1 00:00:00,170 --> 00:00:03,980 S1: Say goodbye to countless spreadsheets and endless email threats when 2 00:00:03,980 --> 00:00:07,580 S1: you automate compliance with Fanta. Fanta automates up to 90% 3 00:00:07,580 --> 00:00:11,660 S1: of the work for Soc2, ISO 27,001, HIPAA, and more 4 00:00:11,690 --> 00:00:15,020 S1: getting you audit ready results in weeks instead of months. 5 00:00:15,020 --> 00:00:18,410 S1: Customers report saving over 300 hours of manual work and 6 00:00:18,410 --> 00:00:22,759 S1: up to 85% of associated costs. And with 200 plus integrations, 7 00:00:22,760 --> 00:00:25,490 S1: you can easily monitor and secure the tools your business 8 00:00:25,489 --> 00:00:29,300 S1: relies on. Join the 5000 fast growing companies that leverage 9 00:00:29,300 --> 00:00:32,720 S1: Vanta to manage risks and prove security in real time. 10 00:00:32,750 --> 00:00:36,290 S1: Vanta scales with your business, helping you successfully enter new markets, 11 00:00:36,320 --> 00:00:40,519 S1: land bigger deals, and earn customer trust. Vanta Unsupervised Learning 12 00:00:40,520 --> 00:00:45,559 S1: listeners get $1,000 off Vanta at vanta. Com slash unsupervised. 13 00:00:45,560 --> 00:00:52,370 S1: That's v a n t a.com/unsupervised. Welcome to Unsupervised Learning, 14 00:00:52,400 --> 00:00:55,400 S1: a security, AI and meaning focused podcast that looks at 15 00:00:55,400 --> 00:00:58,490 S1: how best to thrive as humans in a post AI world. 16 00:00:59,000 --> 00:01:02,990 S1: It combines original ideas, analysis and mental models to bring 17 00:01:02,990 --> 00:01:05,510 S1: not just the news, but why it matters and how 18 00:01:05,510 --> 00:01:13,700 S1: to respond. All right, welcome to unsupervised learning. This is 19 00:01:13,700 --> 00:01:17,600 S1: Daniel okay. I'm going to start off with something that 20 00:01:17,600 --> 00:01:22,160 S1: just happened. So strawberry just launched. It is being called 21 00:01:22,190 --> 00:01:26,240 S1: zero one, which I assume the O might mean Orion 22 00:01:26,240 --> 00:01:29,720 S1: because people were saying that it might have been called Orion. 23 00:01:29,720 --> 00:01:32,750 S1: So this is the new model from OpenAI, and I've 24 00:01:32,750 --> 00:01:36,200 S1: been messing with it for a couple hours already. So, uh, 25 00:01:36,200 --> 00:01:38,480 S1: first thing is I gave it a task of building 26 00:01:38,480 --> 00:01:40,970 S1: a business plan for something I'm working on, and it 27 00:01:40,970 --> 00:01:43,760 S1: produced output that was far and above better than Ford 28 00:01:43,790 --> 00:01:48,410 S1: or Sonnet 3.5. Yeah, it was really quite, quite good. 29 00:01:48,500 --> 00:01:51,710 S1: Very detailed. It took quite a while. There's no streaming 30 00:01:51,710 --> 00:01:54,800 S1: in the API, so it feels a little rough compared 31 00:01:54,800 --> 00:01:58,730 S1: to the current models, but Whatever that. That will come 32 00:01:58,730 --> 00:02:02,150 S1: with time. It's quite expensive. So basically I did a 33 00:02:02,150 --> 00:02:08,990 S1: couple of conversation analysis analyses by passing in, um, you know, 34 00:02:09,020 --> 00:02:12,590 S1: conversations like transcripts from podcasts. And I think I did 35 00:02:12,620 --> 00:02:16,010 S1: 2 or 3 of those and it was almost a dollar. 36 00:02:16,010 --> 00:02:19,880 S1: And there's also a mini version which is way less expensive, 37 00:02:19,910 --> 00:02:22,550 S1: but I'm trying to test the capabilities, so I'm using 38 00:02:22,550 --> 00:02:27,440 S1: the full model. But yeah, a few requests for a dollar, 39 00:02:27,470 --> 00:02:32,390 S1: whereas I would say probably many dozen or a couple 40 00:02:32,419 --> 00:02:37,850 S1: of hundred requests are normally like a few dollars. So 41 00:02:37,850 --> 00:02:42,530 S1: it's many factors more expensive. So just something to consider. 42 00:02:42,650 --> 00:02:45,829 S1: As with most models, you don't need the biggest, best 43 00:02:45,830 --> 00:02:48,079 S1: or latest. This is a tweet I just put out, 44 00:02:48,290 --> 00:02:52,250 S1: so I'm going through it. So this does one particular 45 00:02:52,250 --> 00:02:55,860 S1: thing well which is in better than anything else, which 46 00:02:55,860 --> 00:03:00,179 S1: is pausing to think and actually going step by step. 47 00:03:00,180 --> 00:03:02,550 S1: That's kind of like the magic sauce. Here is the 48 00:03:02,550 --> 00:03:05,460 S1: chain of thought reasoning. So if you don't need that 49 00:03:05,460 --> 00:03:08,640 S1: for what you're trying to do, you definitely shouldn't use 50 00:03:08,639 --> 00:03:11,940 S1: this because it's more expensive, takes longer to run. All 51 00:03:11,940 --> 00:03:14,790 S1: those sorts of reasons, this type of model and similar 52 00:03:14,790 --> 00:03:19,380 S1: ones going forward are going to massively benefit from high 53 00:03:19,380 --> 00:03:23,669 S1: quality prompting. So things like we use with fabric, which 54 00:03:23,669 --> 00:03:26,700 S1: is open source on GitHub, if you're not familiar, but 55 00:03:26,700 --> 00:03:30,090 S1: you probably are if you're listening to this. But essentially, 56 00:03:30,090 --> 00:03:31,860 S1: the more you know what you want and the better 57 00:03:31,860 --> 00:03:34,739 S1: you can articulate that, the better this is going to perform, 58 00:03:34,740 --> 00:03:38,670 S1: because it is a chain of thought sort of concept. 59 00:03:38,670 --> 00:03:40,890 S1: So the more you give it to help with that, 60 00:03:40,890 --> 00:03:43,230 S1: the better. Okay, sorry about that. I was just checking 61 00:03:43,260 --> 00:03:46,920 S1: to make sure I wasn't doxxing anyone by showing my messages, 62 00:03:46,920 --> 00:03:51,630 S1: but I was not, so I don't have to rerecord. Okay, so, um, 63 00:03:51,660 --> 00:03:57,510 S1: continuing on here and going to expand this window fully. Okay. So, um, yeah, 64 00:03:57,510 --> 00:04:00,270 S1: the better you can articulate all of this. And by 65 00:04:00,270 --> 00:04:02,340 S1: the way, I want to do an edit there for 66 00:04:02,340 --> 00:04:05,340 S1: the team. So the better you can articulate this stuff 67 00:04:05,370 --> 00:04:08,940 S1: in exactly what you want, the better things are. That's 68 00:04:08,940 --> 00:04:10,980 S1: the bottom line here. So a lot of people are 69 00:04:11,010 --> 00:04:15,330 S1: going to question is this AGI or not? Uh, Sam 70 00:04:15,330 --> 00:04:18,900 S1: Altman already responded. He's like, yeah, this absolutely is not. 71 00:04:19,140 --> 00:04:23,040 S1: So that that should end it in terms of the 72 00:04:23,040 --> 00:04:26,370 S1: actual creator of this thing saying it's not. I also 73 00:04:26,370 --> 00:04:29,730 S1: don't think it is either, whatever that matters for. But 74 00:04:29,910 --> 00:04:32,520 S1: bottom line is, anyone who's making the claim of like 75 00:04:32,550 --> 00:04:36,270 S1: this is or isn't AGI. Here's my request to the internet. Basically, 76 00:04:36,270 --> 00:04:40,289 S1: anyone claiming something is or is not should also provide 77 00:04:40,290 --> 00:04:44,190 S1: a concise and achievable definition of what that means. And 78 00:04:44,190 --> 00:04:46,560 S1: I have one here, of course, which is I've talked 79 00:04:46,560 --> 00:04:50,789 S1: about before, whether the ability of an AI, whether a 80 00:04:50,790 --> 00:04:53,369 S1: model or a product or a system to perform the 81 00:04:53,370 --> 00:04:58,350 S1: work of an average US based knowledge worker in 2002, 82 00:04:58,350 --> 00:05:02,370 S1: and I say 2002 because that's pre GPT four right. 83 00:05:02,370 --> 00:05:06,270 S1: So basically pre AI in these terms anyway. So yeah 84 00:05:06,300 --> 00:05:09,690 S1: anyone who's talking about AGI make sure they have a definition. 85 00:05:09,690 --> 00:05:13,529 S1: Otherwise you're just wasting your time because the entire conversation 86 00:05:13,529 --> 00:05:16,320 S1: will be about definitions. And you might not even figure 87 00:05:16,320 --> 00:05:19,770 S1: that out until fucking two hours later. Sorry for the cussing. 88 00:05:20,160 --> 00:05:22,289 S1: All right. One of the most important changes to me 89 00:05:22,290 --> 00:05:25,440 S1: with this model. This this is massive, okay? This is 90 00:05:25,440 --> 00:05:28,110 S1: the first model that does this. It's the first model 91 00:05:28,110 --> 00:05:31,320 S1: of its kind to do this. Very very interesting. It's 92 00:05:31,350 --> 00:05:36,180 S1: actually spending tokens to think, okay, before you had input 93 00:05:36,180 --> 00:05:39,779 S1: and you had output and you were being charged in, 94 00:05:39,779 --> 00:05:41,850 S1: the amount of work that was being done was based 95 00:05:41,850 --> 00:05:44,669 S1: on the number of tokens coming in and the number 96 00:05:44,670 --> 00:05:47,250 S1: of tokens coming out, and that that was the extent 97 00:05:47,250 --> 00:05:49,530 S1: of it. What's happening now is you have tokens coming 98 00:05:49,560 --> 00:05:53,640 S1: in and you have tokens coming out, but there are 99 00:05:53,640 --> 00:05:59,580 S1: tokens being spent while it's thinking. It's actually thinking and 100 00:05:59,580 --> 00:06:03,540 S1: reasoning through how to solve the problem. And what's really 101 00:06:03,540 --> 00:06:07,740 S1: fascinating about this is that you now have multiple factors here. Okay. 102 00:06:07,770 --> 00:06:11,849 S1: So you can do better prompting. This is the next 103 00:06:11,850 --> 00:06:16,409 S1: piece here. Number seven. You could do better prompting. You 104 00:06:16,410 --> 00:06:19,320 S1: could use a smarter model. Or you could have the 105 00:06:19,320 --> 00:06:23,190 S1: model think harder on the problem. And these are all 106 00:06:23,220 --> 00:06:27,360 S1: going to be levers and knobs that we have to 107 00:06:27,390 --> 00:06:29,910 S1: get better results from AI. And this is the first 108 00:06:29,910 --> 00:06:33,990 S1: time we have this third level lever of like actually 109 00:06:34,020 --> 00:06:38,700 S1: having it think right. So at inference time, more effort 110 00:06:38,700 --> 00:06:41,160 S1: being spent. And they actually say in the blog post 111 00:06:41,160 --> 00:06:43,260 S1: they're like, hey, look, right now it's taking, you know, 112 00:06:43,290 --> 00:06:45,900 S1: a few seconds to think or whatever, and it's going 113 00:06:45,930 --> 00:06:48,849 S1: to get back great results. But we're thinking, what if 114 00:06:48,850 --> 00:06:52,330 S1: it thinks for minutes? What if it thinks for hours? 115 00:06:52,330 --> 00:06:55,659 S1: What if it thinks for days or weeks? And not 116 00:06:55,660 --> 00:06:59,229 S1: only that, but we give it more compute power to think. 