WEBVTT - UL NO. 441: Substrate, OpenAIs AGI Levels, US Literacy Rates 0:00:00.020 --> 0:00:02.600 Are you prepared for whatever shitstorm may hit your desk 0:00:02.600 --> 0:00:05.360 during the workday? Otto Max has your back. Check out 0:00:05.360 --> 0:00:08.690 the brand new autonomous IT podcast. Listen in as various 0:00:08.690 --> 0:00:12.559 IT experts discuss the latest Patch Tuesday releases, mitigation tips, 0:00:12.560 --> 0:00:16.130 and custom automations to help with CVE remediations make new 0:00:16.130 --> 0:00:20.630 work friends. Listen now to the autonomous IT podcast on Spotify, Apple, 0:00:20.630 --> 0:00:25.310 or wherever you tune in to podcasts. Welcome to Unsupervised Learning, 0:00:25.310 --> 0:00:28.310 a security, AI, and meaning focused podcast that looks at 0:00:28.310 --> 0:00:31.310 how best to thrive as humans in a post AI world. 0:00:31.910 --> 0:00:35.900 It combines original ideas, analysis, and mental models to bring 0:00:35.900 --> 0:00:38.450 not just the news, but why it matters and how 0:00:38.450 --> 0:00:46.280 to respond. All right, welcome to unsupervised learning. This is 0:00:46.280 --> 0:00:51.350 Daniel Miessler. All right. Episode 441 441 of these. That's 0:00:51.350 --> 0:00:54.430 pretty insane. Okay. Friend of UL Ray ulnar is looking 0:00:54.430 --> 0:00:57.400 for a new position as a systems engineer. You should 0:00:57.400 --> 0:01:01.570 reach out to Ray directly and got his LinkedIn there. 0:01:01.570 --> 0:01:06.459 He's one of the best members of UL and, uh, super, 0:01:06.459 --> 0:01:09.460 super smart, super kind and very competent as well as 0:01:09.459 --> 0:01:12.280 you could see from his LinkedIn. And, uh, he just 0:01:12.280 --> 0:01:15.160 came onto the market. So I would grab him before 0:01:15.160 --> 0:01:19.060 he goes, added a couple of new levels to the 0:01:19.060 --> 0:01:22.690 AGI definitions. So if we click through there and we 0:01:22.690 --> 0:01:26.559 go down here, here are the definitions or the levels. 0:01:26.560 --> 0:01:30.640 So AGI one is better but with significant drawbacks. Two 0:01:30.670 --> 0:01:36.070 is competent but imperfect. Three is a full $80,000 a 0:01:36.069 --> 0:01:41.050 year white collar worker replacement four okay, this is where 0:01:41.050 --> 0:01:44.620 it gets cool. for is a world class employee such 0:01:44.620 --> 0:01:49.420 as Andre Karpathy or Jeff Dean. So imagine 1% of 1% 0:01:49.420 --> 0:01:53.470 and vision creativity programming, execution ability. This is like top 0:01:53.470 --> 0:01:57.160 top tier. So imagine there's only like 100 or 1000 0:01:57.160 --> 0:02:00.040 of these employees on the whole entire planet at any 0:02:00.040 --> 0:02:02.440 given time. Wherever you want to set that bar, it's 0:02:02.440 --> 0:02:07.560 something like that. Level five is a pinnacle human employee. Okay, 0:02:07.560 --> 0:02:11.190 so this is somebody like the smartest person who's ever lived. 0:02:11.190 --> 0:02:14.160 So a John von Neumann or an Isaac Newton or 0:02:14.160 --> 0:02:17.460 a Richard Feynman or a Claude Shannon. So what this 0:02:17.460 --> 0:02:21.180 offers over tier four is the ability to invent completely 0:02:21.180 --> 0:02:23.850 new things when they don't exist, or see and explain 0:02:23.850 --> 0:02:26.910 the world in a completely new way. So Newton didn't 0:02:26.910 --> 0:02:31.470 have calculus, so he invented it. That's always like, astounded me. 0:02:31.500 --> 0:02:34.980 Feynman was a supreme teacher, like a lot of people 0:02:34.980 --> 0:02:37.020 consider him to be one of the best teachers ever. 0:02:37.050 --> 0:02:42.180 Von Neumann innovated in mathematics, physics, engineering, game theory, like 0:02:42.180 --> 0:02:45.209 tons of stuff, and Claude Shannon gave us information theory, 0:02:45.210 --> 0:02:49.290 the foundations of cryptography, and a million other things. So basically, 0:02:49.290 --> 0:02:51.540 at this level, you not only have the creativity and 0:02:51.540 --> 0:02:55.139 execution of a top, like, whatever, only 100 or 1000 0:02:55.139 --> 0:02:57.690 in the entire world at any given moment. But you 0:02:57.690 --> 0:03:01.770 also have the once in a generation level innovation capabilities. 