1 00:00:01,010 --> 00:00:04,370 S1: Welcome to Unsupervised Learning, a security, AI and meaning focused 2 00:00:04,370 --> 00:00:07,250 S1: podcast that looks at how best to thrive as humans 3 00:00:07,250 --> 00:00:11,479 S1: in a post AI world. It combines original ideas, analysis, 4 00:00:11,480 --> 00:00:14,720 S1: and mental models to bring not just the news, but 5 00:00:14,720 --> 00:00:23,919 S1: why it matters and how to respond. All right. Episode 437. 6 00:00:23,920 --> 00:00:27,490 S1: What are we doing here? Okay, so I'm delivering my 7 00:00:27,490 --> 00:00:33,370 S1: augmented AI course again live. And this is on July 26th. 8 00:00:33,370 --> 00:00:37,699 S1: It's actually my birthday. And so that'll be next month. 9 00:00:37,700 --> 00:00:42,409 S1: And the theme is moving towards human 3.0. How to 10 00:00:42,409 --> 00:00:46,860 S1: survive and thrive in the world full of AI. And yeah, 11 00:00:46,860 --> 00:00:50,700 S1: I got human 3.0 conversation, skill sets and mental frames, 12 00:00:50,700 --> 00:00:53,519 S1: integrating AI into your life and work and a whole 13 00:00:53,520 --> 00:00:56,940 S1: bunch of resources. Essentially, it's going to be a lot 14 00:00:56,940 --> 00:00:59,280 S1: of hands on, actually showing you the way that I 15 00:00:59,280 --> 00:01:03,779 S1: use AI and adding a whole new section above and 16 00:01:03,780 --> 00:01:08,460 S1: beyond the previous class, which talks about how to basically 17 00:01:08,459 --> 00:01:11,100 S1: get ready, what things do you actually have to learn. 18 00:01:11,100 --> 00:01:13,020 S1: And a big part of this is so that you 19 00:01:13,020 --> 00:01:15,660 S1: could tell the other people that you care about, especially 20 00:01:15,660 --> 00:01:18,870 S1: young people like people in high school, people in about 21 00:01:18,870 --> 00:01:22,260 S1: to go into college, people in college, or people changing 22 00:01:22,260 --> 00:01:25,949 S1: careers who just have massive anxiety from all this AI stuff. 23 00:01:25,950 --> 00:01:30,209 S1: This is basically going to give my outline of what 24 00:01:30,209 --> 00:01:32,070 S1: I think is coming and how to get ready for it, 25 00:01:32,069 --> 00:01:35,550 S1: like how to not just survive it, but actually be 26 00:01:35,550 --> 00:01:39,059 S1: antifragile so that you actually benefit from it. So that 27 00:01:39,060 --> 00:01:43,650 S1: is next month, the end of July. My friend Monica 28 00:01:43,650 --> 00:01:48,930 S1: is launching her cybersecurity leadership Masterclass. Goes over tons of stuff, 29 00:01:48,930 --> 00:01:51,210 S1: which I talk about here, and it's just a really 30 00:01:51,210 --> 00:01:54,480 S1: good course. I think she really has a good perspective 31 00:01:54,480 --> 00:01:56,220 S1: on this, and I got a link here to sign 32 00:01:56,220 --> 00:02:00,600 S1: up for the class doing a webinar with Elastic Security. 33 00:02:01,050 --> 00:02:04,680 S1: And that's coming up Tuesday, June 25th. And I changed 34 00:02:04,950 --> 00:02:08,340 S1: some headers. Doesn't really matter. All right. My work. So 35 00:02:08,340 --> 00:02:11,190 S1: few new essays this week. So the first one is 36 00:02:11,190 --> 00:02:15,180 S1: talking about ratings and basically how I believe AI is 37 00:02:15,180 --> 00:02:19,860 S1: going to basically do like this. This 123 punch one 38 00:02:19,860 --> 00:02:24,810 S1: is teaching us things. Second is testing us to see 39 00:02:24,810 --> 00:02:27,300 S1: how well we know that material. And the third one 40 00:02:27,300 --> 00:02:31,020 S1: is giving these scores. And, uh, I think it's going 41 00:02:31,020 --> 00:02:33,000 S1: to be really crazy. You should you should check out 42 00:02:33,000 --> 00:02:35,700 S1: the whole post. This is the post here. And I 43 00:02:35,700 --> 00:02:38,460 S1: basically talk about like the different scores. I talk about, 44 00:02:38,460 --> 00:02:42,450 S1: like ways and components of the score that are going 45 00:02:42,450 --> 00:02:46,440 S1: to be pretty crazy. And it starts to get a 46 00:02:46,440 --> 00:02:49,859 S1: little bit scary, quite, quite a bit scary actually. Okay. 47 00:02:49,950 --> 00:02:52,620 S1: Imagine you're walking around, you have like the score above 48 00:02:52,620 --> 00:02:55,260 S1: your head because because it's not just going to be 49 00:02:55,260 --> 00:02:57,420 S1: for a test that you took or whatever, right? It's 50 00:02:57,419 --> 00:03:00,150 S1: also going to be for like, hey, instead of me 51 00:03:00,180 --> 00:03:02,820 S1: having to vet every single person that I want to 52 00:03:02,820 --> 00:03:06,390 S1: date or a college having to vet incoming people, why 53 00:03:06,389 --> 00:03:09,780 S1: don't they just send you to do this Astra test, 54 00:03:09,780 --> 00:03:12,870 S1: whatever that is? This is some random company, a made 55 00:03:12,870 --> 00:03:15,060 S1: up company, right? But they're going to get your blood 56 00:03:15,060 --> 00:03:18,360 S1: work and they're going to do all of this, your 57 00:03:18,360 --> 00:03:21,390 S1: personal life history there. And keep in mind, this is 58 00:03:21,389 --> 00:03:25,049 S1: an AI avatar asking you these things. There's a camera 59 00:03:25,050 --> 00:03:28,800 S1: on you watching your facial expressions and your body language. 