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 --> 00:00:07,310 S1: podcast that looks at how best to thrive as humans 3 00:00:07,310 --> 00:00:11,540 S1: in a post AI world. It combines original ideas, analysis, 4 00:00:11,570 --> 00:00:14,780 S1: and mental models to bring not just the news, but 5 00:00:14,780 --> 00:00:22,279 S1: why it matters and how to respond. All right, welcome 6 00:00:22,280 --> 00:00:25,610 S1: to unsupervised learning. This is Daniel Miessler. Okay. What do 7 00:00:25,610 --> 00:00:29,450 S1: we have here? All right. This is, uh, when CSA 8 00:00:29,480 --> 00:00:31,460 S1: makes you release a patch for your end of life 9 00:00:31,460 --> 00:00:34,310 S1: product because the vulnerability is so bad. 10 00:00:34,340 --> 00:00:39,740 S2: I have abandoned my child. Say it. I abandoned my child. 11 00:00:39,770 --> 00:00:43,580 S2: Say it louder. Say it louder! I've abandoned my child. 12 00:00:43,610 --> 00:00:47,659 S2: I've abandoned my child. I've abandoned my boy. 13 00:00:48,409 --> 00:00:49,070 S3: That's so. 14 00:00:49,070 --> 00:00:53,930 S1: Good. Okay. Uh, fabric is now supporting OpenAI's new model 15 00:00:54,020 --> 00:00:57,650 S1: zero one preview. So you basically just pass it. The 16 00:00:57,830 --> 00:01:01,520 S1: switch R flag, which it sends the request using user 17 00:01:01,520 --> 00:01:06,169 S1: rather than system, because O1 preview is not allowing system access, 18 00:01:06,170 --> 00:01:08,600 S1: so you only have to send it in. User results 19 00:01:08,600 --> 00:01:11,780 S1: aren't that much different, honestly, and you also can't send 20 00:01:11,780 --> 00:01:15,860 S1: the temperature parameter which we were doing before. So this 21 00:01:15,860 --> 00:01:19,700 S1: raw option allows you to use O1 preview with fabric 22 00:01:19,700 --> 00:01:23,149 S1: and that works great. Got an insane cookbook use case 23 00:01:23,150 --> 00:01:27,020 S1: for O1 preview, so this person actually used it to. 24 00:01:27,050 --> 00:01:30,289 S1: I'm going to open this up to actually create synthetic 25 00:01:30,290 --> 00:01:34,610 S1: data and then validate whether the data was accurate or not. 26 00:01:34,640 --> 00:01:38,330 S1: Really really cool. Really really cool stuff. And just one 27 00:01:38,330 --> 00:01:40,190 S1: of the things that you could do for O1 preview 28 00:01:40,220 --> 00:01:43,880 S1: that you can't actually do with other models. Okay. So 29 00:01:43,910 --> 00:01:47,600 S1: let's see here. Back to kickboxing I'll be kickboxing today, 30 00:01:47,600 --> 00:01:51,110 S1: so hopefully it won't be dead later tonight. I got 31 00:01:51,110 --> 00:01:55,130 S1: a connect on LinkedIn button here in the newsletter. So 32 00:01:55,130 --> 00:01:57,710 S1: if you want to go and connect there, if you 33 00:01:57,740 --> 00:02:01,780 S1: connect with me I will mutual connect. Okay. Last week's 34 00:02:01,780 --> 00:02:06,310 S1: comments on current AI advances. So last podcast, if you're 35 00:02:06,310 --> 00:02:07,990 S1: listening to this one, you might have already listened to 36 00:02:08,020 --> 00:02:10,120 S1: that one, but I did a real deep dive on 37 00:02:10,120 --> 00:02:12,730 S1: like the current state of like the path to AGI. 38 00:02:12,760 --> 00:02:16,450 S1: I talked about the oh one preview. Um, and I 39 00:02:16,450 --> 00:02:18,610 S1: just did a whole bunch of analysis, but you should 40 00:02:18,610 --> 00:02:20,530 S1: check that out if you didn't listen to that one. 41 00:02:20,530 --> 00:02:23,560 S1: And I wrote a new essay called the Art Quality 42 00:02:23,590 --> 00:02:27,100 S1: Tier List. Pull this one up. And this is basically 43 00:02:27,100 --> 00:02:29,830 S1: a way of rating art. And keep in mind this 44 00:02:29,830 --> 00:02:32,080 S1: is for if you're a beginner, if you're already like 45 00:02:32,110 --> 00:02:35,710 S1: an advanced art person, you already know art back and forward, 46 00:02:35,740 --> 00:02:38,800 S1: then like you need to teach me stuff. I'm not 47 00:02:38,800 --> 00:02:42,490 S1: teaching you anything. This is for someone who is thoughtful 48 00:02:42,490 --> 00:02:45,130 S1: but is not an art expert. Even though I did 49 00:02:45,160 --> 00:02:48,370 S1: take an art history class in college. But anyway, I 50 00:02:48,370 --> 00:02:52,150 S1: don't really remember any of it, and I've basically just 51 00:02:52,150 --> 00:02:55,090 S1: not understood what art was for the longest time. And 52 00:02:55,090 --> 00:02:57,970 S1: I finally came up with a definition which I think 53 00:02:57,970 --> 00:03:01,630 S1: is pretty decent. It is indirect expression of something that 54 00:03:01,630 --> 00:03:06,400 