0:03:01.770 --> 0:03:05.370 And once you're at Pinnacle Human, okay, the level after 0:03:05.370 --> 0:03:08.730 that is where I cross over into ASI, which is 0:03:08.730 --> 0:03:12.450 defined logically as a level of general AI that's smarter 0:03:12.450 --> 0:03:15.690 and more capable than any human that's ever lived. And 0:03:15.690 --> 0:03:17.700 that's the reason I added these, is so we have 0:03:17.700 --> 0:03:22.040 a smoother transition from the end of AGI right into 0:03:22.040 --> 0:03:25.370 the beginning of ASI one. Right. And that's where we 0:03:25.370 --> 0:03:31.160 have right here. ASI one. So that was that. Uh, 0:03:31.160 --> 0:03:33.919 I feel like Apple Notes is my actual operating system 0:03:33.919 --> 0:03:36.650 and Mac OS is just the window manager, just, uh, 0:03:36.650 --> 0:03:41.510 stupid jokes. Um, how many notes do I have in 0:03:41.510 --> 0:03:44.240 Apple Notes? I'm going to look right now to try 0:03:44.240 --> 0:03:50.630 to emphasize this point. 3867 Apple notes my operating system. 0:03:50.630 --> 0:03:53.510 Not really, but you see what I'm saying? All right, 0:03:53.510 --> 0:03:58.970 let's get into it. Okay. So substrate I've been thinking 0:03:58.970 --> 0:04:01.610 about and working on this one for months now and 0:04:01.610 --> 0:04:04.730 finally announced it. And actually there's a separate whole substrate 0:04:04.730 --> 0:04:08.360 video coming out very soon you should check out. So yeah, 0:04:08.360 --> 0:04:10.040 go look at that when it comes out. I'm not 0:04:10.040 --> 0:04:12.170 going to go into detail here because I just did 0:04:12.170 --> 0:04:14.600 in that other piece, got a new piece on dynamic 0:04:14.600 --> 0:04:16.850 content summaries and how I think they're going to be 0:04:16.850 --> 0:04:19.310 the way we view content in the future. Essentially, the 0:04:19.310 --> 0:04:23.690 idea here is that rather than get raw content from 0:04:23.690 --> 0:04:27.679 different places, I will produce content for us dynamic. So 0:04:27.680 --> 0:04:30.680 imagine that we love to see Richard Feynman explain things 0:04:30.680 --> 0:04:34.120 to us. Well, anytime anything complex happens in the world, 0:04:34.120 --> 0:04:36.669 we're going to have a video. It's going to be 0:04:36.670 --> 0:04:39.820 a YouTube video or something like a YouTube video, and 0:04:39.820 --> 0:04:43.930 it's going to be Richard Feynman fully deepfaked, fully Deepfaked. 0:04:43.930 --> 0:04:46.989 It looks exactly like Richard Feynman. And he's actually got 0:04:46.990 --> 0:04:50.620 a piece of chalk, and he's actually in his Caltech 0:04:50.620 --> 0:04:55.930 classroom with a giant whiteboard, and he's dynamically whiteboarding out 0:04:55.930 --> 0:04:58.690 this concept to you and explaining with his voice and 0:04:58.690 --> 0:05:01.930 his jokes and his humor or whatever. And he's explaining 0:05:01.930 --> 0:05:04.360 everything to you. In fact, it might be like if 0:05:04.360 --> 0:05:09.280 it's science related or complexity related or whatever, something Feynman ish, 0:05:09.279 --> 0:05:12.190 then use Feynman. But maybe you've got this other like 0:05:12.190 --> 0:05:15.729 cute influencer girl who you like to explain, like modern 0:05:15.730 --> 0:05:18.789 cultural topics or whatever, and she is deepfaked to do 0:05:18.790 --> 0:05:20.680 that for you there. And of course, there's going to 0:05:20.680 --> 0:05:23.020 be all sorts of considerations around like, well, do they 0:05:23.020 --> 0:05:25.030 have permission to do that? And a lot of people 0:05:25.029 --> 0:05:27.310 will say, yeah, go ahead, do it. As long as 0:05:27.310 --> 0:05:30.940 it falls within these criteria. Some people won't care. They'll 0:05:30.940 --> 0:05:33.940 just deepfake anybody and do that. Um, other times it'll 0:05:33.940 --> 0:05:36.849 be like the avatar of your actual Da. So your 0:05:36.850 --> 0:05:39.670 Da is going to your digital assistant, your personal AI 0:05:39.670 --> 0:05:42.670 is going to have an avatar as well. So maybe 0:05:42.670 --> 0:05:45.029 you like that avatar to tell it to you, right? 0:05:45.029 --> 0:05:48.390 Because maybe that one is of a coach, or honestly, 0:05:48.390 --> 0:05:50.250 it's just going to become like one of your best friends. 0:05:50.250 --> 0:05:53.640 It knows you the closest. It's very, you know, intimately 0:05:53.640 --> 0:05:56.910 connected with you. So maybe it's the one teaching you. 0:05:56.910 --> 0:05:59.910 Maybe it's you and your own voice teaching you. Okay. 0:05:59.910 --> 0:06:02.700 It's going to be very easy to deepfake yourself, so 0:06:02.700 --> 0:06:06.420 maybe you're explaining it to yourself like you created the content. 