1 00:00:22,363 --> 00:00:25,633 S1: Welcome to Unsupervised Learning, a security, AI and meaning focused 2 00:00:25,633 --> 00:00:28,513 S1: podcast that looks at how best to thrive as humans 3 00:00:28,513 --> 00:00:32,743 S1: in a post AI world. It combines original ideas, analysis, 4 00:00:32,743 --> 00:00:35,953 S1: and mental models to bring not just the news, but 5 00:00:35,953 --> 00:00:43,143 S1: why it matters and how to respond. All right. Welcome 6 00:00:43,143 --> 00:00:46,353 S1: to unsupervised learning, Mr. Daniel Missler. All right, let's get 7 00:00:46,353 --> 00:00:49,503 S1: into it here. So, yeah, doing a whole bunch of 8 00:00:49,503 --> 00:00:52,773 S1: new stuff with the desktop swapped out Kitty as the 9 00:00:52,773 --> 00:00:56,313 S1: new terminal got rid of alacrity, which was cool for 10 00:00:56,313 --> 00:00:59,043 S1: a hot minute there, but I think that lasted about 11 00:00:59,043 --> 00:01:01,293 S1: a week until I realized it had a lot of problems. 12 00:01:01,293 --> 00:01:05,553 S1: Someone recommended Kitty in the GitHub project for alacrity, and 13 00:01:05,553 --> 00:01:07,293 S1: I ended up switching over. And I just like a 14 00:01:07,293 --> 00:01:09,753 S1: better for lots of different reasons. And both of them, 15 00:01:09,753 --> 00:01:12,093 S1: in my opinion, are better than I term two. Both 16 00:01:12,093 --> 00:01:15,783 S1: of them have text configuration for for configuration, you don't 17 00:01:15,783 --> 00:01:18,963 S1: have to go into like settings with a UI as oxide. 18 00:01:18,963 --> 00:01:23,913 S1: You buy for a window management SCT for managing shortcuts 19 00:01:23,913 --> 00:01:27,393 S1: for your buy sketchy bar for this new bar. You 20 00:01:27,393 --> 00:01:30,843 S1: can see it up here and uh, Sto for syncing files. 21 00:01:30,843 --> 00:01:33,363 S1: Pretty cool stuff. And we're going to be talking about 22 00:01:33,363 --> 00:01:36,363 S1: this a whole bunch during a UL meeting on Thursday. 23 00:01:36,363 --> 00:01:38,943 S1: So that is coming up this Thursday if you are 24 00:01:38,943 --> 00:01:41,913 S1: a member. And yeah, I haven't been this excited about 25 00:01:41,913 --> 00:01:45,063 S1: a desktop configuration. Like I don't want to if I 26 00:01:45,063 --> 00:01:47,463 S1: start flipping around right now to show you it's going 27 00:01:47,463 --> 00:01:49,383 S1: to become a whole video and we don't have time 28 00:01:49,383 --> 00:01:51,903 S1: for that. So I'm just going to get into the show. 29 00:01:51,903 --> 00:01:54,933 S1: But I'm more excited than I've been for a very 30 00:01:54,933 --> 00:01:58,893 S1: long time about desktop configuration and just optimization. I'm just 31 00:01:58,893 --> 00:02:01,263 S1: I don't want to leave my home keys, right? I 32 00:02:01,263 --> 00:02:04,293 S1: want to vim all the things right. And that means 33 00:02:04,293 --> 00:02:07,623 S1: like screen management. I'm looking at three screens here. I've 34 00:02:07,623 --> 00:02:12,153 S1: got a, you know, the Apple Z1 in the middle 35 00:02:12,153 --> 00:02:14,583 S1: and I've got two LG five KS, so it's six 36 00:02:14,583 --> 00:02:16,683 S1: K in the middle, two five KS on the side. 37 00:02:16,683 --> 00:02:19,563 S1: I can move between them, I can move between desktops 38 00:02:19,563 --> 00:02:22,923 S1: inside of it inside, you know the Mac OS management 39 00:02:22,923 --> 00:02:27,273 S1: of desktops as well. And most importantly applications and windows 40 00:02:27,273 --> 00:02:30,603 S1: inside of the interface. But again, I don't want to 41 00:02:30,603 --> 00:02:33,663 S1: talk about it too much. Already have. All right. So 42 00:02:33,663 --> 00:02:37,713 S1: fabric now supports clod opus and opus is actually the 43 00:02:37,713 --> 00:02:41,343 S1: first model that I've seen that is actually better. I've 44 00:02:41,343 --> 00:02:45,273 S1: seen tangibly be better than GPT four. So I didn't 45 00:02:45,273 --> 00:02:47,853 S1: think that was going to happen before GPT five came out, 46 00:02:47,853 --> 00:02:52,053 S1: but it did. Opus is actually better in numerous cases 47 00:02:52,053 --> 00:02:55,023 S1: that I've actually physically seen, and one of them is 48 00:02:55,023 --> 00:02:58,173 S1: actually finding hidden message, which I talk about right here. 49 00:02:58,173 --> 00:03:02,043 S1: So