WEBVTT 1 00:00:00.160 --> 00:00:06.280 So all the tokens will go on to Fetch, so they will become a 2 00:00:06.400 --> 00:00:13.240 Fetch token, but Fetch token will be renamed as Aside token, So for 3 00:00:13.359 --> 00:00:20.679 Fetch holders, it's a name change and a network upgrade. For Ocean and 4 00:00:20.800 --> 00:00:29.719 Singularity tokens, it's a conversion whereby their tokens will be converted into Aside token. 5 00:00:38.079 --> 00:00:40.880 This content is brought to you by Uphold, which is one of the 6 00:00:40.920 --> 00:00:46.000 top crypto platforms out there which allows you to easily buy cell trade Bitcoin and 7 00:00:46.039 --> 00:00:50.320 the top or coins in the market. Uphold lists two hundred and sixty plus 8 00:00:50.320 --> 00:00:54.719 cryptocurrencies. They also allow you to trade precious metals such as gold, silver, 9 00:00:54.799 --> 00:00:58.240 palladium, and platinum. I've been a user of Uphold since twenty eighteen, 10 00:00:58.280 --> 00:01:02.359 so I can voucher this platform. They have full transparency reports and they 11 00:01:02.359 --> 00:01:06.200 don't commingle or lend out your funds. They also have a great product called 12 00:01:06.280 --> 00:01:11.519 Vault, which is an assisted self custody product. A vault allows you to 13 00:01:11.640 --> 00:01:17.200 maintain custody of your funds and the keys are split, so Uphold holds one 14 00:01:17.280 --> 00:01:21.480 and you hold two, and if there's any issues if you lose one of 15 00:01:21.519 --> 00:01:26.439 your keys, Uphold will help restore these keys and you can maintain access to 16 00:01:26.480 --> 00:01:30.239 your funds. It's twenty four to seven instant trading, it's trustless, so 17 00:01:30.400 --> 00:01:34.480 this is a great feature that can give you peace of mind where you don't 18 00:01:34.519 --> 00:01:40.079 have to worry about if you lose your private keys and much more so. 19 00:01:40.120 --> 00:01:42.400 If you'd like to learn more about uphold, please visit the link in the 20 00:01:42.439 --> 00:01:48.359 description. Welcome to the Thinking Crypto podcast, your home for cryptocurrency news and 21 00:01:48.480 --> 00:01:53.239 interviews with me. Today's Humayun Shake. Who's the CEO and founder of fetch 22 00:01:53.319 --> 00:01:57.560 dot ai and the chairman of the Artificial Superintelligence Alliance. Who are you in 23 00:01:57.599 --> 00:02:01.760 great to have you on. I appreciate your invite, Thank you, thanks 24 00:02:01.760 --> 00:02:06.079 for having me Whomy, and i am so excited to speak with you. 25 00:02:06.280 --> 00:02:09.599 And I'm not saying that because I interview people, but I'm a big fan 26 00:02:09.639 --> 00:02:15.759 of fetch dot ai. I'm very fascinated by this merging of these three AI 27 00:02:15.919 --> 00:02:19.960 driven tokens, and we're going to talk about all that. But before we 28 00:02:20.000 --> 00:02:23.280 get there, tell us about yourself. Where you're from and what's your professional 29 00:02:23.319 --> 00:02:28.240 background. Yeah, I'm from Pakistan. I've been in the UK twenty five 30 00:02:28.319 --> 00:02:32.680 years now, so I came as a student and then I stayed here. 31 00:02:34.879 --> 00:02:38.800 A background just to give you some some professional background a little bit. I 32 00:02:38.840 --> 00:02:46.000 trained as a computer scientist, joined here in the UK, joined the gaming 33 00:02:46.039 --> 00:02:53.159 industry quite early on, and then from there I started. I met Demis 34 00:02:53.240 --> 00:02:59.719 Hasavis, who was the founder of DeepMind, so based on that we kind 35 00:02:59.759 --> 00:03:08.280 of got together and I got into discussing artificial intelligence and AGI concepts and working 36 00:03:08.319 --> 00:03:14.080 on some of those core concepts at that point, and gaming was really a 37 00:03:14.120 --> 00:03:19.240 great place to try these things out, as you probably already know, so 38 00:03:19.280 --> 00:03:23.800 we started trying those out. Then I became part of Google deep Mind. 39 00:03:23.879 --> 00:03:28.560 Early on, I was the first investor in deep Mind. I was also 40 00:03:28.599 --> 00:03:34.240 looking after the commercialization of AI in deep Mind, so that's how I got 41 00:03:34.319 --> 00:03:42.879 into really into AI. And then once Google bought DeepMind and I exited from 42 00:03:42.919 --> 00:03:47.439 deep Mind, I started kind of putting together my thoughts on what was right, 43 00:03:47.719 --> 00:03:52.520 what happened good, what happened bad in deep Mind. And one of 44 00:03:52.520 --> 00:03:57.039 the things which was quite interesting which I felt I needed to work on, 45 00:03:57.360 --> 00:04:05.159 was the commercialization and actually break AIM too much more absorb absorbable and granular kind 46 00:04:05.199 --> 00:04:11.240 of components so people can actually interact with it and use it. That's how 47 00:04:11.520 --> 00:04:15.920 I kind of came to fetch and the agent based framework. So that's kind 48 00:04:15.959 --> 00:04:20.000 of the story behind it a little bit. And I'm curious, So you 49 00:04:20.079 --> 00:04:25.560 have a plathora of experience with AI, and that's a big part of your 50 00:04:25.680 --> 00:04:30.199 experience. But where did you encounter blockchain and crypto? And then you know, 51 00:04:30.199 --> 00:04:34.199 obviously now there's a conversience of those two technologies, but what was your 52 00:04:34.279 --> 00:04:38.439 like first encounter bitcoin? Where did you hear about it? And you know, 53 00:04:38.480 --> 00:04:43.120 what was your harm on it? Yeah, so I've known, I 54 00:04:43.199 --> 00:04:46.240 must admit, I mean, I didn't get into bitcoin upfront, but I 55 00:04:46.279 --> 00:04:53.600 did get into a theoeum quite early on. And so so I'm always interested 56 00:04:53.639 --> 00:04:57.639 in new technology, So you know, I was always looking for new technology. 57 00:04:57.720 --> 00:05:01.639 AI was one. Blockchain and crypto was really kind of maturing and growing, 58 00:05:01.720 --> 00:05:05.519 and if you think about ten years ago, fifteen years ago, so 59 00:05:05.600 --> 00:05:12.439 I was I was watching the you know, the space quite considerably. I'm 60 00:05:12.519 --> 00:05:18.120 not overly financial and I'm not into kind of financial engineering as such, but 61 00:05:19.000 --> 00:05:24.879 it was quite interesting what bitcoin was doing. But it became really exciting when 62 00:05:24.959 --> 00:05:30.279 Ethereum came about and we looked at how you could actually execute code and record 63 00:05:30.360 --> 00:05:38.639 it and create you know, actual functioning computing systems on blockchain. So that's 64 00:05:38.680 --> 00:05:42.399 when I really got interested. So I got interested quite quite a lot in 65 00:05:42.399 --> 00:05:46.040 ethereum and looking at what you could use it for. And on the other 66 00:05:46.120 --> 00:05:50.800 side, although I was interested in it, my main focus was still looking 67 00:05:50.839 --> 00:05:58.439 at commercialization of AI. And when I was thinking about how can you actually 68 00:05:58.800 --> 00:06:02.560 make AI into smaller chunks which people can interact with, and how can you 69 00:06:02.639 --> 00:06:06.199 actually use machine learning models? I mean machine learning models have been around, 70 00:06:08.199 --> 00:06:11.920 you know, tens of years, so I was thinking, Okay, at 71 00:06:11.959 --> 00:06:14.959 the moment, somebody has to build it, somebody has to deploy it, 72 00:06:15.000 --> 00:06:18.759 and it all goes back into somebody who has the data who can use it, 73 00:06:19.240 --> 00:06:24.519 and generally it's these big corporations who can use it. Again, an 74 00:06:24.519 --> 00:06:29.040 example is Google and Microsoft and the likes. And when we exit it to 75 00:06:29.120 --> 00:06:32.959 Google, it became quite clear that training these models quite expensive, and that 76 00:06:33.160 --> 00:06:39.399 really was one of the core things that we needed to be aware of that 77 00:06:39.480 --> 00:06:45.399 without compute you can't really do anything, and the compute is what is expensive. 