1 00:00:17,267 --> 00:00:20,267 S1: All right. Welcome to episode 45. Posted a bunch of 2 00:00:20,267 --> 00:00:23,347 S1: fabric extractions of stuff that I've been watching and reading 3 00:00:23,347 --> 00:00:27,227 S1: and stuff like that. So like YouTube videos, blog posts, 4 00:00:27,627 --> 00:00:30,107 S1: academic papers, and I think I'm going to do more 5 00:00:30,107 --> 00:00:33,107 S1: of this because I'm actually building this into my Pi, 6 00:00:33,147 --> 00:00:37,107 S1: which is pi personal AI infrastructure. And this is the 7 00:00:37,107 --> 00:00:40,307 S1: infrastructure you've heard me talking about in the augmented class. 8 00:00:40,987 --> 00:00:43,987 S1: And just like all over X and in the newsletter 9 00:00:43,987 --> 00:00:46,987 S1: over the years, this is basically the back end core 10 00:00:46,987 --> 00:00:51,347 S1: infrastructure that I basically runs for my life. So it's 11 00:00:51,346 --> 00:00:56,987 S1: basically finding all my content. It's categorizing all my content, labeling, sorting, 12 00:00:57,027 --> 00:01:02,867 S1: doing things like that, extracting summaries, um, you know, characterizing. 13 00:01:02,867 --> 00:01:06,147 S1: And I'm basically going to use that to continue enhancing 14 00:01:06,147 --> 00:01:08,467 S1: my life. Right. So I'm always adding to this thing. 15 00:01:09,267 --> 00:01:11,947 S1: So one of the core components of that is actually 16 00:01:13,507 --> 00:01:17,607 S1: finding new content, um, in an automated way. So basically 17 00:01:17,607 --> 00:01:21,167 S1: there's I'll give you a good example. There's on arXiv. 18 00:01:21,207 --> 00:01:22,967 S1: I think that's the way you pronounce this. I've heard 19 00:01:22,967 --> 00:01:27,207 S1: people say this, I used to call it arXiv because whatever. 20 00:01:27,207 --> 00:01:32,767 S1: I'm silly. arXiv, I guess it's called arXiv. It's a 21 00:01:32,767 --> 00:01:36,727 S1: cool name. It's a better name than arXiv, although that's 22 00:01:36,727 --> 00:01:40,206 S1: how it's spelled. But whatever on arXiv there is a 23 00:01:40,607 --> 00:01:48,287 S1: RSS feed for. CS I so it's papers coming out 24 00:01:48,287 --> 00:01:52,647 S1: of arXiv that are both CS related and AI related. 25 00:01:52,687 --> 00:01:55,447 S1: And obviously this is a sweet spot for me. So 26 00:01:56,047 --> 00:01:57,967 S1: I want to be able to go and get those papers. 27 00:01:57,967 --> 00:02:01,247 S1: But I also want to know how high quality they are, like, 28 00:02:01,247 --> 00:02:04,487 S1: what are the interesting claims that they're making? Like how 29 00:02:04,487 --> 00:02:08,687 S1: good was their methodology, their testing? Like, was it a 30 00:02:08,686 --> 00:02:12,007 S1: meta study? Did they do their own independent research for 31 00:02:12,007 --> 00:02:15,937 S1: the paper, like all those combinations. And there's already a 32 00:02:15,936 --> 00:02:18,337 S1: fabric pattern for this, which is analyze paper. And I 33 00:02:18,337 --> 00:02:20,417 S1: think I have a new version of it as well. 34 00:02:20,737 --> 00:02:23,377 S1: But anyway, I want to do that type of automated 35 00:02:23,377 --> 00:02:27,337 S1: analysis on every paper I might want to read. Okay. 36 00:02:27,376 --> 00:02:30,096 S1: This is kind of giving you hints of like what 37 00:02:30,096 --> 00:02:33,897 S1: my personal AI infrastructure does. And by the way, I'm 38 00:02:33,897 --> 00:02:35,776 S1: aware of the fact that I'm saying like a little 39 00:02:35,776 --> 00:02:40,376 S1: bit too much. So I'm trying to tamper that down. 40 00:02:40,417 --> 00:02:43,537 S1: I like the word. I think it's a cool transition word, 41 00:02:43,536 --> 00:02:45,417 S1: but I don't want to use it 40 times in 42 00:02:45,417 --> 00:02:50,697 S1: every sentence. Anyway, this is the infrastructure that I'm actually building, 43 00:02:51,017 --> 00:02:53,577 S1: and this is one of the modules that's really, really 44 00:02:53,577 --> 00:02:56,416 S1: important to me because it's not just papers. I want 45 00:02:56,457 --> 00:02:58,417 S1: to see this for open source intelligence. I want to 46 00:02:58,417 --> 00:03:00,897 S1: see this for all the different stuff. So these are 47 00:03:00,897 --> 00:03:04,656 S1: all the independent modules that I'm building. And actually those 48 00:03:04,656 --> 00:03:07,977 S1: essentially turn into products for me. Right. So my Intel 49 00:03:07,977 --> 00:03:10,897 S1: product is going to be the output of one of these. 50 00:03:11,137 --> 00:03:15,397 S1: Threshold is already the output of several of these, right? 51 00:03:15,517 --> 00:03:21,637 S1: So anyway, that's just a brief little thing on my pie. Pie. 52 00:03:23,957 --> 00:03:27,677 S1: All that to say that these fabric extractions are essentially 53 00:03:27,677 --> 00:03:31,637 S1: outputs from these modules where it's found a piece of 54 00:03:31,637 --> 00:03:34,436 S1: really unique content, or I manually found it and I 55 00:03:34,436 --> 00:03:39,796 S1: just manually ran, uh, fabric against it. And it outputted 56 00:03:39,797 --> 00:03:42,077 S1: this output so you could check it on X, you 57 00:03:42,077 --> 00:03:44,197 S1: could see it on, you know, it. I do a 58 00:03:44,197 --> 00:03:46,517 S1: few of these in the newsletter every once in a while, 59 00:03:46,517 --> 00:03:50,397 S1: but I find it really useful to go and extract predictions, 60 00:03:50,637 --> 00:03:54,517 S1: to extract wisdom, to extract insights, or is another one 61 00:03:54,517 --> 00:03:58,197 S1: of my favorite. There's also one that just extracts quotes, 62 00:03:58,557 --> 00:04:02,597 S1: which can also be really interesting. There's um, I think 63 00:04:02,597 --> 00:04:07,197 S1: there's a one that just extracts like data or recommendations. 