60 00:03:28,800 --> 00:03:31,410 S1: And keep in mind, this is like GPT six or 61 00:03:31,410 --> 00:03:34,980 S1: GPT seven or whatever. So it's like extremely advanced AI. 62 00:03:35,010 --> 00:03:37,080 S1: So it could read your body language. It knows when 63 00:03:37,080 --> 00:03:39,869 S1: you're lying, it knows when you're uncomfortable about a thing, 64 00:03:39,870 --> 00:03:42,510 S1: but it's going to ask you about your traumas. It's 65 00:03:42,510 --> 00:03:45,840 S1: going to ask you about your like your beliefs about 66 00:03:45,840 --> 00:03:48,570 S1: different things. It's going to ask you about morality. It's 67 00:03:48,570 --> 00:03:51,360 S1: also going to test your technical skills, ask you to 68 00:03:51,360 --> 00:03:54,780 S1: actually build things. It's going to deep dive into your 69 00:03:54,780 --> 00:03:57,540 S1: opinions on like tons of different things. It's going to 70 00:03:57,540 --> 00:04:01,380 S1: look at your base core knowledge, math, physics, biology, history, economics, 71 00:04:01,380 --> 00:04:05,130 S1: all these other disciplines. It's basically going to understand like 72 00:04:05,130 --> 00:04:08,760 S1: how smart you are, all the different stuff that you know, 73 00:04:08,790 --> 00:04:13,500 S1: your maturity across lots of different areas like emotional maturity, 74 00:04:13,500 --> 00:04:17,220 S1: your wisdom, things like that. And then it's going to 75 00:04:17,220 --> 00:04:20,130 S1: ask about your finances. It's going to ask about like 76 00:04:20,130 --> 00:04:22,860 S1: how much money do you make? What is your earning potential, 77 00:04:22,860 --> 00:04:25,350 S1: why do you think you can make that much money, blah, blah, blah. 78 00:04:25,380 --> 00:04:28,260 S1: All of that combined think of like a three day 79 00:04:28,260 --> 00:04:31,140 S1: test where eight hours a day, you're talking to this 80 00:04:31,140 --> 00:04:34,169 S1: AI avatar, and the whole time you're on camera and 81 00:04:34,170 --> 00:04:36,270 S1: a lot of people are thinking, why would anyone sign 82 00:04:36,270 --> 00:04:39,270 S1: up for that? The answer is because they want to 83 00:04:39,270 --> 00:04:42,750 S1: know that about themselves. They also have some measure of 84 00:04:42,750 --> 00:04:45,540 S1: trust that the videos aren't going to go out. That 85 00:04:45,540 --> 00:04:48,720 S1: which isn't going to happen. Those companies are absolute going 86 00:04:48,720 --> 00:04:51,990 S1: to get hacked. Right? So that's that's a mess. But 87 00:04:52,260 --> 00:04:54,690 S1: it doesn't matter. Everyone's going to get hacked in this way. 88 00:04:54,690 --> 00:04:57,750 S1: So it kind of just like that doesn't matter. So 89 00:04:57,750 --> 00:05:00,750 S1: you have the videos, the AI does the analysis. And 90 00:05:00,750 --> 00:05:04,920 S1: keep in mind, the AI knows everything about human psychology. 91 00:05:04,920 --> 00:05:09,810 S1: It knows all the subject matter that you're talking about history, physics, 92 00:05:09,810 --> 00:05:12,840 S1: you know, security, like all the different things that you're 93 00:05:12,839 --> 00:05:15,839 S1: good at. It knows how good you are at it 94 00:05:15,839 --> 00:05:18,420 S1: based on the fact that it did the the interview 95 00:05:18,420 --> 00:05:21,210 S1: with you. So it's like taking you into deep water 96 00:05:21,210 --> 00:05:24,300 S1: to see how deep you can go. Right? Well, now, 97 00:05:24,300 --> 00:05:28,109 S1: now think of companies who want to hire somebody. They 98 00:05:28,110 --> 00:05:31,110 S1: don't have to do all this deep vetting, deep vetting, 99 00:05:31,110 --> 00:05:34,710 S1: a three day set of interviews, or a two week 100 00:05:34,710 --> 00:05:38,010 S1: set of interviews with like five different groups of people. 101 00:05:38,010 --> 00:05:41,580 S1: And they're asking like dumbass questions. That does not work. 102 00:05:41,580 --> 00:05:44,580 S1: You see how bad people can be even after they 103 00:05:44,580 --> 00:05:48,380 S1: go through the interview process. Okay. In current hiring methods, 104 00:05:48,380 --> 00:05:51,560 S1: there's like a gauntlet of all these interviews. That gauntlet 105 00:05:51,589 --> 00:05:56,480 S1: of interviews does not yield great people a high percentage 106 00:05:56,480 --> 00:05:59,750 S1: of the time. It does sometimes, but not nearly as 107 00:05:59,750 --> 00:06:04,250 S1: much as it actually should, given like how extensive that is. 108 00:06:04,250 --> 00:06:08,240 S1: Now compare that with somebody who's like, hey, look, if 109 00:06:08,240 --> 00:06:10,700 S1: you have an astro score, I would love for you 110 00:06:10,700 --> 00:06:12,740 S1: to share it. You don't have to. We're not going 111 00:06:12,740 --> 00:06:17,120 S1: to ask you to do it. But of course, hiring managers, 112 00:06:17,120 --> 00:06:21,260 S1: companies assuming it's legal, that's going to be a problem. 