1 00:00:19,212 --> 00:00:27,852 S1: Hey. What's up? All right, episode 497. Yeah. Hope things 2 00:00:27,852 --> 00:00:31,052 S1: are starting off well this week. Um, got some lab 3 00:00:31,052 --> 00:00:35,772 S1: results back. This is definitely too much information, but, uh, 4 00:00:37,531 --> 00:00:42,052 S1: c metabolic is pristine. Liver is great. Kidneys not so good. 5 00:00:43,692 --> 00:00:46,812 S1: Got to improve kidney function. Uh, turns out too much 6 00:00:46,812 --> 00:00:51,132 S1: creatine is a bad thing. I was actually getting dizziness 7 00:00:51,732 --> 00:00:55,972 S1: from ten grams of creatine, so I think I'm going 8 00:00:56,012 --> 00:01:01,012 S1: to have to go back down to five, maybe even lower. Uh, 9 00:01:01,012 --> 00:01:05,972 S1: because it turns out creatine is not great for the kidneys. And, um, 10 00:01:06,732 --> 00:01:09,732 S1: it's not like real bad, but it's just like, not optimal. 11 00:01:09,732 --> 00:01:15,172 S1: So I need to optimize that, uh, HDL a little 12 00:01:15,172 --> 00:01:18,182 S1: bit too low, LDL a little bit too high. So 13 00:01:19,262 --> 00:01:22,742 S1: I've basically been super sedentary for like the last 4 14 00:01:22,742 --> 00:01:26,982 S1: or 5 months because of working on Kai and all the, uh, 15 00:01:28,062 --> 00:01:30,941 S1: all the crazy stuff going on with AI and projects and, 16 00:01:31,102 --> 00:01:36,302 S1: you know, all the typical stuff, all the crazy world stuff. So, um, 17 00:01:36,301 --> 00:01:39,741 S1: starting to try to adjust to this, uh, stand up 18 00:01:39,742 --> 00:01:45,982 S1: alarms for standing up with my standing desk, um, foam 19 00:01:45,982 --> 00:01:50,262 S1: seating pad with, like, a cutout. So I'm not, like, sitting, um. 20 00:01:52,302 --> 00:01:57,582 S1: In an uncomfortable position all day. Mandatory cardio twice a week. Uh, 21 00:01:57,582 --> 00:02:04,662 S1: table tennis or rucking weights at least twice a week. Uh, 22 00:02:04,662 --> 00:02:06,902 S1: more coffee to balance the Celsius. So that was the 23 00:02:06,902 --> 00:02:11,382 S1: other problem. Wasn't just creatine. I'm drinking like two Celsius 24 00:02:11,422 --> 00:02:17,142 S1: a day. Sometimes more, but usually only 1 to 2, 25 00:02:18,062 --> 00:02:22,942 S1: and energy drinks and creatine are two of the things 26 00:02:22,942 --> 00:02:28,262 S1: on the list for, um, things to avoid. If you 27 00:02:28,262 --> 00:02:35,262 S1: have kidney function that's, uh, not optimal. And exercise. Yeah. 28 00:02:35,302 --> 00:02:38,182 S1: Raises HDL. So let me know if you're on a 29 00:02:38,181 --> 00:02:41,542 S1: similar journey. That's kind of where I'm at. Oh sugar 30 00:02:41,542 --> 00:02:45,662 S1: was great. So my metabolic health was ideal. It was 31 00:02:45,662 --> 00:02:48,662 S1: like perfect. So that was that was a great story. 32 00:02:51,062 --> 00:02:56,662 S1: Okay Kai. Update massive context loading. Breakthrough. So I am 33 00:02:56,662 --> 00:03:01,901 S1: now doing a really cool thing where I'm basically I 34 00:03:01,942 --> 00:03:06,742 S1: changed the hook to point to load dynamic context, and 35 00:03:06,782 --> 00:03:10,101 S1: I move that to point two load dynamic context MD 36 00:03:10,181 --> 00:03:14,142 S1: so now my routing is actually in the MD file, 37 00:03:14,142 --> 00:03:19,232 S1: which means I can actually navigate that whole thing contextually. 38 00:03:19,392 --> 00:03:23,272 S1: Instead of having to edit TypeScript code, I am now 39 00:03:23,272 --> 00:03:28,672 S1: able to write it out in text in markdown, basically describing. Yeah. 40 00:03:28,712 --> 00:03:32,912 S1: If you see this type of question, um, that means 41 00:03:32,912 --> 00:03:34,752 S1: I want you to use the researcher. If you see 42 00:03:34,752 --> 00:03:37,232 S1: this type of thing, that means dynamically load this piece 43 00:03:37,232 --> 00:03:40,152 S1: of context. If you see me talk about this project, 44 00:03:40,152 --> 00:03:44,032 S1: that means go load this piece of context, which means 45 00:03:44,032 --> 00:03:49,392 S1: my primary boot load just gets my core context right. 46 00:03:49,432 --> 00:03:51,752 S1: So this whole UFC system that I talked about in 47 00:03:51,752 --> 00:03:57,112 S1: the video is, um, it's all based on only loading 48 00:03:57,112 --> 00:03:59,672 S1: what you need to load at that particular moment and 49 00:03:59,672 --> 00:04:05,672 S1: keeping the context extremely clean. So, uh, this is just 50 00:04:05,952 --> 00:04:08,512 S1: it's really powerful and it's working perfectly. Like I'm not 51 00:04:08,512 --> 00:04:12,472 S1: having any misses. It's actually dynamically loading exactly what I need. 52 00:04:12,752 --> 00:04:15,352 S1: And it's not just the context files, but it actually 53 00:04:15,352 --> 00:04:20,032 S1: loads the correct researcher as well. Or I'm sorry, the 54 00:04:20,032 --> 00:04:26,192 S1: correct agent, which could be researcher, engineer, designer or pentester. 55 00:04:26,392 --> 00:04:30,072 S1: It can load all those different agents according to the 56 00:04:30,072 --> 00:04:33,392 S1: type of asset that, um, that I give it. Right. 57 00:04:33,432 --> 00:04:38,592 S1: So pretty. Pretty cool. And, um, speaking of Chi Pi, 58 00:04:38,632 --> 00:04:44,551 S1: speaking of the whole Pi system, I am open sourcing it. 59 00:04:44,592 --> 00:04:46,992 S1: In fact, I already have. In fact, it's already live. 60 00:04:47,552 --> 00:04:49,592 S1: And I was just going to put the framework up 61 00:04:49,592 --> 00:04:51,712 S1: there and say, hey, look, I'm going to add stuff later. 