WEBVTT 00:00:00.180 --> 00:00:04.440 You know, jobs are not lost to technology. Jobs are lost to recessions, 00:00:04.440 --> 00:00:06.800 and then they don't come back before technology because of technology. 00:00:06.800 --> 00:00:08.690 If you have a hammer, everything looks like a nail. 00:00:09.420 --> 00:00:12.970 And so we really want to solve climate change with software. 00:00:13.570 --> 00:00:16.350 Climate change is not a software problem. Climate change is a material science 00:00:16.350 --> 00:00:19.510 problem, a policy problem, and a capital problem. There's tremendous investments 00:00:19.510 --> 00:00:20.690 by the large language models. 00:00:21.110 --> 00:00:24.680 They are the equivalent of the telco providers. They are spending all the money 00:00:24.680 --> 00:00:27.860 for the training. They hope to capture the value. They might not. 00:00:27.860 --> 00:00:29.640 And they're establishing the infrastructure. 00:00:29.760 --> 00:00:33.260 And they were overbuilt. We always overbuilt. Hey, everyone. 00:00:33.260 --> 00:00:36.700 I'm Malak Fouad, and welcome again to What I Did Next from A&T Media. 00:00:37.350 --> 00:00:40.630 Every now and again, we have an encore episode where we check in with some of 00:00:40.630 --> 00:00:42.140 your favorite alumni guests. 00:00:42.960 --> 00:00:49.130 Today, we're sitting down again with Ramy Adeeb, founder and managing partner of 1984 Ventures. 00:00:49.260 --> 00:00:53.930 When we last spoke in 2023, Ramy was focused on the tech sector in general. 00:00:54.350 --> 00:00:57.740 Today, everything he looks at has to be AI first. 00:00:57.740 --> 00:01:00.930 There is a saying that is often attributed to, of all people, 00:01:00.930 --> 00:01:04.480 Vladimir Lenin. And I don't know. We don't know if he said it or he didn't say it. 00:01:05.140 --> 00:01:09.610 That there are decades when nothing happens. But there are weeks when decades happen. 00:01:10.590 --> 00:01:15.400 And right now we're going through this period where decades are happening in weeks. 00:01:15.760 --> 00:01:24.610 I think what has grabbed my attention is really trying to stay abreast of where 00:01:24.790 --> 00:01:26.020 the technology is today. 00:01:26.850 --> 00:01:29.630 And what the capabilities allow us to do. 00:01:31.230 --> 00:01:32.800 What has also kept my attention, 00:01:34.880 --> 00:01:42.140 is trying to find the signal from the noise. There is so much noise when a new 00:01:42.140 --> 00:01:44.100 technology comes from scratch. 00:01:44.900 --> 00:01:49.200 And eventually that noise subsides and the signal is what remains. 00:01:50.130 --> 00:01:54.050 That's a very interesting way of putting it. I have never heard of it that way. 00:01:54.050 --> 00:01:55.950 It's a really nice way to explain... 00:01:57.120 --> 00:02:00.700 What we'll end up with. It's a signal, not the noise. Yeah. 00:02:00.700 --> 00:02:04.510 Exactly, exactly. And right now, it's really hard to distill the signal from 00:02:04.510 --> 00:02:08.200 the noise. And then last, it's just been really busy building the firm. 00:02:08.710 --> 00:02:11.920 We're investing out of our third fund now, building a team. 00:02:12.570 --> 00:02:16.750 We have five partners working actively on meeting those entrepreneurs. 00:02:17.690 --> 00:02:20.950 You know, there's really two types of VC firms. There's like top tier and everybody else. 00:02:21.420 --> 00:02:25.190 And we want to make sure we're a top tier firm where every founder is going to work with us. 00:02:25.400 --> 00:02:31.790 And it's ultimately a customer service business, one where you just continuously work for your founders. 00:02:32.240 --> 00:02:38.280 And then that compounds with time until you become a firm that everybody wants to work with. 00:02:38.440 --> 00:02:42.040 And from the very beginning, Ramy, when you set up your VC firm, 00:02:42.040 --> 00:02:43.600 it was always tech focused. 00:02:44.630 --> 00:02:47.940 I remember that that was very much your ethos at the time. 00:02:48.850 --> 00:02:52.400 So today, your third fund, you're talking about your third fund. 00:02:52.400 --> 00:02:56.750 And obviously, it's AI heavy, I'm assuming, since that's where your attention is. 00:02:57.290 --> 00:03:02.880 So how have you kind of evolved from the first fund, the first years into where you are now? 00:03:02.880 --> 00:03:05.460 When we started the fund, we were mainly focused on software, 00:03:06.110 --> 00:03:10.970 pure software. So good old what we call SaaS, software as a service. 00:03:11.520 --> 00:03:16.610 And we avoided all the hype trends. There was a very big hype trend with cryptocurrency, 00:03:16.610 --> 00:03:19.720 and we totally ignored that. We just stuck to our knitting and did software 00:03:19.720 --> 00:03:21.090 upending antiquated industries. 