this it's actually much better on this particular thing. Yeah. 50 00:03:02,043 --> 00:03:04,713 S1: So this is essentially how you use it. So you 51 00:03:04,713 --> 00:03:08,223 S1: do PB paste or whatever the input is. You take 52 00:03:08,223 --> 00:03:11,343 S1: any input from stdin or standard in and then you 53 00:03:11,343 --> 00:03:14,433 S1: send it into fabric extract ideas, and then you set 54 00:03:14,433 --> 00:03:17,013 S1: it to Model cloud three opus and you give the 55 00:03:17,013 --> 00:03:19,143 S1: model number. And we're actually about to make it. So 56 00:03:19,143 --> 00:03:20,703 S1: you don't need the model number. You could just send 57 00:03:20,703 --> 00:03:23,343 S1: it to cloud three opus and they'll use the latest one. 58 00:03:23,343 --> 00:03:26,013 S1: Also added a new pattern. That's basically how you use 59 00:03:26,013 --> 00:03:29,913 S1: cloud but added a new pattern called Extract Predictions. So 60 00:03:29,913 --> 00:03:33,123 S1: I'm going to basically send this content from whoever. And 61 00:03:33,123 --> 00:03:35,853 S1: it's going to rate including my own stuff, and it's 62 00:03:35,853 --> 00:03:38,343 S1: going to rate like how certain I am about the, 63 00:03:38,343 --> 00:03:41,373 S1: the prediction and also what the actual prediction was. And 64 00:03:41,373 --> 00:03:43,353 S1: when I said it would happen by and I'm going 65 00:03:43,353 --> 00:03:45,393 S1: to collect this for a group of people and then 66 00:03:45,393 --> 00:03:48,783 S1: retroactively look back and say, you know, I'm not sure 67 00:03:48,783 --> 00:03:51,123 S1: like how good they were at doing this, because it 68 00:03:51,123 --> 00:03:54,483 S1: looks like all these things are wrong. And with automation 69 00:03:54,483 --> 00:03:56,733 S1: plus AI, you have the ability to just like, go 70 00:03:56,733 --> 00:03:59,313 S1: get everything they ever said and look at all their 71 00:03:59,313 --> 00:04:02,433 S1: different predictions and see how valuable they were. And fine 72 00:04:02,433 --> 00:04:05,793 S1: hidden message is now more effective. And it actually comes 73 00:04:05,793 --> 00:04:08,373 S1: back and says, here's a really cynical view of what 74 00:04:08,373 --> 00:04:11,373 S1: they said, a normal view and then like a favorable view. 75 00:04:11,373 --> 00:04:13,653 S1: So that's pretty cool. I wrote a couple of new 76 00:04:13,653 --> 00:04:16,863 S1: essays this week. AI is already becoming like reading. Kind 77 00:04:16,863 --> 00:04:18,903 S1: of depressing. Not going to get into that. AI is 78 00:04:18,903 --> 00:04:22,083 S1: worse if you think it's someone who's fault, somebody fault. 79 00:04:22,083 --> 00:04:25,263 S1: So yeah, that one a little bit about framing essentially. 80 00:04:25,263 --> 00:04:29,103 S1: All right. Security. So had a Google engineer indicted for 81 00:04:29,103 --> 00:04:33,033 S1: stealing AI trade secrets to benefit China and basically took 82 00:04:33,033 --> 00:04:38,943 S1: over 500 files related to TPUs. Essentially tensor processing units 83 00:04:38,943 --> 00:04:41,583 S1: is what it was. And he actually lives not too 84 00:04:41,583 --> 00:04:44,193 S1: far from here, a few streets from here actually, like 85 00:04:44,193 --> 00:04:47,163 S1: a five minute drive and, uh, yeah, pretty, pretty sad. 86 00:04:47,163 --> 00:04:50,343 S1: And what I talked about here was basically, if you 87 00:04:50,343 --> 00:04:54,063 S1: don't want to like, go crazy with what's going on 88 00:04:54,063 --> 00:04:58,683 S1: with like, how this keeps happening with Chinese employees because 89 00:04:58,683 --> 00:05:02,643 S1: China is overreaching. China loves espionage. I mean, this is 90 00:05:02,643 --> 00:05:06,363 S1: just common knowledge. And they love to tap into their, 91 00:05:06,363 --> 00:05:10,233 S1: you know, people who are deployed all over the place. Right? 92 00:05:10,233 --> 00:05:13,653 S1: And it doesn't I don't know the exact details of this, 93 00:05:13,653 --> 00:05:17,403 S1: but it's not like they're actually deployed like a military unit. 94 00:05:17,403 --> 00:05:19,353 S1: It's more like, hey, I could tap you on the 95 00:05:19,353 --> 00:05:24,003 S1: shoulder at any given moment. Somebody named Andrew Bustamante talked 96 00:05:24,003 --> 00:05:27,393 S1: about this in extreme detail. At one point he's like, look, 97 00:05:27,393 --> 00:05:31,263 S1: this is the biggest and scariest spy service in the 98 00:05:31,263 --> 00:05:35,613 S1: world because they consider all of their, all of their 99 00:05:35,613 --> 00:05:39,123 S1: citizens to basically be an extension of the spy network. 