540 cheese, but I was just thinking 550 00:32:39,500 --> 00:32:42,640 Chris Wragg: of sharks die from choking on humans. 551 00:32:43,300 --> 00:32:45,400 Fraser McGruer: Yeah, and no one talks about that. I'm glad you 552 00:32:45,400 --> 00:32:48,160 brought that up. But the second thing, it's all, it's all very 553 00:32:48,160 --> 00:32:51,520 human centric and Earth centric, you know, because us worrying 554 00:32:51,520 --> 00:32:54,100 about, you know, the destruction of our planet, Earth and so on. 555 00:32:54,460 --> 00:32:56,980 But the universe doesn't, kind of doesn't really care, right? 556 00:32:57,160 --> 00:33:00,480 And it's, I don't know, it just occurs to me, you know, the 557 00:33:00,840 --> 00:33:02,760 Nick Hare: only the kind of nihilist who pressed the red 558 00:33:02,760 --> 00:33:07,200 button, yeah, yeah, universe out. There's probably some other 559 00:33:07,200 --> 00:33:07,980 aliens. They'll be all 560 00:33:07,980 --> 00:33:10,020 Fraser McGruer: right, yeah. Now, to be fair, I get it. It's 561 00:33:10,020 --> 00:33:12,420 what we've got. All we've got is ourselves, our lives, our 562 00:33:12,420 --> 00:33:16,500 planet, etc. But, you know, it's sort of, ultimately, it's true, 563 00:33:16,500 --> 00:33:18,840 though, isn't it? Depressingly, it kind of doesn't really 564 00:33:18,840 --> 00:33:21,000 matter. But then, yes, well, if that doesn't matter, then 565 00:33:21,000 --> 00:33:23,300 nothing does. Exactly, that's why I'm an honest why I'm 566 00:33:23,720 --> 00:33:28,940 honest. So So look, we do need to finish off. So I go to Peter 567 00:33:28,940 --> 00:33:31,280 first, and then, you know, we need to wrap this up. 568 00:33:31,280 --> 00:33:34,160 Peter Coghill: Yeah, so just picking up on Chris's point that 569 00:33:34,640 --> 00:33:37,760 considering exercise at risk is not in the public sort of 570 00:33:38,420 --> 00:33:41,080 discourse. I don't think it's nearly enough. It needs to be 571 00:33:41,080 --> 00:33:43,120 much more in the public discourse. The fact Fraser 572 00:33:43,120 --> 00:33:47,020 hadn't heard of mirror life is damning indictment of Fraser 573 00:33:47,020 --> 00:33:50,680 sort of awareness of important, important things in the world. 574 00:33:50,740 --> 00:33:54,400 Nick Hare: So the fact, luckily, he's got absolutely no power to 575 00:33:54,400 --> 00:33:54,580 do 576 00:33:55,540 --> 00:33:56,800 Fraser McGruer: anything if a 577 00:33:58,540 --> 00:33:59,620 Unknown: mirror Fraser was 578 00:34:00,580 --> 00:34:04,320 Peter Coghill: no one person does, but as your population 579 00:34:04,860 --> 00:34:08,280 gets more aware of this as a problem, then things start to 580 00:34:08,280 --> 00:34:11,820 change. So like the the fact that everyone now worries about 581 00:34:11,820 --> 00:34:15,720 recycling and global warming is starting to shift the perception 582 00:34:15,720 --> 00:34:19,200 and the and the expectation on our political leaders. So the 583 00:34:19,560 --> 00:34:22,700 fact that it's an existential risk. Basically means, if it 584 00:34:22,700 --> 00:34:28,160 come, if it happens, the cost is infinite. The cost of that 585 00:34:28,160 --> 00:34:33,080 catastrophe is infinite, which means, if the probability is at 586 00:34:33,080 --> 00:34:37,460 all, non zero, which is because it could happen, then it must, 587 00:34:37,460 --> 00:34:41,560 then it should, dominate all other discussions, 588 00:34:41,800 --> 00:34:42,760 Unknown: Pascal's Wager 589 00:34:42,759 --> 00:34:45,579 Peter Coghill: style, yeah, so it's like the fam, whereas now, 590 00:34:45,579 --> 00:34:48,519 yeah, we still, we still worry about who's going to be the next 591 00:34:49,900 --> 00:34:52,240 Nick Hare: trivial, actually infinite, because we're not 592 00:34:52,240 --> 00:34:55,240 going to devote an infinite amount of resources to stopping 593 00:34:55,240 --> 00:34:59,500 it. So, you know, it has to be bounded. The existence of life 594 00:34:59,500 --> 00:35:02,880 still has. A bounded value. We're not really agreed on what 595 00:35:02,880 --> 00:35:05,520 that is. Well, nothing, according to Fraser, yeah, zero, 596 00:35:05,520 --> 00:35:08,460 according to him, but and, but somewhere between zero and 597 00:35:08,460 --> 00:35:13,320 infinity. So we pinned it down quite nicely. Yeah? 598 00:35:13,680 --> 00:35:16,800 Fraser McGruer: Okay, that feels like a very satisfying, nice 599 00:35:16,800 --> 00:35:20,600 conclusion there. Yeah, nailed it. Okay, look, I got a 600 00:35:20,600 --> 00:35:23,420 question, yeah, let's do it from fiction. Probably doesn't have 601 00:35:23,420 --> 00:35:27,080 to be necessarily favourite existential risk, or, I suppose, 602 00:35:27,140 --> 00:35:28,640 favourite apocalypse. 