1 00:00:00,150 --> 00:00:02,400 S1: What do you call an endpoint security product that works 2 00:00:02,400 --> 00:00:06,360 S1: perfectly but makes users miserable? A failure? The old approach 3 00:00:06,360 --> 00:00:09,150 S1: to endpoint security is to lock down employee devices and 4 00:00:09,150 --> 00:00:13,320 S1: rollout changes through forced restarts, but it just doesn't work. 5 00:00:13,350 --> 00:00:16,260 S1: It is miserable because they got a mountain of support tickets. 6 00:00:16,260 --> 00:00:19,409 S1: Employees start using personal devices just to get their work done, 7 00:00:19,410 --> 00:00:21,930 S1: and executives up out the very first time. It makes 8 00:00:21,930 --> 00:00:24,270 S1: them late for a meeting. You can't have a successful 9 00:00:24,270 --> 00:00:28,050 S1: security implementation unless you work with end users, and that's 10 00:00:28,050 --> 00:00:31,890 S1: where collide comes in. Their user first device trust solution 11 00:00:31,890 --> 00:00:34,680 S1: notifies users as soon as it detects an issue on 12 00:00:34,680 --> 00:00:37,170 S1: their device, and teaches them how to solve it without 13 00:00:37,170 --> 00:00:40,350 S1: needing help from it. That way, untrusted devices are blocked 14 00:00:40,350 --> 00:00:44,280 S1: from authenticating, but users don't stay blocked. Collide is designed 15 00:00:44,280 --> 00:00:47,640 S1: for companies with Okta, and it works on Mac OS, windows, 16 00:00:47,640 --> 00:00:50,820 S1: Linux and mobile devices. So if you have Okta and 17 00:00:50,820 --> 00:00:54,420 S1: you're looking for a device trust solution that respects your team, 18 00:00:54,420 --> 00:00:59,520 S1: visit callide.com/unsupervised learning to watch a demo and see how 19 00:00:59,520 --> 00:01:08,910 S1: it works. That's callide.com collide.com/unsupervised learning. Welcome to Unsupervised Learning, 20 00:01:08,910 --> 00:01:11,970 S1: a security, AI and meaning focused podcast that looks at 21 00:01:11,970 --> 00:01:14,790 S1: how best to thrive as humans in a post AI world. 22 00:01:15,569 --> 00:01:19,559 S1: It combines original ideas, analysis, and mental models to bring 23 00:01:19,560 --> 00:01:22,110 S1: not just the news, but why it matters and how 24 00:01:22,110 --> 00:01:29,310 S1: to respond. All right. Welcome to unsupervised learning. This is 25 00:01:29,310 --> 00:01:32,490 S1: Daniel Missler. I have been traveling like a crazy person, 26 00:01:32,490 --> 00:01:36,810 S1: but have got to get this podcast out. Don't like 27 00:01:36,810 --> 00:01:41,550 S1: skipping weeks, so let's jump in. So it has been 28 00:01:41,550 --> 00:01:44,760 S1: a crazy week. Started for the last episode. Bit overexposed, 29 00:01:44,760 --> 00:01:48,330 S1: been doing lots of travel, interacting with humans and stuff 30 00:01:48,330 --> 00:01:51,510 S1: like that. Got some minor con flu or whatever it is, 31 00:01:51,510 --> 00:01:53,700 S1: but I didn't actually get it. I didn't get it bad. 32 00:01:53,700 --> 00:01:57,330 S1: It's just slightly congested. Can't tell if it's allergies or whatever, 33 00:01:57,330 --> 00:02:00,960 S1: but everything's fine. Got a new pattern here. Analyze presentation. 34 00:02:01,050 --> 00:02:03,419 S1: Say a new fabric pattern that looks at the transcript 35 00:02:03,420 --> 00:02:06,150 S1: of a presentation and tells you how much the presenter 36 00:02:06,150 --> 00:02:09,060 S1: is like trying to brag versus how much they're trying 37 00:02:09,060 --> 00:02:12,239 S1: to entertain versus how much they're actually trying to inform. 38 00:02:12,240 --> 00:02:15,869 S1: So it tells you essentially how self versus other centered 39 00:02:15,870 --> 00:02:19,470 S1: that they are. I spoke at equilibrium conference last week. 40 00:02:19,470 --> 00:02:22,080 S1: It was fantastic. Got to meet Chris Hughes for the 41 00:02:22,080 --> 00:02:25,889 S1: first time in person and appreciate the invite from Rohit. 