62 00:04:51,712 --> 00:04:56,592 S1: But then I just started adding stuff. Um, and now 63 00:04:56,592 --> 00:04:58,792 S1: I've added a whole bunch of stuff. It's kind of 64 00:04:58,791 --> 00:05:01,272 S1: like the basic structure. It doesn't have the voice stuff 65 00:05:01,272 --> 00:05:05,632 S1: in there yet, but I do have my basic outline 66 00:05:05,632 --> 00:05:08,912 S1: for the main context file and the main tools file, 67 00:05:10,152 --> 00:05:14,002 S1: and then the pointers to where like, it's pretty obvious 68 00:05:14,002 --> 00:05:16,642 S1: where to go. Put the other context for the other pieces. 69 00:05:16,642 --> 00:05:20,482 S1: So that is a lot of infrastructure there. I said 70 00:05:20,481 --> 00:05:23,962 S1: in one of the git commits, I'm probably definitely going 71 00:05:24,002 --> 00:05:27,681 S1: to leak something because I'm putting so much of like 72 00:05:27,682 --> 00:05:31,402 S1: my actual stuff out there, and it's like, no secrets. 73 00:05:31,442 --> 00:05:35,442 S1: Like I'm literally sharing the actual files. A couple of 74 00:05:35,442 --> 00:05:39,282 S1: things are like redacted for like keys and stuff like that. 75 00:05:39,282 --> 00:05:42,121 S1: But other than that, like, this is seriously the exact 76 00:05:42,122 --> 00:05:45,202 S1: infrastructure that Cai is using, and you're going to have 77 00:05:45,202 --> 00:05:47,601 S1: to tweak it for that reason, right? Because I still 78 00:05:47,601 --> 00:05:50,522 S1: have Cai's name in there. I still have like exactly 79 00:05:50,522 --> 00:05:53,882 S1: what I'm saying to him, exactly what what his personality 80 00:05:53,882 --> 00:05:56,601 S1: is and stuff like that. So you'll have to customize 81 00:05:56,601 --> 00:06:00,642 S1: some stuff. But the whole purpose of this is going forward. 82 00:06:00,642 --> 00:06:04,122 S1: When you're watching videos on YouTube that I'm talking about, 83 00:06:05,162 --> 00:06:07,202 S1: I'm not going to say, oh, here's like a gist 84 00:06:07,202 --> 00:06:11,412 S1: file or like, here's some code in the YouTube description. 85 00:06:11,692 --> 00:06:14,171 S1: It's going to be like, hey, go check the Pi 86 00:06:14,212 --> 00:06:17,812 S1: repository and you will see it. And the Pi repository 87 00:06:17,812 --> 00:06:21,972 S1: is set up with a very clear mission. It's like, look, 88 00:06:22,012 --> 00:06:26,012 S1: enable humans to have this functionality. It's the same as fabric, 89 00:06:26,412 --> 00:06:30,092 S1: the same mission as fabric. And now fabric basically falls 90 00:06:30,092 --> 00:06:33,772 S1: under this, right? It falls under pi because fabric is 91 00:06:33,772 --> 00:06:38,452 S1: just one of the sets of tools that I have available. Um, 92 00:06:38,452 --> 00:06:41,492 S1: so Kai has full access to fabric, has full access 93 00:06:41,492 --> 00:06:45,291 S1: to my foundry stuff, which is the fobs and the 94 00:06:45,291 --> 00:06:49,652 S1: custom commands, the custom cloud commands, like all the different stuff, 95 00:06:49,652 --> 00:06:53,731 S1: it's all in the commands directory and in mic, and 96 00:06:53,731 --> 00:06:56,572 S1: Kai can use all of it. And that is the 97 00:06:56,572 --> 00:06:59,812 S1: whole purpose of this entire project. And it is now 98 00:06:59,812 --> 00:07:03,931 S1: open source. So it's capital Pi is the name of 99 00:07:03,932 --> 00:07:13,132 S1: the project on GitHub. Okay. Oh, yeah. Just thought of 100 00:07:13,132 --> 00:07:15,452 S1: a cool. Um. I was sitting here trying to write 101 00:07:15,452 --> 00:07:19,812 S1: the newsletter, uh, and just kept adding more stuff. So 102 00:07:19,812 --> 00:07:22,412 S1: I've got this one. I just added it. So sick. 103 00:07:22,452 --> 00:07:26,972 S1: It's called capture learning. It's a command. And so the 104 00:07:26,972 --> 00:07:31,532 S1: idea is, if I have a I'm working on something 105 00:07:31,532 --> 00:07:35,532 S1: with Kai. We make something new or we're troubleshooting something 106 00:07:35,532 --> 00:07:37,332 S1: and we finally get it to work. I'm just going 107 00:07:37,372 --> 00:07:41,452 S1: to say, hey, write this down or document this, or 108 00:07:41,892 --> 00:07:45,292 S1: capture learning or whatever. And that same thing that I 109 00:07:45,292 --> 00:07:50,052 S1: told you about before, um, that dynamically loads context. Well, 110 00:07:50,052 --> 00:07:55,012 S1: it could also listen because it's being triggered on the prompt, 111 00:07:55,412 --> 00:07:59,532 S1: the user prompt submit that hook is what's actually triggering this. 112 00:07:59,812 --> 00:08:04,452 S1: So if Kai hears anything like that, Kai now knows 113 00:08:04,452 --> 00:08:08,092 S1: to go and execute capture learning. What does that do? 114 00:08:08,382 --> 00:08:13,342 S1: that drops it into the Learning Context directory. The learning 115 00:08:13,342 --> 00:08:19,862 S1: context directory includes a full report of the lesson, the 116 00:08:20,022 --> 00:08:25,102 S1: false assumption that we had, the mistake that we were making, 117 00:08:25,102 --> 00:08:28,382 S1: or whatever talks through the problem, talks through the process 118 00:08:28,382 --> 00:08:31,262 S1: because it has the whole log of our conversation, the 119 00:08:31,262 --> 00:08:34,222 S1: process of how we got there and then what the 120 00:08:34,222 --> 00:08:39,142 S1: solution was and then what the takeaway is. So that's 121 00:08:39,142 --> 00:08:41,622 S1: documented for each one. I'm going to have that going 122 00:08:41,622 --> 00:08:44,422 S1: forward now. And guess what you could do. You could 123 00:08:44,422 --> 00:08:51,902 S1: now say something like Harvest Learnings and then have that 124 00:08:51,902 --> 00:08:55,142 S1: go and compare your current documentation, your current workflows, your 125 00:08:55,142 --> 00:09:00,102 S1: current pipelines, your current context and instructions and everything, and 126 00:09:00,102 --> 00:09:02,862 S1: compare it with the lessons learned and say, is there 127 00:09:02,862 --> 00:09:05,742 S1: anything here that we should update based on the lessons 128 00:09:05,742 --> 00:09:08,502 S1: learned over the last week or the last year or whatever. 