1 00:00:01,129 --> 00:00:04,460 S1: Welcome to Unsupervised Learning, a security, AI and meaning focused 2 00:00:04,460 --> 00:00:07,340 S1: podcast that looks at how best to thrive as humans 3 00:00:07,340 --> 00:00:11,570 S1: in a post AI world. It combines original ideas, analysis, 4 00:00:11,570 --> 00:00:14,810 S1: and mental models to bring not just the news, but 5 00:00:14,810 --> 00:00:22,470 S1: why it matters and how to respond. All right. Welcome 6 00:00:22,470 --> 00:00:28,710 S1: to episode 438. This is Daniel Miessler. Okay. What do 7 00:00:28,710 --> 00:00:35,200 S1: we have here? So registration for augmented V2 July 26th 8 00:00:35,200 --> 00:00:38,680 S1: is now open. About half the slots are about gone. 9 00:00:39,210 --> 00:00:43,560 S1: So sign up by clicking down below in the newsletter. 10 00:00:43,560 --> 00:00:48,479 S1: Or you can just go to augmented dot Unsupervised learning.com 11 00:00:48,479 --> 00:00:54,130 S1: and sign up. And, uh, if you become a UL member, 12 00:00:54,130 --> 00:00:59,830 S1: you actually get $250 off. So inside of the chat, 13 00:00:59,830 --> 00:01:03,130 S1: you can get the discount code, which is $250 off, 14 00:01:03,130 --> 00:01:08,560 S1: and becoming a member is only $99 for the first year. So. 15 00:01:09,910 --> 00:01:14,030 S1: You could do the math there. Let's see here. Fabric 16 00:01:14,030 --> 00:01:19,100 S1: now supports Anthropic's sonnet 3.5, which is really, really good. 17 00:01:19,100 --> 00:01:22,940 S1: So if you update and then you do fabric list models, 18 00:01:23,120 --> 00:01:26,540 S1: it will show you that, uh, Claude 3.5 is in 19 00:01:26,540 --> 00:01:30,290 S1: there and this is the actual tag for it. Claude 20 00:01:30,290 --> 00:01:36,470 S1: dash three, dash five. Sonnet dash 20240620. Uh, one thing 21 00:01:36,500 --> 00:01:39,200 S1: to note, though, is sonnet and haiku. They throw a 22 00:01:39,200 --> 00:01:43,669 S1: lot of copyright failures like it won't actually read podcast transcripts, 23 00:01:43,670 --> 00:01:45,980 S1: which is super annoying because it's one of my favorite 24 00:01:45,980 --> 00:01:48,620 S1: things to do is put long form conversations through these 25 00:01:48,620 --> 00:01:53,090 S1: things and like pull out like lists of references that 26 00:01:53,090 --> 00:01:55,790 S1: they mentioned if I want to order the books or whatever. 27 00:01:55,790 --> 00:01:57,980 S1: So that's kind of annoying that it doesn't allow me 28 00:01:57,980 --> 00:02:01,250 S1: to do that with one of the lower models. Opus 29 00:02:01,250 --> 00:02:03,860 S1: does allow me to do that, and GPT four does 30 00:02:03,860 --> 00:02:06,740 S1: allow me to do that, but I can't do it 31 00:02:06,740 --> 00:02:11,140 S1: in sonnet or haiku. So I'm really waiting for them 32 00:02:11,139 --> 00:02:14,530 S1: to either fix that with the lower models or to 33 00:02:14,530 --> 00:02:18,850 S1: give me opus 3.5. I cannot wait to use that. 34 00:02:19,150 --> 00:02:22,300 S1: Definitely gearing up for Vegas black Hat Defcon. Hope to 35 00:02:22,300 --> 00:02:24,790 S1: see you there! I'm going to have stickers, but, uh, 36 00:02:24,790 --> 00:02:28,989 S1: don't really have anything else other than stickers. Might eventually 37 00:02:28,990 --> 00:02:32,079 S1: do a t shirt or something with like a superhero 38 00:02:32,080 --> 00:02:36,100 S1: logo of just, uh, the UL logo, but no text 39 00:02:36,100 --> 00:02:38,980 S1: or anything. So I've been feeling quite grateful for my 40 00:02:38,980 --> 00:02:43,270 S1: life lately and somehow, like, super powered. Not really, because, uh, 41 00:02:43,270 --> 00:02:46,750 S1: things are going well, which, uh, they're definitely going pretty 42 00:02:46,750 --> 00:02:49,690 S1: well in terms of, like, the business and everything like that, 43 00:02:49,690 --> 00:02:52,809 S1: but it's more so that I feel like I'm living 44 00:02:52,810 --> 00:02:55,720 S1: an authentic life. I feel like the stuff that I'm 45 00:02:55,720 --> 00:02:58,780 S1: doing is all oriented in the direction, and I'm working 46 00:02:58,780 --> 00:03:02,350 S1: on a, like a essays and a course around this 47 00:03:02,350 --> 00:03:05,050 S1: and stuff like that. But it's essentially like, do you 48 00:03:05,050 --> 00:03:07,900 S1: know what you're trying to accomplish? And are you heading 49 00:03:07,900 --> 00:03:10,840 S1: in that direction? And what I've been trying to do 50 00:03:10,840 --> 00:03:14,140 S1: is like, hack together these systems where I say no 51 00:03:14,139 --> 00:03:17,020 S1: to things that are not in the direction of these goals. 52 00:03:17,020 --> 00:03:19,390 S1: And one thing that came up in a conversation with 53 00:03:19,389 --> 00:03:23,290 S1: my buddy recently is, um, the cool thing about this. Oh, 54 00:03:23,290 --> 00:03:26,830 S1: and I also got this from a David Perell analysis 55 00:03:26,830 --> 00:03:31,040 S1: of a bunch of Seinfeld stuff. It's like fail. At 56 00:03:31,040 --> 00:03:34,810 S1: things that you wanted to do anyway. And it really 57 00:03:34,810 --> 00:03:38,050 S1: doesn't feel like failure, right? Because I'm doing a lot 58 00:03:38,050 --> 00:03:41,290 S1: of different stuff. Maybe too much sometimes, sometimes not enough. 59 00:03:41,290 --> 00:03:45,130 S1: But in general, I'm doing lots of different things. And 60 00:03:45,130 --> 00:03:48,460 S1: the question is, how many do you say no to? 61 00:03:48,460 --> 00:03:51,640 S1: And what is your criteria for saying yes or no? 62 00:03:51,670 --> 00:03:56,920 S1: And what I've come to basically figure out is that 63 00:03:57,010 --> 00:03:59,740 S1: if you look at that Seinfeld thing, he basically said, 64 00:03:59,740 --> 00:04:03,670 S1: fail doing it, doing what you want the way that 65 00:04:03,670 --> 00:04:08,260 S1: you want. And this really resonated with me, because I 66 00:04:08,260 --> 00:04:11,380 S1: think about people who are kind of chasing other people. 