64 00:04:07,197 --> 00:04:11,157 S1: I think there's definitely a recommendations. One, actually, now that 65 00:04:11,157 --> 00:04:13,447 S1: I think about it, I need to have one that 66 00:04:13,447 --> 00:04:16,847 S1: extracts like a book list or a watch list. Um, 67 00:04:16,847 --> 00:04:18,607 S1: I'm not sure if I already made that one or 68 00:04:18,607 --> 00:04:21,487 S1: if I need to go add that to fabric. Anyway, 69 00:04:21,487 --> 00:04:27,207 S1: all this to say, I want my pi Pi always 70 00:04:27,207 --> 00:04:31,607 S1: running in the background, always collecting, always finding, always discovering, 71 00:04:31,607 --> 00:04:36,287 S1: always rating. And before you think, well hold on, I 72 00:04:36,327 --> 00:04:39,607 S1: is doing too much work for you. No, it is 73 00:04:39,607 --> 00:04:42,407 S1: not doing too much work for me. And here's why. 74 00:04:43,687 --> 00:04:48,367 S1: Part of this entire thing is the determination of what 75 00:04:48,367 --> 00:04:52,207 S1: I should go and do slowly. Okay? Anything that makes 76 00:04:52,207 --> 00:04:56,487 S1: it into that bucket, I must sit down and manually 77 00:04:56,487 --> 00:05:00,807 S1: enjoy and listen to and process and use my pen 78 00:05:00,807 --> 00:05:04,447 S1: and my index cards and write down notes or Apple 79 00:05:04,447 --> 00:05:06,367 S1: notes or whatever it is if I have to pause 80 00:05:06,367 --> 00:05:09,607 S1: this thing. But I still think the absolute best way 81 00:05:09,607 --> 00:05:14,347 S1: to experience anything is in the slow form, right? What's 82 00:05:14,347 --> 00:05:17,387 S1: still the best possible way to spend an evening is 83 00:05:17,427 --> 00:05:22,107 S1: with friends and family and smart people or whatever, and 84 00:05:22,107 --> 00:05:26,827 S1: you're basically having conversation with actual humans. So to me, 85 00:05:26,827 --> 00:05:29,827 S1: all this AI stuff, the whole purpose of it is 86 00:05:29,827 --> 00:05:33,387 S1: to enhance your ability to do those things. And this 87 00:05:33,387 --> 00:05:36,307 S1: is still this is all part of human 3.0. It's 88 00:05:36,307 --> 00:05:38,987 S1: all part of like, what should we actually be doing? 89 00:05:39,307 --> 00:05:42,987 S1: And what we should actually be doing is magnifying human experience. 90 00:05:43,427 --> 00:05:48,627 S1: Slow experience. Right. All this speeding up and all this 91 00:05:48,627 --> 00:05:52,147 S1: collection and all this, I should be in service of that, 92 00:05:52,507 --> 00:05:57,827 S1: not in opposition to that. All right. That's how we 93 00:05:57,827 --> 00:06:00,227 S1: get stuck on the first item. Not even getting into 94 00:06:00,227 --> 00:06:04,187 S1: the newsletter. Okay. Yeah. Go sign up for the podcast again, 95 00:06:04,187 --> 00:06:06,787 S1: which is silly because you're already listening to it. Uh, 96 00:06:06,787 --> 00:06:11,197 S1: let's see here my friend Tracy Talbot is presenting Tell 97 00:06:11,237 --> 00:06:17,437 S1: No Lies. Teaching AI to know when it doesn't know. Yeah, 98 00:06:17,437 --> 00:06:19,737 S1: to know when it doesn't know. And this is June 99 00:06:19,737 --> 00:06:22,877 S1: 26th in Austin, and she's got a LinkedIn post that 100 00:06:22,877 --> 00:06:25,957 S1: I linked in the newsletter. I worked with her at Apple. 101 00:06:25,957 --> 00:06:31,676 S1: She was extraordinary there at Apple. She was doing data quality, uh, 102 00:06:31,677 --> 00:06:37,517 S1: on our team. And yeah, just really, really great person and, um, 103 00:06:37,597 --> 00:06:42,197 S1: really thoughtful about data quality and data security and all 104 00:06:42,197 --> 00:06:48,197 S1: these different aspects of data and basically data management at 105 00:06:48,197 --> 00:06:51,557 S1: scale and thinking about what it means to trust data. 106 00:06:51,997 --> 00:06:56,197 S1: So now she's getting into AI, and she's been doing 107 00:06:56,197 --> 00:06:57,957 S1: a whole lot of really cool stuff with that. So 108 00:06:57,957 --> 00:07:01,837 S1: definitely recommend going to see her talk again. That's June 109 00:07:01,837 --> 00:07:05,797 S1: 26th in Austin. All right. Places I'm going to be 110 00:07:05,797 --> 00:07:09,777 S1: speaking in 2025, and I realized I missed an item 111 00:07:09,777 --> 00:07:13,657 S1: on the list. Various talks and panels. Blackhat USA in 112 00:07:13,657 --> 00:07:17,657 S1: Vegas speaking at the Swiss Cyberstorm in Bern, Switzerland. It's 113 00:07:17,657 --> 00:07:20,777 S1: going to be an AI workshop. Actually, I'm going to 114 00:07:20,777 --> 00:07:22,937 S1: do a workshop for people, which is going to be 115 00:07:22,937 --> 00:07:27,697 S1: a combination of like augmented fabric slash, essentially like personal 116 00:07:27,697 --> 00:07:30,177 S1: AI infrastructure is kind of what that thing's going to 117 00:07:30,177 --> 00:07:34,817 S1: be about. Keynoting Appsec USA in DC in November and 118 00:07:34,817 --> 00:07:40,537 S1: speaking at Blackhat Mia in Saudi Arabia in December. And 119 00:07:40,537 --> 00:07:43,017 S1: the one that I missed is I'm probably going to 120 00:07:43,017 --> 00:07:47,697 S1: be doing something in Sweden as well. Um, I can't 121 00:07:47,737 --> 00:07:50,737 S1: remember the exact date for that one, but, uh, yeah, 122 00:07:51,057 --> 00:07:53,737 S1: I'll probably add that to the list for this next 123 00:07:53,737 --> 00:07:59,577 S1: coming newsletter. On Monday, personal tech Stack updates I'm all 124 00:07:59,577 --> 00:08:02,457 S1: in on cloud code. I'm not going to go into why, 125 00:08:02,497 --> 00:08:08,147 S1: but ultimately it's because here I am going into why, uh, 126 00:08:08,307 --> 00:08:13,787 S1: anthropic uses this as their agentic coding agent, and they've 127 00:08:13,827 --> 00:08:18,027 S1: gone all in on Agentic coding, which means this thing 128 00:08:18,027 --> 00:08:22,427 S1: isn't like hacking together agents. It's actually using agents to 129 00:08:22,467 --> 00:08:25,907 S1: execute the various tasks. And as of today, I just 130 00:08:25,907 --> 00:08:29,027 S1: updated cloud code and now it has planning built in. 131 00:08:29,947 --> 00:08:32,666 S1: So as I've been talking about pretty much everywhere, I 132 00:08:32,667 --> 00:08:38,667 S1: think scaffolding around AI is the superpower that's about to 133 00:08:38,707 --> 00:08:42,867 S1: unlock this whole thing, because the models are already smart. 134 00:08:42,867 --> 00:08:45,547 S1: The problem is they can't think for long. They have 135 00:08:45,547 --> 00:08:49,666 S1: short memories. So we have all these context problems. So 136 00:08:50,347 --> 00:08:55,147 S1: any any person, any intelligent being gets really stupid when 137 00:08:55,146 --> 00:09:00,146 S1: it can't remember longer than, you know, a short conversation. 