WEBVTT - UL NO. 497: STANDARD EDITION | More NPM Shenanigans, I Open Sourced Kai, Blood Work Results, Finding Vulns in a 10-line Prompt, and more... 0:00:19.212 --> 0:00:27.852 Hey. What's up? All right, episode 497. Yeah. Hope things 0:00:27.852 --> 0:00:31.052 are starting off well this week. Um, got some lab 0:00:31.052 --> 0:00:35.772 results back. This is definitely too much information, but, uh, 0:00:37.531 --> 0:00:42.052 c metabolic is pristine. Liver is great. Kidneys not so good. 0:00:43.692 --> 0:00:46.812 Got to improve kidney function. Uh, turns out too much 0:00:46.812 --> 0:00:51.132 creatine is a bad thing. I was actually getting dizziness 0:00:51.732 --> 0:00:55.972 from ten grams of creatine, so I think I'm going 0:00:56.012 --> 0:01:01.012 to have to go back down to five, maybe even lower. Uh, 0:01:01.012 --> 0:01:05.972 because it turns out creatine is not great for the kidneys. And, um, 0:01:06.732 --> 0:01:09.732 it's not like real bad, but it's just like, not optimal. 0:01:09.732 --> 0:01:15.172 So I need to optimize that, uh, HDL a little 0:01:15.172 --> 0:01:18.182 bit too low, LDL a little bit too high. So 0:01:19.262 --> 0:01:22.742 I've basically been super sedentary for like the last 4 0:01:22.742 --> 0:01:26.982 or 5 months because of working on Kai and all the, uh, 0:01:28.062 --> 0:01:30.941 all the crazy stuff going on with AI and projects and, 0:01:31.102 --> 0:01:36.302 you know, all the typical stuff, all the crazy world stuff. So, um, 0:01:36.301 --> 0:01:39.741 starting to try to adjust to this, uh, stand up 0:01:39.742 --> 0:01:45.982 alarms for standing up with my standing desk, um, foam 0:01:45.982 --> 0:01:50.262 seating pad with, like, a cutout. So I'm not, like, sitting, um. 0:01:52.302 --> 0:01:57.582 In an uncomfortable position all day. Mandatory cardio twice a week. Uh, 0:01:57.582 --> 0:02:04.662 table tennis or rucking weights at least twice a week. Uh, 0:02:04.662 --> 0:02:06.902 more coffee to balance the Celsius. So that was the 0:02:06.902 --> 0:02:11.382 other problem. Wasn't just creatine. I'm drinking like two Celsius 0:02:11.422 --> 0:02:17.142 a day. Sometimes more, but usually only 1 to 2, 0:02:18.062 --> 0:02:22.942 and energy drinks and creatine are two of the things 0:02:22.942 --> 0:02:28.262 on the list for, um, things to avoid. If you 0:02:28.262 --> 0:02:35.262 have kidney function that's, uh, not optimal. And exercise. Yeah. 0:02:35.302 --> 0:02:38.182 Raises HDL. So let me know if you're on a 0:02:38.181 --> 0:02:41.542 similar journey. That's kind of where I'm at. Oh sugar 0:02:41.542 --> 0:02:45.662 was great. So my metabolic health was ideal. It was 0:02:45.662 --> 0:02:48.662 like perfect. So that was that was a great story. 0:02:51.062 --> 0:02:56.662 Okay Kai. Update massive context loading. Breakthrough. So I am 0:02:56.662 --> 0:03:01.901 now doing a really cool thing where I'm basically I 0:03:01.942 --> 0:03:06.742 changed the hook to point to load dynamic context, and 0:03:06.782 --> 0:03:10.101 I move that to point two load dynamic context MD 0:03:10.181 --> 0:03:14.142 so now my routing is actually in the MD file, 0:03:14.142 --> 0:03:19.232 which means I can actually navigate that whole thing contextually. 0:03:19.392 --> 0:03:23.272 Instead of having to edit TypeScript code, I am now 0:03:23.272 --> 0:03:28.672 able to write it out in text in markdown, basically describing. Yeah. 0:03:28.712 --> 0:03:32.912 If you see this type of question, um, that means 0:03:32.912 --> 0:03:34.752 I want you to use the researcher. If you see 0:03:34.752 --> 0:03:37.232 this type of thing, that means dynamically load this piece 0:03:37.232 --> 0:03:40.152 of context. If you see me talk about this project, 0:03:40.152 --> 0:03:44.032 that means go load this piece of context, which means 0:03:44.032 --> 0:03:49.392 my primary boot load just gets my core context right. 0:03:49.432 --> 0:03:51.752 So this whole UFC system that I talked about in 0:03:51.752 --> 0:03:57.112 the video is, um, it's all based on only loading 0:03:57.112 --> 0:03:59.672 what you need to load at that particular moment and 0:03:59.672 --> 0:04:05.672 keeping the context extremely clean. So, uh, this is just 0:04:05.952 --> 0:04:08.512 it's really powerful and it's working perfectly. Like I'm not 0:04:08.512 --> 0:04:12.472 having any misses. It's actually dynamically loading exactly what I need. 0:04:12.752 --> 0:04:15.352 And it's not just the context files, but it actually 0:04:15.352 --> 0:04:20.032 loads the correct researcher as well. Or I'm sorry, the 0:04:20.032 --> 0:04:26.192 correct agent, which could be researcher, engineer, designer or pentester. 0:04:26.392 --> 0:04:30.072 It can load all those different agents according to the 0:04:30.072 --> 0:04:33.392 type of asset that, um, that I give it. Right. 0:04:33.432 --> 0:04:38.592 So pretty. Pretty cool. And, um, speaking of Chi Pi, 0:04:38.632 --> 0:04:44.551 speaking of the whole Pi system, I am open sourcing it. 0:04:44.592 --> 0:04:46.992 In fact, I already have. In fact, it's already live. 0:04:47.552 --> 0:04:49.592 And I was just going to put the framework up 0:04:49.592 --> 0:04:51.712 there and say, hey, look, I'm going to add stuff later. 