603 00:35:29,720 --> 00:35:33,320 Chris Wragg: Yeah, I guess I'll go with Douglas Adams and the 604 00:35:33,320 --> 00:35:38,120 idea that we're a sort of traffic, or rather, we're 605 00:35:38,120 --> 00:35:41,920 roadworks being being performed to clear, clear a path through 606 00:35:41,920 --> 00:35:43,060 the through the universe? 607 00:35:43,240 --> 00:35:45,280 Fraser McGruer: Yes, so you're picking up on my that's like, on 608 00:35:45,280 --> 00:35:46,960 the sort of spectrum of where I am, 609 00:35:46,959 --> 00:35:51,939 Chris Wragg: I like, I like the idea that we sit here and we 610 00:35:52,959 --> 00:35:57,999 consider all of these serious type risks, and that in the end, 611 00:35:58,179 --> 00:36:02,819 we might be wiped out by something very trivial and 612 00:36:02,819 --> 00:36:03,299 amusing. 613 00:36:03,300 --> 00:36:05,880 Fraser McGruer: Maybe it's embarrassing. Maybe, yeah, like 614 00:36:05,880 --> 00:36:08,280 road works through our part of the solar system, galaxy, or 615 00:36:08,280 --> 00:36:12,360 someone forgetting not firing all the telephone cleaners or 616 00:36:12,360 --> 00:36:14,340 something like that, right? Good one. 617 00:36:14,340 --> 00:36:16,740 Peter Coghill: Chris, I just realised how, how normally try 618 00:36:16,740 --> 00:36:19,500 to keep it light at the end. This is not the likeness of 619 00:36:19,560 --> 00:36:20,960 Fraser McGruer: question, be fair. It's a pretty heavy 620 00:36:20,960 --> 00:36:23,000 subject to start with, but yeah, 621 00:36:23,540 --> 00:36:26,180 Peter Coghill: my favourite? Well, my not my favourite, not 622 00:36:26,180 --> 00:36:28,520 the one I worry the most about, either one that occupies the 623 00:36:28,520 --> 00:36:31,520 most time in my head, but the one I kind of think is 624 00:36:32,120 --> 00:36:36,980 potentially quite lols. Is the way if humanity were to evolve 625 00:36:36,980 --> 00:36:41,200 into a less brainy, more fertile version of ourselves, kind of 626 00:36:41,200 --> 00:36:41,860 imagine the 627 00:36:42,760 --> 00:36:44,200 Unknown: Idiocracy, yeah. 628 00:36:45,340 --> 00:36:50,560 Peter Coghill: Movie, yeah. And, or the or there is kind of like, 629 00:36:50,620 --> 00:36:54,100 wall e the movie. The Wall e the movie where everything's every 630 00:36:54,220 --> 00:36:57,040 ever we've advanced our technologies, where all of our 631 00:36:57,040 --> 00:36:59,980 requirements are kind of taken care of automatically for us, 632 00:37:00,160 --> 00:37:02,640 which means we can just sit around watching, yeah, 633 00:37:02,640 --> 00:37:04,680 Nick Hare: that's like the kind of Brave New World type of 634 00:37:04,680 --> 00:37:07,320 dystopia where everyone's actually really happy, but it's 635 00:37:07,320 --> 00:37:08,100 dystopian, 636 00:37:10,260 --> 00:37:14,520 Peter Coghill: sometimes called a dysgenic situation. But yeah, 637 00:37:14,520 --> 00:37:18,900 we're but, and maybe AI will will facilitate Yeah, for us, we 638 00:37:18,900 --> 00:37:20,420 can just sit around eating and 639 00:37:20,420 --> 00:37:22,580 Chris Wragg: fucking Yeah. So it doesn't, it doesn't, doesn't 640 00:37:22,580 --> 00:37:26,540 kill us by actually killing us. It just makes us Yeah, 641 00:37:26,540 --> 00:37:29,000 Peter Coghill: it puts us in a little box, makes us dumb. 642 00:37:29,180 --> 00:37:31,160 Fraser McGruer: Sounds, all right, yeah, Nick, 643 00:37:32,600 --> 00:37:34,820 Nick Hare: I think my fact, because this actually mirror 644 00:37:34,820 --> 00:37:37,580 life, made me think of this. I think it's really imaginative. 