67 00:04:11,380 --> 00:04:14,170 S1: They're chasing like a high school bully, or they're chasing 68 00:04:14,170 --> 00:04:19,450 S1: like their boss or whatever, some, some competitor for some reason, 69 00:04:19,450 --> 00:04:23,710 S1: and they're chasing them, and they might actually not be 70 00:04:23,710 --> 00:04:28,600 S1: thinking about what really drives them, what purpose they are chasing, 71 00:04:28,630 --> 00:04:32,230 S1: the thing that they actually want to do deeply. That's 72 00:04:32,230 --> 00:04:35,800 S1: part of their own personal identity. Instead, they might be going, well, 73 00:04:35,800 --> 00:04:39,100 S1: the thing I'm doing isn't working. The thing they're doing 74 00:04:39,100 --> 00:04:41,650 S1: is working, so I'm going to copy them. Or if 75 00:04:41,650 --> 00:04:45,430 S1: someone gives them advice and they're like, hey! You know, 76 00:04:45,430 --> 00:04:48,159 S1: this thing that I'm doing works really, really well. You 77 00:04:48,160 --> 00:04:50,560 S1: should do that. And one thing that comes to mind 78 00:04:50,560 --> 00:04:54,460 S1: around this is Alex Hormozi. He is, like, obsessed with 79 00:04:54,460 --> 00:04:57,490 S1: making money. Him and his wife are obsessed with making money, 80 00:04:57,490 --> 00:05:00,969 S1: which I respect. They're also way younger than me. So 81 00:05:00,970 --> 00:05:04,359 S1: they're like in their late 20s, early 30s. Or maybe 82 00:05:04,360 --> 00:05:07,360 S1: he's like mid 30s or whatever, but he wants to 83 00:05:07,360 --> 00:05:10,510 S1: crush business. And guess what? He is crushing business. So 84 00:05:10,510 --> 00:05:11,859 S1: what I'm going to do is take some of that 85 00:05:11,860 --> 00:05:14,650 S1: DNA from crushing business and try to sprinkle it into 86 00:05:14,650 --> 00:05:16,810 S1: my stuff. But what I won't be doing is making 87 00:05:16,810 --> 00:05:20,469 S1: the mistake of trying to be Alex Hormozi and switching 88 00:05:20,470 --> 00:05:24,670 S1: my focus and my telescope to point towards this distant 89 00:05:24,670 --> 00:05:27,460 S1: thing of like, oh, I want to make $100 million, 90 00:05:27,460 --> 00:05:29,890 S1: and he doesn't want to make $100 million. He wants 91 00:05:29,890 --> 00:05:32,740 S1: to be a billionaire. I guarantee you that's his goal 92 00:05:32,740 --> 00:05:35,770 S1: is like $1 billion in revenue or wealth or something. 93 00:05:35,770 --> 00:05:40,540 S1: So the thing is, you have to know what direction 94 00:05:40,540 --> 00:05:43,900 S1: your stuff is pointing in, and then you can more 95 00:05:43,900 --> 00:05:47,560 S1: easily say no to things because are they going to 96 00:05:47,560 --> 00:05:50,200 S1: get you there sort of accidentally? And that's the way 97 00:05:50,200 --> 00:05:52,480 S1: I've sort of oriented things, is like, I want to 98 00:05:52,480 --> 00:05:55,990 S1: do things so that I can accidentally make money from them, 99 00:05:55,990 --> 00:05:58,750 S1: like do stuff that you would do anyway and hook 100 00:05:58,750 --> 00:06:00,609 S1: it up to a stripe page and hope it works out. 101 00:06:00,610 --> 00:06:03,160 S1: And I know that's not the way to make a 102 00:06:03,160 --> 00:06:05,500 S1: lot of money. And I know you should, like, go 103 00:06:05,500 --> 00:06:08,380 S1: all in on a thing that's like, got a giant 104 00:06:08,380 --> 00:06:11,200 S1: market and, you know, you spam the hell out of 105 00:06:11,200 --> 00:06:14,500 S1: the ads or whatever, and you just try to make 106 00:06:14,500 --> 00:06:16,750 S1: all your money from that. But I feel like it's 107 00:06:16,750 --> 00:06:20,590 S1: just I feel like it's a bad game. And another 108 00:06:20,589 --> 00:06:22,900 S1: one of my favorite things, which I've mentioned here before, 109 00:06:22,900 --> 00:06:26,230 S1: is like, the worst way to lose is to win 110 00:06:26,230 --> 00:06:28,539 S1: playing the wrong game. And so this is all sort 111 00:06:28,540 --> 00:06:31,630 S1: of like flowing into that same thing. And I've been 112 00:06:31,630 --> 00:06:34,120 S1: happy with the newsletter lately because I've just been more 113 00:06:34,120 --> 00:06:37,930 S1: of my authentic self, and I feel like that's something 114 00:06:37,930 --> 00:06:41,890 S1: I just want for everybody. I want everyone to be 115 00:06:41,890 --> 00:06:44,589 S1: in tune with who they are and what they want, 116 00:06:44,589 --> 00:06:48,969 S1: and have them actually chasing those things and talking about 117 00:06:48,970 --> 00:06:52,240 S1: those things, talking about it is super important. It's like 118 00:06:52,240 --> 00:06:57,010 S1: I'm really worried about people becoming invisible after AI, because 119 00:06:57,010 --> 00:06:59,560 S1: AI is going to be finding the coolest stuff and 120 00:06:59,560 --> 00:07:02,620 S1: the best stuff. Well, if you're not saying anything, AI 121 00:07:02,620 --> 00:07:04,960 S1: is not going to find you. And also, if you're 122 00:07:04,960 --> 00:07:08,200 S1: not saying anything, you might not be thinking the things enough, 123 00:07:08,200 --> 00:07:11,560 S1: or you might not be in practice enough with coming 124 00:07:11,560 --> 00:07:14,470 S1: up with the ideas, crystallizing them, and writing them down. 125 00:07:14,470 --> 00:07:18,850 S1: Because I'm not advocating that people try to become successful authors, 126 00:07:18,850 --> 00:07:22,120 S1: or they try to become successful streamers or YouTubers or 127 00:07:22,120 --> 00:07:25,000 S1: anything like that for the purpose of being successful. I'm 128 00:07:25,000 --> 00:07:28,120 S1: advocating that they do it because it will make them 129 00:07:28,120 --> 00:07:32,650 S1: feel good to know themselves and to articulate themselves and 130 00:07:32,650 --> 00:07:36,010 S1: to emote themselves. And that's why I'm obsessed with this 131 00:07:36,010 --> 00:07:39,340 S1: whole thing. That's why I'm obsessed with this human 3.0 132 00:07:39,340 --> 00:07:41,740 S1: and all the stuff that I'm talking about. So anyway, 133 00:07:41,740 --> 00:07:44,110 S1: I hope you've noticed the difference in terms of me. 134 00:07:44,110 --> 00:07:47,110 S1: But really, the thing I'm trying to do, the whole 135 00:07:47,110 --> 00:07:49,690 S1: purpose of doing this for myself is to like, battle 136 00:07:49,690 --> 00:07:51,790 S1: test it so I can help other people do it. 137 00:07:51,790 --> 00:07:56,020 S1: All right. Stream of consciousness, collection of thoughts on US politics. 138 00:07:56,020 --> 00:07:58,120 S1: Just skip it if you don't like politics. But I 139 00:07:58,120 --> 00:08:00,640 S1: basically lay down a whole bunch of stuff there. Oh, 140 00:08:00,640 --> 00:08:03,760 S1: this is just become a member. Yeah, yeah, become a 141 00:08:03,760 --> 00:08:06,670 S1: member if you can. I would appreciate it. All right. 142 00:08:06,670 --> 00:08:10,060 S1: CSA held its first. This is, uh, security stuff here 143 00:08:10,060 --> 00:08:13,210 S1: or it's the New stories section. But, yeah, I always 144 00:08:13,210 --> 00:08:18,040 S1: start with security. CIA held its first AI security tabletop exercise, 145 00:08:18,040 --> 00:08:22,270 S1: over 50 experts, and they're looking to simulate responses to 146 00:08:22,270 --> 00:08:25,660 S1: AI security incidents. I love that they're thinking about it. 147 00:08:25,660 --> 00:08:29,410 S1: Extremists in the US are using AI to spread hate speech, 148 00:08:29,410 --> 00:08:33,040 S1: recruit and radicalize faster than ever. And they got a 149 00:08:33,040 --> 00:08:37,150 S1: couple of examples here President Biden using using racial slurs 150 00:08:37,150 --> 00:08:40,150 S1: and Emma Watson reading Mein Kampf. I kind of want 151 00:08:40,150 --> 00:08:42,910 S1: to go watch these, but I also don't want to 152 00:08:42,910 --> 00:08:46,900 S1: Google for them. So yeah, I do not need that 153 00:08:46,900 --> 00:08:50,050 S1: garbage in my YouTube feed. The US has banned the 154 00:08:50,050 --> 00:08:54,820 S1: sale of Kaspersky anti-virus software, citing security risks from Russia. 