1 00:00:00,170 --> 00:00:02,810 S1: Two zero day exploits and supply chain attacks keep you 2 00:00:02,810 --> 00:00:05,540 S1: up at night. Worry no more. You can harden your 3 00:00:05,540 --> 00:00:10,010 S1: security with threatlocker. Worldwide, companies like JetBlue trust Threatlocker to 4 00:00:10,039 --> 00:00:13,069 S1: secure their data and keep their business operations flying high. 5 00:00:13,100 --> 00:00:16,670 S1: Threatlocker allows companies to take a proactive deny by default 6 00:00:16,670 --> 00:00:21,139 S1: approach to cybersecurity, blocking every action, process and user unless 7 00:00:21,140 --> 00:00:25,009 S1: specifically authorized by your team. Threatlocker helps you do this 8 00:00:25,010 --> 00:00:28,010 S1: and provides a full audit of every action allowed or 9 00:00:28,010 --> 00:00:32,720 S1: blocked for risk management and compliance. Onboarding and operation is 10 00:00:32,720 --> 00:00:35,989 S1: fully supported by their US based support team. Stop the 11 00:00:35,990 --> 00:00:39,710 S1: exploitation of trusted applications within your organization to keep you 12 00:00:39,710 --> 00:00:44,060 S1: running efficiently and secure. Protected from ransomware. To learn more 13 00:00:44,060 --> 00:00:47,629 S1: about how Threatlocker can mitigate unknown threats and ensure compliance 14 00:00:47,630 --> 00:00:57,290 S1: for your organization, visit threatlocker. Communal. That's Threatlocker communal. Welcome 15 00:00:57,290 --> 00:01:00,990 S1: to Unsupervised Learning, a security, I and meaning focused podcast 16 00:01:01,020 --> 00:01:03,300 S1: that looks at how best to thrive as humans in 17 00:01:03,300 --> 00:01:07,440 S1: a post AI world. It combines original ideas, analysis and 18 00:01:07,440 --> 00:01:10,650 S1: mental models to bring not just the news, but why 19 00:01:10,650 --> 00:01:17,970 S1: it matters and how to respond. All right, welcome to 20 00:01:17,970 --> 00:01:22,410 S1: unsupervised learning. This is Daniel. Okay. The first AI breaches. 21 00:01:22,410 --> 00:01:26,069 S1: What do we have here? Oh. Saw the comet last night. Yeah. 22 00:01:26,069 --> 00:01:27,809 S1: This is really cool. It's been out for a little while. 23 00:01:27,810 --> 00:01:30,750 S1: It's actually going to be out for a few more days, 24 00:01:30,750 --> 00:01:34,080 S1: and I think it's getting a little less bright, but. 25 00:01:34,080 --> 00:01:36,510 S1: And coming out a little bit later each day, I 26 00:01:36,510 --> 00:01:39,450 S1: think maybe like 10 or 15 minutes later each day. 27 00:01:39,600 --> 00:01:45,600 S1: But my girl took this very, very cool, very, very cool. Uh, 28 00:01:45,600 --> 00:01:49,560 S1: she actually spotted it with her, um, naked eyes at first. 29 00:01:49,590 --> 00:01:52,800 S1: She was the first one to see it, and then 30 00:01:52,800 --> 00:01:57,730 S1: we all started pointing cameras and binoculars at it, but 31 00:01:57,730 --> 00:02:02,410 S1: she actually captured this with, um, her iPhone 16 Pro, 32 00:02:02,410 --> 00:02:06,070 S1: so not bad at all. Yeah, basically, the iPhone catches 33 00:02:06,100 --> 00:02:09,520 S1: a lot more than the naked eye, and she used 34 00:02:09,520 --> 00:02:14,320 S1: light or night mode and also long exposure. And yeah, 35 00:02:14,350 --> 00:02:18,610 S1: it looks pretty good. Uh, it's an 80 000 year 36 00:02:18,940 --> 00:02:24,670 S1: orbit time for this particular comet. And, uh, the head there, 37 00:02:24,669 --> 00:02:29,739 S1: which was quite bright, is about two miles across, and 38 00:02:29,740 --> 00:02:34,089 S1: the tail is something like 18 million miles long. And 39 00:02:34,090 --> 00:02:37,720 S1: for reference, it's 8 million miles to the sun. So it's. 40 00:02:37,750 --> 00:02:41,290 S1: What is that? Almost three times two and a half times. 41 00:02:41,290 --> 00:02:43,989 S1: The tail is two and a half times the distance 42 00:02:43,990 --> 00:02:46,480 S1: from the Earth to the sun. And the head is 43 00:02:46,480 --> 00:02:50,079 S1: two miles wide. And it won't be back for another 44 00:02:50,080 --> 00:02:53,020 S1: 80,000 years. She's like, I guess we won't see it again. 45 00:02:53,020 --> 00:02:55,429 S1: I'm like, I don't know, eat your greens, gotta Good 46 00:02:55,460 --> 00:03:00,859 S1: exercise every day. You never know how long you can live. Okay, so, uh. Yeah. 47 00:03:00,889 --> 00:03:03,890 S1: Wrote a tutorial on how to use any hugging face 48 00:03:03,889 --> 00:03:06,590 S1: model with a llama. So I want to open this 49 00:03:06,590 --> 00:03:11,209 S1: up real quick. It's a super quick tutorial. I'm actually 50 00:03:11,240 --> 00:03:14,480 S1: I think I already recorded the video as well. Yeah, 51 00:03:14,480 --> 00:03:16,520 S1: I think I already recorded the video as well, so 52 00:03:16,520 --> 00:03:19,340 S1: it should be a video soon. But look at this. 53 00:03:19,370 --> 00:03:21,650 S1: It's super simple. You go to the models page on 54 00:03:21,680 --> 00:03:25,220 S1: hugging Face. You basically get the URL. You click on 55 00:03:25,220 --> 00:03:28,250 S1: this right here, files and versions, which you can see here. 56 00:03:28,250 --> 00:03:31,340 S1: You get one of these files like I got the 57 00:03:31,340 --> 00:03:35,570 S1: smaller one, the four gig one. You edit a model 58 00:03:35,570 --> 00:03:38,180 S1: file and this is all you have to put in there. 59 00:03:38,210 --> 00:03:42,110 S1: Look at this. It's insane. So you basically do um, 60 00:03:42,350 --> 00:03:47,270 S1: from is all you need actually is from. Okay. And 61 00:03:47,270 --> 00:03:50,360 S1: then you just put the path to the file and 62 00:03:50,360 --> 00:03:53,060 S1: then if you want to add a system, um, prompt 63 00:03:53,060 --> 00:03:56,680 S1: you can. So that's that's pretty much it. And you 64 00:03:56,680 --> 00:03:59,740 S1: don't need a system prompt. Okay. You can just use 65 00:03:59,740 --> 00:04:02,830 S1: the user prompt which is what you send into the model. 66 00:04:02,860 --> 00:04:07,030 S1: But then you do this this command here. So so 67 00:04:07,030 --> 00:04:09,730 S1: again we're trying to get a llama to run all 68 00:04:09,730 --> 00:04:13,360 S1: these models. So once you have this model file it's saved. 69 00:04:13,420 --> 00:04:16,599 S1: You just do this little command right here and then 70 00:04:16,600 --> 00:04:20,140 S1: this this name that you give it. Lexi writer or Lexi, 71 00:04:20,140 --> 00:04:23,140 S1: writer or writer or whatever you want to call it. 72 00:04:23,380 --> 00:04:27,159 S1: You just give it a llama create writer. Okay, so 73 00:04:27,160 --> 00:04:29,560 S1: you just made its own model. This takes like one 74 00:04:29,560 --> 00:04:32,410 S1: second to complete this little piece right here. And then 75 00:04:32,410 --> 00:04:35,380 S1: you run llama run in the name of the model 76 00:04:35,380 --> 00:04:39,490 S1: you created. Llama run. Writer. Llama run. Lexi. Writer, whatever 77 00:04:39,490 --> 00:04:42,370 S1: it is. And now you're actually talking to that model, 78 00:04:42,370 --> 00:04:45,190 S1: that hugging face model that you just downloaded. So this 79 00:04:45,190 --> 00:04:49,240 S1: entire process might be like three minutes. Five minutes, right? 