1 00:00:01,129 --> 00:00:04,460 S1: Welcome to Unsupervised Learning, a security, AI and meaning focused 2 00:00:04,460 --> 00:00:07,310 S1: podcast that looks at how best to thrive as humans 3 00:00:07,310 --> 00:00:11,540 S1: in a post AI world. It combines original ideas, analysis, 4 00:00:11,570 --> 00:00:14,780 S1: and mental models to bring not just the news, but 5 00:00:14,780 --> 00:00:16,730 S1: why it matters and how to respond. 6 00:00:21,710 --> 00:00:25,730 S2: So hello everyone, welcome to this AI at Ypo session. 7 00:00:25,730 --> 00:00:28,550 S2: We're delighted to have Daniel Miessler here with us today 8 00:00:28,550 --> 00:00:31,100 S2: to speak on the skills and mental frames required to 9 00:00:31,130 --> 00:00:34,580 S2: thrive in an AI world. Daniel Miessler is the founder 10 00:00:34,580 --> 00:00:37,940 S2: of Unsupervised Learning, which is a company that focuses on 11 00:00:37,940 --> 00:00:42,620 S2: building products that help companies, organizations, and people identify, articulate, 12 00:00:42,620 --> 00:00:45,710 S2: and execute on their purpose in the world. Daniel has 13 00:00:45,710 --> 00:00:49,159 S2: over 20 years experience in cybersecurity, and has spent the 14 00:00:49,159 --> 00:00:52,310 S2: last several years focused on applying AI to business and 15 00:00:52,310 --> 00:00:58,310 S2: human problems. Daniel has held senior positions at Apple, Robinhood, Ioactive, HPE, 16 00:00:58,310 --> 00:01:01,510 S2: and many other companies, as well as consulted for or 17 00:01:01,510 --> 00:01:04,600 S2: been embedded in hundreds of others in the fortune 500 18 00:01:04,600 --> 00:01:07,660 S2: and global 1000. So thank you very much for being 19 00:01:07,660 --> 00:01:10,030 S2: here with us today, Daniel. The floor is now yours. 20 00:01:10,060 --> 00:01:14,080 S1: All right. Yeah. Thank you for having me. So what 21 00:01:14,080 --> 00:01:16,119 S1: I want to talk about today is what I believe 22 00:01:16,150 --> 00:01:19,780 S1: is coming for us with AI, and what we can 23 00:01:19,780 --> 00:01:22,660 S1: do to get ready for it. So I want to 24 00:01:22,660 --> 00:01:27,520 S1: start with some background, some analysis and predictions, and then 25 00:01:27,550 --> 00:01:31,150 S1: give some recommendations on how to get ready for this 26 00:01:31,150 --> 00:01:33,850 S1: thing that I believe is coming. So I don't normally 27 00:01:33,850 --> 00:01:39,039 S1: do intro slides because I feel like I would rather 28 00:01:39,040 --> 00:01:41,290 S1: just get into the ideas, but in this case, I'm 29 00:01:41,290 --> 00:01:43,540 S1: going to make a little bit of an exception, because 30 00:01:43,540 --> 00:01:45,730 S1: I think it's important to see like the background of 31 00:01:45,730 --> 00:01:48,610 S1: where this is coming from. So I won't be sharing 32 00:01:48,610 --> 00:01:50,860 S1: any CV stuff, but just want to talk about some 33 00:01:50,860 --> 00:01:53,860 S1: stuff that I think is relevant. So I've been thinking 34 00:01:53,860 --> 00:01:58,550 S1: and writing online about things in the future since around 1999, 35 00:01:58,550 --> 00:02:04,340 S1: and started a podcast about security called Unsupervised Learning in 2015, 36 00:02:04,340 --> 00:02:06,230 S1: and I wrote a short book on what I saw 37 00:02:06,260 --> 00:02:10,490 S1: as being the future in around 2016. And in that book, 38 00:02:10,490 --> 00:02:15,530 S1: I basically talked about a future of digital assistants, basically 39 00:02:15,530 --> 00:02:19,460 S1: helping us as humans, a world made up of APIs, 40 00:02:19,460 --> 00:02:24,620 S1: and basically having those digital assistants mediate the world for us, 41 00:02:24,650 --> 00:02:29,840 S1: be like a broker in between us and that API world. 42 00:02:29,840 --> 00:02:32,870 S1: And I think this is already starting to happen now. 43 00:02:32,870 --> 00:02:37,370 S1: So I studied ML pretty deeply in 2017, especially Andrew 44 00:02:37,400 --> 00:02:40,220 S1: Ng's full course, which I still recommend. I believe he 45 00:02:40,220 --> 00:02:44,149 S1: has a Coursera version now as well, but the the 46 00:02:44,150 --> 00:02:47,269 S1: YouTube version is quite good. I then worked at Apple 47 00:02:47,270 --> 00:02:50,780 S1: in 2018 for a few years, starting off in a 48 00:02:50,780 --> 00:02:55,070 S1: machine learning group there, and ended up building another product 49 00:02:55,070 --> 00:02:58,900 S1: there in the security group, and I've been continuing on 50 00:02:58,900 --> 00:03:03,010 S1: this path since like the middle of 22. I basically 51 00:03:03,010 --> 00:03:07,630 S1: went full time into this. So, uh, I basically see 52 00:03:07,660 --> 00:03:11,620 S1: two problems in the world, which is humans losing a 53 00:03:11,620 --> 00:03:15,250 S1: sense of meaning and AI making that much, much worse. 54 00:03:15,250 --> 00:03:18,579 S1: So what I'm essentially doing now is trying to help 55 00:03:18,610 --> 00:03:21,460 S1: solve that. So the point of all this is not 56 00:03:21,460 --> 00:03:24,970 S1: to say that I'm right about everything that I'm about 57 00:03:24,970 --> 00:03:27,520 S1: to share, but rather that I've been thinking about it 58 00:03:27,520 --> 00:03:30,160 S1: for quite some time. Um, another good thing to mention 59 00:03:30,160 --> 00:03:33,700 S1: here is, uh, anything about predictions is like, it's, uh, 60 00:03:33,730 --> 00:03:36,850 S1: you shouldn't really believe them because it's really hard to 61 00:03:36,850 --> 00:03:40,720 S1: predict the future. Especially, yeah, when you're talking about a 62 00:03:40,720 --> 00:03:44,020 S1: thing that hasn't happened yet. Um, so I'll talk with 63 00:03:44,020 --> 00:03:47,170 S1: a decent amount of confidence here, but, um, it's really 64 00:03:47,170 --> 00:03:50,410 S1: because if you hedge all the time, it's not really compelling. 