155 00:08:54,820 --> 00:08:57,490 S1: And they're basically using strong language here. They're like, do 156 00:08:57,490 --> 00:09:01,030 S1: not use this stuff, stop using it immediately, get off 157 00:09:01,030 --> 00:09:04,060 S1: of it. And I think they have a deadline in July, 158 00:09:04,059 --> 00:09:06,939 S1: and there's a new set of high security vulnerabilities in Asus. 159 00:09:06,940 --> 00:09:11,440 S1: Routers let you get complete control without any user interaction. 160 00:09:11,440 --> 00:09:14,050 S1: And you want to patch these immediately. And the Richter 161 00:09:14,050 --> 00:09:23,710 S1: scale is 9.8. Thank you to dropzone for sponsoring Dmitri Alperovitch. Yeah. 162 00:09:23,710 --> 00:09:29,020 S1: Dmitri Alperovitch why is that difficult? That's not difficult. Imagined 163 00:09:29,020 --> 00:09:33,490 S1: a detailed scenario of China invading Taiwan in 2028, focusing 164 00:09:33,490 --> 00:09:38,560 S1: on aerosol strategy through the rough Taiwan Strait waters. This 165 00:09:38,559 --> 00:09:42,069 S1: one is fascinating to me because I don't have strong 166 00:09:42,070 --> 00:09:46,530 S1: intuitions myself I don't have. Expertise myself in like this 167 00:09:46,530 --> 00:09:51,540 S1: whole naval warfare island attack type situation. It's just it's 168 00:09:51,540 --> 00:09:54,120 S1: not something I've ever studied. I don't know the stuff, 169 00:09:54,120 --> 00:09:57,690 S1: but what's really fascinating to me is that the super, 170 00:09:57,690 --> 00:10:01,200 S1: super experts on this, they disagree with each other, which 171 00:10:01,200 --> 00:10:03,840 S1: reminds me a lot of the AI situation there. I 172 00:10:03,840 --> 00:10:08,940 S1: do have strong intuitions and some expertise and a lot 173 00:10:08,940 --> 00:10:11,910 S1: of expertise in some areas, and some expertise in others, 174 00:10:11,910 --> 00:10:16,410 S1: but it's like I feel way more grounded and solid 175 00:10:16,410 --> 00:10:20,729 S1: and confident in at least my intuitions on the AI side. 176 00:10:20,730 --> 00:10:24,540 S1: But here I have no idea what's going on. And 177 00:10:24,540 --> 00:10:28,830 S1: in both cases, the absolute best people, which we're supposed 178 00:10:28,830 --> 00:10:30,660 S1: to be looking up to, they do not agree with 179 00:10:30,660 --> 00:10:35,310 S1: each other and they disagree massively with each other. Tyler 180 00:10:35,309 --> 00:10:40,380 S1: Cowan had someone this woman from, she was Bulgarian on 181 00:10:40,380 --> 00:10:44,370 S1: his podcast recently and it was so good. And her 182 00:10:44,370 --> 00:10:47,550 S1: view was completely different than so many other people. So 183 00:10:47,550 --> 00:10:51,030 S1: I'm watching like Peter Zeihan. I'm listening to her. I'm 184 00:10:51,030 --> 00:10:54,480 S1: listening to the Goodfellas, which is a podcast of a 185 00:10:54,480 --> 00:10:58,650 S1: bunch of security and like really smart people. And I'm 186 00:10:58,650 --> 00:11:02,069 S1: trying to triangulate between all these different people and basically 187 00:11:02,070 --> 00:11:05,280 S1: come up with my own thoughts and opinions based on, 188 00:11:05,280 --> 00:11:08,010 S1: you know, not being an expert on it. I just 189 00:11:08,010 --> 00:11:10,860 S1: I find it so interesting that they can't agree. Oh, 190 00:11:10,860 --> 00:11:16,200 S1: same with, um, same with Ukraine. Right. Vastly different predictions 191 00:11:16,200 --> 00:11:17,970 S1: on whether or not it was going to happen in 192 00:11:17,970 --> 00:11:22,709 S1: the first place, when it's going to end, vastly different predictions. 193 00:11:22,710 --> 00:11:27,120 S1: And that excites me that that massively excites me because 194 00:11:27,120 --> 00:11:30,780 S1: I'm like, there is signal in here somewhere. The question is, 195 00:11:30,780 --> 00:11:35,090 S1: is it available second? Are people not getting the signal 196 00:11:35,090 --> 00:11:41,510 S1: or are they misinterpreting it? How are the best experts wrong? 197 00:11:41,510 --> 00:11:44,510 S1: And it raises the question of like, what else are 198 00:11:44,510 --> 00:11:48,140 S1: the best experts wrong on? And I just like to 199 00:11:48,140 --> 00:11:53,209 S1: have that constant tension and questioning in my mind as 200 00:11:53,210 --> 00:11:56,360 S1: I think about everything, including the things that I. I 201 00:11:56,360 --> 00:11:59,630 S1: believe the strongest US is moving to limit investments in 202 00:11:59,630 --> 00:12:04,760 S1: China's AI, semiconductor and quantum computing sectors to curb China's 203 00:12:04,760 --> 00:12:09,560 S1: tech advancements. And now they're looking at private equity venture 204 00:12:09,559 --> 00:12:14,720 S1: capital funds. Great stuff. And then I say, yeah, cool. 205 00:12:14,720 --> 00:12:17,750 S1: But we need to be building a lot more energy production. 206 00:12:17,750 --> 00:12:21,710 S1: Look at this. Look at this graphic from Leopold. This 207 00:12:21,710 --> 00:12:24,800 S1: is from situational awareness, which I just printed out, by 208 00:12:24,800 --> 00:12:27,350 S1: the way. It's like, I don't know, two inches thick 209 00:12:27,350 --> 00:12:29,959 S1: front and back. It's it's a beast. And I'm just 210 00:12:29,960 --> 00:12:32,540 S1: going to keep it in my car and read it. Uh, 211 00:12:32,540 --> 00:12:37,309 S1: whenever I'm parked and whatever, listening to podcasts. I won't 212 00:12:37,309 --> 00:12:39,380 S1: do them both at the same time. But anyway, I'm 213 00:12:39,380 --> 00:12:41,420 S1: going to keep it in the car and read through it. 214 00:12:41,420 --> 00:12:43,730 S1: I've already read a whole bunch online and of course, 215 00:12:43,730 --> 00:12:50,000 S1: watch the video, but. Yeah. All right. National. Yeah. Nationwide 216 00:12:50,000 --> 00:12:53,359 S1: armed militia from this guy named Jake Lang, who is 217 00:12:53,360 --> 00:12:58,520 S1: a January 6th writer. And they're using telegram for coordination. 218 00:12:58,820 --> 00:13:02,300 S1: And this is a wired article. Study says staring directly 219 00:13:02,300 --> 00:13:06,170 S1: at the camera during online job interviews can significantly boost 220 00:13:06,170 --> 00:13:11,089 S1: your evaluation scores. And this is from University of Hiroshima. 