S1: matters to humans. And I basically got a hierarchy here 55 00:03:06,400 --> 00:03:08,740 S1: of like the different ways you could look at it. 56 00:03:08,740 --> 00:03:11,980 S1: So going from not art to like proto art to 57 00:03:12,010 --> 00:03:15,190 S1: like hollow art, which is kind of like AI art 58 00:03:15,220 --> 00:03:19,570 S1: is one example where it might actually look pretty cool, 59 00:03:19,570 --> 00:03:22,359 S1: or it might actually make you feel a certain way, 60 00:03:22,360 --> 00:03:25,690 S1: but the artist didn't feel anything. And there was a 61 00:03:25,690 --> 00:03:28,570 S1: little bit of a caveat here. If the artist feels something, 62 00:03:28,570 --> 00:03:31,299 S1: but they're using AI to try to make the art, 63 00:03:31,360 --> 00:03:33,640 S1: bring it out. So there's a little bit of a 64 00:03:33,669 --> 00:03:37,180 S1: caveat there. And then decent art is where the artist 65 00:03:37,180 --> 00:03:41,440 S1: feels it and the audience feels it. Great art is like, yeah, 66 00:03:41,470 --> 00:03:45,190 S1: and people like it. And then S-tier or Brilliant is like, 67 00:03:45,220 --> 00:03:49,900 S1: it's recognized as like a world talent or world class 68 00:03:49,900 --> 00:03:53,710 S1: art and that's, that's basically it. So six different levels. 69 00:03:53,710 --> 00:03:58,390 S1: And I even have a methodology for enjoying art, and 70 00:03:58,390 --> 00:04:04,360 S1: this might be especially useful, especially useful for beginners. So 71 00:04:04,360 --> 00:04:07,810 S1: start by feeling the piece. No thinking, no analyzing. Just 72 00:04:07,810 --> 00:04:10,480 S1: take it in and experience how it makes you feel. 73 00:04:10,510 --> 00:04:13,990 S1: Do that for a number of seconds or minutes, however 74 00:04:13,990 --> 00:04:15,550 S1: long you want to take. Now that you have a 75 00:04:15,550 --> 00:04:18,010 S1: sense of how it affected you, analyze the message you 76 00:04:18,010 --> 00:04:20,380 S1: believe is trying to convey. Then, once you have the message, 77 00:04:20,410 --> 00:04:23,320 S1: think about how it was transmitted to you. Like was 78 00:04:23,320 --> 00:04:26,560 S1: it a song or was it paint or whatever it was? 79 00:04:26,560 --> 00:04:29,229 S1: And this one's really important. If you're with a friend, 80 00:04:29,260 --> 00:04:32,530 S1: do this in silence. Do one through three in silence 81 00:04:32,529 --> 00:04:34,839 S1: for however long you want to spend on it for 82 00:04:34,839 --> 00:04:38,049 S1: each piece, and then discuss your results with them. It's 83 00:04:38,050 --> 00:04:40,479 S1: a really fun way to do this, you know, get 84 00:04:40,480 --> 00:04:42,850 S1: a glass of wine or whatever it is. The first 85 00:04:42,850 --> 00:04:44,739 S1: time I ended up doing this, I was doing it 86 00:04:44,740 --> 00:04:49,390 S1: with my partner and with my friend Sasha, and we 87 00:04:49,390 --> 00:04:51,640 S1: were doing this, I think it was the the Big 88 00:04:51,640 --> 00:04:55,240 S1: Art Museum in New York, and we had lots of fun, like, 89 00:04:55,270 --> 00:04:58,440 S1: we all felt like novices, but we did something like 90 00:04:58,440 --> 00:05:00,870 S1: this methodology and it made it a lot of fun. Okay, 91 00:05:00,870 --> 00:05:04,290 S1: so that is that newsletter just went out, by the way, 92 00:05:04,290 --> 00:05:10,200 S1: and security. So US is evidently super reliant on Chinese cranes. 93 00:05:10,200 --> 00:05:15,570 S1: This is freaking me out, particularly from this one company, Zpmc. 94 00:05:15,570 --> 00:05:20,549 S1: And this report says 80% of US ship to shore 95 00:05:20,550 --> 00:05:23,760 S1: cranes are owned by this Chinese company. I thought it 96 00:05:23,760 --> 00:05:26,729 S1: was going to be like 25 or 50%. But yeah, 97 00:05:26,760 --> 00:05:28,890 S1: I hope someone is actually tracking this stuff in a 98 00:05:28,890 --> 00:05:31,500 S1: war room somewhere. And I'm pretty sure they are doing 99 00:05:31,529 --> 00:05:35,340 S1: doing war room exercises. Uh, Fortinet has confirmed a data 100 00:05:35,339 --> 00:05:38,219 S1: breach after someone called. I'm not going to say that word. 101 00:05:38,220 --> 00:05:42,120 S1: Something naughty claimed to have stolen 440 gigs off of 102 00:05:42,150 --> 00:05:45,089 S1: their SharePoint server. And they're talking about third party risk 103 00:05:45,089 --> 00:05:48,150 S1: as being the problem. Uh, GitLab release critical updates to 104 00:05:48,180 --> 00:05:52,440 S1: fix multiple vulnerabilities. Got 9.9 on the Richter scale here. 105 00:05:52,440 --> 00:05:58,380 S1: Enable remote PlayStation with minimal user interaction and low privileges. 