65 00:03:50,410 --> 00:03:53,020 S1: And I do add caveats a lot. So a couple 66 00:03:53,050 --> 00:03:55,920 S1: of things I won't be talking about. I have other 67 00:03:55,920 --> 00:04:00,420 S1: presentations on existential risk from AI, and lots to say 68 00:04:00,420 --> 00:04:05,580 S1: about AGI and superintelligence specifically, as well as some thoughts 69 00:04:05,580 --> 00:04:08,340 S1: on which countries might do what or whatever, but those 70 00:04:08,340 --> 00:04:10,170 S1: are big topics on their own, so I won't be 71 00:04:10,170 --> 00:04:12,780 S1: talking about those here. With that out of the way, 72 00:04:12,990 --> 00:04:15,780 S1: I want to jump in. So what I want to 73 00:04:15,780 --> 00:04:22,080 S1: talk about is, um, how I see AI affecting companies, uh, startups, 74 00:04:22,230 --> 00:04:29,490 S1: the reduction of creative friction, hiring, broadcasting oneself, and an 75 00:04:29,490 --> 00:04:33,780 S1: unfortunate separation between groups of people as a result of this. 76 00:04:33,810 --> 00:04:35,940 S1: And for each of these, I'm going to start with 77 00:04:35,940 --> 00:04:39,690 S1: an observation and then give a prediction. So the first 78 00:04:39,690 --> 00:04:43,650 S1: major concept is a new way of thinking about companies. 79 00:04:43,650 --> 00:04:47,400 S1: And I think AI is going to magnify this effect 80 00:04:47,400 --> 00:04:50,280 S1: quite a bit. So one of the major uses of 81 00:04:50,279 --> 00:04:55,790 S1: AI will be to optimize processes within businesses. And a 82 00:04:55,790 --> 00:05:00,260 S1: company I would argue is really just components of action 83 00:05:00,260 --> 00:05:04,100 S1: combined with process. So what's going to happen is consultancies 84 00:05:04,100 --> 00:05:06,800 S1: are about to come in and basically turn your business 85 00:05:06,800 --> 00:05:09,799 S1: into a diagram similar to like what you see over 86 00:05:09,800 --> 00:05:12,770 S1: here on the right. And I've got a full posting 87 00:05:12,770 --> 00:05:15,680 S1: on this which you can go read. It's called companies 88 00:05:15,680 --> 00:05:19,010 S1: are just a graph of algorithms. And when a company 89 00:05:19,010 --> 00:05:22,040 S1: does this, when they make your company visible to you 90 00:05:22,040 --> 00:05:25,670 S1: in this way, AI thrives off of that, right? Once 91 00:05:25,700 --> 00:05:28,549 S1: I can understand what your business can do, it can 92 00:05:28,550 --> 00:05:33,140 S1: then start making some really powerful predictions. Um, and that's 93 00:05:33,140 --> 00:05:36,349 S1: that's exactly what I think is going to happen. And really, 94 00:05:36,350 --> 00:05:42,080 S1: really crazy stat here. 40% of McKinsey's business in 2024 95 00:05:42,110 --> 00:05:45,470 S1: is already AI essentially doing this kind of thing. So 96 00:05:45,470 --> 00:05:48,380 S1: the prediction here is that AI will find massive waste 97 00:05:48,380 --> 00:05:54,210 S1: in most companies, middle management and bureaucracies will come under 98 00:05:54,210 --> 00:05:59,850 S1: extreme pressure and companies will get smaller and far more efficient. 99 00:05:59,850 --> 00:06:02,970 S1: And the good news is how this will affect startups 100 00:06:02,970 --> 00:06:05,850 S1: and moving on to startups. A lot of people are 101 00:06:05,850 --> 00:06:09,420 S1: talking about the concept of the first one person billion 102 00:06:09,450 --> 00:06:12,870 S1: dollar company, which I think is a really crazy concept. 103 00:06:12,990 --> 00:06:15,240 S1: Just as we saw the efficiency that you can get 104 00:06:15,240 --> 00:06:17,550 S1: in a very large company, we're going to see that 105 00:06:17,550 --> 00:06:19,710 S1: in tiny companies as well. In fact, it's going to 106 00:06:19,710 --> 00:06:22,380 S1: be more magnified in tiny companies because you're not dealing 107 00:06:22,380 --> 00:06:27,570 S1: with the the previous technical debt. So there will soon 108 00:06:27,600 --> 00:06:31,740 S1: be many, many AI based services for doing every aspect 109 00:06:31,740 --> 00:06:35,310 S1: of a business which will allow a single person to 110 00:06:35,339 --> 00:06:39,090 S1: essentially launch the idea. The paperwork can be done, the 111 00:06:39,089 --> 00:06:43,140 S1: customer service can be done, sales, marketing, analytics, optimization, all 112 00:06:43,140 --> 00:06:46,380 S1: these different things can just be like AI services that 113 00:06:46,380 --> 00:06:49,880 S1: they add on, and that's going to allow very small 114 00:06:49,880 --> 00:06:53,270 S1: teams to do really powerful things. So that's the prediction 115 00:06:53,270 --> 00:06:57,890 S1: here is the friction to creating a business falls dramatically. 116 00:06:57,890 --> 00:07:00,680 S1: So there'll be a lot more 1 to 10 person 117 00:07:00,680 --> 00:07:03,920 S1: companies that don't really have to grow larger than that. 118 00:07:03,920 --> 00:07:07,130 S1: And I think the result of this is going to 119 00:07:07,130 --> 00:07:11,690 S1: be an extraordinary multiplication of global productivity. Hard to know 120 00:07:11,690 --> 00:07:14,720 S1: how much exactly that will be, but I think five 121 00:07:14,750 --> 00:07:17,900 S1: X is probably safe in like the next ten years. 122 00:07:17,900 --> 00:07:20,450 S1: So one of the most exciting things to me is 123 00:07:20,450 --> 00:07:24,350 S1: the reduction of creative friction. So if you think about 124 00:07:24,380 --> 00:07:27,020 S1: like how many Steven Spielberg's there are in the world 125 00:07:27,020 --> 00:07:31,250 S1: right now, the number is tiny, right? Very, very few. 126 00:07:31,280 --> 00:07:36,980 S1: But if you only factor in the combination of the 127 00:07:36,980 --> 00:07:40,760 S1: genius with the fact that they're in LA and they 128 00:07:40,760 --> 00:07:43,250 S1: know the people in Hollywood and they have access to 129 00:07:43,280 --> 00:07:46,270 S1: all this money and they have access to all the connections. 