78 00:06:45.439 --> 00:06:47.120 And you know, if you think about what Google has, it's a 79 00:06:47.160 --> 00:06:55.959 cloud infrastructure compute infrastructure. So how can you then bring it to a place 80 00:06:56.079 --> 00:07:00.600 where you know, normal people can interact with it or build solutions and smaller 81 00:07:00.639 --> 00:07:06.079 companies can build solutions. And it kind of really fitted quite well with having 82 00:07:06.160 --> 00:07:14.839 a blockchain which could actually record your participation. It could actually orchestrate the agent 83 00:07:15.040 --> 00:07:21.000 based system because and we talk about why agent based systems, I felt to 84 00:07:21.160 --> 00:07:26.439 really get AI to the masses, we need to have agent based system. 85 00:07:27.000 --> 00:07:33.279 And to manage an agent based system, which is that expansionary and that all 86 00:07:33.399 --> 00:07:39.199 encompassing, you need to have a ledger technology which actually keeps a record of 87 00:07:39.279 --> 00:07:42.560 everything. Now you could do that in a centralized way. You can give 88 00:07:42.560 --> 00:07:46.560 it to a big centralized company, or you could actually make it a bit 89 00:07:46.600 --> 00:07:51.959 more decentralized where you know, the communities can manage it and you can still 90 00:07:53.000 --> 00:07:58.920 build solutions and you can still deploy them. So that's when I got quite 91 00:07:58.920 --> 00:08:03.639 excited about bringing the two together. And this is now eight seven eight years 92 00:08:03.639 --> 00:08:09.160 ago, and that's when we started looking at it. Wow, that's quite 93 00:08:09.160 --> 00:08:13.199 a journey. It was a while ago. And you know, especially in 94 00:08:13.279 --> 00:08:18.480 these industries, the technology moves so fast on both sides. So tell us 95 00:08:18.519 --> 00:08:24.519 about FETCH and what's its mission capabilities and what folks can do with it. 96 00:08:26.199 --> 00:08:30.240 Yeah, so so just kind of tying into what I was saying before. 97 00:08:30.920 --> 00:08:35.600 It was quite clear that agent based systems are going to become to me, 98 00:08:35.720 --> 00:08:41.120 it was quite clear that agent based systems will become the norm for commercializing any 99 00:08:41.399 --> 00:08:46.519 application building in AI space. But what was missing? So Fetch came about 100 00:08:46.519 --> 00:08:52.679 around five years ago, six years probably if you think about the inception as 101 00:08:52.720 --> 00:08:58.200 well. So six years ago, we didn't have this LLLM craze. Although 102 00:08:58.279 --> 00:09:01.639 LLM technology, the transformative knowledgy was there. We also use that in Deep 103 00:09:01.720 --> 00:09:07.279 Mind. We used to train you know, game agents to play the game 104 00:09:07.559 --> 00:09:13.440 just based on reading instruction set. So we were doing that already, but 105 00:09:13.559 --> 00:09:18.879 not to that scale where you know, the elms became all encompassing and what 106 00:09:20.120 --> 00:09:24.559 Opening Eye did was quite amazing in terms of how they opened up this whole 107 00:09:24.559 --> 00:09:28.720 commercialization. But but we I did have this thinking that yes, we're going 108 00:09:28.759 --> 00:09:37.720 to have one these models which will fully understand what you're saying to them. 109 00:09:37.879 --> 00:09:41.679 Now, I don't want to go into the sentient or not sentient, or 110 00:09:41.039 --> 00:09:45.960 how much they actually understand and not understand. But let's just assume that they 111 00:09:46.039 --> 00:09:50.679 understand you well enough to, you know, very convincingly give you arguments and 112 00:09:52.279 --> 00:09:56.279 make their point and they understand what your instruction set is or what your objective 113 00:09:56.320 --> 00:10:01.919 could look like. But what also was clear was okay, what next, 114 00:10:01.000 --> 00:10:03.039 right, So what are you going to do with it? Right? You 115 00:10:03.120 --> 00:10:07.240 understand the language, you can talk, and you can talk the talk, 116 00:10:07.279 --> 00:10:09.799 but can you do the walk? And that was the kind of the moment 117 00:10:09.840 --> 00:10:16.519 when you think, right, so somebody will understand the language, but how 118 00:10:16.559 --> 00:10:20.159 can you kind of create a chain of thought where you can actually start putting 119 00:10:20.720 --> 00:10:26.720 actions together. And what that also opened up as an understanding was that the 120 00:10:26.799 --> 00:10:33.679 search and discovery will change because at the moment, we're just searching for information 121 00:10:33.919 --> 00:10:39.679 and when we can come to search for more actionable items. So for example, 122 00:10:39.960 --> 00:10:43.519 if you are going to come and book for a ticket on an airplane 123 00:10:45.639 --> 00:10:48.519 a flight, or you want to book a hotel, these are action. 124 00:10:48.919 --> 00:10:52.679 Or you want to order a cab, that's action. It also became very 125 00:10:52.679 --> 00:10:58.200 clear that Google wasn't really doing the job right. Hence you needed all these 126 00:10:58.240 --> 00:11:03.080 aggregators to come in because these aggregators were putting all this information together and actually 127 00:11:03.159 --> 00:11:07.919 enabling you to create these actionable searches. Right, so if you want to 128 00:11:09.159 --> 00:11:11.919 book a hotel, you go to booking dot com. If you want to 129 00:11:11.919 --> 00:11:15.840 book a flight, you go to booking dot com. You know, there 130 00:11:15.879 --> 00:11:20.919 were incumbents like Amadaeus, which was aggregating all of these things. How do 131 00:11:20.039 --> 00:11:30.200 you then think the paradigm shift will come when the language understanding is really good? 132 00:11:31.919 --> 00:11:37.600 But how do you integrate this with their legacy systems so you can actually 133 00:11:37.960 --> 00:11:43.039 understand your query and actually you do the search rather than just dumping thirty thousand 134 00:11:43.120 --> 00:11:48.200 websites in front of you. So that was the AHA moment really when we 135 00:11:48.320 --> 00:11:52.799 thought, okay, you know this, I could see with LLLMS, we're 136 00:11:52.840 --> 00:11:56.759 going to change the way we search, and ultimately we're going to change the 137 00:11:56.799 --> 00:12:03.360 way the actionable search also happened, and you know when we started Fetch, 138 00:12:03.440 --> 00:12:07.440 that was the concept behind it. That was the thinking behind it. Of 139 00:12:07.480 --> 00:12:11.320 course, I mean it takes time for things to kind of come to fruition, 140 00:12:11.799 --> 00:12:16.840 and you know when open Aye released it, well kind of that was 141 00:12:16.399 --> 00:12:20.279 that was my moment of saying, okay, I told you so, look 142 00:12:20.360 --> 00:12:26.399 this is coming and that kind of so we were thinking slightly ahead of that 143 00:12:26.559 --> 00:12:31.200 because we also felt that l ELM is a big boys game. You need 144 00:12:31.240 --> 00:12:35.399 a lot of capital, because we had already seen that in the deep mind 145 00:12:35.480 --> 00:12:41.600 days, training models is not cheap. So what also became quite clear is 146 00:12:41.600 --> 00:12:48.039 that if we're going to as a smallish entity going to make a difference, 147 00:12:48.559 --> 00:12:52.200 we need to sit above it. And l elms will become very commoditized because 148 00:12:52.200 --> 00:12:56.720 it will be a game of three or four people, which you're now already 149 00:12:56.759 --> 00:13:01.480 seeing. You can see you know, inflection, know things, things are 150 00:13:01.559 --> 00:13:07.120 changing, what's been happening with open ai. It's unless you get the Microsoft 151 00:13:07.240 --> 00:13:11.799 so the Googles of the world funding it, financing it and videos of the 152 00:13:11.799 --> 00:13:16.279 world, you can't really train a foundational LLM to the right extent. Of 153 00:13:16.320 --> 00:13:20.600 course, techniques are going to