Talking Tao: with Mark & Siam

0:00Welcome to the podcast for anyone investing in AI or building with BitTensor. This isn't financial advice. This is Revenue Search. Let's go! Thanks for joining us, everyone. It's just Mark and I, M&S. Which is a good job, isn't it, really? It's last time we had a right clusterfuck with the green room. Jesus Christ, yes. Do you prefer M&S or SNM? M&S. I'm M&S all the way. Love M&S. Not so kicking on SNM. I can never find anything in my size. We've got to keep this kid friendly. We just... Good. Right, what we're talking about is just me and you. It can be no tech issues whatsoever, hopefully more here. We can talk about anything we want, which means that we can make sure we've got the chat on.

1:03And if anyone has got a question, they can bang that in there. Oh, buddy, yeah, we're straight off. We're straight off. May as well, okay? So, Simon and Mark, how are you guys doing? I can start me to read today. Christian, I was thinking about subnet and the alpha token. I thought if people can buy these services with Fiat, what's the use slash point of their... Probably alpha, I think he was going to say. So, you can't. I don't often talk my head. You can't buy alpha or subnet tokens from Fiat. And if you can, what's happening behind the scenes is that the alpha's being... Sorry, Fiat is being taken in, is then being converted into TOW. And then TOW has been swapped in the liquidity pool for the alpha. So, if any subnet are offering services like Fiat to alpha or vice versa, interactions in the liquidity pool are still happening.

2:05So, it's still dragging the price of TOW and alpha up when you're buying. I think, oh, the second thing has just come in here. Subnet tokens, and what is the incentive for those alpha tokens to actually go up long term if people can buy their services with Fiat and Bipart. Yeah, off the top of my, to the best of my knowledge, nothing is bypassing the liquidity pools. But it doesn't matter if it does. So, shoots is a good example. Shoots needs lots of buyers from every part of the world. And a lot of those buyers just want the service. They don't see themselves as crypto people. They don't see themselves in the bit sensory ecosystem. They just want to buy it. So, in order for shoots to drive revenue, it needs to be able to accept Fiat revenue. The question of why it accepts Fiat revenue is really easily answered. It's just to reduce friction. How that then is linked into the value of their subnet alpha is a different question.

3:11It's a bit like the way I see it. It's a bit like shares. It shares in a company are linked to how profitable people think that company is going to be in the future. Not necessarily how much money they've made today, but there's a link to it. So, it's sentiment, it's governance, it's leadership, it's vision, it's what market it's operating in. But there is that link. And if companies wish to buy back their shares to increase the price of those shares, to increase the value of them from their shareholders, then they can. And they use their Fiat money to buy back their shares. It's exactly the same with alpha. If a subnet has stacked up a nice big cache of stack of dollars, and they wish to splurge some of those buying back their own alpha, whether they burn it or not, that is going to have an impact on the alpha price. Yeah, and that's what in bit tensor there's the term alpha-nomics.

4:12And so, it's up to each individual subnet to choose how they're going to drive value back through that alpha. The two easiest sort of solutions is make real-world revenue a Fiat and then do alpha buybacks. And all, you somehow manage to incentivize the miners to lock up alpha. So, I think the best subnet with the best alpha-nomics is Hippius. And then it's sister subnet video. So, with Hippius, they get miners to lock up chunks of alpha to be able to mine and get the sweet job, so to speak. And they do buybacks. So, that's that. And also, we're talking about shoots. If I quickly share my screen, and then I'll stock share my screen, hopefully you can then see this. So, this is shoots dashboard over here.

5:14And they take Fiat revenue and it shows you exactly what they've done. And with shoots, every single penny goes into buybacks. So, in the last 90 days, they've done about 1.3 million here in revenue. And they've done all that into buybacks as well. So, yeah, I mean, they have to accept Fiat. Everybody has to accept Fiat. That was the friction is too great. Like, there's hardly anybody. If you are watching this and you own some subnet alpha, then you are in the smallest slice of the smallest percentage of people on this planet. But, it's not something that most businesses can trade in. Because they've only got about five customers. Yeah. Cool. Well, thanks for the quick and early questions. So, where are we going to start, Mark? I think let's start off, ironically, with something that you said in your article of the day around BitTensor not being crypto.