113 00:06:21,260 --> 00:06:23,510 S1: There's going to be some regulation against stuff like this, 114 00:06:23,510 --> 00:06:26,089 S1: because it's going to be so good that it's going 115 00:06:26,089 --> 00:06:30,500 S1: to instantly put great people way, way above. And because 116 00:06:30,500 --> 00:06:34,700 S1: it's so good and it's because it's so merit meritocratic, 117 00:06:34,700 --> 00:06:38,510 S1: there's going to be massive pushback. But I think that 118 00:06:38,510 --> 00:06:41,390 S1: it's going to be so useful and so powerful for 119 00:06:41,390 --> 00:06:46,130 S1: startups that they're going to do it anyway, especially this 120 00:06:46,160 --> 00:06:50,480 S1: especially matters in a world of AI, because people are 121 00:06:50,480 --> 00:06:52,460 S1: going to have much smaller teams. It's going to be 122 00:06:52,460 --> 00:06:56,330 S1: like me, the founder, and I need a total, total 123 00:06:56,360 --> 00:06:59,510 S1: badass to be like my chief of staff because they're 124 00:06:59,510 --> 00:07:03,110 S1: going to run my 13 different companies, which are mostly 125 00:07:03,110 --> 00:07:05,989 S1: AI powered. But I need a total killer to be 126 00:07:05,990 --> 00:07:08,930 S1: that person, which means I need a high astro score 127 00:07:08,930 --> 00:07:14,930 S1: specifically in like fluid intelligence, raw IQ, communication skills, like 128 00:07:14,930 --> 00:07:18,470 S1: all these particular things. And what's crazy is I can 129 00:07:18,470 --> 00:07:22,190 S1: describe in an interview to AI exactly what I need, 130 00:07:22,190 --> 00:07:24,620 S1: exactly what my product does. In fact, the AI will 131 00:07:24,620 --> 00:07:27,200 S1: just go get all that stuff. It will then understand 132 00:07:27,200 --> 00:07:30,110 S1: that perfectly. Then it will look at someone's astro score 133 00:07:30,110 --> 00:07:34,100 S1: and just basically find matches and be like, look, this 134 00:07:34,100 --> 00:07:37,790 S1: person is pre-vetted. Their astro score perfectly matches I validated 135 00:07:37,790 --> 00:07:41,420 S1: the Astro score is real, which Astro will also do. 136 00:07:41,420 --> 00:07:43,280 S1: And by the way, Astro, this is a made up company. 137 00:07:43,280 --> 00:07:47,390 S1: It's not real, but there will be many Astros. So 138 00:07:47,390 --> 00:07:50,630 S1: now it's like hey, do you want to talk to them? Yeah, sure. 139 00:07:50,630 --> 00:07:53,270 S1: So you get on a call with them and you're like, hey. Yeah. 140 00:07:53,270 --> 00:07:56,870 S1: So I mean, the AI says that you're really good. 141 00:07:56,870 --> 00:08:01,460 S1: My my da, my da. My personal digital assistant says, 142 00:08:01,460 --> 00:08:04,820 S1: you seem to be really good on paper, which means 143 00:08:04,820 --> 00:08:07,640 S1: through all this AI vetting. So you have a conversation 144 00:08:07,640 --> 00:08:10,850 S1: and what do you know? You start talking about books. 145 00:08:10,850 --> 00:08:14,150 S1: You start talking about what? The way you see the future, 146 00:08:14,150 --> 00:08:16,820 S1: you start talking about ambition. And they're like, oh yeah, 147 00:08:16,820 --> 00:08:20,929 S1: I'm reading those same books. I'm listening to those same podcasts. Yeah, 148 00:08:20,930 --> 00:08:23,150 S1: I just worked out this morning and you're like, Holy crap, 149 00:08:23,150 --> 00:08:27,470 S1: this person's just like me. Total badass. Yeah. Perfect match. 150 00:08:27,470 --> 00:08:30,650 S1: What do you know? What are the odds? Quite good, actually. 151 00:08:30,650 --> 00:08:33,890 S1: Quite good because of the Astro score. So now think 152 00:08:33,890 --> 00:08:39,530 S1: of companies can just rely on the score. What about dating? Dating? Well, 153 00:08:39,590 --> 00:08:42,290 S1: look at all these questionnaires people are filling out and 154 00:08:42,290 --> 00:08:44,990 S1: they're not happy with the dates. And Astro score is 155 00:08:44,990 --> 00:08:47,390 S1: going to be a way better measure of a person 156 00:08:47,390 --> 00:08:51,140 S1: than this. And of course, there's going to be nasty 157 00:08:51,140 --> 00:08:54,349 S1: gross astro scores. Keep in mind it's going to be like, oh, 158 00:08:54,350 --> 00:08:58,730 S1: how pretty is this, this person, how rich is this person? 159 00:08:58,730 --> 00:09:01,969 S1: And they're going to have like whatever, an Omni score 160 00:09:01,970 --> 00:09:06,710 S1: and it's going to be like Douchebaggery score essentially. So 161 00:09:06,710 --> 00:09:10,160 S1: you shouldn't think that just because it's AI generated that 162 00:09:10,160 --> 00:09:13,640 S1: it's a good measure of a person. That's not true. 163 00:09:13,640 --> 00:09:16,730 S1: What I'm talking about is something that's really deep across 164 00:09:16,730 --> 00:09:20,240 S1: multiple facets of a person, where it actually is kind 165 00:09:20,240 --> 00:09:24,620 S1: of capturing how good they are and how valuable they are, 166 00:09:24,620 --> 00:09:29,030 S1: and importantly, how useful they are for a particular application, 167 00:09:29,030 --> 00:09:32,840 S1: such as working inside this company, working on your particular problems, 168 00:09:32,840 --> 00:09:35,270 S1: or if you're trying to build a family and you 169 00:09:35,270 --> 00:09:37,340 S1: want to know, like how good of a dad is 170 00:09:37,340 --> 00:09:39,079 S1: this person going to be? How good of a mom 171 00:09:39,080 --> 00:09:42,350 S1: is this person going to be? Well, now it's focusing 172 00:09:42,350 --> 00:09:45,620 S1: on those areas which were covered in the interview. The 173 00:09:45,620 --> 00:09:51,650 S1: point is, many, many things in society require and desire 174 00:09:52,040 --> 00:09:56,059 S1: to vet. This is universal vetting. That's what this is. 175 00:09:56,059 --> 00:09:59,689 S1: This is universal vetting. You need to vet people to 176 00:09:59,690 --> 00:10:02,060 S1: work on a company with you, to do a project 177 00:10:02,059 --> 00:10:05,060 S1: with you, to be a babysitter for you, to be 178 00:10:05,059 --> 00:10:08,420 S1: a dog sitter, for you to be a life partner. Right. 