106 00:05:58,410 --> 00:06:01,979 S1: Lazarus Group, which is North Korea, has been targeting Python 107 00:06:01,980 --> 00:06:06,270 S1: developers with a piece of malware disguised as coding tests. 108 00:06:06,300 --> 00:06:08,700 S1: They've been doing this for about a year now, thanks 109 00:06:08,700 --> 00:06:13,320 S1: to material for sponsoring. Mastercard is buying Recorded Future from 110 00:06:13,320 --> 00:06:18,180 S1: Insight Partners for $2.65 billion, and they bought it for 111 00:06:18,180 --> 00:06:24,180 S1: 780 million. So nice ROI on that one. And what 112 00:06:24,180 --> 00:06:27,029 S1: I noticed here is like this motion from okay, it's 113 00:06:27,029 --> 00:06:30,029 S1: a startup that does this piece of functionality. Now we're 114 00:06:30,029 --> 00:06:32,550 S1: going to bring it into a platform. In this case 115 00:06:32,550 --> 00:06:35,100 S1: the platform is like Mastercard. So you have good ideas 116 00:06:35,100 --> 00:06:38,339 S1: and execution. And their natural home is within some sort 117 00:06:38,339 --> 00:06:40,529 S1: of ecosystem. And I think this is a good way 118 00:06:40,529 --> 00:06:43,770 S1: to think of startups. It's like a petri dish for 119 00:06:43,770 --> 00:06:47,580 S1: features that will live inside of something bigger, a bigger platform, 120 00:06:47,580 --> 00:06:52,799 S1: bigger ecosystem or a framework. Security Canary Maturity Model is 121 00:06:52,800 --> 00:06:55,650 S1: a framework designed to help organizations assess and improve their 122 00:06:55,650 --> 00:06:59,430 S1: security posture by using Canary tokens. I love this concept 123 00:06:59,430 --> 00:07:02,910 S1: of detection maturity model. I was just talking to dropzone 124 00:07:02,910 --> 00:07:05,640 S1: about this actually. So here's like a percentage of your 125 00:07:05,640 --> 00:07:08,940 S1: most likely minor behaviors that you're likely to see based 126 00:07:08,940 --> 00:07:11,100 S1: on who you are. And you know, what percentage of 127 00:07:11,100 --> 00:07:13,860 S1: these can you detect. And thanks to Defend defy for 128 00:07:13,860 --> 00:07:18,000 S1: sponsoring as well. Australia is set to criminalise doxing with 129 00:07:18,000 --> 00:07:21,270 S1: penalties up to seven years in jail as part of 130 00:07:21,270 --> 00:07:26,130 S1: a new legislation aimed at modernizing the Privacy Act criminalizing 131 00:07:26,130 --> 00:07:30,180 S1: doxing seven years in jail. I like it. I like it. 132 00:07:30,210 --> 00:07:33,930 S1: I think, yeah, it does really come down to there 133 00:07:33,930 --> 00:07:36,540 S1: are times that this happens accidentally. In fact, last week 134 00:07:36,540 --> 00:07:39,120 S1: I was worried I accidentally did it, which I could 135 00:07:39,150 --> 00:07:42,030 S1: have just rerecorded, but I caught it and realised it 136 00:07:42,030 --> 00:07:44,220 S1: wasn't doxing so I didn't have to rerecord. But the 137 00:07:44,220 --> 00:07:47,610 S1: point is, hopefully they're doing it for purely malicious people, 138 00:07:47,610 --> 00:07:50,610 S1: which I'm sure they are. This piece discusses how AI 139 00:07:50,640 --> 00:07:54,739 S1: powered autonomous weapons are changing warfare. So a bunch of 140 00:07:54,740 --> 00:07:58,580 S1: Abrams tanks had to pull back after being targeted by 141 00:07:58,580 --> 00:08:03,500 S1: kamikaze drones run by the Russians. Yeah, definitely. Go read 142 00:08:03,500 --> 00:08:07,100 S1: kill decision by Daniel Suarez. This is the book. It 143 00:08:07,100 --> 00:08:10,700 S1: is extraordinary. Uh, what? Did I read this? I read 144 00:08:10,700 --> 00:08:15,410 S1: it in 2016. Yeah, really? Really good. Russia's naval activity 145 00:08:15,410 --> 00:08:19,850 S1: around undersea cables is freaking people out in the US. Uh, 146 00:08:19,850 --> 00:08:24,290 S1: they're worried that they're going to sabotage underwater infrastructure through 147 00:08:24,290 --> 00:08:29,330 S1: a unit called, uh, goochie. Goochie goochie something. So this 148 00:08:29,330 --> 00:08:34,309 S1: unit operates submarines, surface vessels and naval drones and has 149 00:08:34,309 --> 00:08:38,960 S1: been spotted near critical deep sea cables. And these carry 95% 150 00:08:38,960 --> 00:08:42,980 S1: of international data. US is drafting a New York joint 151 00:08:43,010 --> 00:08:49,430 S1: statement to bolster security of submarine communication cables. Focus on 152 00:08:49,429 --> 00:08:53,360 S1: excluding Chinese firms from the supply chain? Yeah. I think 153 00:08:53,360 --> 00:08:57,830 S1: we need comprehensive critical infrastructure dependency analysis. And like I 154 00:08:57,830 --> 00:09:00,050 S1: said before, I'm pretty sure this is