138 00:09:00,546 --> 00:09:03,027 S1: And then you have to go back into this stored 139 00:09:03,026 --> 00:09:06,967 S1: memory bank. And it's not perfect. I'm talking about rag, 140 00:09:06,967 --> 00:09:10,607 S1: for example. Or if you have actual humans and they 141 00:09:10,607 --> 00:09:13,327 S1: struggle with long term memory or even short term memory, 142 00:09:13,766 --> 00:09:16,766 S1: obviously they're going to be less functional, right? They're going 143 00:09:16,807 --> 00:09:18,927 S1: to be less capable than they used to be before 144 00:09:18,926 --> 00:09:25,327 S1: they lost those abilities. So the ability to um, as 145 00:09:25,367 --> 00:09:29,167 S1: Dwarkesh talks about, learn on the job. Okay. You think 146 00:09:29,166 --> 00:09:32,247 S1: about a knowledge worker. They go in, they do orientation, 147 00:09:32,247 --> 00:09:34,807 S1: they learn on the job. Okay. Guess what? After that 148 00:09:34,807 --> 00:09:38,566 S1: first week, they're way better than when they started because 149 00:09:38,567 --> 00:09:41,806 S1: now they have some knowledge. Well, what happens after six months? 150 00:09:41,807 --> 00:09:44,567 S1: What happens after a year? What happens after they've been 151 00:09:44,567 --> 00:09:48,207 S1: there for 20 years. They now have all this stored 152 00:09:48,207 --> 00:09:52,287 S1: knowledge about this company, which is now part of their context. 153 00:09:53,247 --> 00:09:57,766 S1: But the difference is with a human, they are naturally, fluidly. 154 00:09:57,766 --> 00:10:01,887 S1: Every time they make a decision, they are thinking of 155 00:10:02,016 --> 00:10:06,497 S1: their stored knowledge. They incorporate all of their decisions with 156 00:10:06,497 --> 00:10:10,417 S1: their stored knowledge every time they make any decision. That 157 00:10:10,416 --> 00:10:14,416 S1: is a thing that I currently cannot do. It is struggling. 158 00:10:14,416 --> 00:10:16,576 S1: It's like, oh, I've got to go grab all this 159 00:10:16,577 --> 00:10:20,296 S1: context and stuff it into the prompt. How do you 160 00:10:20,296 --> 00:10:24,416 S1: grab context of 25 years of learning at a job? 161 00:10:25,296 --> 00:10:27,056 S1: How do you know what that even means? How do 162 00:10:27,057 --> 00:10:32,217 S1: you even know what to query to stuff into your 163 00:10:32,256 --> 00:10:37,177 S1: decision making process? This is completely transparent with humans. When 164 00:10:37,176 --> 00:10:40,256 S1: we are like, oh, let me approach this problem from 165 00:10:40,256 --> 00:10:43,097 S1: a holistic standpoint. Let me talk to talk to all 166 00:10:43,097 --> 00:10:46,057 S1: these disparate teams across the org. And I'm going to 167 00:10:46,057 --> 00:10:47,776 S1: gather notes and I'm going to take notes. I'm going 168 00:10:47,776 --> 00:10:50,577 S1: to write them down or whatever. Your brain is going 169 00:10:50,617 --> 00:10:53,417 S1: crazy with all of its knowledge that you've gathered over 170 00:10:53,416 --> 00:10:57,536 S1: these 25 years, and it's making its own queries and 171 00:10:57,536 --> 00:11:01,597 S1: pulling back that context. And these are coming back to 172 00:11:01,597 --> 00:11:04,837 S1: you as kind of like insights or thoughts or creativity. 173 00:11:05,557 --> 00:11:08,156 S1: You're not getting to see exactly what the queries look 174 00:11:08,156 --> 00:11:12,197 S1: like or even what the returned results look like. We 175 00:11:12,197 --> 00:11:15,756 S1: do not have this with AI. We must manually go 176 00:11:16,156 --> 00:11:20,797 S1: and make manual queries into context, like with rag or 177 00:11:20,837 --> 00:11:25,756 S1: like with stuffing other context or context summaries into the 178 00:11:25,756 --> 00:11:29,756 S1: prompts themselves, right? This is the scaffolding that we need. 179 00:11:29,796 --> 00:11:35,277 S1: We need to emulate the process that humans execute naturally 180 00:11:35,437 --> 00:11:39,237 S1: every day, you know, hundreds or thousands or millions of times. 181 00:11:39,237 --> 00:11:42,556 S1: However many times it is that we're making decisions and 182 00:11:42,557 --> 00:11:47,357 S1: little tiny, you know, nano decisions and micro decisions. So 183 00:11:48,756 --> 00:11:53,357 S1: I don't remember how I got onto this, but scaffolding massively, 184 00:11:53,357 --> 00:11:57,997 S1: massively important. And, uh, that's why I think context is, 185 00:11:58,036 --> 00:12:01,646 S1: is so important to all of this stuff. Um. Oh, 186 00:12:01,687 --> 00:12:05,406 S1: I remember how Claude Cote. So they are all in 187 00:12:05,406 --> 00:12:08,446 S1: on the agent stuff, and they are all in on 188 00:12:08,487 --> 00:12:12,047 S1: the planning component. Okay, this planning component is yet another 189 00:12:12,046 --> 00:12:15,807 S1: piece of what I just talked about. It is, you know, 190 00:12:15,847 --> 00:12:19,047 S1: keeping in mind what we're trying to do overall and 191 00:12:19,046 --> 00:12:21,926 S1: then checking off little boxes of like, yeah, we've done this. Yeah, 192 00:12:21,926 --> 00:12:27,886 S1: we've done this. It's all about, uh, context management. Now 193 00:12:27,926 --> 00:12:31,326 S1: context is a little bit, um, overloaded here. I would 194 00:12:31,327 --> 00:12:36,967 S1: say context planning management. Um, learning on the job management, 195 00:12:37,607 --> 00:12:41,127 S1: long term task management. These are the types of things 196 00:12:41,127 --> 00:12:47,766 S1: that really good knowledge workers, like a principle engineer who's 197 00:12:47,766 --> 00:12:51,327 S1: been somewhere for 25 years. These are the types of 198 00:12:51,327 --> 00:12:54,006 S1: things that they are really good at. And these are 199 00:12:54,006 --> 00:12:56,847 S1: the things in my opinion, this is a prediction and 200 00:12:56,847 --> 00:13:01,427 S1: this is just my prediction, is that we're going to 201 00:13:01,426 --> 00:13:05,307 S1: see major advances in those areas, which are actually kind 202 00:13:05,307 --> 00:13:08,227 S1: of like tricks. This is what I've been telling my friend, uh, 203 00:13:08,426 --> 00:13:12,266 S1: Joel ever since 23, that these types of tricks are 204 00:13:12,266 --> 00:13:14,426 S1: the things that are going to push us over the edge. 205 00:13:14,627 --> 00:13:16,667 S1: These are the types of tricks that are going to 206 00:13:17,307 --> 00:13:23,306 S1: hyper jump. I way further ahead. And I think what 207 00:13:23,347 --> 00:13:26,867 S1: I think my theory has been borne out through things 208 00:13:26,867 --> 00:13:33,347 S1: like Post-training post-training is not fundamentally better neural nets. It's 209 00:13:33,347 --> 00:13:38,427 S1: better ways of shaping and getting neural nets to do 210 00:13:38,426 --> 00:13:43,146 S1: what we actually want them to do. Right. So just 211 00:13:43,146 --> 00:13:46,627 S1: zooming out, I think Cloud Code has this figured out. 212 00:13:46,627 --> 00:13:51,067 S1: I think anthropic really deeply understands this. And they are 213 00:13:51,067 --> 00:13:56,847 S1: building agents and planning management and context Management. They're doing really, 214 00:13:56,847 --> 00:14:00,207 S1: really well with it in a way that I'm just 215 00:14:00,247 --> 00:14:02,967 S1: really excited to build things with cloud code in a 216 00:14:02,967 --> 00:14:05,967 S1: way that's not really hitting for me. Uh, on the 217 00:14:05,967 --> 00:14:12,287 S1: other platforms, my secondary backup for cloud code is still cursor. Um, 218 00:14:12,286 --> 00:14:15,687 S1: I was messing with, uh, windsurf for a while and 219 00:14:15,687 --> 00:14:18,167 S1: some other tools, and I still dabble with all the 220 00:14:18,166 --> 00:14:21,527 S1: other tools when I see something, uh, come out or new, 221 00:14:21,527 --> 00:14:24,927 S1: I go and tinker with it. But right now, cloud 222 00:14:24,927 --> 00:14:29,727 S1: code is my main, with cursor being secondary. I'm also 223 00:14:29,727 --> 00:14:33,207 S1: switched over to using Dia as my main daily driver 224 00:14:33,207 --> 00:14:37,567 S1: for my browser. And um, I'm on all the Mac 225 00:14:37,567 --> 00:14:43,206 S1: OS 26 betas. They are fantastic. And uh, specifically having 226 00:14:43,207 --> 00:14:47,847 S1: your phone app on the desktop is fantastic. And, uh, 227 00:14:47,847 --> 00:14:51,807 S1: the podcast app is much improved. You can actually go 228 00:14:51,847 --> 00:14:53,527 S1: above the two x limit. You can go up to 229 00:14:53,527 --> 00:14:57,217 S1: three x now. And yeah, overall I would say Apple 230 00:14:57,217 --> 00:15:01,976 S1: just really did a lot of quality updates. For example 231 00:15:02,017 --> 00:15:07,137 S1: AirPod handoffs are much better. So that's it for updates 232 00:15:07,137 --> 00:15:11,017 S1: I think. And let's get into cybersecurity. Apple quietly fixed 233 00:15:11,017 --> 00:15:16,377 S1: an iPhone zero Day that was used against journalists. Echo 234 00:15:16,377 --> 00:15:22,857 S1: leak using uses markdown syntax to bypass Microsoft 365 Copilot security. 235 00:15:23,337 --> 00:15:25,897 S1: So some researchers found a smart way to steal data 236 00:15:25,897 --> 00:15:31,456 S1: from Copilot by using obscure markdown link formats that Microsoft 237 00:15:31,457 --> 00:15:37,697 S1: did not filter. Researchers turn 2 a.m. Tokyo hotel room 238 00:15:37,697 --> 00:15:43,537 S1: chat into a Netflix RC. So Chubbs, who is a 239 00:15:43,897 --> 00:15:51,337 S1: friend of mine and very, very good, um, attacker, uh, attacker. Um, 240 00:15:51,817 --> 00:15:56,237 S1: what are we saying? Researcher. Bug. Bounty guy. Um. Security 241 00:15:56,277 --> 00:16:00,677 S1: guy focused on offensive security and automated testing. He's also 242 00:16:00,717 --> 00:16:04,317 S1: the creator of Asset note. So he's just one of 243 00:16:04,317 --> 00:16:08,157 S1: the top, you know, top 1% or 1% of the 1%. 244 00:16:08,477 --> 00:16:10,837 S1: I would put him in Sam Curry like up in 245 00:16:10,877 --> 00:16:15,917 S1: that top, top echelon. And, uh, basically they found, uh, 246 00:16:17,797 --> 00:16:21,237 S1: the way they found a way to attack an unclaimed 247 00:16:21,237 --> 00:16:25,677 S1: internal package name called NFC logger. And they use that 248 00:16:25,677 --> 00:16:30,517 S1: to get Archie actually on a Netflix, uh, system or 249 00:16:30,517 --> 00:16:34,757 S1: within Netflix itself. And I'm sure they fixed that. Otherwise 250 00:16:34,757 --> 00:16:39,037 S1: we wouldn't be talking about it. All right. Going to 251 00:16:39,037 --> 00:16:43,477 S1: skip over the Tliose sessions. Uh, offer. That was in 252 00:16:43,477 --> 00:16:46,797 S1: the thing. Uh, don't really want to talk about that. Um, 253 00:16:47,157 --> 00:16:51,087 S1: HTML spec now escapes angle brackets and attributes. Tributes, so 254 00:16:51,087 --> 00:16:53,887 S1: Chrome and other browsers are now going to start escaping 255 00:16:54,847 --> 00:16:59,647 S1: left and right. Caret characters in HTML attributes to prevent 256 00:16:59,647 --> 00:17:06,647 S1: mutation XSS attacks Interpol dismantles 20,000 malicious IPS linked to 257 00:17:06,687 --> 00:17:13,167 S1: 69 malware variants. Cursor security rules project tackles unsafe AI 258 00:17:13,207 --> 00:17:20,206 S1: generated code, Europol says Europol yeah, Europol says stolen data 259 00:17:20,207 --> 00:17:24,526 S1: has become the new underground currency. Phishing messages this is 260 00:17:24,527 --> 00:17:27,927 S1: a quote created by Llms have higher success rate than 261 00:17:27,927 --> 00:17:31,567 S1: those written by humans. And this is a euro pool. 262 00:17:33,367 --> 00:17:41,806 S1: Io CTA report US airlines quietly selling flight data to DHS. Yeah. 263 00:17:41,847 --> 00:17:45,527 S1: Government agencies can search 39 months of flight data, including 264 00:17:45,527 --> 00:17:50,346 S1: passenger names, itineraries, travel dates and credit card numbers, and 265 00:17:50,347 --> 00:17:54,786 S1: other people buying this data are include like the Secret Service, SEC, DEA, 266 00:17:55,147 --> 00:18:01,427 S1: and US Marshals Service. And that's in addition to the 267 00:18:01,466 --> 00:18:06,267 S1: DHS and all its various services. Amazon's AI agents can 268 00:18:06,267 --> 00:18:12,187 S1: build cyber defense signatures in minutes. National security thoughts on 269 00:18:12,547 --> 00:18:16,067 S1: Israel's action against Iran. I'm actually not going to go 270 