00:03:21.570 --> 00:03:25.860 Then there was Web 3.0, a lot of buzzwords, a lot of noise around how, 00:03:26.380 --> 00:03:32.300 you know, cryptocurrency meeting virtual reality would create these modalities. 00:03:32.890 --> 00:03:36.390 And it wasn't really, it wasn't much. It was just a lot of noise. 00:03:36.980 --> 00:03:39.000 And that didn't pan out in the end, right? 00:03:39.330 --> 00:03:41.810 Yeah, it didn't pan out. We do get the credit of like, there's an article written 00:03:41.810 --> 00:03:46.600 about us in 2022, actually an interview from Fortune magazine where they're, 00:03:47.620 --> 00:03:50.980 just actually featuring us as a fund that has always panned crypto. 00:03:52.160 --> 00:03:54.480 Because we were one of the very few funds that said no to crypto from day ones. 00:03:54.480 --> 00:03:55.780 And they asked us, what are you doing? And we're just like, yeah, 00:03:55.960 --> 00:03:57.810 we're doing AI and other stuff. 00:03:57.810 --> 00:04:02.760 What was the reason, Ramy? Why were you averse to crypto? It was so trendy. 00:04:03.180 --> 00:04:05.650 What was your gut feeling? I mean, why did you say no to that? 00:04:06.320 --> 00:04:11.760 Because we are venture capital. We have a particular mode, right? 00:04:12.060 --> 00:04:16.580 Take technological risk early on, on a small team of founders and hope that 00:04:16.580 --> 00:04:20.720 they build massive businesses that can then disrupt the incumbents. 00:04:21.260 --> 00:04:24.010 And we know how to do that with software. We know how to do that with consumer. 00:04:24.240 --> 00:04:25.770 We even could do that with hardware. 00:04:25.890 --> 00:04:31.860 Crypto was an entirely different asset class. It was like commodity trading. 00:04:32.600 --> 00:04:35.610 I don't know how to do commodity trading. Sure, sure. It's not really commodity. 00:04:35.610 --> 00:04:37.160 It's a commodity means gambling. 00:04:38.300 --> 00:04:41.130 And the way to make money in crypto, you don't need a venture capitalist. 00:04:41.890 --> 00:04:43.810 Just go pick a basket of coins and own it. 00:04:43.810 --> 00:04:45.540 It was a different beast. 00:04:46.260 --> 00:04:51.850 Yeah. It was a different beast. there is a almost folk you know a story about 00:04:51.850 --> 00:04:57.460 about steve jobs that when steve jobs created the very first macintosh and there was a 00:04:58.380 --> 00:05:02.380 there was a big event like a gala where he was demoing the computer and all 00:05:02.380 --> 00:05:07.170 of these old folks would come by seasoned you know entrepreneurs seasoned 00:05:07.170 --> 00:05:10.480 financiers it was actually new york city and ask him what is this and he like 00:05:10.480 --> 00:05:12.120 he would talk to them for two minutes and they would leave he would talk to 00:05:12.120 --> 00:05:13.160 them for two minutes and they would leave, 00:05:13.990 --> 00:05:17.930 And then an eight-year-old came to him And he spoke to him for 40 minutes He 00:05:17.930 --> 00:05:19.480 spoke to the eight-year-old for 40 minutes. 00:05:21.810 --> 00:05:24.380 And then a journalist came to him and said, why is it that you spoke to the 00:05:24.380 --> 00:05:28.080 eight-year-old for 40 minutes and to all the old people for like a couple of 00:05:28.080 --> 00:05:30.220 minutes? And he goes, well, they all asked me, what is it? 00:05:30.880 --> 00:05:33.840 And then the little kid said, hey, what can I do with this? It's a really an 00:05:33.840 --> 00:05:36.510 important question where a lot of the folks who are really, 00:05:38.570 --> 00:05:41.680 you know, endowed in finance and in politics and government ask, 00:05:41.920 --> 00:05:44.780 what is this? And you say, oh, it's like artificial intelligence. 00:05:45.350 --> 00:05:48.180 The real question is, what can I do with it? What are the applications? 00:05:48.890 --> 00:05:52.680 What does this unlock? And for crypto, the primary application, 00:05:53.020 --> 00:05:55.560 you know, was a little bit of gambling plus evading capital controls, 00:05:55.560 --> 00:05:56.700 which is a real application. 00:05:57.150 --> 00:06:00.570 You know, if you're in a country with a fastly deteriorating currency, 00:06:00.570 --> 00:06:03.320 you want to actually use dollars. It's just not an application that, 00:06:03.710 --> 00:06:04.870 you know, we're set out to solve. 00:06:06.720 --> 00:06:09.460 With the Web 3.0, it was really hard to see what the application is. 00:06:10.240 --> 00:06:14.440 But with AI, I mean, all it takes is just launching chatty between your cell 00:06:14.440 --> 00:06:18.470 phone, uploading your medical record like everybody in my family does these days. 00:06:19.020 --> 00:06:22.780 And speaking with it about it and realizing what the possibilities are. 00:06:22.780 --> 00:06:25.240 So those possibilities are endless. That's exciting. 