0:04:51.712 --> 0:04:56.592 But then I just started adding stuff. Um, and now 0:04:56.592 --> 0:04:58.792 I've added a whole bunch of stuff. It's kind of 0:04:58.791 --> 0:05:01.272 like the basic structure. It doesn't have the voice stuff 0:05:01.272 --> 0:05:05.632 in there yet, but I do have my basic outline 0:05:05.632 --> 0:05:08.912 for the main context file and the main tools file, 0:05:10.152 --> 0:05:14.002 and then the pointers to where like, it's pretty obvious 0:05:14.002 --> 0:05:16.642 where to go. Put the other context for the other pieces. 0:05:16.642 --> 0:05:20.482 So that is a lot of infrastructure there. I said 0:05:20.481 --> 0:05:23.962 in one of the git commits, I'm probably definitely going 0:05:24.002 --> 0:05:27.681 to leak something because I'm putting so much of like 0:05:27.682 --> 0:05:31.402 my actual stuff out there, and it's like, no secrets. 0:05:31.442 --> 0:05:35.442 Like I'm literally sharing the actual files. A couple of 0:05:35.442 --> 0:05:39.282 things are like redacted for like keys and stuff like that. 0:05:39.282 --> 0:05:42.121 But other than that, like, this is seriously the exact 0:05:42.122 --> 0:05:45.202 infrastructure that Cai is using, and you're going to have 0:05:45.202 --> 0:05:47.601 to tweak it for that reason, right? Because I still 0:05:47.601 --> 0:05:50.522 have Cai's name in there. I still have like exactly 0:05:50.522 --> 0:05:53.882 what I'm saying to him, exactly what what his personality 0:05:53.882 --> 0:05:56.601 is and stuff like that. So you'll have to customize 0:05:56.601 --> 0:06:00.642 some stuff. But the whole purpose of this is going forward. 0:06:00.642 --> 0:06:04.122 When you're watching videos on YouTube that I'm talking about, 0:06:05.162 --> 0:06:07.202 I'm not going to say, oh, here's like a gist 0:06:07.202 --> 0:06:11.412 file or like, here's some code in the YouTube description. 0:06:11.692 --> 0:06:14.171 It's going to be like, hey, go check the Pi 0:06:14.212 --> 0:06:17.812 repository and you will see it. And the Pi repository 0:06:17.812 --> 0:06:21.972 is set up with a very clear mission. It's like, look, 0:06:22.012 --> 0:06:26.012 enable humans to have this functionality. It's the same as fabric, 0:06:26.412 --> 0:06:30.092 the same mission as fabric. And now fabric basically falls 0:06:30.092 --> 0:06:33.772 under this, right? It falls under pi because fabric is 0:06:33.772 --> 0:06:38.452 just one of the sets of tools that I have available. Um, 0:06:38.452 --> 0:06:41.492 so Kai has full access to fabric, has full access 0:06:41.492 --> 0:06:45.291 to my foundry stuff, which is the fobs and the 0:06:45.291 --> 0:06:49.652 custom commands, the custom cloud commands, like all the different stuff, 0:06:49.652 --> 0:06:53.731 it's all in the commands directory and in mic, and 0:06:53.731 --> 0:06:56.572 Kai can use all of it. And that is the 0:06:56.572 --> 0:06:59.812 whole purpose of this entire project. And it is now 0:06:59.812 --> 0:07:03.931 open source. So it's capital Pi is the name of 0:07:03.932 --> 0:07:13.132 the project on GitHub. Okay. Oh, yeah. Just thought of 0:07:13.132 --> 0:07:15.452 a cool. Um. I was sitting here trying to write 0:07:15.452 --> 0:07:19.812 the newsletter, uh, and just kept adding more stuff. So 0:07:19.812 --> 0:07:22.412 I've got this one. I just added it. So sick. 0:07:22.452 --> 0:07:26.972 It's called capture learning. It's a command. And so the 0:07:26.972 --> 0:07:31.532 idea is, if I have a I'm working on something 0:07:31.532 --> 0:07:35.532 with Kai. We make something new or we're troubleshooting something 0:07:35.532 --> 0:07:37.332 and we finally get it to work. I'm just going 0:07:37.372 --> 0:07:41.452 to say, hey, write this down or document this, or 0:07:41.892 --> 0:07:45.292 capture learning or whatever. And that same thing that I 0:07:45.292 --> 0:07:50.052 told you about before, um, that dynamically loads context. Well, 0:07:50.052 --> 0:07:55.012 it could also listen because it's being triggered on the prompt, 0:07:55.412 --> 0:07:59.532 the user prompt submit that hook is what's actually triggering this. 0:07:59.812 --> 0:08:04.452 So if Kai hears anything like that, Kai now knows 0:08:04.452 --> 0:08:08.092 to go and execute capture learning. What does that do? 0:08:08.382 --> 0:08:13.342 that drops it into the Learning Context directory. The learning 0:08:13.342 --> 0:08:19.862 context directory includes a full report of the lesson, the 0:08:20.022 --> 0:08:25.102 false assumption that we had, the mistake that we were making, 0:08:25.102 --> 0:08:28.382 or whatever talks through the problem, talks through the process 0:08:28.382 --> 0:08:31.262 because it has the whole log of our conversation, the 0:08:31.262 --> 0:08:34.222 process of how we got there and then what the 0:08:34.222 --> 0:08:39.142 solution was and then what the takeaway is. So that's 0:08:39.142 --> 0:08:41.622 documented for each one. I'm going to have that going 0:08:41.622 --> 0:08:44.422 forward now. And guess what you could do. You could 0:08:44.422 --> 0:08:51.902 now say something like Harvest Learnings and then have that 0:08:51.902 --> 0:08:55.142 go and compare your current documentation, your current workflows, your 0:08:55.142 --> 0:09:00.102 current pipelines, your current context and instructions and everything, and 0:09:00.102 --> 0:09:02.862 compare it with the lessons learned and say, is there 0:09:02.862 --> 0:09:05.742 anything here that we should update based on the lessons 0:09:05.742 --> 0:09:08.502 learned over the last week or the last year or whatever. 