0:06:06.420 --> 0:06:09.279 That would be insane. Point is, if there is a 0:06:09.279 --> 0:06:14.200 four hour video explaining this very complex topic, okay, it's 0:06:14.200 --> 0:06:16.570 a new type of machine learning. It's a new type 0:06:16.570 --> 0:06:19.210 of AI. It's a new hacking technique, but it's an 0:06:19.210 --> 0:06:21.609 hour long. It's three hours long. It's four hours long. 0:06:21.610 --> 0:06:25.900 It's five hours long. It's a 20 video course. The 0:06:25.900 --> 0:06:28.840 AI will go and watch it, and then it'll be like, hey, um, 0:06:28.839 --> 0:06:31.120 I created a thing. I know you don't have much 0:06:31.120 --> 0:06:35.450 time this week, but I created a 14 minute summary 0:06:35.450 --> 0:06:39.140 of this 20 video series, and I'm going to show 0:06:39.140 --> 0:06:42.260 it to you. It's actually done by Richard Feynman. So 0:06:42.260 --> 0:06:46.010 now Richard Feynman explains that entire 20 videos to you 0:06:46.010 --> 0:06:49.310 in 14 minutes, perfectly taught by Richard Feynman. So this 0:06:49.310 --> 0:06:52.370 is the concept of dynamic content summaries that the raw 0:06:52.370 --> 0:06:55.430 version of a thing is usually not going to be 0:06:55.430 --> 0:06:58.280 ideal for a person. The ideal will be to have 0:06:58.279 --> 0:07:01.250 it pitched in the way they like to receive things. Okay, 0:07:01.250 --> 0:07:03.409 because it won't always be a video. Sometimes it will 0:07:03.410 --> 0:07:05.839 be an audio book. It'll just be audio in your ear. 0:07:05.839 --> 0:07:08.000 Other times it'll be like a visual story, and it 0:07:08.000 --> 0:07:10.580 won't actually have an avatar or a person teaching it 0:07:10.580 --> 0:07:13.520 on a whiteboard. It'll just be like animation, and maybe 0:07:13.520 --> 0:07:15.500 people will learn better from that. Maybe people will learn 0:07:15.500 --> 0:07:18.230 better from a story, whatever they prefer. That content will 0:07:18.230 --> 0:07:22.130 be dynamically created by eye to best give you the 0:07:22.130 --> 0:07:25.220 content from the raw version. And of course, I'm a 0:07:25.220 --> 0:07:28.130 huge fan of slow, slow learning, and there's a whole 0:07:28.130 --> 0:07:31.160 bunch of stuff in fabric already for this. Like sometimes 0:07:31.160 --> 0:07:34.190 you don't want that summary, and sometimes you should stop 0:07:34.190 --> 0:07:37.520 watching those summaries and say in the AI should say, 0:07:37.520 --> 0:07:40.190 you know what? I created you a 14 minute summary 0:07:40.190 --> 0:07:42.710 of this thing, but this you need to go watch 0:07:42.710 --> 0:07:45.410 the full thing because oftentimes because I've been using these 0:07:45.430 --> 0:07:50.739 summaries with fabric for almost two years now and I often, well, 0:07:50.740 --> 0:07:54.220 pretty much always I always get something more from doing 0:07:54.220 --> 0:07:56.950 the slow version, from watching the four hour version, from 0:07:56.950 --> 0:08:00.490 watching their facial expressions. I get value from the summary itself, 0:08:00.490 --> 0:08:03.700 but I get more value from watching it happen slowly 0:08:03.700 --> 0:08:06.670 and watching it seep in slowly. So maybe you do both, 0:08:06.670 --> 0:08:08.890 maybe you do one or the other, but don't give 0:08:08.890 --> 0:08:10.960 up the slow just because you could do the fast. 0:08:10.960 --> 0:08:13.690 Exploring the idea of personal and business brands. Yeah, that 0:08:13.690 --> 0:08:17.020 was another short piece. Okay. Kaspersky is shut down US 0:08:17.050 --> 0:08:21.250 operations because of banning. AT&T says nearly all cellular customers 0:08:21.250 --> 0:08:24.370 and some landline users have had their data stolen. But 0:08:24.370 --> 0:08:30.340 Kim Zetter at Wired said somebody got paid like $300,000 0:08:30.340 --> 0:08:34.990 from the hacking team by AT&T to delete what was 0:08:34.990 --> 0:08:39.189 supposedly a copy of all the data, and supposedly the 0:08:39.190 --> 0:08:42.100 only copy of all the data. So we'll see how 0:08:42.100 --> 0:08:44.860 that plays out. But good reporting from Kim Zetter. Russia 0:08:44.860 --> 0:08:49.479 is using AI enhanced software called Meliorator to create fake 0:08:49.480 --> 0:08:54.370 online personas for disinformation campaigns. Ooh, love this, because the 