improve, you're not going to need that much 154 00:13:20.639 --> 00:13:24.720 compute, you're going to not need that much data, but at this point 155 00:13:24.720 --> 00:13:28.720 in time they're going to be commoditized. Some big corporation will do it, 156 00:13:28.039 --> 00:13:35.559 offer it as a package alongside the other cloud compute for example. So you 157 00:13:35.639 --> 00:13:39.200 know I've been I've been talking quite a bit. So if you have anything 158 00:13:39.240 --> 00:13:43.679 to kind of ask or ad please do no, no, that's that's great 159 00:13:43.000 --> 00:13:48.039 insight. I love that because it gives a lot of perspective and on the 160 00:13:48.840 --> 00:13:52.679 problems you're trying to solve and the dynamics of being in the market with the 161 00:13:52.759 --> 00:14:01.360 big players like you mentioned. Tell us about the merge of FETCH, Ocean 162 00:14:01.399 --> 00:14:05.320 Protocol and s net to form the Artificial Superintelligence Alliance. How did this come 163 00:14:05.360 --> 00:14:15.480 about and what's what's the goal of merging. Yeah, so so obviously we've 164 00:14:15.519 --> 00:14:18.039 spent I mean, as I said, fetches four or five years old. 165 00:14:20.120 --> 00:14:24.679 We've been working quite aggressively on the agent framework. We did some release it. 166 00:14:24.120 --> 00:14:28.360 I mean, when you when you're kind of the first few, there's 167 00:14:28.360 --> 00:14:33.559 a lot to explore. And that's what we were doing. And what was 168 00:14:33.639 --> 00:14:37.919 quite clear was that there's a lot of components if you're going to take this 169 00:14:39.039 --> 00:14:43.639 technology, if you're going to commercialize the technology. My focus has always been 170 00:14:43.759 --> 00:14:48.960 how do we commercialize AI, how do we take it right to the consumer. 171 00:14:50.600 --> 00:14:54.600 And it became also very clear was that you need compute because yes, 172 00:14:54.759 --> 00:14:58.840 although foundational models are sitting there, you still need compute because you're going to 173 00:14:58.879 --> 00:15:03.679 be new models. So give you an example. You know, I want 174 00:15:05.559 --> 00:15:09.559 to go from point A to point B, right that's my objective. I 175 00:15:09.600 --> 00:15:15.559 want to go from London to Paris. Right now, it's it's a statement, 176 00:15:15.600 --> 00:15:22.600 but it has many components. Right, what's my address is it? 177 00:15:24.080 --> 00:15:26.480 What's the weather like? How am I going to travel from point A to 178 00:15:26.480 --> 00:15:31.519 point B? Am I going to take a cab? A train? Am 179 00:15:31.559 --> 00:15:35.600 I going to need an umbrella? Am I going to wear warm clothes? 180 00:15:35.240 --> 00:15:39.240 You know, not warm clothes? And what's the weather going to do? 181 00:15:39.600 --> 00:15:43.879 Where am I going to stay? All these components first need to be understood. 182 00:15:43.440 --> 00:15:48.639 So foundational M you can say that and it can give you a beautiful 183 00:15:48.720 --> 00:15:52.120 prose and it says, you know, how wonderful is going to be in 184 00:15:52.360 --> 00:15:54.159 Paris? Is romantic and you know it will give you all those things, 185 00:15:54.159 --> 00:15:58.559 but you know the useful things like well do I need to you know, 186 00:15:58.799 --> 00:16:03.200 kind of wear warm clothes right now like tomorrow? And how do I order 187 00:16:03.240 --> 00:16:06.720 a cab? It won't be able to do. I'm not saying it can't 188 00:16:06.799 --> 00:16:11.879 do forever. It's just right now it needs those components to be brought in 189 00:16:12.000 --> 00:16:17.320 somehow. So you still need to Although you don't need a foundational LM, 190 00:16:17.480 --> 00:16:22.120 you still need lllms, which are specialists, you know, specialist ALM could 191 00:16:22.120 --> 00:16:26.879 be specialist knowledge based, action based, you know, context based, you 192 00:16:26.919 --> 00:16:30.240 can Actually you still need to train quite a lot of these specialist stellar lems, 193 00:16:30.240 --> 00:16:33.320 and I think that's where the world is going to go next. Because 194 00:16:33.320 --> 00:16:37.960 you have the data, you're going to train them, So you need compute. 195 00:16:37.039 --> 00:16:41.679 So that's one the first layer. You need data. How are you 196 00:16:41.720 --> 00:16:45.159 going to put bring data to train these models? Who is going to provide 197 00:16:45.200 --> 00:16:48.120 you that? Why are they going to provide you that? How do you 198 00:16:48.240 --> 00:16:52.039 make sure it's not getting stolen? How do you make sure it's safe and 199 00:16:52.080 --> 00:16:56.759 secure to kind of train these models with that data. So that's the second 200 00:16:56.799 --> 00:17:03.039 component layer, which is the data component. Then you need to build new 201 00:17:03.159 --> 00:17:07.119 types of models. So you know, it's not like just because we now 202 00:17:07.119 --> 00:17:10.799 have foundational lms, let's just pack up and just leave it to them. 203 00:17:10.920 --> 00:17:15.359 No, because you're going to have new let's say, neurosymbolic lms which are 204 00:17:15.359 --> 00:17:22.359 giving you more deterministic, less hallucination, more actionable kind of information. So 205 00:17:22.400 --> 00:17:26.839 you need new models. You need new technologies which use less data more efficient. 206 00:17:26.920 --> 00:17:30.680 How do you get to that? So that's you have this AI layer 207 00:17:32.079 --> 00:17:37.079 or as if you want to extend that further, it's the artificial general intelligence, 208 00:17:37.079 --> 00:17:41.559 because how do you make these models understand properly? And then from there 209 00:17:41.599 --> 00:17:45.759 on you want to go to the artificial superintelligence layer, which is, you 210 00:17:45.799 --> 00:17:48.119 know, way better than humans. It's not just humans, it's way better 211 00:17:48.160 --> 00:17:56.839 than humans. More specialists, more faster, better in kind of making assumptions, 212 00:17:56.880 --> 00:18:00.079 predictions, objectives, all of that. So that's that layer in the 213 00:18:00.079 --> 00:18:03.720 middle. So now if I translate all of that, and then then the 214 00:18:03.799 --> 00:18:08.559 last layer is Okay, now I've got these beautiful models and I've got data 215 00:18:08.599 --> 00:18:12.279 which I've trained it on. But what so what you know, how do 216 00:18:12.319 --> 00:18:17.920 you get to the consumer? So the consumer you have you know, personal 217 00:18:17.960 --> 00:18:21.680 assistance like Siri, like Alexa. I mean you can talk to them, 218 00:18:22.119 --> 00:18:26.559 but for them to do something and have your agency and do something, it's 219 00:18:26.640 --> 00:18:30.720 very difficult. They still can't connect unless you connect them one to one. 220 00:18:30.400 --> 00:18:36.480 So that's where our agent based technology comes in, which is the fetchiest technology, 221 00:18:36.519 --> 00:18:41.039 which which enables you to granularize actions. It kind of takes all the 222 00:18:41.119 --> 00:18:45.359 legacy systems and you can connect an agent to the legacy systems and they can 223 00:18:45.400 --> 00:18:51.119 execute tasks. So now if I rewind back, so we have a layer, 224 00:18:51.279 --> 00:18:56.960 which is the network which Fetch has where you can think about logging transactions, 225 00:18:59.160 --> 00:19:06.599 interactions, orchestration based on what your commercialization model looks like, based on 226 00:19:06.680 --> 00:19:10.839 which AI model you use looks like, what data you use look like. 227 00:19:10.920 --> 00:19:15.799 So you need a recording system. So that's the Fetch network layer. Then 228 00:19:15.839 --> 00:19:19.480 you on top of that is as I said, data Ocean brings that data 229 00:19:19.559 --> 00:19:22.720 technology, which is how do you secure it, how do you make it 230 00:19:22.799 --> 00:19:26.640 safe, how do you monetize it, how do you use the data, 231 00:19:26.680 --> 00:19:30.000 how do you bring compute to data or data to compute, how can you 232 00:19:30.039 --> 00:19:36.359 do multi party computation? All of those things Ocean is very good at. 233 00:19:36.599 --> 00:19:38.359 So we felt, you know, that's the second layer which you need. 234 00:19:38.680 --> 00:19:42.359 And then the third layer, which is Singularity is really good with research. 