100 00:05:39,123 --> 00:05:41,233 S1: So at any given moment, they could. Just tap you 101 00:05:41,233 --> 00:05:43,663 S1: on the shoulder and say, hey, it is your duty 102 00:05:43,663 --> 00:05:47,413 S1: as a Chinese citizen to help China. And that does 103 00:05:47,413 --> 00:05:50,443 S1: not mean that anyone's going to respond to that, right? 104 00:05:50,563 --> 00:05:52,963 S1: I absolutely don't believe that people have been here for 105 00:05:52,963 --> 00:05:55,543 S1: a long time, have loyalty to the US, that they're 106 00:05:55,543 --> 00:05:57,193 S1: going to do that. Right. And that's why I talk 107 00:05:57,193 --> 00:06:00,193 S1: about this. It takes a lot of courage and wisdom 108 00:06:00,193 --> 00:06:06,643 S1: to simultaneously realize how bad Chinese government espionage is without 109 00:06:06,643 --> 00:06:09,523 S1: giving in to racism, right? And a really good answer 110 00:06:09,523 --> 00:06:12,103 S1: to this is to have a really good insider threat 111 00:06:12,103 --> 00:06:16,303 S1: program that looks at behavior rather than characteristics. But the 112 00:06:16,303 --> 00:06:19,453 S1: problem here is that those programs are really difficult to 113 00:06:19,453 --> 00:06:22,903 S1: build and run and maintain, and so they tend to 114 00:06:22,903 --> 00:06:26,203 S1: only exist in big companies like Google. And this was 115 00:06:26,203 --> 00:06:29,893 S1: actually Google where this this was discovered. So the question 116 00:06:29,893 --> 00:06:33,223 S1: is like where is it also happening but not being discovered? 117 00:06:33,433 --> 00:06:38,053 S1: Big story. Midnight Blizzard is basically pounding on Microsoft. And 118 00:06:38,053 --> 00:06:42,883 S1: that is a Russian group. Qnap alerts more Criticals. I mean, 119 00:06:43,333 --> 00:06:47,863 S1: never under any circumstances put a NAS online. It's like 120 00:06:47,953 --> 00:06:49,963 S1: like I say here, it's a perfect storm of the 121 00:06:49,963 --> 00:06:52,063 S1: most critical data that you don't want to lose or 122 00:06:52,063 --> 00:06:56,623 S1: have stolen. And then the worst possible code just absolutely nasty. 123 00:06:56,623 --> 00:07:00,373 S1: And thanks to Plex Track for sponsoring North Korean spies 124 00:07:00,373 --> 00:07:05,053 S1: hacked into South Korean chipmakers stealing designs A flipper zero 125 00:07:05,053 --> 00:07:07,153 S1: device was used to break into a Tesla, but it 126 00:07:07,153 --> 00:07:10,333 S1: was a little bit of like kind of a stretch here. 127 00:07:10,333 --> 00:07:12,253 S1: First of all, they didn't have to use a flipper zero. 128 00:07:12,253 --> 00:07:15,193 S1: They could have used anything that just stole Wi-Fi because 129 00:07:15,193 --> 00:07:18,223 S1: it was like a fake Ssid. And then you had 130 00:07:18,223 --> 00:07:20,683 S1: to have multiple things sort of stacked together. You had 131 00:07:20,683 --> 00:07:22,633 S1: to be right next to it, like a bunch of 132 00:07:22,633 --> 00:07:25,363 S1: stuff that didn't make it super serious, which is why 133 00:07:25,363 --> 00:07:27,943 S1: I think Tesla didn't take it that serious either. Scammers 134 00:07:27,943 --> 00:07:31,273 S1: are increasingly using AI to mimic the voices of loved ones, 135 00:07:31,273 --> 00:07:34,213 S1: and my advice here is to let your most vulnerable 136 00:07:34,213 --> 00:07:38,143 S1: family members and friends know that scammers can now fake 137 00:07:38,143 --> 00:07:41,713 S1: voices and everything, and that if they think you're in trouble, 138 00:07:41,713 --> 00:07:44,023 S1: they should try to get in contact with you or 139 00:07:44,023 --> 00:07:46,723 S1: get in contact with someone who knows that you're okay, 140 00:07:46,723 --> 00:07:50,023 S1: someone around you, in your circle and in scammers. Do, 141 00:07:50,023 --> 00:07:52,513 S1: I think, try to do this? I know this for 142 00:07:52,513 --> 00:07:54,583 S1: a fact that they do. Not all can do this 143 00:07:54,583 --> 00:07:58,003 S1: because this takes time and effort. It it's what we 144 00:07:58,003 --> 00:08:01,183 S1: call in security world. Expensive to try to determine what 145 00:08:01,183 --> 00:08:04,003 S1: your current state is, but the more advanced the attack is, 146 00:08:04,003 --> 00:08:05,743 S1: the more money they would get from it, the more 147 00:08:05,743 --> 00:08:08,743 S1: likely it is to happen. But the point is, scammers 148 00:08:08,743 --> 00:08:11,533 S1: do try to figure out how to do this when 149 00:08:11,533 --> 00:08:13,903 S1: they know you won't be able to get in contact 150 00:08:13,903 --> 00:08:16,663 S1: with them. Okay, if you're on a vacation or you're, 151 00:08:16,663 --> 00:08:19,723 S1: you know, whitewater rafting or something and they know you're 152 00:08:19,723 --> 00:08:23,113 S1: on the river, again, this is less likely. But if 153 00:08:23,113 --> 00:08:25,183 S1: they know you're on a river, you don't have cell 154 00:08:25,183 --> 00:08:28,093 S1: phone connection. Well, that's the perfect time to call grandma. 155 00:08:28,513 --> 00:08:31,783 S1: And here's the problem. People post that they're whitewater rafting. 156 00:08:31,783 --> 00:08:36,073 S1: People post and say, I'm going whitewater rafting. I will 157 00:08:36,073 --> 00:08:40,813 S1: be unapproachable, unreachable for five hours. See you when I 158 00:08:40,813 --> 00:08:44,023 S1: get back. Well, thanks for mentioning that because now I 159 00:08:44,023 --> 00:08:46,723 S1: can call grandma and say, hey, they had a rafting 160 00:08:46,723 --> 00:08:50,653 S1: accident and the whole family needs surgery and we need 161 00:08:50,653 --> 00:08:54,763 S1: $80,000 to get the best doctors. And guess what? Grandma 162 00:08:54,763 --> 00:08:58,213 S1: immediately calls you and gets no answer. Or even worse, 163 00:08:58,393 --> 00:09:01,243 S1: the phone is out of service. Holy crap, now they're 164 00:09:01,243 --> 00:09:04,123 S1: freaking out. And then they use these tactics to just 165 00:09:04,123 --> 00:09:06,103 S1: make it more and more pressure. It's like, oh, the 166 00:09:06,103 --> 00:09:08,533 S1: doctors really worried. They said, they have to start now. 167 00:09:08,563 --> 00:09:10,963 S1: I need your credit card. And so that that sense 168 00:09:10,963 --> 00:09:13,543 S1: of urgency is really, really bad. So you've got to 169 00:09:13,543 --> 00:09:16,603 S1: train the more vulnerable people in your life to like, 170 00:09:16,603 --> 00:09:19,753 S1: get ready for this. It's especially bad. And this is 171 00:09:19,753 --> 00:09:21,583 S1: kind of the point of this one is like, what 172 00:09:21,583 --> 00:09:24,283 S1: if it's in your voice? Okay, it's a voice message 173 00:09:24,283 --> 00:09:28,393 S1: from you saying, please wire money to this account. Otherwise 174 00:09:28,393 --> 00:09:31,333 S1: these kidnappers are going to kill me, okay? And it is. 175 00:09:31,333 --> 00:09:34,003 S1: It sounds exactly like you. I mean, that that is 176 00:09:34,003 --> 00:09:36,133 S1: not in the future. That's like right now they can 177 00:09:36,133 --> 00:09:39,553 S1: make something sound like anyone. So you've got to prepare 178 00:09:39,553 --> 00:09:43,513 S1: your family and friends for this US sanctioned individual and 179 00:09:43,513 --> 00:09:46,603 S1: entities behind predator spyware. I like the fact that they're 180 00:09:46,603 --> 00:09:49,843 S1: going after people like this. Cloudflare has a new firewall 181 00:09:49,843 --> 00:09:54,133 S1: for AI and looking at protecting yeah, basically AI from 182 00:09:54,133 --> 00:09:58,693 S1: security threats. And my my analysis here is so nimble. 183 00:09:58,693 --> 00:10:01,993 S1: These guys, they they basically seep into all the cracks. 184 00:10:02,023 --> 00:10:06,643 S1: They're slowly becoming the internet. I honestly believe this. Cloudflare 185 00:10:06,643 --> 00:10:09,763 S1: is quietly becoming the internet, right. If Google like gets 186 00:10:09,763 --> 00:10:12,793 S1: rid of Gmail and YouTube out of sheer stupidity, which 187 00:10:12,793 --> 00:10:16,033 S1: is not out of question, and Akamai gets bought by 188 00:10:16,033 --> 00:10:19,453 S1: whatever Johnson and Johnson and boom, suddenly Cloudflare is like 189 00:10:19,453 --> 00:10:23,083 S1: the internet. Brian Krebs has a really cool analysis of radar, 190 00:10:23,113 --> 00:10:26,533 S1: a data broker that sells all sorts of American data, 191 00:10:26,533 --> 00:10:29,443 S1: and they've got ties to Russia and all sorts of badness. 