179 00:10:08,420 --> 00:10:12,230 S1: So these types of scores, the better they are. Let's 180 00:10:12,230 --> 00:10:14,510 S1: say this Astro one is the best one. It's like 181 00:10:14,510 --> 00:10:19,370 S1: a seven day on site, in person talk with all 182 00:10:19,370 --> 00:10:22,610 S1: these different eyes, and they're hitting you from different angles. 183 00:10:22,760 --> 00:10:25,850 S1: They're watching you do jumping jacks. I mean, it's going 184 00:10:25,850 --> 00:10:27,710 S1: to be crazy. Oh, by the way, you submitted a 185 00:10:27,710 --> 00:10:31,309 S1: blood sample. Right now it's looking at your genome. Yeah, yeah. 186 00:10:31,309 --> 00:10:34,130 S1: And that's why when I came up with the name, 187 00:10:34,130 --> 00:10:39,109 S1: I used AI for that Astra. I forget what it means, but. Yeah. Uh, 188 00:10:39,110 --> 00:10:43,310 S1: what's that movie with Ethan Hawke? What's the one where, like, 189 00:10:43,309 --> 00:10:47,030 S1: the guy wasn't chosen to go on the. A flight anyway. 190 00:10:47,030 --> 00:10:51,950 S1: It's dystopian. Okay, 100%. All this stuff is dystopian. All 191 00:10:51,950 --> 00:10:56,420 S1: this awesome stuff is simultaneously dystopian. The question is, can 192 00:10:56,420 --> 00:10:59,630 S1: we take it in good directions? Right. That's the only question. 193 00:10:59,630 --> 00:11:02,210 S1: We know it's going to go in bad directions. That's 194 00:11:02,210 --> 00:11:04,970 S1: a guarantee. I mean, have you seen humanity? Of course 195 00:11:04,970 --> 00:11:07,460 S1: it's going to go in bad directions. The question is, 196 00:11:07,460 --> 00:11:10,189 S1: can we limit some of those? Can we control some 197 00:11:10,190 --> 00:11:14,120 S1: of those? Can we discourage some of those and steer 198 00:11:14,120 --> 00:11:18,530 S1: them in positive like human 3.0 directions? Like ideally you 199 00:11:18,530 --> 00:11:21,560 S1: would have something like an astro score that's super deep 200 00:11:21,559 --> 00:11:25,250 S1: and super knowledgeable, but it's not like, man, not as 201 00:11:25,250 --> 00:11:29,150 S1: attractive man. Oh, there could be a flaw here. We 202 00:11:29,150 --> 00:11:32,810 S1: don't talk to flawed people. A real astro score is 203 00:11:32,809 --> 00:11:35,360 S1: going to be like, hey, this person is nuanced and 204 00:11:35,360 --> 00:11:38,900 S1: this person has, you know, warts. And this person is 205 00:11:38,900 --> 00:11:43,610 S1: like flawed and and complex and multiple layers and blah, blah, blah. 206 00:11:43,610 --> 00:11:46,670 S1: But I think at the core level, you guys could 207 00:11:46,670 --> 00:11:50,030 S1: be soul mates. And here's the list of questions to 208 00:11:50,030 --> 00:11:53,630 S1: like initiate that initial discovery period or whatever. It's not 209 00:11:53,630 --> 00:11:56,570 S1: about perfection. Not in the human based world, not in 210 00:11:56,570 --> 00:11:59,179 S1: the human 3.0 based world. It's not based on perfection. 211 00:11:59,179 --> 00:12:04,290 S1: It's based on deep matches. On the things that matter. 212 00:12:05,120 --> 00:12:11,600 S1: And spending less time or getting confused by surface level things. 213 00:12:12,120 --> 00:12:15,750 S1: He's very attractive. He's very rich. But you find out 214 00:12:15,750 --> 00:12:19,950 S1: three years later, you know, he's in your mind. He's garbage, right? 215 00:12:19,950 --> 00:12:24,720 S1: Because he doesn't treat you well like he treats people poorly. Right. 216 00:12:24,720 --> 00:12:27,450 S1: And the whole thing that he postured with you was 217 00:12:27,450 --> 00:12:29,400 S1: all an act, and you just wasted three years of 218 00:12:29,400 --> 00:12:31,440 S1: your life, and you were supposed to be having kids 219 00:12:31,440 --> 00:12:34,170 S1: by now. So, like, there's a positive way to go 220 00:12:34,170 --> 00:12:38,189 S1: with this. And I think we need to realize, one, 221 00:12:38,190 --> 00:12:43,000 S1: that there are positive opportunities to. Bad things are going 222 00:12:43,000 --> 00:12:46,150 S1: to happen from these types of scores and not be 223 00:12:46,150 --> 00:12:49,660 S1: confused in thinking that because bad things happen, mean the 224 00:12:49,660 --> 00:12:52,150 S1: whole thing is a wash. First of all, doesn't matter 225 00:12:52,150 --> 00:12:54,160 S1: if it's a wash because it's going to happen anyway. 226 00:12:54,340 --> 00:12:57,850 S1: So either way you should be prepared. So lots of 227 00:12:57,850 --> 00:13:00,970 S1: different ways to get value from this type of analysis. 228 00:13:01,210 --> 00:13:04,990 S1: And here we are. We're we're like 15 20 minutes 229 00:13:04,990 --> 00:13:09,430 S1: in or whatever ten 15 minutes in. And uh that 230 00:13:09,429 --> 00:13:13,959 S1: was one essay okay. Next one the fast slow problem. 