235 00:19:42.400 --> 00:19:45.599 I mean Ben has been working in this space for twenty I mean he kind 236 00:19:45.599 --> 00:19:52.440 of coined with Shane leg who was a colleague in deep Mind and Ben. 237 00:19:52.559 --> 00:19:56.720 They coined the term AGI together, so so he knows what he's kind of 238 00:19:56.240 --> 00:20:00.119 wanting the AGI to look like, and he's been researching and his team has 239 00:20:00.160 --> 00:20:06.119 been researching. So we need that research to improve the models. And then 240 00:20:06.200 --> 00:20:11.759 finally the Fetch agent framework of the agent platform, which you can connect multiple 241 00:20:11.920 --> 00:20:18.480 llms, machine learning models, legacy systems, APIs, and it automatically finds 242 00:20:18.559 --> 00:20:23.279 a way to connect with them execute the task. That's what the fetch layer 243 00:20:23.440 --> 00:20:30.319 brings. So now if you think it makes logical sense that if a builder 244 00:20:30.480 --> 00:20:34.079 a developer is going to come and build a solution, he or she needs 245 00:20:34.200 --> 00:20:40.880 all of them, all of those components. If you can save them time 246 00:20:41.039 --> 00:20:45.160 in the sense that they don't have to worry about which token do I need 247 00:20:45.200 --> 00:20:48.480 for this service? Which token do I need for this service, and you 248 00:20:48.519 --> 00:20:53.480 can just do full integration, a fully vertically integrated system, which makes their 249 00:20:53.519 --> 00:20:59.839 life very easy, then it absolutely makes sense. Why wouldn't you do it 250 00:21:00.079 --> 00:21:03.319 the question, So that's how I kind of came about with it. And 251 00:21:03.359 --> 00:21:07.920 then obviously I've known both projects, and you know, we worked with Ocean 252 00:21:07.960 --> 00:21:15.200 a lot in terms of the German space where we worked with automotive sector part 253 00:21:15.240 --> 00:21:21.680 of several initiatives in Europe, so we knew them. We knew Singularity quite 254 00:21:21.720 --> 00:21:26.200 while back, so we've been in the space, we've been speaking to each 255 00:21:26.240 --> 00:21:29.359 other, so it kind of made perfect sense to think, Okay, you 256 00:21:29.400 --> 00:21:33.759 know, we need to build a general intelligence, which then moves on to 257 00:21:33.960 --> 00:21:40.559 superintelligence. How are we going to actually get to that point? That's so, 258 00:21:40.640 --> 00:21:45.400 that's what triggered it, and then from there on we put everything together 259 00:21:45.559 --> 00:21:49.480 to make sure that, you know, we can actually get to a point 260 00:21:49.559 --> 00:21:57.079 where you know, it's not it's we didn't want to change anything in terms 261 00:21:57.079 --> 00:22:04.720 of how people manage the foundations, but we just felt that technologically we needed 262 00:22:06.200 --> 00:22:10.839 something to change, and putting them all together makes sense. You gave a 263 00:22:10.839 --> 00:22:17.359 great rundown of the why of how you know these projects are merging together and 264 00:22:17.680 --> 00:22:21.680 the reasons right, and the benefits and so forth. Tell us about the 265 00:22:21.720 --> 00:22:26.200 logistics though, what happens to the three blockchains? Are they merged into one 266 00:22:26.319 --> 00:22:32.039 or they will coexist in the ecosystem? And then also how did the tokens 267 00:22:32.160 --> 00:22:36.960 come into play? And let's say, as a Fetch token holder, what 268 00:22:37.160 --> 00:22:44.799 happens to my tokens under the new ASI system. Yeah, so it's a 269 00:22:44.920 --> 00:22:49.559 merge of tokens and Fetch chain, which is a Cosmo space chain at the 270 00:22:49.599 --> 00:22:56.720 moment. But we have different consensus mechanism we're triling out. There's different features 271 00:22:56.720 --> 00:23:04.400 we're trialing out for AI kind of application build and development. So all the 272 00:23:04.440 --> 00:23:11.839 tokens will go on too Fetch, so they will become a Fetch token, 273 00:23:11.920 --> 00:23:18.799 but Fetch token will be renamed as air Side token. So for Fetch holders, 274 00:23:18.880 --> 00:23:26.279 it's a name change and network upgrade. For Ocean and Singularity tokens, 275 00:23:26.880 --> 00:23:34.559 it's a conversion whereby their tokens will be converted into Aside token based on the 276 00:23:34.599 --> 00:23:40.119 conversion rate that we announced, so I think it's point four three three, 277 00:23:40.359 --> 00:23:45.559 So for every Ocean token or Singularity token, you roughly get point four three 278 00:23:45.720 --> 00:23:53.400 three Aside token. But that means that the number of tokens VEAT holders have 279 00:23:53.960 --> 00:24:00.759 remains the same. But to accommodate and merge those other tokens, we have 280 00:24:00.839 --> 00:24:04.680 to issue new tokens, and the ratio, as I said, is point 281 00:24:04.720 --> 00:24:07.839 four three to three. So I think I do get this question a lot. 282 00:24:07.960 --> 00:24:11.079 Is it going to be valution were it's not because you're actually adding more 283 00:24:11.079 --> 00:24:15.880 water, You're actually adding more value into the pot, so there's no dilution 284 00:24:17.039 --> 00:24:22.119 as such. But you do have those communities joining and merging their token into 285 00:24:22.160 --> 00:24:26.559 this. So at the moment, let's say the you know, the market 286 00:24:26.759 --> 00:24:30.440 cap is, let's you know, I don't know what the detail is. 287 00:24:30.480 --> 00:24:34.319 But let's say the market cap is three billion, and once those two projects 288 00:24:34.440 --> 00:24:37.880 join us, it will become you know, six or seven billion, So 289 00:24:38.319 --> 00:24:41.640 I think that's that's how it would be. So the pot becomes bigger, 290 00:24:42.240 --> 00:24:47.680 the cake becomes bigger, the piece stays the same. So ultimately, in 291 00:24:47.720 --> 00:24:52.839 a absolute way, the value is the same, but as a percentage, 292 00:24:52.880 --> 00:24:59.519 you become percentage off the value. What's your timeline for this and also what's 293 00:24:59.559 --> 00:25:04.119 on your twenty four road map once the merger is completed. Yeah, so 294 00:25:04.680 --> 00:25:11.720 we are carrying out the vote right now. I mean, fetch we kind 295 00:25:11.759 --> 00:25:15.279 of the first vote is in and then the second vote is going to go 296 00:25:15.359 --> 00:25:19.079 in, which will be you know, kind of acceptance of the mergers. 297 00:25:21.559 --> 00:25:30.480 So we expect by in roughly around thirty days that the merger will be approved. 298 00:25:30.039 --> 00:25:37.000 And we're already working quite hard on getting the network upgrades ready. We're 299 00:25:37.039 --> 00:25:41.440 speaking to the exchanges already so that the transition is as smooth as possible. 300 00:25:41.480 --> 00:25:47.720 So over the next thirty to forty days, we expect to move across FAT 301 00:25:48.039 --> 00:25:55.759 to ASI upgrade the network, plus bring in all the other two communities and 302 00:25:55.799 --> 00:26:02.240 they're token onto ASI issue the token or there will be a token exchange for 303 00:26:02.319 --> 00:26:07.319 them and then so that so that kind of takes care of the token. 304 00:26:07.960 --> 00:26:14.640 We then move on to and we're already working together to build the tech kind 305 00:26:14.640 --> 00:26:18.960 of vertical integration, so all the services will just use a SI tokens. 