0:09:09.142 --> 0:09:11.302 So I just think this is a really cool sort 0:09:11.342 --> 0:09:18.462 of meta improvement concept. All right. Cybersecurity. Npm's most used 0:09:18.462 --> 0:09:23.862 UI packages got backdoored to hijack crypto. Yeah. This, uh, 0:09:24.422 --> 0:09:28.222 aikido company evidently found this, and, um, put out a 0:09:28.222 --> 0:09:35.262 report about it. It's 18 different packages and, uh. Yeah, 0:09:35.302 --> 0:09:40.662 rewritten to steal crypto. Uh, lots of supply chain stuff 0:09:40.662 --> 0:09:44.702 going on. Uh, I really like the idea of using. 0:09:46.982 --> 0:09:49.822 AI for many eyes. I feel like many eyes is 0:09:49.822 --> 0:09:53.901 like my my big, uh, thing I keep coming back 0:09:53.902 --> 0:09:57.381 to with with AI. And I think I talked about 0:09:57.381 --> 0:10:01.262 this maybe last week or the week before, but SSL, 0:10:01.422 --> 0:10:05.342 you remember the giant SSL vulnerability? I think we still 0:10:05.342 --> 0:10:08.832 called it SSL back then. Maybe it was already TLS, 0:10:08.832 --> 0:10:12.552 but anyway, that was a huge vuln and everyone was like, hey, 0:10:12.552 --> 0:10:14.992 how could this possibly happen? This is open source. And 0:10:14.992 --> 0:10:18.631 it's like, well, turns out open source is just a guy. 0:10:18.752 --> 0:10:21.552 I heard someone say that recently. Open source is just 0:10:21.552 --> 0:10:25.912 a guy. You know, it's just Chris Richards or whatever 0:10:25.912 --> 0:10:29.632 who lives in Idaho. Turns out that whole open source, 0:10:29.631 --> 0:10:33.312 many eyes things. Turns out it's just Chris and Chris 0:10:33.312 --> 0:10:35.472 hurt his ankle and he hasn't been online in a while. 0:10:35.472 --> 0:10:39.592 You know what I mean? So it's like Many Eyes 0:10:39.631 --> 0:10:41.631 didn't turn out to be what we thought it was 0:10:41.631 --> 0:10:45.592 going to be. It's because what it means, what many eyes, 0:10:45.832 --> 0:10:48.511 what we thought many eyes was going to be was 0:10:48.752 --> 0:10:52.192 many eyes are looking at the project, but what it 0:10:52.192 --> 0:10:55.032 turned out to be was many eyes could look at 0:10:55.032 --> 0:10:59.791 the project, but they weren't actually. And so this is 0:10:59.792 --> 0:11:05.762 why I love the idea of AI. Many eyes. Millions 0:11:05.761 --> 0:11:09.922 of eyes. Right. Cheap eyes that don't get tired, don't 0:11:09.922 --> 0:11:16.162 get sick or whatever. And you just stick them on 0:11:16.162 --> 0:11:19.602 the internet. You're just like, hey, go watch Chris's project. 0:11:20.042 --> 0:11:23.602 You know, watch the code, you know, see if anything 0:11:23.602 --> 0:11:27.002 hasn't been updated, like, maybe reach out and ping them 0:11:27.002 --> 0:11:29.522 and say, hey, do you want to update something? Maybe 0:11:29.522 --> 0:11:33.842 offer to help, uh, you know, maintain the project or something? 0:11:34.522 --> 0:11:38.762 Or more importantly, if something nasty is in there, you 0:11:38.761 --> 0:11:41.802 can actually find it, go and report it to him 0:11:41.802 --> 0:11:46.362 or whatever. Or at the very least, maybe submit an 0:11:46.362 --> 0:11:49.642 issue and say there's a vulnerability here, right? It's like 0:11:50.082 --> 0:11:54.162 these are the eyes that we need. And I guess 0:11:54.162 --> 0:11:57.722 the naive assumption was, well, we'll just train people up 0:11:57.722 --> 0:12:00.722 in the school or in the trade schools, and eventually 0:12:00.722 --> 0:12:04.641 we'll have millions of developers, millions and millions and they'll 0:12:04.642 --> 0:12:09.242 all be security trained. And guess what? It didn't happen 0:12:10.202 --> 0:12:14.761 because one, there's not that many of them. And two, 0:12:16.722 --> 0:12:19.602 I mean, who's got time for that? Who's got time 0:12:19.602 --> 0:12:22.842 for going and watching all the internets, you know, open 0:12:22.842 --> 0:12:26.602 source projects and making sure there's nothing nasty in them 0:12:26.881 --> 0:12:29.722 or nothing nasty is being done to them. The answer 0:12:29.722 --> 0:12:35.242 is nobody. And it's likely to be nobody forever. Um, 0:12:35.242 --> 0:12:40.082 so I think this AI agents thing is a really 0:12:40.082 --> 0:12:43.682 good way to achieve this whole mini AIS thing. My 0:12:43.682 --> 0:12:46.322 buddy Clint Gibler and team found tons of vulns with 0:12:46.322 --> 0:12:48.922 a ten line cloud code prompt, so I think found 0:12:48.922 --> 0:12:51.602 like 400 vulns. I think he said it was something 0:12:51.602 --> 0:12:57.362 like 20 something Vulns were real, and it's actually fairly easy, 0:12:57.402 --> 0:13:02.852 I think, um, to whittle that down and get rid 0:13:02.892 --> 0:13:05.372 of a lot of those false positives. But the important 0:13:05.372 --> 0:13:08.852 point to me is that ten lines of, you know, 0:13:08.892 --> 0:13:14.771 a prompt in cloud code, and it actually found real vulns. 