231 00:13:14,230 --> 00:13:15,910 S1: You know what? This one's short enough. I'm going to 232 00:13:15,910 --> 00:13:20,500 S1: go ahead and perform this one. Is that what it's called, performing? Okay, 233 00:13:20,500 --> 00:13:23,770 S1: so I've been obsessed lately with the concept of slow 234 00:13:23,770 --> 00:13:28,120 S1: versus fast. I'm calling it the fast. Slow problem refers 235 00:13:28,120 --> 00:13:31,030 S1: to the speed and amount of dopamine that you get 236 00:13:31,030 --> 00:13:33,700 S1: from a thing. So in the distant past, an apple 237 00:13:33,700 --> 00:13:36,400 S1: would have been a treat, and especially an apple pie, 238 00:13:36,400 --> 00:13:39,580 S1: because it took so long to prepare and enhance those 239 00:13:39,580 --> 00:13:43,000 S1: apples which were by themselves not easy to come by. 240 00:13:43,000 --> 00:13:47,199 S1: That's slow from the difficulty of finding the thing, and 241 00:13:47,200 --> 00:13:50,560 S1: from the long process of creating an apple pie. A 242 00:13:50,559 --> 00:13:54,610 S1: box of donuts, on the other hand, is fast. First 243 00:13:54,610 --> 00:13:56,860 S1: of all, you could just walk in and say, give 244 00:13:56,860 --> 00:14:00,189 S1: me a dozen donuts. And second of all, a donut 245 00:14:00,220 --> 00:14:04,630 S1: has been engineered to taste better than a hundred crates 246 00:14:04,630 --> 00:14:07,390 S1: of apples. So if you eat like six donuts, one 247 00:14:07,390 --> 00:14:10,210 S1: after another and then you bite into an apple, it 248 00:14:10,210 --> 00:14:14,559 S1: tastes like a 3D piece of like printed styrofoam. Like 249 00:14:14,559 --> 00:14:17,050 S1: it just does not compare to a donut. And there's 250 00:14:17,050 --> 00:14:19,600 S1: lots of other examples of this, like porn being the 251 00:14:19,600 --> 00:14:23,770 S1: fast extracted version of sex, TikTok being the fast extracted 252 00:14:23,770 --> 00:14:27,280 S1: version of reading a book or watching a movie. And 253 00:14:27,280 --> 00:14:31,570 S1: I worry that this is what's happening with human relationships, 254 00:14:31,570 --> 00:14:34,479 S1: like dating. They say that young people don't want to 255 00:14:34,480 --> 00:14:37,540 S1: get into relationships or have sex with each other anymore, 256 00:14:37,540 --> 00:14:40,540 S1: and evidently it's because they would rather do other things. 257 00:14:40,540 --> 00:14:44,410 S1: And this feels very much like a fast, slow problem. 258 00:14:44,410 --> 00:14:48,820 S1: What if this is as simple as teenagers and young 259 00:14:48,820 --> 00:14:52,030 S1: adults having more fast options on their phones and TVs, 260 00:14:52,030 --> 00:14:54,820 S1: and they are so good that the slow options of 261 00:14:54,820 --> 00:14:59,980 S1: companionship and relationships simply cannot compete. It's one thing when 262 00:14:59,980 --> 00:15:03,310 S1: a donut competes with an apple, but in this case, 263 00:15:03,310 --> 00:15:06,670 S1: the fast versus slow problem might actually be affecting the 264 00:15:06,670 --> 00:15:09,670 S1: future of the species. I'm not sure what the complete 265 00:15:09,670 --> 00:15:12,310 S1: answer is here, because I'm sure there are places where 266 00:15:12,310 --> 00:15:16,450 S1: fast versions are simply better and not just faster. But 267 00:15:16,450 --> 00:15:18,280 S1: what I do know is that I think we should 268 00:15:18,280 --> 00:15:22,540 S1: start being very cautious about replacing slow versions of things 269 00:15:22,690 --> 00:15:27,740 S1: with the fast versions of things. There is meaning in 270 00:15:27,740 --> 00:15:32,660 S1: doing things slowly. Slow walks while holding hands. Reading physical 271 00:15:32,660 --> 00:15:37,310 S1: books in natural light and making apple pies from scratch. 272 00:15:37,340 --> 00:15:39,800 S1: I think we should do our best to figure out 273 00:15:39,800 --> 00:15:43,430 S1: what those slow, meaningful things are and ensure that we 274 00:15:43,430 --> 00:15:46,820 S1: keep as many of them in our lives as possible. 275 00:15:46,820 --> 00:15:50,660 S1: That was that one. All right list of hard won 276 00:15:50,660 --> 00:15:53,270 S1: life lessons. I think I'm going to do a separate 277 00:15:53,270 --> 00:15:58,100 S1: one for this, but essentially it's a list of pretty 278 00:15:58,100 --> 00:16:02,480 S1: decent wisdom here. Yeah, I'll give you one. In practice, 279 00:16:02,480 --> 00:16:07,280 S1: weak people behave very similarly to evil people. Don't tie 280 00:16:07,280 --> 00:16:10,430 S1: yourself to either. So that's the vibe there. It's a 281 00:16:10,430 --> 00:16:12,830 S1: whole bunch of those. So I do want to move 282 00:16:12,830 --> 00:16:14,870 S1: to a model where I support a lot more of 283 00:16:14,870 --> 00:16:19,430 S1: my work with memberships, courses and revenue from apps that 284 00:16:19,430 --> 00:16:23,240 S1: I'm building. So I would love for you all to 285 00:16:23,240 --> 00:16:27,170 S1: actually subscribe and become members of UL. So you just 286 00:16:27,170 --> 00:16:32,180 S1: go to Daniel meister.com/upgrade and you can become a member. 