306 00:26:19.000 --> 00:26:25.680 So you can imagine all the all the staking rewards, all the all the 307 00:26:25.720 --> 00:26:30.920 features and functions that the other projects have, apart from what Vetch already has, 308 00:26:30.440 --> 00:26:36.920 all of that will become part of the fetch Slash a side chain, 309 00:26:37.799 --> 00:26:41.559 and then suddenly you will have you'll see a lot more volume going through the 310 00:26:41.599 --> 00:26:48.680 air side chain and ultimately it will benefit us as a whole ASI community. 311 00:26:48.720 --> 00:26:52.759 It will benefit us because the the transactions will go up, the tvls will 312 00:26:52.759 --> 00:26:59.960 go up, and and you know, I can't commit that this will stay 313 00:27:00.079 --> 00:27:03.799 as a Cosmos chain or it will have other features in the Cosmos chain, 314 00:27:03.920 --> 00:27:11.240 but the aside chain will be the coordination chain for AI and that's our objective. 315 00:27:11.319 --> 00:27:15.640 So over the next you know, twelve months, we want a very 316 00:27:17.960 --> 00:27:23.200 reasonably tight integration with all the tech stack. There there are other technologies which 317 00:27:23.240 --> 00:27:30.400 we still need which will probably will will find and integrate more. There is 318 00:27:30.480 --> 00:27:38.039 a lot more coherent integration stack integrated stack, which will be available so you 319 00:27:38.079 --> 00:27:42.559 can use data service, you can use different machine learning models, you can 320 00:27:42.559 --> 00:27:48.000 actually commercialize using agents, you can actually record all of that for audered purpose, 321 00:27:48.599 --> 00:27:52.839 and it will all be done via one token, which is the Airside 322 00:27:52.839 --> 00:27:59.000 token. So that's that's the vision, that's the objective, and we're already 323 00:27:59.000 --> 00:28:04.720 working on it. We all starting to make huge progress because it's just such 324 00:28:04.799 --> 00:28:11.319 complementary technologies that it makes sense and it automatically kind of fits in. They're 325 00:28:11.319 --> 00:28:17.680 not they're not like competing. We're not competing technologies. We're complementary technologies. 326 00:28:18.039 --> 00:28:23.880 So that's that's the great news. And we hope over the next twelve months 327 00:28:23.960 --> 00:28:29.000 to be the chain for AI. And that's really the objective, and the 328 00:28:29.079 --> 00:28:34.880 technology for AI, the platform for deploying AI applications. So that's so those 329 00:28:34.920 --> 00:28:40.359 are the key components that we're technically, we're looking to do in terms of 330 00:28:40.440 --> 00:28:48.640 token. We we expect with the mergers and hopefully as we see more adoption 331 00:28:48.880 --> 00:28:52.440 on our system, we expect to be, you know, hopefully among the 332 00:28:52.480 --> 00:29:00.599 top you know, twenty twenty five tokens. Price goes up, price comes 333 00:29:00.640 --> 00:29:03.119 down. I can't really comment on that, But but I think at least 334 00:29:03.279 --> 00:29:11.000 as a as a project value as a project should you know, multiply rather 335 00:29:11.119 --> 00:29:21.000 than it's it's rather than just additive, So it's it's it's it's an alternative 336 00:29:21.279 --> 00:29:29.960 to the centralized kind of solutions that are available. So we wanted to provide 337 00:29:30.160 --> 00:29:37.440 a decentralized alternative to the AI deployment technology, not just not just deployment, 338 00:29:37.480 --> 00:29:42.079 building, delivering, deploying testing. So the whole text that takes care of 339 00:29:42.119 --> 00:29:47.240 time. The question on that. Obviously, the way you guys are doing 340 00:29:47.279 --> 00:29:52.359 things is decentralized and there's transparency and much more. And in the tech world 341 00:29:52.519 --> 00:29:56.880 we have two things running in parallel because of crypto and blockchain. You have 342 00:29:56.960 --> 00:30:02.160 this decentralized movement, but you also have centralized movements with companies like Google and 343 00:30:02.240 --> 00:30:06.359 open ai and so forth. So I guess my question for you, and 344 00:30:06.400 --> 00:30:10.359 this may be a hard question to answer, do you see the future ultimately 345 00:30:10.680 --> 00:30:14.279 leading the path that you guys are going with where things are more decentralized. 346 00:30:14.319 --> 00:30:18.400 Once again, given that there's the macro of blockchain and crypto taking us to 347 00:30:19.319 --> 00:30:26.119 that in that direction, but you still have your incumbent centralized companies like your 348 00:30:26.119 --> 00:30:32.119 Googles and so forth. How do you see that dynamic playing out? I 349 00:30:32.119 --> 00:30:34.079 don't think we have to take that bet. Is it going to be one 350 00:30:34.160 --> 00:30:38.079 or the other? It could be a combination of both. It could be 351 00:30:38.079 --> 00:30:42.799 some people just prefer using one then as compared to the others. Now, 352 00:30:44.519 --> 00:30:48.799 what we're trying to do here is there is an inherent problem in decentralization. 353 00:30:48.960 --> 00:30:52.319 The inherent problem is it's a lot tougher to do. It's not easy to 354 00:30:52.359 --> 00:30:59.680 deliver decision making putting things together. You have multiple different decentralized systems running around 355 00:31:00.160 --> 00:31:03.160 which you have to bring together. What about the governance, you know, 356 00:31:03.440 --> 00:31:10.839 So we're trying to make that much better by putting all these projects together, 357 00:31:10.880 --> 00:31:15.240 because then one governance, four or five technologies within that stack much easier. 358 00:31:15.319 --> 00:31:19.359 Right, you don't have to go to ten different projects use ten different tokens. 359 00:31:19.400 --> 00:31:22.759 I mean you don't do that in centralized entities, right, you just 360 00:31:22.920 --> 00:31:29.480 use your fead currency. So we're kind of removing that friction, trying to 361 00:31:29.599 --> 00:31:37.000 be in a much more competitive way available as a secondary solution, not just 362 00:31:37.079 --> 00:31:42.519 the secondary solution, but an alternative solution. So that in itself has huge 363 00:31:42.599 --> 00:31:48.839 value for not just for AI space, but also for the crypto space, 364 00:31:48.880 --> 00:31:55.279 because one thing, one of the biggest problems in crypto space is we're finding 365 00:31:56.279 --> 00:32:00.519 difficult to get applications on board it. So it's very much like, well, 366 00:32:00.559 --> 00:32:06.119 we have this wonderful solution, but we don't have a problem. So 367 00:32:06.119 --> 00:32:09.599 so there is that element which happens. So DeFi financial, you know, 368 00:32:09.960 --> 00:32:15.599 that's all there, but that's it. We haven't moved past that gaming or 369 00:32:15.640 --> 00:32:21.079 defile, right, so we're struggling to add new use cases to this. 370 00:32:21.559 --> 00:32:27.440 Now AI is very different. It opens up that space to application building, 371 00:32:27.880 --> 00:32:34.400 It opens up the space to bring users in a different way onto crypto. 372 00:32:34.519 --> 00:32:39.359 Now we don't have to keep telling everybody that we use crypto, because crypto 373 00:32:39.599 --> 00:32:45.680 the whole point is yes, it's decentralized, but it exists in the in 374 00:32:45.720 --> 00:32:49.440 the background, and it does what you need to do. And until we 375 00:32:49.480 --> 00:32:53.799 get to that point when you actually see a real application running on crypto, 376 00:32:54.519 --> 00:33:00.640 then you will start seeing a huge amount of traction. And I'm hoping AI 377 00:33:00.839 --> 00:33:05.759 is going to be that catalyst to bring that traction, which you know, 378 00:33:06.000 --> 00:33:08.359 initially we had a lot of traction because of DeFi, but then kind of 379 00:33:08.400 --> 00:33:14.160 slowed down. DeFi has other problems. The regulation, now that's another issue. 380 00:33:14.240 --> 00:33:19.640 But let's now compare that to AI. AI will also have regulation. 