80 00:04:49,270 --> 00:04:53,440 S1: Depends how long it takes you to download The the file. 81 00:04:53,440 --> 00:04:56,980 S1: So you just went from having like a couple of 82 00:04:56,980 --> 00:05:01,719 S1: dozen models available inside of a llama to having tens 83 00:05:01,720 --> 00:05:05,680 S1: of thousands or hundreds of thousands. So really, really cool. 84 00:05:05,680 --> 00:05:11,589 S1: And my buddy Marcus Hutchins and I disagree about Elon, 85 00:05:11,770 --> 00:05:14,710 S1: about a couple of things about his politics. Um, well, 86 00:05:14,710 --> 00:05:17,740 S1: we we agree on his politics. I just think Elon's 87 00:05:17,740 --> 00:05:19,599 S1: going to come back. I think he's still liberal. I'm 88 00:05:19,600 --> 00:05:23,320 S1: hoping he's still liberal. Um, I would say I'm hoping 89 00:05:23,320 --> 00:05:27,040 S1: he's liberal. Not a liberal. Not a current liberal. Um, 90 00:05:27,040 --> 00:05:31,060 S1: but I'm hoping he's more balanced. I believe he's more balanced. 91 00:05:31,060 --> 00:05:35,020 S1: And he's going to return. Marcus thinks he is Sith. 92 00:05:35,050 --> 00:05:38,920 S1: He is turned Sith. He is straight, far right, and 93 00:05:38,920 --> 00:05:42,700 S1: he's not coming back. And I want to say, I 94 00:05:42,700 --> 00:05:45,339 S1: do agree that that's possible and I'm worried about it, 95 00:05:45,339 --> 00:05:47,620 S1: but I'm hoping it's not the case. And I think 96 00:05:47,620 --> 00:05:50,260 S1: it's not the case. I think after a bunch of 97 00:05:50,260 --> 00:05:53,120 S1: this craziness blows over in the next year or two 98 00:05:53,150 --> 00:05:58,130 S1: that he will start showing his his more liberal and 99 00:05:58,130 --> 00:06:01,339 S1: balanced side, that that's my hope. He also believes that 100 00:06:01,370 --> 00:06:04,370 S1: Ellen is basically I don't want to put too many 101 00:06:04,370 --> 00:06:06,859 S1: words in his mouth, but I think he believes that 102 00:06:06,860 --> 00:06:10,460 S1: he's not a true innovator. He got lucky. A lot 103 00:06:10,460 --> 00:06:14,300 S1: of his stuff is garbage. Um, it's just like there's 104 00:06:14,330 --> 00:06:16,430 S1: there's no way any of this is going to go 105 00:06:16,460 --> 00:06:19,520 S1: well for him. Tesla is going to be basically bankrupt 106 00:06:19,520 --> 00:06:21,860 S1: within a year or two, I believe, he said. And 107 00:06:21,860 --> 00:06:23,630 S1: I've got a link to the whole conversation we had. 108 00:06:23,660 --> 00:06:26,000 S1: And we were we were texting on the side as 109 00:06:26,000 --> 00:06:29,900 S1: well during this. Um, and then basically I'm like, hey, look, 110 00:06:30,200 --> 00:06:33,919 S1: should we bet on this? Um, and so he offered 111 00:06:33,920 --> 00:06:37,250 S1: a couple of bets and I said, I need these 112 00:06:37,250 --> 00:06:41,000 S1: to be measurable. So I offered three measurable bets. One, 113 00:06:41,000 --> 00:06:44,450 S1: the Tesla stock would hit 250. Um, you know, it 114 00:06:44,450 --> 00:06:47,659 S1: would hit that point by June 30th. So middle of 115 00:06:47,660 --> 00:06:50,970 S1: next year it would hit 250. End of next year 116 00:06:50,970 --> 00:06:54,600 S1: it will have hit $300 at least once. It would 117 00:06:54,630 --> 00:06:57,000 S1: have got to that point. Um, and I actually think 118 00:06:57,000 --> 00:06:59,580 S1: it's going to be higher, but I just wanted to 119 00:06:59,580 --> 00:07:03,390 S1: have it 250 and 300. And then that Elon would, 120 00:07:03,390 --> 00:07:07,890 S1: in a liberal way, in a principled way that defends 121 00:07:07,890 --> 00:07:14,160 S1: something around anti-authoritarianism, something freedom oriented. In other words, Trump 122 00:07:14,160 --> 00:07:16,470 S1: would be on the wrong side of this and Elon 123 00:07:16,470 --> 00:07:18,660 S1: would be on the right side of it, the correct 124 00:07:18,660 --> 00:07:22,080 S1: side of it. And I argued that that would happen 125 00:07:22,080 --> 00:07:24,750 S1: by the end of next year. Okay. So that's three 126 00:07:24,750 --> 00:07:30,360 S1: different bets, each for $1,000. And uh, he accepted. So 127 00:07:30,570 --> 00:07:33,510 S1: we got three bets out there. Kind of reminds me 128 00:07:33,510 --> 00:07:36,180 S1: of this thing that I've been obsessed with lately, which 129 00:07:36,180 --> 00:07:40,470 S1: are these prediction markets. Which one is manifold? I forget 130 00:07:40,470 --> 00:07:43,680 S1: the name of the other one starts with a P, but, um, yeah, 131 00:07:43,710 --> 00:07:47,770 S1: pretty excited about these. I've barely started messing around with them, 132 00:07:47,770 --> 00:07:50,290 S1: but I feel like I want to use them more. Okay. 133 00:07:50,320 --> 00:07:53,290 S1: I did a talk for a UN group called Ypo, 134 00:07:53,320 --> 00:07:57,550 S1: which is a intellectual property, and it went really well. 135 00:07:57,550 --> 00:08:01,690 S1: And I want to thank Olivia Fabrizi, I believe, is 136 00:08:01,690 --> 00:08:04,390 S1: how you pronounce her last name for being not just 137 00:08:04,390 --> 00:08:07,090 S1: a great host, but actually, really, she has some really 138 00:08:07,090 --> 00:08:10,120 S1: great questions. She seems very interested in all this AI stuff, 139 00:08:10,120 --> 00:08:14,470 S1: very knowledgeable, and she was kind enough to host us 140 00:08:14,470 --> 00:08:19,360 S1: in this. And yeah, it was great. It was great conversation. 