130 00:07:46,270 --> 00:07:48,940 S1: That's the reason the number is low. So the question 131 00:07:48,940 --> 00:07:51,430 S1: is how many more Spielbergs are there on the entire 132 00:07:51,430 --> 00:07:55,210 S1: planet with as good or better ideas who simply aren't 133 00:07:55,210 --> 00:07:57,910 S1: in LA or don't have access to these resources? Or 134 00:07:57,910 --> 00:07:59,500 S1: maybe they don't know how to draw, or they don't 135 00:07:59,500 --> 00:08:01,720 S1: know how to write a script or something like that. 136 00:08:01,720 --> 00:08:04,270 S1: And I think there's a lot of opportunity there. So 137 00:08:04,270 --> 00:08:07,600 S1: what I see happening, starting very soon is basically an 138 00:08:07,600 --> 00:08:11,590 S1: explosion of new Spielberg's. So I will be able to 139 00:08:11,590 --> 00:08:16,360 S1: help the people with the ideas, but they have one 140 00:08:16,360 --> 00:08:20,140 S1: or more bottlenecks. You know, maybe they're in sub-Saharan Africa, 141 00:08:20,170 --> 00:08:25,270 S1: maybe they're in, you know, some apartment in Scotland, and 142 00:08:25,270 --> 00:08:27,580 S1: they just don't have access to this stuff. And they 143 00:08:27,580 --> 00:08:31,060 S1: didn't really get formally trained in these different skills. So 144 00:08:31,060 --> 00:08:35,470 S1: we're talking about an explosion of books, film, poetry. Artists 145 00:08:35,470 --> 00:08:39,370 S1: will just be able to talk to their AI and say, hey, 146 00:08:39,400 --> 00:08:42,280 S1: you know, I have this thing, here's the story and 147 00:08:42,280 --> 00:08:45,569 S1: go off into extreme detail about the story, and the 148 00:08:45,570 --> 00:08:48,600 S1: AI is basically taking notes, and you combine that with 149 00:08:48,600 --> 00:08:52,440 S1: the ability to generate art, the ability to generate video 150 00:08:52,470 --> 00:08:54,750 S1: like we're seeing. And pretty soon the AI is going 151 00:08:54,780 --> 00:08:56,310 S1: to be able to come back and say, oh, do 152 00:08:56,309 --> 00:08:59,070 S1: you mean like this? And they're like, no, no, no 153 00:08:59,100 --> 00:09:02,010 S1: more of this influence, more of that influence, and basically 154 00:09:02,010 --> 00:09:06,300 S1: have this interactive conversation with an AI which results in 155 00:09:06,300 --> 00:09:08,460 S1: an art output that I think is going to rival 156 00:09:08,490 --> 00:09:10,680 S1: the best people in the world. And just imagine that 157 00:09:10,679 --> 00:09:13,650 S1: for 8 billion people on the planet. So the next 158 00:09:13,650 --> 00:09:16,770 S1: one here is hiring. I think hiring is going to 159 00:09:16,800 --> 00:09:20,130 S1: change significantly. And there's a number of pitfalls here that 160 00:09:20,130 --> 00:09:22,110 S1: we need to watch out for. But we've known for 161 00:09:22,110 --> 00:09:26,340 S1: a long time that interviews are not super predictive of performance. 162 00:09:26,340 --> 00:09:30,360 S1: So I think it's absolutely ripe for disruption. And what 163 00:09:30,360 --> 00:09:33,330 S1: I see happening here, and I've already seen multiple companies 164 00:09:33,330 --> 00:09:36,870 S1: working on this, is basically AI goes out and does 165 00:09:36,870 --> 00:09:41,100 S1: full collection of everything about this person, this candidate, and 166 00:09:41,100 --> 00:09:43,510 S1: it's basically looking at all their blog posts. It's looking 167 00:09:43,510 --> 00:09:47,650 S1: at the papers that they've written, it's looking at their GitHub, 168 00:09:47,650 --> 00:09:51,010 S1: it's looking at commentary, it's looking at basically anything they've 169 00:09:51,010 --> 00:09:54,280 S1: put anywhere that's available to the eye, basically in public 170 00:09:54,280 --> 00:09:56,829 S1: to read. And it's basically turning that into like a, 171 00:09:56,830 --> 00:10:00,520 S1: a profile or a, a psychological profile or a work 172 00:10:00,520 --> 00:10:05,620 S1: history profile. And then it can also do interviews with 173 00:10:05,620 --> 00:10:09,220 S1: the person or will be able to very soon. So 174 00:10:09,250 --> 00:10:12,939 S1: it first, it knows everything you've ever written online that's public. 175 00:10:12,940 --> 00:10:15,370 S1: And then it can have a conversation with you about 176 00:10:15,370 --> 00:10:18,550 S1: the pertinent skills of what that company is looking for. 177 00:10:18,550 --> 00:10:21,040 S1: And it also is going to be doing ML magic 178 00:10:21,040 --> 00:10:23,410 S1: to say, okay, here are the things we actually know 179 00:10:23,410 --> 00:10:27,309 S1: that are predictive of success in this in this job. 180 00:10:27,640 --> 00:10:31,300 S1: So this is going to be really powerful because it's 181 00:10:31,300 --> 00:10:34,720 S1: then going to result in a score and a recommendation 182 00:10:34,720 --> 00:10:36,790 S1: of like, yes, we should hire this person or yes 183 00:10:36,820 --> 00:10:39,730 S1: or no, we shouldn't. Now one thing to mention here 184 00:10:39,730 --> 00:10:44,220 S1: is that anytime you have ML involved and AI involved, 185 00:10:44,220 --> 00:10:46,710 S1: it's a little bit scary to have a thumbs up 186 00:10:46,710 --> 00:10:49,709 S1: or thumbs down or a score when you're not exactly 187 00:10:49,710 --> 00:10:53,760 S1: sure what the criteria are. And we absolutely need to 188 00:10:53,760 --> 00:10:58,559 S1: protect against bias in this case. Um, so transparent AI 189 00:10:58,590 --> 00:11:00,690 S1: is going to really help here. The next thing I 190 00:11:00,690 --> 00:11:05,370 S1: see happening is a surging requirement to be visible in 191 00:11:05,370 --> 00:11:08,250 S1: the world, to be resilient in this in this world 192 00:11:08,250 --> 00:11:14,010 S1: that's coming up. And authenticity essentially becomes signal. So ideas, 193 00:11:14,280 --> 00:11:18,000 S1: ideas and personality become the things that people care about. 