381 00:33:19.799 --> 00:33:27.640 But what's the great bit is being decentralized, being everything you know, auditable 382 00:33:27.799 --> 00:33:34.680 on chain, being being a solution which you know the regulator can come and 383 00:33:34.759 --> 00:33:37.319 monitor all the time. It gives us an edge, It doesn't give us 384 00:33:37.359 --> 00:33:43.279 a disadvantage. And I think that we want to make sure all the tools 385 00:33:43.319 --> 00:33:47.640 exist that the regulator can actually see. There is inclusion, there is safety, 386 00:33:47.680 --> 00:33:55.839 there's security, and we're not building this artificial superintelligence in isolation. It's 387 00:33:55.880 --> 00:33:59.839 not in a black box. You know, you can see what's happening, 388 00:34:00.200 --> 00:34:05.240 training, whose data is coming in, when the action is taken on agent 389 00:34:05.359 --> 00:34:09.360 based system, which agent took, what action it's recorded on the chain, 390 00:34:09.880 --> 00:34:15.239 what financial implication? You know. The biggest danger in AI at the moment 391 00:34:15.440 --> 00:34:21.079 is if you put it in you know, big corporations control and they get 392 00:34:21.079 --> 00:34:22.760 it wrong. They might not want to get it wrong, but they can 393 00:34:22.800 --> 00:34:25.679 get it wrong. I mean, look at what happened with Google's you know, 394 00:34:28.440 --> 00:34:31.000 the video and picture software, right. You know they didn't intend to 395 00:34:31.000 --> 00:34:35.679 do it. It just can happen and if a big corporation like Google can 396 00:34:36.360 --> 00:34:39.440 have that problem, then why would others not have that problem? They would 397 00:34:39.480 --> 00:34:44.559 certainly have that problem. And once it's out of the bag, once, 398 00:34:44.639 --> 00:34:49.519 I mean he now imagine making such a mistake with the AGI, which is 399 00:34:49.559 --> 00:34:54.599 self controlling, self learning, self assessing, and it kind of builds the 400 00:34:54.679 --> 00:35:00.760 wrong model in itself. Then how do we deal with it? Because it's 401 00:35:00.760 --> 00:35:04.920 in a centralized control, nobody can see it. So I think regulation and 402 00:35:05.119 --> 00:35:10.360 AI kind of are going to be together, but I think we put us 403 00:35:10.360 --> 00:35:20.000 in a very good place to actually showcase how There is an alternative question for 404 00:35:20.039 --> 00:35:23.880 you on the relationship between AI and blockchain. Do you see it being like 405 00:35:24.000 --> 00:35:28.239 symbiotic where to help each other. I think you touched a bit on it. 406 00:35:28.280 --> 00:35:31.840 But one item I've been paying attention to is the rise of AI defate 407 00:35:32.000 --> 00:35:38.159 content and certain platforms have been looking to put content on the blockchain so you 408 00:35:38.199 --> 00:35:43.119 can verify where did this content come from, when, when was it published, 409 00:35:43.119 --> 00:35:45.320 and so on and so forth. And there could be many other use 410 00:35:45.360 --> 00:35:51.280 cases where blockchain maybe is helping to police AI and AI, like you said, 411 00:35:51.559 --> 00:35:54.000 helping to open doors for adoption and use cases. How do you see 412 00:35:54.000 --> 00:36:00.360 that relationship, you know, continuing growing in the future. I think there's 413 00:36:00.440 --> 00:36:07.920 definitely a use case here for blockchain to manage that visibility, that auditability, 414 00:36:08.559 --> 00:36:13.519 And as you said, you know you could put that on chain. You 415 00:36:13.559 --> 00:36:16.760 can, you can do many things. But here's the problem. The problem 416 00:36:16.920 --> 00:36:22.679 is it's too cumbersome. It's not efficient. Unless we make it efficient, 417 00:36:22.800 --> 00:36:27.559 Unless we make it less cumbersome, unless it's easy for people to use it, 418 00:36:27.559 --> 00:36:31.199 it's going to the adoption will not come. So really, one of 419 00:36:31.280 --> 00:36:36.760 the key objectives that we have here is, yes, we want to build 420 00:36:36.960 --> 00:36:43.960 artificial superintelligence, but that's down the line and in building, in building that, 421 00:36:44.199 --> 00:36:47.679 in training that, we need visibility, we need auditability, we need 422 00:36:47.719 --> 00:36:52.039 ease of use, and we're putting all those components together. So while we 423 00:36:52.119 --> 00:36:57.639 are trying to go from here to there, we're also looking at commercialization, 424 00:36:57.840 --> 00:37:05.719 which is open, decentralized, which can actually provide visibility and all the things 425 00:37:05.719 --> 00:37:10.679 that you just mentioned. You could easily take care of that using the blockchain. 426 00:37:10.719 --> 00:37:14.280 But that's not to say you can't do it in a centralized way. 427 00:37:14.320 --> 00:37:19.320 But then who controls it is the question, and I think that's where we 428 00:37:19.360 --> 00:37:24.840 start to feel. But one of My biggest objective is unless we are at 429 00:37:24.920 --> 00:37:30.800 least as efficient as a centralized entities technology stack, we're not going to get 430 00:37:30.840 --> 00:37:35.400 the adoption because it makes it more expensive, it makes it more cumbersome, 431 00:37:35.880 --> 00:37:39.159 and it's difficult. But one thing which is really exciting is that if you 432 00:37:39.239 --> 00:37:45.760 have seen the developer community in crypto, you would know they are very passionate 433 00:37:45.760 --> 00:37:50.639 about it right, and they would do you know, they will spend nights 434 00:37:50.679 --> 00:37:54.039 and days to build something. And I think that's where we're going to have 435 00:37:54.159 --> 00:38:02.360 an edge over the centralized communities because now creating these way we're bringing these developers 436 00:38:02.480 --> 00:38:08.880 who who are so engaged and who are so passionate about building this new thing 437 00:38:09.639 --> 00:38:15.039 that we just on that basis, if we can get it right and we 438 00:38:15.079 --> 00:38:20.440 can excite our communities in the right way, we will we will have a 439 00:38:20.559 --> 00:38:23.880 very big advantage. M hmm. Yeah, absolutely. I do want to 440 00:38:23.880 --> 00:38:30.280 ask you about AI impact on the macro and this is going to deviate a 441 00:38:30.280 --> 00:38:35.880 bit off of the blockchain crypto conversation, but because of your experience, I 442 00:38:36.079 --> 00:38:42.119 absolutely have to ask you. There's a ton of benefits to AI, right, 443 00:38:42.280 --> 00:38:45.679 and I'm excited to see all these things, But are there anything, 444 00:38:45.559 --> 00:38:51.239 any items that keep you up at night where there may be downsides that impacts 445 00:38:51.239 --> 00:38:55.920 our society or culture or civilization. Is there anything that you think might be 446 00:38:58.400 --> 00:39:01.360 a threat. Whatever new technology, you have a threat to the old one, 447 00:39:01.800 --> 00:39:07.519 right. So you know, when when we had telephones, we had 448 00:39:07.519 --> 00:39:14.719 a threat to telegram we had and then we had a threat to mail system 449 00:39:14.760 --> 00:39:17.599 a little bit. And then when when we became uh, you know, 450 00:39:17.679 --> 00:39:22.400 mobile and five G, there was a threat to fiber optic network and there 451 00:39:22.480 --> 00:39:27.960 was a threat to all the other network and the telephone landline network. So 452 00:39:27.960 --> 00:39:32.280 so you're always going to have that problem, right. So Google came to 453 00:39:32.400 --> 00:39:37.719 disrupt Yahoo, Yahoo came to disrupt Yellow Pages, and you know it's going 454 00:39:37.760 --> 00:39:42.679 to happen. But that doesn't mean it's a negative impact always. It's you 455 00:39:42.679 --> 00:39:45.719 know, you could look at it in two different ways, whereby you can 456 00:39:45.760 --> 00:39:50.480 think, you know that we're going to improve in certain things, and we 457 00:39:50.519 --> 00:39:54.039 can you know, like, actually, I think this this whole new wave 458 00:39:54.280 --> 00:40:00.639 of AI and the ability to code will kind of reduce the U use cases 459 00:40:00.679 --> 00:40:05.360 for developers. Right, it's coming, right, like