already happening. I 155 00:09:00,050 --> 00:09:03,080 S1: just hope we're using really good red teamers to emulate 156 00:09:03,080 --> 00:09:07,310 S1: the capabilities of these adversaries. US House has voted to 157 00:09:07,309 --> 00:09:11,210 S1: block the purchase of new drones from DJI. So much 158 00:09:11,210 --> 00:09:14,720 S1: coverage of Counter China stuff lately. Seems like leadership is 159 00:09:14,720 --> 00:09:17,479 S1: getting the message here, which is great. State Department has 160 00:09:17,480 --> 00:09:21,680 S1: declared that Russia's state owned RT news has become a 161 00:09:21,679 --> 00:09:25,370 S1: key player in the Kremlin's military intelligence operations. I mean, 162 00:09:25,400 --> 00:09:28,100 S1: I remember a friend of mine talking about how awesome 163 00:09:28,340 --> 00:09:30,469 S1: he was because he was hearing stories he's never heard 164 00:09:30,470 --> 00:09:33,590 S1: anywhere before, and he's like, yeah, you know, the US 165 00:09:33,620 --> 00:09:36,620 S1: has really messed up. And I'm finding out how messed 166 00:09:36,620 --> 00:09:39,109 S1: up the US is from this channel. And I'm like, okay, 167 00:09:39,110 --> 00:09:42,080 S1: what's the channel? And it's RT and I'm like, dude, 168 00:09:42,080 --> 00:09:45,890 S1: you realize this is like, this is just a Russian 169 00:09:45,890 --> 00:09:49,939 S1: state media, media, right? This is like this is pretty 170 00:09:49,940 --> 00:09:52,700 S1: close to propaganda. And he started watching more and he's like, 171 00:09:52,730 --> 00:09:55,700 S1: oh yeah, it was absolutely. But he didn't know he 172 00:09:55,700 --> 00:09:58,490 S1: was being manipulated until it was called out to him. 173 00:09:58,520 --> 00:10:02,060 S1: And this is like so long ago, this, this might 174 00:10:02,090 --> 00:10:05,870 S1: have been like 2015 or something. All right. Siri flash. 175 00:10:05,900 --> 00:10:10,429 S1: Best Chronos Dave is a civilian radio enthusiast who's become 176 00:10:10,429 --> 00:10:14,030 S1: a key figure in Ukraine's drone defense strategy against Russia. 177 00:10:14,059 --> 00:10:18,650 S1: Operating from a mobile intelligence center in his Volkswagen van, 178 00:10:18,710 --> 00:10:22,520 S1: he monitors Russian radio transmissions and shares his findings with 179 00:10:22,520 --> 00:10:28,280 S1: over 127,000 followers, including soldiers and government officials. I'm not 180 00:10:28,280 --> 00:10:29,990 S1: sure I'd be giving this guy a lot of press. 181 00:10:29,990 --> 00:10:33,500 S1: It is basically putting a bounty on him. I and 182 00:10:33,500 --> 00:10:37,850 S1: tech newspaper had humans and AI create novel research ideas, 183 00:10:37,850 --> 00:10:41,390 S1: and then had human experts rate the ideas to see 184 00:10:41,390 --> 00:10:44,449 S1: which ones they liked, and they actually liked the AI 185 00:10:44,480 --> 00:10:47,300 S1: ideas better. And I think this is the way to 186 00:10:47,330 --> 00:10:51,520 S1: measure the abilities of AI, not with like standalone benchmarks, 187 00:10:51,520 --> 00:10:54,220 S1: but like what you do is you have I results, 188 00:10:54,220 --> 00:10:57,250 S1: you have human results and then you have a rating 189 00:10:57,250 --> 00:11:00,460 S1: system which is humans. And later we could use AI 190 00:11:00,490 --> 00:11:03,670 S1: for that. But in the meantime we have humans doing that. 191 00:11:03,670 --> 00:11:07,449 S1: And then we essentially say, which ones did you like better? 192 00:11:07,450 --> 00:11:09,459 S1: And they're blinded. Of course, they don't know which one 193 00:11:09,460 --> 00:11:13,000 S1: is which. So when they pick the AI ones, that's 194 00:11:13,000 --> 00:11:18,280 S1: a pretty good estimate estimation that the AI ones are better. 195 00:11:18,309 --> 00:11:20,020 S1: And I think that's how we should do a whole 196 00:11:20,020 --> 00:11:21,730 S1: bunch of these. Otherwise it's going to be a lot 197 00:11:21,730 --> 00:11:25,390 S1: of bias. OpenAI releases their new zero one preview model 198 00:11:25,390 --> 00:11:29,829 S1: focused on reasoning. We've talked a bunch about this one. Um, yeah. 199 00:11:29,860 --> 00:11:32,320 S1: Got my thoughts on it so far in the podcast. 200 00:11:32,320 --> 00:11:34,569 S1: And the big difference I just want to point this 201 00:11:34,570 --> 00:11:36,429 S1: one out. That's why I kept the story in here. 202 00:11:36,429 --> 00:11:40,090 S1: Big difference is that it uses chain of thought reasoning 203 00:11:40,090 --> 00:11:43,600 S1: and the fact that it actually spends time and actually tokens, 204 00:11:43,600 --> 00:11:48,460 S1: which means money thinking before responding. Because before it would 205 00:11:48,460 --> 00:11:50,410 S1: be tokens in and tokens out, and that was pretty 206 00:11:50,410 --> 00:11:54,640 S1: much it. Klarna's CEO says that I could replace enterprise. 