287 00:16:32,180 --> 00:16:35,210 S1: And there's just so many benefits. The biggest one is 288 00:16:35,210 --> 00:16:38,510 S1: the actual community, which it's just the most amazing people 289 00:16:38,510 --> 00:16:41,210 S1: that I've ever met. It's uh it's great. We got 290 00:16:41,210 --> 00:16:46,430 S1: a book club, monthly meetups, significant discounts. So that augmented 291 00:16:46,460 --> 00:16:49,160 S1: course that's going to be $200 off if you become 292 00:16:49,160 --> 00:16:51,440 S1: a member. And guess what? Becoming a member is only 293 00:16:51,440 --> 00:16:54,590 S1: 100 bucks a year. So you kind of like make 294 00:16:54,590 --> 00:16:57,140 S1: up for it instantly. And there's just a whole bunch 295 00:16:57,140 --> 00:17:00,200 S1: of other stuff. We actually have physical meetups. It's pretty cool. 296 00:17:00,200 --> 00:17:04,220 S1: All right. Stories House just passed a bill that bans 297 00:17:04,220 --> 00:17:10,220 S1: DJI drones from using FCC frequencies. Disgruntled ex-employee costs over 298 00:17:10,220 --> 00:17:15,470 S1: $600,000 by deleting 180 test servers. Using a script that 299 00:17:15,470 --> 00:17:18,260 S1: he found on Google, Wells Fargo fired over a dozen 300 00:17:18,260 --> 00:17:24,109 S1: employees for faking keyboard activity to look busy. Yeah, AWS 301 00:17:24,140 --> 00:17:28,190 S1: has added Fido Passkeys for MFA, and root users must 302 00:17:28,190 --> 00:17:31,909 S1: enable this by July of 2024. I love this, I 303 00:17:31,910 --> 00:17:36,560 S1: don't think there's anything better recently for security than passkeys. 304 00:17:36,560 --> 00:17:39,770 S1: Like maybe the like the last ten years. Like it's 305 00:17:39,770 --> 00:17:42,650 S1: just really good. New York Times source code was stolen 306 00:17:42,650 --> 00:17:46,580 S1: using an exposed GitHub token like this is an older story. 307 00:17:46,580 --> 00:17:49,910 S1: Maybe it happened again. iOS 18 will let you automatically 308 00:17:49,910 --> 00:17:53,419 S1: record and transcribe phone calls through the phone app. I 309 00:17:53,420 --> 00:17:56,450 S1: have the beta. I need to get that going. Thanks 310 00:17:56,450 --> 00:18:01,190 S1: to a project discovery for sponsoring, Canada is proposing potentially 311 00:18:01,190 --> 00:18:05,750 S1: life in prison for hate speech, and maybe even preemptively 312 00:18:05,750 --> 00:18:09,530 S1: restrict their freedom based on anticipated crimes. Based on what 313 00:18:09,530 --> 00:18:12,440 S1: I saw happen to Jordan Peterson, who I'm not a 314 00:18:12,440 --> 00:18:16,670 S1: complete fan of, by the way, especially lately. But what 315 00:18:16,670 --> 00:18:19,970 S1: I saw that Canada did to him, not a huge 316 00:18:19,970 --> 00:18:24,980 S1: fan linked out. LinkedIn is rolling out AI career coaches. Yeah, 317 00:18:24,980 --> 00:18:27,380 S1: see my essay above on the Astro scores. That's what 318 00:18:27,380 --> 00:18:31,879 S1: we just talked about. AI and LMS are transforming cyber insurance. Yeah, 319 00:18:31,880 --> 00:18:33,830 S1: I really needed to do a whole piece on this. 320 00:18:33,830 --> 00:18:38,330 S1: But real time risk adjustments and pricing adjustments based on 321 00:18:38,330 --> 00:18:43,100 S1: that risk I plus insurance. Yeah that's spicy. It's going 322 00:18:43,100 --> 00:18:44,990 S1: to be good. And when I say good I mean 323 00:18:44,990 --> 00:18:48,230 S1: good for insurance companies. Scale came out with a controversial 324 00:18:48,230 --> 00:18:52,399 S1: stance that they will hire on merit and merit alone. 325 00:18:52,400 --> 00:18:56,689 S1: This thing is worth reading. Okay, skip that one. Tesla 326 00:18:56,690 --> 00:19:02,120 S1: shareholders have backed Elon Musk's $56 billion pay package. Got 327 00:19:02,119 --> 00:19:05,360 S1: two things here where I left the subtitle in there 328 00:19:05,359 --> 00:19:09,360 S1: do not like. Okay. Yeah. So they basically approved $56 329 00:19:09,359 --> 00:19:13,460 S1: billion pay package. This feels very Atlas Shrugged to me. 330 00:19:13,460 --> 00:19:17,300 S1: Basically give the actual builders anything they want because there's 331 00:19:17,300 --> 00:19:20,119 S1: so few of them left and they're so special. I 332 00:19:20,119 --> 00:19:24,320 S1: really like how it's like pro entrepreneur, but I feel 333 00:19:24,320 --> 00:19:27,560 S1: like it comes with a lot of toxicity. Like give 334 00:19:27,590 --> 00:19:31,369 S1: Buck Reinhold a harem because Buck deserves it and there's 335 00:19:31,369 --> 00:19:34,429 S1: cheering in the background. I think we need like a 336 00:19:34,430 --> 00:19:41,270 S1: healthy hybrid, right where we have this massive respect for founders, builders, creators, entrepreneurs. 337 00:19:41,270 --> 00:19:45,740 S1: But we build in the assumption that, look, it's not 338 00:19:45,740 --> 00:19:49,850 S1: just the 1%, they're not special people. Why isn't everyone 339 00:19:49,850 --> 00:19:53,210 S1: doing that? And that's what that's what human 3.0 is, 340 00:19:53,240 --> 00:19:57,109 S1: is getting everyone into this category. So guess what? There's 341 00:19:57,109 --> 00:20:00,350 S1: no more God complexes because everyone's supposed to be that way. Yeah, 342 00:20:00,380 --> 00:20:04,400 S1: great timing for this one. Evidently, Musk has had sexual 343 00:20:04,400 --> 00:20:09,500 S1: relationships with multiple SpaceX employees, including a former intern, which 344 00:20:09,500 --> 00:20:12,169 S1: he hired on to his executive team. And this is 345 00:20:12,170 --> 00:20:16,220 S1: according to Wall Street Journal. Now, yeah, I got complex 346 00:20:16,220 --> 00:20:20,659 S1: thoughts on this. I think radical honesty and like voicing 347 00:20:20,660 --> 00:20:23,690 S1: your desire to be with somebody, all of that is like, 348 00:20:23,690 --> 00:20:25,949 S1: I think Musk is actually doing that better than. And 349 00:20:25,950 --> 00:20:30,480 S1: most people. The problem is he works with these people. Right? 