00:18:16,067 --> 00:18:22,107 S1: super into this one. It's an evolving situation. As everyone knows, 271 00:18:22,427 --> 00:18:26,907 S1: new stuff is happening like every few hours on this, uh, go, 272 00:18:26,907 --> 00:18:28,627 S1: go read the newsletter if you want to see my 273 00:18:28,627 --> 00:18:33,867 S1: take on this. Army is commissioning Big Tech executives as 274 00:18:33,907 --> 00:18:38,547 S1: lieutenant colonel. So for Silicon Valley executives from meta, Palantir 275 00:18:38,907 --> 00:18:43,266 S1: and OpenAI are being brought in as lieutenant colonels to 276 00:18:43,307 --> 00:18:47,997 S1: speed up the government's adoption of technology. And I'm overall 277 00:18:47,997 --> 00:18:52,677 S1: for this, but I also have my spider senses going 278 00:18:52,677 --> 00:18:58,596 S1: off in terms of like dystopia type situation. Again, I, 279 00:18:59,357 --> 00:19:03,956 S1: I see the value of Palantir. I see the value 280 00:19:03,956 --> 00:19:10,077 S1: of winning at almost any cost against China. Right? That 281 00:19:10,077 --> 00:19:12,436 S1: is why I support this type of move from the 282 00:19:12,436 --> 00:19:17,877 S1: military and why I support overall. In some ways, companies 283 00:19:17,877 --> 00:19:23,917 S1: like Palantir and Andriel and you know why these companies exist? Yes, 284 00:19:23,917 --> 00:19:29,077 S1: we need to move super fast. Unfortunately, the cultures and 285 00:19:29,077 --> 00:19:34,957 S1: the ambitions and a whole bunch of externalities often come 286 00:19:35,277 --> 00:19:40,077 S1: with moving really fast and in really ambitious ways. In 287 00:19:40,117 --> 00:19:43,797 S1: with technology companies where you're making tons of money supporting 288 00:19:43,797 --> 00:19:47,456 S1: the military. I don't think I need to spell it out. 289 00:19:47,696 --> 00:19:52,976 S1: There's possibility for harm here, right? There's possibility for harm here. 290 00:19:53,177 --> 00:19:57,537 S1: And I think we have to take that risk. I 291 00:19:57,537 --> 00:19:59,897 S1: think this is the right move to take that risk. 292 00:20:00,177 --> 00:20:03,137 S1: But holy crap, is it scary? Yes. And do we 293 00:20:03,137 --> 00:20:06,937 S1: need oversight? Yes. Do we need to be extremely careful? Yes. 294 00:20:06,976 --> 00:20:09,417 S1: Do we need to not turn our eyes away and 295 00:20:09,417 --> 00:20:15,737 S1: just hope it all works out? Absolutely. Cheap drones will 296 00:20:15,736 --> 00:20:21,457 S1: massively disrupt current large military dominance. Yeah. Israel and Ukraine 297 00:20:21,456 --> 00:20:26,057 S1: are showing that $500 drones can destroy extremely expensive military gear, 298 00:20:26,456 --> 00:20:31,137 S1: which could upend the advantage of countries like, you know, Russia, 299 00:20:31,137 --> 00:20:39,016 S1: the US. We're already seeing it with Russia in Ukraine. Uh, 300 00:20:39,017 --> 00:20:42,016 S1: big scope aside on this one. I'm curious about how 301 00:20:42,017 --> 00:20:45,506 S1: this is going to start trickling into consumer and personal security. 302 00:20:45,507 --> 00:20:48,827 S1: Like when will executive production teams all need to have 303 00:20:48,867 --> 00:20:53,226 S1: like Anti-drone tech as part of their package? I think 304 00:20:53,267 --> 00:20:58,347 S1: probably sooner rather than later. I that apple paper saying 305 00:20:58,387 --> 00:21:02,667 S1: eyes don't reason was highly flawed, and someone named Alex 306 00:21:02,706 --> 00:21:06,706 S1: Lawson says that the paper basically didn't do a good 307 00:21:06,706 --> 00:21:11,226 S1: job and it was easy to make better examples. Although 308 00:21:11,226 --> 00:21:15,507 S1: I've heard someone did a paper like this, this might 309 00:21:15,547 --> 00:21:18,106 S1: have been Alex Lawson. Actually, I don't know if it 310 00:21:18,107 --> 00:21:22,826 S1: was this paper. I heard somebody say that that whole paper, 311 00:21:22,827 --> 00:21:27,466 S1: this whole takedown paper was actually Claude generated, and it 312 00:21:27,466 --> 00:21:30,907 S1: wasn't a real paper. And it was just like, I 313 00:21:30,907 --> 00:21:33,787 S1: don't know, fake or hype or whatever, and that they 314 00:21:33,787 --> 00:21:36,987 S1: didn't expect it to blow up so much. I don't 315 00:21:36,986 --> 00:21:40,186 S1: know if it was this one, but interesting to note. 316 00:21:40,547 --> 00:21:42,706 S1: Also interesting to note, and I'm going to beat up 317 00:21:42,706 --> 00:21:47,686 S1: on myself here for a second. I don't know if 318 00:21:47,686 --> 00:21:51,007 S1: this was the paper that said that it doesn't actually 319 00:21:51,007 --> 00:21:55,047 S1: matter if I know, or it doesn't actually matter if 320 00:21:55,047 --> 00:21:58,167 S1: it was the paper or not. The fact that I 321 00:21:58,167 --> 00:22:03,847 S1: put it in the newsletter is potentially a strike against me. 322 00:22:04,087 --> 00:22:09,407 S1: Think about this. I counted as a strike against myself. Um, 323 00:22:09,726 --> 00:22:12,607 S1: the thing is, I'm not sure if I would do 324 00:22:12,607 --> 00:22:16,686 S1: it differently. It's just maybe I, I don't know. This 325 00:22:16,686 --> 00:22:19,167 S1: is weird. Okay. Because when I put something in the 326 00:22:19,167 --> 00:22:21,927 S1: newsletter and I talk about it, how much of a 327 00:22:21,927 --> 00:22:26,686 S1: high how much have I endorsed this thing and said 328 00:22:26,686 --> 00:22:29,246 S1: for sure that this is a real paper? For sure. 329 00:22:29,247 --> 00:22:33,407 S1: These are real results because I read the paper fully. 330 00:22:33,807 --> 00:22:35,806 S1: Not only that, but I looked at the all the 331 00:22:35,807 --> 00:22:41,407 S1: testing methodologies and I evaluated the data and I confirmed 332 00:22:41,696 --> 00:22:44,897 S1: that their testing was not a lie. That is a 333 00:22:44,936 --> 00:22:47,857 S1: level of depth that I cannot do. For the amount 334 00:22:47,857 --> 00:22:51,777 S1: of papers and stories that I'm doing. So what I'm 335 00:22:51,777 --> 00:22:55,737 S1: saying is it is possible for me to be duped. Um, 336 00:22:55,777 --> 00:22:59,137 S1: and this one very well could be, uh, in fact, 337 00:22:59,177 --> 00:23:00,417 S1: you know what? I'm going to click on a link 338 00:23:00,417 --> 00:23:07,097 S1: here and see if there's possible, like updates. I don't 339 00:23:07,097 --> 00:23:09,976 S1: think so. I don't think so. I don't think it applied. 