192 00:10:29,443 --> 00:10:32,353 S1: And Krebs does a great job as as always taking 193 00:10:32,353 --> 00:10:35,863 S1: that down. And Russia, setting up fake news sites. Can't 194 00:10:35,863 --> 00:10:39,553 S1: wait to actually start consuming these things. I'm going to 195 00:10:39,553 --> 00:10:43,973 S1: start consuming. Tons of online content and actually doing multiple 196 00:10:43,973 --> 00:10:46,223 S1: filters on it. And one of them is going to 197 00:10:46,223 --> 00:10:49,433 S1: be looking for propaganda. So I'm hoping to detect these 198 00:10:49,433 --> 00:10:53,813 S1: types of networks before anyone else does. China is increasing 199 00:10:53,813 --> 00:10:57,833 S1: its defense budget by 7.2%, and this more than doubles 200 00:10:57,833 --> 00:11:03,203 S1: the military budget under Xi's leadership. Technology for AI is 201 00:11:03,203 --> 00:11:05,813 S1: a cold calling AI service. You have got to check 202 00:11:05,813 --> 00:11:08,303 S1: this thing out. It is the craziest thing and I've 203 00:11:08,303 --> 00:11:10,763 S1: got links here to the video. I don't know, I 204 00:11:10,763 --> 00:11:12,473 S1: might start doing a longer version of this where I 205 00:11:12,473 --> 00:11:15,293 S1: can actually play the videos, but this, this one here. 206 00:11:15,293 --> 00:11:20,183 S1: New voice demo. This one actually has a female southern 207 00:11:20,183 --> 00:11:23,423 S1: person calling and you can hear the accent. It's really endearing. 208 00:11:23,423 --> 00:11:26,003 S1: I spent like, whatever, 12 years in Georgia, got my 209 00:11:26,003 --> 00:11:28,733 S1: girl from there. And like, the accent is just amazing 210 00:11:28,733 --> 00:11:31,733 S1: and it just kind of makes you open up and like, 211 00:11:31,733 --> 00:11:34,103 S1: listen a little bit more, which is the whole key 212 00:11:34,103 --> 00:11:38,153 S1: to a cold call. And anyway, this thing is it's insane. 213 00:11:38,153 --> 00:11:42,353 S1: It's the coolest. Like sales automation thing I've ever seen. 214 00:11:42,353 --> 00:11:44,663 S1: And it yeah, it uses AI. All right. So it's 215 00:11:44,663 --> 00:11:49,883 S1: been giving AIS different IQ tests actually this matrix IQ test. 216 00:11:49,883 --> 00:11:52,943 S1: But they're giving different LMS the matrix IQ test. And 217 00:11:52,943 --> 00:11:56,453 S1: Claude three just broke over 100 for the first time. 218 00:11:56,453 --> 00:12:00,053 S1: And 100 is like the average human IQ. Apple has 219 00:12:00,053 --> 00:12:04,193 S1: quickly shifted from a passive stance in AI to going 220 00:12:04,193 --> 00:12:06,323 S1: crazy with it, and can't wait to see what they 221 00:12:06,323 --> 00:12:11,033 S1: do in September. Apple podcasts now have auto generated transcripts. 222 00:12:11,033 --> 00:12:14,693 S1: Research is saying that models will soon be able to 223 00:12:14,693 --> 00:12:18,353 S1: optimize prompts better than people, so prompt engineering might be 224 00:12:18,353 --> 00:12:22,253 S1: just going away. I think that's kind of right, but 225 00:12:22,253 --> 00:12:25,973 S1: at the same time, you can't just give it total garbage, right? 226 00:12:26,153 --> 00:12:28,463 S1: I don't see some random person be able to just 227 00:12:28,463 --> 00:12:31,433 S1: bark idiocy at a model and have it say, oh, 228 00:12:31,433 --> 00:12:35,333 S1: you were referring to Feynman's third principle. Indeed. Let us proceed. 229 00:12:35,333 --> 00:12:37,043 S1: Don't think it's going to be able to do that, 230 00:12:37,043 --> 00:12:40,313 S1: but if you do voice something that's halfway decent, they 231 00:12:40,313 --> 00:12:42,083 S1: might be able to like, clean it up and be like, 232 00:12:42,083 --> 00:12:46,643 S1: I think you meant this. Global trust in AI is falling, waning, 233 00:12:46,643 --> 00:12:51,373 S1: dropping off with a significant drop from 61% in 2019 234 00:12:51,373 --> 00:12:55,253 S1: to 53% now. And this is why I wrote one 235 00:12:55,253 --> 00:12:57,743 S1: of the essays that I was talking about up above 236 00:12:57,743 --> 00:13:02,273 S1: France just put the right to abortion directly in the Constitution, 237 00:13:02,273 --> 00:13:04,973 S1: and somebody set a new rule capping credit card late 238 00:13:04,973 --> 00:13:07,553 S1: fees at $8. This is going to cost banks a 239 00:13:07,553 --> 00:13:10,433 S1: lot of money. A lot of people are basically saying, hey, look, 240 00:13:10,433 --> 00:13:12,563 S1: they're just going to find ways to get that money 241 00:13:12,563 --> 00:13:14,963 S1: back with other fees, but it doesn't matter. I like 242 00:13:14,963 --> 00:13:17,603 S1: the fact that the government is actually pushing back on 243 00:13:17,603 --> 00:13:23,243 S1: this really cool, raw image. Look at this. This is insane. 