WEBVTT - UL NO. 429: Build Your Career Around Problems 0:00:01.230 --> 0:00:04.500 Welcome to Unsupervised Learning, a security, AI and meaning focused 0:00:04.500 --> 0:00:07.380 podcast that looks at how best to thrive as humans 0:00:07.380 --> 0:00:11.610 in a post AI world. It combines original ideas, analysis, 0:00:11.610 --> 0:00:14.820 and mental models to bring not just the news, but 0:00:14.820 --> 0:00:21.939 why it matters and how to respond. All right. Welcome 0:00:21.940 --> 0:00:25.150 to supervised learning. This is Daniel Missler. All right. In 0:00:25.150 --> 0:00:28.630 this one. What have we got here. This is episode 429. 0:00:28.780 --> 0:00:34.450 So llama three came out last week and it was 0:00:34.450 --> 0:00:38.230 super impressive. I'm getting more and more use out of it, actually. 0:00:38.229 --> 0:00:41.379 I'm finding it's doing a lot of stuff as well 0:00:41.380 --> 0:00:46.540 as GPT four, and sometimes it just completely fails. And 0:00:46.540 --> 0:00:49.600 it's worse than like GPT three. So I think that's 0:00:49.600 --> 0:00:52.210 the case with all these different models. Like, you never know, 0:00:52.210 --> 0:00:54.280 like how good it's going to be at certain things 0:00:54.280 --> 0:00:56.410 and how bad it's going to be at other things. 0:00:56.410 --> 0:00:59.410 And what I really hope we get before too long 0:00:59.410 --> 0:01:03.790 is like this really robust testing framework. So we know 0:01:03.790 --> 0:01:06.580 what things are good at and what they're not good at. 0:01:06.580 --> 0:01:08.980 And we could just have this continuous testing because a 0:01:08.980 --> 0:01:12.490 lot of the benchmarks that are used to test, you know, 0:01:12.490 --> 0:01:15.400 to show the quality of models, I feel like people 0:01:15.400 --> 0:01:18.250 are actually building their models to score good on them, 0:01:18.459 --> 0:01:24.640 and they're pretty useless. There's also this cool bolt AI application, 0:01:24.640 --> 0:01:27.190 which this link is actually bad, I think. I think 0:01:27.190 --> 0:01:31.360 this is the wrong thing, but if we do bolt AI, 0:01:31.630 --> 0:01:35.260 this should actually pull us up into it. Yeah, this 0:01:35.260 --> 0:01:37.720 is the one. So this is a pretty cool interface 0:01:37.720 --> 0:01:41.920 for actually dealing with working with LMS. It looks almost 0:01:41.920 --> 0:01:45.910 like ChatGPT, but it's a standalone app and you could 0:01:45.910 --> 0:01:48.970 use all your different models, right? So you got like 0:01:49.000 --> 0:01:52.240 you could use GPT four, you could use local models, 0:01:52.390 --> 0:01:54.310 you could put in your own model. This is just 0:01:54.310 --> 0:01:58.000 really good. And see here you got different commands AI 0:01:58.000 --> 0:02:01.360 and line. Yeah, it's just really powerful. I like this 0:02:01.360 --> 0:02:05.080 little lap here. So that's one one nice app to 0:02:05.080 --> 0:02:08.380 take a look at. Again it's called bolt AI. That 0:02:08.380 --> 0:02:11.320 link I have here, like I said it's not the 0:02:11.320 --> 0:02:16.029 correct link but whatever. Let's see here. Yeah llama three 0:02:16.030 --> 0:02:18.550 really impressive. At the seven B level there's also an 0:02:18.550 --> 0:02:22.089 eight B version which I don't really mess with. I 0:02:22.090 --> 0:02:24.820 only messed with the 70 B version. By the way, 0:02:24.820 --> 0:02:28.090 the 70 B version actually all the different versions of 0:02:28.090 --> 0:02:31.480 llama three are also in fabric, so you can use 0:02:31.480 --> 0:02:35.470 fabric with local models of course, and that includes llama three. 0:02:35.470 --> 0:02:38.650 And one thing I've noticed is that it's way less restricted. 