117 00:06:59,230 --> 00:07:01,960 S1: And the example they gave, I think this was an 118 00:07:01,960 --> 00:07:05,320 S1: OpenAI post. The example they gave here was how much 119 00:07:05,320 --> 00:07:07,839 S1: do you want to solve cancer? What if you could 120 00:07:07,870 --> 00:07:10,210 S1: build a data center? What if you had one data 121 00:07:10,240 --> 00:07:13,270 S1: center just for working on cancer and one data center 122 00:07:13,270 --> 00:07:17,770 S1: just for working on aging and so on? Okay. And 123 00:07:17,770 --> 00:07:21,040 S1: you basically have models like this that scale with the 124 00:07:21,040 --> 00:07:24,100 S1: inference difficulty based on the amount of difficulty of the, 125 00:07:24,130 --> 00:07:27,010 S1: of the thinking. And then of course, you have a 126 00:07:27,010 --> 00:07:30,310 S1: smart model and a good neural net and all that. Right. 127 00:07:30,370 --> 00:07:32,980 S1: Scalability of the of the neural net. So maybe that's 128 00:07:32,980 --> 00:07:37,180 S1: GPT five, GPT six, whatever. Combined with the good prompting, 129 00:07:37,180 --> 00:07:42,310 S1: combined with this thinking capability and combined with, you know, 130 00:07:42,340 --> 00:07:48,160 S1: all those things unified into the combined with having that 131 00:07:48,160 --> 00:07:52,210 S1: giant infrastructure to run it so that that's insane. And 132 00:07:52,210 --> 00:07:54,220 S1: the scales all the way down to like the smallest 133 00:07:54,220 --> 00:07:58,900 S1: stupid problem where it's just like whatever GPT three and 134 00:07:58,900 --> 00:08:02,200 S1: you get back the answer almost instantaneously. In fact, forget 135 00:08:02,230 --> 00:08:05,590 S1: GPT three. It's some local model that only does one 136 00:08:05,590 --> 00:08:09,340 S1: thing well. You're spending almost no resources whatsoever. It just 137 00:08:09,340 --> 00:08:12,489 S1: goes to your phone, bounces back immediately. It doesn't go anywhere, 138 00:08:12,520 --> 00:08:16,390 S1: barely costs any cycles of a GPU or a CPU 139 00:08:16,420 --> 00:08:19,300 S1: because you don't need those resources to run because it's 140 00:08:19,300 --> 00:08:22,330 S1: just an easy thing to answer. So now we're talking 141 00:08:22,330 --> 00:08:26,770 S1: about AI that scales with the difficulty of the problem, right? With, 142 00:08:26,800 --> 00:08:31,570 S1: you know, cancer, aging, getting out of the solar system, 143 00:08:31,570 --> 00:08:35,980 S1: escaping the sun, expanding, ultimately heat, death of the universe. 144 00:08:35,980 --> 00:08:40,420 S1: That's a big one, right? Because entropy kills everything. So ultimately, 145 00:08:40,420 --> 00:08:42,010 S1: we're going to need a way out of here at 146 00:08:42,010 --> 00:08:45,760 S1: some point, assuming we survive that long. Not happening anytime soon. 147 00:08:45,760 --> 00:08:48,370 S1: I wouldn't worry about that. But these are the types 148 00:08:48,370 --> 00:08:51,460 S1: of things that are really exciting. You know, the size 149 00:08:51,460 --> 00:08:54,580 S1: of the problem being being a factor, for which I 150 00:08:54,610 --> 00:08:57,339 S1: you point at it with lots and lots of different 151 00:08:57,370 --> 00:09:01,900 S1: knobs and levers controlling that decision. So I think that's 152 00:09:01,900 --> 00:09:05,020 S1: really cool. Another important thing to mention is that the 153 00:09:05,020 --> 00:09:08,620 S1: innovation seems independent of what we were waiting for for 154 00:09:08,620 --> 00:09:12,370 S1: GPT five. So based on all I read, all the 155 00:09:12,370 --> 00:09:15,820 S1: releases from OpenAI, and I've seen all the rumors and, 156 00:09:15,850 --> 00:09:18,100 S1: you know, talked to a bunch of people who've been 157 00:09:18,100 --> 00:09:22,540 S1: speculating about this. And this seems completely independent from, oh, 158 00:09:22,570 --> 00:09:25,480 S1: is this GPT four oh, is it for oh, is 159 00:09:25,480 --> 00:09:29,320 S1: it five? Is it an early version of five. Doesn't 160 00:09:29,320 --> 00:09:32,290 S1: really matter. It's like a separate axis. This is like 161 00:09:32,290 --> 00:09:35,920 S1: a capability. This is like thinking capability which is on 162 00:09:35,920 --> 00:09:39,760 S1: a separate axis from how big or smart is the 163 00:09:39,760 --> 00:09:42,910 S1: neural net, right? Or how big or smart is the 164 00:09:42,940 --> 00:09:46,450 S1: is the model. So really, really cool to think about 165 00:09:46,450 --> 00:09:48,880 S1: those being two separate things because now we can start 166 00:09:48,880 --> 00:09:52,179 S1: thinking about, okay, well if GPT five is still going 167 00:09:52,210 --> 00:09:54,220 S1: to come out, you know, later this year or be 168 00:09:54,220 --> 00:09:56,020 S1: getting in next year or whenever it's going to come 169 00:09:56,020 --> 00:09:58,990 S1: out and whatever they're going to call it. Well, imagine 170 00:09:58,990 --> 00:10:03,850 S1: GPT five with this thinking capability. That's cool. So presumably 171 00:10:03,850 --> 00:10:06,189 S1: this is just a feature that you can add onto 172 00:10:06,220 --> 00:10:09,970 S1: any model, which is what we're just talking about. And 173 00:10:09,970 --> 00:10:13,810 S1: I think this is okay. This is really, really crucial here. 174 00:10:13,809 --> 00:10:16,150 S1: I've been talking for a long time about slack in 175 00:10:16,150 --> 00:10:19,929 S1: the rope and tricks that we're going to use to 176 00:10:19,960 --> 00:10:24,400 S1: jump ahead in, um, advancement of AI, so so check 177 00:10:24,400 --> 00:10:26,199 S1: this out. A lot of people are like, oh, we're 178 00:10:26,200 --> 00:10:29,050 S1: running into a data wall. Neural nets are only so 179 00:10:29,050 --> 00:10:31,960 S1: good they can only get so good. We've already hit 180 00:10:31,960 --> 00:10:34,750 S1: a thing. I mean, so many, so many people are 181 00:10:34,750 --> 00:10:40,000 S1: saying things like this that just sound absolutely ridiculous to me. 182 00:10:40,010 --> 00:10:42,380 S1: First of all, they were the ones saying we wouldn't 183 00:10:42,380 --> 00:10:47,090 S1: be here. And so now we are here and everyone's 184 00:10:47,090 --> 00:10:49,790 S1: surprised and they're like, well, here's what we know for 185 00:10:49,790 --> 00:10:52,070 S1: sure is we're not going to get any better. How 186 00:10:52,070 --> 00:10:55,310 S1: can I believe you if you didn't predict any of 187 00:10:55,309 --> 00:10:59,540 S1: this and you were absolutely certain back then, and now 188 00:10:59,540 --> 00:11:03,080 S1: you're absolutely certain it's not going to jump ahead again, right? 189 00:11:03,140 --> 00:11:06,650 S1: Leopold talks about this in his paper. There's lots of 190 00:11:06,650 --> 00:11:11,150 S1: different ways to get better. There's the architecture of the model. 191 00:11:11,150 --> 00:11:13,849 S1: There's the size of the model. I forget what all 192 00:11:13,850 --> 00:11:16,400 S1: levers he had, but it was the architecture of the model, 193 00:11:16,429 --> 00:11:18,170 S1: the size of the model. And I think it was 194 00:11:18,170 --> 00:11:21,920 S1: hobbling was the other one, which is what I called 195 00:11:22,130 --> 00:11:25,040 S1: like a year ago. Slack in the rope or tricks 196 00:11:25,040 --> 00:11:26,660 S1: we're going to. This is what I told a friend 197 00:11:26,660 --> 00:11:29,210 S1: of mine who's really smart in this stuff. I said, 198 00:11:29,210 --> 00:11:32,960 S1: watch this. We're going to find multiple tricks where we're 199 00:11:32,960 --> 00:11:36,410 S1: messing around in percentage points, and then we find a 200 00:11:36,410 --> 00:11:39,320 S1: thing and it jumps us two or 3 or 5 201 00:11:39,320 --> 00:11:43,160 S1: or 10 x or 100 x ahead. And I actually 202 00:11:43,160 --> 00:11:46,490 S1: learned this from him. Uh, I actually learned this from him. 203 00:11:46,490 --> 00:11:48,590 S1: He was like, hey, you know, there are things that 204 00:11:48,590 --> 00:11:51,770 S1: jump you ahead. Um, and I think he gave me 205 00:11:51,770 --> 00:11:54,830 S1: example from some public paper or whatever. And it was 206 00:11:54,830 --> 00:11:57,950 S1: an example of like a big jump. And my natural 207 00:11:57,950 --> 00:12:01,130 S1: intuition was, there's going to be a lot more of those. 208 00:12:01,130 --> 00:12:05,030 S1: And they're not coming from pursuing along this axis, which 209 00:12:05,030 --> 00:12:07,849 S1: is difficult. They are actually just hanging off to the side. 210 00:12:07,850 --> 00:12:10,130 S1: It's like, oh, did you know if you just changed 211 00:12:10,130 --> 00:12:12,050 S1: the color of this? Hey, did you know if you 212 00:12:12,050 --> 00:12:15,589 S1: just orient the data backward instead of forward? Hey, did 213 00:12:15,590 --> 00:12:17,600 S1: you know if you just prune the data in this 214 00:12:17,600 --> 00:12:21,080 S1: way or if you add this particular data set or. 215 00:12:21,110 --> 00:12:24,770 S1: And I'm just making up these examples, but simple things 216 00:12:24,770 --> 00:12:28,310 S1: that you wouldn't think would work. And this is why 217 00:12:28,340 --> 00:12:32,449 S1: Leopold talks about if you automate an AI engineer or 218 00:12:32,450 --> 00:12:35,420 S1: an AI researcher, is what he called it. That's when 219 00:12:35,420 --> 00:12:38,449 S1: it gets completely silly because they have the ability to 220 00:12:38,480 --> 00:12:40,459 S1: now go and try a whole bunch of these things, 221 00:12:40,460 --> 00:12:43,760 S1: including these tricks. All this to say that the slack 222 00:12:43,760 --> 00:12:47,450 S1: in the rope or this series of tricks is going 223 00:12:47,450 --> 00:12:51,200 S1: to keep multiplying our advances. And that's at the same 224 00:12:51,200 --> 00:12:55,370 S1: time that we're working on the algorithms. Oh, that was 225 00:12:55,370 --> 00:12:58,070 S1: the other. That was the other factor is algorithms. That 226 00:12:58,070 --> 00:12:59,929 S1: was this is going to happen. At the same time 227 00:12:59,929 --> 00:13:02,780 S1: we're working on the algorithms to make those better. We're 228 00:13:02,780 --> 00:13:05,690 S1: also working on the size of the neural net and 229 00:13:05,690 --> 00:13:08,150 S1: the quality and the structure. And everything about the neural 230 00:13:08,150 --> 00:13:10,579 S1: net is going to get bigger and more powerful, but 231 00:13:10,580 --> 00:13:13,790 S1: mostly just a matter of size, number of parameters. But 232 00:13:13,790 --> 00:13:16,520 S1: all those things are changing at the same time as 233 00:13:16,550 --> 00:13:20,780 S1: we're finding all these tricks. Right? So we're talking about 234 00:13:21,500 --> 00:13:25,040 S1: this is just begun. And this is what people don't realize. 235 00:13:25,040 --> 00:13:27,560 S1: This is just now starting. We're going to look back 236 00:13:27,559 --> 00:13:30,349 S1: in two years and be like, what was that? That 237 00:13:30,350 --> 00:13:34,160 S1: was silly, right? And so I really want to warn 238 00:13:34,160 --> 00:13:37,280 S1: people against thinking we're hitting some kind of a wall. 239 00:13:37,309 --> 00:13:40,040 S1: Think of it this way. We just found alien technology. 