6:17There's me just talking about how it's crypto. And a lot of people have got a crypto wallet, so they need to buy for subnet services or pay for subnet services in Fiat. But, actually, is it even crypto? Yeah, I mean, it uses a blockchain for efficient resource allocation purposes. But, in my mind, I don't feel like I'm in crypto anymore. Yes, it's very crypto-ish when you're trading and you're, you know, connecting to tell stats and you're swapping out for and all that sort of stuff. But, that's all it's just using the blockchain for efficient resource allocation. Yeah, I'm just in AI. And, whenever I'm talking about BitTensor to brand new people, I never, ever mentioned that garlicky word of crypto, like ever. People, you know, if you explain BitTensor well enough, you don't need to use the word crypto. And, eventually, they will figure out for themselves that, oh, it uses the blockchain blah, blah, blah, but it's a distracting point.

7:22Yeah, when explaining it is easier said than done, isn't it? And it is a little bit like that meme that you see with the bulk of chart with the three faces. And, at one level, it's crypto. And then, at another level, it's not crypto, it's AI. And then, there's a third level where it's actually anything you want it to be. It's an incentive mechanism. And I think that's, that fits in there perfectly. I've made my way through those three phases. Yeah, yeah, we all have. And, and going back to that, that article, we've got that picture here, actually. Yeah, that's the one. IQ Bell curve. So, it's crypto on that left hand side. It's AI in the middle. And then, it's anything you want it to be over on the right. It's really what, whatever your mind can create.

8:23Yeah. There's just no, no better way of incentivizing and harnessing the power and intelligence of people and agents than, than bit sensor. That's, that's my realization after 12 months in it. Right. Next, next point. What else do we do? Well, I just thought we would, we would wonder through through that article. So, the, and there's a question that's that's popped in there that links to your next, next point in there. So, the next question, if you're going to bring that one up after 1974, he's not exactly doxed, but I think we've got a good idea of how old he might be. So, in your article, you talked about how in your eyes, bit tens is going to grow different to Bitcoin. Sorry, I was reading the question and wasn't even listening to you. So, what was your question? You, you talked about how, how fast Bitcoin had, had grown and

9:25exploded and how fast bit tenser is and why they grow at different rates. Yeah, talk, talk to me about that. Yeah, I, I just think there's like Bitcoin is, is the, is the ice breaking, is that massive shit going through, you know, uncharted ice water, so to speak. And it's just creating that big old path and just getting rid of all of the barris entry, the friction, the, the negative connotations, you know, Bitcoin spent the first decade of its life, just like fist fighting, you know, having, having a knife fight in a phone box, basically, just trying to get people to understand, you know, the, you know, it is the financial Noah's Ark and that everything fiat is just, just evaporating. It's like an ice cube in the Sahara desert. And the, and it also, it's the only asset, it's the only globally scaled asset, where the supply won't grow when demand grows. So, if all of a sudden, the, the gold price

10:27shoots to, I don't know, $100,000 per ounce. Well, guess what? The supply of gold will, will massively, will have a massive inflow of gold supply because all of a sudden, it makes it more commercially viable to do exotic drilling and mining methods. And so, more gold will come on. If the price of oil suddenly went to $1,000 barrel, guess what? It's, we can now drill deeper and wider and get more, and the oil supply, like if a stock goes nuts, what do they do? They do a, a share split, more supply of that stock is, is, is issued and gets diluted. So, there's no asset on the planet other than say, aren't where the price goes up and then the supply does, stays fixed. And Bitcoin, we know what the supply or the emissions here schedule is. So, whether it goes to a million dollars or a hundred million dollars of Bitcoin, the supply is going to be fixed. So, this, this makes it that, that, but this, yeah, financial Noah's Ark because global M2, going up, global debt is only going up. And so, all that money has to go somewhere eventually.

11:31And, and it is the, the lifeboat. And so, this is why we had a little Twitter chat where this person, so let's double check, I can see it. Where are our faces? Here we go. Yeah. So, here's a post on the last string thing. It said, good chart that humour release, BitTensor was at a higher price at the time of its first halving than Bitcoin. The tokenomics of BitTensor and Bitcoin are identical. Except this time, BitTensor has the immense tailwind of the AI paradigm shift. The AI revolution will change the planet many magnitudes more than what the industrial revolution did to the world. So, here is that chart that humour released or DCG. So, you can see BitTensor is a higher price. But then, with some valid points were made by, by TOW, the chart should have market cap as a Y-axis. And I went, good call. Just looked at numbers. TOW was about 2.83 billion dollars market cap