0:02:38.650 --> 0:02:41.170 It actually will tell you a lot of stuff that 0:02:41.169 --> 0:02:43.930 llama two would not, and a lot of the other 0:02:43.930 --> 0:02:46.750 pinnacle models will not talk to you about at all. 0:02:46.750 --> 0:02:51.130 It's not quite uncensored, but it is a lot more 0:02:51.130 --> 0:02:54.220 similar to an uncensored, I would say 50% there or something. 0:02:54.220 --> 0:02:58.329 And of course, there's already dolphin versions of llama three 0:02:58.330 --> 0:03:02.919 that are uncensored. So, um, not just dolphin, but other versions. 0:03:02.919 --> 0:03:06.250 So definitely take a look at that if you need that. 0:03:06.250 --> 0:03:08.710 And one thing that's really interesting to me is like 0:03:08.710 --> 0:03:12.489 how good llama three is getting or local models in general, 0:03:12.490 --> 0:03:16.420 because Apple is now releasing open source models. Snowflake just 0:03:16.419 --> 0:03:19.540 released an open source model. And here's one thing I 0:03:19.540 --> 0:03:22.960 really think is interesting about this. If you think about 0:03:22.960 --> 0:03:26.290 the daily tasks that humans go through every day with like, oh, 0:03:26.290 --> 0:03:29.170 find me a car, get me a grocery list, give 0:03:29.169 --> 0:03:31.690 me a really good workout, play the best song for 0:03:31.690 --> 0:03:34.210 right now. You know, get a present for my partner 0:03:34.210 --> 0:03:37.690 because it's our anniversary or whatever. All these tasks that 0:03:37.690 --> 0:03:41.770 are like everyday things they have, this is a hypothesis, 0:03:41.770 --> 0:03:44.560 but they have a limit to how good they need 0:03:44.560 --> 0:03:48.610 to be. Right? So if you exceed that limit of like, 0:03:48.610 --> 0:03:51.820 they just get the most amazing flowers every single time 0:03:51.820 --> 0:03:55.030 they pick the best song or music every single time 0:03:55.030 --> 0:03:59.290 and it's really good. Then if that was GPT five 0:03:59.290 --> 0:04:02.830 that made it that good, then GPT seven isn't really needed. 0:04:02.830 --> 0:04:06.340 And if local models exceed what GPT five could do, 0:04:06.340 --> 0:04:08.350 which of course I think they will within a year 0:04:08.350 --> 0:04:11.170 or two or whatever, however long that takes. The point 0:04:11.170 --> 0:04:16.180 is the models keep getting better, including the local models, 0:04:16.300 --> 0:04:20.290 and that a task that a human wants to have 0:04:20.290 --> 0:04:25.090 done kind of hits a level of diminishing return. And 0:04:25.089 --> 0:04:27.760 I think for a lot of human tasks, we're about 0:04:27.760 --> 0:04:31.419 to hit that level with our pinnacle models. And then 0:04:31.420 --> 0:04:34.630 you have your local models that are like lagging behind it. 0:04:34.630 --> 0:04:39.580 Once that exceeds it, you have local models everywhere that 0:04:39.580 --> 0:04:44.770 could do most tasks. So for example, home automation, home management, 0:04:44.770 --> 0:04:48.339 automation of tasks around the home. Like I said, groceries, 0:04:48.339 --> 0:04:52.600 everyday things, those things can all be done almost instantly, 0:04:52.600 --> 0:04:56.920 basically instantly and locally on the device because the local 0:04:56.920 --> 0:04:59.919 chipset will be good enough, or there'll be like a 0:04:59.920 --> 0:05:02.980 local chipset in the house somewhere that you outsource to 0:05:02.980 --> 0:05:07.090 or whatever, but you don't need pinnacle models for those. Now, 0:05:07.089 --> 0:05:08.950 it could be that some of those tasks would actually 0:05:09.250 --> 0:05:12.909 benefit from a much bigger model, but I think so 0:05:12.910 --> 0:05:16.779 many of them actually won't. You won't need them. And 0:05:16.779 --> 0:05:20.450 that's really exciting because when local models can. You most 0:05:20.450 --> 0:05:23.630 of what we need. That's an insane world. And my 0:05:23.630 --> 0:05:26.270 buddy Joseph Thacker keeps