00:06:27.470 --> 00:06:30.740 As an investor, though, as I mentioned, there's a lot of nice. 00:06:31.760 --> 00:06:36.570 Because we don't know at the end of AI who will capture the value. 00:06:37.830 --> 00:06:44.430 Who will come out ahead? What business models will be enduring and what business models will die? 00:06:45.480 --> 00:06:51.100 And anyone who claims to have an end state, their guess is as good as mine. 00:06:51.330 --> 00:06:52.270 Sure, sure, sure. 00:06:52.320 --> 00:06:54.960 And objectively, the hardest part isn't predicting an end state. 00:06:54.960 --> 00:06:58.910 It's just helping founders navigate that transition while it's underway. 00:06:58.910 --> 00:07:02.230 You're talking about that we are at an inflection point. 00:07:02.230 --> 00:07:07.830 We're about to be, where I think, this is my reading of it as a layperson, 00:07:09.550 --> 00:07:15.720 we could see a slight collapse in the industry of maybe extra weight or companies 00:07:15.720 --> 00:07:16.740 that are not performing. 00:07:18.220 --> 00:07:24.110 And then from that, you will get those diamond companies that will endure. 00:07:24.430 --> 00:07:27.480 It's the same with the internet. The system didn't end. 00:07:28.320 --> 00:07:34.200 We got rid of a lot of extra weight companies that were not functioning properly. 00:07:34.690 --> 00:07:38.180 But the industry went on to amazing things after that. 00:07:38.600 --> 00:07:45.220 I couldn't have drawn a better analogy. Actually, I want to go back to the late 90s for a moment. 00:07:45.220 --> 00:07:47.970 It's very in vogue, by the way, now, Ramy, the 90s. 00:07:47.970 --> 00:07:50.630 It is a little bit in vogue. It is a little bit in vogue. It's also, 00:07:50.630 --> 00:07:54.860 you know, the style, the Reebok shoes are back in vogue again. 00:07:54.860 --> 00:08:00.130 It's all back, yeah. There are two lessons from the 90s that I want to touch 00:08:00.130 --> 00:08:02.250 on. One around uncertainty. 00:08:04.550 --> 00:08:08.980 You know, and the other around value capture. But let's start with the first one. 00:08:10.990 --> 00:08:15.580 I remember sitting in a panel at Harvard in 1999 that was called, 00:08:16.170 --> 00:08:18.560 Who Will Capture Value from the Internet? 00:08:19.240 --> 00:08:22.170 Who Will Capture Value from the Internet? And then we had three speakers. 00:08:23.130 --> 00:08:29.200 One was arguing that the telco providers, the companies that laid out the fiber, 00:08:29.560 --> 00:08:32.910 the internet in the ground, were the ones that would capture all the value. 00:08:32.910 --> 00:08:37.440 The second speaker was arguing that it would be the browsers, the Netscapes. 00:08:37.850 --> 00:08:41.940 And it made sense because at the time, Netscape and Microsoft were literally 00:08:42.130 --> 00:08:46.600 at the Supreme Court at each other's throat, where Microsoft had bundled a free, 00:08:47.760 --> 00:08:51.730 Internet Explorer into Windows operating system. 00:08:52.240 --> 00:08:55.860 And Netscape sued it for antitrust. And the thing went all the way to the Supreme 00:08:55.860 --> 00:09:02.970 Court. And the prevailing model for making money from software in 1999 was to 00:09:02.970 --> 00:09:06.550 pay money for somebody to install a piece of software in your computer. 00:09:07.030 --> 00:09:10.790 So if Netscape software was the gateway to this internet, it made sense that 00:09:10.790 --> 00:09:11.720 it would capture all the value. 00:09:12.800 --> 00:09:17.890 The third speaker, who was the least convincing, was arguing that it would be the content providers. 00:09:18.850 --> 00:09:22.740 The websites, the folks who actually ended up capturing all the value from the 00:09:22.740 --> 00:09:24.760 internet. But the reason it wasn't obvious is in. 00:09:26.760 --> 00:09:28.550 1999, you couldn't accept credit 00:09:28.550 --> 00:09:32.830 cards online. The only way to make money online was via display ads. 00:09:33.060 --> 00:09:35.560 Remember those long ads on Yahoo? 00:09:36.670 --> 00:09:41.970 And as a business, you could only make, you know, you could only make maybe 00:09:41.970 --> 00:09:44.160 five bucks, 10 bucks for a million ads. 00:09:44.600 --> 00:09:46.030 So how many ads will I possibly see 00:09:46.030 --> 00:09:49.790 in a month to make it worthwhile for businesses to make a lot of money? 00:09:50.820 --> 00:09:56.120 So the point of the story is when you are going through this technological inflection point, 00:09:57.290 --> 00:09:59.680 There is a tremendous investment by everybody. There is tremendous, 00:09:59.680 --> 00:10:03.470 today there is a very similar analogy. There's tremendous investments by the large language models. 