0:08:54.370 --> 0:08:56.229 better this gets, the more it's going to look like 0:08:56.230 --> 0:08:59.520 a real human. And then if the AI starts acting 0:08:59.520 --> 0:09:02.339 like a human with like clicking on things, being able 0:09:02.340 --> 0:09:03.930 to tell the difference is going to get a whole 0:09:03.929 --> 0:09:05.729 lot more difficult. And this is actually one of the 0:09:05.730 --> 0:09:08.670 things I'm most worried about from AI is disinformation bots, 0:09:08.670 --> 0:09:11.490 not just the sheer number of them, but also their 0:09:11.490 --> 0:09:14.610 how advanced they are, how sophisticated they are. Basically, the 0:09:14.610 --> 0:09:18.180 better AI gets, and especially from agent frameworks, the more 0:09:18.179 --> 0:09:21.660 it's going to be exactly the same, or almost exactly 0:09:21.660 --> 0:09:26.490 the same, as if our enemies have millions of employees 0:09:26.490 --> 0:09:29.640 in their intelligence agencies. Okay, let me say that again. 0:09:29.640 --> 0:09:31.920 That was that was too convoluted. One of the things 0:09:31.920 --> 0:09:35.010 I'm most worried about is that the better AI gets, 0:09:35.010 --> 0:09:42.360 it'll be like our foreign intelligence agencies for our enemies China, Russia, Iran, 0:09:42.360 --> 0:09:46.170 North Korea. Instead of having a few hundred elite people, 0:09:46.170 --> 0:09:48.810 they're going to have a few hundred elite people and 0:09:48.809 --> 0:09:52.800 40,000 right below them, elite people. And then the better 0:09:52.800 --> 0:09:56.010 the AI gets, those move up into the elite, and 0:09:56.010 --> 0:09:59.520 then the elite humans become augmented with that AI and 0:09:59.520 --> 0:10:02.430 they become like super elite. And the better the AI gets, 0:10:02.429 --> 0:10:05.130 I mean, it just keeps moving up that chain. So 0:10:05.130 --> 0:10:09.060 if you think it's bad to have ten, 50, 100 0:10:09.059 --> 0:10:13.520 elite people in these different countries who are both Intel 0:10:13.520 --> 0:10:17.540 trained and manipulation trained, but they're also like super hackers, 0:10:17.540 --> 0:10:20.600 and they're good at all these different skills. But there's 0:10:20.600 --> 0:10:23.150 I mean, there's only so many in the world. Well, 0:10:23.150 --> 0:10:28.370 now imagine them with a thousand of them, 10,000 of them, 100,000, 0:10:28.370 --> 0:10:32.540 100 million. And they're working consistently with no sleep to 0:10:32.540 --> 0:10:35.840 go after all your targets. That that's what I'm personally 0:10:35.840 --> 0:10:39.260 worried about right now. And if they are producing content 0:10:39.260 --> 0:10:42.170 and the content is all manipulative, the internet is going 0:10:42.170 --> 0:10:45.200 to have to switch from block list to an allow list. 0:10:45.200 --> 0:10:47.569 I mean, think about that. The internet currently is an 0:10:47.570 --> 0:10:52.580 allow list. I'm sorry, it's a block list. Everything is available, 0:10:52.580 --> 0:10:55.010 and you have to specifically go into a place and say, 0:10:55.010 --> 0:10:56.540 I don't want to see this. I don't want to 0:10:56.540 --> 0:10:58.640 see this. I don't want to see this because of 0:10:58.640 --> 0:11:01.110 this problem right here that I'm talking about. We're going 0:11:01.110 --> 0:11:03.150 to have to switch. We're going to have to switch 0:11:03.150 --> 0:11:06.990 to the other method, which is to say my I, 0:11:07.020 --> 0:11:13.980 my internet platform, my mobile OS, okay, Android, macOS, iOS, 0:11:13.980 --> 0:11:16.620 they're not going to allow anything in unless you have 0:11:16.620 --> 0:11:18.809 the green check mark. What is the green check mark? 0:11:18.809 --> 0:11:22.890 Green check mark is an agreement between the the tech platform, 0:11:22.890 --> 0:11:27.130 the mobile platform, the OS and trusted providers like a YouTube, 0:11:27.130 --> 0:11:32.860 which has approved sources from approved individuals and approved organizations. 0:11:32.860 --> 0:11:34.870 So your news is going to have to come that 0:11:34.870 --> 0:11:37.150 way because it's going to have to be vetted. So 0:11:37.150 --> 0:11:40.540 it gets the green checkmark. This might be something that's required. 