0:13:14.852 --> 0:13:17.972 These are real projects that are sitting online. Who knows 0:13:17.972 --> 0:13:20.891 if they've been assessed before, but either way, it found 0:13:20.892 --> 0:13:24.372 real violence. Now it was noisy. But again, I think 0:13:24.372 --> 0:13:30.732 that's a filter. Um, that is likely to, uh, be 0:13:30.732 --> 0:13:33.492 pretty easy to build to just, um, you know, take 0:13:33.492 --> 0:13:37.372 that in stages and clean those up. Um, yeah. And 0:13:37.372 --> 0:13:41.132 there might be a blog post about that. Uh, maybe 0:13:41.131 --> 0:13:43.972 a follow up would be nice, um, to maybe talk 0:13:43.972 --> 0:13:47.972 about that whole cleanup process, but, uh, very cool work. 0:13:49.612 --> 0:13:58.572 Hiring fraud flips zero. Trust from network to identity. So, yeah, 0:13:58.612 --> 0:14:04.692 somebody's talking about here Hiring fake people, like they don't 0:14:04.692 --> 0:14:07.052 have to break in and they don't have to phish 0:14:07.052 --> 0:14:11.372 you if they can just get onboarded as regular employees 0:14:11.892 --> 0:14:15.732 with full credentials. So this whole fake hiring thing is 0:14:15.772 --> 0:14:19.372 becoming a serious issue. It's funny that it's happening at 0:14:19.372 --> 0:14:24.692 the same time that people can't get hired. So it's like, okay, well, 0:14:24.692 --> 0:14:27.972 if nobody's getting hired, and obviously it's not nobody. But 0:14:28.172 --> 0:14:30.172 if it's hard to get hired right now, why are 0:14:30.172 --> 0:14:36.892 we hiring all these North Koreans anyway? Yeah. Identity becoming 0:14:37.252 --> 0:14:42.452 more important than zero trust. Or like the headline said, 0:14:43.012 --> 0:14:47.772 zero trust moves to identity instead of, you know, perimeter 0:14:47.772 --> 0:14:53.972 or whatever. Critical fixes for Chrome definitely go update their 0:14:54.532 --> 0:15:01.222 scattered spider now targets browsers as enterprises move operations there. Yeah. So, uh, 0:15:01.262 --> 0:15:04.942 my buddy Jason Haddox has been talking about AI moving 0:15:04.982 --> 0:15:09.342 to the browser as the next phase. And he's right. 0:15:09.662 --> 0:15:12.182 He's right. I didn't think it would go this heavy, 0:15:12.182 --> 0:15:15.942 but it's going heavy like tons of browsers coming out 0:15:15.942 --> 0:15:20.702 with AI built directly in. And, uh, yeah. And now 0:15:21.062 --> 0:15:24.582 attackers are going after it as well because. And this 0:15:24.582 --> 0:15:27.742 was Jason. This is what Jason was explaining to me 0:15:27.742 --> 0:15:29.582 why he thought this was going to happen. It's just 0:15:29.582 --> 0:15:31.942 like it is the focal point. Like we are usually 0:15:31.942 --> 0:15:35.142 in the browser. We need to log into things with 0:15:35.142 --> 0:15:37.782 the browser. So we've got like our passkeys set up. 0:15:37.782 --> 0:15:40.022 We got like our password set up. And it's just 0:15:40.022 --> 0:15:42.742 like it's a focal point of so much that we do. 0:15:43.822 --> 0:15:49.902 So it's a good attack point. Attackers can serve poison 0:15:49.902 --> 0:15:53.662 websites only to agents. Yeah, I love this. This is 0:15:53.702 --> 0:15:59.312 a really cool one. So you can actually. Yeah. Have 0:15:59.312 --> 0:16:02.032 a site that looks fine and acts fine if it's 0:16:02.032 --> 0:16:04.432 a human browsing it and using it. But if it's 0:16:04.432 --> 0:16:08.592 an AI, they basically get fed, uh, prompt AI stuff 0:16:08.712 --> 0:16:16.032 that takes them astray. Kind of like steganography. Actually, think 0:16:16.032 --> 0:16:20.952 about it. US puts $10 million bounties on three Russian 0:16:20.952 --> 0:16:25.152 FSB hackers. Yeah. This administration. Trump is a little bit 0:16:25.152 --> 0:16:28.232 going after Putin, a little bit in a way that 0:16:28.232 --> 0:16:33.472 I quite like. Seems like the relationship is fraying and 0:16:33.472 --> 0:16:36.872 I'm very happy with that. And it looks like he's 0:16:36.872 --> 0:16:41.632 going after some of their, uh, hacker groups. Yeah. Good news. 0:16:41.752 --> 0:16:46.792 CISOs face growing pressure to hide breaches Cisa pushes universal 0:16:46.792 --> 0:16:52.272 software ingredients lists. Yeah, this is great. This is all 0:16:52.272 --> 0:16:54.672 part of the whole supply chain stuff. My buddy, uh, 0:16:54.672 --> 0:17:00.952 Sasha's dealer is, over at crosspoint and he talks tons 0:17:00.952 --> 0:17:06.752 about software supply chain stuff. Uh, and definitely in relation to, uh, 0:17:06.752 --> 0:17:10.071 reversing labs, which, as far as I could tell, is 0:17:10.112 --> 0:17:13.232 the best company doing this stuff. I'm not affiliated with them, 0:17:13.232 --> 0:17:18.832 but they are, uh, really cool at this stuff. And, um, yeah, 0:17:18.912 --> 0:17:23.352 probably should get, uh, more affiliated with them. Um, but yeah. 0:17:26.472 --> 0:17:30.192 CSA's talking about this. Uh, we got the NPM story. 0:17:30.672 --> 0:17:33.192 Lots of people have been working on this. It goes 0:17:33.192 --> 0:17:35.312 back to what I was just saying about Many Eyes. 0:17:36.232 --> 0:17:39.832 It reminds me of also vendor lists, like I used to, uh, 0:17:40.152 --> 0:17:43.272 work a lot with vendor lists when I was at Apple, and. 0:17:43.272 --> 0:17:46.992 Holy crap, you you think you got a lot of vendors? 0:17:46.992 --> 0:17:49.831 You should see Apple. They got a lot of vendors. Um, 0:17:50.032 --> 0:17:54.712 but how many eyes do you need to watch all 0:17:54.752 --> 0:17:56.962 the vendors that you have? Let's say you're a small startup. 