141 00:08:19,480 --> 00:08:21,940 S1: I listened back to the talk which which I hadn't 142 00:08:21,940 --> 00:08:24,670 S1: done before, and it seemed like it went pretty well 143 00:08:24,700 --> 00:08:26,980 S1: and good questions at the end. So I encourage you 144 00:08:26,980 --> 00:08:30,310 S1: to check that out. Thanks to Nudge Security for sponsoring 145 00:08:30,310 --> 00:08:36,339 S1: and let's get into security. Okay, so an attacker accessed moi. 146 00:08:36,610 --> 00:08:39,820 S1: I don't know how you pronounce that. Moi, moi. I 147 00:08:40,210 --> 00:08:42,010 S1: feel like you could have done a better job with 148 00:08:42,010 --> 00:08:47,819 S1: the name, but, um. AI chatbot database. Okay, so this 149 00:08:47,820 --> 00:08:49,590 S1: is I've been talking about this for a long time. 150 00:08:49,620 --> 00:08:51,780 S1: A lot of people who are anti I have been 151 00:08:51,780 --> 00:08:54,780 S1: talking about this, but I am obviously pro AI. But 152 00:08:54,780 --> 00:08:58,829 S1: I absolutely believe this is going to happen. It already 153 00:08:58,830 --> 00:09:01,949 S1: has happened and it's going to get worse. This AI 154 00:09:01,980 --> 00:09:06,839 S1: chatbot database is full of like basically AI girlfriends is 155 00:09:06,870 --> 00:09:10,740 S1: is the vibe that I'm getting. In fact, let's let's 156 00:09:10,740 --> 00:09:12,330 S1: go pull it up. What are we doing? This is 157 00:09:12,330 --> 00:09:15,540 S1: why we have video. All right. Select your login. Don't 158 00:09:15,540 --> 00:09:21,720 S1: think so. Okay. Yeah. Notice any, um, notice any trends here? 159 00:09:21,750 --> 00:09:24,090 S1: How many guys do we have here? Is this a guy? 160 00:09:24,120 --> 00:09:26,490 S1: I think the bottom left. We got a guy. Top 161 00:09:26,490 --> 00:09:29,579 S1: right or top left? We have a guy. Uh, I'm 162 00:09:29,580 --> 00:09:31,590 S1: not sure on this one. And then the rest are 163 00:09:31,590 --> 00:09:35,700 S1: all girls looking a little young. Honestly, uh, if I'm 164 00:09:35,700 --> 00:09:41,400 S1: being frank here, but anyway. Um. Avatars. Okay, so basically, 165 00:09:41,400 --> 00:09:46,449 S1: it's avatars. It's Companions. This whole scene is getting really 166 00:09:46,450 --> 00:09:50,860 S1: big an eye for lots of reasons. There's companionship, but obviously, 167 00:09:50,890 --> 00:09:55,900 S1: I mean, this whole girlfriend sex like fantasy angle is 168 00:09:55,929 --> 00:09:58,959 S1: is so obvious and it's so obvious that that's where 169 00:09:58,960 --> 00:10:01,210 S1: a lot of attention was going to go, both from 170 00:10:01,210 --> 00:10:03,610 S1: the user side but also on the builder side. And 171 00:10:03,610 --> 00:10:09,340 S1: so this company, their database got breached sensitive user interactions. 172 00:10:09,340 --> 00:10:12,550 S1: It had the chat history okay. So the user sign 173 00:10:12,580 --> 00:10:16,209 S1: up you look look at this. You sign up with 174 00:10:16,300 --> 00:10:21,520 S1: you look look it's got federated logins okay. These are 175 00:10:21,520 --> 00:10:24,790 S1: your actual user accounts okay. So people sign up with 176 00:10:24,790 --> 00:10:27,069 S1: their user accounts. Like if I click on this okay 177 00:10:27,100 --> 00:10:28,900 S1: I'm going to click on it I'm signing up with 178 00:10:28,900 --> 00:10:31,809 S1: my real stuff okay. If I sign up with this. 179 00:10:31,809 --> 00:10:34,930 S1: And then I have this chat with this, this avatar, 180 00:10:34,960 --> 00:10:38,709 S1: this AI over there and whatever I say on there. 181 00:10:38,800 --> 00:10:44,300 S1: And now that's what gets compromised. Okay. So including sexual fantasies, 182 00:10:44,300 --> 00:10:47,569 S1: and the user accounts were linked to people's personal addresses. 183 00:10:47,600 --> 00:10:49,459 S1: I mean, it's so obvious that this was going to 184 00:10:49,460 --> 00:10:52,130 S1: happen and it is now happening. It's going to be 185 00:10:52,160 --> 00:10:55,610 S1: it's going to be nasty, right? So input validation issues, 186 00:10:55,610 --> 00:10:58,220 S1: prompt injection, all of that. But the vast majority of 187 00:10:58,220 --> 00:11:00,530 S1: damage this is what I'm saying afterwards will come from 188 00:11:00,530 --> 00:11:04,969 S1: customers giving their souls essentially to small startups in the 189 00:11:04,970 --> 00:11:08,240 S1: AI assistant AI girlfriend space. This doesn't have to be 190 00:11:08,270 --> 00:11:11,179 S1: like a sexual fantasy thing. It could be trauma. It 191 00:11:11,179 --> 00:11:14,870 S1: could be somebody using it as there's other sites just 192 00:11:14,870 --> 00:11:17,750 S1: like this, but it's more therapy oriented. So it's like 193 00:11:17,780 --> 00:11:20,000 S1: you have a best friend to talk to you and 194 00:11:20,000 --> 00:11:22,730 S1: you're talking about your favorite show and blah, blah, blah, 195 00:11:22,730 --> 00:11:25,640 S1: and like, you develop this relationship with it, but you 196 00:11:25,640 --> 00:11:27,709 S1: might have been saying some really nasty stuff about your 197 00:11:27,710 --> 00:11:32,120 S1: parents or about whatever, maybe about your friends, right? Maybe 198 00:11:32,120 --> 00:11:34,580 S1: you're really sad and you think your friend stabbed you 199 00:11:34,580 --> 00:11:36,949 S1: in the back or whatever. So you're talking all this 200 00:11:36,950 --> 00:11:40,160 S1: crap about them and that gets released. I mean, we're 201 00:11:40,160 --> 00:11:42,420 S1: talking about intimacy here. It doesn't have to be of 202 00:11:42,420 --> 00:11:45,150 S1: a sexual, you know, sort of flavor. Right. So we're 203 00:11:45,150 --> 00:11:49,770 S1: talking about the the closest form of intimacy, which is 204 00:11:49,770 --> 00:11:52,110 S1: what you have with like this friend, you you can 205 00:11:52,110 --> 00:11:55,050 S1: tell anything and you can't tell anyone else. A lot 206 00:11:55,080 --> 00:11:56,850 S1: of people, they don't have any friends. So this is 207 00:11:56,850 --> 00:11:59,910 S1: like their only friend. So you give it your trauma, 208 00:11:59,940 --> 00:12:05,340 S1: your political opinions. Yeah. Fantasies, whatever. And they totally get you. 209 00:12:05,340 --> 00:12:08,790 S1: But this, it might be like a nine person startup. 210 00:12:08,790 --> 00:12:12,690 S1: They have no security. Right. I mean, startups get hacked 211 00:12:12,690 --> 00:12:15,540 S1: so easily. It's so easy to hack one of these 212 00:12:15,540 --> 00:12:17,940 S1: little startups. They have no security, they don't know how 213 00:12:17,940 --> 00:12:20,670 S1: to do security. And once they do, I mean, think 214 00:12:20,670 --> 00:12:24,510 S1: about the extortion campaigns here, right? Assuming the people have money, 215 00:12:24,510 --> 00:12:26,280 S1: which a lot some of them will and some of 216 00:12:26,280 --> 00:12:28,830 S1: them won't. But but I mean, you could just be like, hey, 217 00:12:28,830 --> 00:12:31,620 S1: here's a clip from this thing you were talking with 218 00:12:31,620 --> 00:12:34,410 S1: your avatar or whatever. You shared this trauma that you had. 219 00:12:34,410 --> 00:12:39,250 S1: You shared this talking crap about your friend. Um, I mean, 220 00:12:39,280 --> 00:12:43,630 S1: there's just so much possibility here for malicious use by 221 00:12:43,630 --> 00:12:46,660 S1: an attacker who gets hold of one of these databases. 