244 00:13:23,243 --> 00:13:25,943 S1: This is why I want to do a video podcast 245 00:13:25,943 --> 00:13:29,813 S1: instead of a audio. I mean, look at this. I've 246 00:13:29,813 --> 00:13:32,543 S1: already got the Hubble version of this, but this is 247 00:13:32,543 --> 00:13:35,453 S1: the new one from the Webb telescope. Look at this thing. 248 00:13:35,453 --> 00:13:41,363 S1: This is ridiculous. Every single thing you see is a galaxy, 249 00:13:41,393 --> 00:13:45,113 S1: not a star, a galaxy. Actually, there might be some 250 00:13:45,113 --> 00:13:48,863 S1: stars that that happens sometimes, but yeah, really cool. So 251 00:13:48,863 --> 00:13:50,993 S1: I'm going to basically order this on a metal print 252 00:13:50,993 --> 00:13:53,573 S1: like I have for the Hubble. One single dose of 253 00:13:53,573 --> 00:14:00,233 S1: LSD shows promising results for treating anxiety 48% remission at 254 00:14:00,233 --> 00:14:03,683 S1: 12 weeks Sweden is now a NATO. New York has 255 00:14:03,683 --> 00:14:09,143 S1: rolled out National Guard troops to protect its subways. That's disturbing. 256 00:14:09,143 --> 00:14:14,723 S1: And Scotus ruled unanimously to keep Trump on state ballots. Yeah, 257 00:14:14,723 --> 00:14:18,623 S1: I consider what Colorado and some other people tried to 258 00:14:18,623 --> 00:14:22,283 S1: do to get him off the ballot as a Brexit move, 259 00:14:22,283 --> 00:14:24,623 S1: and I call it a Brexit move because it's the 260 00:14:24,623 --> 00:14:26,933 S1: type of thing that you try to do. You get 261 00:14:26,933 --> 00:14:29,363 S1: it done and then you immediately regret that you did 262 00:14:29,363 --> 00:14:32,963 S1: it right. You don't take people you dislike off the ballot. 263 00:14:32,963 --> 00:14:35,513 S1: You don't do that. It's not American. It's a bad thing. 264 00:14:35,513 --> 00:14:38,633 S1: And like I said down here, the moment you do 265 00:14:38,633 --> 00:14:41,303 S1: it and it actually works well when you're not in 266 00:14:41,303 --> 00:14:43,973 S1: control and it comes time to put up your own candidate, 267 00:14:43,973 --> 00:14:45,893 S1: the opposition is just going to do it to you 268 00:14:45,893 --> 00:14:47,423 S1: and it's going to get worse. And pretty soon you 269 00:14:47,423 --> 00:14:51,413 S1: have Banana Republic. It's not a real democracy. And here's 270 00:14:51,413 --> 00:14:54,833 S1: the other point. If somebody can be elected and you 271 00:14:54,833 --> 00:14:58,403 S1: find them so detestable, but they actually get elected like 272 00:14:58,403 --> 00:15:02,423 S1: they they legally are allowed to run in the country 273 00:15:02,423 --> 00:15:05,753 S1: votes for them. That's the game. That is the game. 274 00:15:05,753 --> 00:15:08,993 S1: And I do not like Trump. I'm not a Trump guy. 275 00:15:09,023 --> 00:15:11,573 S1: Doesn't matter. And we're not going into politics here. But 276 00:15:11,573 --> 00:15:14,153 S1: he's not my guy. The point is, if the country 277 00:15:14,153 --> 00:15:18,173 S1: votes him in and that very well may happen, then 278 00:15:18,173 --> 00:15:22,043 S1: that is American. That is American. For somebody to vote 279 00:15:22,043 --> 00:15:25,163 S1: in in a free election, of course, which I think 280 00:15:25,163 --> 00:15:28,043 S1: it will be. And it's like, that's what that's what 281 00:15:28,043 --> 00:15:31,253 S1: America wanted. So either change the country or leave. I mean, 282 00:15:31,253 --> 00:15:33,683 S1: those are my options at that point. All right. Ideas 283 00:15:33,683 --> 00:15:38,033 S1: and analysis. I was troubled with Harare's analysis on Colbert. 