129 00:09:09,142 --> 00:09:11,302 S1: So I just think this is a really cool sort 130 00:09:11,342 --> 00:09:18,462 S1: of meta improvement concept. All right. Cybersecurity. Npm's most used 131 00:09:18,462 --> 00:09:23,862 S1: UI packages got backdoored to hijack crypto. Yeah. This, uh, 132 00:09:24,422 --> 00:09:28,222 S1: aikido company evidently found this, and, um, put out a 133 00:09:28,222 --> 00:09:35,262 S1: report about it. It's 18 different packages and, uh. Yeah, 134 00:09:35,302 --> 00:09:40,662 S1: rewritten to steal crypto. Uh, lots of supply chain stuff 135 00:09:40,662 --> 00:09:44,702 S1: going on. Uh, I really like the idea of using. 136 00:09:46,982 --> 00:09:49,822 S1: AI for many eyes. I feel like many eyes is 137 00:09:49,822 --> 00:09:53,901 S1: like my my big, uh, thing I keep coming back 138 00:09:53,902 --> 00:09:57,381 S1: to with with AI. And I think I talked about 139 00:09:57,381 --> 00:10:01,262 S1: this maybe last week or the week before, but SSL, 140 00:10:01,422 --> 00:10:05,342 S1: you remember the giant SSL vulnerability? I think we still 141 00:10:05,342 --> 00:10:08,832 S1: called it SSL back then. Maybe it was already TLS, 142 00:10:08,832 --> 00:10:12,552 S1: but anyway, that was a huge vuln and everyone was like, hey, 143 00:10:12,552 --> 00:10:14,992 S1: how could this possibly happen? This is open source. And 144 00:10:14,992 --> 00:10:18,631 S1: it's like, well, turns out open source is just a guy. 145 00:10:18,752 --> 00:10:21,552 S1: I heard someone say that recently. Open source is just 146 00:10:21,552 --> 00:10:25,912 S1: a guy. You know, it's just Chris Richards or whatever 147 00:10:25,912 --> 00:10:29,632 S1: who lives in Idaho. Turns out that whole open source, 148 00:10:29,631 --> 00:10:33,312 S1: many eyes things. Turns out it's just Chris and Chris 149 00:10:33,312 --> 00:10:35,472 S1: hurt his ankle and he hasn't been online in a while. 150 00:10:35,472 --> 00:10:39,592 S1: You know what I mean? So it's like Many Eyes 151 00:10:39,631 --> 00:10:41,631 S1: didn't turn out to be what we thought it was 152 00:10:41,631 --> 00:10:45,592 S1: going to be. It's because what it means, what many eyes, 153 00:10:45,832 --> 00:10:48,511 S1: what we thought many eyes was going to be was 154 00:10:48,752 --> 00:10:52,192 S1: many eyes are looking at the project, but what it 155 00:10:52,192 --> 00:10:55,032 S1: turned out to be was many eyes could look at 156 00:10:55,032 --> 00:10:59,791 S1: the project, but they weren't actually. And so this is 157 00:10:59,792 --> 00:11:05,762 S1: why I love the idea of AI. Many eyes. Millions 158 00:11:05,761 --> 00:11:09,922 S1: of eyes. Right. Cheap eyes that don't get tired, don't 159 00:11:09,922 --> 00:11:16,162 S1: get sick or whatever. And you just stick them on 160 00:11:16,162 --> 00:11:19,602 S1: the internet. You're just like, hey, go watch Chris's project. 161 00:11:20,042 --> 00:11:23,602 S1: You know, watch the code, you know, see if anything 162 00:11:23,602 --> 00:11:27,002 S1: hasn't been updated, like, maybe reach out and ping them 163 00:11:27,002 --> 00:11:29,522 S1: and say, hey, do you want to update something? Maybe 164 00:11:29,522 --> 00:11:33,842 S1: offer to help, uh, you know, maintain the project or something? 165 00:11:34,522 --> 00:11:38,762 S1: Or more importantly, if something nasty is in there, you 166 00:11:38,761 --> 00:11:41,802 S1: can actually find it, go and report it to him 167 00:11:41,802 --> 00:11:46,362 S1: or whatever. Or at the very least, maybe submit an 168 00:11:46,362 --> 00:11:49,642 S1: issue and say there's a vulnerability here, right? It's like 169 00:11:50,082 --> 00:11:54,162 S1: these are the eyes that we need. And I guess 170 00:11:54,162 --> 00:11:57,722 S1: the naive assumption was, well, we'll just train people up 171 00:11:57,722 --> 00:12:00,722 S1: in the school or in the trade schools, and eventually 172 00:12:00,722 --> 00:12:04,641 S1: we'll have millions of developers, millions and millions and they'll 173 00:12:04,642 --> 00:12:09,242 S1: all be security trained. And guess what? It didn't happen 174 00:12:10,202 --> 00:12:14,761 S1: because one, there's not that many of them. And two, 175 00:12:16,722 --> 00:12:19,602 S1: I mean, who's got time for that? Who's got time 176 00:12:19,602 --> 00:12:22,842 S1: for going and watching all the internets, you know, open 177 00:12:22,842 --> 00:12:26,602 S1: source projects and making sure there's nothing nasty in them 178 00:12:26,881 --> 00:12:29,722 S1: or nothing nasty is being done to them. The answer 179 00:12:29,722 --> 00:12:35,242 S1: is nobody. And it's likely to be nobody forever. Um, 180 00:12:35,242 --> 00:12:40,082 S1: so I think this AI agents thing is a really 181 00:12:40,082 --> 00:12:43,682 S1: good way to achieve this whole mini AIS thing. My 182 00:12:43,682 --> 00:12:46,322 S1: buddy Clint Gibler and team found tons of vulns with 183 00:12:46,322 --> 00:12:48,922 S1: a ten line cloud code prompt, so I think found 184 00:12:48,922 --> 00:12:51,602 S1: like 400 vulns. I think he said it was something 185 00:12:51,602 --> 00:12:57,362 S1: like 20 something Vulns were real, and it's actually fairly easy, 186 00:12:57,402 --> 00:13:02,852 S1: I think, um, to whittle that down and get rid 187 00:13:02,892 --> 00:13:05,372 S1: of a lot of those false positives. But the important 188 00:13:05,372 --> 00:13:08,852 S1: point to me is that ten lines of, you know, 189 00:13:08,892 --> 00:13:14,771 S1: a prompt in cloud code, and it actually found real vulns. 