350 00:20:30,480 --> 00:20:34,560 S1: So I've got a fake example here. Julie, great work 351 00:20:34,560 --> 00:20:37,710 S1: on the Montague project I admire you. You should have 352 00:20:37,710 --> 00:20:41,970 S1: my babies. And keep in mind, this is not a 353 00:20:41,970 --> 00:20:44,939 S1: real quote, but this is kind of the vibe, right? 354 00:20:44,940 --> 00:20:49,290 S1: This would be wrong if nothing happened from it. Because 355 00:20:49,470 --> 00:20:52,050 S1: even if that worked out, if that conversation worked out, 356 00:20:52,050 --> 00:20:55,920 S1: her career was great. His career was great. Like no problems. 357 00:20:55,920 --> 00:20:58,410 S1: It would still be wrong because it has so many 358 00:20:58,410 --> 00:21:01,560 S1: chances of going wrong. And this is why this is 359 00:21:01,560 --> 00:21:05,580 S1: not allowed, because the the allegation here, which nobody knows 360 00:21:05,580 --> 00:21:08,340 S1: if this is true, but the allegation is that some 361 00:21:08,340 --> 00:21:12,750 S1: of the women supposedly said that they were given this offer, 362 00:21:12,750 --> 00:21:15,930 S1: they declined and then they got screwed because of it. Right? 363 00:21:15,930 --> 00:21:18,900 S1: They didn't get the promotion, they got fired. Whatever. Whatever 364 00:21:18,900 --> 00:21:21,900 S1: the case was that's being alleged. All right. California has 365 00:21:21,900 --> 00:21:24,989 S1: the biggest wage gap in the US. This seems to 366 00:21:24,990 --> 00:21:28,020 S1: make sense to me. It's like it's the most extreme 367 00:21:28,020 --> 00:21:29,700 S1: version of a big city. So you're going to have 368 00:21:29,700 --> 00:21:32,610 S1: higher highs and lower lows. So that makes sense. We 369 00:21:32,609 --> 00:21:36,600 S1: just broke ground on America's first next gen nuclear facility. 370 00:21:36,600 --> 00:21:39,300 S1: And this is a piece by Bill gates, a really 371 00:21:39,300 --> 00:21:42,060 S1: cool to see him working on energy the way he is. 372 00:21:42,060 --> 00:21:47,280 S1: Russia replaced the dollar and the euro with the Chinese currency, 373 00:21:47,280 --> 00:21:51,570 S1: the yuan yuan, yuan, something like that. There's a fungus 374 00:21:51,570 --> 00:21:55,800 S1: that can eat plastic. I'm worried. What else it can eat? 375 00:21:55,800 --> 00:21:57,689 S1: That's just me. Apple left out a lot of the 376 00:21:57,690 --> 00:22:04,379 S1: small updates in WWDC keynote, so. Oh, I already used this. 377 00:22:04,380 --> 00:22:08,489 S1: So I speak Spanish pretty well and I type in 378 00:22:08,490 --> 00:22:12,150 S1: Spanish a lot, and it's not auto correcting me, it's 379 00:22:12,150 --> 00:22:16,080 S1: letting me type in Spanish mixed with English. Absolutely love it. 380 00:22:16,080 --> 00:22:19,980 S1: So that must be already turned on. Flashlight app has 381 00:22:19,980 --> 00:22:25,109 S1: new animations. Widgets are easier to resize. The Vision Pro 382 00:22:25,109 --> 00:22:29,730 S1: shows your keyboard in VR, and Mac OS has some 383 00:22:29,730 --> 00:22:34,530 S1: new nostalgic wallpapers. Apple Vision Pro lets you watch House 384 00:22:34,530 --> 00:22:37,470 S1: of the Dragon, which is the new Game of Thrones thing, 385 00:22:37,470 --> 00:22:43,980 S1: in an updated, immersive Iron Throne room environment. Spatial personas 386 00:22:43,980 --> 00:22:48,480 S1: envision two I just updated mine and recorded a new persona. 387 00:22:48,510 --> 00:22:51,209 S1: Someone wants a FaceTime? Let me know. And got an 388 00:22:51,210 --> 00:22:54,330 S1: argument here that Jupyter notebooks are the fast food of coding. 389 00:22:54,330 --> 00:22:59,070 S1: Convenient but often unhealthy for long term projects. And Jensen 390 00:22:59,070 --> 00:23:03,750 S1: Huang told Caltech grads to pursue $0 billion markets, which 391 00:23:03,750 --> 00:23:07,109 S1: are markets with no current value but huge future potential. 392 00:23:07,170 --> 00:23:08,880 S1: And I just figured out what I love so much 393 00:23:08,880 --> 00:23:12,420 S1: about Jensen. He's like a permanently nice version of Elon 394 00:23:12,420 --> 00:23:16,379 S1: and we've all seen Elon be nice. I actually love Elon. 395 00:23:16,380 --> 00:23:19,080 S1: I think he's amazing when I see him talk about 396 00:23:19,080 --> 00:23:23,399 S1: space or the future or saving humanity. I'm like this 397 00:23:23,400 --> 00:23:26,040 S1: guy is literally my hero. Then I see him posting 398 00:23:26,040 --> 00:23:29,899 S1: memes on Twitter about politics and I'm like. Do not 399 00:23:29,900 --> 00:23:34,220 S1: be this person. Do not let people make you into 400 00:23:34,220 --> 00:23:37,790 S1: this person. And I get very angry at him. So 401 00:23:37,790 --> 00:23:39,860 S1: it's great to see that you can have like an 402 00:23:39,859 --> 00:23:45,350 S1: Elan type, like an Atlas shrug type who's also nice. 403 00:23:45,350 --> 00:23:49,679 S1: It's amazing. Got someone predicting that generative AI will actually 404 00:23:49,680 --> 00:23:52,710 S1: increase the demand for software engineers over the next 20 years, 405 00:23:52,710 --> 00:23:56,250 S1: not decrease it. I think there's some validity to that. 406 00:23:56,250 --> 00:23:58,949 S1: I think we'll also decrease it, but also increase. The 407 00:23:58,950 --> 00:24:02,580 S1: question is which one wins out? I think probably increase 408 00:24:02,580 --> 00:24:05,520 S1: wins out in the beginning. Widely held view that sperm 409 00:24:05,520 --> 00:24:08,280 S1: counts and men are dropping around the world may be wrong. 