it or not, 460 00:40:05.519 --> 00:40:09.199 It's going to happen, accept it. I'm not saying it's happening right now 461 00:40:09.239 --> 00:40:13.119 because it can't write the code, can't test the code. You still need 462 00:40:13.119 --> 00:40:20.400 the logics in the right way, but the actual developers, the efficiency is 463 00:40:20.400 --> 00:40:23.119 going to multiply greatly, right, so you can actually the same one person 464 00:40:23.559 --> 00:40:29.360 can probably develop ten people's work right using these tools. So that's the step 465 00:40:29.400 --> 00:40:31.920 one. The second step is you probably don't need to write that much code 466 00:40:31.960 --> 00:40:37.280 anyway because the llms will take care of And that's what I was kind of 467 00:40:37.320 --> 00:40:40.800 alluding to. If you think about you, rather than writing code to understand 468 00:40:40.840 --> 00:40:45.760 your objective, you can just give your objective and the objective goes on to 469 00:40:45.920 --> 00:40:50.199 our agent based framework and actually assembles all the agents that you need, so 470 00:40:50.639 --> 00:40:53.079 you don't really need to write code, right, And now you have the 471 00:40:53.199 --> 00:40:58.719 user interface which understands you completely, which is you know, natural language user 472 00:40:58.760 --> 00:41:01.159 interface, so you can you can speak to your washing machine and tell it, 473 00:41:01.760 --> 00:41:07.400 you know term turn yourself on at this time, save me ten dollars 474 00:41:07.400 --> 00:41:09.599 worth of energy. It will just do what you wanted to do. That's 475 00:41:09.679 --> 00:41:13.760 your user interface. You don't have to have another app for it, right, 476 00:41:14.159 --> 00:41:17.280 So that model is going to for sure is going to change, and 477 00:41:17.320 --> 00:41:22.000 that will disrupt quite a few industries, and that will disrupt and that will 478 00:41:22.039 --> 00:41:28.440 have a negative impact on some of the people who work in those industries. 479 00:41:28.840 --> 00:41:34.239 But that will then open up some new areas of interest. And you know, 480 00:41:34.320 --> 00:41:37.280 for example, in my mind, biotech is a really great example. 481 00:41:37.719 --> 00:41:42.079 We're not going to solve that problem just with what we have right now. 482 00:41:42.119 --> 00:41:45.760 There's a lot more development, there's a lot more understanding needs to happen. 483 00:41:45.039 --> 00:41:50.679 So if anything, what would happen is that our focus will go back to 484 00:41:51.360 --> 00:41:58.679 more science research, which then is going to also be accelerated with this AI. 485 00:41:59.199 --> 00:42:04.360 But then comes this how can you bring multiple things together to develop a 486 00:42:04.400 --> 00:42:07.480 solution? How can you tell the AI to develop that solution? So I 487 00:42:07.519 --> 00:42:12.679 think one which which in these days is prompt engineering, but it will be 488 00:42:12.800 --> 00:42:16.599 much more than prompt engineering. It will be probably objective engineering. So how 489 00:42:16.639 --> 00:42:21.559 do you create the right objectives for this AI to kind of deliver the solution. 490 00:42:22.440 --> 00:42:28.760 I think we still will need quite a bit of manual kind of work 491 00:42:28.840 --> 00:42:35.760 will need to be done still in the foreseeable future, because robots are great, 492 00:42:36.079 --> 00:42:39.760 But still I think we have some time to get to a perfection level. 493 00:42:40.079 --> 00:42:45.079 So there will be still that But but I think some some areas for 494 00:42:45.159 --> 00:42:49.519 sure, will have a negative impact. Like entertainment, for example, you 495 00:42:49.519 --> 00:42:52.880 could just you could just write drama series and you can play them out and 496 00:42:52.920 --> 00:42:57.199 you don't need all the actors. And that's coming very soon, and we 497 00:42:57.239 --> 00:43:00.840 saw that in Hollywood, and I think a lot of a lot of investment 498 00:43:00.960 --> 00:43:06.599 has been delayed or stopped because of those reasons. So you are going to 499 00:43:06.639 --> 00:43:12.079 see some negative impact. But ultimately, we as humans will adapt and will 500 00:43:12.119 --> 00:43:16.960 come come back to you know, different and better ways of doing things for 501 00:43:17.039 --> 00:43:20.719 sure. You know, as you were saying that with biotech, I'm just 502 00:43:20.760 --> 00:43:23.880 thinking, like we as humans right now, with the technology we have, 503 00:43:23.920 --> 00:43:29.440 we're not able to cure cancer, but AI could possibly help us cure that 504 00:43:29.519 --> 00:43:31.360 and other diseases. I mean, you saw what Alpha fold is doing. 505 00:43:31.440 --> 00:43:37.800 You saw how accelerated development there will be. So yes, of course that 506 00:43:37.840 --> 00:43:43.360 will that will come. All right, let's jump back to crypto. Now, 507 00:43:43.360 --> 00:43:45.920 you're based in the UK, and I want to get your thoughts on 508 00:43:45.280 --> 00:43:50.320 you know, the UK working on cryptoregulations, EU pass crypto regulations. Here 509 00:43:50.320 --> 00:43:52.679 in the United States we're still waiting, but I would love to get your 510 00:43:52.719 --> 00:43:57.960 perspective on those regulations ever passed and how they impact the work you're doing as 511 00:43:57.960 --> 00:44:02.920 an innovator. I I still feel we don't have huge amounts of clarity here 512 00:44:04.000 --> 00:44:09.440 still, and we have a we have a double whamy effect because we are 513 00:44:09.480 --> 00:44:14.159 in Ai as well, which is now getting to be regulated, and there 514 00:44:14.199 --> 00:44:19.000 is crypto which needs to be regulated. So so I think there is there's 515 00:44:19.039 --> 00:44:25.519 a lot of misconceptions still and I don't find UK as a very conducive place 516 00:44:25.800 --> 00:44:30.920 for crypto at this point in time, and haven't done for quite a while. 517 00:44:31.119 --> 00:44:37.000 But I think there is improvement which is coming. There is there's a 518 00:44:37.079 --> 00:44:44.679 lot more understanding that is coming. But I feel some parts of you know, 519 00:44:44.760 --> 00:44:49.199 like for example, we moved our crypto operations to Dubai, which was 520 00:44:49.239 --> 00:44:54.559 probably the right thing to do because Dubai is much more welcoming to the crypto 521 00:44:54.679 --> 00:45:00.639 industry and has, you know, a regulation we set up in Singapore, 522 00:45:00.760 --> 00:45:06.760 which the Foundation was in Singapore initially because there was some clarity around how the 523 00:45:06.880 --> 00:45:15.079 token issuance should work. Although UK is making very good progress, the problem 524 00:45:15.320 --> 00:45:21.880 comes with banking effectively, because because scrypto has got such a bad kind of 525 00:45:21.920 --> 00:45:25.760 stigma attached to it, the banks you know, still that don't feel that 526 00:45:25.920 --> 00:45:32.039 they can comfortably allow crypto companies to operate with a bank account. So most 527 00:45:32.079 --> 00:45:37.280 of the high street banks, the mainstream banks, they don't want to touch 528 00:45:37.320 --> 00:45:40.280 crypto. It's a bad thing. It doesn't matter what you're doing. You 529 00:45:40.360 --> 00:45:45.199 might be running, you know, a company or legitimate with all the approvals, 530 00:45:45.880 --> 00:45:50.880 they just still feel very uncomfortable with it. So unless I think the 531 00:45:51.840 --> 00:46:00.360 regulators start kind of easing off their fears, I think we'll still see still 532 00:46:00.159 --> 00:46:07.199 won't feel the true value that could come here anytime soon. Yeah, I 533 00:46:07.199 --> 00:46:10.119 mean to your point, though they are they're making some progress, right, 534 00:46:10.159 --> 00:46:15.960 so at least in're moving in the right Yeah, but it's still got ways 535 00:46:15.960 --> 00:46:20.679 to go. And I envy a bit of the UK, even the EU, 536 00:46:20.760 --> 00:46:22.239 because you guys are at least a bit further ahead. And then we 537 00:46:22.280 --> 00:46:27.440 are here in the United States because here we're dealing with the SEC and all 538 00:46:27.480 --> 00:46:31.039 the litigation and so forth. But you know, we'll eventually get there. 