190 00:13:14,852 --> 00:13:17,972 S1: These are real projects that are sitting online. Who knows 191 00:13:17,972 --> 00:13:20,891 S1: if they've been assessed before, but either way, it found 192 00:13:20,892 --> 00:13:24,372 S1: real violence. Now it was noisy. But again, I think 193 00:13:24,372 --> 00:13:30,732 S1: that's a filter. Um, that is likely to, uh, be 194 00:13:30,732 --> 00:13:33,492 S1: pretty easy to build to just, um, you know, take 195 00:13:33,492 --> 00:13:37,372 S1: that in stages and clean those up. Um, yeah. And 196 00:13:37,372 --> 00:13:41,132 S1: there might be a blog post about that. Uh, maybe 197 00:13:41,131 --> 00:13:43,972 S1: a follow up would be nice, um, to maybe talk 198 00:13:43,972 --> 00:13:47,972 S1: about that whole cleanup process, but, uh, very cool work. 199 00:13:49,612 --> 00:13:58,572 S1: Hiring fraud flips zero. Trust from network to identity. So, yeah, 200 00:13:58,612 --> 00:14:04,692 S1: somebody's talking about here Hiring fake people, like they don't 201 00:14:04,692 --> 00:14:07,052 S1: have to break in and they don't have to phish 202 00:14:07,052 --> 00:14:11,372 S1: you if they can just get onboarded as regular employees 203 00:14:11,892 --> 00:14:15,732 S1: with full credentials. So this whole fake hiring thing is 204 00:14:15,772 --> 00:14:19,372 S1: becoming a serious issue. It's funny that it's happening at 205 00:14:19,372 --> 00:14:24,692 S1: the same time that people can't get hired. So it's like, okay, well, 206 00:14:24,692 --> 00:14:27,972 S1: if nobody's getting hired, and obviously it's not nobody. But 207 00:14:28,172 --> 00:14:30,172 S1: if it's hard to get hired right now, why are 208 00:14:30,172 --> 00:14:36,892 S1: we hiring all these North Koreans anyway? Yeah. Identity becoming 209 00:14:37,252 --> 00:14:42,452 S1: more important than zero trust. Or like the headline said, 210 00:14:43,012 --> 00:14:47,772 S1: zero trust moves to identity instead of, you know, perimeter 211 00:14:47,772 --> 00:14:53,972 S1: or whatever. Critical fixes for Chrome definitely go update their 212 00:14:54,532 --> 00:15:01,222 S1: scattered spider now targets browsers as enterprises move operations there. Yeah. So, uh, 213 00:15:01,262 --> 00:15:04,942 S1: my buddy Jason Haddox has been talking about AI moving 214 00:15:04,982 --> 00:15:09,342 S1: to the browser as the next phase. And he's right. 215 00:15:09,662 --> 00:15:12,182 S1: He's right. I didn't think it would go this heavy, 216 00:15:12,182 --> 00:15:15,942 S1: but it's going heavy like tons of browsers coming out 217 00:15:15,942 --> 00:15:20,702 S1: with AI built directly in. And, uh, yeah. And now 218 00:15:21,062 --> 00:15:24,582 S1: attackers are going after it as well because. And this 219 00:15:24,582 --> 00:15:27,742 S1: was Jason. This is what Jason was explaining to me 220 00:15:27,742 --> 00:15:29,582 S1: why he thought this was going to happen. It's just 221 00:15:29,582 --> 00:15:31,942 S1: like it is the focal point. Like we are usually 222 00:15:31,942 --> 00:15:35,142 S1: in the browser. We need to log into things with 223 00:15:35,142 --> 00:15:37,782 S1: the browser. So we've got like our passkeys set up. 224 00:15:37,782 --> 00:15:40,022 S1: We got like our password set up. And it's just 225 00:15:40,022 --> 00:15:42,742 S1: like it's a focal point of so much that we do. 226 00:15:43,822 --> 00:15:49,902 S1: So it's a good attack point. Attackers can serve poison 227 00:15:49,902 --> 00:15:53,662 S1: websites only to agents. Yeah, I love this. This is 228 00:15:53,702 --> 00:15:59,312 S1: a really cool one. So you can actually. Yeah. Have 229 00:15:59,312 --> 00:16:02,032 S1: a site that looks fine and acts fine if it's 230 00:16:02,032 --> 00:16:04,432 S1: a human browsing it and using it. But if it's 231 00:16:04,432 --> 00:16:08,592 S1: an AI, they basically get fed, uh, prompt AI stuff 232 00:16:08,712 --> 00:16:16,032 S1: that takes them astray. Kind of like steganography. Actually, think 233 00:16:16,032 --> 00:16:20,952 S1: about it. US puts $10 million bounties on three Russian 234 00:16:20,952 --> 00:16:25,152 S1: FSB hackers. Yeah. This administration. Trump is a little bit 235 00:16:25,152 --> 00:16:28,232 S1: going after Putin, a little bit in a way that 236 00:16:28,232 --> 00:16:33,472 S1: I quite like. Seems like the relationship is fraying and 237 00:16:33,472 --> 00:16:36,872 S1: I'm very happy with that. And it looks like he's 238 00:16:36,872 --> 00:16:41,632 S1: going after some of their, uh, hacker groups. Yeah. Good news. 239 00:16:41,752 --> 00:16:46,792 S1: CISOs face growing pressure to hide breaches Cisa pushes universal 240 00:16:46,792 --> 00:16:52,272 S1: software ingredients lists. Yeah, this is great. This is all 241 00:16:52,272 --> 00:16:54,672 S1: part of the whole supply chain stuff. My buddy, uh, 242 00:16:54,672 --> 00:17:00,952 S1: Sasha's dealer is, over at crosspoint and he talks tons 243 00:17:00,952 --> 00:17:06,752 S1: about software supply chain stuff. Uh, and definitely in relation to, uh, 244 00:17:06,752 --> 00:17:10,071 S1: reversing labs, which, as far as I could tell, is 245 00:17:10,112 --> 00:17:13,232 S1: the best company doing this stuff. I'm not affiliated with them, 246 00:17:13,232 --> 00:17:18,832 S1: but they are, uh, really cool at this stuff. And, um, yeah, 247 00:17:18,912 --> 00:17:23,352 S1: probably should get, uh, more affiliated with them. Um, but yeah. 248 00:17:26,472 --> 00:17:30,192 S1: CSA's talking about this. Uh, we got the NPM story. 249 00:17:30,672 --> 00:17:33,192 S1: Lots of people have been working on this. It goes 250 00:17:33,192 --> 00:17:35,312 S1: back to what I was just saying about Many Eyes. 251 00:17:36,232 --> 00:17:39,832 S1: It reminds me of also vendor lists, like I used to, uh, 252 00:17:40,152 --> 00:17:43,272 S1: work a lot with vendor lists when I was at Apple, and. 253 00:17:43,272 --> 00:17:46,992 S1: Holy crap, you you think you got a lot of vendors? 254 00:17:46,992 --> 00:17:49,831 S1: You should see Apple. They got a lot of vendors. Um, 255 00:17:50,032 --> 00:17:54,712 S1: but how many eyes do you need to watch all 256 00:17:54,752 --> 00:17:56,962 S1: the vendors that you have? Let's say you're a small startup. 