0:11:40.540 --> 0:11:42.370 It's going to happen anyway, but it might be something 0:11:42.370 --> 0:11:47.080 that's required within settings and within operating systems because it's 0:11:47.080 --> 0:11:49.840 simply too hard to drown out everything. Because this problem 0:11:49.840 --> 0:11:53.950 right here, the amount of disinformation bots, pretty soon it's 0:11:53.950 --> 0:11:58.060 going to be like, whatever, 50% of all content is 0:11:58.059 --> 0:12:00.760 going to be manipulation, then it's going to be like 75%, 0:12:00.760 --> 0:12:03.130 then it's going to be 95%. So if you are 0:12:03.130 --> 0:12:05.920 naive and you just walk onto the internet, especially as 0:12:05.920 --> 0:12:08.860 like a regular person or like a grandma or something, 0:12:08.860 --> 0:12:11.109 you're going to be screwed. You're going to be just 0:12:11.110 --> 0:12:13.780 like pulled around, Like all your money will disappear because 0:12:13.780 --> 0:12:16.449 you had to save your kids from some horrible accident. 0:12:16.450 --> 0:12:18.640 And all of a sudden there was an asteroid that 0:12:18.640 --> 0:12:21.099 hit the thing and aliens came. And pretty soon you're 0:12:21.100 --> 0:12:23.410 in a bunker and you have no money, and it's 0:12:23.410 --> 0:12:26.079 like none of it was real. The amount of and 0:12:26.080 --> 0:12:30.250 quality of manipulation. Think about that as well. The quality 0:12:30.250 --> 0:12:34.360 of the pitches and the manipulation and the deepfakes and 0:12:34.360 --> 0:12:36.780 the trickery is going to get so good. It's going 0:12:36.780 --> 0:12:39.300 to trick us to it's going to trick experts, not 0:12:39.300 --> 0:12:42.720 just regular people. There's a new XM vulnerability with a 0:12:42.720 --> 0:12:46.319 Cvss of 9.1. Google is now offering Passkeys for high 0:12:46.320 --> 0:12:49.740 risk users who join Advanced Protection. Previously, they had to 0:12:49.740 --> 0:12:56.309 get a physical token foreign influence campaigns analysis from US intelligence. Basically, 0:12:56.309 --> 0:12:59.939 they say Russia is backing Trump, probably mostly because of Ukraine. 0:12:59.940 --> 0:13:03.000 Iran is acting as a chaos agent in its campaigns, 0:13:03.000 --> 0:13:06.809 focusing on exploiting US political and social tensions rather than 0:13:06.809 --> 0:13:10.590 backing one side or another. And China's mostly staying out 0:13:10.590 --> 0:13:15.000 of the elections just trying to data collect for future 0:13:15.000 --> 0:13:20.130 influence operations. Think OPM, uh, Marriott. Yeah, though don't get 0:13:20.130 --> 0:13:23.100 me started on those. Thanks to auto Max for sponsoring. 0:13:23.100 --> 0:13:27.140 GitLab has a critical flaw in its CI CD pipelines. 0:13:27.170 --> 0:13:32.360 Thanks to Project Discovery for sponsoring and OpenAI's AGI levels, 0:13:32.360 --> 0:13:35.630 OpenAI has published their five tier ladder for AI progress. 0:13:35.630 --> 0:13:38.449 I'm honestly not a fan. Other than level five, I 0:13:38.450 --> 0:13:41.360 find level five really interesting. But the problem is they're 0:13:41.360 --> 0:13:44.420 going from like chat bots to human level reasoners and 0:13:44.420 --> 0:13:47.300 then to AI agents, then to innovators that can aid 0:13:47.300 --> 0:13:50.329 in invention. So level two and level four are way 0:13:50.330 --> 0:13:55.790 too close. Okay. So human level reasoners to agents. It's 0:13:55.790 --> 0:13:59.719 not a big enough jump. Uh, and both are already possible. 0:13:59.720 --> 0:14:02.090 That's the other problem. Level two and level four. They're 0:14:02.090 --> 0:14:04.580 kind of already possible. Then you have this really interesting 0:14:04.580 --> 0:14:06.950 jump at level five to something that can do the 0:14:06.950 --> 0:14:11.180 work of an organization. So for OpenAI level five, AGI 0:14:11.179 --> 0:14:14.450 is something that could do the work of an entire organization, 0:14:14.450 --> 0:14:18.380 but that's not super helpful. Okay, there are lots of 0:14:18.380 --> 0:14:22.400 very dumb organizations, okay? An organization could be for people. 