222 00:12:46,660 --> 00:12:50,439 S1: So something to think about, Castillo says. A ransomware attack 223 00:12:50,440 --> 00:12:56,620 S1: led to a theft of sensitive data. I mean, 200GB. 224 00:12:56,620 --> 00:12:59,800 S1: They said it wasn't credit card info, but basically, Castillo 225 00:12:59,830 --> 00:13:05,440 S1: got breached. Miter introduced caldera bug bounty hunter plugin horizon three. 226 00:13:05,440 --> 00:13:08,949 S1: Researchers detailed how they identified new vulns in Palo Alto, 227 00:13:08,950 --> 00:13:12,490 S1: and they went all the way to full system compromise. 228 00:13:12,580 --> 00:13:17,230 S1: And there's patches out for that. Wayback machine got breached 229 00:13:17,230 --> 00:13:22,240 S1: six gigabytes of a SQL database, 31 million user records, 230 00:13:22,240 --> 00:13:24,880 S1: sites still down, but they're working to get it back up. 231 00:13:24,880 --> 00:13:30,760 S1: Researchers from Eset have discovered two sophisticated tool sets used 232 00:13:30,760 --> 00:13:35,710 S1: by nation state groups to go after airgapped devices. I 233 00:13:35,710 --> 00:13:38,480 S1: see Airgap, I think Israel, but they're saying this is 234 00:13:38,480 --> 00:13:43,880 S1: possibly Russian drop zone. Thanks again for sponsoring. Cybernews says 235 00:13:43,880 --> 00:13:47,660 S1: Google's Pixel nine Pro sends data packets to Google every 236 00:13:47,660 --> 00:13:51,620 S1: 15 minutes, including location, email and phone number, even with 237 00:13:51,620 --> 00:13:55,939 S1: GPS off. And they're using Wi-Fi to to estimate location. 238 00:13:55,940 --> 00:14:00,949 S1: UNODC warns that Southeast Asian scammers are using deepfakes to 239 00:14:00,980 --> 00:14:05,809 S1: enhance pig pig butchering. Yea, pig butchering these. These are nasty. 240 00:14:05,840 --> 00:14:10,309 S1: It's basically someone builds a relationship with you over a 241 00:14:10,309 --> 00:14:14,990 S1: period of time, and basically, especially for older people who 242 00:14:14,990 --> 00:14:19,070 S1: are lonely and who have money, obviously. Um, so basically 243 00:14:19,070 --> 00:14:23,030 S1: you become their friend, then this person gets in distress 244 00:14:23,030 --> 00:14:25,820 S1: and you're like, oh, I can send you money. And 245 00:14:25,820 --> 00:14:29,270 S1: that's kind of how it's done. Usually. Salt typhoon Chinese 246 00:14:29,270 --> 00:14:33,770 S1: group exploited backdoors meant for lawful data requests Ukraine has 247 00:14:33,770 --> 00:14:38,250 S1: sentenced two hackers linked to Russia's FSB and the Armageddon 248 00:14:38,250 --> 00:14:42,600 S1: Group to 15 years for cyber attacks on state institutions 249 00:14:42,600 --> 00:14:45,480 S1: in absentia. So they're not there. They don't have these people, 250 00:14:45,480 --> 00:14:49,830 S1: but they have called them out and sentenced them regardless. 251 00:14:49,860 --> 00:14:54,600 S1: OpenAI has stopped another 20 foreign operations using its stuff 252 00:14:54,600 --> 00:14:57,870 S1: to sway political opinions and meddle in elections. They said 253 00:14:58,140 --> 00:15:02,010 S1: they haven't seen any really nasty malware being created because 254 00:15:02,010 --> 00:15:04,050 S1: they have a lot of a lot of protections built 255 00:15:04,050 --> 00:15:07,140 S1: in for that. So that's that's good news. But still 256 00:15:07,170 --> 00:15:12,930 S1: being used for. Yeah, to try to influence campaigns. Basically 257 00:15:12,930 --> 00:15:17,100 S1: private intelligence groups like Recorded Future and Flashpoint are changing 258 00:15:17,100 --> 00:15:20,370 S1: intelligence by leveraging tons of data from the internet, including 259 00:15:20,370 --> 00:15:23,880 S1: dark web. And I love this. I love this vibe 260 00:15:23,880 --> 00:15:27,930 S1: of small companies competing with corporations and small Intel groups 261 00:15:27,930 --> 00:15:31,830 S1: competing with, like, you know, big Intel groups. In some ways, 262 00:15:31,830 --> 00:15:35,650 S1: study says popular car brands like Hyundai, Kia and Tesla 263 00:15:35,650 --> 00:15:39,640 S1: are collecting driver data, including voice recognition and camera footage, 264 00:15:39,640 --> 00:15:41,650 S1: and sharing it with third parties. I'm going to look 265 00:15:41,650 --> 00:15:45,370 S1: more into this because I'm curious how much Tesla is 266 00:15:45,370 --> 00:15:48,760 S1: sharing because I have a Tesla myself. Pentagon said the 267 00:15:48,760 --> 00:15:52,960 S1: US will send a third missile defense system to Israel, 268 00:15:52,960 --> 00:15:56,620 S1: along with 100 troops, just to operate it to improve 269 00:15:56,650 --> 00:16:01,300 S1: defense against Iran. Okay, so AI tech, if you use ChatGPT, 270 00:16:01,330 --> 00:16:03,970 S1: try this prompt just for fun. So this is a 271 00:16:03,970 --> 00:16:07,060 S1: meme or whatever that's going around. And essentially what it 272 00:16:07,060 --> 00:16:12,040 S1: is is it's saying go into ChatGPT and ask it 273 00:16:12,040 --> 00:16:14,260 S1: from all of our interactions together. What is the one 274 00:16:14,260 --> 00:16:16,570 S1: thing you could tell me about myself that I may 275 00:16:16,570 --> 00:16:18,790 S1: not know? And the idea is it's going to come 276 00:16:18,790 --> 00:16:21,250 S1: back with something and you're going to learn some insights 277 00:16:21,250 --> 00:16:23,500 S1: or whatever. And after it gives you an answer, you 278 00:16:23,500 --> 00:16:26,350 S1: ask it for another thing. What else may I not know? 279 00:16:26,350 --> 00:16:30,070 S1: And then what areas might I be holding myself back? 280 00:16:30,070 --> 00:16:32,810 S1: And here's what I got back. You seem to have 281 00:16:32,810 --> 00:16:36,860 S1: a constant drive to synthesize ideas across different domains, almost 282 00:16:36,860 --> 00:16:39,620 S1: like a mental architect. And as soon as I read, 283 00:16:39,620 --> 00:16:43,340 S1: like the first sentence, I'm like, uh, this feels gross. 284 00:16:43,370 --> 00:16:45,650 S1: I don't like it. And so this is what I 285 00:16:45,650 --> 00:16:50,000 S1: wrote about it in the newsletter yesterday. Basically, it feels 286 00:16:50,000 --> 00:16:54,110 S1: very complimentary and a bit like a horoscope. It feels 287 00:16:54,110 --> 00:16:56,420 S1: like a scam, right? Like it's just trying to make 288 00:16:56,420 --> 00:17:00,800 S1: me feel good about myself now. Simon. Um, I. I 289 00:17:00,800 --> 00:17:03,140 S1: can't remember his last name. Um, I always mess up 290 00:17:03,140 --> 00:17:05,540 S1: his last name. But Simon, he talks a lot about 291 00:17:05,570 --> 00:17:09,680 S1: AI stuff, and he basically had the same exact observation. 