0:17:56.962 --> 0:17:59.922 Let's say you're a medium sized company. Whatever. You don't 0:17:59.922 --> 0:18:03.042 have to be Apple or Google, right? How many eyes 0:18:03.042 --> 0:18:13.962 do you need to watch all the different contracts to watch. Conduct, request. Read, parse, 0:18:14.362 --> 0:18:17.161 make sense of and do something with all the different 0:18:17.162 --> 0:18:21.682 assessments that have to be happening continuously. Then how about 0:18:21.682 --> 0:18:25.962 determining where that stuff is installed on your network and 0:18:25.962 --> 0:18:30.002 what impact that might have to the business? How many 0:18:30.002 --> 0:18:31.642 people you need for that? Do you need to hire 0:18:31.642 --> 0:18:34.802 three extra people? Because that's going to be a hard ask. 0:18:34.962 --> 0:18:37.242 Like we're trying to reduce headcount right now. We can 0:18:37.242 --> 0:18:42.162 maybe get you one. How much is that going to help? Right. 0:18:42.642 --> 0:18:46.362 I mean, it's just ridiculous. It's a ridiculous question. So. 0:18:48.962 --> 0:18:54.362 We need millions of extra eyes who can watch continuously 0:18:54.561 --> 0:18:59.602 and constantly like, it's just like, you know, the solution 0:18:59.602 --> 0:19:03.522 is not, you know, an extra person. So I love 0:19:03.522 --> 0:19:07.121 this combination of many eyes with the supply chain problem. 0:19:10.042 --> 0:19:14.642 You can find bugs by reading code. Yeah. Really cool. Um, 0:19:14.642 --> 0:19:20.722 article here by Alex. Over 1100 exposed Olama servers found 0:19:20.722 --> 0:19:24.482 via Shodan. And that's messed up because the question is, 0:19:24.482 --> 0:19:29.162 what models are they running? What models are they pointing to? 0:19:29.642 --> 0:19:33.682 Are we using up their GPUs. Like yeah what infrastructure 0:19:33.722 --> 0:19:37.282 is that on. You can start asking questions like are 0:19:37.282 --> 0:19:40.442 they vulnerable? Can you pivot. But you can also ask 0:19:40.442 --> 0:19:47.081 resource exhaustion questions. National security US launches a Department of 0:19:47.082 --> 0:19:51.962 War site. So our guy was supposed to be like 0:19:51.962 --> 0:19:56.292 the isolationist guy. sky. Getting us out of this stuff. 0:19:56.692 --> 0:20:02.292 Not going to go political here, but. Yeah. Department of war. Cool. 0:20:03.452 --> 0:20:07.212 America steps back from the information fight. This one is 0:20:07.212 --> 0:20:12.452 disturbing to me. Anne Applebaum says the current team is 0:20:12.492 --> 0:20:18.171 gutting US overseas media and Anti-propaganda systems, which in my opinion, 0:20:18.172 --> 0:20:21.612 is basically handing that over to China and Russia to 0:20:21.652 --> 0:20:25.851 have like, full control over the narrative and story about 0:20:25.852 --> 0:20:29.852 the West, West versus East, you know, China versus US, 0:20:30.412 --> 0:20:35.932 Russia versus the West, whatever. I mean, this work is 0:20:35.932 --> 0:20:38.772 nasty regardless. It's nasty. If the US is doing it 0:20:38.772 --> 0:20:41.851 and of course we have done it and we continue 0:20:41.852 --> 0:20:45.252 to do it, I'm sure. So it's kind of gross 0:20:45.252 --> 0:20:51.292 no matter who's doing it, but. I don't know. That's 0:20:51.292 --> 0:20:55.012 a longer conversation and definitely, uh, somewhat political. So I'm 0:20:55.012 --> 0:21:01.012 going to skip it. But I would say if you 0:21:01.012 --> 0:21:04.412 don't want China and Russia to win, maybe don't allow 0:21:04.412 --> 0:21:08.612 them to control the narrative in every other country other 0:21:08.612 --> 0:21:15.212 than the US. F-35's head to Puerto Rico as US 0:21:15.252 --> 0:21:23.891 Venezuela standoff heats up. Hmm. Yeah. Sending ten F-35s to 0:21:23.932 --> 0:21:33.772 Puerto Rico after Venezuelan F-16s buzzed a Navy destroyer. And. Yeah, 0:21:33.772 --> 0:21:36.612 and then that's when that, uh, cartel boat got hit. 0:21:39.092 --> 0:21:45.212 Strange times. Uh, State Department says non-immigrant visa applicants now 0:21:45.212 --> 0:21:51.262 need to interview from their actual country or their actual residence, residents, 0:21:51.302 --> 0:21:57.062 not inside the US. Sudan closes oil facilities after drone strikes. 0:21:57.502 --> 0:22:01.622 DHS wants $100 million for counter drone systems to protect America. 0:22:01.902 --> 0:22:05.222 I'm on board. I'm on board. I think we need 0:22:05.222 --> 0:22:12.222 to stop spending money on F-35s or whatever modern fighters. 0:22:12.222 --> 0:22:15.702 I think we need to move to the next phase. 0:22:16.821 --> 0:22:24.222 And China is using its private sector to advance military capabilities. Uh, yes. Yes. 0:22:24.582 --> 0:22:27.782 And it's explicit and it's like direct. And it's kind 0:22:27.782 --> 0:22:32.622 of super gross and nasty, but also not really. It's nationalistic. 0:22:32.662 --> 0:22:36.782 It's like, I mean, doesn't super bother me. What bothers 0:22:36.782 --> 0:22:40.422 me is when they claim it's not happening and they're 0:22:40.422 --> 0:22:45.942 trying to, you know, spread across the world and basically 0:22:46.502 --> 0:22:51.071 get control of everything. And they're hiding this and people 0:22:51.071 --> 0:22:54.752 are accepting them in like with Huawei, for example. Like 0:22:54.912 --> 0:22:58.632 that's the part that bothers me, them being nationalistic and 0:22:58.672 --> 0:23:02.552 having unity between their private sector and the goals of 0:23:02.552 --> 0:23:08.952 the country. And it doesn't really bother me. AI OpenAI 0:23:08.992 --> 0:23:13.672 launches an academy and jobs board to reskill displaced workers. 