207 00:11:54,670 --> 00:11:59,560 S1: Enterprise software giants like Salesforce and Workday claims that conversational AI, 208 00:11:59,590 --> 00:12:04,480 S1: like OpenAI's strawberry, can handle natural language commands to build 209 00:12:04,480 --> 00:12:08,230 S1: custom apps. 100%, this is what I've been saying Sbca 210 00:12:08,260 --> 00:12:11,860 S1: that's going to happen. And yeah, this is my Fpca piece, 211 00:12:11,860 --> 00:12:15,700 S1: and this is what I think future software looks like 212 00:12:15,730 --> 00:12:18,309 S1: you've got, uh, see, I can find the thing. There 213 00:12:18,309 --> 00:12:22,000 S1: we go. State policy questions, action. I think that is 214 00:12:22,000 --> 00:12:25,450 S1: the vibe. AI powered SAR satellites are now capable of 215 00:12:25,450 --> 00:12:29,140 S1: detecting aircraft from space due to new radar tech, which 216 00:12:29,170 --> 00:12:32,260 S1: enables real time monitoring of air traffic, good for both 217 00:12:32,260 --> 00:12:37,450 S1: civilian and military uses. Cardio tech? That's a good name. 218 00:12:37,630 --> 00:12:42,790 S1: Leveraging AI to tackle cardiovascular disease. Leading cause of death. 219 00:12:42,790 --> 00:12:47,650 S1: Partnering with 65 hospitals to build a massive human heart tissue. 220 00:12:47,679 --> 00:12:53,650 S1: Multi-omics data set to identify new drug candidates. Super exciting 221 00:12:53,650 --> 00:12:57,699 S1: because the AI needs data to form its model of 222 00:12:57,700 --> 00:13:00,100 S1: the world. And this is a whole bunch of that 223 00:13:00,100 --> 00:13:03,070 S1: data to model it, because then you could figure out 224 00:13:03,070 --> 00:13:05,800 S1: that's kind of the punchline here. You can figure out 225 00:13:05,800 --> 00:13:09,220 S1: if drugs might work or that's the plan anyway. If 226 00:13:09,220 --> 00:13:12,160 S1: you have a really good model of how the heart works, 227 00:13:12,160 --> 00:13:17,320 S1: you can have drugs, potentially synthetic testing of drugs to 228 00:13:17,350 --> 00:13:20,020 S1: see if they will work without doing trials or without 229 00:13:20,020 --> 00:13:23,830 S1: doing as many trials, or for as long or as expensive. 230 00:13:23,860 --> 00:13:27,160 S1: Salesforce just launched Agent Force, a suite of AI powered 231 00:13:27,160 --> 00:13:32,559 S1: agents designed to enhance human workers across various business functions. Yeah, yeah, 232 00:13:32,590 --> 00:13:36,640 S1: definitely designed to enhance human workers. Not designed at all 233 00:13:36,670 --> 00:13:40,630 S1: to replace them. Yep, yep. Waymo's latest data shows that 234 00:13:40,660 --> 00:13:44,920 S1: human drivers are responsible for most serious collisions involving its 235 00:13:44,920 --> 00:13:50,100 S1: driverless cars. 16 out of 23 severe crashes being rear 236 00:13:50,130 --> 00:13:54,780 S1: ended by human driven vehicles. Tesla Cybertruck is spiking in 237 00:13:54,780 --> 00:13:59,939 S1: the truck segment, while electric truck segment with 61% of 238 00:13:59,940 --> 00:14:05,850 S1: sales surge in July, a surge of 61% in July 239 00:14:05,880 --> 00:14:10,079 S1: did way better than the R1t's and the Ford F-150 lightning. 240 00:14:10,080 --> 00:14:12,270 S1: I think we didn't. We just see a story saying 241 00:14:12,270 --> 00:14:14,880 S1: the lightning was crushing it. I don't know why there's 242 00:14:14,880 --> 00:14:16,500 S1: a big jump right now, but I have seen a 243 00:14:16,500 --> 00:14:21,480 S1: whole bunch more very recently. Anecdotally, USPS has rolled out 244 00:14:21,480 --> 00:14:24,210 S1: a new delivery vehicle. Let's see if we can get 245 00:14:24,210 --> 00:14:26,490 S1: a picture of this thing, which I saw. Yeah, look 246 00:14:26,490 --> 00:14:29,490 S1: at that. Looks like a duck. Looks like a duck, Bill, 247 00:14:29,520 --> 00:14:32,730 S1: doesn't it? Yeah. Evidently the drivers love them. They're very, 248 00:14:32,730 --> 00:14:37,140 S1: very practical, but they look very strange. Dmitri Greenberg has 249 00:14:37,140 --> 00:14:40,080 S1: managed to run Linux and Ultrix on a business card, 250 00:14:40,110 --> 00:14:43,260 S1: turning it into a tiny computer, eight kilobytes of Ram 251 00:14:43,260 --> 00:14:48,420 S1: and 32kB of flash storage. New studies showing that debunk bot, 252 00:14:48,450 --> 00:14:54,060 S1: an AI chatbot can effectively persuade users to abandon conspiracy theories. 