292 00:17:09,680 --> 00:17:13,520 S1: And he pointed out, in addition, that it's not actually 293 00:17:13,520 --> 00:17:18,530 S1: doing introspection on all your conversations. Okay, so so what 294 00:17:18,530 --> 00:17:20,030 S1: you actually have to do, if you want to do that, 295 00:17:20,030 --> 00:17:22,189 S1: you have to go get the conversations and send it 296 00:17:22,190 --> 00:17:26,630 S1: in this thing that it built here. I know specifically, 297 00:17:26,630 --> 00:17:29,180 S1: I knew immediately when it came back. It's working off 298 00:17:29,180 --> 00:17:32,100 S1: of my system prompt. So if you are a paid 299 00:17:32,100 --> 00:17:37,020 S1: member and you have anything inside of your your config 300 00:17:37,020 --> 00:17:41,520 S1: for ChatGPT, it's basically here's who I am. Here's how 301 00:17:41,520 --> 00:17:43,800 S1: I would like you to interact with me. Right? So 302 00:17:43,800 --> 00:17:47,340 S1: I have stuff in there that mentions human 3.0. It 303 00:17:47,340 --> 00:17:50,820 S1: mentions like the type of stuff that I'm interested in 304 00:17:50,820 --> 00:17:53,129 S1: because I'm using that to have it shape how it 305 00:17:53,130 --> 00:17:55,050 S1: responds to me when I ask it a question about 306 00:17:55,050 --> 00:17:58,080 S1: whatever I want it to answer me in a specific way. 307 00:17:58,080 --> 00:18:02,250 S1: So all of that is what it's basically going off of. 308 00:18:02,250 --> 00:18:05,670 S1: And it's essentially hallucination. It's making up all of this 309 00:18:05,670 --> 00:18:09,090 S1: stuff to make you feel good. And in fact, if 310 00:18:09,090 --> 00:18:12,720 S1: you go and ask now, um, I think they might 311 00:18:12,750 --> 00:18:16,170 S1: have modified the output and basically said that, no, we 312 00:18:16,170 --> 00:18:19,020 S1: can't give you deep analysis. We don't actually have the 313 00:18:19,020 --> 00:18:22,470 S1: history of all your conversations readily available. You would need 314 00:18:22,470 --> 00:18:24,629 S1: to put them in. It's giving some kind of a 315 00:18:24,630 --> 00:18:28,990 S1: warning to counteract people doing this meme and thinking the 316 00:18:28,990 --> 00:18:31,990 S1: wrong thing about it. But yeah, here's what I wrote 317 00:18:31,990 --> 00:18:35,920 S1: about it yesterday. It's like, well, everyone's going to agree. 318 00:18:35,920 --> 00:18:38,890 S1: If you know, something comes back and it's like, you know, 319 00:18:38,920 --> 00:18:41,109 S1: I just think you're so insightful and you're so smart 320 00:18:41,109 --> 00:18:43,330 S1: and you're trying to help the world and all this, 321 00:18:43,330 --> 00:18:47,290 S1: and it's like, oh, wow, that it's so accurate. It's 322 00:18:47,290 --> 00:18:49,989 S1: so accurate. It really gets me right. This is the 323 00:18:50,020 --> 00:18:53,109 S1: same trick that horoscope and palm reader people use against 324 00:18:53,109 --> 00:18:57,580 S1: you is to, like, just basically find special ways and 325 00:18:57,580 --> 00:18:59,920 S1: unique ways to compliment you. And it's like, oh, I've 326 00:18:59,920 --> 00:19:03,369 S1: never seen how awesome I was before. Thank you. So 327 00:19:03,400 --> 00:19:07,030 S1: I asked people, you know, curious what you get. And 328 00:19:07,030 --> 00:19:08,830 S1: I got back a whole bunch of responses and it 329 00:19:08,830 --> 00:19:11,230 S1: was basically the exact same thing. And a couple of 330 00:19:11,230 --> 00:19:13,480 S1: people responded and said, look, it's just going off the 331 00:19:13,480 --> 00:19:17,619 S1: system prompt. It's just hallucinating. So I think we all 332 00:19:17,650 --> 00:19:20,859 S1: sort of collectively figured out it was kind of lame. 333 00:19:20,890 --> 00:19:25,000 S1: Apple's AI researchers found that llms from meta and OpenAI 334 00:19:25,030 --> 00:19:28,310 S1: struggle with basic reasoning, and they've got a whole paper 335 00:19:28,310 --> 00:19:31,909 S1: basically going into that. I think it's more accurate to 336 00:19:31,940 --> 00:19:35,030 S1: say that it's easy to disrupt the reasoning than to 337 00:19:35,060 --> 00:19:38,510 S1: say it doesn't have any. Geoffrey Hinton, often dubbed godfather 338 00:19:38,510 --> 00:19:41,719 S1: of AI, has won the Nobel Prize in Physics for 339 00:19:41,720 --> 00:19:45,590 S1: early work in neural networks. Congrats to him. And one 340 00:19:45,590 --> 00:19:48,590 S1: thing to notice is that he is in the Doomer camp. 341 00:19:48,590 --> 00:19:54,050 S1: And anytime somebody gets a Nobel Prize for crazy work 342 00:19:54,050 --> 00:19:58,700 S1: in something like really groundbreaking work, you should listen, especially 343 00:19:58,700 --> 00:20:01,670 S1: if they disagree with you. He disagrees with me, so 344 00:20:01,670 --> 00:20:07,670 S1: I'm listening. Musk unveiled Tesla's new robotaxi self-driving electric vehicle 345 00:20:07,670 --> 00:20:11,270 S1: without a steering wheel or pedals at the We Robot event, 346 00:20:11,270 --> 00:20:14,510 S1: and there was so much hate about this. In the 347 00:20:14,510 --> 00:20:17,149 S1: first version that I recorded yesterday, I went into a 348 00:20:17,150 --> 00:20:20,810 S1: whole like political thing, which was too much. I don't 349 00:20:20,810 --> 00:20:22,399 S1: want to do that again here. So I'm going to 350 00:20:22,400 --> 00:20:25,590 S1: try to overview instead of it. I would say that 351 00:20:25,590 --> 00:20:31,680 S1: everyone or most people, most people are very anti like 352 00:20:31,680 --> 00:20:35,790 S1: militantly anti Elon or they think he's like the new 353 00:20:35,790 --> 00:20:38,820 S1: Jesus and I don't feel like there's very many people 354 00:20:38,820 --> 00:20:42,510 S1: who see the subtlety and nuance. And if you go 355 00:20:42,510 --> 00:20:45,360 S1: all the way down to the bottom here, I think 356 00:20:45,359 --> 00:20:49,980 S1: it's because not enough people are reading biographies, right? Biographies, 357 00:20:49,980 --> 00:20:53,820 S1: by definition, are written by or written about people who 358 00:20:53,820 --> 00:20:56,790 S1: have done extraordinary things. And if you go and look 359 00:20:56,820 --> 00:21:01,230 S1: at a whole bunch of inventors like Edison and Tesla 360 00:21:01,230 --> 00:21:05,190 S1: and people like that, um, they tend Ben Franklin, they 361 00:21:05,190 --> 00:21:09,360 S1: tend to be crazy and weird and, like, not take 362 00:21:09,390 --> 00:21:13,170 S1: showers and have weird social interactions and, like, sleep under 363 00:21:13,170 --> 00:21:19,140 S1: their desk and just smell, like, absolutely horrible and all 364 00:21:19,140 --> 00:21:21,210 S1: sorts of things like this. Right? They tend to be 365 00:21:21,210 --> 00:21:25,690 S1: very strange and oftentimes they do bad things, right? And 366 00:21:25,690 --> 00:21:28,600 S1: also they're down on themselves and they have these up 367 00:21:28,600 --> 00:21:32,649 S1: cycles like manic depressive type vibes, or they have a 368 00:21:32,650 --> 00:21:36,040 S1: Jesus complex, or they hate themselves and they think they're 369 00:21:36,040 --> 00:21:39,190 S1: not good enough and they're horribly jealous of everyone. Like 370 00:21:39,220 --> 00:21:45,730 S1: you're not finding horribly stable people, like very regular people, right, 371 00:21:45,760 --> 00:21:49,419 S1: who get biographies written about them. Usually not always the case, 372 00:21:49,420 --> 00:21:53,080 S1: but usually they tend to be kind of weird. And 373 00:21:53,109 --> 00:21:56,020 S1: I think that's important when you look at somebody like Elon, 374 00:21:56,020 --> 00:21:58,150 S1: because I feel like people are looking at it through 375 00:21:58,150 --> 00:22:03,310 S1: the lens of today and what's like considered normal on 376 00:22:03,310 --> 00:22:08,649 S1: X or in, you know, in California or specifically in 377 00:22:08,650 --> 00:22:13,239 S1: San Francisco, in California, like what is considered acceptable behavior. 