00:10:03.930 --> 00:10:07.470 They are the equivalent of the telco providers. They are spending all the money 00:10:07.470 --> 00:10:10.670 for the training. They hope to capture the value. They might not. 00:10:10.670 --> 00:10:12.430 And they're establishing the infrastructure. 00:10:12.850 --> 00:10:14.870 But they're establishing the infrastructure. And without them, 00:10:14.870 --> 00:10:17.080 we wouldn't have it. And they were overbuilt. 00:10:17.460 --> 00:10:22.880 We always overbuilt, right? No matter, just because the technology is here to 00:10:22.880 --> 00:10:26.150 last, doesn't mean we won't overbuilt. 00:10:26.770 --> 00:10:30.070 We always do. We overbuilt with the internet. You know, we overbuilt. 00:10:30.450 --> 00:10:31.700 And so they will overbuild a little bit. 00:10:32.150 --> 00:10:34.820 Now, you can't, of course, time it, right? But they will overbuild. 00:10:34.820 --> 00:10:37.570 And then you have the folks who are actually providing the application late. 00:10:38.370 --> 00:10:41.400 And then you have the incumbents, the people who have what we call the systems 00:10:41.400 --> 00:10:46.470 of record, the databases, the software, the large software companies that they use today. 00:10:46.910 --> 00:10:50.530 Which one of those three will capture value? Really unclear, right? 00:10:52.230 --> 00:10:55.710 Now, for us, though, as an early stage fund, what's exciting is we get to work 00:10:55.710 --> 00:11:00.600 with very small companies who don't have the innovator's dilemma of the existing 00:11:00.600 --> 00:11:02.980 incumbents, right, and who are very nimble. 00:11:02.980 --> 00:11:08.880 And so they are, in a way, part of writing the future, in a sense. 00:11:09.120 --> 00:11:13.420 Many of them will fail, but the ones that succeed do end up capturing a lot 00:11:13.420 --> 00:11:16.570 of value because that's what we see with every one of these technological inflection 00:11:16.570 --> 00:11:21.030 points is new startups are way more nimble, come in with fresh ideas, 00:11:21.550 --> 00:11:24.890 and are able to just, you know, rewrite the rule of the game. 00:11:25.750 --> 00:11:29.690 But as you navigate it, it is both exciting, it's thrilling. 00:11:29.690 --> 00:11:33.620 And it's obviously, especially for large tech, for big tech companies, 00:11:34.040 --> 00:11:35.150 it's very challenging times. 00:11:35.950 --> 00:11:41.780 Yeah. What would you look back at the last couple of years and think of as you 00:11:41.780 --> 00:11:46.730 and the company's biggest achievements in the space you're working in? 00:11:46.730 --> 00:11:52.240 I mean, I think the company's biggest achievements is when... 00:11:53.350 --> 00:11:59.300 A business that was not built on AI has managed to transform itself to be an AI-first business. 00:11:59.440 --> 00:12:02.260 We have a number of examples that have been very successful. 00:12:02.260 --> 00:12:07.860 I think most notably a company called PostHog that is probably one of our better 00:12:07.860 --> 00:12:12.890 known businesses at 1984 Ventures, where they were the number one product analytics platform. 00:12:13.320 --> 00:12:15.150 The company started it in 2020. 00:12:15.590 --> 00:12:16.960 What do they do, Ramy? 00:12:16.960 --> 00:12:22.580 They provide data and analytics for other software businesses to monitor and 00:12:22.580 --> 00:12:24.130 track how their products are used. 00:12:24.770 --> 00:12:28.720 But it includes everything from how many visitors came to my website, 00:12:29.050 --> 00:12:33.420 to session recordings, to running experiments, to attributing value. 00:12:33.420 --> 00:12:36.790 It's effectively the operating system for online businesses, 00:12:36.790 --> 00:12:41.840 if you will. And it's very successful. It's one of our, you know, best known unicorns. 00:12:42.610 --> 00:12:45.660 It was not an AI company. It was started in 2020. 00:12:46.810 --> 00:12:52.490 And in 2020, beginning of 2020, end of 2024, we were having a lot of conversations 00:12:52.490 --> 00:12:55.080 with the founders and they're starting to see AI. 00:12:55.450 --> 00:12:59.380 And in the beginning of 2025, just something lit in the mind of the founder. 00:13:00.010 --> 00:13:04.820 And he recognized we have to totally rebuild everything, rethink our business 00:13:05.250 --> 00:13:10.130 to first and foremost address AI. You know, every sub-company says we're AI, we're AI, we're AI. 00:13:10.310 --> 00:13:14.660 And what they do, you slap a chat interface on top of your existing product 00:13:14.660 --> 00:13:17.220 and you say, hour AI, you know, the chat interface, you can talk to it. 00:13:17.400 --> 00:13:18.110 Yeah, yeah. 00:13:18.240 --> 00:13:19.270 Nobody cares, right? 