194 00:11:18,000 --> 00:11:20,790 S1: And it becomes more important than medium. It becomes more 195 00:11:20,790 --> 00:11:23,939 S1: important than, you know, being being on a TV show. 196 00:11:23,970 --> 00:11:27,120 S1: As we're already seeing like CNN is not really competing 197 00:11:27,120 --> 00:11:31,350 S1: with YouTube anymore. So ideas and personality become the things 198 00:11:31,350 --> 00:11:34,080 S1: that people care about. And it's more important than medium, 199 00:11:34,080 --> 00:11:39,500 S1: and it's becoming more important because the creator content is 200 00:11:39,500 --> 00:11:41,660 S1: going to become more direct. It's going to be like 201 00:11:41,690 --> 00:11:45,050 S1: an influencer, create something and it goes directly to the consumer, 202 00:11:45,050 --> 00:11:47,929 S1: and they're not even really going to care what medium 203 00:11:47,929 --> 00:11:50,210 S1: it was on. And that's going to be even more 204 00:11:50,210 --> 00:11:53,600 S1: true when it's actually a person's eye that's going to 205 00:11:53,630 --> 00:11:56,720 S1: collect the stuff and bring it back to the principle. 206 00:11:56,720 --> 00:12:00,170 S1: So the prediction here is that I will work for 207 00:12:00,170 --> 00:12:04,069 S1: people to find the best content for them to read. So, 208 00:12:04,070 --> 00:12:06,620 S1: for example, my eye will be going out while I'm 209 00:12:06,620 --> 00:12:10,819 S1: sleeping to collect the coolest new ideas, the coolest new stories, 210 00:12:10,820 --> 00:12:15,050 S1: the coolest new opinions. And in order to find those 211 00:12:15,230 --> 00:12:19,490 S1: for a given person like, say, somebody who's listening to this, 212 00:12:19,490 --> 00:12:21,740 S1: you will have had to have put that out there. 213 00:12:21,740 --> 00:12:24,410 S1: So you have to get your ideas out there into 214 00:12:24,410 --> 00:12:27,980 S1: the world. So basically, if you're not visible, you're vulnerable. 215 00:12:27,980 --> 00:12:31,490 S1: And this brings me to something I'm quite concerned with 216 00:12:31,490 --> 00:12:34,730 S1: about AI, which is we already have a major divide 217 00:12:34,730 --> 00:12:39,400 S1: in terms of like overall success in life around being 218 00:12:39,400 --> 00:12:43,270 S1: a voracious reader or not. And there's there's a decent 219 00:12:43,270 --> 00:12:46,300 S1: amount of data on this which I don't have here, 220 00:12:46,300 --> 00:12:50,980 S1: but I can include later. But essentially there's lots of 221 00:12:50,980 --> 00:12:54,160 S1: anecdote anecdote around this as well. So Charlie Munger has 222 00:12:54,160 --> 00:12:56,770 S1: this great quote. In my whole life I have known 223 00:12:56,770 --> 00:13:00,550 S1: no wise people who didn't read all the time, none, zero. 224 00:13:00,550 --> 00:13:04,329 S1: And I think that really resonates. I think it's a 225 00:13:04,330 --> 00:13:07,209 S1: powerful concept, and I'm worried that AI is going to 226 00:13:07,210 --> 00:13:11,320 S1: be a massive magnifier of this. It's going to be 227 00:13:11,320 --> 00:13:14,950 S1: just like reading, but much, much worse. And here's the problem. 228 00:13:14,950 --> 00:13:18,100 S1: The the problem is we're basically going to have two groups, right? 229 00:13:18,130 --> 00:13:23,319 S1: One group that has say, 10,000 AI assistants working for them. 230 00:13:23,320 --> 00:13:26,890 S1: So they learn the AI, they learn the programming, they 231 00:13:26,890 --> 00:13:30,100 S1: deep dive into this whole world, and they hire all 232 00:13:30,100 --> 00:13:33,490 S1: these very, very cheap AIS to do all these different 233 00:13:33,490 --> 00:13:37,910 S1: things for them. So they're constantly modifying their their finances. 234 00:13:37,910 --> 00:13:41,239 S1: They're constantly finding better books. They're constantly summarizing books for 235 00:13:41,240 --> 00:13:46,010 S1: them and basically turning this person that they're working for into, 236 00:13:46,040 --> 00:13:49,550 S1: like a super knowledgeable, like almost like a superhuman brain. 237 00:13:49,550 --> 00:13:52,580 S1: And some top percentage of the world is going to 238 00:13:52,610 --> 00:13:56,570 S1: function in this way, whether that's 1% or 5% or 10%, 239 00:13:56,600 --> 00:13:59,510 S1: whatever you want to call that. But another group is 240 00:13:59,510 --> 00:14:02,510 S1: going to essentially use it to find entertainment, or they're 241 00:14:02,510 --> 00:14:08,000 S1: going to be influenced by, uh, potentially malicious actors through 242 00:14:08,000 --> 00:14:10,189 S1: that medium, but they're not going to be doing this 243 00:14:10,190 --> 00:14:12,710 S1: thing that the first group is doing, and this is 244 00:14:12,710 --> 00:14:16,970 S1: going to separate groups more than ever. So with that, 245 00:14:16,970 --> 00:14:19,220 S1: I want to talk about what I think we can 246 00:14:19,220 --> 00:14:23,540 S1: do to make ourselves resilient to this. So I want 247 00:14:23,570 --> 00:14:26,720 S1: to talk about a concept called human 3.0. And my 248 00:14:26,720 --> 00:14:31,340 S1: first recommendation here is that you essentially want to combine 249 00:14:31,340 --> 00:14:35,590 S1: all this together to understand your full authentic self, to 250 00:14:35,620 --> 00:14:37,900 S1: know what you want and to share it with the world. 251 00:14:37,930 --> 00:14:40,840 S1: We used to essentially be our resumes or we kind 252 00:14:40,870 --> 00:14:43,450 S1: of still are now. I consider what we're in now 253 00:14:43,480 --> 00:14:46,420 S1: to be human 2.0, where the value of your you 254 00:14:46,420 --> 00:14:48,280 S1: as a human is what you can put in a 255 00:14:48,280 --> 00:14:51,550 S1: CV or a resume. And I think the value of 256 00:14:51,550 --> 00:14:56,020 S1: this human 3.0 concept is that you are your full 257 00:14:56,020 --> 00:15:00,640 S1: spectrum self. You know, your your authenticity. Your personality is 258 00:15:00,640 --> 00:15:03,160 S1: now not a thing that you never include in your CV. 259 00:15:03,190 --> 00:15:05,380 S1: It's now the thing that is able to cut through 260 00:15:05,380 --> 00:15:07,180 S1: and make people want to listen. So if you don't 261 00:15:07,180 --> 00:15:10,870 S1: broadcast your full self, you won't be able to differentiate. 