340 00:23:09,976 --> 00:23:12,857 S1: It doesn't actually matter if it applied. This paper that 341 00:23:12,857 --> 00:23:16,496 S1: I included might be completely legit. And I think it is, 342 00:23:16,497 --> 00:23:19,537 S1: but it doesn't actually matter because this strike against me 343 00:23:19,577 --> 00:23:23,537 S1: still counts, right? So this is a question from me 344 00:23:23,537 --> 00:23:27,857 S1: to myself and also from you to me and from 345 00:23:27,857 --> 00:23:32,777 S1: me to you, to me. You got to watch what 346 00:23:32,777 --> 00:23:35,417 S1: people are posting, and you got to watch with how 347 00:23:35,456 --> 00:23:38,977 S1: sure that they are that that thing is actually real 348 00:23:40,117 --> 00:23:42,877 S1: And whether or not they're just using something that sounds 349 00:23:42,877 --> 00:23:47,316 S1: real to support their own opinion. This one happened to 350 00:23:47,357 --> 00:23:50,437 S1: go with my opinion. Oh, so what do you know? Magically, 351 00:23:50,517 --> 00:23:55,037 S1: I just included in the newsletter as support for my opinion. Well, 352 00:23:55,037 --> 00:23:57,956 S1: what if it turns out there was complete garbage? Well 353 00:23:57,956 --> 00:24:01,476 S1: then I'm just like anyone else who is posting a 354 00:24:01,476 --> 00:24:05,236 S1: random garbage. I would argue I'm not like that, but 355 00:24:05,277 --> 00:24:07,757 S1: I'm saying it looks kind of the same, so I 356 00:24:07,757 --> 00:24:11,277 S1: don't really care how much the difference is. So just 357 00:24:11,277 --> 00:24:14,037 S1: keep in mind. And not just for me, but for yourself. 358 00:24:14,677 --> 00:24:17,877 S1: You got to watch for whether or not evidence that's 359 00:24:17,877 --> 00:24:22,357 S1: coming in is only air quote evidence, because how much 360 00:24:22,476 --> 00:24:26,117 S1: time have you actually spent looking at the evidence, redoing 361 00:24:26,117 --> 00:24:31,597 S1: the paper yourself, you know. And so how much are 362 00:24:31,597 --> 00:24:37,637 S1: we using external validation or citations pointing to a paper 363 00:24:37,637 --> 00:24:43,216 S1: Or as support for the paper being legit. And I 364 00:24:43,257 --> 00:24:46,577 S1: think it gets more and more serious the further we go. Um, 365 00:24:46,617 --> 00:24:49,657 S1: the good news is, a big part of my personal 366 00:24:49,936 --> 00:24:53,897 S1: AI structure is going to be to find stuff like this. Uh, 367 00:24:53,897 --> 00:24:55,737 S1: I want to be able to find stuff like this 368 00:24:55,736 --> 00:24:59,177 S1: with my AI infrastructure where it's like, I don't know, 369 00:24:59,417 --> 00:25:01,777 S1: seems like they use a lot of, uh, catchy AI 370 00:25:01,817 --> 00:25:04,137 S1: terms in here, and it looks like they were really 371 00:25:04,177 --> 00:25:07,457 S1: hand-wavy with their examples. Like, I want to be able 372 00:25:07,456 --> 00:25:10,617 S1: to catch this stuff for myself. Just something to keep 373 00:25:10,617 --> 00:25:16,657 S1: in mind. OpenAI's O3 Pro is very smart, but context hungry. Agreed. 374 00:25:17,177 --> 00:25:21,537 S1: Man killed by police after spiraling into a ChatGPT driven psychosis. 375 00:25:22,657 --> 00:25:26,737 S1: Google plans to cut ties with scale AI because basically 376 00:25:26,736 --> 00:25:30,337 S1: meta bought like half of them or something. CIO wants 377 00:25:30,337 --> 00:25:34,897 S1: to clone staff as digital twins and AI agents. UC 378 00:25:34,897 --> 00:25:39,067 S1: San Diego CIO wants to extract knowledge from experienced IT 379 00:25:39,107 --> 00:25:45,907 S1: staff and create digital twins that handle routine problems. Well, yeah, obviously, 380 00:25:46,027 --> 00:25:50,667 S1: you know, you got somebody, uh, greybeard, you know, Chris Myers, 381 00:25:51,147 --> 00:25:54,667 S1: who's been at the company for 41 years, and they 382 00:25:54,667 --> 00:25:58,267 S1: know everything. And anyone who, you know, finds something broken 383 00:25:58,267 --> 00:26:01,867 S1: that nobody could fix, they all call Chris Myers. So 384 00:26:01,867 --> 00:26:05,586 S1: the CEO is like, let's go find Chris Myers. Let's 385 00:26:05,587 --> 00:26:10,067 S1: go interview him for seven weeks in a dark, padded 386 00:26:10,067 --> 00:26:12,987 S1: room and pull out all of his knowledge and put 387 00:26:12,986 --> 00:26:15,667 S1: that into an AI agent. And now you could just 388 00:26:15,706 --> 00:26:18,947 S1: ask a virtual Chris Myers. Well, obviously we need this 389 00:26:18,946 --> 00:26:21,347 S1: for everything, right? We need this for the best doctors. 390 00:26:21,347 --> 00:26:24,387 S1: We need this for the best psychiatrists. We need this 391 00:26:24,387 --> 00:26:27,627 S1: for everything. This is you know, this is the problem. When, 392 00:26:28,107 --> 00:26:31,987 S1: you know, extremely experienced people leave a company, you lose 393 00:26:31,986 --> 00:26:36,087 S1: all the knowledge that they, uh, took with them. And 394 00:26:36,087 --> 00:26:41,966 S1: unfortunately also for I, it's also knowledge that you can't 395 00:26:41,966 --> 00:26:47,647 S1: possibly extract from them efficiently, right? You can't possibly ask 396 00:26:47,647 --> 00:26:51,726 S1: enough questions to to gather their knowledge. Their knowledge is, 397 00:26:52,486 --> 00:26:56,887 S1: let's call it infinite. 41 years of knowledge essentially baked 398 00:26:56,887 --> 00:27:01,726 S1: into an AI model. That's a great analogy actually it. 399 00:27:02,206 --> 00:27:06,167 S1: How many questions do you have to ask? You know, uh, 400 00:27:06,647 --> 00:27:11,367 S1: O3 to get all of O3 knowledge out of it? 401 00:27:11,686 --> 00:27:15,046 S1: The answer is an infinite number of questions. So it 402 00:27:15,047 --> 00:27:18,367 S1: doesn't matter how much you interview Chris Myers, who's been 403 00:27:18,367 --> 00:27:21,686 S1: there for 41 years, you're not going to get enough. Right? 