539 00:46:32.400 --> 00:46:36.719 I want to get your thoughts on the market as a whole and how it's 540 00:46:36.840 --> 00:46:40.360 grown. Obviously, you got bigcoin ETFs out in the market. There's talks 541 00:46:40.360 --> 00:46:45.239 of tokenization, some of the biggest institutions in the world, from Wall Street 542 00:46:45.280 --> 00:46:47.440 and so forth, are getting involved. What are your thoughts on how the 543 00:46:47.480 --> 00:46:52.760 industry has grown. I think the industry is maturing, of course, and 544 00:46:52.800 --> 00:46:59.119 that's very clear because you know, initially it was not going to last for 545 00:46:59.199 --> 00:47:01.400 a year, and it was not going to last for you know, two 546 00:47:01.519 --> 00:47:06.920 years or five years or this. This this downcycle is the end of it. 547 00:47:07.079 --> 00:47:12.440 But you know it survived and it's right, so every time it comes 548 00:47:12.440 --> 00:47:15.920 back, it comes back stronger. So I think I think the way I 549 00:47:15.960 --> 00:47:21.639 look at it is bitcoin is now in a way, it kind of is 550 00:47:21.719 --> 00:47:25.159 now a commodity. I don't I don't know if it's a store value or 551 00:47:25.199 --> 00:47:30.079 not, but you know, I believe it is, but some people don't. 552 00:47:30.079 --> 00:47:34.679 But I don't want to argue that case. But it's definitely something which 553 00:47:34.760 --> 00:47:39.039 is becoming part of financial instruments. It's becoming part of financial portfolios. So 554 00:47:39.119 --> 00:47:45.079 it has kind of grown to a point where you can't ignore it. And 555 00:47:45.119 --> 00:47:52.920 that's that's probably my kind of only the limit which I can say, because 556 00:47:52.960 --> 00:47:58.639 I anything else would be just an opinion. So it's gone grown to become 557 00:47:58.760 --> 00:48:06.840 part of a financial kind of market and a value we have usefulness in terms 558 00:48:06.920 --> 00:48:10.719 of ethereum as we have seen and several other changes we have seen, so 559 00:48:10.840 --> 00:48:15.400 the usefulness is also coming, and I think people are starting to build solutions. 560 00:48:15.400 --> 00:48:20.639 But I did I did raise that point that unless we start to see 561 00:48:20.679 --> 00:48:25.599 some real traction in real use cases, we are going to struggle and you 562 00:48:25.639 --> 00:48:30.679 know, having another layer one where you do exactly the same as you did 563 00:48:30.760 --> 00:48:36.000 last year, or another layer one but slightly better, and then you can 564 00:48:36.079 --> 00:48:39.119 have meme points. I've got nothing against dogs, right, It's just that, 565 00:48:39.159 --> 00:48:44.159 I mean, I love dogs, I love cats. Meme coins great, 566 00:48:45.039 --> 00:48:50.119 you know, but they're not for me because I feel I feel that's 567 00:48:50.199 --> 00:48:54.280 what is kind of partly giving it a bad kind of image, because they 568 00:48:54.480 --> 00:48:59.280 you know, there's this up and there's this down, and people talk about 569 00:48:59.280 --> 00:49:04.320 it and people and then everybody feels that every project in this space is just 570 00:49:04.519 --> 00:49:07.480 going to do the same and it doesn't, and then you have this negativity 571 00:49:07.599 --> 00:49:13.480 amongst community, which then gets picked up by the outsiders and then they write 572 00:49:13.519 --> 00:49:15.880 about it because they think, well, you know, it's a bunch of 573 00:49:15.920 --> 00:49:22.119 people who are just going out and buying dog tokens and waiting for it to 574 00:49:22.159 --> 00:49:27.840 pump and then dump. So I think that's something which I don't appreciate in 575 00:49:28.639 --> 00:49:32.599 our industry, which I think, but what that still does. It brings 576 00:49:32.639 --> 00:49:37.559 attention. It's negative or positive, I don't care, but I mean it's 577 00:49:37.880 --> 00:49:42.239 an opinion, and you know, a lot of people would probably disagree with 578 00:49:42.280 --> 00:49:47.760 my opinion. But I feel if we start delivering some serious use cases, 579 00:49:49.960 --> 00:49:55.960 we have a lot to gain in the next few years. And I feel 580 00:49:57.159 --> 00:50:04.400 that AI might be a really, really good target technology which could open up 581 00:50:04.400 --> 00:50:09.880 that door because you will see automation coming, which brings ease of use of 582 00:50:10.039 --> 00:50:22.159 crypto. We have real use case of building AI applications on decentralized Leger technology, 583 00:50:22.159 --> 00:50:25.400 which is going to bring attention, but also real solution. I mean 584 00:50:25.719 --> 00:50:30.400 what I was explaining was agent based solution. We have two sites of agent 585 00:50:30.440 --> 00:50:35.639 based solution. One runs decentralized, one is centralized. And we're feel in 586 00:50:35.800 --> 00:50:39.159 a lot more traction in the decentralized space because people don't want to share a 587 00:50:39.159 --> 00:50:44.719 lot of their data. They want to maintain their own control on the data. 588 00:50:44.960 --> 00:50:52.159 They don't want some other company to run parts of the machine learning students 589 00:50:52.239 --> 00:50:54.480 want to build machine learning models and they want to monetize them, which they 590 00:50:54.519 --> 00:51:00.320 can't do on a centralized system, although centralized system like hugging faces also a 591 00:51:00.440 --> 00:51:04.079 very good example, but that's where we're going to start seeing traction. So 592 00:51:04.599 --> 00:51:08.800 of course I'm biased to my own space, which is AI, and I 593 00:51:08.800 --> 00:51:12.840 feel it can actually do quite a lot of good. And I think this 594 00:51:13.360 --> 00:51:20.920 whole alliance is just part of bringing that motion onto this space and saying you 595 00:51:20.920 --> 00:51:24.719 can build real world use cases. The developers who are building solutions in centralized 596 00:51:24.719 --> 00:51:30.119 space can build it here, and Crypto has a real use case. Well 597 00:51:30.159 --> 00:51:35.079 put, I got some wrap up questions here for you. First, if 598 00:51:35.119 --> 00:51:37.719 you could create your own metaverse, what would the theme be? And I'm 599 00:51:37.760 --> 00:51:43.000 assuming it might be AI. Well, I think you say that, But 600 00:51:43.079 --> 00:51:50.000 what I want to kind of point out towards is our alliance partner, Singularity 601 00:51:50.079 --> 00:51:55.639 has a Sophia Verse which is around AI, and I don't know if you've 602 00:51:55.639 --> 00:51:59.400 seen this, but it's pretty good, and I think you're going to have 603 00:51:59.559 --> 00:52:02.920 Agent running in the Sofia verse, so it'll be Agent would be my theme 604 00:52:04.360 --> 00:52:12.519 AI Agents. Got some rapid fire questions for your versus favorite food, Korean, 605 00:52:13.559 --> 00:52:22.360 favorite musician or band Metallica, I love Metallica, A favorite movie Showshank 606 00:52:22.440 --> 00:52:30.480 Redemption. It's a good one. Favorite book super Intelligence Nick Bostrom. Mm 607 00:52:30.559 --> 00:52:34.280 hmmm. And what do you do for fun as a hobby or pastime. 608 00:52:36.000 --> 00:52:42.639 Well, I like watching cars race, so Formula one is one of my 609 00:52:42.719 --> 00:52:49.679 favorites. Formula one or any car sports awesome, a pleasure chatting with you. 610 00:52:49.800 --> 00:52:53.519 I gotta have you back on because I'm super fascinated with AI and the 611 00:52:53.639 --> 00:52:58.760 convergence with blockchain and so forth. So we'll have to set up another interview 612 00:52:58.760 --> 00:53:00.599 and go deeper in somebody these AI topics. But thank you so much for 613 00:53:00.679 --> 00:53:05.320 joining me. Appreciate it. Thank you, Tony. It was a pleasure.