262 00:15:10,870 --> 00:15:14,350 S1: And this starts with knowing really who you are. What 263 00:15:14,350 --> 00:15:17,440 S1: is your mission? What are your goals? Really, really importantly, 264 00:15:17,440 --> 00:15:19,690 S1: what do you think are the most important problems in 265 00:15:19,690 --> 00:15:22,090 S1: the world and what are your metrics? How do you 266 00:15:22,090 --> 00:15:24,280 S1: know if you're doing well as a as a person 267 00:15:24,280 --> 00:15:27,340 S1: or not? And if you can't define yourself in this way, 268 00:15:27,340 --> 00:15:30,400 S1: you're going to be more replaceable from I. So the 269 00:15:30,420 --> 00:15:33,750 S1: next piece of this is getting in the habit. And 270 00:15:33,750 --> 00:15:36,750 S1: this goes back to the the broadcasting of self. This 271 00:15:36,750 --> 00:15:38,580 S1: is essentially how you do it. You got to get 272 00:15:38,580 --> 00:15:41,970 S1: in the habit of being able to articulate yourself clearly. 273 00:15:41,970 --> 00:15:44,700 S1: So one way to think about this is a great 274 00:15:44,700 --> 00:15:47,550 S1: quote that I found from Paul Graham, but it's actually 275 00:15:47,550 --> 00:15:53,010 S1: from Leslie Lamport. If you think without writing, you only 276 00:15:53,010 --> 00:15:56,460 S1: think you're thinking. If you're thinking without writing, you only 277 00:15:56,460 --> 00:16:00,000 S1: think that you're thinking, I really, really love this. And 278 00:16:00,000 --> 00:16:04,020 S1: it essentially the way I encapsulate this is you've got 279 00:16:04,020 --> 00:16:08,610 S1: to get good at thinking, writing and presenting. Um, because 280 00:16:08,610 --> 00:16:11,970 S1: it's no longer for special people, for authors, this is 281 00:16:11,970 --> 00:16:13,950 S1: the only way that you're going to be able to 282 00:16:13,980 --> 00:16:18,000 S1: become visible and stay visible to the world, especially when 283 00:16:18,030 --> 00:16:21,420 S1: I is the one crawling and looking for signal and content. 284 00:16:21,420 --> 00:16:24,330 S1: So you've got to get really good at articulating what 285 00:16:24,330 --> 00:16:26,940 S1: you're about, what you're working on, and why you're working 286 00:16:26,940 --> 00:16:30,230 S1: on those things. And these are the primary skills that 287 00:16:30,230 --> 00:16:33,500 S1: I recommend that you cultivate. So this is really an 288 00:16:33,500 --> 00:16:37,310 S1: encapsulation of the whole thing. It's clear articulation of self. 289 00:16:37,310 --> 00:16:42,230 S1: It's being insatiably curious. This is another big thing that 290 00:16:42,230 --> 00:16:45,170 S1: Charlie Munger talks about. And that's the reason he reads 291 00:16:45,170 --> 00:16:49,580 S1: so much. It's continuous learning, deep integration of AI to 292 00:16:49,610 --> 00:16:54,170 S1: accelerate that learning, discipline and passion towards chasing your curiosity, 293 00:16:54,200 --> 00:16:58,850 S1: focusing on useful work, and staying focused on the problems 294 00:16:58,850 --> 00:17:02,240 S1: that you're solving, and using those problems as your your 295 00:17:02,270 --> 00:17:06,950 S1: guiding light. And in terms of practical skills being good 296 00:17:06,950 --> 00:17:09,139 S1: at programming, you don't have to be an expert programmer. 297 00:17:09,140 --> 00:17:11,270 S1: You just have to know the concepts and be able 298 00:17:11,270 --> 00:17:15,470 S1: to interact with AI, because AI does a lot of programming. Uh, 299 00:17:15,470 --> 00:17:17,119 S1: but you still have to be able to tell it 300 00:17:17,119 --> 00:17:19,850 S1: and understand the concepts. So I would say programming is 301 00:17:19,850 --> 00:17:23,660 S1: important AI tool integration. You want to learn AI and 302 00:17:23,660 --> 00:17:27,750 S1: most importantly, thinking, writing and presenting and all those really 303 00:17:27,750 --> 00:17:32,669 S1: are writing and presenting really are just ways to ensure 304 00:17:32,670 --> 00:17:36,869 S1: that your thinking is clear. And so bringing it all together, 305 00:17:36,869 --> 00:17:41,310 S1: this is what I recommend we do tactically and immediately. 306 00:17:41,310 --> 00:17:46,350 S1: So that's capturing yourself in terms of purpose, goals, metrics, mission, 307 00:17:46,350 --> 00:17:51,000 S1: that sort of thing. Get really good at explaining your 308 00:17:51,000 --> 00:17:54,420 S1: encapsulation of that to other people. So that's writing, speaking, 309 00:17:54,420 --> 00:17:56,940 S1: presenting that sort of thing. Learn to program and learn 310 00:17:56,940 --> 00:18:00,300 S1: to use. I probably read 2 to 4 hours a day. 311 00:18:00,330 --> 00:18:05,100 S1: High quality content. Build your presence online. So get your 312 00:18:05,220 --> 00:18:09,300 S1: website going. Essentially talk about it doesn't matter. A lot 313 00:18:09,300 --> 00:18:11,370 S1: of people say, look, I have nothing to say because 314 00:18:11,369 --> 00:18:14,340 S1: I'm still learning. Be visible, learn in public. And that 315 00:18:14,340 --> 00:18:16,290 S1: will allow you to do the next piece, which is 316 00:18:16,290 --> 00:18:18,840 S1: to connect with others who are doing the same. Because 317 00:18:18,840 --> 00:18:21,210 S1: you can have ten different people who are talking about 318 00:18:21,210 --> 00:18:24,060 S1: the same stuff, but if they are being their full selves, 319 00:18:24,060 --> 00:18:27,380 S1: those ten people will look and sound very different and 320 00:18:27,380 --> 00:18:31,100 S1: they will find different audiences. And most importantly, you want 321 00:18:31,130 --> 00:18:34,760 S1: to do this authentically as your full self and share 322 00:18:34,760 --> 00:18:36,770 S1: this with the world. Thanks for your time. 323 00:18:36,770 --> 00:18:39,800 S2: Well, thank you so much, Daniel. That was incredibly profound 324 00:18:39,800 --> 00:18:43,280 S2: and a lot deeper than I was expecting, which was amazing. Yeah, 325 00:18:43,310 --> 00:18:45,140 S2: I have a few questions myself, but I think I'll 326 00:18:45,140 --> 00:18:48,379 S2: start by handing it over to my colleague Daniel, who 327 00:18:48,380 --> 00:18:52,880 S2: is Wipo's information security engineer. And yeah, he will cover 328 00:18:52,880 --> 00:18:56,840 S2: some questions that were pre-submitted from yeah, people who initially 329 00:18:56,840 --> 00:18:59,540 S2: registered for this talk. So Daniel Jeremiah, the talk is 330 00:18:59,540 --> 00:19:01,909 S2: now yours. Sorry. The floor is now yours. 