410 00:24:08,280 --> 00:24:12,090 S1: And according to a new study and University of Manchester. 411 00:24:12,090 --> 00:24:14,790 S1: So I tend to go by the quality of these 412 00:24:14,790 --> 00:24:19,199 S1: things to determine if I include it or not. And 413 00:24:19,200 --> 00:24:21,000 S1: if I really want to look, I run it through 414 00:24:21,000 --> 00:24:25,140 S1: fabric and see like how it scores the paper. Curiosity 415 00:24:25,140 --> 00:24:30,060 S1: yeah argues that perfectionism is about optimizing at the wrong scale. 416 00:24:30,270 --> 00:24:33,330 S1: I love that, I love that statement. I love the sentence. 417 00:24:33,330 --> 00:24:37,619 S1: Perfection is about optimizing at the wrong scale. This reminds 418 00:24:37,619 --> 00:24:40,590 S1: me of another thing that I really love. The worst 419 00:24:40,590 --> 00:24:43,530 S1: way to lose is to win at the wrong game. 420 00:24:43,890 --> 00:24:48,480 S1: And that's what this reminds me of. Perfectionism. Perfectionism is 421 00:24:48,480 --> 00:24:52,050 S1: optimizing at the wrong scale. Love it. 38% of web 422 00:24:52,050 --> 00:24:57,750 S1: pages that existed in 2013 are no longer accessible. Derek 423 00:24:57,750 --> 00:25:01,320 S1: Sivers explains why you should create a now page. I've 424 00:25:01,320 --> 00:25:03,780 S1: been doing this for a long time. I don't currently 425 00:25:03,780 --> 00:25:06,179 S1: have one up. That's because I'm going to have it 426 00:25:06,180 --> 00:25:09,450 S1: in my personal API soon, which will be all AI powered. 427 00:25:09,450 --> 00:25:13,170 S1: Basically a Da going to be cool stuff that's going 428 00:25:13,170 --> 00:25:15,179 S1: to take a little bit longer. I'm not actively working 429 00:25:15,180 --> 00:25:17,969 S1: on it yet. Waiting for the tech to evolve a 430 00:25:17,970 --> 00:25:21,840 S1: few more months. Okay, ideas. Increased production of goods and 431 00:25:21,840 --> 00:25:25,980 S1: services after AI, and how does that make sense? If 90% 432 00:25:25,980 --> 00:25:28,320 S1: of people don't have money to buy the stuff? So 433 00:25:28,320 --> 00:25:31,410 S1: I'm talking about the economics there. Got someone arguing that 434 00:25:31,410 --> 00:25:33,660 S1: in a world dominated by AI and robots, the key 435 00:25:33,660 --> 00:25:37,650 S1: to human survival isn't being smarter, but being more mediocre. 436 00:25:37,680 --> 00:25:43,980 S1: I included it just because I like alternative views. Pop cultures, oligopoly, 437 00:25:43,980 --> 00:25:46,889 S1: dream machine, AI model. Yeah, lumen labs this thing is 438 00:25:46,890 --> 00:25:52,320 S1: blowing up. Nvidia Warp Infinite Content ideas generator can now 439 00:25:52,320 --> 00:25:57,120 S1: design and manufacture your own chips. LM Mojo Olama new 440 00:25:57,119 --> 00:26:02,669 S1: version enhanced GPU discovery SQLite is likely the most used 441 00:26:02,670 --> 00:26:06,780 S1: database engine. Pretty cool argument there. There's a new Starlink 442 00:26:06,780 --> 00:26:09,420 S1: coming out, a tiny one that's really good for like 443 00:26:09,420 --> 00:26:13,440 S1: camping and RVs and stuff. And, uh, Max lighter has 444 00:26:13,440 --> 00:26:16,770 S1: a simple but powerful tip ship something every day doesn't 445 00:26:16,770 --> 00:26:18,840 S1: have to be big, just something you can point to. 446 00:26:18,840 --> 00:26:21,899 S1: Recommendation of the week. I recommend you check out my 447 00:26:21,900 --> 00:26:24,810 S1: list of hard won life lessons I was talking about earlier. 448 00:26:24,810 --> 00:26:27,330 S1: The full list is pretty good, and the aphorism of 449 00:26:27,330 --> 00:26:29,400 S1: the week. The sad truth is that most evil is 450 00:26:29,400 --> 00:26:32,400 S1: done by people who never make up their minds to 451 00:26:32,400 --> 00:26:35,970 S1: be good or evil. The sad truth is that most 452 00:26:35,970 --> 00:26:38,130 S1: evil is done by people who never make up their 453 00:26:38,130 --> 00:26:44,010 S1: minds to be good or evil. Hannah Arendt. Unsupervised Learning 454 00:26:44,010 --> 00:26:46,650 S1: is produced and edited by Daniel Missler on a Neumann 455 00:26:46,650 --> 00:26:51,240 S1: U87 AI microphone using Hindenburg. Intro and outro music is 456 00:26:51,240 --> 00:26:54,480 S1: by zombie with the why and to get the text 457 00:26:54,480 --> 00:26:56,640 S1: and links from this episode, sign up for the newsletter 458 00:26:56,640 --> 00:27:02,010 S1: version of the show at Daniel meisler.com/newsletter. We'll see you 459 00:27:02,010 --> 00:27:02,640 S1: next time.