42 00:02:25,889 --> 00:02:29,399 S1: It was a great event, just absolutely loved it and 43 00:02:29,400 --> 00:02:32,220 S1: got to attend a live UFC event for the first time. 44 00:02:32,220 --> 00:02:36,000 S1: It was so great. It was a UFC 300. It 45 00:02:36,000 --> 00:02:38,520 S1: was supposed to be like the best UFC ever. You 46 00:02:38,520 --> 00:02:40,410 S1: know how they always say that, but it turned out 47 00:02:40,410 --> 00:02:45,120 S1: it actually was. It was unbelievable. Especially the major fight 48 00:02:45,120 --> 00:02:49,350 S1: between Cage and Holloway. It was beyond all expectations. It 49 00:02:49,350 --> 00:02:51,570 S1: was just so great going to be speaking at Hardly 50 00:02:51,570 --> 00:02:55,230 S1: Strictly Security soon, speaking at the Security Frontiers conference that 51 00:02:55,230 --> 00:02:58,530 S1: are actually already happened, and also this one already happened. 52 00:02:58,530 --> 00:03:01,590 S1: This was yesterday and this was super fun. So thanks 53 00:03:01,590 --> 00:03:04,350 S1: to Jane Chief for the invite there and got to 54 00:03:04,350 --> 00:03:07,020 S1: meet and hang out with song for quite a while. 55 00:03:07,020 --> 00:03:10,200 S1: That was super fun to talk with her. She's super 56 00:03:10,200 --> 00:03:13,769 S1: bright and awesome and yeah, gearing up for RSA and 57 00:03:13,770 --> 00:03:16,350 S1: just finish reading A Sense of Style by Steven Pinker 58 00:03:16,350 --> 00:03:19,290 S1: for the third time. Highly recommend you put that in 59 00:03:19,290 --> 00:03:22,140 S1: your rotation for every few years. It's my favorite writing 60 00:03:22,139 --> 00:03:24,600 S1: and style guide, so we talked about this one. Analyze 61 00:03:24,600 --> 00:03:28,260 S1: presentation didn't really do any other essays or videos this 62 00:03:28,260 --> 00:03:30,600 S1: week just because all the travel and stuff, but they 63 00:03:30,600 --> 00:03:33,359 S1: are accumulating and I'm going to have a lot more 64 00:03:33,630 --> 00:03:36,390 S1: content coming out soon, and it'll be both an essay 65 00:03:36,390 --> 00:03:41,280 S1: form and YouTube. I'm actually transitioning really heavy towards YouTube, 66 00:03:41,280 --> 00:03:44,430 S1: so go subscribe to the channel if you're not subscribed yet. 67 00:03:44,430 --> 00:03:46,800 S1: It is. Unsupervised learning is the name of the channel 68 00:03:46,800 --> 00:03:50,220 S1: on YouTube. All right. Security. Microsoft says Chinese hackers are 69 00:03:50,220 --> 00:03:54,600 S1: using AI to inflame social tensions in the US. And 70 00:03:54,600 --> 00:03:56,640 S1: every time I hear about these, I can't wait to 71 00:03:56,640 --> 00:04:00,090 S1: build my propaganda tracker, because if somebody starts a campaign, 72 00:04:00,090 --> 00:04:02,580 S1: I want to see where it spreads, where it becomes viral. 73 00:04:02,580 --> 00:04:05,850 S1: And we have it's like Dan Kaminsky used to say, 74 00:04:05,850 --> 00:04:08,640 S1: we have the technology. I don't know if that's from 75 00:04:08,640 --> 00:04:12,840 S1: Bionic Man. Think it might be from Bionic Man. But anyway, yeah, 76 00:04:12,840 --> 00:04:15,180 S1: we have the technology to do this type of stuff, 77 00:04:15,180 --> 00:04:18,930 S1: especially once we start incorporating AI agents. So I'm really 78 00:04:18,930 --> 00:04:21,810 S1: looking forward to this. Yeah. And so the thing about 79 00:04:21,810 --> 00:04:25,110 S1: it is we need a combination of government and private 80 00:04:25,110 --> 00:04:28,529 S1: for this. Right. Because we need the thing to be trusted. 