331 00:19:02,000 --> 00:19:05,600 S3: All right. Thank you so much for the eye opening presentation. 332 00:19:05,600 --> 00:19:07,730 S3: I don't know about you, Olivia, but I know what 333 00:19:07,730 --> 00:19:10,010 S3: I'm going to be busy with for the next couple 334 00:19:10,040 --> 00:19:10,610 S3: of years. 335 00:19:11,840 --> 00:19:12,530 S4: Yeah. 336 00:19:13,100 --> 00:19:16,040 S3: We have received many, many questions. Colleagues, thank you for 337 00:19:16,040 --> 00:19:19,250 S3: your interest. And I believe that several of them could 338 00:19:19,250 --> 00:19:22,219 S3: probably be grouped under different headings, different buckets related to 339 00:19:22,660 --> 00:19:27,159 S3: I and our jobs. The integration at Waco and to 340 00:19:27,160 --> 00:19:30,490 S3: what level we can trust. I. So here's the first one. 341 00:19:31,060 --> 00:19:33,970 S3: Will I replace human labor in the near future? 342 00:19:34,000 --> 00:19:38,890 S1: Yeah. Um, I think there's a couple of interesting, um, 343 00:19:38,890 --> 00:19:42,610 S1: caveats or sort of pivot points inside of the question. Uh, 344 00:19:42,609 --> 00:19:46,990 S1: replace implies a one or a zero that it's just yes, 345 00:19:46,990 --> 00:19:49,810 S1: it's replaced or no, it's not replaced. That's going to 346 00:19:49,810 --> 00:19:53,350 S1: be gradual. And when near future is also like that 347 00:19:53,350 --> 00:19:55,750 S1: as well. It's like that. What does that mean. Does 348 00:19:55,750 --> 00:19:59,080 S1: it mean one year? Does it mean three years? The 349 00:19:59,080 --> 00:20:01,630 S1: the way I would answer that is to say that 350 00:20:02,050 --> 00:20:05,980 S1: will AI replace human jobs? The answer is yes. The 351 00:20:05,980 --> 00:20:08,800 S1: question is how much and how quickly and what types 352 00:20:08,800 --> 00:20:13,480 S1: of jobs. And I would say that in general, what 353 00:20:13,480 --> 00:20:15,669 S1: it's going to replace first is things that can be 354 00:20:15,670 --> 00:20:21,180 S1: done easily that don't have human authenticity and human Uniqueness 355 00:20:21,210 --> 00:20:24,270 S1: inside of the work. And this is why this whole 356 00:20:24,270 --> 00:20:28,380 S1: presentation is oriented around getting that into what you do, 357 00:20:28,410 --> 00:20:32,310 S1: so that it is harder to replace. So I think 358 00:20:32,310 --> 00:20:34,980 S1: there are millions of jobs that will be replaced in 359 00:20:34,980 --> 00:20:38,760 S1: the next 5 to 10 years and many, many more 360 00:20:39,030 --> 00:20:42,060 S1: over the course of ten years versus five. The good 361 00:20:42,060 --> 00:20:44,730 S1: news is it's also going as we talked about in 362 00:20:44,730 --> 00:20:49,080 S1: the presentation. It's also going to create lots more jobs. 363 00:20:49,109 --> 00:20:52,530 S1: In fact, I believe that it's going to kind of 364 00:20:52,560 --> 00:20:56,730 S1: shut down the human 2.0 economy where we're basically doing 365 00:20:56,730 --> 00:21:00,150 S1: raw tasks. And it's going to move us towards this 366 00:21:00,150 --> 00:21:04,470 S1: human 3.0 economy, which is more human to human interaction. 367 00:21:04,500 --> 00:21:07,800 S1: So the answer is yes, it will replace a lot 368 00:21:07,800 --> 00:21:10,290 S1: of jobs, but I don't think that's all bad news. 369 00:21:10,320 --> 00:21:12,450 S3: Well, let's hope for the enhancement of the. 370 00:21:12,480 --> 00:21:13,290 S1: Yes. 371 00:21:13,859 --> 00:21:18,090 S3: Thanks. All right. Thank you. The next one would be, uh, 372 00:21:18,090 --> 00:21:21,790 S3: if you have any advice, what advice would you have 373 00:21:21,820 --> 00:21:25,810 S3: for Ypo other organizations looking to go looking to integrate 374 00:21:25,810 --> 00:21:28,600 S3: AI in our current work? Because we know, you know, 375 00:21:28,630 --> 00:21:33,340 S3: we have several initiatives and ideas and explorations, but actually 376 00:21:33,340 --> 00:21:36,850 S3: the the substance of the work is not yet touched. Yes. 377 00:21:37,210 --> 00:21:40,780 S1: Yeah. So really good question. I think the answer is 378 00:21:40,780 --> 00:21:43,840 S1: to do what I was talking about. The AI consultancies 379 00:21:43,840 --> 00:21:46,600 S1: are going to do to your company. So you essentially 380 00:21:46,600 --> 00:21:50,770 S1: want to go to a very large whiteboard and write, 381 00:21:50,830 --> 00:21:54,190 S1: document your company, capture it as a giant graph, and 382 00:21:54,190 --> 00:21:57,399 S1: break out every single little piece. Okay. The customer sends 383 00:21:57,400 --> 00:22:01,629 S1: this thing in it does this following thing is done 384 00:22:01,630 --> 00:22:05,500 S1: with Sarah's team that sends over to Chris's team. And 385 00:22:05,500 --> 00:22:07,570 S1: here's what the teams look like. And here's exactly what 386 00:22:07,570 --> 00:22:09,790 S1: they do. And here's how many people are doing that. 387 00:22:09,820 --> 00:22:13,210 S1: You basically want to break that out and understand your 388 00:22:13,210 --> 00:22:17,610 S1: business perfectly the way that it is today before I. 389 00:22:17,609 --> 00:22:20,879 S1: And then you want to say, okay, now what parts 390 00:22:20,880 --> 00:22:23,699 S1: of this look inefficient? What parts of this should we 391 00:22:23,730 --> 00:22:26,310 S1: not be doing at all? That's the first question I 392 00:22:26,340 --> 00:22:29,070 S1: is going to ask is what can we cut? What 393 00:22:29,070 --> 00:22:32,340 S1: are the extra layers. So you want to think of 394 00:22:32,340 --> 00:22:37,619 S1: the I as basically just a very harsh lens that's 395 00:22:37,619 --> 00:22:39,750 S1: going to look at everything you're doing. So the very 396 00:22:39,750 --> 00:22:43,170 S1: first step is understanding exactly what you are doing. And 397 00:22:43,170 --> 00:22:45,330 S1: that's going to make it clear where you should apply. 