0:14:22.400 --> 0:14:26.360 An organization can be Apple, an organization could be OpenAI 0:14:26.360 --> 0:14:30.740 in the year 2027. So that is not super descriptive 0:14:30.740 --> 0:14:33.410 in my opinion. And even worse, the problem is they're 0:14:33.410 --> 0:14:37.070 mixing criteria here. Right? So one is I don't know 0:14:37.070 --> 0:14:41.320 what what criteria number one is Reasoners is about thinking 0:14:41.320 --> 0:14:45.010 quality agents is just an attribute, which means the thing 0:14:45.010 --> 0:14:46.960 can take an action. It doesn't say anything about the 0:14:46.960 --> 0:14:50.890 quality of the action. So that's not necessarily above one 0:14:50.890 --> 0:14:54.700 or below four or above four. It's it doesn't put 0:14:54.700 --> 0:14:57.850 it anywhere because it's apples to oranges. Innovators is just 0:14:57.850 --> 0:15:02.050 a descriptive output. So it aids in invention. Here's the 0:15:02.050 --> 0:15:05.710 problem with this chat bots already aid with invention and 0:15:05.710 --> 0:15:07.960 that's what they have as their level. One is chat bots. 0:15:07.990 --> 0:15:11.440 The chat bots already help inventors invent better things. Then 0:15:11.440 --> 0:15:15.010 you have level five, which is actually about scale more 0:15:15.010 --> 0:15:18.340 than quality. Um, yeah. And like I said, scale isn't 0:15:18.340 --> 0:15:21.370 super important because you don't know if you're scaling like 0:15:21.370 --> 0:15:23.920 an ant colony. They're just kind of doing the same thing. 0:15:23.920 --> 0:15:27.580 Or if you're scaling and the scale itself leads to 0:15:27.580 --> 0:15:31.510 smarter things. So just not overall happy with it. I'm 0:15:31.510 --> 0:15:33.910 happy they put it out and I feel bad because 0:15:33.910 --> 0:15:36.040 I'm sure someone put some effort into it. And I'm 0:15:36.040 --> 0:15:38.950 sure they're human and I don't want you to feel bad. 0:15:38.950 --> 0:15:42.610 Just hopefully take this feedback and maybe the next one. 0:15:42.610 --> 0:15:46.550 Just think about these ideas, okay? AI startups raising $100 0:15:46.550 --> 0:15:49.870 million in 2024. Got a full list there. Anthropic has 0:15:49.870 --> 0:15:54.600 more new features, uh, to help automate prompt engineering. And 0:15:54.600 --> 0:15:57.240 I see a lot of my friends switching to cloud 0:15:57.240 --> 0:16:01.860 over ChatGPT right now, especially sonnet 35. I'm actually using 0:16:01.860 --> 0:16:05.370 sonnet 35 as my default within fabric right now, and 0:16:05.370 --> 0:16:07.530 it's a leapfrog game because soon we're about to have 0:16:07.530 --> 0:16:12.690 opus 35. Cannot wait for that llama 3.0 300 billion 0:16:12.690 --> 0:16:17.290 and eventually GPT five, I'm guessing fall to right at 0:16:17.290 --> 0:16:21.609 the very beginning of January. So I'm going to say August. September. 0:16:21.610 --> 0:16:27.190 I'm going to say between October and February, 90% chance 0:16:27.190 --> 0:16:30.850 of GPT five or the next GPT to come out, 0:16:30.850 --> 0:16:32.440 whether they call it 4 or 5 or they call 0:16:32.440 --> 0:16:36.070 it five, whatever, or some new name. And yeah, 2025 0:16:36.070 --> 0:16:38.950 is going to be nuts. So we're going gonna we're 0:16:38.950 --> 0:16:41.560 gonna have a new president. Well, maybe a new president, 0:16:41.560 --> 0:16:43.630 but either way, it's going to be a new whatever. 0:16:43.630 --> 0:16:45.310 There's going to be an election and an outcome. So 0:16:45.310 --> 0:16:47.739 that's going to be crazy by itself. Um, hopefully not 0:16:47.740 --> 0:16:50.890 the bad crazy, but, uh, yeah, there's just going to 0:16:50.890 --> 0:16:54.280 be a whole lot going on in 2025, not the 0:16:54.280 --> 0:16:56.890 least of which is all these new models. New fiber 0:16:56.890 --> 0:17:03.960 optic network transmits data at above 400TB per second. And importantly, 0:17:03.960 --> 0:17:07.590 this is using existing fiber, not some, you know, adamantium 0:17:07.590 --> 0:17:10.410 special stuff. YouTube music is testing. A new feature lets 0:17:10.410 --> 0:17:13.470 you use AI to generate a playlist by just describing 0:17:13.470 --> 0:17:16.020 what you want to hear. I would use that a lot. 0:17:16.020 --> 0:17:19.889 There's a surge in delivery startups like Hail Ify, okay, 0:17:19.890 --> 0:17:24.840 Hail Ify, and they're focusing on shine and TMU, which 0:17:24.840 --> 0:17:28.490 are the Chinese people competing with Amazon? Why women are 0:17:28.490 --> 0:17:31.790 disappearing from tech Houston is on a path to an 0:17:31.790 --> 0:17:35.720 all out power crisis. Tour de France riders are inhaling 0:17:35.720 --> 0:17:38.870 carbon monoxide, which is a poison gas. It's the one 0:17:38.869 --> 0:17:41.330 that kills you, I think, in your car, in a garage, 0:17:41.330 --> 0:17:44.719 if you don't have, like, open ventilation, how how is 0:17:44.720 --> 0:17:50.750 a toxic gas improving performance that okay. Yeah. 130 million 0:17:50.750 --> 0:17:56.030 US adults have low literacy skills, over half of Americans 0:17:56.030 --> 0:18:01.040 aged 16 to 74. Read below. A sixth grade level. 