talking about this thing. It's like, 0:05:26.270 --> 0:05:28.580 why not just have it in everything? And he keeps 0:05:28.580 --> 0:05:30.890 talking about this cool idea, which I want to touch 0:05:30.890 --> 0:05:33.800 on too, which is this concept of an oracle, which 0:05:33.800 --> 0:05:35.719 I've always wondered about too. It's like if you went 0:05:35.720 --> 0:05:38.510 back in time, how useful would you actually be? Could 0:05:38.510 --> 0:05:42.320 you actually describe, you know, chemistry or, you know, satellites 0:05:42.320 --> 0:05:45.560 or combustion engines or any of this stuff? It'd be 0:05:45.560 --> 0:05:47.780 amazing to have like a little soccer ball that talks 0:05:47.779 --> 0:05:50.719 to you. And inside of the soccer ball is this model, 0:05:50.720 --> 0:05:53.570 and you could somehow get power into it or whatever. 0:05:53.570 --> 0:05:57.620 But this one little thing, this tiny little device has 0:05:57.620 --> 0:06:01.070 something like Lama three on it, which is a large 0:06:01.070 --> 0:06:05.570 percentage of all human knowledge ever. And it's local, doesn't 0:06:05.570 --> 0:06:07.700 need to call out to any cloud. And you can 0:06:07.700 --> 0:06:10.339 put that in a soccer ball or a park bench 0:06:10.339 --> 0:06:13.910 or anything like that. It's insane to me that you 0:06:13.910 --> 0:06:18.470 could basically bootstrap humanity to some degree with something that's 0:06:18.470 --> 0:06:20.960 completely local on a chip. All right, more stuff going 0:06:20.960 --> 0:06:24.650 on prepping for RSA. Definitely come by and say hi. 0:06:24.650 --> 0:06:28.070 I might be crazy busy or distracted or socially awkward 0:06:28.070 --> 0:06:31.520 at the time. Whatever. Just come say hi. What have 0:06:31.520 --> 0:06:35.779 we got here? We got hugs, waves, finger guns, always appropriate. 0:06:35.779 --> 0:06:39.890 Fist bumps, whatever. Whatever you got I should be comfortable with. 0:06:39.890 --> 0:06:42.950 Last few talks have gone really well. There's something to 0:06:42.950 --> 0:06:46.640 be said for speaking, to share an idea rather than 0:06:46.640 --> 0:06:50.960 trying to, like, do a presentation. I've crossed over for 0:06:50.960 --> 0:06:53.570 most of my talks. At least I'm trying to cross 0:06:53.570 --> 0:06:57.169 over completely to this thing where I'm just I have 0:06:57.170 --> 0:06:59.690 this idea that I want to share, and the goal 0:06:59.690 --> 0:07:02.240 of the talk has nothing to do with the talk 0:07:02.240 --> 0:07:04.969 or the slides. It has to do with how many 0:07:04.970 --> 0:07:08.659 people actually get converted over to this way of thinking 0:07:08.660 --> 0:07:12.020 at the end. And when you are trying to do that, 0:07:12.020 --> 0:07:15.800 it makes the slides not matter. It makes like if 0:07:15.800 --> 0:07:19.430 whether you stumble, your words not matter. Nervousness kind of 0:07:19.430 --> 0:07:22.550 goes away. And I've got a couple of talks already 0:07:22.580 --> 0:07:25.700 or essays about how to make that happen. Anyway, there's 0:07:25.700 --> 0:07:28.370 a breathing technique for like the tactical version, and then 0:07:28.370 --> 0:07:32.420 there's like this concept for getting rid of the overall 0:07:32.420 --> 0:07:36.050 problem of stage fright. If you simply have something to 0:07:36.050 --> 0:07:39.320 share and you are sharing it, that is not a 0:07:39.320 --> 0:07:42.739 stage fright issue. It's very strange. It's like, what are 0:07:42.740 --> 0:07:45.860 you actually looking at? If you go up on stage 0:07:45.860 --> 0:07:49.790 and you are thinking about your talk in terms of 0:07:49.790 --> 0:07:54.290 like the slides, the slide notes, where are the cameras, 0:07:54.290 --> 0:07:56.929 where's the monitor, where are the people, who are they 0:07:56.930 --> 0:08:00.680 looking at? If you pull into yourself like that, you're 0:08:00.680 --> 0:08:03.080 going to get stressed. You're going to get nervous. The 0:08:03.080 --> 0:08:07.400 way to fix that is to go outside of yourself. 