81 00:04:28,529 --> 00:04:31,710 S1: But we also need it to be dependable and actually 82 00:04:31,710 --> 00:04:34,770 S1: incentivized to work. And of course it needs funding. But 83 00:04:34,770 --> 00:04:35,940 S1: you got to come up with some kind of cool 84 00:04:35,940 --> 00:04:40,440 S1: name prop track or something. Propaganda tracking. Okay, over 92,000 85 00:04:40,440 --> 00:04:45,720 S1: D-Link NAS devices exposed nasty to IP has three high 86 00:04:45,720 --> 00:04:51,180 S1: severity vulnerabilities us just blacklisted for additional Chinese companies for 87 00:04:51,180 --> 00:04:56,549 S1: procuring AI chips for China's AI modernization efforts. And US 88 00:04:56,550 --> 00:05:00,420 S1: and China are racing to dominate with AI driven drone swarms. 89 00:05:00,420 --> 00:05:03,270 S1: Once again, you gotta check out this book by Daniel Suarez. 90 00:05:03,270 --> 00:05:07,620 S1: Kill decision. Really entertaining and also very prescient at the 91 00:05:07,620 --> 00:05:12,630 S1: same time. Google's contract reveals a partnership with Israel's Defense Ministry, 92 00:05:12,630 --> 00:05:16,469 S1: and this caused actually like a walkout or a protest 93 00:05:16,470 --> 00:05:19,380 S1: at Google. And I think it was 23 people. They 94 00:05:19,380 --> 00:05:22,740 S1: all got fired, which was pretty interesting. Technology. Great New 95 00:05:22,740 --> 00:05:26,880 S1: York Times piece on automating finance jobs using AI, and 96 00:05:26,880 --> 00:05:29,070 S1: got a cool quote here. Some of Wall Street's major 97 00:05:29,070 --> 00:05:31,500 S1: banks are asking the same question as they test AI 98 00:05:31,500 --> 00:05:34,800 S1: tools that can largely replace their armies of analysts by 99 00:05:34,800 --> 00:05:38,849 S1: performing in seconds, the work that now takes hours or 100 00:05:38,850 --> 00:05:43,080 S1: a whole weekend. That is a big jump. Seconds versus 101 00:05:43,080 --> 00:05:46,080 S1: hours or a whole weekend. Like like we've been saying, 102 00:05:46,440 --> 00:05:48,299 S1: it's like at this point, if you don't see the 103 00:05:48,300 --> 00:05:50,610 S1: impact of AI and jobs, it's because you don't read 104 00:05:50,610 --> 00:05:52,830 S1: enough or you don't want to see. I mean, this 105 00:05:52,830 --> 00:05:57,120 S1: is willful ignorance at this point. It isn't like the.com boom. 106 00:05:57,270 --> 00:06:00,869 S1: This is what everyone's saying. oh.com crypto. We got a 107 00:06:00,870 --> 00:06:03,390 S1: lot of people saying things like this and it's like, no, 108 00:06:03,390 --> 00:06:07,200 S1: this is tech that's instantly useful. And when Wall Street 109 00:06:07,200 --> 00:06:10,650 S1: is talking about this in very concrete terms of like, yeah, 110 00:06:10,650 --> 00:06:14,250 S1: we're getting rid of lots of people because seconds versus 111 00:06:14,250 --> 00:06:16,859 S1: hours or a whole weekend. I mean, I don't know 112 00:06:16,860 --> 00:06:19,530 S1: how you get more compelling and realistic than that, yo. 113 00:06:19,529 --> 00:06:22,920 S1: He continues to do great work around AI agents. He's 114 00:06:22,920 --> 00:06:26,970 S1: got one that does AI agent log visualization. Let's open 115 00:06:26,970 --> 00:06:30,150 S1: this one up. So basically you click through here and 116 00:06:30,150 --> 00:06:32,340 S1: yeah look at these. You click through and you can 117 00:06:32,339 --> 00:06:35,370 S1: see exactly what the agent did and get like logs 118 00:06:35,370 --> 00:06:39,360 S1: from it. Really cool. Project. Russia's getting 89% of its 119 00:06:39,360 --> 00:06:43,560 S1: chips through China as a way of bypassing sanctions. Gary 120 00:06:43,560 --> 00:06:47,550 S1: Marcus is betting 10 million that we won't have smarter AI, 121 00:06:47,550 --> 00:06:51,299 S1: smarter than humans by 2025. Yeah, the whole trick there 122 00:06:51,300 --> 00:06:54,270 S1: is like how they measure, right, what KPI they're actually using. 