257 00:17:56,962 --> 00:17:59,922 S1: Let's say you're a medium sized company. Whatever. You don't 258 00:17:59,922 --> 00:18:03,042 S1: have to be Apple or Google, right? How many eyes 259 00:18:03,042 --> 00:18:13,962 S1: do you need to watch all the different contracts to watch. Conduct, request. Read, parse, 260 00:18:14,362 --> 00:18:17,161 S1: make sense of and do something with all the different 261 00:18:17,162 --> 00:18:21,682 S1: assessments that have to be happening continuously. Then how about 262 00:18:21,682 --> 00:18:25,962 S1: determining where that stuff is installed on your network and 263 00:18:25,962 --> 00:18:30,002 S1: what impact that might have to the business? How many 264 00:18:30,002 --> 00:18:31,642 S1: people you need for that? Do you need to hire 265 00:18:31,642 --> 00:18:34,802 S1: three extra people? Because that's going to be a hard ask. 266 00:18:34,962 --> 00:18:37,242 S1: Like we're trying to reduce headcount right now. We can 267 00:18:37,242 --> 00:18:42,162 S1: maybe get you one. How much is that going to help? Right. 268 00:18:42,642 --> 00:18:46,362 S1: I mean, it's just ridiculous. It's a ridiculous question. So. 269 00:18:48,962 --> 00:18:54,362 S1: We need millions of extra eyes who can watch continuously 270 00:18:54,561 --> 00:18:59,602 S1: and constantly like, it's just like, you know, the solution 271 00:18:59,602 --> 00:19:03,522 S1: is not, you know, an extra person. So I love 272 00:19:03,522 --> 00:19:07,121 S1: this combination of many eyes with the supply chain problem. 273 00:19:10,042 --> 00:19:14,642 S1: You can find bugs by reading code. Yeah. Really cool. Um, 274 00:19:14,642 --> 00:19:20,722 S1: article here by Alex. Over 1100 exposed Olama servers found 275 00:19:20,722 --> 00:19:24,482 S1: via Shodan. And that's messed up because the question is, 276 00:19:24,482 --> 00:19:29,162 S1: what models are they running? What models are they pointing to? 277 00:19:29,642 --> 00:19:33,682 S1: Are we using up their GPUs. Like yeah what infrastructure 278 00:19:33,722 --> 00:19:37,282 S1: is that on. You can start asking questions like are 279 00:19:37,282 --> 00:19:40,442 S1: they vulnerable? Can you pivot. But you can also ask 280 00:19:40,442 --> 00:19:47,081 S1: resource exhaustion questions. National security US launches a Department of 281 00:19:47,082 --> 00:19:51,962 S1: War site. So our guy was supposed to be like 282 00:19:51,962 --> 00:19:56,292 S1: the isolationist guy. sky. Getting us out of this stuff. 283 00:19:56,692 --> 00:20:02,292 S1: Not going to go political here, but. Yeah. Department of war. Cool. 284 00:20:03,452 --> 00:20:07,212 S1: America steps back from the information fight. This one is 285 00:20:07,212 --> 00:20:12,452 S1: disturbing to me. Anne Applebaum says the current team is 286 00:20:12,492 --> 00:20:18,171 S1: gutting US overseas media and Anti-propaganda systems, which in my opinion, 287 00:20:18,172 --> 00:20:21,612 S1: is basically handing that over to China and Russia to 288 00:20:21,652 --> 00:20:25,851 S1: have like, full control over the narrative and story about 289 00:20:25,852 --> 00:20:29,852 S1: the West, West versus East, you know, China versus US, 290 00:20:30,412 --> 00:20:35,932 S1: Russia versus the West, whatever. I mean, this work is 291 00:20:35,932 --> 00:20:38,772 S1: nasty regardless. It's nasty. If the US is doing it 292 00:20:38,772 --> 00:20:41,851 S1: and of course we have done it and we continue 293 00:20:41,852 --> 00:20:45,252 S1: to do it, I'm sure. So it's kind of gross 294 00:20:45,252 --> 00:20:51,292 S1: no matter who's doing it, but. I don't know. That's 295 00:20:51,292 --> 00:20:55,012 S1: a longer conversation and definitely, uh, somewhat political. So I'm 296 00:20:55,012 --> 00:21:01,012 S1: going to skip it. But I would say if you 297 00:21:01,012 --> 00:21:04,412 S1: don't want China and Russia to win, maybe don't allow 298 00:21:04,412 --> 00:21:08,612 S1: them to control the narrative in every other country other 299 00:21:08,612 --> 00:21:15,212 S1: than the US. F-35's head to Puerto Rico as US 300 00:21:15,252 --> 00:21:23,891 S1: Venezuela standoff heats up. Hmm. Yeah. Sending ten F-35s to 301 00:21:23,932 --> 00:21:33,772 S1: Puerto Rico after Venezuelan F-16s buzzed a Navy destroyer. And. Yeah, 302 00:21:33,772 --> 00:21:36,612 S1: and then that's when that, uh, cartel boat got hit. 303 00:21:39,092 --> 00:21:45,212 S1: Strange times. Uh, State Department says non-immigrant visa applicants now 304 00:21:45,212 --> 00:21:51,262 S1: need to interview from their actual country or their actual residence, residents, 305 00:21:51,302 --> 00:21:57,062 S1: not inside the US. Sudan closes oil facilities after drone strikes. 306 00:21:57,502 --> 00:22:01,622 S1: DHS wants $100 million for counter drone systems to protect America. 307 00:22:01,902 --> 00:22:05,222 S1: I'm on board. I'm on board. I think we need 308 00:22:05,222 --> 00:22:12,222 S1: to stop spending money on F-35s or whatever modern fighters. 309 00:22:12,222 --> 00:22:15,702 S1: I think we need to move to the next phase. 310 00:22:16,821 --> 00:22:24,222 S1: And China is using its private sector to advance military capabilities. Uh, yes. Yes. 311 00:22:24,582 --> 00:22:27,782 S1: And it's explicit and it's like direct. And it's kind 312 00:22:27,782 --> 00:22:32,622 S1: of super gross and nasty, but also not really. It's nationalistic. 313 00:22:32,662 --> 00:22:36,782 S1: It's like, I mean, doesn't super bother me. What bothers 314 00:22:36,782 --> 00:22:40,422 S1: me is when they claim it's not happening and they're 315 00:22:40,422 --> 00:22:45,942 S1: trying to, you know, spread across the world and basically 316 00:22:46,502 --> 00:22:51,071 S1: get control of everything. And they're hiding this and people 317 00:22:51,071 --> 00:22:54,752 S1: are accepting them in like with Huawei, for example. Like 318 00:22:54,912 --> 00:22:58,632 S1: that's the part that bothers me, them being nationalistic and 319 00:22:58,672 --> 00:23:02,552 S1: having unity between their private sector and the goals of 320 00:23:02,552 --> 00:23:08,952 S1: the country. And it doesn't really bother me. AI OpenAI 321 00:23:08,992 --> 00:23:13,672 S1: launches an academy and jobs board to reskill displaced workers. 