0:23:14.712 --> 0:23:17.752 I mean, to give credit to Sam Altman here, he's 0:23:17.752 --> 0:23:21.391 been saying for like the whole time, starting in like 22, 0:23:21.432 --> 0:23:24.832 probably way before that. But I started paying attention mostly 0:23:24.832 --> 0:23:28.311 in 22. And he's like, look, there's going to be 0:23:28.352 --> 0:23:33.112 tremendous job loss from AGI or AI in general or whatever. 0:23:33.112 --> 0:23:38.552 But definitely after AGI, depending on how you define it. 0:23:38.912 --> 0:23:40.472 But he's like, yeah, we're going to have to figure 0:23:40.472 --> 0:23:43.552 out things for people to do. He's been saying this 0:23:43.552 --> 0:23:47.952 over and over in interviews for years and years and years. Um, 0:23:47.952 --> 0:23:49.952 so if you didn't know that, then this is going 0:23:49.992 --> 0:23:52.072 to look kind of crazy. Like all the reports come 0:23:52.071 --> 0:23:55.472 out saying everyone's getting laid off. And then he's like, hey, 0:23:55.512 --> 0:23:59.752 you know, we launched a job board. It's a bulletin board. 0:23:59.752 --> 0:24:01.272 And you can come and you can pin a piece 0:24:01.272 --> 0:24:03.392 of paper to it, and someone can come by and 0:24:03.432 --> 0:24:10.072 offer you a job. Thanks, Sam. Super helpful. That's one 0:24:10.071 --> 0:24:11.672 way to look at it. Another way to look at 0:24:11.672 --> 0:24:14.872 it is obviously you would want to do this to 0:24:14.912 --> 0:24:19.071 help people. I'm not sure if this is the same 0:24:19.232 --> 0:24:24.311 article as the one competing with LinkedIn. Did I include 0:24:24.311 --> 0:24:26.592 that one or not? That was something else I read. 0:24:28.512 --> 0:24:30.272 I think it is. I think this is the one 0:24:30.272 --> 0:24:34.392 that people are saying is competing with LinkedIn. It's probably 0:24:34.432 --> 0:24:38.232 how could they possibly launch two? Anyway, there's another thing 0:24:38.232 --> 0:24:42.752 where they they're trying to compete directly with LinkedIn and 0:24:42.752 --> 0:24:45.832 basically help people find jobs by matching skills, which is 0:24:45.832 --> 0:24:49.962 exactly what LinkedIn Microsoft is doing. A lot of people 0:24:49.962 --> 0:24:56.401 forget that Microsoft owns LinkedIn. I forget sometimes, which is 0:24:56.402 --> 0:25:03.842 why I just said a lot of people. But really, 0:25:03.842 --> 0:25:08.882 it's just me. Anthropic agrees to pay authors at least 0:25:08.882 --> 0:25:13.762 1.5 billion in settlement. So they're doing this to try 0:25:13.762 --> 0:25:17.242 to avoid a much bigger one. This is for authors 0:25:17.242 --> 0:25:27.762 of books. So roughly $3,000 per book. Huh? Can I 0:25:27.762 --> 0:25:30.722 get $3,000? How do I find out if my data 0:25:30.722 --> 0:25:37.002 is in here? Anthropic valued at 183 billion after $13 billion. 0:25:37.002 --> 0:25:41.482 Raise anthropic blocks. Chinese owned firms from Claude. This is 0:25:41.482 --> 0:25:44.482 because of that threat research that came out that basically 0:25:44.482 --> 0:25:47.762 showed a whole bunch of Chinese companies are actually using 0:25:47.762 --> 0:25:51.641 it to do, you know, military oriented stuff and then 0:25:51.682 --> 0:25:57.562 handing it directly to the Chinese government. So they disallow 0:25:58.722 --> 0:26:03.522 listed a bunch of them. DeepMind's AI cuts gravitational wave 0:26:03.522 --> 0:26:11.282 detector noise by 100,000. I'm telling you DeepMind Google stuff, 0:26:11.602 --> 0:26:16.802 they are quiet and they just make these massive jumps. 0:26:17.002 --> 0:26:23.002 And I I'm very long on Google, very long on Google, 0:26:23.202 --> 0:26:26.762 especially as it relates to AI. I think they're just 0:26:27.362 --> 0:26:33.122 they're business mind is there, especially with Sundar. Like he's 0:26:33.122 --> 0:26:36.561 like the real deal. And they woke up on AI 0:26:36.602 --> 0:26:40.162 like they're just they're just crushing it. AI agents could 0:26:40.162 --> 0:26:44.601 reshape entire economies within the next decade. This is a 0:26:44.722 --> 0:26:49.892 arXiv archived paper. Agents on MD emerges as the universal 0:26:49.892 --> 0:26:53.932 standard for AI coding agent instructions. So I'm messing with 0:26:53.972 --> 0:26:59.732 using agents versus coded MD with Chi. I don't know. 0:26:59.772 --> 0:27:02.492 I've gone back and forth twice now. Um, right now 0:27:02.492 --> 0:27:05.371 I'm currently back on Cloudmd because I felt like I 0:27:05.372 --> 0:27:11.091 was getting less enforcement. Yeah, I'm going to watch it closely. 