221 00:13:11,120 --> 00:13:14,929 S1: AI agents and the Raz Revolution results as a service 222 00:13:14,929 --> 00:13:20,570 S1: discussing the evolution from SAS to raz, emphasizing the role 223 00:13:20,570 --> 00:13:25,160 S1: of AI agents. Yeah, I like this a lot. It's 224 00:13:25,160 --> 00:13:27,709 S1: like abstract. This is exactly what I'm building with a 225 00:13:27,710 --> 00:13:29,929 S1: project called substrate, which you're going to hear a lot 226 00:13:29,929 --> 00:13:34,490 S1: about here before too long. But this is exactly what 227 00:13:34,490 --> 00:13:39,109 S1: I'm building. You make a query and agents and a 228 00:13:39,110 --> 00:13:41,960 S1: whole bunch of data work together in a whole bunch 229 00:13:41,960 --> 00:13:45,140 S1: of steps, and the output that you get is the 230 00:13:45,140 --> 00:13:49,939 S1: result of their work. So if you imagine one idea 231 00:13:49,940 --> 00:13:52,970 S1: I have here, which is super exciting, is I'm going 232 00:13:52,970 --> 00:13:56,990 S1: to build intelligence reports this way. So on one hand, 233 00:13:56,990 --> 00:14:00,980 S1: you've got a whole bunch of sources that are maybe 234 00:14:00,980 --> 00:14:04,400 S1: they're hard to find, maybe they're not, but there are 235 00:14:04,400 --> 00:14:07,370 S1: many of them. And it takes some effort and curation 236 00:14:07,370 --> 00:14:11,030 S1: ability to to pick the best ones. Right. You bring 237 00:14:11,030 --> 00:14:14,479 S1: that together, you turn that into a giant summary or 238 00:14:14,480 --> 00:14:18,470 S1: a giant context file, and you basically start producing intelligence 239 00:14:18,470 --> 00:14:21,500 S1: artifacts out of this thing. And that's what I like 240 00:14:21,500 --> 00:14:26,090 S1: about results as a service. Okay. You you ask the question, 241 00:14:26,090 --> 00:14:29,360 S1: what is going on in Ukraine? What are the latest 242 00:14:29,360 --> 00:14:32,570 S1: global trends? Where should I invest my money or whatever? 243 00:14:32,570 --> 00:14:37,100 S1: And there could be hundreds of operations that happen, multiple 244 00:14:37,100 --> 00:14:40,940 S1: different eyes doing multiple different things. But really, it's not 245 00:14:40,940 --> 00:14:43,910 S1: even so much the AI until the end where it's 246 00:14:43,910 --> 00:14:46,820 S1: like putting it together and telling the story. It starts 247 00:14:46,820 --> 00:14:49,640 S1: with more legacy tech. It starts with all this collection, 248 00:14:49,640 --> 00:14:55,190 S1: all this correlation and unification of this data in getting 249 00:14:55,190 --> 00:14:58,100 S1: it into a thing that's ready for the for the 250 00:14:58,100 --> 00:15:03,710 S1: actual AI. And having brushed up against intelligence a little 251 00:15:03,710 --> 00:15:06,200 S1: bit when I was in the Army, I was attached 252 00:15:06,200 --> 00:15:09,500 S1: to an intelligence group and got to see a little 253 00:15:09,500 --> 00:15:12,260 S1: bit of how things work. Very little. I don't want 254 00:15:12,260 --> 00:15:15,500 S1: to oversell it here, but I've also read a whole 255 00:15:15,500 --> 00:15:19,010 S1: bunch about it and seen a whole bunch of this 256 00:15:19,010 --> 00:15:23,120 S1: stuff in on the Intel side and also on the 257 00:15:23,120 --> 00:15:26,780 S1: military side, and it comes down. Oh, I also built 258 00:15:26,780 --> 00:15:32,540 S1: a program around this concept at Apple for the security org. Right. 259 00:15:32,540 --> 00:15:35,150 S1: And this is a product that's still being used today 260 00:15:35,150 --> 00:15:39,260 S1: inside the company, which I'm very proud of. But I 261 00:15:39,260 --> 00:15:45,770 S1: use these concepts there. It's like data information intelligence. And 262 00:15:45,770 --> 00:15:50,270 S1: the intelligence piece is the one where there's intelligence, human 263 00:15:50,270 --> 00:15:54,530 S1: intelligence going into that effort to turn this into a narrative, 264 00:15:54,530 --> 00:15:59,150 S1: into a story for decision support. So at the above intelligence, 265 00:15:59,150 --> 00:16:02,780 S1: you have decision, right? Or maybe maybe it's not above, 266 00:16:02,780 --> 00:16:05,570 S1: maybe it's, um, to the right or whatever. But it's 267 00:16:05,570 --> 00:16:09,050 S1: like the purpose of that intelligence is to distill down. 268 00:16:09,050 --> 00:16:13,220 S1: So the the general is not looking at log files. Right. 269 00:16:13,220 --> 00:16:17,420 S1: Which is like data streaming in. Okay. You have information. 270 00:16:17,420 --> 00:16:19,520 S1: It's like a summary of the log files. And this 271 00:16:19,520 --> 00:16:23,060 S1: might not be a perfect analogy, but it moves up 272 00:16:23,060 --> 00:16:27,920 S1: and up in layers of abstraction with the entire purpose 273 00:16:27,920 --> 00:16:31,190 S1: of a human making a decision. Right? So if you 274 00:16:31,190 --> 00:16:35,390 S1: look at results as a service, I'm thinking intelligence as 275 00:16:35,390 --> 00:16:38,360 S1: a service is what I'm thinking. That's the first thing 276 00:16:38,360 --> 00:16:39,890 S1: in my mind goes to there's a whole bunch of 277 00:16:39,890 --> 00:16:43,460 S1: different implementations of this. Like it could be a a 278 00:16:43,460 --> 00:16:46,790 S1: business decision. Do I cancel this customer? Do I refund 279 00:16:46,790 --> 00:16:50,000 S1: this customer? Do I do whatever? But either way, it's 280 00:16:50,000 --> 00:16:53,000 S1: a whole bunch of legacy things happening with collection and 281 00:16:53,000 --> 00:16:57,320 S1: stuff like that, and IT systems and data sources and everything. 282 00:16:57,320 --> 00:17:01,100 S1: And then there's this thing that traditionally a human is done, 283 00:17:01,100 --> 00:17:06,590 S1: or like Deborah or Chris or Raj or whoever it is, 284 00:17:06,590 --> 00:17:08,750 S1: they look at all the data and they say, you 285 00:17:08,750 --> 00:17:12,590 S1: know what, we are going to issue a payout for 286 00:17:12,590 --> 00:17:16,820 S1: this insurance policy, and that is the decision part. And 287 00:17:16,820 --> 00:17:19,940 S1: the result is a yes or no. Thumbs up, thumbs down. 288 00:17:19,940 --> 00:17:23,510 S1: And that gets us to this results as a service. Right? 