378 00:22:13,270 --> 00:22:16,450 S1: And I think the question is, if somebody is an 379 00:22:16,450 --> 00:22:22,250 S1: extraordinary creator, which we also have to like, agree on that. 380 00:22:22,250 --> 00:22:25,220 S1: I would say he is, but let's say that you 381 00:22:25,220 --> 00:22:28,130 S1: agree he's an extraordinary creator and you're like, yeah, but 382 00:22:28,130 --> 00:22:31,190 S1: it doesn't matter because he has all these flaws. I 383 00:22:31,190 --> 00:22:35,030 S1: ask you again to look backward at all the different 384 00:22:35,030 --> 00:22:37,760 S1: things that have been created in the world, and what 385 00:22:37,760 --> 00:22:41,929 S1: percentage of those were created by essentially a crazy person, 386 00:22:41,960 --> 00:22:45,350 S1: crazy and or broken and or messed up and or 387 00:22:45,350 --> 00:22:50,270 S1: in need of therapy and or gross or in some way, 388 00:22:50,300 --> 00:22:54,830 S1: you know, counter normal. And I'm not saying excuses, right. 389 00:22:54,830 --> 00:22:56,659 S1: I'm not saying it excuses all the people in the 390 00:22:56,660 --> 00:23:00,440 S1: past either. I'm just saying reality tends to look like this. 391 00:23:00,440 --> 00:23:04,130 S1: So we should keep that in mind. All right. Deals, deals, 392 00:23:04,130 --> 00:23:07,610 S1: sales staff given two days notice to return to office 393 00:23:07,609 --> 00:23:11,359 S1: full time. So I've been talking about this quite a bit. Essentially, 394 00:23:11,359 --> 00:23:15,020 S1: return to office is a way of doing firing essentially 395 00:23:15,020 --> 00:23:20,060 S1: riffs right toe equals riff. Not always, but in many, 396 00:23:20,070 --> 00:23:23,429 S1: many cases. It's like it's a cheating way of doing it. 397 00:23:23,850 --> 00:23:31,680 S1: Co-Founder of Robinhood, Baiju is going to collect and transmit 398 00:23:31,680 --> 00:23:36,119 S1: solar energy using infrared lasers. And he's basically going to 399 00:23:36,150 --> 00:23:39,090 S1: I believe he's going to send the energy back down 400 00:23:39,090 --> 00:23:40,980 S1: to earth using a laser as well. So it'd have 401 00:23:40,980 --> 00:23:43,350 S1: to be like a receiver. And I'm like, well, you 402 00:23:43,350 --> 00:23:45,330 S1: need you would need to be able to target this 403 00:23:45,330 --> 00:23:49,560 S1: receiver while in space, moving very fast, right? Going around 404 00:23:49,560 --> 00:23:52,590 S1: the planet like every hour and a half. That's very, 405 00:23:52,590 --> 00:23:56,040 S1: very fast. But you'd still be able to hit these receivers. 406 00:23:56,070 --> 00:23:58,590 S1: Actually I think they would be geosynchronous. So you wouldn't 407 00:23:58,619 --> 00:24:02,850 S1: actually have to they wouldn't be moving in relation to 408 00:24:02,850 --> 00:24:07,920 S1: the receivers. Maybe not sure. Either way, you are technically 409 00:24:07,920 --> 00:24:11,580 S1: building a space laser, and if you're sending enough energy 410 00:24:11,580 --> 00:24:14,190 S1: to be useful on the ground, you're also sending enough 411 00:24:14,190 --> 00:24:17,460 S1: energy that could potentially be used as a weapon. So 412 00:24:17,460 --> 00:24:21,760 S1: just expect to have at least a questionnaire for the 413 00:24:21,760 --> 00:24:27,910 S1: military and or a request for purchase from the military. U.S. 414 00:24:27,910 --> 00:24:31,870 S1: Department of Justice considering breaking up Google after court said 415 00:24:31,869 --> 00:24:38,020 S1: they've crushed competition. Google Chrome, Android. Yeah, lots of different 416 00:24:38,020 --> 00:24:41,740 S1: products they have potentially. I agree with this. I agree 417 00:24:41,740 --> 00:24:44,560 S1: with Scott Galloway that a lot of these services need 418 00:24:44,590 --> 00:24:48,310 S1: to be broken up, potentially even Apple, even though I 419 00:24:48,310 --> 00:24:50,530 S1: love Apple and I don't think they've really done anything 420 00:24:50,530 --> 00:24:54,460 S1: wrong at some point. It's like a reward. It's basically 421 00:24:54,460 --> 00:24:57,940 S1: saying you are too awesome, therefore we have to break 422 00:24:57,940 --> 00:25:01,750 S1: you up and not necessarily like an indication that you're 423 00:25:01,750 --> 00:25:04,570 S1: a bad person. You know what I mean? Ticketmaster is 424 00:25:04,570 --> 00:25:08,200 S1: the first to use Apple's upgraded wallet tickets for iOS 18, 425 00:25:08,230 --> 00:25:13,540 S1: giving us stuff like venue, maps, parking, music playlists, weather forecasts, 426 00:25:13,570 --> 00:25:17,750 S1: stuff like that, stuff like that. So that's good. Anything 427 00:25:17,780 --> 00:25:20,959 S1: to make Ticketmaster suck less? That's what I wrote. New 428 00:25:20,960 --> 00:25:25,280 S1: HBO documentary claims Canadian crypto expert Peter Todd is the 429 00:25:25,280 --> 00:25:30,919 S1: mysterious inventor of Bitcoin. Satoshi Todd says nope, I was 430 00:25:30,920 --> 00:25:32,929 S1: too busy with school and work at the time, so 431 00:25:32,930 --> 00:25:35,330 S1: it couldn't have been me. Which is exactly what Satoshi 432 00:25:35,330 --> 00:25:40,310 S1: would say for Taiwanese employees at Foxconn's Zhengzhou plant. World's 433 00:25:40,310 --> 00:25:43,850 S1: largest iPhone production facility have been detained by Chinese authorities. 434 00:25:43,880 --> 00:25:47,600 S1: Nobody knows why they assume politics. Humans. Looks like Christopher 435 00:25:47,600 --> 00:25:52,790 S1: Columbus was a Sephardic Jew from Western Europe. I think 436 00:25:52,790 --> 00:25:56,119 S1: they're saying Spain can't remember if he was Spanish or 437 00:25:56,119 --> 00:25:58,699 S1: if he was Portuguese. I can't remember that whole vibe. 438 00:25:58,700 --> 00:26:02,540 S1: But they're saying, yeah, Spanish Jew. And I think Sephardic 439 00:26:02,540 --> 00:26:05,660 S1: means I can't remember exactly. I think it means like 440 00:26:05,690 --> 00:26:10,250 S1: Israeli Jew as opposed to European Jew, which is Ashkenazi. 