240 00:13:40,070 --> 00:13:42,949 S1: We have no idea how it works. And we're, like, 241 00:13:42,980 --> 00:13:45,949 S1: poking it with a stick, and it's already spitting out 242 00:13:45,950 --> 00:13:49,130 S1: amazing things. So think about that. Okay, we got a 243 00:13:49,130 --> 00:13:51,800 S1: glowy ball. We don't know how it floats. We don't 244 00:13:51,830 --> 00:13:55,309 S1: know how it's doing. Anti-Gravity, right? We don't know how 245 00:13:55,309 --> 00:13:57,829 S1: it's doing this. We don't know how it's reflecting its surface. 246 00:13:57,860 --> 00:13:59,960 S1: We don't know how it's coming up with these answers. 247 00:13:59,960 --> 00:14:02,179 S1: We don't know how it got here from the other 248 00:14:02,179 --> 00:14:05,150 S1: solar system. We don't know anything about it. You poke 249 00:14:05,150 --> 00:14:07,850 S1: it with a stick and it tells this magic stuff 250 00:14:07,850 --> 00:14:12,410 S1: and we're like, Holy crap, that's amazing. Somebody walks up, 251 00:14:12,440 --> 00:14:14,690 S1: sees you poke it with a stick and goes, yeah, 252 00:14:14,720 --> 00:14:17,240 S1: that's I mean, that's that's all it's ever going to 253 00:14:17,240 --> 00:14:19,640 S1: be able to do. I mean, I've seen you poke 254 00:14:19,640 --> 00:14:22,970 S1: it with a stick twice, and it gave you kind 255 00:14:22,970 --> 00:14:26,060 S1: of a similar answer, which means that's all we could 256 00:14:26,060 --> 00:14:29,690 S1: learn from this alien ball. That's their conclusion. I am 257 00:14:29,690 --> 00:14:32,840 S1: certain that since you poked it with a stick while 258 00:14:32,840 --> 00:14:35,390 S1: I was standing here three times, and it kind of 259 00:14:35,420 --> 00:14:39,590 S1: gave you a similar answer. One it must be stupid. Two, 260 00:14:39,620 --> 00:14:42,710 S1: it's not as smart as us. And three, this is 261 00:14:42,710 --> 00:14:45,860 S1: as as smart as it's ever going to be. This 262 00:14:45,860 --> 00:14:48,350 S1: is the most it has to offer. That is the 263 00:14:48,350 --> 00:14:52,160 S1: claim that's being made by these kind of like denialists, 264 00:14:52,190 --> 00:14:56,510 S1: in my view. And that doesn't mean the current shiny 265 00:14:56,510 --> 00:15:01,130 S1: ball is better than humans, or it should replace humans, 266 00:15:01,130 --> 00:15:04,070 S1: or it could do everything we could do. Like this 267 00:15:04,070 --> 00:15:06,470 S1: is not a competition. Okay, here's a better way to 268 00:15:06,470 --> 00:15:09,410 S1: think about this. This is not like a rock that 269 00:15:09,410 --> 00:15:12,320 S1: we have animated. Think of it this way. If an 270 00:15:12,320 --> 00:15:15,710 S1: alien comes here because someone else was like, hey, this 271 00:15:15,710 --> 00:15:19,460 S1: is not thinking, this is processing. And I'm like, come on, 272 00:15:19,460 --> 00:15:22,850 S1: come on. If you if an alien comes here, let's 273 00:15:22,850 --> 00:15:26,210 S1: assume we know how our brain works. An alien comes 274 00:15:26,210 --> 00:15:28,820 S1: here and we look at its brain, or it shows 275 00:15:28,820 --> 00:15:32,040 S1: us its brain and it looks different. And we're like, oh, 276 00:15:32,070 --> 00:15:36,900 S1: you guys do neurons and synapses different than us? Who's 277 00:15:36,900 --> 00:15:39,479 S1: going to walk over and be like, well, since they're 278 00:15:39,480 --> 00:15:43,290 S1: doing neurons and synapses different than us, they're not thinking. 279 00:15:43,290 --> 00:15:47,280 S1: Only humans can think. And I'm like, they got here. 280 00:15:47,310 --> 00:15:50,070 S1: They got here, didn't they? It's a little shiny ball. 281 00:15:50,070 --> 00:15:52,800 S1: And they got here from whatever part of the galaxy 282 00:15:52,800 --> 00:15:56,460 S1: or universe that they came from. They're obviously doing something right. 283 00:15:56,460 --> 00:15:59,850 S1: And AI is obviously doing something right too. So I 284 00:15:59,850 --> 00:16:02,729 S1: think it's a little bit specious. Is that is that 285 00:16:02,730 --> 00:16:07,170 S1: the name of the word? It's like specious to just 286 00:16:07,170 --> 00:16:10,650 S1: magically assume that we are the best. Only we are 287 00:16:10,680 --> 00:16:14,730 S1: thinking only we are special. Instead of thinking like we 288 00:16:14,730 --> 00:16:19,050 S1: might have this nascent alien intelligence thing going on that 289 00:16:19,050 --> 00:16:21,930 S1: actually is doing things that are very much analogous to us. 290 00:16:21,960 --> 00:16:24,150 S1: It reminds me of the first time that I clicked 291 00:16:24,150 --> 00:16:27,210 S1: around inside of Linux. This is like late 90s. I 292 00:16:27,210 --> 00:16:30,090 S1: was messing with Linux. this must have been like 97, 293 00:16:30,120 --> 00:16:33,630 S1: 98 or something. I'm messing with Linux and I'm clicking 294 00:16:33,630 --> 00:16:36,930 S1: around because I had started with windows and I'm like, oh, 295 00:16:36,960 --> 00:16:39,570 S1: it opens windows and it opens things that I could 296 00:16:39,570 --> 00:16:42,930 S1: click and navigate. Then I'm like, it's it's just like 297 00:16:42,930 --> 00:16:47,790 S1: on Windows Explorer. And this like, blew me away. It 298 00:16:47,820 --> 00:16:50,640 S1: absolutely blew me away that this was just a different 299 00:16:50,640 --> 00:16:53,910 S1: way of doing the same thing. And that underneath this, 300 00:16:53,910 --> 00:16:57,000 S1: there's a universal thing of you need to be able 301 00:16:57,000 --> 00:16:59,790 S1: to browse files, you need to be able to open windows, 302 00:16:59,790 --> 00:17:02,580 S1: you need to be able to close windows. And that 303 00:17:02,580 --> 00:17:04,590 S1: clicked for me. And I'm like, oh, I guess like 304 00:17:04,619 --> 00:17:07,440 S1: all operating systems are going to do this differently. It's 305 00:17:07,440 --> 00:17:09,870 S1: the same with aliens. It's the same with like they 306 00:17:09,869 --> 00:17:13,200 S1: might think differently, but whatever. They have to think, right. 307 00:17:13,200 --> 00:17:16,649 S1: So why would we expect this synthetic intelligence that we've 308 00:17:16,680 --> 00:17:19,410 S1: birthed to do it exactly the same way that we 309 00:17:19,440 --> 00:17:23,640 S1: way that we do? We should not expect that we 310 00:17:23,640 --> 00:17:29,040 S1: got here accidentally stumbling through time due to evolution. And 311 00:17:29,040 --> 00:17:32,550 S1: we've got this version that we have and it's awesome, obviously. 312 00:17:32,550 --> 00:17:36,360 S1: But like that's way different than we invented this thing 313 00:17:36,390 --> 00:17:40,469 S1: five years ago or whenever that was 2017, six years ago. 314 00:17:40,500 --> 00:17:42,840 S1: And I know it goes further back than that. But 315 00:17:42,869 --> 00:17:46,500 S1: you know what I'm saying? Transformers. All right. So that's that. 316 00:17:46,500 --> 00:17:49,950 S1: And this this is becoming a long thing. But whatever 317 00:17:49,950 --> 00:17:53,820 S1: we'll go with it. So yeah, basically we have no 318 00:17:54,119 --> 00:17:57,750 S1: idea how early all of this is. We're likely to 319 00:17:57,780 --> 00:18:00,929 S1: find ten, 20 or 200 more of these holy crap 320 00:18:00,960 --> 00:18:04,919 S1: optimizations like this thinking thing before we start hitting any 321 00:18:04,920 --> 00:18:11,550 S1: limits for neural network architecture or the transformer transformer like. 322 00:18:11,580 --> 00:18:14,790 S1: Plus we could just find something better than a transformer. 323 00:18:14,790 --> 00:18:19,200 S1: You realize how how lucky we were to find the transformer. 324 00:18:19,230 --> 00:18:21,990 S1: Like the people who made that paper. They're like, hey, 325 00:18:21,990 --> 00:18:24,210 S1: this is this is a cool way. We think this 326 00:18:24,210 --> 00:18:26,100 S1: is a cool way of doing something. They didn't know 327 00:18:26,100 --> 00:18:29,220 S1: what they had. Okay. You should watch a Karpathy talk 328 00:18:29,250 --> 00:18:32,520 S1: about the transformer. He's like, this thing is a general 329 00:18:32,520 --> 00:18:36,389 S1: purpose computer. This thing is insanely good at learning. He 330 00:18:36,390 --> 00:18:40,379 S1: talks about different ways that it's better than humans at learning. Okay. 331 00:18:40,410 --> 00:18:45,119 S1: Some some people randomly found this thing, and it shot 332 00:18:45,119 --> 00:18:48,030 S1: us off. Okay. So. So check this out. This is 333 00:18:48,030 --> 00:18:52,950 S1: another example of finding tricks or slack in the rope 334 00:18:52,950 --> 00:18:55,830 S1: just lying on the ground. So we stumble through AI 335 00:18:55,859 --> 00:19:00,240 S1: for decades and decades and decades. And then someone's like, hey, 336 00:19:00,240 --> 00:19:02,700 S1: this is kind of cool about this attention mechanism. Hey, 337 00:19:02,700 --> 00:19:06,479 S1: what do you think about this architecture for a neural net? Boom, 338 00:19:06,480 --> 00:19:10,710 S1: now we have this take off. There's nothing saying somebody 339 00:19:10,710 --> 00:19:13,590 S1: isn't going to be like, I like what you did 340 00:19:13,590 --> 00:19:18,540 S1: with that transformer architecture. What if it looked like this instead? 341 00:19:18,570 --> 00:19:22,410 S1: It might be 20 times better. It might be 2000 342 00:19:22,410 --> 00:19:25,840 S1: times better. It might be 4% better. It doesn't matter. 343 00:19:25,869 --> 00:19:30,280 S1: Like we have only just begun. We have only just begun. 344 00:19:30,280 --> 00:19:33,910 S1: I can absolutely guarantee you that assuming we don't kill 345 00:19:33,910 --> 00:19:36,640 S1: ourselves off as a result of this, like that would 346 00:19:36,640 --> 00:19:38,949 S1: set things back. But I'm trying to get you to 347 00:19:38,980 --> 00:19:43,810 S1: think about things in this way, because it's insane what's 348 00:19:43,810 --> 00:19:45,730 S1: about to happen. And yeah, I'm gonna I'm gonna have 349 00:19:45,730 --> 00:19:47,980 S1: more examples here. I'm working on an example right here 350 00:19:47,980 --> 00:19:50,919 S1: on this other screen. Uh, pretty cool thing I'm building 351 00:19:50,920 --> 00:19:54,369 S1: with it. Um, okay, so that was that. All right, 352 00:19:54,400 --> 00:19:58,629 S1: now back to the show. Okay. We are adding a 353 00:19:58,630 --> 00:20:01,899 S1: whole bunch more content to the podcast. So if you 354 00:20:01,930 --> 00:20:04,990 S1: have not been listening to the podcast, I used to 355 00:20:04,990 --> 00:20:08,770 S1: do this really dry thing. Um, and sometimes I'm still dry, 356 00:20:08,800 --> 00:20:12,310 S1: so whatever. I'm an intellectual guy. Sometimes I just like, 357 00:20:12,340 --> 00:20:15,100 S1: talk about the idea. I'm not overly excited, like the 358 00:20:15,100 --> 00:20:17,859 S1: way I was just now, and it could be a 359 00:20:17,859 --> 00:20:20,290 S1: little boring. I mean, I started this thing back in 360 00:20:20,290 --> 00:20:23,600 S1: 2015 and I was literally like, welcome to the Take 361 00:20:23,630 --> 00:20:27,560 S1: One Security podcast. The podcast where I do everything and 362 00:20:27,560 --> 00:20:30,619 S1: take 1 in 1 take. That's all I do. I 363 00:20:30,619 --> 00:20:35,360 S1: just read the news. Okay. First news. Now that was fire. 364 00:20:35,390 --> 00:20:40,040 S1: The podcast blew up. Why? Because no one had a podcast. Well, 365 00:20:40,040 --> 00:20:43,669 S1: there was Risky Biz. There was like GRC. What's his name? 366 00:20:43,670 --> 00:20:47,900 S1: Steve Gibson. There was that there was a couple, but very, 367 00:20:47,900 --> 00:20:53,090 S1: very few. And actually mine was dry and to the point. 