289 00:17:23,510 --> 00:17:26,060 S1: So ideally you could just take all your different data, 290 00:17:26,060 --> 00:17:29,419 S1: all your different decision points, throw that into context, throw 291 00:17:29,420 --> 00:17:33,140 S1: it up to one of these services, which is one 292 00:17:33,140 --> 00:17:36,230 S1: of these that I'm building as well. And in like 293 00:17:36,230 --> 00:17:38,600 S1: 3 or 4 years, everyone's going to know this, and 294 00:17:38,600 --> 00:17:40,520 S1: this is just going to be the way all software 295 00:17:40,520 --> 00:17:43,640 S1: is built. But I'm already building this and I've already 296 00:17:43,640 --> 00:17:48,080 S1: got pieces of this that work. Um, and it's it's 297 00:17:48,080 --> 00:17:50,689 S1: really exciting. I mean, this is essentially the future of 298 00:17:50,690 --> 00:17:54,560 S1: software is doing this. You ask a question, you include context, 299 00:17:54,560 --> 00:17:57,859 S1: and you get back a result, a very high quality result. 300 00:17:57,859 --> 00:18:01,730 S1: So very excited about this concept and very good piece 301 00:18:01,730 --> 00:18:04,460 S1: here on medium, which I wish nothing was on medium 302 00:18:04,460 --> 00:18:08,570 S1: but whatever. All right. Sam Logan has a new post 303 00:18:08,570 --> 00:18:11,899 S1: on how to get structured output from Llms covering various 304 00:18:11,900 --> 00:18:16,160 S1: frameworks and their pros and cons. Yeah, really enjoyed this one. 305 00:18:16,160 --> 00:18:19,970 S1: Talked about converting English to JSON and a bunch of 306 00:18:19,970 --> 00:18:23,390 S1: other stuff but yeah, enjoyed that. All right, someone got 307 00:18:23,390 --> 00:18:26,659 S1: invited to a project called Rebind where I versions of 308 00:18:26,660 --> 00:18:30,350 S1: authors interact with readers about classic books. These are the 309 00:18:30,350 --> 00:18:34,070 S1: really cool ideas. I just love ideas like this. Coming 310 00:18:34,100 --> 00:18:38,090 S1: to life Fabian Boff from Octo Mind explains why they 311 00:18:38,090 --> 00:18:42,200 S1: stopped using laying chain for their agents, highlighting its rigid, 312 00:18:42,200 --> 00:18:45,920 S1: high level abstractions. So I'm going to say something extra 313 00:18:45,920 --> 00:18:51,530 S1: spicy right now. A lot of these low level libraries 314 00:18:51,530 --> 00:18:57,409 S1: and frameworks for doing really difficult things like Rag, uh, 315 00:18:57,410 --> 00:19:01,550 S1: or agents, a lot of these agent frameworks, things like 316 00:19:01,550 --> 00:19:07,940 S1: laying chain, they are extremely kludgy. If anything was like overhyped, 317 00:19:07,940 --> 00:19:12,920 S1: I would say these first generation efforts at these things 318 00:19:12,920 --> 00:19:16,639 S1: are massively overhyped. And it kind of goes back to 319 00:19:16,640 --> 00:19:19,940 S1: this results as a service thing. Results as a service 320 00:19:19,940 --> 00:19:23,060 S1: is kind of the future of this, because it's not 321 00:19:23,060 --> 00:19:28,790 S1: asking you to mesh stack together these Jenga blocks of 322 00:19:28,790 --> 00:19:33,139 S1: like really kludgy code. Because here's the problem. And it 323 00:19:33,140 --> 00:19:39,109 S1: kind of applies to, um, to, uh, what are those called, um, notebooks. 324 00:19:39,109 --> 00:19:42,950 S1: It kind of applies to notebooks as well, because notebooks 325 00:19:42,950 --> 00:19:47,960 S1: like notebooks and agents and in Lang Chain, they make 326 00:19:47,960 --> 00:19:51,619 S1: great demos, they make great demos. But when you try 327 00:19:51,619 --> 00:19:54,170 S1: to take this stuff and build a production app on it, 328 00:19:54,170 --> 00:19:57,560 S1: you throw real scenarios into it, you throw real data 329 00:19:57,560 --> 00:20:01,790 S1: at it and it just falls over. And what universally, 330 00:20:01,790 --> 00:20:05,360 S1: almost everyone I've seen, and this is definitely what I've done. 331 00:20:05,359 --> 00:20:09,740 S1: You have to rebuild the entire thing from scratch using 332 00:20:09,740 --> 00:20:13,070 S1: smaller parts, which I've always been obsessed with this smaller 333 00:20:13,070 --> 00:20:16,760 S1: parts connected together. So it's like it it's necessary for 334 00:20:16,760 --> 00:20:19,910 S1: AI as well, which is why I'm building my very own, 335 00:20:19,910 --> 00:20:24,050 S1: like complete stack on this, uh, which is called substrate. 336 00:20:24,050 --> 00:20:28,250 S1: And I'm basically using those components to build apps. And 337 00:20:28,250 --> 00:20:32,060 S1: each one of the components is highly trusted and highly 338 00:20:32,060 --> 00:20:36,710 S1: rigorous in what it takes in and processes and what 339 00:20:36,710 --> 00:20:40,610 S1: it outputs. So it's dependable all the way through the pipeline. 340 00:20:40,609 --> 00:20:44,419 S1: And if you don't do this and you just start 341 00:20:44,420 --> 00:20:49,940 S1: piecing together laying chain and agent frameworks. And don't get 342 00:20:49,940 --> 00:20:54,200 S1: me wrong, I am not like, uh, being negative about 343 00:20:54,200 --> 00:20:58,010 S1: these things. Laying chain has done amazing things for AI 344 00:20:58,010 --> 00:21:03,230 S1: and these agent frameworks Autogen crew AI. They're doing amazing 345 00:21:03,230 --> 00:21:07,429 S1: stuff too, like no question. But as someone who's actually 346 00:21:07,430 --> 00:21:10,430 S1: building this stuff, the reality is completely different when you 347 00:21:10,430 --> 00:21:14,270 S1: start throwing production data at this stuff because you need consistency, 348 00:21:14,270 --> 00:21:19,399 S1: you need trust in these systems. And, uh, pretty much 349 00:21:19,400 --> 00:21:23,900 S1: every one that I know, they started building with these 350 00:21:23,900 --> 00:21:28,580 S1: bigger blocks. And then they got a few months in 351 00:21:28,580 --> 00:21:31,220 S1: and they had to tear it all down and start over, 352 00:21:31,220 --> 00:21:35,270 S1: building their own individual components. And so I think that's 353 00:21:35,270 --> 00:21:38,780 S1: where this is all going. Um, and eventually that's going 354 00:21:38,780 --> 00:21:41,660 S1: to merge into results as a service, which is kind 355 00:21:41,660 --> 00:21:45,050 S1: of like the opposite, actually. It's more like somebody did 356 00:21:45,050 --> 00:21:49,580 S1: build that back end, uh, dependable infrastructure, and now you're 357 00:21:49,580 --> 00:21:53,330 S1: just getting the results from it. Um, and the way 358 00:21:53,330 --> 00:21:55,909 S1: to know if it works is does it work? Do 359 00:21:55,910 --> 00:21:58,190 S1: you keep getting the right answer? And if you do, 360 00:21:58,220 --> 00:22:01,220 S1: you'll we will, as a society, build trust in that 361 00:22:01,220 --> 00:22:05,480 S1: over time. And ideally the thing will also be it 362 00:22:05,480 --> 00:22:10,850 S1: will also be publishing information about its components and the 363 00:22:10,850 --> 00:22:14,659 S1: sources that it used. So there'll be like some transparency there. Uh, 364 00:22:14,660 --> 00:22:17,270 S1: that will be part of. The quality of the service 365 00:22:17,270 --> 00:22:20,359 S1: is if you can actually inspect it and inspect like, okay, 366 00:22:20,359 --> 00:22:24,169 S1: which models did you use, which data sources did you use? Right. 