00:13:19.530 --> 00:13:19.860 Yeah. 00:13:20.610 --> 00:13:24.680 AI means, it's actually funny because as a consumer, you have a really good 00:13:24.680 --> 00:13:28.610 instinct of what AI is. AI is Instagram or TikTok. 00:13:29.790 --> 00:13:35.000 Extreme personalization. When you launch your Instagram, and not just because 00:13:35.000 --> 00:13:37.830 of the people you follow or your friends, just when you go to the videos, 00:13:37.830 --> 00:13:43.110 it knows you so intimately, it knows your interests, and it's able to tailor the feed to you, 00:13:44.550 --> 00:13:47.930 Your closest friend, your best friend, if you launched your Instagram, it's different. 00:13:48.720 --> 00:13:53.150 It's not the same. That level of extreme personalization is at the core of AI. 00:13:53.270 --> 00:13:59.770 Beyond just the personalization, we want human-like one-to-one communication, 00:14:00.540 --> 00:14:03.950 across every modality, whether it's voice, text, and soon video. 00:14:04.980 --> 00:14:09.590 And it's one-to-one communication, as in the AI knows who you are when it talks 00:14:09.590 --> 00:14:13.550 to you. It talks to you based on your context and your insights and your data. 00:14:15.130 --> 00:14:20.560 And then last, we need to do all of that was like feedback, data loops and infrastructures. 00:14:20.560 --> 00:14:23.340 And then like then now the product starts to look like an AI product. 00:14:24.400 --> 00:14:29.060 But to do that also, I need to transform my actual organization into being an AI organization. 00:14:29.440 --> 00:14:33.340 What's an AI organization? God, it's not like giving everybody access to Cloud. 00:14:33.340 --> 00:14:38.750 Everybody's like, oh, hey, go take access to Cloud or we give our businesses all access to chat GPT. 00:14:39.690 --> 00:14:43.210 No, it's really about rethinking from the ground up what function should be 00:14:43.210 --> 00:14:46.010 done by humans. and what functions should be done by machine. 00:14:46.260 --> 00:14:50.170 How many companies that you're dealing with are seeing this future? 00:14:50.600 --> 00:14:55.910 Are some resisting? Are some trying to catch up? Are they already there? 00:14:56.430 --> 00:14:59.600 I mean, where have the challenges been for you? 00:15:00.590 --> 00:15:05.590 So the biggest challenge is for an existing company who has a lot of revenue. 00:15:06.750 --> 00:15:09.780 And this is just a textbook example of the innovator's dilemma. 00:15:10.910 --> 00:15:15.540 The innovator's dilemma, the old Clay Christensen case study around how companies, 00:15:16.210 --> 00:15:17.830 don't fail because they're doing the wrong thing. 00:15:18.480 --> 00:15:22.350 They fail because they continue to do the right thing when it's no longer relevant, 00:15:23.530 --> 00:15:26.300 when it's no longer the right thing. And that is sort of the irony. 00:15:26.640 --> 00:15:33.440 And so we started investing from fund one in 2018, 2018 to 2021, fund two is 2021 to 2024. 00:15:34.050 --> 00:15:38.470 I would say our fund one companies and half our fund two companies have had to adapt. Yeah. 00:15:40.760 --> 00:15:46.110 But then investments starting from like 2023 up until today, they're all AI native. 00:15:46.660 --> 00:15:54.300 They're all AI native. Yeah, yeah. So Ramy, you're in this world. You're deep into it. 00:15:54.680 --> 00:15:58.300 You're looking at all these incredible innovative companies looking at the future. 00:15:59.030 --> 00:16:06.750 I'm sitting in Cairo and I am not tech savvy in terms of using AI. 00:16:08.040 --> 00:16:13.440 I'm just beginning to use it on a practical basis but what I am doing is I've 00:16:13.440 --> 00:16:15.220 been educating myself a lot about 00:16:15.220 --> 00:16:20.660 it in terms of what it means for society what it means for us as humans 00:16:21.930 --> 00:16:26.850 I'm reading I've been doing this for the last year and a half and I'm beginning 00:16:26.850 --> 00:16:32.190 to get involved in events and panels where AI is obviously a huge focus. 00:16:33.720 --> 00:16:38.700 One of the big thinkers that I, you know, latched onto, I think partly because 00:16:39.020 --> 00:16:43.740 he writes in a very simple manner for someone like me, who's not tech like you, 00:16:43.740 --> 00:16:47.070 is Mustafa Suleiman, who runs Microsoft AI. 00:16:47.470 --> 00:16:52.130 He did this incredible book in 2020 called The Coming Wave, which he's updated 00:16:52.130 --> 00:16:56.120 since then. Obviously, it's no longer coming. The wave is washing over us. 00:16:56.120 --> 00:16:57.680 It's looked behind us. We're in it. 00:16:59.120 --> 00:17:05.920 We've swam deep into the sea. And what I like about him is he doesn't make it 00:17:05.920 --> 00:17:10.140 look like it's all going to come up roses. He doesn't make it look like it's 00:17:10.140 --> 00:17:11.220 all going to be positive. 