0:08:07.400 --> 0:08:10.550 What's bigger than yourself? What? What is outside of yourself? 0:08:10.550 --> 0:08:14.450 The idea is the thing that matters. You focus on 0:08:14.450 --> 0:08:17.150 that and suddenly you're very calm because what you want 0:08:17.150 --> 0:08:19.550 to do is you want to convert your excitement away 0:08:19.550 --> 0:08:24.380 from anxiousness and into excitement around sharing the idea. And 0:08:24.380 --> 0:08:26.960 you could view this as a hack, or you could 0:08:26.960 --> 0:08:30.800 view it as just a better way of thinking about presenting. Period. 0:08:30.800 --> 0:08:33.260 And I think it is both, but I frame it 0:08:33.260 --> 0:08:35.660 as the second one. Updated the intro to the newsletter 0:08:35.660 --> 0:08:40.190 focusing on human to 2.0 transitioning to 3.0. Let me 0:08:40.190 --> 0:08:43.040 know what you think of that. I put discovery actually 0:08:43.040 --> 0:08:46.880 into each section instead of having a dedicated discovery this time. 0:08:46.880 --> 0:08:51.469 And if you have time, you've got to go listen 0:08:51.470 --> 0:08:55.340 to this conversation between Tyler Cohen and Peter Thiel. I'm 0:08:55.340 --> 0:08:58.580 not a Peter Thiel fan in general. I didn't like 0:08:58.580 --> 0:09:01.429 a lot of his political views. While back they had 0:09:01.429 --> 0:09:05.330 this wide ranging conversation from like Star Wars to the 0:09:05.330 --> 0:09:11.060 Antichrist to the Bible to Shakespeare to economics, political theory, 0:09:11.179 --> 0:09:15.589 just theology. It was an insane conversation. What I like 0:09:15.590 --> 0:09:17.510 most about it, because I've shown a few people and 0:09:17.510 --> 0:09:19.550 they were like, yeah, I don't know why you liked 0:09:19.550 --> 0:09:22.040 it so much. And a few people said, look, they 0:09:22.040 --> 0:09:23.750 just talked about a whole bunch of stuff that I 0:09:23.750 --> 0:09:27.290 didn't understand. Here's my trick for you, and this is 0:09:27.290 --> 0:09:29.750 why I was so excited about this talk. I don't 0:09:29.750 --> 0:09:34.910 often find anyone lately because I read so much who 0:09:34.910 --> 0:09:37.550 says a string of words and actually just a string 0:09:37.550 --> 0:09:42.199 of answers to questions. And in those answers I'm like, 0:09:42.200 --> 0:09:46.280 what does that mean? Like Strachan, for example, Strachan was 0:09:46.370 --> 0:09:48.770 Stross was one of the people that he was talking 0:09:48.770 --> 0:09:51.830 about in this thing. And I'm like, I don't know 0:09:51.830 --> 0:09:54.950 Stross yet, I don't what does that mean? So I 0:09:54.950 --> 0:09:57.740 ended up having to go in Google and look up 0:09:57.740 --> 0:10:00.830 all these different things. Google more like talk to my 0:10:00.830 --> 0:10:03.890 Da and have her explain the stuff to me. And 0:10:03.890 --> 0:10:07.309 then I also put a set of books into audible 0:10:07.309 --> 0:10:09.590 as well. So now I'm going to go read these 0:10:09.590 --> 0:10:13.310 canonical books that came from these thinkers. So I don't 0:10:13.309 --> 0:10:17.630 accept when somebody talks over my head because I essentially say, well, 0:10:17.630 --> 0:10:19.860 you can only do that once because. Now I'm going 0:10:19.860 --> 0:10:22.650 to research all that stuff. And first of all, figure 0:10:22.650 --> 0:10:25.500 out what you were actually saying just so it doesn't 0:10:25.500 --> 0:10:28.800 appear smart when it's actually not smart. But what I 0:10:28.800 --> 0:10:32.069 came away with after deciphering all the stuff and learning 0:10:32.070 --> 0:10:34.650 like a quick primer on this stuff, is I came 0:10:34.650 --> 0:10:38.429 away realizing that Peter Thiel knows a lot of stuff 0:10:38.429 --> 0:10:41.280 and what I'm really attracted to. And I don't know 0:10:41.280 --> 0:10:43.980 if this is something where he changed or where I changed, 0:10:43.980 --> 0:10:48.420 or maybe both or something, but he doesn't seem bound 0:10:48.420 --> 0:10:52.350 to any sort of side. He seems bound to principles 0:10:52.350 --> 0:10:54.810 of like what he wants to see happen. And I 0:10:54.809 --> 0:10:57.300 think that is really powerful. And I talked about that. 0:10:57.300 --> 0:11:00.090 I think in a previous show. I can't remember where 0:11:00.090 --> 0:11:03.359 I talked about that recently, but I think it's really oh, 0:11:03.360 --> 0:11:07.410 it's actually in this newsletter actually. So I would say 0:11:07.410 --> 0:11:10.140 definitely check this out no matter what. Check it out. 0:11:10.140 --> 0:11:12.750 You might not like it, but listen to what I 0:11:12.750 --> 0:11:14.880 was saying before and see if that helps you like 0:11:14.880 --> 0:11:17.250 it more. All right, so I wrote a new essay 0:11:17.250 --> 0:11:20.969 on the old paradigm of planning a career no longer working, 0:11:20.970 --> 0:11:25.620 and basically saying you should plan your career around problems instead. 0:11:25.620 --> 0:11:30.030 And I actually have released a standalone video on this 0:11:30.030 --> 0:11:31.860 as well, which is also going to come out as 0:11:31.860 --> 0:11:35.160 a podcast, so you can check that out on YouTube. 0:11:35.280 --> 0:11:38.340 It should already be live. All right. Security, the House 0:11:38.340 --> 0:11:40.860 just passed a bill making it illegal for the government 0:11:40.860 --> 0:11:44.610 to buy your data without a warrant. Calling it the 0:11:44.610 --> 0:11:47.910 Fourth amendment is not for sale. This is really cool. 0:11:48.150 --> 0:11:51.360 That's bipartisan. It seems like it would. It would have 0:11:51.360 --> 0:11:53.790 had to have been. Not sure about that, but I'm 0:11:53.790 --> 0:11:58.410 pretty sure that's right. Sandworm. Notorious Russian hacking group linked 0:11:58.410 --> 0:12:02.820 to cyber attack on a water facility in Texas. The 0:12:02.820 --> 0:12:06.210 ban on TikTok looks like it's actually going forward. That 0:12:06.210 --> 0:12:09.780 is pretty cool. Biden signed it. So that's bipartisan as well. 0:12:09.780 --> 0:12:13.590 A little bit of light in the darkness there might 0:12:13.590 --> 0:12:17.010 have got compromised among a bunch of other companies by 0:12:17.010 --> 0:12:20.280 another China based attacker, and a flaw and putty lets 0:12:20.280 --> 0:12:24.120 people look at a whole bunch of different keys or 0:12:24.120 --> 0:12:28.290 cryptographic signatures. Basically figure out your private keys, which is 0:12:28.290 --> 0:12:33.900 no bueno. FBI Director Christopher Wray says that China is 0:12:33.900 --> 0:12:39.300 basically shifting to figuring out how to attack US critical infrastructure, 0:12:39.300 --> 0:12:44.340 possibly to go in concert with a move in 2027. 0:12:44.340 --> 0:12:46.800 So it would be like, you know, do these cyber 0:12:46.800 --> 0:12:49.830 or infrastructure things at the same time to kind of 0:12:49.830 --> 0:12:55.260 keep the enemy off balance? Marlinspike moxie basically says he's 0:12:55.260 --> 0:12:58.230 no longer affiliated with signal. I've not been a huge 0:12:58.230 --> 0:12:59.939 fan of signal. I don't like the fact that you 0:12:59.940 --> 0:13:03.180 need to restart it every 45 minutes. I don't like 0:13:03.179 --> 0:13:06.689 being outside of Apple Messages. I prefer to have as 0:13:06.690 --> 0:13:10.530 few messaging apps as possible because I'm on Apple. That 0:13:10.530 --> 0:13:13.439 tends to be my favorite. Not a fan of WhatsApp, 0:13:13.440 --> 0:13:15.900 not a fan of telegram, so