441 00:26:10,369 --> 00:26:13,250 S1: I believe that's the case. JP Morgan and Wells Fargo 442 00:26:13,250 --> 00:26:17,800 S1: dips in profits. So they say it was geopolitical tension. 443 00:26:17,800 --> 00:26:21,310 S1: I saw Bank of America saying they had something similar happen. 444 00:26:21,310 --> 00:26:25,000 S1: So yeah, banks are sort of struggling. Although I did 445 00:26:25,000 --> 00:26:28,480 S1: see Goldman Sachs say they were doing okay, I believe 446 00:26:28,540 --> 00:26:32,290 S1: I can't remember the details, but some mixed results on 447 00:26:32,290 --> 00:26:34,540 S1: the banks. Your brain changes based on what you did 448 00:26:34,570 --> 00:26:36,850 S1: two weeks ago. So this is basically saying that it's 449 00:26:36,850 --> 00:26:39,280 S1: not just what you did yesterday or the day before, 450 00:26:39,280 --> 00:26:44,470 S1: but that there's like accumulated downsides of having crappy diet 451 00:26:44,470 --> 00:26:48,370 S1: and exercise and sleep over like the course of two weeks. 452 00:26:48,369 --> 00:26:51,580 S1: So pretty cool. I mean, essentially, to me this says 453 00:26:51,609 --> 00:26:55,750 S1: keep doing your routine because like it accumulates, the benefits accumulate. 454 00:26:55,780 --> 00:27:00,310 S1: American Heart Association outlines a strict protocol for taking blood pressure. 455 00:27:00,310 --> 00:27:03,460 S1: They're saying you have to sit calmly with an empty bladder, 456 00:27:03,460 --> 00:27:06,820 S1: and that too many people are measuring right after running 457 00:27:06,820 --> 00:27:11,290 S1: around or while you have to pee. And basically this 458 00:27:11,290 --> 00:27:15,740 S1: this is inflating Leading the hypertension numbers. Boeing is cutting 459 00:27:15,740 --> 00:27:19,790 S1: 17,000 jobs, 10% of its workforce. Federal emergency workers in 460 00:27:19,790 --> 00:27:25,280 S1: Rutherford County, North Carolina, were temporarily moved after reports of 461 00:27:25,280 --> 00:27:29,330 S1: an armed militia threatening government personnel, essentially because they think 462 00:27:29,330 --> 00:27:34,760 S1: they helped somehow create the hurricanes. I assume because they 463 00:27:34,760 --> 00:27:38,510 S1: work for the government. Elizabeth Landau says single cell cyanobacteria 464 00:27:38,540 --> 00:27:42,230 S1: can anticipate seasonal changes by sensing day length and preparing 465 00:27:42,230 --> 00:27:46,399 S1: for winter. Single cell organism can sense seasons. That trips 466 00:27:46,400 --> 00:27:50,209 S1: me out and kind of explains how trees start, you know, 467 00:27:50,240 --> 00:27:53,570 S1: turning fall and dropping leaves and all that. United Airlines 468 00:27:53,600 --> 00:27:58,369 S1: is adding new routes to lesser known destinations like Bilbao, Faro, Madeira, 469 00:27:58,400 --> 00:28:02,210 S1: Sicily and Newark trying to get away from crowded hotspots. 470 00:28:02,240 --> 00:28:05,480 S1: Essentially got some journals from Alexei Navalny while he was 471 00:28:05,480 --> 00:28:08,780 S1: in prison, for he died in prison. Retail sales have 472 00:28:08,780 --> 00:28:14,200 S1: dropped from 7.5 to 7% of employment over the last decade, 473 00:28:14,200 --> 00:28:21,070 S1: losing 850,000 positions despite the US adding 19 million jobs overall. 474 00:28:21,100 --> 00:28:24,100 S1: Looks like Wegovy and Zepp bound might have been the 475 00:28:24,100 --> 00:28:28,150 S1: cause of US adult obesity rate, dropping by two percentage 476 00:28:28,150 --> 00:28:33,130 S1: points from 2020 to 2023. And speaking of that, new 477 00:28:33,160 --> 00:28:36,909 S1: GLP one weight loss drugs are coming in pill form 478 00:28:36,910 --> 00:28:42,580 S1: soon to basically replace the weekly injections. Dhairya Karwa Mirza, 479 00:28:42,610 --> 00:28:49,000 S1: self-taught Kurdish astrophotographer. Look at this image that she took. Extraordinary. 480 00:28:49,030 --> 00:28:52,180 S1: Absolutely extraordinary. Look at that. Isn't that cool? Looks kind 481 00:28:52,180 --> 00:28:55,450 S1: of oblong. Not sure why. Maybe the shadow. Anyway, very 482 00:28:55,450 --> 00:29:00,250 S1: cool ideas. Gullibility. Not disinformation. So I don't think the 483 00:29:00,280 --> 00:29:03,160 S1: US has a misinformation problem. I think it has a 484 00:29:03,190 --> 00:29:06,490 S1: gullibility problem. It's not that we're being fed too much crap, 485 00:29:06,490 --> 00:29:09,070 S1: it's the fact that we're actually eating it. Some two 486 00:29:09,070 --> 00:29:11,980 S1: large number of Republicans now believe that the Democrats are 487 00:29:11,980 --> 00:29:14,860 S1: sending hurricanes to Florida because it's election time. That's a 488 00:29:14,860 --> 00:29:19,570 S1: population problem, an education problem, not a conspiracy theory problem. 489 00:29:19,570 --> 00:29:22,990 S1: So in infosec terms, we need to reduce our vulnerability, 490 00:29:23,020 --> 00:29:26,230 S1: not try to get all the threats removed. Right. Because 491 00:29:26,230 --> 00:29:28,090 S1: the threats will always be there. And in fact, the 492 00:29:28,090 --> 00:29:30,040 S1: threats are going to get better. The only chance of 493 00:29:30,040 --> 00:29:34,690 S1: actually fixing this is education about how the world actually works. 494 00:29:34,720 --> 00:29:37,930 S1: And unfortunately, the far left and the far right seem 495 00:29:37,930 --> 00:29:41,140 S1: to have like lost touch with that, like the basic 496 00:29:41,140 --> 00:29:43,990 S1: facts of reality, because it's not just the far right 497 00:29:43,990 --> 00:29:47,560 S1: that's doing this, right. Remember the anti-vax thing that was 498 00:29:47,590 --> 00:29:50,860 S1: far left before it was far right? It was far left. 499 00:29:50,860 --> 00:29:53,320 S1: And then Covid happened and then the far right jumped on. 500 00:29:53,320 --> 00:29:55,330 S1: And I guess now the far left and the far 501 00:29:55,330 --> 00:29:59,770 S1: right are anti-vax. Not really sure, but the point is 502 00:29:59,770 --> 00:30:03,310 S1: we have a vulnerability, which is a lack of education 503 00:30:03,310 --> 00:30:06,100 S1: and a lack of understanding of how the world works. 