0:27:11.532 --> 0:27:14.812 High school senior says AI is destroying real learning for 0:27:14.811 --> 0:27:22.292 his generation. And, uh, yeah. Says classmates constantly use ChatGPT 0:27:22.332 --> 0:27:27.012 for everything from literature annotations to math homework and basically 0:27:27.012 --> 0:27:32.492 copy paste exercises from AI, which. Well, I mean, that's 0:27:32.492 --> 0:27:36.052 to be expected. The question is, are they also doing 0:27:36.052 --> 0:27:40.292 good things with it? And I think from this person's answer, 0:27:40.292 --> 0:27:43.131 the answer is no. And this was an article in 0:27:43.132 --> 0:27:51.102 The Atlantic. Okay. Um, evidence shows AI is already replacing 0:27:51.102 --> 0:27:55.382 entry level workers. Derek Thompson looks at this. Yeah, there's 0:27:55.382 --> 0:27:57.621 just more and more data on this coming out. This. 0:27:57.662 --> 0:28:00.901 I don't think it's really debatable at this point, but 0:28:00.902 --> 0:28:05.462 good to have additional data there in the newsletter. Technology. 0:28:05.582 --> 0:28:09.462 Google gets to keep Chrome but must end exclusive deals 0:28:09.702 --> 0:28:14.061 and share search data. So kind of a mostly a 0:28:14.102 --> 0:28:16.581 win I would say for Google to not get, uh, 0:28:17.702 --> 0:28:22.182 Chrome divested out. OpenAI moves to make its own AI 0:28:22.222 --> 0:28:26.581 chips with Broadcom and UC Santa Cruz. Engineers show Wi-Fi 0:28:26.622 --> 0:28:30.381 signals can measure heart rate. Yeah, I love this stuff. 0:28:30.422 --> 0:28:32.502 You can actually see through walls. You can measure your 0:28:32.502 --> 0:28:34.822 heart rate. You can. Yeah. You could do all sorts 0:28:34.821 --> 0:28:38.582 of stuff with Wi-Fi signals. Can't wait to see more 0:28:38.582 --> 0:28:42.022 of this. Tech comes to the consumer side. SpaceX can 0:28:42.022 --> 0:28:46.862 now launch 120 times a year from Florida humans. Job 0:28:46.862 --> 0:28:51.502 growth stalls as unemployment hits four year high. Looks like 0:28:51.502 --> 0:28:57.142 we added just 22,000 jobs in August. Got a new 0:28:57.142 --> 0:29:02.062 drug that outperforms aspirin for heart attack prevention. Private equity 0:29:02.062 --> 0:29:06.622 rentals make suburbs more diverse, but apply pressure to buyers. 0:29:06.982 --> 0:29:10.622 So sometimes it raises the prices, but sometimes it does 0:29:10.622 --> 0:29:14.542 bring in other people who normally couldn't get into that neighborhood. 0:29:15.302 --> 0:29:21.302 Interesting idea to have like. The houses being bought up 0:29:21.902 --> 0:29:26.342 by giant investment companies. And it's like I guess it's 0:29:26.342 --> 0:29:29.302 just another landlord. So I guess maybe it's not that 0:29:29.302 --> 0:29:33.622 much different than what we already have, but landlords now 0:29:34.222 --> 0:29:38.502 it's just a guy, you know, just a guy or 0:29:38.542 --> 0:29:42.952 a woman or whatever. And it's just like. They have 0:29:42.952 --> 0:29:46.472 like three properties and that's their life. Their landlord feels 0:29:46.472 --> 0:29:49.632 kind of weird to just see houses go off the 0:29:49.632 --> 0:29:53.472 market and it's, you know, Blackrock or whoever it is. 0:29:54.912 --> 0:30:01.112 More people are opting out of news to protect their mood. 40% 0:30:01.152 --> 0:30:06.552 now skip news, according to Reuters. Mathematicians just broke an 0:30:06.592 --> 0:30:12.272 87 year old knot theory, conjecture and attention loops. Quietly 0:30:12.272 --> 0:30:16.032 rewrite your model of reality. Whatever you stare at long 0:30:16.032 --> 0:30:21.911 enough starts echoing back and reshaping you. Yeah. Who was it? 0:30:21.912 --> 0:30:27.152 Mark Twain. Somebody said, uh, whatever. You pay attention, you 0:30:27.152 --> 0:30:29.592 become whatever you pay attention to. So be careful what 0:30:29.592 --> 0:30:32.752 you pay attention to. Not sure if that was Mark 0:30:32.752 --> 0:30:41.032 Twain or not. Phones on the toilet raise hemorrhoid risk. Discovery, 0:30:41.032 --> 0:30:45.192 Rick Rubin's The Way of Code. This is a very 0:30:45.192 --> 0:30:49.872 strange site. Gorgeous. Not sure exactly how it has to 0:30:49.872 --> 0:30:51.552 do with code, but I was just checking it out. 0:30:51.592 --> 0:30:57.712 Quite nice. Muscle memory turns repeated agent tasks into deterministic replays. 0:30:57.712 --> 0:31:01.352 This thing is sick, so it monitors what your AI 0:31:01.352 --> 0:31:04.112 is doing on your behalf when it's calling tools and everything. 0:31:04.832 --> 0:31:07.112 And then it's just like learning how to do that 0:31:08.312 --> 0:31:11.952 and turning that into steps that it can then replay. 0:31:12.832 --> 0:31:16.432 So yeah, it's just kind of like learning from what 0:31:16.472 --> 0:31:21.232 AI does and incorporating that into your workflows. Being good 0:31:21.232 --> 0:31:26.792 isn't enough to win anymore. John Sword's last name is swords. 0:31:26.792 --> 0:31:28.912 Is that the coolest last name? It's one of them. 0:31:29.712 --> 0:31:34.792 Argues that craft alone doesn't carry you. You need distribution, 0:31:34.792 --> 0:31:39.882 timing and taste. Or your good never gets found. This 0:31:39.882 --> 0:31:42.122 is why I'm arguing that everyone needs to become a creator. 0:31:42.122 --> 0:31:44.522 It's not that you need to become what a creator 0:31:44.522 --> 0:31:50.522 is today, but creator just needs to be the new 0:31:50.522 --> 0:31:54.442 way of living, which is thinking, building, sharing. I'm arguing 0:31:54.442 --> 0:31:59.482 that these should be human things, not creator things. Blogs 0:31:59.482 --> 0:32:02.161 used to be weird and personal. So this is a 0:32:02.242 --> 0:32:05.202 blog by Jet Girl. Really really good. You should check 0:32:05.202 --> 0:32:09.402 it out. Cutting end to end test time. 84% with 0:32:09.402 --> 0:32:14.122 Cloud Code SDK. Cloud code. Cloud code SDK. By the way, 0:32:14.122 --> 0:32:16.882 is just you run cloud with a switch of P 0:32:17.602 --> 0:32:22.842 and that is a standalone like AI. It's it's completely insane. 