367 00:22:24,170 --> 00:22:26,390 S1: Because if you're just like, no, trust me, because I'm 368 00:22:26,390 --> 00:22:29,719 S1: awesome and you like the results and don't worry about it, 369 00:22:29,720 --> 00:22:32,120 S1: for a lot of people, that won't be enough. All right? 370 00:22:32,119 --> 00:22:36,410 S1: My buddy Evan Orzolek and long time Ulr, he argues 371 00:22:36,410 --> 00:22:39,919 S1: that our obsession with convenience is making us less human, 372 00:22:39,920 --> 00:22:42,590 S1: and he wrote this on his blog, Snake Eye Software. 373 00:22:42,590 --> 00:22:46,010 S1: Solar generated a fifth of global electricity at midday on 374 00:22:46,010 --> 00:22:51,560 S1: the summer solstice. Solstice? Yep, a fifth of global energy 375 00:22:51,980 --> 00:22:57,690 S1: at midday. On summer solstice. That is interesting. I feel like, 376 00:22:57,900 --> 00:23:00,960 S1: I don't know, I'm not an expert on energy either, 377 00:23:00,960 --> 00:23:05,310 S1: but I, I really just wish we would go way heavier, 378 00:23:05,310 --> 00:23:10,979 S1: way heavier on solar, build, 100 giant mega solar farms 379 00:23:11,310 --> 00:23:15,550 S1: or 500 of them. And just go head to head 380 00:23:15,550 --> 00:23:21,909 S1: against China on solar. Because nuclear reactors, I understand that 381 00:23:21,910 --> 00:23:24,610 S1: they're really good energy and I understand they're awesome and 382 00:23:24,609 --> 00:23:28,990 S1: I understand all that, but they're very complex. They are 383 00:23:28,990 --> 00:23:33,580 S1: fundamentally dangerous, even though I know they're way safer than 384 00:23:33,580 --> 00:23:36,640 S1: they are now. It's possible to make them safe. Here's 385 00:23:36,640 --> 00:23:39,669 S1: another reason they freak people out. So there's going to 386 00:23:39,670 --> 00:23:42,820 S1: be so much more friction to trying to get one 387 00:23:42,820 --> 00:23:46,660 S1: or 2 or 10 of them built. And we are 388 00:23:46,660 --> 00:23:51,400 S1: in a race where speed matters and time matters. So 389 00:23:51,400 --> 00:23:53,890 S1: China can just build a whole bunch of those because 390 00:23:53,890 --> 00:23:58,090 S1: they have nobody to ask. They're just like, move, we move, 391 00:23:58,090 --> 00:24:02,020 S1: we're building stuff here. And everyone's like, yeah, China. Like 392 00:24:02,020 --> 00:24:06,189 S1: they have just this massive advantage. Well, we can we 393 00:24:06,190 --> 00:24:08,290 S1: can move fast too if we do solar and it's 394 00:24:08,290 --> 00:24:12,070 S1: not freaking people out. So let's just go build 500 395 00:24:12,070 --> 00:24:15,700 S1: mega solar plants like Elon has been talking about forever, 396 00:24:15,700 --> 00:24:19,570 S1: and just jump way ahead, way ahead of the game 397 00:24:19,570 --> 00:24:23,440 S1: and still do nuclear, right? And you know where that 398 00:24:23,440 --> 00:24:25,929 S1: makes sense and where we can. But don't let it 399 00:24:25,930 --> 00:24:28,420 S1: slow us down. So when it's talking about using SSH 400 00:24:28,450 --> 00:24:32,350 S1: as a pseudo replacement, uh, basically you can use it 401 00:24:32,350 --> 00:24:37,210 S1: to execute commands as another user. So why not use 402 00:24:37,210 --> 00:24:42,100 S1: that permission model instead of pseudo? Interesting. Apple is renaming 403 00:24:42,100 --> 00:24:46,330 S1: Apple ID to Apple account I like it. A group 404 00:24:46,330 --> 00:24:52,030 S1: of 17 secondary schools in Southwark, London, are going smartphone 405 00:24:52,210 --> 00:24:56,920 S1: free to combat the negative effects of phone use on students, 406 00:24:57,760 --> 00:25:00,700 S1: and I really love that Jonathan Heights work here is 407 00:25:00,700 --> 00:25:03,190 S1: spreading so fast. He put out his new book. He's 408 00:25:03,190 --> 00:25:07,840 S1: been blasting it all over podcasts. And I just love 409 00:25:07,840 --> 00:25:11,080 S1: that his work is catching on. And Gavin Newsom is 410 00:25:11,080 --> 00:25:14,200 S1: now going to ban, uh, smartphones. Or he's trying he's 411 00:25:14,200 --> 00:25:17,409 S1: trying to ban them. And likely this is coming from 412 00:25:17,410 --> 00:25:21,190 S1: impact from height as well. House prices are surging again 413 00:25:21,190 --> 00:25:25,359 S1: with global index up over 3% year on year. Insufficient 414 00:25:25,359 --> 00:25:29,830 S1: sun exposure is now a serious public health issue, potentially 415 00:25:29,830 --> 00:25:32,770 S1: causing hundreds of thousands of deaths annually in the US 416 00:25:32,770 --> 00:25:35,469 S1: and Europe. Isn't this like cheese and eggs and milk 417 00:25:35,470 --> 00:25:39,129 S1: and I don't know, so many, so many things. It's like, uh, 418 00:25:39,130 --> 00:25:42,340 S1: that'll kill you. Well, everyone stops and it's like, yeah, 419 00:25:42,340 --> 00:25:46,000 S1: if you don't do that, that'll kill you. So I 420 00:25:46,000 --> 00:25:49,540 S1: can't wait till Huberman figures this out for us. Uh, 421 00:25:49,540 --> 00:25:52,960 S1: reads all the papers, uh, invites all the experts on. 422 00:25:52,960 --> 00:25:54,730 S1: And I know a lot of people are starting to 423 00:25:54,730 --> 00:25:59,470 S1: doubt Huberman and stuff like that, and I'm. I'm getting 424 00:25:59,470 --> 00:26:02,290 S1: a little more cautious with him. I'm not sure why, 425 00:26:02,290 --> 00:26:04,210 S1: but I'm getting a little more cautious with him. But 426 00:26:04,210 --> 00:26:06,820 S1: I don't know anyone else who is looking at these 427 00:26:06,820 --> 00:26:10,510 S1: papers and doing this great analysis. And I don't fault 428 00:26:10,510 --> 00:26:14,980 S1: people for being wrong sometimes I don't, I fault people 429 00:26:14,980 --> 00:26:21,730 S1: for being overconfident, never looking at themselves, never going after 430 00:26:21,730 --> 00:26:25,419 S1: their own errors, never trying to self-correct, and not being 431 00:26:25,420 --> 00:26:29,320 S1: analytical and pointing that lens at themselves. And I feel 432 00:26:29,320 --> 00:26:32,020 S1: like Huberman is good at that. I would like to 433 00:26:32,020 --> 00:26:34,660 S1: him to come back and do some episodes on, like, 434 00:26:34,660 --> 00:26:37,090 S1: here's stuff I got wrong and why and why I'm 435 00:26:37,090 --> 00:26:40,300 S1: going to fix that in the future. But anyway, he's 436 00:26:40,300 --> 00:26:42,970 S1: the best there is right now. And what I would 437 00:26:42,970 --> 00:26:45,640 S1: like to know about the sun thing is, I feel 438 00:26:45,640 --> 00:26:47,830 S1: better when I go out in the sun in the morning, 439 00:26:47,830 --> 00:26:50,290 S1: and I'm not sure it's only because of the light 440 00:26:50,290 --> 00:26:54,129 S1: going in my eyes. I feel like there's something about 441 00:26:54,130 --> 00:26:57,850 S1: walking on the ground, absorbing sunlight, and a lot of 442 00:26:57,850 --> 00:27:00,910 S1: this feels very emotional. So that's why I actually want 443 00:27:00,910 --> 00:27:03,940 S1: to see data. Like, I don't want this to be 444 00:27:03,940 --> 00:27:07,840 S1: based on vibes, but meanwhile I'm going to do it 445 00:27:07,840 --> 00:27:11,560 S1: based on vibes. Seems like a net good. Oh, in 446 00:27:11,560 --> 00:27:16,869 S1: one caveat, there is, um, between 10 and 2 is 447 00:27:16,869 --> 00:27:20,080 S1: like the worst time because your UV index is the worst. 