398 00:22:45,359 --> 00:22:50,070 S3: I think it's uh, well it's concerning, but a lot 399 00:22:50,070 --> 00:22:53,760 S3: of challenges ahead. Yeah. Um, so, so the for the 400 00:22:53,760 --> 00:22:57,300 S3: next we have a package is, uh, regarding the security. Well, 401 00:22:57,300 --> 00:23:00,240 S3: it's close to hard. To what level can we trust AI? 402 00:23:00,240 --> 00:23:03,840 S3: Is cybersecurity really possible in the face of artificial intelligence? 403 00:23:03,840 --> 00:23:06,570 S3: And mainly for me at least, how do we ensure 404 00:23:06,570 --> 00:23:09,270 S3: that the information we receive has come from a good source? 405 00:23:09,300 --> 00:23:12,690 S3: And how can we best validate that information? 406 00:23:12,720 --> 00:23:16,970 S1: Yeah, it's a great, great question. I am not too 407 00:23:16,970 --> 00:23:21,139 S1: concerned about it, actually. Not because I trust the raw 408 00:23:21,140 --> 00:23:24,139 S1: content coming out of an LLM that is not a 409 00:23:24,140 --> 00:23:27,530 S1: good idea, but essentially what's going to happen and is 410 00:23:27,530 --> 00:23:32,330 S1: already happening is that content being created from an AI 411 00:23:32,480 --> 00:23:35,899 S1: will go through a pipeline of steps depending on the 412 00:23:35,900 --> 00:23:39,439 S1: importance of the task. So if the task is to 413 00:23:39,470 --> 00:23:43,280 S1: say no, you do not have this disease. That is 414 00:23:43,280 --> 00:23:46,879 S1: such a powerful thing to say. It must be checked 415 00:23:46,880 --> 00:23:49,159 S1: in multiple places. So what you're going to have is 416 00:23:49,160 --> 00:23:52,220 S1: you're going to have multiple little eyes which are doing 417 00:23:52,250 --> 00:23:55,159 S1: fact checking. So they're going to be checking Google. They're 418 00:23:55,160 --> 00:23:58,700 S1: going to be checking official medicine publications. It's going to 419 00:23:58,700 --> 00:24:02,360 S1: be maybe consulting with a human expert for a final check. 420 00:24:02,359 --> 00:24:07,100 S1: But imagine a a result coming out of an AI 421 00:24:07,130 --> 00:24:11,300 S1: being evaluated by 15 other AIS. And they all have 422 00:24:11,300 --> 00:24:13,659 S1: to have a green check mark before the answer is 423 00:24:13,660 --> 00:24:16,419 S1: actually returned to the user. So I don't think we 424 00:24:16,420 --> 00:24:19,270 S1: have to worry about a world in which we just 425 00:24:19,270 --> 00:24:22,719 S1: randomly trust. I think that might happen for something that 426 00:24:22,720 --> 00:24:25,870 S1: doesn't matter. Like tell me a funny joke, but it 427 00:24:25,869 --> 00:24:29,890 S1: won't necessarily it won't happen for the things that really matter. Like, 428 00:24:29,920 --> 00:24:34,630 S1: should we turn on this, um, nuclear control system or 429 00:24:34,630 --> 00:24:37,000 S1: should we turn it off? Should we raise the firewall, 430 00:24:37,000 --> 00:24:39,129 S1: lower the firewall? Those sorts of things are going to 431 00:24:39,130 --> 00:24:41,320 S1: have lots of extra scrutiny added on. 432 00:24:41,350 --> 00:24:44,950 S3: Right. So that's that's the most interesting application of democracy 433 00:24:44,950 --> 00:24:48,970 S3: I heard recently if if we have agents voting on 434 00:24:48,970 --> 00:24:50,080 S3: the answers, that's, uh. 435 00:24:50,950 --> 00:24:55,120 S1: Yes, absolutely. And different kinds. This is really important. You'll 436 00:24:55,119 --> 00:24:57,370 S1: have someone who's like an academic, and they have this 437 00:24:57,369 --> 00:25:00,369 S1: one type of scrutiny, and you can have this other 438 00:25:00,640 --> 00:25:03,580 S1: industry expert kind of type of scrutiny, but they can 439 00:25:03,580 --> 00:25:06,070 S1: have different backgrounds, but all of them must say yes 440 00:25:06,070 --> 00:25:08,230 S1: for this kid to get an actual s o. 441 00:25:08,230 --> 00:25:11,230 S3: Great. So thank you very much. I believe that's that's 442 00:25:11,230 --> 00:25:12,050 S3: all I have. 443 00:25:12,619 --> 00:25:13,159 S5: Uh. 444 00:25:13,340 --> 00:25:15,919 S3: Thank you. It's. It was amazing. And a lot of 445 00:25:15,920 --> 00:25:18,680 S3: lessons for me, and I hope for for our colleagues. 446 00:25:18,680 --> 00:25:21,350 S3: So I will hand back to, to Olivia to to 447 00:25:21,380 --> 00:25:22,760 S3: wrap up. Thank you. 448 00:25:22,940 --> 00:25:26,720 S2: Perfect. Thanks. Um, also, maybe Daniel. Do. Daniel, do you 449 00:25:26,720 --> 00:25:28,760 S2: have any questions from your side? Because I've got 1 450 00:25:28,760 --> 00:25:30,080 S2: or 2 I'd like to ask. I think we have 451 00:25:30,080 --> 00:25:31,250 S2: a few more minutes. So. 452 00:25:31,820 --> 00:25:32,690 S6: Uh, so. 453 00:25:32,720 --> 00:25:36,140 S3: My questions were, you know, security and this thing of 454 00:25:36,140 --> 00:25:41,990 S3: trustworthiness and, uh, explainability, but I think I have my answers, so. 455 00:25:41,990 --> 00:25:42,950 S3: So I believe I'm good. 456 00:25:43,340 --> 00:25:45,859 S2: Okay, perfect. I have a quick question about what you mentioned, 457 00:25:45,859 --> 00:25:49,070 S2: about how personalities and ideas are most important and that 458 00:25:49,070 --> 00:25:52,670 S2: we should be online. But what about AI influencers? What 459 00:25:52,670 --> 00:25:53,780 S2: are your thoughts on those? 