I tend to just 0:13:15.900 --> 0:13:19.350 prefer messages. I was okay with signal a little bit 0:13:19.350 --> 0:13:23.160 because it was moxie, but now moxie is out. I'm 0:13:23.160 --> 0:13:25.350 going to try to get people to talk to me 0:13:25.350 --> 0:13:30.840 on messages again. Sacramento International Airport had to stop flights 0:13:30.840 --> 0:13:36.480 because somebody deliberately cut the AT&T internet cable that provided 0:13:36.480 --> 0:13:40.890 internet to the airport. Tale scale now has SSH. DeepMind's 0:13:40.890 --> 0:13:44.460 boss says Google is going to outspend everyone in AI. 0:13:44.490 --> 0:13:47.910 You know, it's not about outspending guys. You've been able 0:13:47.910 --> 0:13:51.060 to outspend and have been outspending people for a long time. 0:13:51.480 --> 0:13:55.860 It's you've lost the ability to ship good products because 0:13:55.860 --> 0:13:58.590 you don't have vision as your primary guide. You have 0:13:58.590 --> 0:14:02.100 basically engineers making stuff and throw them at the walls. 0:14:02.100 --> 0:14:05.850 And then like three years later, if it launches, you 0:14:05.850 --> 0:14:10.170 put it in the graveyard. It's just complete joke. I 0:14:10.170 --> 0:14:13.590 think you need new leadership, honestly, is what you need. 0:14:13.710 --> 0:14:17.819 And Stanford released a quality report on AI models and 0:14:17.820 --> 0:14:21.479 we did a micro summary from fabric. So basically I 0:14:21.480 --> 0:14:26.130 surpasses humans in specific tasks, not in complex reasoning. I 0:14:26.130 --> 0:14:30.720 think that'll probably change soon. US is ahead of model development. 0:14:30.720 --> 0:14:34.170 And yeah, pretty good breakdown here. Interesting argument about how 0:14:34.170 --> 0:14:39.030 search engines, especially Google, can actually move elections. Google fired 0:14:39.030 --> 0:14:42.660 28 employees for protesting and they actually fired some more 0:14:42.660 --> 0:14:46.680 people as well recently. More recently, Google has merged its 0:14:46.680 --> 0:14:51.210 Android and hardware teams. Netflix is using FreeBSD current for 0:14:51.210 --> 0:14:53.790 its edge network. That is surprising to me. I would 0:14:53.790 --> 0:14:57.990 have thought those packages would be old, not well maintained, 0:14:58.380 --> 0:15:03.570 and definitely behind. Yeah, so that's surprising in a good way. 0:15:03.720 --> 0:15:06.300 Reddit showing up a lot more in Google results. They 0:15:06.300 --> 0:15:09.180 actually have some sort of contract together. I think they 0:15:09.270 --> 0:15:13.560 sent Reddit like $30 million or something. So something around there. 0:15:13.680 --> 0:15:16.260 Apple's Airplay is starting to come to hotel rooms. Hopefully 0:15:16.260 --> 0:15:21.340 Marriott's soon tiny S.A. is a budget. Friendly spectrum analyzer 0:15:21.790 --> 0:15:25.690 programming is mostly thinking this it is. This was a 0:15:25.690 --> 0:15:30.310 good essay, I enjoyed it. Broad introduction to AWS log 0:15:30.310 --> 0:15:35.650 sources and events and Gen Z is outperforming previous generations 0:15:35.650 --> 0:15:40.120 at their age. This was also surprising to me. This 0:15:40.120 --> 0:15:44.380 article says that societal decline mirrors the death spiral seen 0:15:44.380 --> 0:15:49.220 in ants. Why everything is becoming a game. This blog 0:15:49.220 --> 0:15:54.320 by the way. Really cool. I love this blog. Gurinder. Yeah, 0:15:54.320 --> 0:15:58.220 this guy is awesome. Definitely want to subscribe. Their study 0:15:58.220 --> 0:16:00.680 found that jobs that require you to think a lot 0:16:00.680 --> 0:16:04.640 or protective against Alzheimer's. I think that needs an H. 0:16:04.790 --> 0:16:08.360 Bayer is doing an experiment where they remove most of 0:16:08.360 --> 0:16:14.060 middle management and let basically like 100,000 employees mostly self-organize. 