368 00:20:53,090 --> 00:20:56,990 S1: So it was like super exciting. Anyway, that that is 369 00:20:56,990 --> 00:21:00,980 S1: the old version that is no longer interesting. People got 370 00:21:00,980 --> 00:21:03,770 S1: tired of that and I lost interest in it because 371 00:21:03,770 --> 00:21:05,990 S1: I just didn't want to talk about the news anymore. 372 00:21:06,020 --> 00:21:09,290 S1: I wanted to move on to other things. So it 373 00:21:09,290 --> 00:21:12,410 S1: is now what it is. The point is, the thing 374 00:21:12,410 --> 00:21:15,770 S1: that I just did about I about the release of strawberry, 375 00:21:15,800 --> 00:21:19,100 S1: that is now a piece which will be cut out 376 00:21:19,100 --> 00:21:21,790 S1: by the team. The team will put that on YouTube 377 00:21:21,790 --> 00:21:25,030 S1: and the team will put out that out on the podcast. 378 00:21:25,030 --> 00:21:28,300 S1: And when I do little clips like that, I don't 379 00:21:28,300 --> 00:21:30,340 S1: know that I'm in a clip in the moment. But 380 00:21:30,340 --> 00:21:33,040 S1: what happens is they take that and put it into 381 00:21:33,040 --> 00:21:36,129 S1: its own piece. They name it, they do a thumbnail maybe. 382 00:21:36,160 --> 00:21:38,260 S1: I'm not sure about that, but they put it out 383 00:21:38,260 --> 00:21:42,490 S1: there and now that's a piece of content, and oftentimes 384 00:21:42,490 --> 00:21:45,399 S1: it's way easier to listen to those. I would prefer 385 00:21:45,400 --> 00:21:47,890 S1: to listen to those. Again, I'm building the thing that 386 00:21:47,920 --> 00:21:51,460 S1: I believe needs to exist because it's what I wish 387 00:21:51,460 --> 00:21:54,609 S1: I was getting from other people. Right? So Alex Hormozi, 388 00:21:54,640 --> 00:21:57,609 S1: Chris Williamson, all these different people, I love the fact 389 00:21:57,609 --> 00:22:01,240 S1: that I can get a single discrete piece of content, 390 00:22:01,240 --> 00:22:04,930 S1: and I don't mean like a YouTube short, like 30s 391 00:22:04,930 --> 00:22:07,720 S1: or a minute. I mean something like, I don't care 392 00:22:07,720 --> 00:22:10,720 S1: if it's 30s or if it's like 3 or 4 minutes. 393 00:22:10,720 --> 00:22:15,040 S1: I just want a discrete idea labeled as that discrete idea. 394 00:22:15,040 --> 00:22:18,399 S1: So that's what I've been doing. I've been taking content 395 00:22:18,430 --> 00:22:21,460 S1: either making it independently or if it happens to come 396 00:22:21,460 --> 00:22:25,030 S1: out like it just did. Talking about strawberry inside of 397 00:22:25,030 --> 00:22:29,080 S1: a big episode that is then brought out and turned 398 00:22:29,080 --> 00:22:32,920 S1: into a full thing. Now, the advantage here is that 399 00:22:32,920 --> 00:22:35,080 S1: the podcast is now going to be full of content. 400 00:22:35,080 --> 00:22:40,090 S1: There's now going to be three, four, five, six pieces 401 00:22:40,090 --> 00:22:46,240 S1: of hopefully decent pieces of content inside the podcast feed. 402 00:22:46,240 --> 00:22:48,340 S1: So if you're a subscriber to the podcast, which you 403 00:22:48,340 --> 00:22:51,550 S1: should definitely go do on Apple or Spotify or whatever 404 00:22:51,550 --> 00:22:56,560 S1: your client, um, you will not only have this big episode, which, 405 00:22:56,560 --> 00:22:59,650 S1: when I mess around and start ranting like this, could 406 00:22:59,650 --> 00:23:03,550 S1: end up being 30 minutes long, right? But in addition 407 00:23:03,550 --> 00:23:05,950 S1: to those, you're going to have these really clean little 408 00:23:05,950 --> 00:23:09,880 S1: pieces of content that's like a single idea. So bottom 409 00:23:09,880 --> 00:23:12,640 S1: line is get back on the podcast, whether that's on 410 00:23:12,640 --> 00:23:18,140 S1: YouTube or on, um, Audio? Definitely. Please go subscribe on 411 00:23:18,140 --> 00:23:24,980 S1: the audio version so Apple Podcasts, Spotify, whatever, Android. And uh, yeah, 412 00:23:25,010 --> 00:23:27,890 S1: tune in because you're probably going to like it a lot. Okay, 413 00:23:27,920 --> 00:23:30,770 S1: new keyboard I'm using. Okay. How do I type without 414 00:23:30,770 --> 00:23:34,580 S1: messing anything up? Um, let's see here. Okay, I'll just do. Okay. 415 00:23:34,609 --> 00:23:36,709 S1: I'm going to type fast. I don't know if you 416 00:23:36,710 --> 00:23:39,680 S1: can hear this. I got a Sennheiser here listening. Okay. 417 00:23:39,680 --> 00:23:46,040 S1: This keyboard is called the Aleph 75. I love this keyboard. 418 00:23:46,040 --> 00:23:48,830 S1: And it's actually super cheap. It's like 70 bucks. I've 419 00:23:48,830 --> 00:23:52,159 S1: purchased so many expensive keyboards that are sitting over there 420 00:23:52,160 --> 00:23:55,700 S1: and I don't use them. I was using this one 421 00:23:55,700 --> 00:23:58,159 S1: for a while, but I heard this sound on a 422 00:23:58,160 --> 00:24:00,890 S1: YouTube video in a review and I was like, ooh, 423 00:24:00,920 --> 00:24:03,800 S1: I like that sound. Hope it sounds like that in person. 424 00:24:03,800 --> 00:24:06,859 S1: And yeah, it does. It sounds like that in person. 425 00:24:06,859 --> 00:24:11,120 S1: Hopefully the post-production will not take out that sound, but, um, 426 00:24:11,150 --> 00:24:13,970 S1: I love the sound of this thing. Um, this particular 427 00:24:13,970 --> 00:24:18,649 S1: one has no key. Uh, has no labels on the 428 00:24:18,650 --> 00:24:20,750 S1: top of the keys. They're actually on the bottom of 429 00:24:20,750 --> 00:24:23,570 S1: the keys, and they're backlit, so it's pretty cool. So 430 00:24:23,570 --> 00:24:27,440 S1: the whole keyboard looks kind of gray. It's really heavy. 431 00:24:27,440 --> 00:24:29,810 S1: I like it, but mostly I like the sound and 432 00:24:29,810 --> 00:24:34,909 S1: the feel. So combine that with Neovim config text based world. 433 00:24:34,940 --> 00:24:38,240 S1: I'm happy. I'm a happy person. Okay, continue to be 434 00:24:38,240 --> 00:24:41,480 S1: blown away by the idea of encapsulating what people think 435 00:24:41,480 --> 00:24:45,290 S1: the biggest problem in the world is using extract. Primary 436 00:24:45,290 --> 00:24:49,879 S1: problem I love this thing. So echo Viktor Frankl's work, 437 00:24:49,970 --> 00:24:53,270 S1: pipe that into extract primary problem and look what it 438 00:24:53,270 --> 00:24:56,540 S1: brings back. The lack of meaning in life leads to 439 00:24:56,570 --> 00:25:02,060 S1: suffering and existential despair. That is beautiful, that is beautiful. 440 00:25:02,060 --> 00:25:05,840 S1: And I'm taking someone's entire body of work and pulling 441 00:25:05,840 --> 00:25:09,770 S1: out what they think the biggest problem in the world is. 442 00:25:09,800 --> 00:25:13,310 S1: That's just astounding to me. Yeah. So the keynote at 443 00:25:13,310 --> 00:25:17,270 S1: Sans went really well. Almost 30 minutes of questions afterwards 444 00:25:17,270 --> 00:25:20,330 S1: was really cool for them to have me appreciate them. Yeah, 445 00:25:20,359 --> 00:25:24,980 S1: I'm experimenting with some micro art fiction. And Tim from 446 00:25:24,980 --> 00:25:26,899 S1: community was like, hey, what are you working on? You're 447 00:25:26,900 --> 00:25:32,060 S1: working on a book? Uh, not really. Kind of. That's 448 00:25:32,060 --> 00:25:36,320 S1: all I'll say about that. Not actively and fiercely, but 449 00:25:36,320 --> 00:25:39,800 S1: I will say that. But yeah, some ideas out there 450 00:25:39,800 --> 00:25:42,260 S1: working on a bunch of flagship content. Right now, I'm 451 00:25:42,260 --> 00:25:45,260 S1: mentioning more and more about human 3.0. So I'm going 452 00:25:45,290 --> 00:25:48,590 S1: to put out like a description of that, like the 453 00:25:48,590 --> 00:25:52,100 S1: structure and what I'm working on there. Uh, a new 454 00:25:52,100 --> 00:25:55,490 S1: piece on security asset management in AI, how to write 455 00:25:55,490 --> 00:25:58,550 S1: fiction using AI. That one's going to be insane and 456 00:25:58,550 --> 00:26:01,400 S1: a whole bunch of other ones. And all right, let's 457 00:26:01,400 --> 00:26:05,750 S1: get into security. Chinese government backed hackers have been infiltrating 458 00:26:05,780 --> 00:26:09,619 S1: US internet service providers to spy on users, according to 459 00:26:09,650 --> 00:26:14,360 S1: a bunch of researchers, Halliburton confirmed a cyberattack where intruders 460 00:26:14,390 --> 00:26:18,620 S1: access an exfiltrated data, and they blamed the Ransom hub Group, 461 00:26:18,650 --> 00:26:25,400 S1: who also claimed responsibility. Predator spyware predator spyware is back 462 00:26:25,400 --> 00:26:27,920 S1: with new features that make it even harder to track, 463 00:26:27,920 --> 00:26:31,520 S1: and looks like it's showing up in Democratic Republic of 464 00:26:31,520 --> 00:26:35,570 S1: Congo and Angola, and there's a lot more anonymity built 465 00:26:35,570 --> 00:26:38,750 S1: into it. So the trick with this kind of malware 466 00:26:38,780 --> 00:26:41,419 S1: is that it's not just the people who make it, 467 00:26:41,420 --> 00:26:44,270 S1: it's the people who sell it. And the fact that 468 00:26:44,270 --> 00:26:48,649 S1: there's a market for governments who want to control their populations. 