504 00:30:06,100 --> 00:30:10,190 S1: That's the vulnerability. Threats are things like misinformation. We can 505 00:30:10,190 --> 00:30:13,670 S1: have a threat and not be vulnerable, and we still 506 00:30:13,670 --> 00:30:18,170 S1: have lower risk. So that's the vibe there. Discovery swarm, 507 00:30:18,170 --> 00:30:23,510 S1: OpenAI's new experimental framework for building and orchestrating multi-agent systems, 508 00:30:23,540 --> 00:30:26,090 S1: feels a lot like long chain to me. I didn't 509 00:30:26,090 --> 00:30:29,630 S1: see so many benefits. I'm going to mess with it. 510 00:30:29,630 --> 00:30:33,170 S1: More command tools. I like really good set of tools here. 511 00:30:33,200 --> 00:30:40,400 S1: RFT Delta tldr s-oxide http I fcf. Yeah, lots of 512 00:30:40,400 --> 00:30:44,600 S1: different tools here Xvm better vim mode for ssh. I'm 513 00:30:44,600 --> 00:30:49,160 S1: sorry for zsh. Love this thing. It's syntax highlighting. It's 514 00:30:49,160 --> 00:30:53,810 S1: basically a full featured version of vim on the command line, 515 00:30:53,810 --> 00:30:57,980 S1: so I use vim mode for my zsh have been 516 00:30:57,980 --> 00:31:00,980 S1: for years, but this gives you like syntax highlighting. It 517 00:31:00,980 --> 00:31:03,560 S1: gives you the ability to use like the surround plugin, 518 00:31:03,560 --> 00:31:08,610 S1: all kinds of stuff. Theano 3.0 AI powered API documentation tool. 519 00:31:08,610 --> 00:31:12,510 S1: I updated my post on dynamic content generation. I think 520 00:31:12,510 --> 00:31:17,520 S1: this is going to be insanely disruptive to so many industries. 521 00:31:17,550 --> 00:31:20,220 S1: Highly recommend you go check that out. Essentially, what it 522 00:31:20,220 --> 00:31:23,790 S1: means is people are going to produce raw content. Like 523 00:31:23,790 --> 00:31:27,780 S1: for example, this is raw content of a full podcast, right? 524 00:31:27,810 --> 00:31:31,350 S1: And this has got me in this bottom right here 525 00:31:31,350 --> 00:31:34,230 S1: talking to the camera. And let's say this, this whole 526 00:31:34,230 --> 00:31:38,280 S1: episode is like 20 minutes long. Well, let's say that, um, 527 00:31:38,310 --> 00:31:43,350 S1: you have a, like a five minute bus ride or 528 00:31:43,350 --> 00:31:45,510 S1: a five minute bike ride, or you're waiting in line 529 00:31:45,510 --> 00:31:48,330 S1: for five minutes or whatever. And and this is like, 530 00:31:48,360 --> 00:31:50,670 S1: I don't know, three, five, five years in the future. 531 00:31:50,670 --> 00:31:53,250 S1: And you're like, hey, I want a summary. What did 532 00:31:53,250 --> 00:31:56,010 S1: Daniel put in his podcast yesterday? Your AI is not 533 00:31:56,010 --> 00:31:58,650 S1: going to go and find the video and speed it up. 534 00:31:58,650 --> 00:32:02,220 S1: In fact, your AI is just going to read, watch, 535 00:32:02,250 --> 00:32:05,720 S1: consume the whole video if it hasn't already, and it's 536 00:32:05,720 --> 00:32:09,530 S1: going to make a version of it. Maybe it's my face, 537 00:32:09,530 --> 00:32:12,290 S1: maybe it's me in a video. Maybe it's only audio 538 00:32:12,290 --> 00:32:15,050 S1: because they can actually look at anything right now. But anyway, 539 00:32:15,050 --> 00:32:17,990 S1: it's going to play in your ear a five minute version. 540 00:32:17,990 --> 00:32:20,120 S1: It's going to pick the best stuff out of here 541 00:32:20,120 --> 00:32:22,010 S1: that you actually care about. How is it going to 542 00:32:22,010 --> 00:32:25,400 S1: do that? Because your eye knows you. Your eye knows 543 00:32:25,400 --> 00:32:27,830 S1: what parts of these which which of these stories are 544 00:32:27,830 --> 00:32:31,070 S1: actually good or most pertinent to you. It's going to 545 00:32:31,070 --> 00:32:33,770 S1: take out the garbage. It doesn't like those stories, thinks 546 00:32:33,770 --> 00:32:37,100 S1: they're dumb, thinks it was filler or whatever, and it's 547 00:32:37,100 --> 00:32:39,890 S1: going to build a clean little clip. It could be 548 00:32:39,890 --> 00:32:42,320 S1: a 32nd clip, but it could be read in the 549 00:32:42,320 --> 00:32:46,490 S1: voice of Richard Feynman. It could be read in the 550 00:32:46,490 --> 00:32:49,790 S1: voice of Agatha Christie. It could be it could be anyone. 551 00:32:49,790 --> 00:32:53,150 S1: It could be Taylor Swift. Right. Telling you the news 552 00:32:53,150 --> 00:32:57,770 S1: that that came out of the Unsupervised Learning podcast. Point is, 553 00:32:57,770 --> 00:33:01,400 S1: your eye is going to be giving you content in 554 00:33:01,400 --> 00:33:05,840 S1: any format, in any presenter deepfaked exactly the way you 555 00:33:05,840 --> 00:33:12,440 S1: want it from 10s to giant series. Okay, the original content, 556 00:33:12,440 --> 00:33:15,710 S1: which lives in a textbook somewhere, or in some YouTube 557 00:33:15,710 --> 00:33:19,670 S1: video or in some blog post. That original content you 558 00:33:19,670 --> 00:33:22,700 S1: never have to mess with again. Your AI is going 559 00:33:22,730 --> 00:33:25,610 S1: to consume all of that and then dynamically build the 560 00:33:25,610 --> 00:33:29,180 S1: content for their principal. So that's the idea there. Augment 561 00:33:29,180 --> 00:33:33,320 S1: UI use AI to prototype front end designs. Software engineer 562 00:33:33,350 --> 00:33:36,950 S1: pay heatmap across the US, the digits of pi are 563 00:33:36,950 --> 00:33:40,160 S1: not random. Passbook lets you create an Apple Wallet, pass 564 00:33:40,160 --> 00:33:43,100 S1: from any QR code and export it to wallet. How 565 00:33:43,100 --> 00:33:46,250 S1: I animate three blue one Brown. It's one of my 566 00:33:46,250 --> 00:33:49,940 S1: favorite YouTubers, describes a whole bunch of science stuff and 567 00:33:49,940 --> 00:33:52,940 S1: recommendation of the week. If you want to calm your 568 00:33:52,940 --> 00:33:56,060 S1: nerves during this next month and a half going into 569 00:33:56,060 --> 00:34:00,200 S1: this election, go read about the Civil Rights movement and 570 00:34:00,200 --> 00:34:03,810 S1: how disruptive that was to politics in the US. And 571 00:34:03,810 --> 00:34:06,390 S1: it was actually, in a lot of ways, way worse 572 00:34:06,390 --> 00:34:10,080 S1: than what we're seeing today. So we survived that really 573 00:34:10,080 --> 00:34:16,020 S1: bad stuff. We probably will again survive this election cycle, probably. 574 00:34:16,020 --> 00:34:18,360 S1: And the aphorism of the week, what is to give 575 00:34:18,390 --> 00:34:23,670 S1: light must endure burning. What is to give light must 576 00:34:23,700 --> 00:34:29,279 S1: endure burning Viktor Frankl. Unsupervised learning is produced and edited 577 00:34:29,280 --> 00:34:33,750 S1: by Daniel Miessler on a Neumann U87 AI microphone using Hindenburg. 578 00:34:34,170 --> 00:34:37,110 S1: Intro and outro music is by zombie with a Y. 579 00:34:37,710 --> 00:34:39,989 S1: And to get the text and links from this episode, 580 00:34:39,989 --> 00:34:42,240 S1: sign up for the newsletter version of the show at 581 00:34:42,239 --> 00:34:47,010 S1: Daniel miessler.com/newsletter. We'll see you next time.