0:18:01.040 --> 0:18:03.020 I'm gonna say that again for the people in the back, 0:18:03.020 --> 0:18:08.000 over half of Americans aged 16 to 74 read below. 0:18:08.000 --> 0:18:11.930 A sixth grade level moving on Colorado poultry workers test 0:18:11.930 --> 0:18:14.659 positive for bird flu. But we're not talking about that 0:18:14.660 --> 0:18:18.010 because we we're tired of talking about viruses. Uh, we're 0:18:18.010 --> 0:18:19.959 tired of talking about Covid. We don't want to talk 0:18:19.960 --> 0:18:23.470 about bird flu. That's just another Covid. It's, uh. We're exhausted. 0:18:23.470 --> 0:18:25.960 You know what? I give people a break. People are exhausted. 0:18:25.960 --> 0:18:28.870 They don't want to talk about negative stuff. They literally 0:18:28.869 --> 0:18:31.900 are just going to ignore this with blinders until it 0:18:31.900 --> 0:18:35.020 literally comes to their front door and, like, makes their 0:18:35.020 --> 0:18:38.290 friends sick. And other than that, any other coverage of this, 0:18:38.290 --> 0:18:40.120 they're just going to be like, yeah, I'll know about 0:18:40.119 --> 0:18:43.510 it if it gets too bad. Yeah. And meanwhile, all 0:18:43.510 --> 0:18:46.540 the experts are like, guys, this is kind of bad. 0:18:46.540 --> 0:18:51.070 It's transferring it humans. It's killed 6 million birds and 0:18:51.070 --> 0:18:55.480 is now infecting dairy cattle across the state, which means 0:18:55.480 --> 0:19:01.270 it's going into milk and dairy workers are getting it. Uh, no. 0:19:01.270 --> 0:19:04.030 Poultry workers are now getting it. So it's like. I mean, 0:19:04.030 --> 0:19:07.179 I kind of feel it along with everyone else. I'm like, uh, 0:19:07.180 --> 0:19:11.380 I guess if my friends start getting it, I'll pay attention. 0:19:11.380 --> 0:19:15.250 Very strange times we live in, we're so, so exhausted 0:19:15.250 --> 0:19:19.419 by negativity in certain types of negativity. Just foreign. 10,000 0:19:19.420 --> 0:19:22.990 galaxies may host intelligent aliens. This is a new study 0:19:22.990 --> 0:19:26.109 basically trying to explain the Drake equation of like why 0:19:26.109 --> 0:19:30.389 we're not seeing aliens, but four out of 10,000 galaxies. 0:19:30.390 --> 0:19:34.919 I believe there's 200 billion galaxies. Unless I'm mixing up 0:19:34.920 --> 0:19:37.560 my numbers. Is it 200 billion galaxies? I think it's 0:19:37.560 --> 0:19:42.510 200 billion. So 4 in 10,000. But there's actually. So 0:19:42.510 --> 0:19:45.990 what is that ten K to 200 billion. That's still 0:19:45.990 --> 0:19:50.130 a lot of galaxies. That's a lot of galaxies. And 0:19:50.130 --> 0:19:52.500 that's just this is just one study okay okay. Okay. 0:19:52.500 --> 0:19:59.130 So it's saying 0.003% to 0.2% of exoplanets. Yeah, but 0:19:59.130 --> 0:20:02.160 there's a lot of exoplanets like this is not super 0:20:02.160 --> 0:20:04.770 depressing to me. But I guess it's interesting if it 0:20:04.770 --> 0:20:09.990 explains the Drake equation. Okay, ideas VC's are buying medical practices. 0:20:09.990 --> 0:20:12.810 I got a buddy who's talking about how VCs are 0:20:12.810 --> 0:20:16.080 just coming into these practices, and they're just moving in 0:20:16.080 --> 0:20:18.119 and being like, you've got to sell all this super 0:20:18.119 --> 0:20:22.260 gross stuff. Order tests that you don't need, like sell 0:20:22.260 --> 0:20:25.169 them miracle drugs that don't actually work. You've got to 0:20:25.170 --> 0:20:28.890 do this to hype up the user base. They're basically 0:20:28.890 --> 0:20:33.900 turning it into like sexy sales and, um, getting them 0:20:33.900 --> 0:20:38.460 to do things other than fundamental, like high quality, proactive 0:20:38.460 --> 0:20:41.720 health like diet and exercise. It's kind of going the 0:20:41.720 --> 0:20:48.169 opposite direction towards more tests, more assessments, pills and things 0:20:48.170 --> 0:20:50.810 that don't actually work. Why? Because they make more money, 0:20:50.810 --> 0:20:54.170 which is good for VC's. So really, really bad. And 0:20:54.170 --> 0:20:57.380 thanks to my buddy for calling that out. Most conspiracies 0:20:57.380 --> 0:21:00.590 come from not realizing