469 00:26:48,650 --> 00:26:52,939 S1: And that's what's especially scary about, like the CCP way 470 00:26:52,940 --> 00:26:55,940 S1: of viewing the world. And if you combine that with AI, 471 00:26:55,940 --> 00:27:00,080 S1: that can kind of control things and see everything and 472 00:27:00,109 --> 00:27:04,640 S1: like determine who's breaking the rules and monitor and surveil 473 00:27:04,640 --> 00:27:07,190 S1: and things like that, and you combine those things together 474 00:27:07,250 --> 00:27:11,870 S1: And you do have a recipe for controlling populations. Latest 475 00:27:11,869 --> 00:27:17,480 S1: version of NIST, CSF, which is two, introduces governance as 476 00:27:17,480 --> 00:27:21,710 S1: a new step and focuses on continuous improvement to adapt 477 00:27:21,710 --> 00:27:25,700 S1: to emerging threats. And it also they also released a 478 00:27:25,700 --> 00:27:31,520 S1: continuous threat exposure management framework, CDM framework. Thanks to Hyper 479 00:27:31,520 --> 00:27:36,530 S1: Proof for sponsoring Maltese security, researchers have been charged after 480 00:27:36,530 --> 00:27:41,810 S1: discovering a flaw in an application called Free Hour. And 481 00:27:41,810 --> 00:27:44,450 S1: I think they are in jail. Yeah, they're in jail 482 00:27:44,450 --> 00:27:48,770 S1: waiting for trial, I believe, because they don't have like, oh, disclosure, 483 00:27:48,859 --> 00:27:52,760 S1: public disclosure or, you know, good citizen or anything like that. 484 00:27:52,760 --> 00:27:55,010 S1: It's like, nope, you did a bad thing. You're in jail. 485 00:27:55,040 --> 00:27:57,830 S1: US Space Force is gearing up for potential conflicts in 486 00:27:57,830 --> 00:28:01,850 S1: space with countries like China and Russia, and they're trying 487 00:28:01,850 --> 00:28:06,870 S1: to predict the stuff in space, obviously, Satellites and other 488 00:28:06,869 --> 00:28:11,159 S1: space assets. And thanks to dropzone for sponsoring as well. 489 00:28:11,160 --> 00:28:14,160 S1: In the US, is offering a $10 million reward for 490 00:28:14,160 --> 00:28:17,580 S1: information on the Russian hacking group called Cadet Blizzard, linked 491 00:28:17,580 --> 00:28:23,490 S1: to Gru's unit 29 155, which has been particularly focused 492 00:28:23,490 --> 00:28:26,280 S1: on disrupting aid to Ukraine. NASA is launching a new 493 00:28:26,280 --> 00:28:31,020 S1: podcast called No Such Podcast. Good name. That's a good name. Evidently, 494 00:28:31,020 --> 00:28:33,900 S1: lots of people use the I forgot My password feature 495 00:28:33,900 --> 00:28:38,010 S1: as a de facto login method. Human behavior A Starlink 496 00:28:38,010 --> 00:28:41,070 S1: satellite dish was used on a US Navy ship for 497 00:28:41,070 --> 00:28:45,060 S1: an illicit Wi-Fi network called stinky, which was used for 498 00:28:45,060 --> 00:28:49,620 S1: streaming and civilian communication. The Navy demoted the senior enlisted 499 00:28:49,650 --> 00:28:54,060 S1: leader responsible for being awesome. Uh, so I didn't write that. 500 00:28:54,060 --> 00:28:58,710 S1: I wrote that right. Obviously. Uh, and I got some pushback. Honestly, 501 00:28:58,710 --> 00:29:02,430 S1: I got some pushback from some military people and some 502 00:29:02,430 --> 00:29:04,450 S1: Navy people, and they're like, hey, this is not a 503 00:29:04,450 --> 00:29:07,060 S1: laughing matter. I know it's not a laughing matter. I 504 00:29:07,060 --> 00:29:10,630 S1: know you shouldn't be allowed to set up satellite receivers 505 00:29:10,630 --> 00:29:15,880 S1: on a warship. I know this. I'm ex-military too. I 506 00:29:15,880 --> 00:29:17,830 S1: just thought it was funny. Like they're trying to give 507 00:29:17,860 --> 00:29:22,959 S1: internet to the people. They're a hero. I know it's bad. Like, sorry. 508 00:29:22,990 --> 00:29:26,740 S1: Not sorry. Um. All right. I tech. Apple released their 509 00:29:26,740 --> 00:29:30,850 S1: September updates yesterday, and they were decent. So I'm going 510 00:29:30,850 --> 00:29:34,750 S1: to jump ahead. I'm basically getting probably I don't think 511 00:29:34,750 --> 00:29:36,460 S1: I'm going to get the new AirPods. I saw a 512 00:29:36,460 --> 00:29:39,280 S1: thing that basically said that the AirPods Pro two are 513 00:29:39,280 --> 00:29:42,910 S1: still better because they have the silicone tip part, and 514 00:29:42,910 --> 00:29:45,430 S1: I'm going to get the black 16 Pro, but not 515 00:29:45,430 --> 00:29:48,190 S1: the max. I can't remember if there was some. Oh, 516 00:29:48,220 --> 00:29:50,410 S1: I might get the watch as well. I might either 517 00:29:50,440 --> 00:29:55,300 S1: get a black Ultra Series two or I might get 518 00:29:55,300 --> 00:29:58,630 S1: the series ten because it's got a bigger screen then 519 00:29:58,630 --> 00:30:00,970 S1: the ultra. I don't know which one I'm going to get. 520 00:30:00,970 --> 00:30:02,760 S1: I'm going to wait and see it in person because 521 00:30:02,760 --> 00:30:05,550 S1: I will be camping on the night of the 19th 522 00:30:05,550 --> 00:30:08,310 S1: in Burlingame. So if anyone wants to come hang out 523 00:30:08,310 --> 00:30:11,460 S1: in person, we can have like an impromptu UL meetup 524 00:30:11,460 --> 00:30:17,670 S1: because I've been there for the last let's see since ten, 2010. 525 00:30:17,790 --> 00:30:22,560 S1: Holy crap. I've been camping at that store for 14 526 00:30:22,560 --> 00:30:27,030 S1: years and I camped the three previous times in other places. 527 00:30:27,030 --> 00:30:30,600 S1: So I've been camping for the iPhone since 2007, but 528 00:30:30,600 --> 00:30:34,080 S1: I've been doing it at that store like first in 529 00:30:34,080 --> 00:30:36,870 S1: line for last 14 years, and I will be there 530 00:30:36,870 --> 00:30:39,780 S1: again on the 19th. If anyone wants to do an 531 00:30:39,780 --> 00:30:43,320 S1: impromptu UL meetup, which it won't be an actual UL meetup, 532 00:30:43,320 --> 00:30:45,780 S1: it'll just be us hanging out because there's coffee shops, 533 00:30:45,780 --> 00:30:48,330 S1: there's a Phil's right there so we can hang out. 534 00:30:48,330 --> 00:30:53,520 S1: We could talk. Um, I usually walk a lot through 535 00:30:53,520 --> 00:30:55,350 S1: the course of the night, maybe try to take a 536 00:30:55,350 --> 00:30:59,190 S1: nap or whatever at night. But, uh, anyway, I'll be there. 537 00:30:59,190 --> 00:31:04,830 S1: Nvidia's RTX 40 series GPUs, including the 5080 and 5090 538 00:31:04,860 --> 00:31:09,090 S1: are getting ready to be finalized and potentially come out 539 00:31:09,090 --> 00:31:12,090 S1: end of the year or beginning of next year. They're 540 00:31:12,090 --> 00:31:14,850 S1: saying these things are going to be massively power hungry. 541 00:31:14,880 --> 00:31:17,310 S1: I've got two 49 seconds in the room next to 542 00:31:17,340 --> 00:31:20,700 S1: me here, and they are beasts. I'm going to have 543 00:31:20,700 --> 00:31:24,330 S1: some FOMO when the 5090 comes out, but I don't 544 00:31:24,330 --> 00:31:27,690 S1: think I'm going to be upgrading my system because that 545 00:31:27,690 --> 00:31:30,210 S1: would be expensive and I'm not sure I need that, 546 00:31:30,210 --> 00:31:34,320 S1: especially when a lot of the AI models are probably 547 00:31:34,320 --> 00:31:38,460 S1: going to run better. I'd rather wait and upgrade my 548 00:31:38,460 --> 00:31:42,000 S1: main desktop on Apple and see what, like the giant 549 00:31:42,030 --> 00:31:47,280 S1: M4 or giant M5 systems have. When I have, let's say, 550 00:31:47,310 --> 00:31:51,720 S1: half a terabyte of system on a chip memory and 551 00:31:51,720 --> 00:31:54,810 S1: all of that can be used for GPU operations. I'm 552 00:31:54,810 --> 00:31:57,510 S1: not sure I'm going to need a dedicated, well, maybe 553 00:31:57,510 --> 00:32:01,060 S1: I'll need a dedicated one of those boxes to do I. 554 00:32:01,060 --> 00:32:03,820 S1: But I'm not sure I'm going to build another dedicated 555 00:32:03,820 --> 00:32:07,300 S1: PC with Nvidia chips in it, which is weird because 556 00:32:07,300 --> 00:32:10,000 S1: I got a lot of Nvidia stock. Anyway, that's a 557 00:32:10,000 --> 00:32:14,350 S1: separate talk show. Uh, okay. Trump is launching a crypto project, 558 00:32:14,350 --> 00:32:17,500 S1: but there are concerns because 70% of the tokens are 559 00:32:17,500 --> 00:32:22,630 S1: being allocated to insiders. Uh, Ilya has a new startup, SSI, 560 00:32:22,630 --> 00:32:25,870 S1: and they raised $1 billion. So a lot of people 561 00:32:25,870 --> 00:32:30,850 S1: have $1 billion valuations. Ilya did it differently. He has 562 00:32:30,850 --> 00:32:35,590 S1: a $1 billion seed round, a $1 billion seed round. Insane. 563 00:32:35,620 --> 00:32:38,560 S1: Visa is set to launch a new account to account 564 00:32:38,590 --> 00:32:41,230 S1: a to a payment service in Europe. So you could 565 00:32:41,230 --> 00:32:45,190 S1: make direct bank to bank transfers without using credit cards. 566 00:32:45,220 --> 00:32:47,920 S1: I wonder if that's a competitor to Swift. I imagine 567 00:32:47,920 --> 00:32:50,020 S1: it is. Or maybe it rides on top of Swift, 568 00:32:50,020 --> 00:32:53,560 S1: I don't know. Engineers from Cornell and Florence University have 569 00:32:53,560 --> 00:32:58,479 S1: developed a biohybrid robot that uses electric signals from a 570 00:32:58,480 --> 00:33:03,130 S1: king trumpet mushroom to move and sense its environment. Reminds 571 00:33:03,130 --> 00:33:08,230 S1: me of Neri Oxman. Neri Oxman doing her, uh, networking 572 00:33:08,260 --> 00:33:13,300 S1: of nature types. Uh, stuff for 2024 annual Work Trend 573 00:33:13,300 --> 00:33:16,540 S1: Index from Microsoft and LinkedIn reveals a shift in employer 574 00:33:16,540 --> 00:33:21,340 S1: preferences 71% of leaders favoring candidates with AI skill over 575 00:33:21,340 --> 00:33:24,880 S1: those with industry experience. Yeah, I found this one kind 576 00:33:24,910 --> 00:33:28,810 S1: of interesting. Um, Wall Street Journal is highlighting a trend 577 00:33:28,840 --> 00:33:32,980 S1: where small startups are increasingly influencing the US economy. So 578 00:33:32,980 --> 00:33:34,780 S1: I've been thinking about this for a while, is getting 579 00:33:34,780 --> 00:33:36,460 S1: ready to write a piece on it. But I want 580 00:33:36,490 --> 00:33:38,860 S1: to I want to state this more forcefully. I think 581 00:33:38,860 --> 00:33:41,290 S1: people are about to realize that the most medium to 582 00:33:41,320 --> 00:33:47,140 S1: large companies, like the size companies, have become ineffective. They 583 00:33:47,140 --> 00:33:50,830 S1: lack vision and focus, too much bureaucracy. They have giant 584 00:33:50,830 --> 00:33:54,070 S1: workforces that are hired for like a worker bee mentality 585 00:33:54,100 --> 00:33:58,060 S1: rather than to have them be exceptional or innovative or 586 00:33:58,060 --> 00:34:01,150 S1: challenging to the structure. And I think this is another 587 00:34:01,150 --> 00:34:03,640 S1: part of the essay that I wrote called The End 588 00:34:03,640 --> 00:34:06,310 S1: of Work, where much of the innovation in the world 589 00:34:06,310 --> 00:34:09,790 S1: moves away from big companies and towards individuals and small, 590 00:34:09,790 --> 00:34:13,270 S1: dynamic startups. And this is also what Marc Andreessen talked 591 00:34:13,270 --> 00:34:16,240 S1: about in his conversation with Huberman, which I thought just 592 00:34:16,239 --> 00:34:18,370 S1: came out. But it turns out that was a year ago. 593 00:34:18,370 --> 00:34:22,480 S1: But that was a fantastic episode. And related to this, 594 00:34:22,480 --> 00:34:27,100 S1: Paul Graham's piece, uh, his latest piece, called Founder Mode, 595 00:34:27,100 --> 00:34:31,990 S1: looks at how bigger companies make the mistakes talked about above, 596 00:34:31,989 --> 00:34:34,990 S1: and how this founder mode, which he didn't even define it, 597 00:34:35,020 --> 00:34:36,880 S1: he just said it like it needs to be defined 598 00:34:36,880 --> 00:34:41,050 S1: because it's really interesting. But, um, yeah, it basically he 599 00:34:41,050 --> 00:34:43,330 S1: says it allows you to stay in a more innovation 600 00:34:43,330 --> 00:34:46,630 S1: focused mindset. Uh, and great read. All this stuff is 601 00:34:46,630 --> 00:34:49,120 S1: a great read. I just wish he, uh, I wish 602 00:34:49,120 --> 00:34:52,210 S1: he locked it down. Yeah, I understand it's not locked 603 00:34:52,239 --> 00:34:54,310 S1: down yet, but I thought he might have tried to 604 00:34:54,310 --> 00:34:58,150 S1: do that. Oakland police are using Tesla's Sentry Mode footage 605 00:34:58,150 --> 00:35:01,180 S1: to aid crime investigations, and if they can't find the owner, 606 00:35:01,180 --> 00:35:03,370 S1: they just tow the car. They just tow the Tesla 607 00:35:03,400 --> 00:35:06,580 S1: because it has the evidence. Waymo is tackling the skepticism 608 00:35:06,580 --> 00:35:09,819 S1: around its autonomous vehicles by launching a new safety hub 609 00:35:09,820 --> 00:35:12,220 S1: with a whole bunch of data and charts showing that 610 00:35:12,219 --> 00:35:16,060 S1: they're more safe than human drivers. Oh, related story that 611 00:35:16,060 --> 00:35:19,360 S1: just came out is that most of the accidents with 612 00:35:19,360 --> 00:35:23,620 S1: Waymo are from human drivers hitting the Waymo. Not surprising 613 00:35:23,650 --> 00:35:28,420 S1: to me. Joshua Austin's a manifesto manifesto for Radical Simplicity 614 00:35:28,420 --> 00:35:32,080 S1: argues for a streamlined approach to software delivery, ditching subjective 615 00:35:32,080 --> 00:35:35,650 S1: metrics like story points in favor of focusing on real 616 00:35:35,650 --> 00:35:40,450 S1: dependencies and outcomes. Bluetooth 6.0 is here. Oh, by the way, 617 00:35:40,450 --> 00:35:43,629 S1: the new iPhones have Wi-Fi seven. I am upgrading my 618 00:35:43,630 --> 00:35:46,660 S1: whole house to Wi-Fi seven. They say it's about five 619 00:35:46,690 --> 00:35:51,440 S1: times faster, so if I'm getting like 800 down right now, 620 00:35:51,440 --> 00:35:55,280 S1: eight times 544 gigabits. It would be nice if I 621 00:35:55,280 --> 00:35:57,380 S1: got four gigabits. I bet you I'm going to get like, 622 00:35:57,410 --> 00:36:01,340 S1: two gigabits streaming to my phone and then to my 623 00:36:01,340 --> 00:36:04,759 S1: laptop once laptops have Wi-Fi seven. Not sure if my 624 00:36:04,760 --> 00:36:08,630 S1: current ones do. A Caritas phone was snatched in London 625 00:36:08,630 --> 00:36:12,560 S1: and despite tracking it with Find My iPhone, he watched 626 00:36:12,560 --> 00:36:17,480 S1: it basically travel around the city and ended up in Shenzhen. Humans. 627 00:36:17,570 --> 00:36:21,590 S1: President XI Jinping has pledged to create over 1 million 628 00:36:21,590 --> 00:36:26,750 S1: jobs in Africa. I cannot stand seeing Africa become an 629 00:36:26,750 --> 00:36:30,560 S1: extension of China. I do not like it, but it's 630 00:36:30,560 --> 00:36:33,500 S1: pretty hard for the West and anyone in the West 631 00:36:33,500 --> 00:36:37,370 S1: to even notice this, given our history, which is not great. 632 00:36:37,370 --> 00:36:40,370 S1: The question is, how long will we let that guilt 633 00:36:40,400 --> 00:36:44,390 S1: be an obstacle to opposing China wherever they go and 634 00:36:44,390 --> 00:36:49,250 S1: basically do colonialism, which is bad, but hard to talk 635 00:36:49,280 --> 00:36:51,770 S1: about how bad it is when you already did it 636 00:36:51,770 --> 00:36:54,410 S1: and got scolded for it, and now you feel bad 637 00:36:54,410 --> 00:36:56,719 S1: about it, and now you're pointing the finger at someone 638 00:36:56,719 --> 00:36:59,150 S1: else doing it. It feels kind of weird, but at 639 00:36:59,150 --> 00:37:02,600 S1: the same time, it also feels weird and cowardly to 640 00:37:02,630 --> 00:37:05,360 S1: not do anything about it just because you feel guilty. 641 00:37:05,570 --> 00:37:08,930 S1: A whole bunch of right wing influencers receive millions of 642 00:37:08,930 --> 00:37:13,040 S1: dollars from Russia in return for promoting pro-Russian talking points, 643 00:37:13,040 --> 00:37:16,610 S1: which is hilarious to me since their whole narrative is like, 644 00:37:16,610 --> 00:37:21,140 S1: I don't believe anything. I'm skeptical. I'm discerning. Except when 645 00:37:21,140 --> 00:37:24,830 S1: it comes to obvious Russian propaganda. Here's another way to 646 00:37:24,860 --> 00:37:27,410 S1: think about it. And I did some Intel in the Army. 647 00:37:27,410 --> 00:37:31,310 S1: I'm not, like heavily Intel trained, but I did have 648 00:37:31,310 --> 00:37:33,440 S1: a decent amount of exposure to it and worked in 649 00:37:33,469 --> 00:37:36,020 S1: like the Intel Group or whatever. But think about this. 650 00:37:36,020 --> 00:37:40,880 S1: Here are probably two unrelated phenomenon one. We know for 651 00:37:40,880 --> 00:37:44,360 S1: absolute certain that Russia is trying to use its significant 652 00:37:44,360 --> 00:37:48,110 S1: propaganda capabilities to influence the right wing in the United 653 00:37:48,110 --> 00:37:51,830 S1: States to be pro-Russian into Ukraine that we know for 654 00:37:51,830 --> 00:37:55,760 S1: absolute certain. Right. That's not even debatable. They are really 655 00:37:55,760 --> 00:37:59,779 S1: good at this, and they are spending at least tens 656 00:37:59,780 --> 00:38:02,390 S1: of millions of dollars, probably hundreds of millions of dollars, 657 00:38:02,390 --> 00:38:06,259 S1: probably not billions. So I'm guessing tens of millions to 658 00:38:06,290 --> 00:38:09,410 S1: hundreds of millions of dollars they are spending on doing this. 659 00:38:09,410 --> 00:38:12,230 S1: And they're really good at it because it's essentially like 660 00:38:12,230 --> 00:38:16,160 S1: the evolution of the KGB. Second point, which is probably unrelated. 661 00:38:16,160 --> 00:38:19,790 S1: The right wing in the United States is now completely, 662 00:38:19,790 --> 00:38:23,989 S1: almost completely pro-Russian Anti-ukraine. And some people reached out to 663 00:38:24,020 --> 00:38:26,149 S1: me and they're like, that's not true. That's not true. 664 00:38:26,180 --> 00:38:28,580 S1: I should have went and found some data and linked 665 00:38:28,580 --> 00:38:31,130 S1: to it. I read the data all the time. I 666 00:38:31,130 --> 00:38:34,190 S1: read the data all the time in polling, where it's like, 667 00:38:34,190 --> 00:38:40,280 S1: how much of the Republican Party, when polled, are actually pro-Putin? 668 00:38:40,310 --> 00:38:42,980 S1: The numbers are ridiculous. I'm sorry, I don't have the 669 00:38:42,980 --> 00:38:45,290 S1: link here. I should have put it in there, but 670 00:38:45,320 --> 00:38:48,980 S1: it's it's pretty easy to Google and you'll find it yourself. 671 00:38:48,980 --> 00:38:51,319 S1: It's actually better if you find it yourself than if 672 00:38:51,320 --> 00:38:54,589 S1: I feed you the link. Just Google that. Look for 673 00:38:54,590 --> 00:38:59,180 S1: your look for the stats, the polling. It's ridiculous. And 674 00:38:59,180 --> 00:39:01,729 S1: a brief political aside. So this is an example of 675 00:39:01,730 --> 00:39:04,730 S1: a little breakout piece. So I already know I'm going 676 00:39:04,760 --> 00:39:07,190 S1: to get hate mail about the point above, because I'm 677 00:39:07,219 --> 00:39:09,919 S1: a crazy liberal and I post a lot of other 678 00:39:09,920 --> 00:39:15,290 S1: stuff that's like anti-left far left and their idiocracy or idiocy. 679 00:39:15,290 --> 00:39:18,469 S1: And I get tons of comments about being way too 680 00:39:18,500 --> 00:39:22,640 S1: right from the left because I attack the left and whatever, 681 00:39:22,640 --> 00:39:24,500 S1: and I just now attack the right and they're going 682 00:39:24,530 --> 00:39:27,140 S1: to be like, oh, you're too left wing. So I 683 00:39:27,140 --> 00:39:30,020 S1: asked you to consider another possibility, which is I'm actively 684 00:39:30,020 --> 00:39:34,370 S1: considering each position from first principles. Not perfect. I could 685 00:39:34,370 --> 00:39:36,920 S1: be wrong, but I put a lot of effort into 686 00:39:36,950 --> 00:39:39,230 S1: having my own opinions that are not part of a 687 00:39:39,230 --> 00:39:42,770 S1: tribe of pre-approved options. So perhaps the best way to 688 00:39:42,800 --> 00:39:46,120 S1: sum me up right now is that I'm liberal in 689 00:39:46,120 --> 00:39:50,290 S1: my goals and somewhat conservative in my approach for how 690 00:39:50,290 --> 00:39:53,140 S1: to achieve them. So meaning, I want a planet full 691 00:39:53,140 --> 00:39:57,640 S1: of lots of different colors and ethnicities of people all 692 00:39:57,640 --> 00:40:02,530 S1: thriving together. A secular society that encourages any religion but 693 00:40:02,530 --> 00:40:06,130 S1: doesn't allow any of them to infringe on government or 694 00:40:06,130 --> 00:40:10,629 S1: the ideals listed here. Gender identity, private sexual behavior between 695 00:40:10,630 --> 00:40:16,569 S1: consenting adults, all personal choices and nobody's business. Basically, the 696 00:40:16,570 --> 00:40:19,270 S1: freedom for everyone to strive to be the best versions 697 00:40:19,270 --> 00:40:22,750 S1: of themselves that they can be, and a society that 698 00:40:22,750 --> 00:40:26,830 S1: sees that as simultaneously a matter of personal responsibility, but 699 00:40:26,830 --> 00:40:31,480 S1: also helps people along that path. So free speech, the 700 00:40:31,480 --> 00:40:35,920 S1: ability to offend people with difficult ideas and have those 701 00:40:35,920 --> 00:40:40,210 S1: debates and be controversial and be offensive, right. The concept 702 00:40:40,210 --> 00:40:45,239 S1: of meritocracy, the emphasis on personal responsibility. Okay. And those 703 00:40:45,239 --> 00:40:48,690 S1: are kind of like currently those are considered like right wing. Right. 