00:17:11.520 --> 00:17:16.930 He really does address the scary side and the need for containment, 00:17:16.930 --> 00:17:19.300 the need for regulation, the need for... 00:17:20.870 --> 00:17:25.190 Controlling this, you know, this tool 00:17:27.350 --> 00:17:32.320 And in my mind i always try and separate this sort of big picture idea of where 00:17:32.320 --> 00:17:36.220 are we going where will we be with this industry 00:17:37.290 --> 00:17:42.230 and then of course drilling down like you do into practical what the companies 00:17:42.230 --> 00:17:47.120 are doing how they're providing these incredible services and tools for us as 00:17:47.620 --> 00:17:51.590 people who want to be more productive, more creative, all of that. 00:17:52.050 --> 00:17:56.620 But I can't help feel really scared, Ramy. 00:17:58.030 --> 00:18:01.950 I'm really quite scared of what's coming. You know, a good example, 00:18:01.950 --> 00:18:07.980 we had the Anthropic issue with the Ministry of Defense or Ministry of War, 00:18:09.460 --> 00:18:15.630 where Anthropic, being a private company, did what a state should do. 00:18:15.800 --> 00:18:18.650 And they said, I'm sorry, I'm not going to sell you this, 00:18:21.990 --> 00:18:26.840 You know, this software, this technology, if you're going to use it to surveil 00:18:26.840 --> 00:18:30.620 Americans, if you're going to use it for autonomous weaponry. 00:18:31.100 --> 00:18:36.870 So, I mean, people who know this world will know the issue. I'm not going to go into the issue. 00:18:38.470 --> 00:18:43.460 So that was one thing that a lot of people suddenly went, light bulb moment went off. 00:18:43.790 --> 00:18:48.220 The state was being the aggressor here and being the sort of the evil one. 00:18:48.700 --> 00:18:55.150 And the corporation was being the one trying to safeguard and protect people, 00:18:55.520 --> 00:18:58.020 which most of the time it's the opposite way around. 00:18:58.020 --> 00:19:02.490 And then, of course, we've got the war, which we're in now, where there's no 00:19:02.490 --> 00:19:08.450 doubt in my mind, autonomous weaponry is being used with either no human supervision 00:19:08.790 --> 00:19:10.660 or the bare minimum supervision. 00:19:11.630 --> 00:19:15.640 How do you think of these issues? 00:19:15.820 --> 00:19:20.140 You know, there's many fears being talked about about AI. There is sort of the 00:19:20.140 --> 00:19:23.120 existential fear of AI taking over the world. 00:19:24.400 --> 00:19:28.760 I think that that might happen. I think there are fears around the use of AI 00:19:28.760 --> 00:19:32.260 surveillance and the use of AI at war. 00:19:33.050 --> 00:19:36.690 I think, I hate to say it, but the US is a laggard in that market. 00:19:37.060 --> 00:19:43.590 China's leading the way. You know, I was in, I think I was in Bahrain in 2019 where China, 00:19:44.110 --> 00:19:49.970 in a conference, where China was touting the face-to-recognition software and 00:19:49.970 --> 00:19:52.310 sort of full government control software. 00:19:53.730 --> 00:19:56.670 They're obviously way ahead of us in drones. They're way ahead of us in the US, 00:19:57.190 --> 00:20:00.280 so like just as somebody in the US. 00:20:02.380 --> 00:20:05.800 I'm not sure if we're really.. I think, if anything, we have some catch up to do, 00:20:07.010 --> 00:20:12.100 we really do have some catch up to do, right? In terms of the future of the war, 00:20:13.170 --> 00:20:15.060 if you think of if you actually go back, 00:20:16.800 --> 00:20:22.850 the US did not win World War II by having the most powerful weapons it's actually 00:20:22.850 --> 00:20:27.760 the opposite you know the US won World War II by having crazy strong supply chain, 00:20:28.790 --> 00:20:33.320 and then building them faster that the Germans can sink them that was the model, 00:20:33.830 --> 00:20:35.190 you're never going to build something as good as, 00:20:37.070 --> 00:20:40.730 a T1 German tank just so much more advanced than we are but we can build six, 00:20:42.010 --> 00:20:44.420 And the same thing with boats. And if you ever actually, if you ever come to 00:20:44.420 --> 00:20:47.230 the US, you go to the, there's a New Orleans, a Museum of World War II in New Orleans. 00:20:47.230 --> 00:20:50.850 It's incredible that it tells a story about World War II and really emphasizes 00:20:51.170 --> 00:20:57.210 this was a supply chain war, a war won by taking Ford, which was the biggest 00:20:57.210 --> 00:21:00.330 automotive company, and saying, stop making cars and make airplanes. 