0:32:22.842 --> 0:32:25.642 I don't know why they call it the SDK. It's 0:32:25.642 --> 0:32:31.922 just a switch P command. Anyway, probably something I'm missing there. 0:32:31.922 --> 0:32:36.282 But anyway, that is it's extremely powerful functionality and I'm 0:32:36.322 --> 0:32:38.892 going to be using it a lot more. Make your 0:32:38.892 --> 0:32:43.372 own handwriting font easily. Claude FAQs tool use unless you 0:32:43.372 --> 0:32:47.372 enforce strict tool calling. I have this in tons of 0:32:47.372 --> 0:32:51.372 my scaffolding and I have this. Well, it's all in 0:32:52.092 --> 0:32:53.892 the Pi repo now you can go check it out. 0:32:54.572 --> 0:32:58.092 AI enables live phone calls across language barriers. Oh, I 0:32:58.092 --> 0:33:02.092 just saw the, um, Apple event to today. Two new 0:33:02.092 --> 0:33:06.812 AirPods live transcription. Talking to anybody. They could just talk 0:33:06.812 --> 0:33:10.612 in their language and you get the translation in your ear. 0:33:11.252 --> 0:33:15.452 Absolutely insane. I wonder if it works at the haircut 0:33:15.452 --> 0:33:22.092 place with Vietnamese people. Like having conversations around you and 0:33:22.092 --> 0:33:26.612 you don't know if it's like. Damn, his feet are big. 0:33:26.612 --> 0:33:29.572 Or if it's like they're trying to figure out lunch. 0:33:29.612 --> 0:33:38.492 Like you're not quite sure. Tech interviews reward complexity over 0:33:38.492 --> 0:33:44.492 practical engineering. Diallo's blog post. Yeah, this is a really 0:33:44.492 --> 0:33:50.532 good blog on kind of the problems with overall interviewing, which, yeah, 0:33:50.532 --> 0:33:57.092 there's just yeah, this is why a lot of AI interviewing, 0:33:57.292 --> 0:34:00.732 recruiting services I think are starting to do fairly well 0:34:01.452 --> 0:34:04.052 is not so much that they're awesome, it's just that 0:34:04.052 --> 0:34:10.332 the status quo is notably bad. Chibi. I think that's 0:34:10.332 --> 0:34:14.452 how you pronounce it, analyzes why users churn. This is 0:34:14.492 --> 0:34:16.852 actually just a straight link to a vendor, and I 0:34:16.852 --> 0:34:20.692 just thought the vendor was pretty cool. And um, no 0:34:20.692 --> 0:34:25.532 sponsorship obviously. Otherwise I'd have like a block for them and, 0:34:25.572 --> 0:34:29.972 you know, the asterisk or sponsorship or something to indicate 0:34:29.972 --> 0:34:32.972 that it's a sponsor. But no, this is just some 0:34:32.972 --> 0:34:35.252 pretty cool tech. I feel like churn is like one 0:34:35.422 --> 0:34:40.142 of the biggest things. It's like not about the gaining 0:34:40.182 --> 0:34:44.942 the users. It's about keeping the users. Random walks almost 0:34:44.942 --> 0:34:47.862 never return home in three dimensions. This is a cool 0:34:47.862 --> 0:34:54.502 physics thing. Recommendation. Whoa. Recommendation of the week. I'm going 0:34:54.502 --> 0:34:57.902 to reuse one that I typically have to ask myself 0:34:57.942 --> 0:35:02.022 and kind of always try to ask myself. And that 0:35:02.022 --> 0:35:06.942 is especially important right now with like extraordinary change, I think. 0:35:06.942 --> 0:35:09.542 What would you do if you're not afraid? Now, normally 0:35:09.542 --> 0:35:11.862 I would follow up and say, you shouldn't be afraid, right? 0:35:11.902 --> 0:35:15.422 Everything's fine. And, you know, you should just be okay 0:35:15.422 --> 0:35:18.142 and think what you would do if you weren't afraid. 0:35:19.302 --> 0:35:24.902 But everything isn't okay. But that doesn't mean it's still 0:35:24.902 --> 0:35:27.022 not a really good time to ask this question. So 0:35:27.062 --> 0:35:30.342 I think the question still stands. What would you do 0:35:30.342 --> 0:35:33.622 if you were not afraid? Here's what I would challenge 0:35:33.622 --> 0:35:39.102 here Just to add to this. Think about what your 0:35:39.102 --> 0:35:44.022 ideal self looks like. Think about what you and your 0:35:44.022 --> 0:35:48.942 perfect Super Saiyan form. What abilities and skills would you have? 0:35:48.942 --> 0:35:52.062 What confidence would you have? What narratives would you be 0:35:52.062 --> 0:35:56.502 able to tell people? And here's the question is that 0:35:56.502 --> 0:36:00.022 person more prepared for what is coming in the future 0:36:00.022 --> 0:36:03.182 than you are? And if that's the case, then you 0:36:03.182 --> 0:36:07.462 need to think up some dramatic ways, some extreme ways 0:36:07.462 --> 0:36:12.102 to possibly migrate to that version of yourself, because that 0:36:12.102 --> 0:36:15.022 version is going to survive a lot better maybe, than 0:36:15.022 --> 0:36:18.622 your current version. And the aphorism of the week, very 0:36:18.622 --> 0:36:21.782 much in line with this, the privilege of a lifetime 0:36:21.782 --> 0:36:25.822 is to become who you truly are. The privilege of 0:36:25.822 --> 0:36:30.382 a lifetime is to become who you truly are. Carl Jung.