284 00:15:38,033 --> 00:15:40,583 S1: So he basically said, down here. He was talking to 285 00:15:40,583 --> 00:15:44,753 S1: Colbert and he basically says, we have no idea what 286 00:15:44,753 --> 00:15:48,083 S1: to teach kids growing up. Now we have no idea 287 00:15:48,083 --> 00:15:50,573 S1: to tell them what kind of skills to have. And 288 00:15:50,573 --> 00:15:52,463 S1: we used to be able to tell them, oh, you know, 289 00:15:52,463 --> 00:15:55,373 S1: shoot a bow, you know, ride a horse, you know, 290 00:15:55,373 --> 00:15:58,373 S1: farm wheat or whatever. And now we don't know what 291 00:15:58,373 --> 00:16:01,973 S1: to say. And I just don't think that was good. 292 00:16:01,973 --> 00:16:04,763 S1: I mean, he's he's the best. I got massive respect 293 00:16:04,763 --> 00:16:07,523 S1: for him. And I also want to give him, you know, 294 00:16:07,523 --> 00:16:12,233 S1: some props because like I say here, the lights are bright, right? 295 00:16:12,293 --> 00:16:15,143 S1: I mean you're encouraged to give like these little soundbites. 296 00:16:15,143 --> 00:16:18,623 S1: So so I'm not like throwing shade, but I just 297 00:16:18,623 --> 00:16:20,723 S1: wish he didn't come out and say, we have no 298 00:16:20,723 --> 00:16:24,083 S1: idea what to tell people because we actually do write 299 00:16:24,083 --> 00:16:28,553 S1: clear thinking, clear communication, understanding the past, understanding the merits 300 00:16:28,553 --> 00:16:32,573 S1: of various arguments, how to disagree, philosophy, critical thinking. All 301 00:16:32,573 --> 00:16:36,203 S1: these things are more important than ever and more obvious 302 00:16:36,203 --> 00:16:38,693 S1: than ever that we need them in the future. So 303 00:16:38,693 --> 00:16:40,733 S1: it's kind of the opposite of what he's saying. So 304 00:16:40,733 --> 00:16:43,433 S1: that's what I basically said in this tweet. All right. 305 00:16:43,433 --> 00:16:46,403 S1: Been drawn back into the Stoics lately, been reading a 306 00:16:46,403 --> 00:16:49,673 S1: whole lot of Marcus Aurelius, and really got a whole 307 00:16:49,673 --> 00:16:52,883 S1: bunch of stuff from Ryan Holiday's Stoic store. And you 308 00:16:52,883 --> 00:16:55,163 S1: should check that out. And oh, yeah, we're doing our 309 00:16:55,163 --> 00:16:58,493 S1: mid-month on optimization. I think I talked about that already, 310 00:16:58,493 --> 00:17:01,133 S1: and I should go sign up. That's the link to 311 00:17:01,133 --> 00:17:05,573 S1: go sign up Daniel missler.com/premium. And yeah, this newsletter was 312 00:17:05,573 --> 00:17:08,603 S1: super fun to do. Really fast, more creative than ever. 313 00:17:08,603 --> 00:17:11,243 S1: Just it's going great. Don't really need to say too 314 00:17:11,243 --> 00:17:14,483 S1: much about that. All right. Discovery obsidian as a graph 315 00:17:14,483 --> 00:17:17,093 S1: database for Rag. This is really cool because people put 316 00:17:17,093 --> 00:17:20,213 S1: so much effort into obsidian. Being able to apply AI 317 00:17:20,243 --> 00:17:23,333 S1: to it is super, super cool. Male in the middle 318 00:17:23,333 --> 00:17:29,003 S1: automate spearfishing. Enough said. Junaid Islam outlines a five step 319 00:17:29,003 --> 00:17:32,663 S1: method for cutting cybersecurity budgets. This was a pretty cool 320 00:17:32,663 --> 00:17:35,813 S1: write up. During World War Two, America fought against damaging 321 00:17:35,813 --> 00:17:40,223 S1: rumors with rumor clinics in newspapers and magazines. I think 322 00:17:40,223 --> 00:17:43,013 S1: we should look and do this right. We should look 323 00:17:43,013 --> 00:17:47,513 S1: into this so debunked lies by fact checking and publishing 324 00:17:47,513 --> 00:17:52,253 S1: the findings, helping to maintain morale and unity. Um, pretty cool. 325 00:17:52,373 --> 00:17:57,593 S1: I get Julia Evans, the surprisingly complex world of Jets head, 326 00:17:57,593 --> 00:18:02,513 S1: revealing its multifaceted roles. I am embarrassed by how often 327 00:18:02,513 --> 00:18:06,443 S1: I realized I don't know anything about Kit. I just 328 00:18:06,533 --> 00:18:09,083 S1: too many times per month, I just delete the whole 329 00:18:09,173 --> 00:18:12,593 S1: I'm like, I'm confused, delete the whole repo, re clone it. 330 00:18:12,593 --> 00:18:15,113 S1: And that's just sad. I need to get class. All right. 