448 00:27:20,080 --> 00:27:24,100 S1: But like super early sun is supposedly fairly safe because 449 00:27:24,100 --> 00:27:27,969 S1: the angle of incidence, uh, like coming into the atmosphere. 450 00:27:27,970 --> 00:27:32,380 S1: So evidently the UV is pretty low, but the visible 451 00:27:32,380 --> 00:27:35,860 S1: light I think is still much higher. I'm not sure 452 00:27:35,859 --> 00:27:39,610 S1: if those are correlated, but visible light is going to, uh, 453 00:27:39,609 --> 00:27:43,840 S1: set your clock, your circadian rhythm. And I feel like 454 00:27:43,840 --> 00:27:48,040 S1: there's just other health benefits there, too. And anyway, what 455 00:27:48,040 --> 00:27:50,679 S1: that study was basically saying is that a lot of 456 00:27:50,680 --> 00:27:53,980 S1: people aren't getting any sun, they're not going outside, they're 457 00:27:53,980 --> 00:27:57,310 S1: not doing anything. And I wonder how mixed that is 458 00:27:57,310 --> 00:28:00,190 S1: with like other things like exercise or whatever. But obviously 459 00:28:00,190 --> 00:28:02,679 S1: they try to filter that out. But the the idea 460 00:28:02,680 --> 00:28:06,280 S1: is basically people aren't going outside and they're depressed. And 461 00:28:06,280 --> 00:28:10,210 S1: guess what? There's probably a connection. They're having more positive 462 00:28:10,210 --> 00:28:12,850 S1: experience in life is linked to lower odds of brain 463 00:28:12,850 --> 00:28:18,850 S1: disorders like Alzheimer's and lower cognitive decline. And, uh, got 464 00:28:18,850 --> 00:28:23,979 S1: this secret Ted Cruz funding documents, slightly redacted. Got this, um, 465 00:28:24,520 --> 00:28:28,330 S1: this guy over here called Capital Press, and he's basically 466 00:28:28,330 --> 00:28:34,090 S1: like a watchdog, like a one person, like journalist group monitoring, 467 00:28:34,090 --> 00:28:39,500 S1: like the goings on of Congress people. And it's not in, like, 468 00:28:39,500 --> 00:28:44,630 S1: a stalkery nasty way. Uh, or a legal way. I 469 00:28:44,630 --> 00:28:48,620 S1: think it's more like Osint. Like what? Things are being 470 00:28:48,620 --> 00:28:51,380 S1: dropped out there in the public, you know, little clues 471 00:28:51,380 --> 00:28:55,220 S1: that can be hoovered up and, uh, and checked out. 472 00:28:55,220 --> 00:28:57,500 S1: And this is one thing I'm really excited about. This 473 00:28:57,500 --> 00:29:00,830 S1: is another thing I'm building in substrate is like an 474 00:29:00,830 --> 00:29:04,340 S1: intelligence collector, uh, which will allow me to do intelligence 475 00:29:04,340 --> 00:29:06,590 S1: reports like the one that I did before. And like this, 476 00:29:06,590 --> 00:29:11,900 S1: this guy is doing so. Yeah. Before. Before basically, this 477 00:29:11,900 --> 00:29:16,070 S1: was spy agencies, right? Journalists and spies were able to 478 00:29:16,070 --> 00:29:18,740 S1: do this because they had the money and they had 479 00:29:18,740 --> 00:29:24,050 S1: the training. Right. Well, I, I is cheap. It has 480 00:29:24,050 --> 00:29:26,209 S1: a lot of knowledge about how to do things. It 481 00:29:26,210 --> 00:29:28,880 S1: also has knowledge about how to turn those things into 482 00:29:28,880 --> 00:29:32,600 S1: a story and see patterns between them. So I'm going 483 00:29:32,600 --> 00:29:36,230 S1: to build something really nasty here. Really, really cool. Nasty. 484 00:29:36,230 --> 00:29:38,630 S1: I mean awesome. Uh, so so I can have like 485 00:29:38,630 --> 00:29:41,900 S1: a daily intelligence report of, like, here's what the smartest 486 00:29:41,900 --> 00:29:45,200 S1: people in the world think about this, and here's where 487 00:29:45,200 --> 00:29:50,060 S1: their ideas diverge from each other. Here's how their opinions 488 00:29:50,060 --> 00:29:54,800 S1: diverge from one another. Here's where their ideas point at 489 00:29:54,800 --> 00:29:57,170 S1: the exact same thing. And I could do like a 490 00:29:57,170 --> 00:30:00,140 S1: heat map of like, opinions and like, I can't wait 491 00:30:00,140 --> 00:30:03,380 S1: to mess with this thing. All right, intergular solar eclipse 492 00:30:03,380 --> 00:30:07,460 S1: photo competition. Look at this thing. This thing is insane. 493 00:30:08,230 --> 00:30:12,190 S1: Look at that. Look at this. That is so beautiful. Purple. 494 00:30:12,190 --> 00:30:14,950 S1: It's got the solar halo. And you got a plane. Oh. 495 00:30:14,950 --> 00:30:18,760 S1: Second place. Yeah. The plane one's way better. Sorry. Yeah. 496 00:30:18,760 --> 00:30:23,440 S1: Look at this. Purple. Oh, man. Got some, uh, vapor 497 00:30:23,440 --> 00:30:29,770 S1: trails there. Yeah. Very cool. All right. Ideas. Toddler intelligence 498 00:30:29,770 --> 00:30:34,580 S1: to PhD in around four years. That's what, uh, Mira 499 00:30:34,580 --> 00:30:38,990 S1: Murati just talked about on a podcast. Very interesting. I've 500 00:30:38,990 --> 00:30:41,870 S1: got this thing. Uh, just don't call it AI if 501 00:30:41,870 --> 00:30:44,030 S1: you're tired of hearing the word, I have a solution 502 00:30:44,030 --> 00:30:47,090 S1: for you. Forget. I forget the word. Forget everything about it. 503 00:30:47,090 --> 00:30:49,880 S1: And now just imagine the addition of hundreds of billions 504 00:30:49,880 --> 00:30:54,170 S1: of little helpers that people and companies can use to 505 00:30:54,170 --> 00:30:57,590 S1: do things like white collar work and writing and editing documents, 506 00:30:57,590 --> 00:31:01,880 S1: responding to emails, doing customer service, making sales calls, etc. 507 00:31:02,360 --> 00:31:05,930 S1: and taking ideas and turning those into full presentations, books 508 00:31:05,930 --> 00:31:08,900 S1: or films or whatever, and millions of other tasks that 509 00:31:08,900 --> 00:31:11,720 S1: otherwise would have not been done at all or would 510 00:31:11,720 --> 00:31:16,850 S1: have been done slowly or poorly by humans. And we're 511 00:31:16,850 --> 00:31:18,800 S1: about to have that. That's what we're about to have. 512 00:31:18,800 --> 00:31:22,700 S1: We're about to have billions of those things, doing tons 513 00:31:22,700 --> 00:31:27,770 S1: of stuff. I mean, just mountain loads of stuff that 514 00:31:27,770 --> 00:31:30,740 S1: has never been done before or was being done worse 515 00:31:30,740 --> 00:31:34,280 S1: by humans. Maybe we should call these things helpers, or 516 00:31:34,280 --> 00:31:37,520 S1: maybe we should call them whatever, but they're definitely going 517 00:31:37,520 --> 00:31:40,190 S1: to be very helpful, especially to employees who want to 518 00:31:40,190 --> 00:31:42,710 S1: do ten x the amount of work, or 100 x 519 00:31:42,710 --> 00:31:45,380 S1: the amount of work, or 1000 times the amount of 520 00:31:45,380 --> 00:31:48,260 S1: work that they were doing before, and have it be 521 00:31:48,260 --> 00:31:52,310 S1: done in a more consistent way or with higher quality. 522 00:31:52,310 --> 00:31:55,940 S1: And remember, the bar for human work is low. It 523 00:31:55,940 --> 00:31:58,220 S1: is very low. If you've ever worked anywhere, if you've 524 00:31:58,220 --> 00:32:02,300 S1: ever worked with anyone. We are fallible. We make mistakes, 525 00:32:02,300 --> 00:32:04,940 S1: we are inconsistent. And if you ask a whole bunch 526 00:32:04,940 --> 00:32:07,160 S1: of experts how to do something, or you have them 527 00:32:07,160 --> 00:32:10,520 S1: actually do something, the quality level will be vastly different 528 00:32:10,520 --> 00:32:14,600 S1: and the average worker is just not that good at 529 00:32:14,600 --> 00:32:18,200 S1: their job and they don't really care that much. I mean, 530 00:32:18,200 --> 00:32:21,470 S1: that's just a fact. Or I would say maybe not 531 00:32:21,470 --> 00:32:24,680 S1: the average one, but like right below the average and 532 00:32:24,680 --> 00:32:27,709 S1: there are millions upon millions of those workers. So we're 533 00:32:27,710 --> 00:32:32,780 S1: talking about replacing that with better versions, more scalable, more 534 00:32:32,780 --> 00:32:38,120 S1: consistent and more upgradable, like instantly upgradable. So that's what 535 00:32:38,120 --> 00:32:41,180 S1: we're about to have billions of those. But don't call 536 00:32:41,180 --> 00:32:45,380 S1: it AI. Nope, nope, don't call it AI. Because AI 537 00:32:45,380 --> 00:32:47,930 S1: is annoying. It's an annoying word, so just don't use 538 00:32:47,930 --> 00:32:50,780 S1: it and you'll be safe. All right. Discovery web check. 539 00:32:50,780 --> 00:32:53,930 S1: Free and open source tool. Oh, this thing is so cool. 540 00:32:54,170 --> 00:32:57,020 S1: This thing is so cool. Yeah, I was playing with 541 00:32:57,020 --> 00:32:59,930 S1: this earlier. Watch this. All right, so first of all, 542 00:32:59,930 --> 00:33:01,850 S1: you can download it and run it yourself. But they 543 00:33:01,850 --> 00:33:05,750 S1: also have a web version. So look at the website websites. Cool. 544 00:33:06,110 --> 00:33:11,120 S1: All right. Is this cool. This is cool. I mean, 545 00:33:11,210 --> 00:33:15,040 S1: is it valuable? I don't know. But I love it. 546 00:33:15,040 --> 00:33:17,380 S1: I mean, it's got so much value in here. It's 547 00:33:17,380 --> 00:33:19,780 S1: just I'm not sure how I'm going to use that value. 548 00:33:20,520 --> 00:33:24,790 S1: Probably parse the whole thing and ask AI to. Tell 549 00:33:24,790 --> 00:33:28,390 S1: me what I need to pay attention to and prioritize 550 00:33:28,390 --> 00:33:32,620 S1: it for me. Give me a turn into a task list. 551 00:33:32,620 --> 00:33:36,400 S1: That's what I probably do anyway. It's really cool. Hays 552 00:33:36,400 --> 00:33:42,520 S1: labs automated red teaming against Llms Agentic LLM Vulnerability Scanner 553 00:33:42,520 --> 00:33:45,820 S1: open source tool for fuzzing and stress testing Llms with 554 00:33:45,820 --> 00:33:50,470 S1: customizable rule sets Incogni. I use another one I think 555 00:33:50,470 --> 00:33:54,340 S1: I use delete me is very similar to this service 556 00:33:54,340 --> 00:33:56,830 S1: that removes your personal info from the web to block 557 00:33:56,830 --> 00:34:02,380 S1: spam calls and protect privacy. Confusion is a muse. This 558 00:34:02,380 --> 00:34:06,800 S1: is a really cool little article. Talk about. Yeah. You 559 00:34:06,800 --> 00:34:09,860 S1: don't want to get rid of confusion because it it 560 00:34:09,860 --> 00:34:14,210 S1: speaks to you, makes you smarter. Complexity expands to fill 561 00:34:14,210 --> 00:34:18,920 S1: the space available. Mhm. Love it. The winter of content 562 00:34:18,920 --> 00:34:24,080 S1: how Game of Thrones changed media driving traffic and homogenizing 563 00:34:24,080 --> 00:34:29,150 S1: journalism a nice salary database created from viral videos. The 564 00:34:29,150 --> 00:34:33,350 S1: internet is now for bots, not humans, and an impossibly 565 00:34:33,380 --> 00:34:37,670 S1: thin fabric can cool you down over 16 degrees. I 566 00:34:37,670 --> 00:34:40,370 S1: hesitate to show these things unless I can buy them, 567 00:34:40,370 --> 00:34:43,279 S1: but this one was cool enough. I made an exception 568 00:34:43,489 --> 00:34:46,160 S1: and the recommendation of the week. Get a copy of 569 00:34:46,160 --> 00:34:49,460 S1: Meditations by Marcus Aurelius and keep it by your bed 570 00:34:49,460 --> 00:34:53,180 S1: and read some before you go to sleep. Two major 571 00:34:53,180 --> 00:34:57,109 S1: benefits one. You'll fall asleep faster. Two you will fall 572 00:34:57,110 --> 00:35:01,460 S1: asleep feeling grateful for your life. I think that's what 573 00:35:01,460 --> 00:35:03,890 S1: stoicism does for me, and I hope it does it 574 00:35:03,890 --> 00:35:06,650 S1: for you as well. And the aphorism of the week 575 00:35:06,650 --> 00:35:11,810 S1: inspired by love, guided by knowledge, inspired by love, guided 576 00:35:11,810 --> 00:35:19,140 S1: by knowledge Bertrand Russell. Unsupervised Learning is produced and edited 577 00:35:19,140 --> 00:35:23,580 S1: by Daniel Meisler on a Neumann U87 AI microphone using Hindenburg. 578 00:35:24,060 --> 00:35:26,880 S1: Intro and outro music is by Zomby with the Y, 579 00:35:27,600 --> 00:35:29,850 S1: and to get the text and links from this episode, 580 00:35:29,850 --> 00:35:32,100 S1: sign up for the newsletter version of the show at 581 00:35:32,100 --> 00:35:34,739 S1: Daniel miessler.com/newsletter. 582 00:35:35,860 --> 00:35:36,850 UU: We'll see you next time.