0:16:14.210 --> 0:16:18.110 And they're hoping to save. I think that was 12 billion, 0:16:18.110 --> 0:16:22.490 not 2 billion. That might be a typo. Unfortunate. I 0:16:22.490 --> 0:16:26.750 think it was $12 billion. I think this is really cool. 0:16:26.750 --> 0:16:29.960 I love the idea of 100,000 people with all these 0:16:29.960 --> 0:16:33.920 layers of middle management, where it's just a giant waste machine. 0:16:34.040 --> 0:16:37.730 And who better to figure that out then? AI who 0:16:37.730 --> 0:16:41.510 figures out, okay, who's doing what, what functions actually need 0:16:41.510 --> 0:16:45.950 to get done? What who whose teams are structured in 0:16:45.950 --> 0:16:48.890 what way? How do they communicate with each other? What 0:16:48.890 --> 0:16:51.860 are the work products that are actually being produced? What 0:16:51.860 --> 0:16:55.670 are those pipelines look like for work being produced? You 0:16:55.670 --> 0:16:58.730 overlay that on the org structure, and you look at 0:16:58.730 --> 0:17:00.980 the budget and everything, and you look at the outcomes 0:17:00.980 --> 0:17:04.130 and actually how fast things move. And AI is going 0:17:04.130 --> 0:17:06.740 to look at it and just be like a horrible mess. 0:17:07.130 --> 0:17:09.740 Why do you have all these managers? What are they 0:17:09.740 --> 0:17:13.340 actually doing now? Humans have already known that we didn't 0:17:13.340 --> 0:17:16.040 need all those layers of middle management like this. It's 0:17:16.040 --> 0:17:20.030 so common. It's just a meme. It's common knowledge. But 0:17:20.030 --> 0:17:21.530 AI is going to be able to look at it 0:17:21.530 --> 0:17:24.230 for any given company and just be like, here's why 0:17:24.230 --> 0:17:28.220 this is a giant waste. You know, Chris isn't doing 0:17:28.220 --> 0:17:31.430 anything right. And he hands it to John, who's not 0:17:31.430 --> 0:17:34.070 doing anything either. And what is Mary doing? Like this 0:17:34.070 --> 0:17:38.840 makes no sense whatsoever. And it's going to recommend look, 0:17:38.869 --> 0:17:41.179 you need your experts and you're going to need you 0:17:41.180 --> 0:17:44.840 need your high level thinkers and you need very little 0:17:44.840 --> 0:17:48.169 middle management, which really is probably going to be AI 0:17:48.170 --> 0:17:52.400 agents very soon. And of course, at some point you 0:17:52.400 --> 0:17:56.060 can even use AI for the top and bottom pieces 0:17:56.060 --> 0:17:59.060 as well. But first they're going to come for those 0:17:59.060 --> 0:18:01.400 middle managers. And I think it's a good thing the 0:18:01.400 --> 0:18:06.139 term brainwashing morphed into a blanket term for unconventional behavior. 0:18:06.140 --> 0:18:09.830 And this came from MKUltra seeing a lot of stuff 0:18:09.830 --> 0:18:13.609 talking about MKUltra. I wonder what the reason for that is. Okay, 0:18:13.609 --> 0:18:18.530 ideas and analysis. All right. I like I said, I 0:18:18.530 --> 0:18:21.260 talked about Peter Thiel. I'm not going to read this 0:18:21.260 --> 0:18:23.240 whole thing. This is a bit of an essay here. 0:18:23.240 --> 0:18:27.050 I think I might turn this into a standalone essay. Okay. 0:18:27.050 --> 0:18:32.060 Let's see here. Okay. Recommendation of the week. Establish your 0:18:32.060 --> 0:18:35.210 ground truth in terms of morality and the society that 0:18:35.210 --> 0:18:37.100 you want to live in. And this is going to 0:18:37.100 --> 0:18:40.879 the essay above. Okay. Establish your ground truth in terms 0:18:40.880 --> 0:18:43.550 of morality and the society that you want to live in. 0:18:43.550 --> 0:18:47.150 Lock that in perfectly. You say, look, this is how 0:18:47.150 --> 0:18:52.280 I think the world should work and that is your morality, okay? 0:18:52.280 --> 0:18:55.790 And it's got no labels on it. Left, right, progressive, conservative. 