322 00:23:14,712 --> 00:23:17,752 S1: I mean, to give credit to Sam Altman here, he's 323 00:23:17,752 --> 00:23:21,391 S1: been saying for like the whole time, starting in like 22, 324 00:23:21,432 --> 00:23:24,832 S1: probably way before that. But I started paying attention mostly 325 00:23:24,832 --> 00:23:28,311 S1: in 22. And he's like, look, there's going to be 326 00:23:28,352 --> 00:23:33,112 S1: tremendous job loss from AGI or AI in general or whatever. 327 00:23:33,112 --> 00:23:38,552 S1: But definitely after AGI, depending on how you define it. 328 00:23:38,912 --> 00:23:40,472 S1: But he's like, yeah, we're going to have to figure 329 00:23:40,472 --> 00:23:43,552 S1: out things for people to do. He's been saying this 330 00:23:43,552 --> 00:23:47,952 S1: over and over in interviews for years and years and years. Um, 331 00:23:47,952 --> 00:23:49,952 S1: so if you didn't know that, then this is going 332 00:23:49,992 --> 00:23:52,072 S1: to look kind of crazy. Like all the reports come 333 00:23:52,071 --> 00:23:55,472 S1: out saying everyone's getting laid off. And then he's like, hey, 334 00:23:55,512 --> 00:23:59,752 S1: you know, we launched a job board. It's a bulletin board. 335 00:23:59,752 --> 00:24:01,272 S1: And you can come and you can pin a piece 336 00:24:01,272 --> 00:24:03,392 S1: of paper to it, and someone can come by and 337 00:24:03,432 --> 00:24:10,072 S1: offer you a job. Thanks, Sam. Super helpful. That's one 338 00:24:10,071 --> 00:24:11,672 S1: way to look at it. Another way to look at 339 00:24:11,672 --> 00:24:14,872 S1: it is obviously you would want to do this to 340 00:24:14,912 --> 00:24:19,071 S1: help people. I'm not sure if this is the same 341 00:24:19,232 --> 00:24:24,311 S1: article as the one competing with LinkedIn. Did I include 342 00:24:24,311 --> 00:24:26,592 S1: that one or not? That was something else I read. 343 00:24:28,512 --> 00:24:30,272 S1: I think it is. I think this is the one 344 00:24:30,272 --> 00:24:34,392 S1: that people are saying is competing with LinkedIn. It's probably 345 00:24:34,432 --> 00:24:38,232 S1: how could they possibly launch two? Anyway, there's another thing 346 00:24:38,232 --> 00:24:42,752 S1: where they they're trying to compete directly with LinkedIn and 347 00:24:42,752 --> 00:24:45,832 S1: basically help people find jobs by matching skills, which is 348 00:24:45,832 --> 00:24:49,962 S1: exactly what LinkedIn Microsoft is doing. A lot of people 349 00:24:49,962 --> 00:24:56,401 S1: forget that Microsoft owns LinkedIn. I forget sometimes, which is 350 00:24:56,402 --> 00:25:03,842 S1: why I just said a lot of people. But really, 351 00:25:03,842 --> 00:25:08,882 S1: it's just me. Anthropic agrees to pay authors at least 352 00:25:08,882 --> 00:25:13,762 S1: 1.5 billion in settlement. So they're doing this to try 353 00:25:13,762 --> 00:25:17,242 S1: to avoid a much bigger one. This is for authors 354 00:25:17,242 --> 00:25:27,762 S1: of books. So roughly $3,000 per book. Huh? Can I 355 00:25:27,762 --> 00:25:30,722 S1: get $3,000? How do I find out if my data 356 00:25:30,722 --> 00:25:37,002 S1: is in here? Anthropic valued at 183 billion after $13 billion. 357 00:25:37,002 --> 00:25:41,482 S1: Raise anthropic blocks. Chinese owned firms from Claude. This is 358 00:25:41,482 --> 00:25:44,482 S1: because of that threat research that came out that basically 359 00:25:44,482 --> 00:25:47,762 S1: showed a whole bunch of Chinese companies are actually using 360 00:25:47,762 --> 00:25:51,641 S1: it to do, you know, military oriented stuff and then 361 00:25:51,682 --> 00:25:57,562 S1: handing it directly to the Chinese government. So they disallow 362 00:25:58,722 --> 00:26:03,522 S1: listed a bunch of them. DeepMind's AI cuts gravitational wave 363 00:26:03,522 --> 00:26:11,282 S1: detector noise by 100,000. I'm telling you DeepMind Google stuff, 364 00:26:11,602 --> 00:26:16,802 S1: they are quiet and they just make these massive jumps. 365 00:26:17,002 --> 00:26:23,002 S1: And I I'm very long on Google, very long on Google, 366 00:26:23,202 --> 00:26:26,762 S1: especially as it relates to AI. I think they're just 367 00:26:27,362 --> 00:26:33,122 S1: they're business mind is there, especially with Sundar. Like he's 368 00:26:33,122 --> 00:26:36,561 S1: like the real deal. And they woke up on AI 369 00:26:36,602 --> 00:26:40,162 S1: like they're just they're just crushing it. AI agents could 370 00:26:40,162 --> 00:26:44,601 S1: reshape entire economies within the next decade. This is a 371 00:26:44,722 --> 00:26:49,892 S1: arXiv archived paper. Agents on MD emerges as the universal 372 00:26:49,892 --> 00:26:53,932 S1: standard for AI coding agent instructions. So I'm messing with 373 00:26:53,972 --> 00:26:59,732 S1: using agents versus coded MD with Chi. I don't know. 374 00:26:59,772 --> 00:27:02,492 S1: I've gone back and forth twice now. Um, right now 375 00:27:02,492 --> 00:27:05,371 S1: I'm currently back on Cloudmd because I felt like I 376 00:27:05,372 --> 00:27:11,091 S1: was getting less enforcement. Yeah, I'm going to watch it closely. 