404 00:27:21,726 --> 00:27:28,327 S1: So so this is a serious challenge. ChatGPT dominates LLM 405 00:27:28,367 --> 00:27:31,927 S1: usage at 80%, 86% market share. I did not know 406 00:27:31,927 --> 00:27:35,216 S1: it was that high technology fabric. Summary of the massive 407 00:27:35,257 --> 00:27:39,297 S1: Google outage. Yeah. This is this is really cool. Um, 408 00:27:40,337 --> 00:27:42,737 S1: so they put out the, you know, Google put out 409 00:27:42,736 --> 00:27:47,617 S1: their their giant paper saying what happened. And when I see, 410 00:27:47,657 --> 00:27:50,817 S1: you know, a 14 page paper or whatever. And it's 411 00:27:50,817 --> 00:27:53,817 S1: about an outage and describing an outage, it's not like 412 00:27:53,817 --> 00:27:57,377 S1: deep human knowledge that I'm getting. That's where I run Extraction's. 413 00:27:57,736 --> 00:28:03,216 S1: So I've got this, um, pattern called create story explanation, 414 00:28:03,696 --> 00:28:09,377 S1: create underscore story story underscore explanation. It is one of 415 00:28:09,377 --> 00:28:14,337 S1: my favorite patterns. It explains anything at like a ninth 416 00:28:14,337 --> 00:28:17,897 S1: grade level in kind of a story format. So this happened, 417 00:28:17,897 --> 00:28:22,296 S1: then this happened, then this happened. And it's just very clear. 418 00:28:22,456 --> 00:28:26,097 S1: So I recommend checking it out if you care about 419 00:28:26,097 --> 00:28:28,977 S1: that outage, but more importantly using it against other stuff. 420 00:28:29,537 --> 00:28:33,797 S1: YouTube officially beats all other streaming platforms. In US viewership, 421 00:28:34,157 --> 00:28:40,437 S1: they have 13% of all US TV viewership. Oh, wow. Okay. 422 00:28:40,477 --> 00:28:44,277 S1: I thought it was 13% of streaming. Okay, I read 423 00:28:44,277 --> 00:28:52,117 S1: that wrong. 13% of all US TV viewership. That's insane. 424 00:28:55,157 --> 00:29:00,877 S1: That's insane. Okay, YouTube by itself, 13% of TV. Yeah, 425 00:29:00,917 --> 00:29:03,037 S1: it seems very high. I bet you that goes up 426 00:29:03,077 --> 00:29:07,437 S1: crazy amounts in the next few years. It's basically the 427 00:29:07,437 --> 00:29:09,837 S1: only TV that I watched. I was actually at a 428 00:29:09,837 --> 00:29:14,437 S1: resort yesterday in Napa, and they only had regular TV, 429 00:29:14,437 --> 00:29:18,036 S1: and it was like it was like a commercial rectangle. 430 00:29:19,197 --> 00:29:22,197 S1: All I did was watch commercials and occasionally they would 431 00:29:22,197 --> 00:29:26,237 S1: show like part of a show. Anyway, Amazon joins the 432 00:29:26,237 --> 00:29:34,207 S1: big nuclear party buying 1.2GW, uh, for AWS. tvOS 26 433 00:29:34,527 --> 00:29:37,847 S1: hints at built in camera for a new Apple TV 4K. 434 00:29:38,567 --> 00:29:40,727 S1: I wish it was an Apple TV 8-K, and I 435 00:29:40,727 --> 00:29:43,967 S1: wish it had ten Gigabit Ethernet. That's what I wish. 436 00:29:44,127 --> 00:29:45,887 S1: But I would like to see a camera because I'd 437 00:29:45,887 --> 00:29:49,727 S1: like to do FaceTime with, um, in the living room 438 00:29:49,727 --> 00:29:52,447 S1: sitting on the couch. That would be nice. I also 439 00:29:52,447 --> 00:29:56,527 S1: don't want a camera facing the living room, so I 440 00:29:56,527 --> 00:30:01,447 S1: would need one of those flip dongle cover things. Apple's 441 00:30:01,447 --> 00:30:05,207 S1: iPad OS 26 finally makes the iPad a real computer. 442 00:30:05,207 --> 00:30:09,967 S1: Everyone is talking about this. People are saying that this 443 00:30:10,007 --> 00:30:14,887 S1: actually allows you to use an iPad as a laptop, 444 00:30:15,407 --> 00:30:19,247 S1: almost like very closely. And it's kind of been universal 445 00:30:19,247 --> 00:30:22,087 S1: that people are saying, Holy crap, this is really, really good. 446 00:30:22,687 --> 00:30:26,007 S1: I have not. I don't use my iPads much. And 447 00:30:26,087 --> 00:30:29,947 S1: the reason is I just didn't like the OS. The 448 00:30:29,947 --> 00:30:32,347 S1: only thing that I use it for is reading and 449 00:30:32,347 --> 00:30:36,187 S1: drawing and sketching and doing stuff like that with like freeform. 450 00:30:36,667 --> 00:30:39,187 S1: And for that I really love it. And for reading, 451 00:30:39,187 --> 00:30:41,227 S1: I really love it. But I don't use it as 452 00:30:41,227 --> 00:30:44,787 S1: a computer because I prefer laptops. Maybe that will change 453 00:30:46,107 --> 00:30:48,667 S1: the argument that it's time to kill Siri. I did 454 00:30:48,667 --> 00:30:52,147 S1: not know that I believe this, but I definitely do. 455 00:30:52,187 --> 00:30:57,267 S1: I think once Apple fixes its AI interface, its universal 456 00:30:57,267 --> 00:31:00,747 S1: AI to life OS that I've been talking about, they 457 00:31:00,747 --> 00:31:05,947 S1: need to call it something new. Siri has too much baggage. I. 458 00:31:07,547 --> 00:31:08,787 S2: I can show them if you ask. 459 00:31:09,307 --> 00:31:15,067 S1: Hey Siri, stop the one time Siri actually listens when 460 00:31:15,067 --> 00:31:20,867 S1: I'm trying to record, um, it essentially that system does 461 00:31:20,867 --> 00:31:24,907 S1: not turn on and turn off lights reliably. It has 462 00:31:24,907 --> 00:31:28,277 S1: so many problems. It's been over ten years. It had 463 00:31:28,277 --> 00:31:31,597 S1: its attempt. It had its chances, and now it needs 464 00:31:31,597 --> 00:31:34,757 S1: a new name. I thoroughly believe this. It's just a 465 00:31:34,757 --> 00:31:38,397 S1: marketing concept. When you have enough bad baggage associated with something, 466 00:31:38,797 --> 00:31:41,237 S1: you got to flip it over. You got to retire it. 467 00:31:41,957 --> 00:31:45,997 S1: Nvidia writes off China revenue in company forecasts because, you 468 00:31:45,997 --> 00:31:50,157 S1: know it's too unpredictable. With the current administration, Waymo rides 469 00:31:50,157 --> 00:31:52,997 S1: cost more than Uber or Lyft, but people are happy 470 00:31:52,997 --> 00:31:56,677 S1: paying it so often 5 or $10 more for a 471 00:31:56,717 --> 00:32:00,757 S1: ride versus like Uber or a taxi. But people are 472 00:32:00,757 --> 00:32:03,277 S1: still paying it and they're still making money because it's 473 00:32:03,277 --> 00:32:08,036 S1: so much of a better