645 00:37:37,580 --> 00:37:41,860 Is ice nine, which I think is Ray. Is it Ray Bradbury, or, I 646 00:37:41,860 --> 00:37:48,460 can't remember if it's him or but it's a story about a new 647 00:37:48,460 --> 00:37:54,340 type of water molecule that is kind of ice at kind of room 648 00:37:54,340 --> 00:37:58,660 temperature. Basically, it's a Okay, and so. So all of the 649 00:37:58,660 --> 00:38:01,560 water in the entire world just freezes, because this one 650 00:38:01,560 --> 00:38:05,160 molecule kind of ends up converting all the other water 651 00:38:05,160 --> 00:38:08,340 into itself. So, yeah, ice nine gradually. So obviously all the 652 00:38:08,340 --> 00:38:10,980 water in your body freezes, everything the earth just turns 653 00:38:10,980 --> 00:38:15,000 into a great big, you know, Ice Cube made of ice nine. I think 654 00:38:15,000 --> 00:38:17,700 it's very cool idea, but it's not unlike the mirror life 655 00:38:17,700 --> 00:38:17,940 thing. 656 00:38:18,840 --> 00:38:20,960 Peter Coghill: There is a, there is a version of that. There's a 657 00:38:20,960 --> 00:38:25,640 good, great cooker video that that's actually physically 658 00:38:25,640 --> 00:38:26,240 plausible. 659 00:38:26,240 --> 00:38:29,180 Unknown: Oh, okay, we should start worrying about 660 00:38:29,720 --> 00:38:33,080 Peter Coghill: matter, which is something to do with the 661 00:38:33,620 --> 00:38:37,100 configuration of the subatomic particles inside the quarks, 662 00:38:37,100 --> 00:38:41,300 which mean you can get a special kind of quark, or too many of 663 00:38:41,300 --> 00:38:44,740 the special kind of quarks together. And if the problem 664 00:38:44,740 --> 00:38:49,240 with that is that any other protons and neutrons that that 665 00:38:49,240 --> 00:38:53,200 that proton, that weird proton, interacts with, get converted to 666 00:38:53,200 --> 00:38:54,280 the same sort 667 00:38:54,340 --> 00:38:56,440 Nick Hare: so we've got to keep an eye on these quarks to stop 668 00:38:56,440 --> 00:38:58,000 them all getting together in the same 669 00:38:58,000 --> 00:39:00,100 Peter Coghill: try and dig it out for the show notes. Yeah, 670 00:39:00,100 --> 00:39:02,460 worth of view. But, yeah, there's a sort of version of 671 00:39:02,520 --> 00:39:05,520 that. So it's like, it becomes, like the Midas touch, that 672 00:39:05,520 --> 00:39:08,700 anything a particle of that touches, turns into more of 673 00:39:08,700 --> 00:39:09,120 that. 674 00:39:09,120 --> 00:39:12,120 Nick Hare: Okay, yeah, well, it's like those prions. They're 675 00:39:12,120 --> 00:39:15,720 another thing that calls kreutzfeldt Jakob disease, you 676 00:39:15,720 --> 00:39:18,900 know, Mad Cow Disease. Okay? They're basically a, I think 677 00:39:18,900 --> 00:39:21,680 they either a mirror image or just a particular kind of 678 00:39:21,680 --> 00:39:25,580 protein or something that that your body can ingest but can't 679 00:39:25,580 --> 00:39:28,340 then process or deal with, yeah? But they also, they're kind of 680 00:39:28,340 --> 00:39:30,920 self replicating, which is why they're so scary. If they get 681 00:39:30,920 --> 00:39:33,200 into the food chain, then they start spreading everywhere, 682 00:39:33,380 --> 00:39:35,900 which is why you shouldn't do that. 683 00:39:35,900 --> 00:39:37,890 Fraser McGruer: That's why they're not a good thing. Yeah, 684 00:39:37,938 --> 00:39:40,876 prawns are bad. Yeah. So I have to say, I mean, I don't really 685 00:39:40,924 --> 00:39:43,625 have one beyond what we've talked about. But I don't like 686 00:39:43,673 --> 00:39:46,469 the sound of everything turning to ice. I like the sound of 687 00:39:46,517 --> 00:39:49,503 that. I have to say, I did quite like the sound of Peter's one, 688 00:39:49,550 --> 00:39:52,489 like just the eating and being caught in a box. Yeah? Patient, 689 00:39:52,536 --> 00:39:55,475 yeah, that sounded All right. So I'll go with that one. Pretty 690 00:39:55,522 --> 00:39:58,414 much your life at the moment. Well, that's why I try. Is what 691 00:39:58,461 --> 00:40:01,115 I aim for. Yeah? Okay, so we're. Stop there. You've been 692 00:40:01,163 --> 00:40:03,912 listening to the Cognitive Engineering podcast, brought to 693 00:40:03,959 --> 00:40:06,661 you by Aleph Insights and produced by me, Fraser McGruer. 694 00:40:06,708 --> 00:40:09,552 If you haven't already, please like and subscribe. We try to 695 00:40:09,600 --> 00:40:12,491 release an episode every week or two. If there are any topics 696 00:40:12,538 --> 00:40:15,430 you'd like us to cover, please do get in touch via email, and 697 00:40:15,477 --> 00:40:17,610 you can find out more about Aleph insights at 698 00:40:17,657 --> 00:40:20,549 Alephinsights.com thanks, as always, for listening until next 699 00:40:20,596 --> 00:40:21,260 time. Goodbye.