331 00:18:15,113 --> 00:18:17,933 S1: It's getting harder to tell humans apart from bots, but 332 00:18:17,933 --> 00:18:21,503 S1: not because bots are getting more smart or more like humans, 333 00:18:21,503 --> 00:18:25,313 S1: but because humans are acting more like bots. Oh that's disturbing. 334 00:18:25,313 --> 00:18:28,163 S1: How to start a home lab? Really good primer here 335 00:18:28,163 --> 00:18:33,293 S1: by Hayden James. And this guy Chen is all excited 336 00:18:33,293 --> 00:18:36,503 S1: about soft skills. He says 80% of failures at Amazon 337 00:18:36,503 --> 00:18:39,743 S1: are due to soft skill issues, not technical ability. And 338 00:18:39,743 --> 00:18:43,793 S1: it says that he got a job at Amazon even 339 00:18:43,793 --> 00:18:46,553 S1: though he wasn't technical because of this skill. I'm not 340 00:18:46,553 --> 00:18:49,013 S1: sure it was actually him that said 80% of failures, 341 00:18:49,013 --> 00:18:53,693 S1: but I wrote that anyway. All right. Tolkien intensely disliked 342 00:18:53,693 --> 00:18:59,213 S1: Dune because Tolkien wanted to talk about inherent acts of goodness, 343 00:18:59,363 --> 00:19:05,393 S1: and Dune talks about consequentialist morality of action judged by 344 00:19:05,393 --> 00:19:08,813 S1: their outcomes. I like it, I like it, and I 345 00:19:08,813 --> 00:19:10,973 S1: also like this word. I wrote a piece about this 346 00:19:10,973 --> 00:19:14,933 S1: a while back. Deontological. The morality of something depends on 347 00:19:14,933 --> 00:19:18,503 S1: the effects that it has as opposed to intentions. Pretty 348 00:19:18,503 --> 00:19:22,853 S1: cool concept. Someone reminisces about the simpler, less polished writing 349 00:19:22,853 --> 00:19:25,823 S1: they used to do. They express a longing. This is 350 00:19:25,823 --> 00:19:28,583 S1: on Reddit, by the way, a longing for their earlier, 351 00:19:28,583 --> 00:19:32,183 S1: unrefined work. Feeling it had charm that their current writing lacks. 352 00:19:32,183 --> 00:19:35,783 S1: This is absolutely true. I've experienced that in the past, 353 00:19:36,083 --> 00:19:38,873 S1: and Kate Hall shares how anyone can learn to be 354 00:19:38,873 --> 00:19:42,023 S1: more energetic. This is pretty cool. It's over on the 355 00:19:42,023 --> 00:19:45,623 S1: site every, which is a really cool site from Dan 356 00:19:45,623 --> 00:19:49,103 S1: Shipper and the recommendation of the week. Trust your routine. 357 00:19:49,103 --> 00:19:51,563 S1: If you spend a lot of time putting together a 358 00:19:51,563 --> 00:19:54,383 S1: routine that helps you feel good and positive, then you 359 00:19:54,383 --> 00:19:56,453 S1: got to follow it. You got to follow your routine. 360 00:19:56,453 --> 00:19:59,243 S1: Whenever I'm out of routine, I'm not feeling good. I'm like, ah, 361 00:19:59,243 --> 00:20:00,953 S1: it's got to be something else. It's got to be. 362 00:20:00,953 --> 00:20:02,483 S1: I got to figure out what it is. Is a 363 00:20:02,483 --> 00:20:06,593 S1: supplement or something? No, it's it's not a supplement. It's 364 00:20:06,593 --> 00:20:09,353 S1: not something special. You're just not on your routine. Right. 365 00:20:09,353 --> 00:20:13,403 S1: What's the routine? Something like this very Huberman esque sleep, 366 00:20:13,403 --> 00:20:17,513 S1: sun exercise, clean food, walking. Talk to your friends and family. 367 00:20:17,513 --> 00:20:20,903 S1: Like be social. It's the basics and there's a reason 368 00:20:20,903 --> 00:20:23,483 S1: you wrote them down, so go do them. And the 369 00:20:23,483 --> 00:20:26,813 S1: aphorism of the week. Choose not to be harmed and 370 00:20:26,813 --> 00:20:30,743 S1: you won't feel harmed. Don't feel harmed and you haven't been. 371 00:20:30,743 --> 00:20:33,443 S1: Choose not to be harmed and you won't feel harmed. 372 00:20:33,443 --> 00:20:36,083 S1: And if you don't feel harmed, then you haven't been. 373 00:20:36,083 --> 00:20:41,163 S1: Marcus Aurelius. Unsupervised learning is produced in. By Daniel Miller 374 00:20:41,163 --> 00:20:45,723 S1: on a Norman 87 AI microphone using Hindenburg. Intro and 375 00:20:45,723 --> 00:20:49,143 S1: outro music is by zombie with the Y, and to 376 00:20:49,143 --> 00:20:51,303 S1: get the text and links from this episode, sign up 377 00:20:51,303 --> 00:20:55,653 S1: for the newsletter version of the show at Daniel miller.com/newsletter. 378 00:20:56,913 --> 00:20:57,783 S1: We'll see you next time.