00:21:00.620 --> 00:21:04.530 Shitty, tiny little thinkable airplanes, literally like the Shahd drones. 00:21:04.690 --> 00:21:05.780 And we just need a lot of them. 00:21:05.780 --> 00:21:08.800 Which is exactly what the Shad drones are doing. And then we're, 00:21:08.800 --> 00:21:11.940 you know, and the US is actually trying to copy the Shahd drones. 00:21:12.060 --> 00:21:16.250 But, you know, just talking about this existential element for a minute, 00:21:16.250 --> 00:21:19.110 we're so deep into this new world now. 00:21:19.760 --> 00:21:25.160 And, you know, people have been talking about we must regulate, 00:21:25.160 --> 00:21:27.870 we must control, we will, we will, we will. 00:21:28.530 --> 00:21:30.590 Hasn't happened. It's not going to happen. 00:21:31.560 --> 00:21:37.720 Governments are not going to get together and do some kind of agreements together. 00:21:38.200 --> 00:21:43.090 This is a race to the finish by big tech companies. 00:21:43.090 --> 00:21:47.830 And, you know, governments around the world have proven to be completely, 00:21:49.110 --> 00:21:57.960 you know, unable to manage their current issues, let alone what we're beginning to move into. 00:21:58.740 --> 00:22:04.210 So no, no, no, it is but I also you know I liken it again to the nuclear race in 00:22:04.210 --> 00:22:08.920 the late 40s early 50s but nobody was going to sit down and do a nuclear non-perforation 00:22:08.920 --> 00:22:12.550 treaty people are like I gotta get there first, you know, I gotta it's. 00:22:12.550 --> 00:22:16.390 Different, Ramy. It's different because I thought of that as a good analogy but 00:22:16.390 --> 00:22:23.160 the difference is is that nuclear weapons were it was one form of tech and it had one use. 00:22:23.980 --> 00:22:29.170 AI is omni-tech. So it's available throughout everything we're going to be doing 00:22:29.170 --> 00:22:31.690 for the rest of our lives for the next forever. 00:22:32.440 --> 00:22:35.540 So how do you regulate something that's omni-use? 00:22:38.480 --> 00:22:40.760 You can't, you can't, Ramy, you can't. 00:22:40.760 --> 00:22:43.760 I mean, look, you can count in the Europeans to over-regulate it. 00:22:46.520 --> 00:22:49.320 You can count in China to sort of, 00:22:51.680 --> 00:22:55.040 state control it, not regulate it, but state control it. And then you can count 00:22:55.040 --> 00:23:00.080 in the US to like kind of slightly under-regulate, slightly over-regulate. 00:23:00.100 --> 00:23:01.180 Fudge it, yeah. 00:23:01.960 --> 00:23:04.010 To kind of fudge it, right? I mean, it's basically where you're going to net 00:23:04.010 --> 00:23:07.410 out. But I want to go back to the employment for a moment. 00:23:08.080 --> 00:23:14.920 Yes. Because, like, you know, will AI, like, destroy humanity as we know it? 00:23:16.230 --> 00:23:18.230 It's a possibility, but also, 00:23:20.580 --> 00:23:24.930 that same applies to nuclear weapons, the same applies to chemical weapons, 00:23:24.930 --> 00:23:27.280 the same applies to many other things. It's definitely a threat. 00:23:27.280 --> 00:23:28.700 It's just a threat that we have to understand. 00:23:29.940 --> 00:23:31.560 But does AI have the potential? 00:23:33.300 --> 00:23:37.720 To cause economic havoc in the next five years. In the next five years, 00:23:37.720 --> 00:23:39.150 not like the next 20 years. 00:23:39.530 --> 00:23:40.800 Maybe even less. 00:23:41.160 --> 00:23:41.740 For sure. 00:23:41.890 --> 00:23:42.950 Yeah, yeah, yeah. 00:23:43.400 --> 00:23:45.890 And a lot of folks are saying, a lot of folks, the counter argument, 00:23:46.040 --> 00:23:47.680 and I really worry about this. 00:23:48.190 --> 00:23:48.440 Yeah. 00:23:48.580 --> 00:23:55.830 You know, we have our entire world is built today around 2% inflation or 5% 00:23:55.830 --> 00:23:59.440 inflation around economic growth, around prosperity. 00:23:59.640 --> 00:24:01.460 It's almost like a massive pyramid scheme. 00:24:02.440 --> 00:24:03.900 And it's so precarious. 00:24:03.900 --> 00:24:06.780 One thing changes and the whole thing is wobbly. 00:24:07.060 --> 00:24:10.630 A lot of folks, the counter argument to these concerns on unemployment is always 00:24:10.630 --> 00:24:13.790 like, oh, don't worry. The technology always creates new jobs. 00:24:14.590 --> 00:24:17.120 I'm sitting here scratching and I'm like, yeah, I get it. I get it. 00:24:17.120 --> 00:24:23.130 Like when we switched from the horse carriages to the automotive, it's a classic story. 00:24:23.130 --> 00:24:25.670 Like, you know, all the people that worked in the horse carriages, 00:24:25.670 --> 00:24:28.930 like lots of jobs, but then you got mechanics and drivers and yet, 00:24:29.290 --> 00:24:33.480 yes, that took freaking 30 years. It took 30 years, right? 