253 00:14:54,060 --> 00:14:58,710 S1: The bot made significant progress in changing people's beliefs, challenging 254 00:14:58,710 --> 00:15:03,240 S1: the notion that facts and logic cannot combat conspiracies. I mean, 255 00:15:03,270 --> 00:15:05,220 S1: this makes sense to me. So this is what I 256 00:15:05,220 --> 00:15:09,180 S1: wrote you. You can convince. If you can convince somebody 257 00:15:09,180 --> 00:15:11,609 S1: of something, you can convince them as well. This is 258 00:15:11,610 --> 00:15:15,600 S1: why I'm optimistic about having AI on us all the time, 100%. 259 00:15:15,600 --> 00:15:18,120 S1: It can be Orwellian, or it could be a defender 260 00:15:18,120 --> 00:15:21,870 S1: or protector or tutor or coach, etc. that's always with us. 261 00:15:21,900 --> 00:15:25,260 S1: Community college had to cancel its CS career fair because 262 00:15:25,260 --> 00:15:28,680 S1: no companies reached out. Super sad, but super expected. If 263 00:15:28,680 --> 00:15:31,320 S1: you have people coming out of college with four year 264 00:15:31,320 --> 00:15:34,590 S1: degrees and master's in CS and they can't find jobs, 265 00:15:34,590 --> 00:15:39,060 S1: what do you think? Do you think companies are rejecting? Right? 266 00:15:39,090 --> 00:15:41,430 S1: The companies who would come to the job fair, they're 267 00:15:41,580 --> 00:15:44,550 S1: rejecting the master's degree. People, but they're going to go 268 00:15:44,550 --> 00:15:46,950 S1: to community college. I don't think so. This is why 269 00:15:46,950 --> 00:15:50,970 S1: we need human 3.0. Future is connecting directly to individuals 270 00:15:50,970 --> 00:15:55,080 S1: not going through these credential forms. Google has officially killed 271 00:15:55,080 --> 00:15:58,020 S1: off cache links, which is sad because that's how I 272 00:15:58,020 --> 00:16:00,450 S1: used to see old versions of the site. And you 273 00:16:00,450 --> 00:16:03,360 S1: can see the site even if it was down, like, whatever. 274 00:16:03,390 --> 00:16:07,890 S1: United Airlines is partnering with SpaceX to bring free Starlink 275 00:16:07,890 --> 00:16:11,010 S1: Wi-Fi to all of its planes. I'm a united person. 276 00:16:11,010 --> 00:16:14,040 S1: I'm so happy that I'm a united person right now, man, 277 00:16:14,070 --> 00:16:17,460 S1: because I fly JSX as well and the Starlink is 278 00:16:17,460 --> 00:16:19,860 S1: so good. It's better than a lot of my friends. 279 00:16:19,860 --> 00:16:23,100 S1: Wi-Fi at their house. Ukraine just launched its biggest drone 280 00:16:23,100 --> 00:16:27,510 S1: attack on Moscow yet, hitting the region with 144 drones. 281 00:16:27,510 --> 00:16:30,720 S1: I still don't understand how Ukraine could possibly be winning 282 00:16:30,720 --> 00:16:33,810 S1: this war. Completely insane to me, but in a good way. 283 00:16:33,840 --> 00:16:37,650 S1: Sweden is increasing how much it's paying migrants to go home. 284 00:16:37,650 --> 00:16:41,760 S1: It's now up to $34,000, roughly paying them to just 285 00:16:41,760 --> 00:16:47,479 S1: leave NASA's advanced composite solar sail successfully been deployed. Its 286 00:16:47,480 --> 00:16:51,470 S1: ultrathin solar sail is now in low Earth orbit means 287 00:16:51,470 --> 00:16:54,200 S1: it's visible in the night sky from various locations. It 288 00:16:54,200 --> 00:16:58,040 S1: could be as bright as serious. Sirius is the brightest 289 00:16:58,040 --> 00:17:01,130 S1: star out there. The dog star. It is very bright, 290 00:17:01,280 --> 00:17:03,800 S1: so I can't wait to go find this thing. And 291 00:17:03,800 --> 00:17:05,840 S1: I wonder. I mean, it would be crazy to see 292 00:17:05,840 --> 00:17:08,780 S1: something as bright as Sirius actually moving, so I can't 293 00:17:08,780 --> 00:17:13,220 S1: wait to see this thing. Got a comment? They're calling 294 00:17:13,220 --> 00:17:16,970 S1: it the comet of the century. I feel like these astronomer, uh, 295 00:17:16,970 --> 00:17:20,540 S1: journalists are always saying it's the brightest. You will not 296 00:17:20,540 --> 00:17:24,920 S1: see another solar eclipse for 900 years or whatever. And 297 00:17:24,920 --> 00:17:28,940 S1: then it turns out, well, not one that comes directly here. 298 00:17:28,970 --> 00:17:31,820 S1: Not one that falls on a Friday. Not one that 299 00:17:31,820 --> 00:17:34,700 S1: you could see while standing upside down or whatever. And 300 00:17:34,700 --> 00:17:36,530 S1: then you look and you're like, okay, when is the 301 00:17:36,530 --> 00:17:40,580 S1: actual next eclipse? Oh, nine months from now or whatever. Anyway, 302 00:17:40,580 --> 00:17:41,899 S1: this is going to be a cool one. It's going 303 00:17:41,900 --> 00:17:45,109 S1: to be bright. So I'm going to try to