how often things fail. The tweet 0:21:00.590 --> 0:21:04.430 thread I put out recently on X therapy, rumination, and 0:21:04.430 --> 0:21:08.550 Untying knots. This is about therapy. Recommend you click on it. 0:21:08.580 --> 0:21:10.500 We're already going long so I'm not going to talk 0:21:10.500 --> 0:21:15.359 about it here. Discovery fluff on Lambda. Really cool. Uh 0:21:15.359 --> 0:21:18.750 really cool project by this guy here. Um, definitely want 0:21:18.750 --> 0:21:21.930 to check it out. Bullfrog GitHub action secures your workflows 0:21:21.930 --> 0:21:25.649 by controlling outbound network connections. Everything you see is a 0:21:25.650 --> 0:21:29.370 computational process if you know how to look. Um. Love it. 0:21:29.369 --> 0:21:33.879 Sean Purry emotional elicitor. This is a prompt by Moritz 0:21:33.910 --> 0:21:37.120 Kremp who's a really cool, uh, doing some really cool 0:21:37.119 --> 0:21:40.720 prompt stuff. Got to check him out. WTF happened to blogs? 0:21:40.720 --> 0:21:44.050 As an employee you are disposable. You never control the 0:21:44.050 --> 0:21:47.620 arc of your career. Smoking versus lung cancer deaths. I'm 0:21:47.619 --> 0:21:49.629 going to show this one. This is insane. Look at 0:21:49.630 --> 0:21:52.480 this thing. Hmm. Wonder if there's a correlation. Well, we 0:21:52.480 --> 0:21:56.590 see a correlation. I wonder if there's causation. That's that's compelling. 0:21:56.590 --> 0:22:01.210 Look at this. Taxes increasing bans taxes. Look at this. 0:22:01.210 --> 0:22:04.900 These are deaths going down. Great depression, World War two 0:22:04.900 --> 0:22:08.890 boom US banned cigarette ads on the radio. And surgeon 0:22:08.890 --> 0:22:12.820 General report linked smoking to death from cancer. That's the top. 0:22:12.820 --> 0:22:14.740 And it starts going down from there with all these 0:22:14.740 --> 0:22:18.969 other measures. Really interesting learning multiple concepts from a SQL image, 0:22:18.970 --> 0:22:22.870 change detection in satellite imagery. This is a hobby. I 0:22:22.869 --> 0:22:25.060 would be doing a lot more. I'm actually waiting for 0:22:25.060 --> 0:22:27.010 the AI to get really good at this. I want 0:22:27.010 --> 0:22:30.640 to create a bot that basically watches the latest satellite 0:22:30.640 --> 0:22:35.680 images from choice locations, uses AI in a pipeline, and 0:22:35.800 --> 0:22:40.119 sends me a constant flow of updates of what changed. 0:22:40.119 --> 0:22:44.700 Like it looks like they did something underground. Um, a 0:22:44.700 --> 0:22:46.890 whole bunch of trucks moved in. They appear to be 0:22:46.890 --> 0:22:50.429 this kind of truck. Here are the possible explanations for that. Again, 0:22:50.430 --> 0:22:52.979 going back to the Intel thing that I'm building, I 0:22:52.980 --> 0:22:58.170 want continuous flows of information flowing into me based on 0:22:58.170 --> 0:23:03.419 capturing artifacts, in this case images, and doing AI analysis 0:23:03.420 --> 0:23:06.660 on them. 89 things I Know About Git Commits I 0:23:06.660 --> 0:23:10.740 got the most comments about this one and substrate this week, 0:23:10.740 --> 0:23:14.100 so definitely check this one out by Jamie Tanner and 0:23:14.100 --> 0:23:16.920 the recommendation of the week. Check on your friends if 0:23:16.920 --> 0:23:20.340 you haven't heard from them in a while, send them 0:23:20.340 --> 0:23:23.370 a text. It's free and they will appreciate being thought 0:23:23.369 --> 0:23:26.520 of and be consistent about it. Like push it right? 0:23:26.520 --> 0:23:31.200 They they will appreciate that in the age of infinite leverage, 0:23:31.200 --> 0:23:35.820 judgment is the most important skill. Naval has been crushing 0:23:35.820 --> 0:23:38.460 it lately. Absolutely crushing it. I'm going to say this 0:23:38.460 --> 0:23:42.330 one again. In the age of infinite leverage, judgment is 0:23:42.330 --> 0:23:48.330 the most important skill. Naval. Unsupervised learning is produced and 0:23:48.330 --> 0:23:51.870 edited by Daniel Miessler on a Neumann U87 AI microphone 0:23:51.869 --> 0:23:55.850 using Hindenburg. Intro and outro music is by Zomby with 0:23:55.850 --> 0:23:58.670 the Y, and to get the text and links from 0:23:58.670 --> 0:24:00.830 this episode, sign up for the newsletter version of the 0:24:00.830 --> 0:24:06.350 show at Daniel miessler.com/newsletter. We'll see you next time.