00:24:33.850 --> 00:24:38.530 When women, when we switched from typing on a typewriter into using a computer, 00:24:38.530 --> 00:24:40.800 into a personal computer, people had to reskill. 00:24:41.140 --> 00:24:45.530 They had a tenure to stop using, you know, to stop using typewriter and learn 00:24:45.530 --> 00:24:46.400 how to use a word processor. 00:24:47.910 --> 00:24:53.830 I see developers today, software engineers, people who should be at the forefront of using technology, 00:24:54.660 --> 00:24:59.200 who are struggling to use the most advanced AI coding, struggling to use it, 00:24:59.830 --> 00:25:03.510 because it changes their entire framework of how I write code. 00:25:04.150 --> 00:25:06.300 It changes their sense of self-worth. 00:25:07.200 --> 00:25:12.210 That previously, I was an engineer whose work was precious. Now I just tell 00:25:12.210 --> 00:25:14.570 Claude to do something and it does it for me. 00:25:14.910 --> 00:25:16.430 And your work was in demand and needed. 00:25:16.430 --> 00:25:19.530 I mean, it's really hilarious. But if I show you this, this is just my actual 00:25:19.530 --> 00:25:23.510 library. This isn't a point to every book here is irrelevant. 00:25:23.880 --> 00:25:26.820 Like this is something called, you know, the mythical man month. 00:25:27.130 --> 00:25:29.700 This was the ultimate book about software engineering in groups. 00:25:30.100 --> 00:25:31.430 Totally, totally irrelevant. 00:25:31.850 --> 00:25:35.310 This is a book in like compiler, modern compiler work. Oh, that knowledge is irrelevant. 00:25:35.310 --> 00:25:37.930 It's gone out the window. It's gone out the window. And it's shocking. 00:25:38.290 --> 00:25:40.890 And then you have to rescale. And that's a software engineer. 00:25:41.460 --> 00:25:44.440 And then rescale to a world. Now, will we want more software engineers? 00:25:45.110 --> 00:25:48.630 Sure. We probably would want more software to be built. And we will need more 00:25:48.630 --> 00:25:51.220 infrastructure and so forth. But I think about lawyers. 00:25:52.560 --> 00:25:56.170 I think about accountants. I think about brokers. I think about the entire world of the economy. 00:25:56.890 --> 00:26:01.070 And then I think of, and again, I think, frankly, we're fortunate in countries 00:26:01.070 --> 00:26:03.640 like Egypt where we haven't yet switched to a services economy. 00:26:04.640 --> 00:26:08.470 We're still a goods economy to some extent. And these are way more resilient. 00:26:08.470 --> 00:26:09.030 Pencil and paper economy. 00:26:10.050 --> 00:26:10.670 Yeah, yeah, yeah. 00:26:11.070 --> 00:26:13.260 Pencil and paper economy, exactly. And a lot of people, and, 00:26:13.260 --> 00:26:17.400 you know, it's not the, you know, money isn't being made by low-region accounts. 00:26:17.400 --> 00:26:21.290 It's being made like from real estate and from food. So there is a chart philanthropic 00:26:21.290 --> 00:26:25.380 that I really think is worth taking a look back at. And the chart lists every job. 00:26:26.840 --> 00:26:35.810 And how much of it, you know, is being, you know, automated today and how much 00:26:35.810 --> 00:26:38.770 we think we might be able to automate in the future, right? 00:26:39.440 --> 00:26:43.700 And at the top of this list, at the top of this list, you see computer and math. 00:26:44.630 --> 00:26:50.020 Computer programming, 96% automatable, right? Business and finance, 94% automatable. 00:26:50.730 --> 00:26:55.860 Legal, 88% automatable. And then at the very bottom, you see agriculture, 00:26:56.500 --> 00:26:58.590 you see food and servicing, you see construction. 00:26:59.570 --> 00:27:03.010 And you can very quickly recognize, well, how much of my GDP comes from where? 00:27:03.010 --> 00:27:04.240 Yeah, yeah, yeah. 00:27:04.310 --> 00:27:08.360 And instantly get an assessment of where it is false. And the more advanced 00:27:08.360 --> 00:27:12.010 the society, the bigger the challenge. So we need to reskill. 00:27:12.760 --> 00:27:16.260 But we don't know what the new skills are. And we have to reskill very quickly. 00:27:16.450 --> 00:27:21.530 So that brings me to a question for you, Ramy. So I have, you met my children. 00:27:21.990 --> 00:27:24.850 Ali is one more year of university after this one. 00:27:25.430 --> 00:27:30.490 And he is, like all his friends, struggling to find internships, 00:27:31.340 --> 00:27:33.530 doesn't know if he's going to have a job at the end of this. 00:27:34.630 --> 00:27:38.730 What is he employable for? It's a real, real problem. 00:27:39.280 --> 00:27:42.980 My younger son is about to go to university in September, in August. 00:27:43.570 --> 00:27:47.680 And, you know, I'm thinking, you know, what should he be studying? 00:27:47.680 --> 00:27:50.920