remember 304 00:17:45,109 --> 00:17:47,090 S1: these times. I'm sure they'll say more about it at 305 00:17:47,090 --> 00:17:49,399 S1: the time so I can go take a look. US 306 00:17:49,430 --> 00:17:53,300 S1: is closing a trade loophole that e-commerce people like Timo 307 00:17:53,300 --> 00:17:58,580 S1: and Shine, shine, Shine have been exploiting. Loophole allows them 308 00:17:58,580 --> 00:18:02,600 S1: to ship goods directly to customers without paying tariffs. So 309 00:18:02,600 --> 00:18:04,700 S1: the US government is going after that. There's a leaked 310 00:18:04,700 --> 00:18:08,659 S1: PDF that details Mr. Beast's unique company culture, and it 311 00:18:08,660 --> 00:18:11,359 S1: got a one page summary of it here. Pretty, pretty 312 00:18:11,390 --> 00:18:14,899 S1: smart business stuff. Although he is, I think, being investigated 313 00:18:14,900 --> 00:18:18,920 S1: for some some stuff. See here. Uh, person says sunlight 314 00:18:18,950 --> 00:18:21,920 S1: cured their migraines. It's not a study. It's really just 315 00:18:21,920 --> 00:18:24,950 S1: somebody talking about it. But I figure most people who 316 00:18:24,950 --> 00:18:27,920 S1: have migraines have tried everything else. Why not try this? 317 00:18:27,950 --> 00:18:30,740 S1: It's basically sunlight getting lots of sunlight, I think in 318 00:18:30,740 --> 00:18:34,970 S1: the morning. Love this concept from Laura Hogan. Be a thermostat, 319 00:18:34,970 --> 00:18:38,480 S1: not a thermometer. That's that's cool. I mean, it's cool 320 00:18:38,480 --> 00:18:41,090 S1: if you don't even read the article. Be a thermostat, 321 00:18:41,119 --> 00:18:44,810 S1: not a thermometer. Content driven development is a strategy for 322 00:18:44,840 --> 00:18:48,590 S1: making progress on side projects by focusing on small, shareable 323 00:18:48,590 --> 00:18:50,929 S1: pieces of work so you finish a thing you share, 324 00:18:50,930 --> 00:18:53,930 S1: you finish the thing you share. In 1913, Vienna was 325 00:18:53,930 --> 00:18:58,550 S1: quite a place to hang out. Hitler, Leon, Trotsky, Tito, 326 00:18:58,550 --> 00:19:03,380 S1: Sigmund Freud, Joseph Stalin all hanging out in the same city. 327 00:19:03,380 --> 00:19:06,619 S1: And by the way, this is before Hitler. And I 328 00:19:06,650 --> 00:19:09,800 S1: guess Stalin turned evil. And whoever else in that list 329 00:19:09,800 --> 00:19:13,010 S1: was evil. I'm not sure how evil Trotsky was, because 330 00:19:13,010 --> 00:19:16,670 S1: that's how poorly read I am about Russian history. Discovery. 331 00:19:16,700 --> 00:19:19,790 S1: Merkle map CLI command line tool lets you search and 332 00:19:19,790 --> 00:19:25,790 S1: enumerate subdomains using the Merkle map API. Okay, 71 terabyte 333 00:19:25,820 --> 00:19:32,060 S1: zfs NAS 24 four terabyte drives. It has lasted without 334 00:19:32,060 --> 00:19:35,899 S1: a single drive failure for over a decade, thanks to 335 00:19:35,930 --> 00:19:39,290 S1: a strategy of keeping the server off when not in use. Cheating. 336 00:19:39,290 --> 00:19:43,450 S1: I feel like it's cheating. Doctor Mordechai Guri has unveiled 337 00:19:43,450 --> 00:19:47,470 S1: a new side channel attack called Rambo, which uses radio 338 00:19:47,470 --> 00:19:53,260 S1: signals from a device's Ram to exfiltrate data over Airgapped networks. 339 00:19:53,290 --> 00:19:56,140 S1: I was like, I know this is Israeli. I know 340 00:19:56,140 --> 00:19:59,470 S1: for sure, I guess Tel Aviv, but it was actually 341 00:19:59,470 --> 00:20:04,660 S1: a different Israeli university. But yes, it was Israeli. Israelis 342 00:20:04,660 --> 00:20:07,899 S1: are the side channel goats. Six techniques I use to 343 00:20:07,930 --> 00:20:10,960 S1: create a great user experience for shell scripts. This was 344 00:20:10,960 --> 00:20:13,960 S1: a great read. User friendly shell scripts with techniques like 345 00:20:13,960 --> 00:20:19,389 S1: comprehensive error handling, colorful output, and progress reporting. These things 346 00:20:19,390 --> 00:20:21,340 S1: really do make a difference. They make it feel like 347 00:20:21,340 --> 00:20:23,590 S1: an actual app. Two people made a tool that lets 348 00:20:23,590 --> 00:20:28,480 S1: you transfer playlists between Apple Music and Spotify and YouTube music. 349 00:20:28,510 --> 00:20:31,000 S1: Those are kind of the three big ones. Nothing is 350 00:20:31,000 --> 00:20:34,300 S1: a timer that celebrates the art of doing absolutely nothing. 