123 00:06:54,270 --> 00:06:57,930 S1: But I think this person's, uh, yeah, this is going 124 00:06:57,930 --> 00:07:00,780 S1: to come down to definition competition, I guarantee you. But 125 00:07:00,779 --> 00:07:03,540 S1: that is a bad bet. I think Mr. Ole just 126 00:07:03,540 --> 00:07:08,430 S1: released a eight times 22 billion model via torrent, which 127 00:07:08,430 --> 00:07:12,090 S1: is pretty cool. It's a really big files and distributed 128 00:07:12,090 --> 00:07:16,290 S1: and yeah, and the more people downloading it, the faster 129 00:07:16,290 --> 00:07:19,050 S1: it gets. So this is all the reason that transformers 130 00:07:19,350 --> 00:07:22,050 S1: or no torrents were made in the first place. So cool. 131 00:07:22,260 --> 00:07:25,050 S1: This paper says transformers capable of acting like a general 132 00:07:25,050 --> 00:07:30,030 S1: purpose computer. OpenAI is shipped new GPT four turbo. Although 133 00:07:30,030 --> 00:07:33,750 S1: I have tested it, and in my most difficult AI 134 00:07:33,750 --> 00:07:37,200 S1: jobs that I have, which are within fabric like detect 135 00:07:37,200 --> 00:07:41,070 S1: hidden message, for example, that is a extremely difficult human 136 00:07:41,070 --> 00:07:45,930 S1: cognitive function, like advanced logic is really needed there. And 137 00:07:45,930 --> 00:07:49,830 S1: opus still crushes, uh, GPT four in that particular one 138 00:07:49,830 --> 00:07:53,070 S1: and others that are similar. And after I beat them, 139 00:07:53,070 --> 00:07:55,890 S1: professional go players didn't just catch up, they became better 140 00:07:55,890 --> 00:07:59,310 S1: and more creative. I love this thing where it's like 141 00:07:59,310 --> 00:08:01,920 S1: we learned and we got better, but I don't want 142 00:08:01,920 --> 00:08:03,630 S1: people to think that it means we'll be able to 143 00:08:03,630 --> 00:08:07,440 S1: keep up. It really doesn't, but we still can find 144 00:08:07,440 --> 00:08:11,160 S1: things interesting and useful, even if it's just human to human. 145 00:08:11,160 --> 00:08:14,880 S1: And we're nowhere near as good as human versus computer 146 00:08:14,880 --> 00:08:18,780 S1: or computer versus computer. That's fine. Right? It's not. One 147 00:08:18,780 --> 00:08:20,640 S1: does not have to mean the other. So look at chess. 148 00:08:20,640 --> 00:08:22,950 S1: Chess is more popular than ever. And we got crushed 149 00:08:22,950 --> 00:08:25,890 S1: by computers in 97. And we're not getting much better. 150 00:08:25,890 --> 00:08:30,510 S1: Nanotech Onyx is rolling out Cube Fab, a moderate, a 151 00:08:30,510 --> 00:08:36,510 S1: modular chip fab powered by Nvidia to democratize semiconductor manufacturing. 152 00:08:36,660 --> 00:08:41,490 S1: Human pin got flamed by Mkbhd, and I think he 153 00:08:41,490 --> 00:08:43,350 S1: could say whatever he wants. I just think he was 154 00:08:43,350 --> 00:08:45,959 S1: a little harsh on this thing. And to me, it's 155 00:08:45,960 --> 00:08:48,599 S1: not like I want to control what he's saying. I 156 00:08:48,600 --> 00:08:51,329 S1: was just a little bit let down by the vibe. 157 00:08:51,360 --> 00:08:54,630 S1: I tend to think of him like as this nice, 158 00:08:54,630 --> 00:08:59,340 S1: balanced person. So when I see a headline that's just 159 00:08:59,340 --> 00:09:03,150 S1: like worst product ever, it just disappointed me at a 160 00:09:03,150 --> 00:09:07,410 S1: personality level, not at an analysis level. Nothing about his 161 00:09:07,410 --> 00:09:11,069 S1: analysis was wrong, and he actually was somewhat balanced in 162 00:09:11,070 --> 00:09:16,080 