experience. Humans. Trump ends protection 474 00:32:08,037 --> 00:32:12,157 S1: for Afghans as Congress scrambles to intervene. I'm really upset 475 00:32:12,157 --> 00:32:18,397 S1: about this, really upset that these Afghanis who had helped 476 00:32:18,397 --> 00:32:22,597 S1: the US military during the war and who cannot go 477 00:32:22,597 --> 00:32:26,537 S1: back because they are being hunted by the Taliban and 478 00:32:26,537 --> 00:32:32,857 S1: because their families are under like watch by the Taliban, 479 00:32:32,857 --> 00:32:36,336 S1: they are in grave danger. Their families are and they are. 480 00:32:36,537 --> 00:32:38,296 S1: And if they go back and try to link with 481 00:32:38,297 --> 00:32:41,897 S1: their families because they're forcibly kicked out of the US, 482 00:32:42,897 --> 00:32:47,977 S1: it is essentially damning them and their families, potentially to die. 483 00:32:48,857 --> 00:32:53,297 S1: And this is after they served and sacrificed essentially for 484 00:32:53,297 --> 00:32:56,817 S1: the US military. I do not believe we should treat 485 00:32:56,817 --> 00:33:03,217 S1: allies like that. Robin I system makes first autonomous scientific discovery. 486 00:33:03,217 --> 00:33:06,857 S1: So this is future houses. Robin I so I talked 487 00:33:06,857 --> 00:33:13,177 S1: about this story of automated, you know, science assistance. And 488 00:33:13,177 --> 00:33:16,017 S1: this is even more it's more end to end the 489 00:33:16,017 --> 00:33:21,737 S1: entire research process. So it, um, it discovered that ripasudil 490 00:33:22,057 --> 00:33:28,306 S1: could treat dry macular degeneration by orchestrating multiple specialized agents 491 00:33:28,307 --> 00:33:33,347 S1: to handle the entire research process autonomously in just 2.5 months. 492 00:33:34,107 --> 00:33:37,667 S1: And who knows how much of that is like the 493 00:33:37,667 --> 00:33:41,747 S1: entire air quotes entire research process, but supposedly it went 494 00:33:41,747 --> 00:33:46,546 S1: end to end and actually worked. So this is the 495 00:33:46,547 --> 00:33:49,347 S1: type of thing that I think is really, really exciting. 496 00:33:50,067 --> 00:33:55,067 S1: Ultra black paint might solve satellite light pollution crisis. Although 497 00:33:55,067 --> 00:33:57,907 S1: I'm curious, wouldn't you just see black dots moving through 498 00:33:57,947 --> 00:34:03,347 S1: your pictures, blocking out stars and, you know, the moon 499 00:34:03,347 --> 00:34:07,987 S1: or whatever? I guess the black dots wouldn't be as annoying. 500 00:34:09,867 --> 00:34:12,667 S1: It seems like they could still overwrite stars. If you 501 00:34:12,707 --> 00:34:14,907 S1: are trying to take a picture of the starry sky, 502 00:34:15,627 --> 00:34:18,787 S1: you might still see artifacts, but it won't be nearly 503 00:34:18,787 --> 00:34:22,967 S1: as bad as like a bright white line Him drawing 504 00:34:22,967 --> 00:34:28,447 S1: over your picture, actually multiple bright white lines. But, uh, 505 00:34:28,447 --> 00:34:31,727 S1: I wonder how possible this is, like, are we going 506 00:34:31,727 --> 00:34:33,727 S1: to pull all the ones that are up there down? 507 00:34:33,727 --> 00:34:37,647 S1: Are we going to paint the ones that are up there? Doubtful. 508 00:34:38,967 --> 00:34:41,287 S1: So this would involve painting all the new ones that 509 00:34:41,287 --> 00:34:46,367 S1: are going up, which unfortunately is many thousands. And the 510 00:34:46,367 --> 00:34:49,327 S1: Pentagon has evidently been pushing Americans to believe in UFOs 511 00:34:49,327 --> 00:34:53,887 S1: for decades. And evidently, according to this report, this is 512 00:34:53,887 --> 00:34:57,487 S1: to cover up actual secret weapons programs. So they basically 513 00:34:57,487 --> 00:35:00,767 S1: circulate this stuff on purpose to get people talking about 514 00:35:01,327 --> 00:35:07,967 S1: a red herring discovery. Fiddle ITM brings malicious traffic detection 515 00:35:07,967 --> 00:35:10,327 S1: to man of the middle proxy. Oh, I'm not going 516 00:35:10,327 --> 00:35:11,847 S1: to actually read all of these, so I'm just going 517 00:35:11,887 --> 00:35:14,047 S1: to scroll through and look for some really cool ones, 518 00:35:14,167 --> 00:35:17,007 S1: or if any are from my friends. Let's see. Boom 519 00:35:17,007 --> 00:35:20,407 S1: boom boom boom. Oh, Sam Altman's gentle singularity. You probably 520 00:35:20,407 --> 00:35:22,417 S1: want to read this one. I've also got a fabric 521 00:35:22,457 --> 00:35:25,297 S1: extracted predictions that comes out of it in the newsletter, 522 00:35:25,297 --> 00:35:34,137 S1: so check that out. And yeah, I think that's about it. Okay, 523 00:35:34,137 --> 00:35:36,057 S1: this is the end of the standard edition of the podcast, 524 00:35:36,057 --> 00:35:38,257 S1: which includes just the news items for the week to 525 00:35:38,297 --> 00:35:40,377 S1: get the rest of the episode, which includes much more 526 00:35:40,377 --> 00:35:43,817 S1: of my analysis, the ideas section, and the weekly member essay. 527 00:35:43,817 --> 00:35:46,497 S1: Please consider becoming a member. As a member, you get 528 00:35:46,537 --> 00:35:49,297 S1: access to all sorts of stuff, most importantly, access to 529 00:35:49,297 --> 00:35:52,457 S1: our extraordinary community of over a thousand brilliant and kind 530 00:35:52,497 --> 00:35:56,897 S1: people in industries like cybersecurity, AI, and the humanities. You 531 00:35:56,897 --> 00:35:59,577 S1: also get access to the UL Book Club, dedicated member 532 00:35:59,577 --> 00:36:02,577 S1: content and events, and lots more. Plus, you'll get a 533 00:36:02,577 --> 00:36:04,897 S1: dedicated podcast feed that you can put into your client 534 00:36:04,897 --> 00:36:07,417 S1: that gets you the full member edition of the podcast 535 00:36:07,417 --> 00:36:09,457 S1: that basically doesn't have this in it, and just goes 536 00:36:09,457 --> 00:36:11,937 S1: all the way through with all the different sections. So 537 00:36:11,977 --> 00:36:13,817 S1: to become a member and get all that, just head 538 00:36:13,817 --> 00:36:18,737 S1: over to Daniel Music.com. That's Daniel Miessler, and we'll see 539 00:36:18,737 --> 00:36:19,377 S1: you next time.