377 00:27:11,532 --> 00:27:14,812 S1: High school senior says AI is destroying real learning for 378 00:27:14,811 --> 00:27:22,292 S1: his generation. And, uh, yeah. Says classmates constantly use ChatGPT 379 00:27:22,332 --> 00:27:27,012 S1: for everything from literature annotations to math homework and basically 380 00:27:27,012 --> 00:27:32,492 S1: copy paste exercises from AI, which. Well, I mean, that's 381 00:27:32,492 --> 00:27:36,052 S1: to be expected. The question is, are they also doing 382 00:27:36,052 --> 00:27:40,292 S1: good things with it? And I think from this person's answer, 383 00:27:40,292 --> 00:27:43,131 S1: the answer is no. And this was an article in 384 00:27:43,132 --> 00:27:51,102 S1: The Atlantic. Okay. Um, evidence shows AI is already replacing 385 00:27:51,102 --> 00:27:55,382 S1: entry level workers. Derek Thompson looks at this. Yeah, there's 386 00:27:55,382 --> 00:27:57,621 S1: just more and more data on this coming out. This. 387 00:27:57,662 --> 00:28:00,901 S1: I don't think it's really debatable at this point, but 388 00:28:00,902 --> 00:28:05,462 S1: good to have additional data there in the newsletter. Technology. 389 00:28:05,582 --> 00:28:09,462 S1: Google gets to keep Chrome but must end exclusive deals 390 00:28:09,702 --> 00:28:14,061 S1: and share search data. So kind of a mostly a 391 00:28:14,102 --> 00:28:16,581 S1: win I would say for Google to not get, uh, 392 00:28:17,702 --> 00:28:22,182 S1: Chrome divested out. OpenAI moves to make its own AI 393 00:28:22,222 --> 00:28:26,581 S1: chips with Broadcom and UC Santa Cruz. Engineers show Wi-Fi 394 00:28:26,622 --> 00:28:30,381 S1: signals can measure heart rate. Yeah, I love this stuff. 395 00:28:30,422 --> 00:28:32,502 S1: You can actually see through walls. You can measure your 396 00:28:32,502 --> 00:28:34,822 S1: heart rate. You can. Yeah. You could do all sorts 397 00:28:34,821 --> 00:28:38,582 S1: of stuff with Wi-Fi signals. Can't wait to see more 398 00:28:38,582 --> 00:28:42,022 S1: of this. Tech comes to the consumer side. SpaceX can 399 00:28:42,022 --> 00:28:46,862 S1: now launch 120 times a year from Florida humans. Job 400 00:28:46,862 --> 00:28:51,502 S1: growth stalls as unemployment hits four year high. Looks like 401 00:28:51,502 --> 00:28:57,142 S1: we added just 22,000 jobs in August. Got a new 402 00:28:57,142 --> 00:29:02,062 S1: drug that outperforms aspirin for heart attack prevention. Private equity 403 00:29:02,062 --> 00:29:06,622 S1: rentals make suburbs more diverse, but apply pressure to buyers. 404 00:29:06,982 --> 00:29:10,622 S1: So sometimes it raises the prices, but sometimes it does 405 00:29:10,622 --> 00:29:14,542 S1: bring in other people who normally couldn't get into that neighborhood. 406 00:29:15,302 --> 00:29:21,302 S1: Interesting idea to have like. The houses being bought up 407 00:29:21,902 --> 00:29:26,342 S1: by giant investment companies. And it's like I guess it's 408 00:29:26,342 --> 00:29:29,302 S1: just another landlord. So I guess maybe it's not that 409 00:29:29,302 --> 00:29:33,622 S1: much different than what we already have, but landlords now 410 00:29:34,222 --> 00:29:38,502 S1: it's just a guy, you know, just a guy or 411 00:29:38,542 --> 00:29:42,952 S1: a woman or whatever. And it's just like. They have 412 00:29:42,952 --> 00:29:46,472 S1: like three properties and that's their life. Their landlord feels 413 00:29:46,472 --> 00:29:49,632 S1: kind of weird to just see houses go off the 414 00:29:49,632 --> 00:29:53,472 S1: market and it's, you know, Blackrock or whoever it is. 415 00:29:54,912 --> 00:30:01,112 S1: More people are opting out of news to protect their mood. 40% 416 00:30:01,152 --> 00:30:06,552 S1: now skip news, according to Reuters. Mathematicians just broke an 417 00:30:06,592 --> 00:30:12,272 S1: 87 year old knot theory, conjecture and attention loops. Quietly 418 00:30:12,272 --> 00:30:16,032 S1: rewrite your model of reality. Whatever you stare at long 419 00:30:16,032 --> 00:30:21,911 S1: enough starts echoing back and reshaping you. Yeah. Who was it? 420 00:30:21,912 --> 00:30:27,152 S1: Mark Twain. Somebody said, uh, whatever. You pay attention, you 421 00:30:27,152 --> 00:30:29,592 S1: become whatever you pay attention to. So be careful what 422 00:30:29,592 --> 00:30:32,752 S1: you pay attention to. Not sure if that was Mark 423 00:30:32,752 --> 00:30:41,032 S1: Twain or not. Phones on the toilet raise hemorrhoid risk. Discovery, 424 00:30:41,032 --> 00:30:45,192 S1: Rick Rubin's The Way of Code. This is a very 425 00:30:45,192 --> 00:30:49,872 S1: strange site. Gorgeous. Not sure exactly how it has to 426 00:30:49,872 --> 00:30:51,552 S1: do with code, but I was just checking it out. 427 00:30:51,592 --> 00:30:57,712 S1: Quite nice. Muscle memory turns repeated agent tasks into deterministic replays. 428 00:30:57,712 --> 00:31:01,352 S1: This thing is sick, so it monitors what your AI 429 00:31:01,352 --> 00:31:04,112 S1: is doing on your behalf when it's calling tools and everything. 430 00:31:04,832 --> 00:31:07,112 S1: And then it's just like learning how to do that 431 00:31:08,312 --> 00:31:11,952 S1: and turning that into steps that it can then replay. 432 00:31:12,832 --> 00:31:16,432 S1: So yeah, it's just kind of like learning from what 433 00:31:16,472 --> 00:31:21,232 S1: AI does and incorporating that into your workflows. Being good 434 00:31:21,232 --> 00:31:26,792 S1: isn't enough to win anymore. John Sword's last name is swords. 435 00:31:26,792 --> 00:31:28,912 S1: Is that the coolest last name? It's one of them. 436 00:31:29,712 --> 00:31:34,792 S1: Argues that craft alone doesn't carry you. You need distribution, 437 00:31:34,792 --> 00:31:39,882 S1: timing and taste. Or your good never gets found. This 438 00:31:39,882 --> 00:31:42,122 S1: is why I'm arguing that everyone needs to become a creator. 439 00:31:42,122 --> 00:31:44,522 S1: It's not that you need to become what a creator 440 00:31:44,522 --> 00:31:50,522 S1: is today, but creator just needs to be the new 441 00:31:50,522 --> 00:31:54,442 S1: way of living, which is thinking, building, sharing. I'm arguing 442 00:31:54,442 --> 00:31:59,482 S1: that these should be human things, not creator things. Blogs 443 00:31:59,482 --> 00:32:02,161 S1: used to be weird and personal. So this is a 444 00:32:02,242 --> 00:32:05,202 S1: blog by Jet Girl. Really really good. You should check 445 00:32:05,202 --> 00:32:09,402 S1: it out. Cutting end to end test time. 84% with 446 00:32:09,402 --> 00:32:14,122 S1: Cloud Code SDK. Cloud code. Cloud code SDK. By the way, 447 00:32:14,122 --> 00:32:16,882 S1: is just you run cloud with a switch of P 448 00:32:17,602 --> 00:32:22,842 S1: and that is a standalone like AI. It's it's completely insane. 