S1: the quite balanced inside of the actual review. My thing 163 00:09:16,080 --> 00:09:19,710 S1: is like when you put that spin of like, worst ever, 164 00:09:19,710 --> 00:09:22,829 S1: it just automatically has like this negative valence to it. 165 00:09:22,830 --> 00:09:27,060 S1: And it turned me off. As a fan of his reviews, 166 00:09:27,059 --> 00:09:28,800 S1: but I don't think he was wrong, and I don't 167 00:09:28,800 --> 00:09:31,320 S1: think anyone should be telling him what he can or 168 00:09:31,320 --> 00:09:34,620 S1: cannot do. Dyson's new Clean Trace feature lets you use 169 00:09:34,620 --> 00:09:37,290 S1: AR to see all the spots you missed. Okay, but 170 00:09:37,290 --> 00:09:38,820 S1: what do you have to wear? Do you have to 171 00:09:38,820 --> 00:09:45,209 S1: wear a resume, an Oculus, or an AVP Sequoia's arc product? 172 00:09:45,210 --> 00:09:48,690 S1: This one I haven't messed with yet introduces three unique 173 00:09:48,690 --> 00:09:53,400 S1: archetypes to help startups navigate their way to market success. 174 00:09:53,910 --> 00:09:55,860 S1: I'm going to keep this tab open and you need 175 00:09:55,860 --> 00:09:58,590 S1: to go play with that more. Tesla just dropped the 176 00:09:58,590 --> 00:10:03,120 S1: FSD subscription from I think it was like 250 or 300, 177 00:10:03,120 --> 00:10:05,699 S1: something like that, and now it's 100 bucks a month. Cool. 178 00:10:05,700 --> 00:10:08,010 S1: So I hope mine that I was already paying for 179 00:10:08,010 --> 00:10:10,230 S1: just goes down to 99 naturally. And I don't have 180 00:10:10,230 --> 00:10:14,339 S1: to do something like cancel it and renew. All right, humans, 181 00:10:14,340 --> 00:10:16,530 S1: right after the solar eclipse, a whole bunch of people 182 00:10:16,530 --> 00:10:20,309 S1: were googling my eyes hurt, which, uh, further diminishes my 183 00:10:20,309 --> 00:10:23,100 S1: hope for humanity. It's not so much that they looked 184 00:10:23,100 --> 00:10:26,070 S1: at the sun. It's the two hit combo of them 185 00:10:26,070 --> 00:10:29,580 S1: looking at the sun and then wondering why they couldn't see. 186 00:10:29,580 --> 00:10:32,670 S1: And one of the books that will absolutely change your 187 00:10:32,670 --> 00:10:35,880 S1: mind on like, so much stuff is Everybody Lies. It 188 00:10:35,880 --> 00:10:38,520 S1: talks about the difference between what people say is true 189 00:10:38,520 --> 00:10:42,599 S1: or what people say they believe or like versus what 190 00:10:42,600 --> 00:10:45,780 S1: they actually Google for. And it's just like it reveals 191 00:10:45,780 --> 00:10:49,500 S1: so much interesting stuff. It's just a fantastic book. Job 192 00:10:49,500 --> 00:10:53,189 S1: interviews are pretty terrible at predicting job performance, but what's 193 00:10:53,190 --> 00:10:56,099 S1: the alternative? That's my question, is, like, I see a 194 00:10:56,100 --> 00:10:58,170 S1: lot of things that say it doesn't do a good job, 195 00:10:58,170 --> 00:11:03,000 S1: but if you don't have a better proxy, then so what? 196 00:11:03,000 --> 00:11:05,340 S1: Why are we talking about it? Over half of women 197 00:11:05,340 --> 00:11:10,440 S1: experience sexual harassment at work. Homicides are on a surprising 198 00:11:10,440 --> 00:11:14,820 S1: decline across major cities. I feel like we're this close 199 00:11:14,820 --> 00:11:17,760 S1: and very close to an onion article where far right 200 00:11:17,760 --> 00:11:20,940 S1: people are. Like, we used to have good murder rates, 201 00:11:20,940 --> 00:11:24,900 S1: and now the whole country is like feminized. And the 202 00:11:24,900 --> 00:11:27,830 S1: murder rates keep. Going down. And that's just further evidence 203 00:11:27,830 --> 00:11:30,560 S1: of America's decline. Not trying to make fun of the right. 