704 00:40:48,690 --> 00:40:53,220 S1: But also the acknowledgement that some people in groups need 705 00:40:53,250 --> 00:40:57,150 S1: help getting to this point where their personal responsibility can 706 00:40:57,150 --> 00:41:00,239 S1: actually take root and help them thrive, and that it 707 00:41:00,239 --> 00:41:05,100 S1: is society's responsibility to give that to them. And that's 708 00:41:05,100 --> 00:41:07,620 S1: where it's liberal. So in other words, if everyone had 709 00:41:07,620 --> 00:41:11,670 S1: the same opportunity, I'd be fiercely all about meritocracy. But 710 00:41:11,670 --> 00:41:15,300 S1: not everyone has the same opportunity. And that's the role 711 00:41:15,300 --> 00:41:19,830 S1: of society and charity and kindness to help them get 712 00:41:19,830 --> 00:41:23,400 S1: to the place where their hard work can actually benefit them. 713 00:41:23,400 --> 00:41:26,040 S1: So the problem right now, in my mind, is that 714 00:41:26,040 --> 00:41:30,779 S1: the far right and the far left are in opposition 715 00:41:30,780 --> 00:41:33,239 S1: to that model. I just laid out the far right 716 00:41:33,239 --> 00:41:35,970 S1: because they actually want the wrong things, and the far 717 00:41:35,969 --> 00:41:38,280 S1: left because they are so confused about how the world 718 00:41:38,280 --> 00:41:41,250 S1: works that they're causing more harm than good. So anyway, 719 00:41:41,250 --> 00:41:44,010 S1: that's a short version of where I currently stand. And 720 00:41:44,010 --> 00:41:46,529 S1: if you ever get confused about like if I'm crazy 721 00:41:46,530 --> 00:41:50,250 S1: right or crazy left, just go read that. And here's 722 00:41:50,250 --> 00:41:53,610 S1: a really important point. I also consider you to make 723 00:41:53,610 --> 00:41:57,690 S1: your own North Star paragraph or set of bullet points 724 00:41:57,690 --> 00:42:00,300 S1: like I just did, and have that be your thing. 725 00:42:00,300 --> 00:42:04,020 S1: And then realize that this this, the left narrative or 726 00:42:04,020 --> 00:42:08,100 S1: the right narrative or centrist narrative with all these different narratives, 727 00:42:08,100 --> 00:42:10,200 S1: they don't really matter. What matters is what is your 728 00:42:10,230 --> 00:42:12,840 S1: North Star look like? Like I just laid out. And 729 00:42:12,840 --> 00:42:15,509 S1: then what you'll find is, like all these different narratives 730 00:42:15,510 --> 00:42:19,230 S1: and tribes and everything they might have. Venn diagram overlaps 731 00:42:19,230 --> 00:42:21,810 S1: with some part of your thing, but they shouldn't be 732 00:42:21,810 --> 00:42:25,320 S1: your thing. If you have a perfect Venn diagram, overlap 733 00:42:25,320 --> 00:42:29,670 S1: with the extreme right or the extreme left, you don't 734 00:42:29,670 --> 00:42:33,210 S1: have your own opinions. You are using someone else's opinions 735 00:42:33,210 --> 00:42:36,660 S1: and that's a problem. So North Star plus First principles 736 00:42:36,660 --> 00:42:39,779 S1: is far better than picking a tribe and endorsing everything 737 00:42:39,780 --> 00:42:42,989 S1: they say. All right. Sweden's health authority has issued new 738 00:42:42,989 --> 00:42:47,940 S1: guidelines advising the children under two should have no screen time, 739 00:42:47,940 --> 00:42:51,120 S1: while teenagers should be limited to three hours a day. 740 00:42:51,150 --> 00:42:53,340 S1: A lot of people are starting to say, and this 741 00:42:53,340 --> 00:42:55,920 S1: is supported by a bunch of different studies, that exercise 742 00:42:55,920 --> 00:42:58,950 S1: could be the most potent medical intervention that we know of. 743 00:42:58,950 --> 00:43:01,800 S1: And I got a buddy who's a cardiologist named Jonathan, 744 00:43:01,800 --> 00:43:06,330 S1: and he agrees with this. David Brooks discusses TEDx Goias. 745 00:43:06,360 --> 00:43:10,800 S1: Gioia's essay on the decline of American culture, where art 746 00:43:10,800 --> 00:43:14,400 S1: is overshadowed by entertainment. And now even entertainment is being 747 00:43:14,400 --> 00:43:18,480 S1: consumed by distraction from platforms like TikTok and Instagram. Got 748 00:43:18,480 --> 00:43:22,560 S1: a photographer documenting life and beauty of America's last old 749 00:43:22,560 --> 00:43:25,890 S1: growth forests? Got an article here that explores the belief 750 00:43:25,890 --> 00:43:27,959 S1: that there's a place for everyone, and suggesting that every 751 00:43:27,960 --> 00:43:32,430 S1: person has a unique purpose and value. Marco Giancotti argues 752 00:43:32,430 --> 00:43:35,969 S1: that with millions of books available, only select a few 753 00:43:36,030 --> 00:43:40,109 S1: what he calls damn good books, which are truly life changing. 754 00:43:40,110 --> 00:43:42,450 S1: And these are the books that transform you and not 755 00:43:42,450 --> 00:43:45,870 S1: to get bogged down with, like, every book out there. 756 00:43:45,989 --> 00:43:48,060 S1: And this goes along with the idea of like, you 757 00:43:48,060 --> 00:43:50,700 S1: should be willing to put down bad books, just put 758 00:43:50,700 --> 00:43:52,859 S1: it down. Go on to a better one. Phoenix just 759 00:43:52,860 --> 00:43:57,000 S1: hit 100 consecutive days of 100 degree heat, which beats 760 00:43:57,000 --> 00:44:01,080 S1: the record that was set in 1993. And discovery lmao 761 00:44:01,080 --> 00:44:06,300 S1: swsh a bash wrapper around Python's ML whisper. It's basically 762 00:44:06,300 --> 00:44:10,110 S1: a really fast whisper for Mac. NN term lets you 763 00:44:10,110 --> 00:44:13,740 S1: browse Hacker News from your terminal and dungeon dash a 764 00:44:13,739 --> 00:44:18,300 S1: remote know a command line RPG where you dive into dungeons, 765 00:44:18,300 --> 00:44:21,990 S1: battle enemies, and collect loot to up level and become 766 00:44:21,989 --> 00:44:24,750 S1: the ultimate hero. I want to play this a whole 767 00:44:24,750 --> 00:44:27,000 S1: bunch more, and I actually want to make some of these. 768 00:44:27,030 --> 00:44:30,060 S1: These these just seem super fun and I love the terminal. 769 00:44:30,060 --> 00:44:33,719 S1: If you didn't already know that NSA's national Cryptographic school 770 00:44:33,760 --> 00:44:39,370 S1: television catalog from 1991 has surfaced, listing around 600 training 771 00:44:39,370 --> 00:44:43,569 S1: videos on CommSec and Sigint. I love these treasure troves 772 00:44:43,570 --> 00:44:47,020 S1: of old things. When I was in the Army, I 773 00:44:47,020 --> 00:44:49,630 S1: got to work in a library for a certain amount 774 00:44:49,630 --> 00:44:52,570 S1: of time, and I got all the Special Forces manuals 775 00:44:52,570 --> 00:44:55,660 S1: and all the CommSec and Sigint and stuff like that, 776 00:44:55,660 --> 00:44:59,469 S1: and I would just sit and read these things, like voraciously. Okay. 777 00:44:59,500 --> 00:45:02,860 S1: List of ideas. Extraordinarily simple template you could use to 778 00:45:02,890 --> 00:45:05,860 S1: orient your life. Okay. I love this thing. I believe 779 00:45:05,860 --> 00:45:08,380 S1: one of the biggest issues in the world is this, 780 00:45:08,410 --> 00:45:11,529 S1: which is why I'm looking to solve it using this, 781 00:45:11,560 --> 00:45:15,010 S1: which is why I'm doing these projects. And I'm measuring 782 00:45:15,010 --> 00:45:19,360 S1: my success using these metrics. This is how to introduce 783 00:45:19,360 --> 00:45:22,600 S1: yourself at a party. This is potentially how to introduce 784 00:45:22,600 --> 00:45:25,300 S1: yourself to a mate. This is how to explain yourself 785 00:45:25,300 --> 00:45:29,049 S1: in an interview. It's your LinkedIn profile. Look at this. 786 00:45:29,050 --> 00:45:31,959 S1: I believe this is the biggest problem which I think 787 00:45:31,960 --> 00:45:34,240 S1: we should solve by doing this, which is why I'm 788 00:45:34,239 --> 00:45:39,100 S1: doing this work which I am measuring using these metrics. 789 00:45:39,100 --> 00:45:43,000 S1: That is purpose, that is direction, and it explains it 790 00:45:43,000 --> 00:45:46,150 S1: to others. And it's done using a simple sentence like, 791 00:45:46,180 --> 00:45:48,969 S1: you could turn this into one complex sentence. And if 792 00:45:48,969 --> 00:45:52,060 S1: you can answer this, I would argue you're better off 793 00:45:52,060 --> 00:45:54,550 S1: than most people in the workforce and most people on 794 00:45:54,550 --> 00:45:56,230 S1: the planet. The more I think about it, the more 795 00:45:56,230 --> 00:46:00,190 S1: I think major career for creators going forward will be 796 00:46:00,190 --> 00:46:04,210 S1: building entire realities for people to live inside of. So 797 00:46:04,239 --> 00:46:08,710 S1: think post, AGI or ASI and post UBI and where 798 00:46:08,739 --> 00:46:11,770 S1: games are extraordinarily immersive. So I think there will be 799 00:46:11,770 --> 00:46:15,489 S1: a huge market for creative people, creative people building the 800 00:46:15,489 --> 00:46:19,300 S1: storylines and stat systems that look and feel like entire 801 00:46:19,300 --> 00:46:22,120 S1: worlds that people will live inside of. So imagine how 802 00:46:22,120 --> 00:46:25,660 S1: everyone loved Game of Thrones for all those years, except 803 00:46:25,660 --> 00:46:28,600 S1: like you're actually one of the characters. In fact, you're 804 00:46:28,600 --> 00:46:32,140 S1: one of the heroes and you're navigating through the the difficulty. Right? 805 00:46:32,170 --> 00:46:34,450 S1: Or it could be like a superhero themed thing or 806 00:46:34,450 --> 00:46:38,230 S1: like a romance themed thing or like fantasy or whatever 807 00:46:38,230 --> 00:46:41,560 S1: the structure is. So imagine basically your favorite book or 808 00:46:41,560 --> 00:46:45,310 S1: TV show or series or movie or whatever. We basically 809 00:46:45,310 --> 00:46:50,260 S1: build fully immersive realities based on those for people to 810 00:46:50,290 --> 00:46:54,400 S1: spend their time inside of. And the problem is, it 811 00:46:54,400 --> 00:46:57,219 S1: can't all be entertainment because people need a purpose. They 812 00:46:57,219 --> 00:46:59,469 S1: need some sort of drive. I think we're going to 813 00:46:59,469 --> 00:47:03,430 S1: find ways to create that purpose inside of these realities. 814 00:47:03,430 --> 00:47:05,800 S1: I have a piece from I need to go dig 815 00:47:05,800 --> 00:47:09,550 S1: this thing up. It's from like 2002. And I basically 816 00:47:09,550 --> 00:47:12,850 S1: said that the games of the future will be people 817 00:47:12,880 --> 00:47:16,960 S1: going in and doing the work and the jobs inside there. 