351 00:20:34,300 --> 00:20:38,050 S1: And how to create a rag pipeline with pinecone and 352 00:20:38,050 --> 00:20:42,970 S1: Semantic image search CLI ideas. I love it when experts 353 00:20:42,970 --> 00:20:46,570 S1: completely disagree about a really important thing. Like for example, 354 00:20:46,570 --> 00:20:50,290 S1: I have no idea if China is thriving right now 355 00:20:50,290 --> 00:20:52,419 S1: and is about to take over the world, or if 356 00:20:52,420 --> 00:20:55,480 S1: they are about to have this demographic collapse and their 357 00:20:55,480 --> 00:20:58,929 S1: economy is tanking, and they've got the housing crisis and 358 00:20:58,930 --> 00:21:02,679 S1: they're just in major upheaval, and she is facing pressure 359 00:21:02,680 --> 00:21:05,950 S1: and time is running out. And they have brain drain 360 00:21:05,950 --> 00:21:08,710 S1: going to the west. And they're super screwed because they're 361 00:21:08,710 --> 00:21:11,020 S1: slowing down AI and they're going to get passed up. 362 00:21:11,020 --> 00:21:13,150 S1: I don't know if that's true or if they're about 363 00:21:13,150 --> 00:21:16,180 S1: to invent their own chips. They're about to not need Nvidia. 364 00:21:16,210 --> 00:21:19,000 S1: They're about to take over everything because the West is stupid. 365 00:21:19,000 --> 00:21:22,390 S1: And I could find the world's best experts who tell 366 00:21:22,420 --> 00:21:24,729 S1: me both of those things and tell me that the 367 00:21:24,730 --> 00:21:28,270 S1: other set of experts is completely wrong. I find that fascinating, 368 00:21:28,270 --> 00:21:31,210 S1: and I've read a whole bunch of books. I've read 369 00:21:31,210 --> 00:21:36,850 S1: probably 12 deep, like large books about China or dedicated 370 00:21:36,850 --> 00:21:40,060 S1: just to China and its development, probably the most important 371 00:21:40,060 --> 00:21:42,340 S1: one being like Chip Wars. Although that was more about 372 00:21:42,340 --> 00:21:44,770 S1: chips than it was about China. But it definitely talked 373 00:21:44,770 --> 00:21:46,510 S1: about China a lot. But I read a whole bunch 374 00:21:46,510 --> 00:21:49,449 S1: that were dedicated just to China. And I mean, it 375 00:21:49,450 --> 00:21:51,490 S1: didn't tell me the answer. I mean, I have all 376 00:21:51,490 --> 00:21:53,920 S1: these opinions and I listen to these opinions of, like, 377 00:21:53,950 --> 00:21:56,740 S1: the smartest people, but the fact that I still don't know, 378 00:21:56,740 --> 00:21:59,920 S1: after reading all those books and seeing all the data 379 00:21:59,920 --> 00:22:02,200 S1: and listening to all these smart experts and I still 380 00:22:02,200 --> 00:22:05,770 S1: don't know, that is exhilarating to me. I'm like, wow. 381 00:22:05,800 --> 00:22:09,220 S1: I mean, I love the idea of like and I'm 382 00:22:09,250 --> 00:22:13,150 S1: happy with myself for not like picking one and just saying, oh, 383 00:22:13,180 --> 00:22:15,220 S1: this is what I think is going to happen. I 384 00:22:15,220 --> 00:22:18,040 S1: tend it's like strong opinions, loosely held, you know, and 385 00:22:18,040 --> 00:22:22,000 S1: the recommendation of the week actively guard against age related 386 00:22:22,000 --> 00:22:24,850 S1: lock in, which starts around 30. I'm guessing I made 387 00:22:24,850 --> 00:22:27,550 S1: that up, but roughly around there, maybe even late 20s. 388 00:22:27,580 --> 00:22:29,980 S1: And for some people it doesn't kick in. I've been 389 00:22:29,980 --> 00:22:33,580 S1: very resistant to it, although I am writing this also 390 00:22:33,580 --> 00:22:37,530 S1: to myself, because you actively have to fight it. Listen 391 00:22:37,530 --> 00:22:41,250 S1: to new music. Read new books with new ideas in them. 392 00:22:41,280 --> 00:22:44,400 S1: Talk to new people. Go to strange restaurants. Try new foods. 393 00:22:44,400 --> 00:22:48,000 S1: Don't let your experiences reduce into a tighter and tighter 394 00:22:48,000 --> 00:22:51,359 S1: death spiral. Variation keeps your mind young and the aphorism 395 00:22:51,359 --> 00:22:54,060 S1: of the week. Choosing not to read great books has 396 00:22:54,060 --> 00:22:57,300 S1: the same effect as not being allowed to. Choosing not 397 00:22:57,300 --> 00:23:00,780 S1: to read great books has the same effect as not 398 00:23:00,780 --> 00:23:05,340 S1: being allowed to. Unsupervised learning is produced and edited by 399 00:23:05,369 --> 00:23:09,659 S1: Daniel Miessler on a Neumann U87 AI microphone using Hindenburg. 400 00:23:10,050 --> 00:23:13,020 S1: Intro and outro music is by zombie with a Y. 401 00:23:13,680 --> 00:23:15,899 S1: And to get the text and links from this episode, 402 00:23:15,900 --> 00:23:18,119 S1: sign up for the newsletter version of the show at 403 00:23:18,119 --> 00:23:22,920 S1: Daniel miessler.com/newsletter. We'll see you next time.