204 00:11:30,559 --> 00:11:33,980 S1: They're just, uh. I thought that was funny. I recently 205 00:11:33,980 --> 00:11:37,580 S1: re subscribed to onion, so I'm hearing onion jokes everywhere 206 00:11:37,580 --> 00:11:40,220 S1: I look. Now, U.S. steel is about to become part 207 00:11:40,220 --> 00:11:44,270 S1: of Japan's Nippon Steel. I think that's how you pronounce that. 208 00:11:44,720 --> 00:11:48,230 S1: And Harvard is back to SAT and Act scores. It's 209 00:11:48,230 --> 00:11:50,360 S1: almost like they were a good idea in the first place, 210 00:11:50,630 --> 00:11:52,730 S1: which is why we had them in the first place. 211 00:11:52,730 --> 00:11:56,150 S1: And my comment here, I wish the Overton Pendulum had 212 00:11:56,150 --> 00:11:59,540 S1: more than two settings, which is like woke and racist. Like, 213 00:11:59,540 --> 00:12:01,250 S1: can we get a middle ground? Can we get a 214 00:12:01,250 --> 00:12:06,589 S1: middle setting? Compassion, logical, something like secular humanism or something, 215 00:12:06,590 --> 00:12:10,609 S1: or even not secular humanism, just logical humanism getting a 216 00:12:10,610 --> 00:12:14,329 S1: lot more open to some moderate religion? I think it's 217 00:12:14,330 --> 00:12:17,809 S1: got a lot of positive functions. In fact, the research 218 00:12:17,809 --> 00:12:21,320 S1: by Jonathan Haidt on this is pretty compelling, basically showing 219 00:12:21,320 --> 00:12:26,030 S1: that kids of religious parents are doing so much better 220 00:12:26,030 --> 00:12:28,670 S1: in mental health. And yeah, I'm trying to track all 221 00:12:28,670 --> 00:12:31,400 S1: these threads and find a way to unify all this. 222 00:12:31,400 --> 00:12:33,800 S1: Dumb phones are making comeback. You basically get to stay 223 00:12:33,800 --> 00:12:36,439 S1: in contact but not be distracted. It's going to be 224 00:12:36,440 --> 00:12:39,590 S1: a Dungeons and Dragons show at Madison Square Garden, which 225 00:12:39,590 --> 00:12:42,590 S1: is supposedly going to sell out. That does not surprise me. 226 00:12:42,620 --> 00:12:45,319 S1: I think this whole thing, which, by the way, tonight 227 00:12:45,320 --> 00:12:46,970 S1: is a gaming night and I have to leave soon, 228 00:12:46,970 --> 00:12:49,400 S1: so I guess I'm going to speed up. Yeah, I 229 00:12:49,400 --> 00:12:52,760 S1: played D&D with my friends. It's fantastic. I'm so glad 230 00:12:52,760 --> 00:12:56,000 S1: to see D&D rising up. Actually, it doesn't really matter 231 00:12:56,000 --> 00:12:59,480 S1: if it's exactly D&D just getting together, doing some sort 232 00:12:59,480 --> 00:13:04,490 S1: of intellectually, creatively challenging role playing game I think is 233 00:13:04,490 --> 00:13:08,450 S1: super valuable. What top performers do and how to deal 234 00:13:08,450 --> 00:13:13,189 S1: with their critics and detractors. Ernest Hemingway transform from a 235 00:13:13,190 --> 00:13:16,400 S1: humble journalist to a celebrity, losing himself to the fame 236 00:13:16,400 --> 00:13:20,450 S1: and power that came with success. Abraham Lincoln was shaped 237 00:13:20,450 --> 00:13:24,260 S1: by Aesop's Fables more than any other book, including the Bible. 