818 00:47:16,960 --> 00:47:19,299 S1: So you'll go in there and you'll be a cop, okay? 819 00:47:19,330 --> 00:47:21,430 S1: And you'll see someone coming down the thing and the 820 00:47:21,430 --> 00:47:24,100 S1: music is too loud, or they're beating somebody up, and 821 00:47:24,100 --> 00:47:28,330 S1: you'll go and actually engage in a fight with this person, um, 822 00:47:28,360 --> 00:47:30,819 S1: to try to restrain them or whatever. Or maybe you're 823 00:47:30,850 --> 00:47:32,950 S1: the bad guy and you're just going to go beat 824 00:47:32,950 --> 00:47:35,680 S1: somebody up, and it's someone else's job to be the 825 00:47:35,680 --> 00:47:38,649 S1: good guy, and they're actually paid or rewarded in some 826 00:47:38,650 --> 00:47:41,410 S1: sort of way. So other people inside the game can 827 00:47:41,410 --> 00:47:45,790 S1: give them something, or the system can reward them in 828 00:47:45,790 --> 00:47:48,220 S1: a way that's good on the outside as well. Now, 829 00:47:48,219 --> 00:47:52,989 S1: in most scenarios, your house and your game subscription and 830 00:47:52,989 --> 00:47:55,930 S1: your game gear and everything is being paid by UBI 831 00:47:55,930 --> 00:47:58,330 S1: in this world that I'm thinking of, most of the 832 00:47:58,330 --> 00:48:01,120 S1: jobs are gone and people are being paid essentially to 833 00:48:01,150 --> 00:48:05,860 S1: live inside these realities, which is massively dystopian. I want 834 00:48:05,890 --> 00:48:09,100 S1: to be very clear about that, if that is the 835 00:48:09,100 --> 00:48:12,190 S1: only option. Okay. What I'm saying is that when the 836 00:48:12,190 --> 00:48:14,380 S1: jobs go away, there will be a lot of people 837 00:48:14,380 --> 00:48:17,800 S1: who want to live these lives because we are creative. 838 00:48:17,830 --> 00:48:21,730 S1: We read these books, we see these movies. We we 839 00:48:21,730 --> 00:48:24,700 S1: know that that's on offer. That's the thing we could 840 00:48:24,700 --> 00:48:27,560 S1: potentially have. So if games make it where we can 841 00:48:27,560 --> 00:48:30,140 S1: actually have it, we can actually be that hero. We 842 00:48:30,140 --> 00:48:32,330 S1: will want to do it, but we won't want to 843 00:48:32,360 --> 00:48:35,719 S1: do it. It will be empty and weak and shallow 844 00:48:35,719 --> 00:48:40,759 S1: and lame and horribly destructive. If it's nothing but just 845 00:48:40,760 --> 00:48:44,900 S1: like violence and porn, which there definitely will be, that right? 846 00:48:44,930 --> 00:48:47,239 S1: There will be games that are just nothing but that. 847 00:48:47,239 --> 00:48:50,810 S1: But it's all entertainment. It's empty. It's hollow, it's nothing. 848 00:48:50,810 --> 00:48:54,680 S1: The better version of this, which I hope happens and 849 00:48:54,680 --> 00:48:58,669 S1: definitely could happen with, like if we're wielding ASI or 850 00:48:58,700 --> 00:49:01,190 S1: if ASI takes over and it's doing this in a 851 00:49:01,190 --> 00:49:05,600 S1: benign way, it would build meaning into the system, but 852 00:49:05,600 --> 00:49:09,230 S1: it wouldn't require that you do it. You could also 853 00:49:09,230 --> 00:49:14,149 S1: take off the gear, put it aside, go outside, start 854 00:49:14,150 --> 00:49:18,890 S1: a community, start a farm. Like, uh, till the land, 855 00:49:18,920 --> 00:49:22,340 S1: toe the land. What's the name? What's the verb? Um, 856 00:49:22,340 --> 00:49:25,710 S1: you know, till I think it's till anyway you tend 857 00:49:25,710 --> 00:49:29,430 S1: to the land, you grow crops, you make food, people 858 00:49:29,430 --> 00:49:32,280 S1: eat the food. You have families, you have kids like. 859 00:49:32,280 --> 00:49:35,940 S1: So imagine this. There are societies that's kind of like Amish. 860 00:49:35,940 --> 00:49:38,670 S1: They don't play the games, so they're kind of Amish. 861 00:49:38,790 --> 00:49:42,540 S1: And maybe there's like more advanced technological versions of Amish 862 00:49:42,540 --> 00:49:45,900 S1: where it's like, I don't know, the 1950s or something, right? 863 00:49:45,930 --> 00:49:48,840 S1: In this ideal kind of world, except for it's everywhere 864 00:49:48,840 --> 00:49:53,040 S1: and everyone can participate. I would argue that the actual 865 00:49:53,040 --> 00:49:54,840 S1: reality is going to look like there's going to be 866 00:49:54,840 --> 00:49:58,469 S1: some communities like that, some communities that are super Amish, 867 00:49:58,469 --> 00:50:01,529 S1: they're like Ted Kaczynski, like tech is bad and they 868 00:50:01,530 --> 00:50:04,950 S1: just do their own thing. Government should leave them alone. 869 00:50:04,950 --> 00:50:07,920 S1: They are productive members of society. They have their own schools. 870 00:50:07,920 --> 00:50:10,080 S1: They have their own religion. They have their own stuff. 871 00:50:10,110 --> 00:50:12,839 S1: Like they're making their own food. Leave them alone. Everything's fine. 872 00:50:12,840 --> 00:50:14,640 S1: They don't have to play the games. And a bunch 873 00:50:14,640 --> 00:50:17,160 S1: of people like maybe some of their kids, and that'll 874 00:50:17,190 --> 00:50:18,870 S1: be bad, but a bunch of people are going to 875 00:50:18,870 --> 00:50:20,490 S1: be like, are you kidding me? I want to be 876 00:50:20,489 --> 00:50:23,370 S1: a superhero. I want to be a princess. I want 877 00:50:23,400 --> 00:50:25,169 S1: to be, you know, I want to take over the world. 878 00:50:25,170 --> 00:50:27,780 S1: I want to do all these things. Fine. Go find 879 00:50:27,780 --> 00:50:31,290 S1: your meaning inside the game system. Importantly here, there is 880 00:50:31,290 --> 00:50:33,690 S1: going to be a period of time in which we're 881 00:50:33,690 --> 00:50:36,900 S1: going to need something like this as a transition period, 882 00:50:37,020 --> 00:50:39,870 S1: because so many people are going to lose their meaning, 883 00:50:39,870 --> 00:50:41,370 S1: they're going to have to go into this thing. I 884 00:50:41,370 --> 00:50:43,710 S1: think there's going to be a requirement for an ASI, 885 00:50:43,739 --> 00:50:46,500 S1: for a government, for something to be like, look, we 886 00:50:46,500 --> 00:50:49,110 S1: need to give out money and we need to give 887 00:50:49,110 --> 00:50:57,750 S1: them something awesome. I think game companies, whoever it is, Ubisoft, Microsoft, Sony, 888 00:50:57,780 --> 00:51:00,120 S1: all these people, they're actually going to get hired by 889 00:51:00,120 --> 00:51:04,049 S1: the government to create these giant scale games that people 890 00:51:04,050 --> 00:51:06,840 S1: can live inside of. There's going to be haptic gear 891 00:51:06,870 --> 00:51:09,839 S1: they think of like a military push to invent the 892 00:51:09,840 --> 00:51:13,290 S1: bomb or whatever, but the military push is going to 893 00:51:13,290 --> 00:51:17,250 S1: be to invent really immersive technology because we need it, 894 00:51:17,250 --> 00:51:20,400 S1: because guess what? It is actually a military concern. It 895 00:51:20,400 --> 00:51:24,919 S1: is actually a military level requirement that we do not 896 00:51:24,920 --> 00:51:29,300 S1: have hundreds of millions of people, also known as billions 897 00:51:29,300 --> 00:51:33,380 S1: of people with no meaning and no money, because they 898 00:51:33,380 --> 00:51:35,840 S1: will in fact light everything on fire and that will 899 00:51:35,870 --> 00:51:38,239 S1: be bad. So there will be a period where we 900 00:51:38,239 --> 00:51:41,870 S1: actually need to invent something like this to get us 901 00:51:41,870 --> 00:51:46,340 S1: through this period. That kind of calms everyone down. But 902 00:51:46,340 --> 00:51:49,879 S1: if you don't do it well, they will be satiated 903 00:51:49,880 --> 00:51:53,090 S1: and they will be not violent, but it will start 904 00:51:53,090 --> 00:51:56,390 S1: to rot. It will rot the human soul. Because what 905 00:51:56,390 --> 00:51:58,700 S1: are they doing for creativity? What are they making? What 906 00:51:58,700 --> 00:52:01,310 S1: are they building? Are they building? Are they making families? 907 00:52:01,310 --> 00:52:05,180 S1: Who's making the kids? Right? I mean, the games didn't 908 00:52:05,180 --> 00:52:08,240 S1: solve aging. So a bunch of people hooked into the games, 909 00:52:08,239 --> 00:52:11,870 S1: they don't meet anyone. And two generations later, we don't 910 00:52:11,870 --> 00:52:14,600 S1: have any people. Or actually, technically, I think everyone dies 911 00:52:14,600 --> 00:52:18,109 S1: in one generation, right? If you're not reproducing. So like 912 00:52:18,140 --> 00:52:22,050 S1: that is only a temporary solution to not have the 913 00:52:22,050 --> 00:52:24,330 S1: world blow up. But ultimately you have to solve the 914 00:52:24,330 --> 00:52:28,200 S1: meaning problem, which either happens outside the game, which I 915 00:52:28,200 --> 00:52:31,290 S1: think that is going to happen and should happen, or 916 00:52:31,290 --> 00:52:34,680 S1: inside the game. And I think what's most likely in 917 00:52:34,680 --> 00:52:38,610 S1: a benign situation is that you'll have both recommendation of 918 00:52:38,610 --> 00:52:42,299 S1: the week. I've been a bit obsessed with the problem 919 00:52:42,300 --> 00:52:45,839 S1: definition lately, like I talked about with the fabric pattern. 920 00:52:45,840 --> 00:52:49,380 S1: And here's my recommendation get really good at articulating and 921 00:52:49,380 --> 00:52:53,880 S1: prioritizing your problems, like write them out in vast detail. 922 00:52:53,910 --> 00:52:57,900 S1: Make yourself an expert in your problems. It takes away 923 00:52:57,900 --> 00:53:00,750 S1: their power. Kind of like staring directly at anger when 924 00:53:00,750 --> 00:53:03,540 S1: you're meditating. Like it takes the spin off the anger. 925 00:53:03,540 --> 00:53:05,550 S1: If you look away, the anger hurts you again. You 926 00:53:05,550 --> 00:53:07,680 S1: look directly at the anger and you're like, hmm, that's 927 00:53:07,680 --> 00:53:10,049 S1: not so bad. Happens the same with your problems. They 928 00:53:10,050 --> 00:53:14,070 S1: become approachable, and this also happens to be the key 929 00:53:14,070 --> 00:53:17,910 S1: to brilliant I prompting because it's an extension of like 930 00:53:17,910 --> 00:53:21,359 S1: know yourself and then you can articulate yourself and the 931 00:53:21,360 --> 00:53:23,940 S1: aphorism of the week when I have one week to 932 00:53:23,969 --> 00:53:28,200 S1: solve a seemingly impossible problem, I spend six days defining it, 933 00:53:28,200 --> 00:53:31,650 S1: and then the solution becomes obvious. When I have one 934 00:53:31,650 --> 00:53:35,550 S1: week to solve a seemingly impossible problem, I spend six 935 00:53:35,550 --> 00:53:41,400 S1: days defining it, and then the solution becomes obvious. Albert Einstein. 936 00:53:42,719 --> 00:53:45,960 S1: Unsupervised learning is produced and edited by Daniel Miessler on 937 00:53:45,960 --> 00:53:50,520 S1: a Neumann U87 AI microphone using Hindenburg. Intro and outro 938 00:53:50,550 --> 00:53:53,879 S1: music is by zombie with a Y. And to get 939 00:53:53,880 --> 00:53:55,980 S1: the text and links from this episode, sign up for 940 00:53:55,980 --> 00:54:00,450 S1: the newsletter version of the show at Daniel missler.com/newsletter. 941 00:54:01,440 --> 00:54:02,490 UU: We'll see you next time.