449 00:32:22,842 --> 00:32:25,642 S1: I don't know why they call it the SDK. It's 450 00:32:25,642 --> 00:32:31,922 S1: just a switch P command. Anyway, probably something I'm missing there. 451 00:32:31,922 --> 00:32:36,282 S1: But anyway, that is it's extremely powerful functionality and I'm 452 00:32:36,322 --> 00:32:38,892 S1: going to be using it a lot more. Make your 453 00:32:38,892 --> 00:32:43,372 S1: own handwriting font easily. Claude FAQs tool use unless you 454 00:32:43,372 --> 00:32:47,372 S1: enforce strict tool calling. I have this in tons of 455 00:32:47,372 --> 00:32:51,372 S1: my scaffolding and I have this. Well, it's all in 456 00:32:52,092 --> 00:32:53,892 S1: the Pi repo now you can go check it out. 457 00:32:54,572 --> 00:32:58,092 S1: AI enables live phone calls across language barriers. Oh, I 458 00:32:58,092 --> 00:33:02,092 S1: just saw the, um, Apple event to today. Two new 459 00:33:02,092 --> 00:33:06,812 S1: AirPods live transcription. Talking to anybody. They could just talk 460 00:33:06,812 --> 00:33:10,612 S1: in their language and you get the translation in your ear. 461 00:33:11,252 --> 00:33:15,452 S1: Absolutely insane. I wonder if it works at the haircut 462 00:33:15,452 --> 00:33:22,092 S1: place with Vietnamese people. Like having conversations around you and 463 00:33:22,092 --> 00:33:26,612 S1: you don't know if it's like. Damn, his feet are big. 464 00:33:26,612 --> 00:33:29,572 S1: Or if it's like they're trying to figure out lunch. 465 00:33:29,612 --> 00:33:38,492 S1: Like you're not quite sure. Tech interviews reward complexity over 466 00:33:38,492 --> 00:33:44,492 S1: practical engineering. Diallo's blog post. Yeah, this is a really 467 00:33:44,492 --> 00:33:50,532 S1: good blog on kind of the problems with overall interviewing, which, yeah, 468 00:33:50,532 --> 00:33:57,092 S1: there's just yeah, this is why a lot of AI interviewing, 469 00:33:57,292 --> 00:34:00,732 S1: recruiting services I think are starting to do fairly well 470 00:34:01,452 --> 00:34:04,052 S1: is not so much that they're awesome, it's just that 471 00:34:04,052 --> 00:34:10,332 S1: the status quo is notably bad. Chibi. I think that's 472 00:34:10,332 --> 00:34:14,452 S1: how you pronounce it, analyzes why users churn. This is 473 00:34:14,492 --> 00:34:16,852 S1: actually just a straight link to a vendor, and I 474 00:34:16,852 --> 00:34:20,692 S1: just thought the vendor was pretty cool. And um, no 475 00:34:20,692 --> 00:34:25,532 S1: sponsorship obviously. Otherwise I'd have like a block for them and, 476 00:34:25,572 --> 00:34:29,972 S1: you know, the asterisk or sponsorship or something to indicate 477 00:34:29,972 --> 00:34:32,972 S1: that it's a sponsor. But no, this is just some 478 00:34:32,972 --> 00:34:35,252 S1: pretty cool tech. I feel like churn is like one 479 00:34:35,422 --> 00:34:40,142 S1: of the biggest things. It's like not about the gaining 480 00:34:40,182 --> 00:34:44,942 S1: the users. It's about keeping the users. Random walks almost 481 00:34:44,942 --> 00:34:47,862 S1: never return home in three dimensions. This is a cool 482 00:34:47,862 --> 00:34:54,502 S1: physics thing. Recommendation. Whoa. Recommendation of the week. I'm going 483 00:34:54,502 --> 00:34:57,902 S1: to reuse one that I typically have to ask myself 484 00:34:57,942 --> 00:35:02,022 S1: and kind of always try to ask myself. And that 485 00:35:02,022 --> 00:35:06,942 S1: is especially important right now with like extraordinary change, I think. 486 00:35:06,942 --> 00:35:09,542 S1: What would you do if you're not afraid? Now, normally 487 00:35:09,542 --> 00:35:11,862 S1: I would follow up and say, you shouldn't be afraid, right? 488 00:35:11,902 --> 00:35:15,422 S1: Everything's fine. And, you know, you should just be okay 489 00:35:15,422 --> 00:35:18,142 S1: and think what you would do if you weren't afraid. 490 00:35:19,302 --> 00:35:24,902 S1: But everything isn't okay. But that doesn't mean it's still 491 00:35:24,902 --> 00:35:27,022 S1: not a really good time to ask this question. So 492 00:35:27,062 --> 00:35:30,342 S1: I think the question still stands. What would you do 493 00:35:30,342 --> 00:35:33,622 S1: if you were not afraid? Here's what I would challenge 494 00:35:33,622 --> 00:35:39,102 S1: here Just to add to this. Think about what your 495 00:35:39,102 --> 00:35:44,022 S1: ideal self looks like. Think about what you and your 496 00:35:44,022 --> 00:35:48,942 S1: perfect Super Saiyan form. What abilities and skills would you have? 497 00:35:48,942 --> 00:35:52,062 S1: What confidence would you have? What narratives would you be 498 00:35:52,062 --> 00:35:56,502 S1: able to tell people? And here's the question is that 499 00:35:56,502 --> 00:36:00,022 S1: person more prepared for what is coming in the future 500 00:36:00,022 --> 00:36:03,182 S1: than you are? And if that's the case, then you 501 00:36:03,182 --> 00:36:07,462 S1: need to think up some dramatic ways, some extreme ways 502 00:36:07,462 --> 00:36:12,102 S1: to possibly migrate to that version of yourself, because that 503 00:36:12,102 --> 00:36:15,022 S1: version is going to survive a lot better maybe, than 504 00:36:15,022 --> 00:36:18,622 S1: your current version. And the aphorism of the week, very 505 00:36:18,622 --> 00:36:21,782 S1: much in line with this, the privilege of a lifetime 506 00:36:21,782 --> 00:36:25,822 S1: is to become who you truly are. The privilege of 507 00:36:25,822 --> 00:36:30,382 S1: a lifetime is to become who you truly are. Carl Jung.