238 00:13:24,260 --> 00:13:27,559 S1: That was a really interesting read and for my idea 239 00:13:27,559 --> 00:13:30,800 S1: and analysis this week, I heard somewhere on a podcast 240 00:13:30,800 --> 00:13:34,100 S1: it might have been Alex Hormones. It might have been like, no, 241 00:13:34,100 --> 00:13:37,130 S1: it was so many on Chris Williams and maybe the 242 00:13:37,130 --> 00:13:40,610 S1: gamer doctor or yeah, maybe. I'm not sure exactly who 243 00:13:40,610 --> 00:13:43,099 S1: it was, but they didn't exactly say it this way. 244 00:13:43,100 --> 00:13:45,710 S1: But it was essentially like, if you have an idea 245 00:13:45,710 --> 00:13:48,679 S1: that's really exciting to you, you basically don't want to 246 00:13:48,679 --> 00:13:51,440 S1: share that or you don't want to get any wins 247 00:13:51,440 --> 00:13:55,010 S1: from the idea that aren't wins from actually doing the work. 248 00:13:55,010 --> 00:13:58,610 S1: And so I'm expanding that out here to basically say 249 00:13:58,610 --> 00:14:01,730 S1: energy you have for the idea is limited and precious, 250 00:14:01,730 --> 00:14:04,520 S1: and even telling someone about it, I think they did say, 251 00:14:04,520 --> 00:14:06,980 S1: actually say this. I think the guests might have said this. 252 00:14:06,980 --> 00:14:10,250 S1: Even telling somebody about it removes some of its power. 253 00:14:10,250 --> 00:14:12,950 S1: So instead of telling people about it, you should work 254 00:14:12,950 --> 00:14:16,460 S1: on it immediately. Use that power to go right into 255 00:14:16,460 --> 00:14:19,310 S1: the actual work. Otherwise, you might end up not doing 256 00:14:19,310 --> 00:14:23,570 S1: anything because your your body, like evolution, thinks that you 257 00:14:23,570 --> 00:14:26,630 S1: did the work because you talked about it. It's like 258 00:14:26,630 --> 00:14:28,880 S1: it's a fake win and you don't want to do that. 259 00:14:28,880 --> 00:14:31,400 S1: This one I'm probably going to do as a standalone essay. 260 00:14:31,610 --> 00:14:33,710 S1: Reason to worry, reason to build. So I'm going to 261 00:14:33,710 --> 00:14:36,380 S1: skip that one. And we got a colab notebook here 262 00:14:36,380 --> 00:14:40,940 S1: that uses cloud to turn images into magic cards. Karpathy 263 00:14:41,150 --> 00:14:45,890 S1: launched LMK shell history. This thing is really cool. Adder 264 00:14:45,890 --> 00:14:49,430 S1: turns your terminal into a pair programming session with AI, 265 00:14:49,460 --> 00:14:54,740 S1: DNS viz cohere Rerank three boosts enterprise search and drag systems. 266 00:14:54,740 --> 00:14:58,700 S1: Terminal tweaks UDL turns your text prompts into music. This 267 00:14:58,700 --> 00:15:03,320 S1: thing is insane. Collection of content ideas that have resonated 268 00:15:03,320 --> 00:15:08,120 S1: on TikTok. Pretty cool list for creators. Fire crawls Wdev 269 00:15:08,150 --> 00:15:13,040 S1: turns entire websites into data sets for LMS. Zeek turns 270 00:15:13,040 --> 00:15:17,570 S1: your command line into a powerful plaintext Zettelkasten or personal 271 00:15:17,570 --> 00:15:20,840 S1: wiki keeper. Kind of an alternative to some of the 272 00:15:20,840 --> 00:15:23,840 S1: other ones like obsidian. And yeah, recommendation of the week 273 00:15:23,840 --> 00:15:25,700 S1: is actually to read that piece, which I'm going to 274 00:15:25,700 --> 00:15:28,369 S1: do a standalone episode on, and the aphorism of the 275 00:15:28,370 --> 00:15:32,090 S1: week being deeply loved by someone gives you strength, but 276 00:15:32,090 --> 00:15:36,410 S1: loving someone deeply gives you courage. Being deeply loved by 277 00:15:36,410 --> 00:15:40,220 S1: someone gives you strength, while loving someone deeply gives you courage. 278 00:15:40,220 --> 00:15:44,479 S1: Lao Tzu Unsupervised Learning is produced and edited by Daniel 279 00:15:44,480 --> 00:15:49,820 S1: Missler on a 1987 AI microphone using Hindenburg. Intro and outro. 280 00:15:49,820 --> 00:15:53,150 S1: Music is by zombie with a Y, and to get 281 00:15:53,150 --> 00:15:55,250 S1: the text and links from this episode, sign up for 282 00:15:55,250 --> 00:16:00,860 S1: the newsletter version of the show at Daniel missler.com/newsletter. We'll 283 00:16:00,860 --> 00:16:01,580 S1: see you next time.