460 00:25:53,810 --> 00:25:57,859 S1: Yeah, it's, uh, it depends how good they get. Uh, 461 00:25:57,859 --> 00:26:01,850 S1: if they get really, really good, um, and they're allowed 462 00:26:01,850 --> 00:26:04,610 S1: to continue, uh, because it could be that, um, some 463 00:26:04,609 --> 00:26:08,960 S1: governments decide that they're too good and they're disrupting human 464 00:26:08,980 --> 00:26:12,610 S1: capabilities to actually be creative. At which point they could 465 00:26:12,609 --> 00:26:15,220 S1: be shut down. But if they are not shut down 466 00:26:15,220 --> 00:26:18,639 S1: and they are really, really good, the bad news there 467 00:26:18,640 --> 00:26:22,000 S1: is that they might quickly become much, much better than 468 00:26:22,000 --> 00:26:26,050 S1: any human like. They'll just be more attractive. They will 469 00:26:26,050 --> 00:26:29,650 S1: speak more attractively. The ideas will be better, they will 470 00:26:29,680 --> 00:26:34,419 S1: be funnier. And it's like the human creators, first of all, 471 00:26:34,420 --> 00:26:37,720 S1: they have to sleep so they won't be creating as much, 472 00:26:37,720 --> 00:26:40,389 S1: and they might move down in the rankings. And the 473 00:26:40,390 --> 00:26:43,990 S1: most popular influencers and the most popular shows and everything. 474 00:26:44,020 --> 00:26:46,930 S1: It might be all I created. I do believe that 475 00:26:46,930 --> 00:26:50,409 S1: that is the type of thing. It's like the last. 476 00:26:50,650 --> 00:26:53,740 S1: It's the last stand for humanity is to be this 477 00:26:53,740 --> 00:26:57,700 S1: creative thing. And if if they take that away, I'm 478 00:26:57,700 --> 00:27:00,790 S1: not sure what else we have. So I could honestly 479 00:27:00,790 --> 00:27:05,679 S1: see that being the place that regulation happens to say 480 00:27:05,680 --> 00:27:10,230 S1: we these must be validated humans to be creators. 481 00:27:10,770 --> 00:27:15,419 S2: Well, yeah. It's interesting. That's a scary thought. Um, and also, yeah, 482 00:27:15,450 --> 00:27:18,510 S2: on the line of being out there and being visible online, 483 00:27:18,510 --> 00:27:20,760 S2: what are your thoughts when it comes to data privacy risks? 484 00:27:20,760 --> 00:27:23,250 S2: For example, if I put all this information of me 485 00:27:23,250 --> 00:27:25,410 S2: out there and cybersecurity risks as well. 486 00:27:25,500 --> 00:27:30,930 S1: Yeah, it's a great question. I believe the answer is 487 00:27:30,960 --> 00:27:38,370 S1: it won't matter that much. Uh, unfortunately or fortunately, I think, um, 488 00:27:38,460 --> 00:27:40,740 S1: the way this is going to move is people are 489 00:27:40,740 --> 00:27:43,680 S1: going to give so much data to their AI, their 490 00:27:43,680 --> 00:27:48,840 S1: personal AI. And because those are largely startup companies, they're 491 00:27:48,840 --> 00:27:51,690 S1: all going to get hacked. Um, so these startup companies 492 00:27:51,690 --> 00:27:54,180 S1: are going to get hacked. And rather than just having 493 00:27:54,180 --> 00:27:58,830 S1: your financial information, it's going to be essentially like your soul. 494 00:27:58,830 --> 00:28:01,710 S1: It's going to be like, I had these traumatic experiences. 495 00:28:01,710 --> 00:28:05,100 S1: Here are all my relationships. It's going to know all 496 00:28:05,100 --> 00:28:07,609 S1: your health data. When that gets hacked, it's going to 497 00:28:07,609 --> 00:28:13,040 S1: be unbelievably traumatic. But watch this only for the first 498 00:28:13,040 --> 00:28:16,820 S1: few people, only for the first year or two, because 499 00:28:16,850 --> 00:28:22,070 S1: we're not used to having that level of personal exposure online. 500 00:28:22,070 --> 00:28:26,359 S1: But soon we will. And then we're just going to 501 00:28:26,359 --> 00:28:29,360 S1: know that we're all the same. We're all equally flawed. 502 00:28:29,359 --> 00:28:33,470 S1: We all have these equally weird things about us. And 503 00:28:33,500 --> 00:28:36,109 S1: there's one possible way for it to go is that 504 00:28:36,109 --> 00:28:41,150 S1: it just becomes so, uh, regular for this to happen, 505 00:28:41,180 --> 00:28:45,050 S1: that it becomes the new baseline and nobody cares. Yeah. 506 00:28:45,320 --> 00:28:47,990 S2: Well, so, yeah, if everything's out there for everyone to see, 507 00:28:48,020 --> 00:28:50,300 S2: then they can't use it as a weapon. That makes sense. 508 00:28:50,540 --> 00:28:51,410 S1: Exactly. 509 00:28:51,890 --> 00:28:55,160 S2: Okay, well, that was it from my side. Um, thank 510 00:28:55,190 --> 00:28:58,490 S2: you very much, Daniel. This was fantastic. And thank you 511 00:28:58,490 --> 00:29:02,630 S2: for your time today. Um, to everyone watching, we do 512 00:29:02,630 --> 00:29:05,310 S2: have an AI at Ypo Viper Teams channel. So if 513 00:29:05,310 --> 00:29:08,580 S2: you'd like to join the community, please do so. Um, yeah. 514 00:29:08,610 --> 00:29:11,100 S2: You can find out about upcoming AI repo sessions and 515 00:29:11,130 --> 00:29:13,680 S2: other un AI events there, as well as well as 516 00:29:13,680 --> 00:29:17,640 S2: industry news, and you can join general AI discussions. And 517 00:29:17,640 --> 00:29:20,280 S2: please feel free to reach out or consult the AI 518 00:29:20,430 --> 00:29:23,610 S2: repo intranet page if you'd like to learn more. So 519 00:29:23,610 --> 00:29:26,700 S2: thank you again, Daniel Mazia. Thank you Daniel Jeremiah and 520 00:29:26,700 --> 00:29:27,930 S2: have a good day everyone. 521 00:29:29,490 --> 00:29:32,670 S1: Unsupervised learning is produced and edited by Daniel Miessler on 522 00:29:32,700 --> 00:29:37,260 S1: a Neumann U87 AI microphone using Hindenburg. Intro and outro 523 00:29:37,290 --> 00:29:40,620 S1: music is by zombie with a Y. And to get 524 00:29:40,620 --> 00:29:42,690 S1: the text and links from this episode, sign up for 525 00:29:42,690 --> 00:29:45,630 S1: the newsletter version of the show at Daniel Missler Comm 526 00:29:45,690 --> 00:29:49,230 S1: Slash newsletter. We'll see you next time.