0:18:55.790 --> 0:18:58.670 Nobody cares. You don't put anything on that. You just say, 0:18:58.670 --> 0:19:02.240 here's how I think the world should work. Then you say, okay, 0:19:02.240 --> 0:19:05.240 I'm going to listen to any ideas, right? I'm curious 0:19:05.240 --> 0:19:09.500 how those ideas or those structures or those religions or whatever. 0:19:09.920 --> 0:19:12.649 How would those affect this world that I want to 0:19:12.650 --> 0:19:16.940 see come to pass? And then you don't discard ideas 0:19:16.940 --> 0:19:19.969 just because you disagree with something else that they say 0:19:19.970 --> 0:19:23.570 somewhere else. So if people have like this idea, a 0:19:23.570 --> 0:19:27.800 constellation or belief constellation, you are free to pick and 0:19:27.800 --> 0:19:31.909 choose for where they they are smart in your mind, 0:19:31.910 --> 0:19:34.580 at least where they're smart about things, where they're dumb 0:19:34.580 --> 0:19:38.629 about things. I get so many comments where people are like, hey, listen, 0:19:38.630 --> 0:19:42.500 I absolutely love your stuff. I don't agree with everything, 0:19:42.500 --> 0:19:45.320 but I agree with you on most things, and I 0:19:45.320 --> 0:19:47.870 kind of feel like I'm happy when I hear that 0:19:47.960 --> 0:19:52.160 because I feel like they're parsing everything. They're deciding because 0:19:52.160 --> 0:19:55.610 they've already done number one. They've decided what they believe. 0:19:55.760 --> 0:19:58.580 And I've had other people also say that over time 0:19:58.580 --> 0:20:01.820 they came to believe something that I also believed, which 0:20:01.820 --> 0:20:05.030 is cool. And sometimes I'll switch and, you know, believe 0:20:05.030 --> 0:20:07.790 what they believe. The point is, you don't have to 0:20:07.790 --> 0:20:11.989 accept everything a person says to pull wisdom out of 0:20:11.990 --> 0:20:15.109 what they say, right? Maybe they're only good at making 0:20:15.109 --> 0:20:19.100 the best, you know, Reuben sandwiches. And maybe that's all 0:20:19.100 --> 0:20:21.680 I get from them, because the person's name is Tucker 0:20:21.680 --> 0:20:25.939 Carlson and they don't know anything other than sandwiches. That's fine, 0:20:25.940 --> 0:20:28.910 I will listen. I probably won't spend time like trying 0:20:28.910 --> 0:20:31.070 to seek out that content, but it doesn't matter. That's 0:20:31.070 --> 0:20:34.280 the point of this. Feel free to label people as 0:20:34.280 --> 0:20:37.639 bad overall or stupid, but realize it doesn't mean they're 0:20:37.640 --> 0:20:41.210 wrong about everything. So again, same concept. Be willing to 0:20:41.210 --> 0:20:45.409 take truth from anywhere. And basically you want to regularly 0:20:45.410 --> 0:20:49.280 revisit your number one, which. Is your basis of how 0:20:49.280 --> 0:20:54.260 you think the world should work and then regularly re-evaluate, right? 0:20:54.260 --> 0:20:58.610 Pull in lots of different opinions and see if it changes. 0:20:58.609 --> 0:21:02.030 Number one and the aphorism of the week, best way 0:21:02.030 --> 0:21:05.570 to have good ideas is to have lots of ideas. 0:21:05.570 --> 0:21:07.910 The best way to have good ideas is to have 0:21:07.910 --> 0:21:13.209 lots of ideas. Linus Pauling. Unsupervised Learning is produced and 0:21:13.210 --> 0:21:16.810 edited by Daniel Missler on a Neumann 87 AI microphone 0:21:16.810 --> 0:21:20.770 using Hindenburg. Intro and outro music is by zombie with 0:21:20.770 --> 0:21:23.590 a Y, and to get the text and links from 0:21:23.590 --> 0:21:25.720 this episode, sign up for the newsletter version of the 0:21:25.720 --> 0:21:30.570 show at Daniel Missler. Com slash newsletter. We'll see you 0:21:30.570 --> 0:21:31.139 next time.