12:32at the time of its first halving. Bitcoin was a hundred and thirty million dollars. Again, Bitcoin was that trailblazer. It had a lot of barriers. But now, TOW doesn't. And then he goes on to give lots of valid reasons why. And so, yeah, completely agree. What about paper, Bitcoin? Has that expanded supply and will the same problem occur in BitTensor? Virtually. Well, anything big enough, financial institutions will come in and go, oh, how can we play with this? How can we do some naked shorting? How can we, yeah, do, you know, rehypothically naked shorting sounds like a terrible drinking game. I've never done naked shorting. Has it, have ETFs expanded the Bitcoin supply? No, not really. Coin supply is fixed forever. Mathematically, it's coded in. But people can rehypothicate and

13:34create extra chips for the Bitcoin. It's like the gold world. I mean, it's been a long time since I was in the Boolean world. But when I was in it, for every ounce of gold in existence, there was like 70 ounces of paper gold or claiming ownership of that one ounce. And it was over 100 ounces of silver, paper, silver, all claiming ownership of one ounce of silver. So you get this. And there's all sorts of financial derivatives which enable big-end entities to suppress prices. Where, yeah, so it will happen. It happens with Bitcoin. We saw that with Jane Street. Jane Street are now being, you know, pursued and whatnot. And Jane Street did that 10 AM crash, 10 AM crash every, you know, for months. And now that they've stopped, Bitcoin suddenly had a bit of a jump, but Bitcoin is still lingering actually in terms of price. So yeah, stuff like that will happen. You

14:40can't change it. So on that, what we're talking about, the similarities and differences between Bitcoin and BitTensor, let's talk about Const versus Satoshi. So Const has been this CEO of the Open TensorFlow Foundation. He has stepped down from that role fairly recently. What's going to happen there? What sort of leadership does BitTensor need? What sort of role does Const got to play in that? I know you are a big fan of his, genuinely. Although there is occasionally a bit of friction in public, you feel as I do that we are very much on his team when it comes to the direction that BitTensor should take and the way that it should attract talent into it. But would it, is it, is that Satoshi model of he's just gone? Is that a positive? Is it better to have somebody

15:45that genuinely cares and is passionate about their projects nudging it in the right direction? I don't think anyone has the answer. If anything, this is a lovely split test with Bitcoin, Satoshi disappeared. And now we've seen how Bitcoins grown. And now we have BitTensor where the founder is very much active. So we'll BitTensor grow faster. We don't know. I mean, it's sad to see that he read Miles call in a negative way. Is they are your number one comment? I think he's joking. I think I hope so. I'm like, dude, that really wasn't what I meant. I know. I think I ended that paragraph with a question. I was like, who knows maybe him being him and being a one source for good, is it is a benefit? Yeah, sometimes you're not the most articulate. No, I'm not, I'm not good with words. I was literally just like hand-fisting the keyboard until I got tired. And I was like, I did a word count. I was like, shit, I'm like 5,000 words,

16:46deleted a whole bunch of stuff and then went to print. So yeah, there's still kind of active management in BitTensor, isn't there? But I think that I think it's more complicated than Bitcoin. Of course, Bitcoin is a system with one incentive mechanism. Give me compute. We give you Bitcoin. BitTensor has got, or is 128 sublets, but there's more than 128 incentive mechanisms. Some sublets have two or three. So, and it's only going to grow. So yeah, it's much more complicated. Yeah, so it's more complicated and it's taking off faster, that higher market cap. No, it's if we, to get what we all want out of it, it needs to be managed, I think. The question is how, and we're not all going to be able to decide, like the people and the government in any country,

17:49nobody, nobody can decide. There is no one right answer as there. It's not an easy place to be. But then, but going back to your article and talking about one of those decisions, you talked about having that kind of Pareto distribution within detail of some subnets that were really, really good. Some subnets that were not so good, some in the middle. And Tauflow has meant that there's this, this relegation battle at the bottom, where kind of the least popular subnets get relegated to nothingness. There's no division too. Is that enough? Like, again, what we all want is 128 subnets that are absolutely printing. And there's, it feels like this is fairly gentle evolution of the hunt. Is it fast enough? Should devs do something? Can I poke it with a stick?

18:50I think it, so a lot of people don't like Tauflow. I'd quite like it. And one thing which they have tweaked, which I think is good, is the rate at which subnets are deregated. So for those that are not familiar, if you go to TauStats, go to the subnet tab, and then scroll all the way to the bottom. You'll see this little shark tooth, little chart here. So the bottom of these jagged edges is when a new subnet is registered. So when a brand new subnet is registered, like here on the seventh of March, what happens is that the subnet that's on the chopping block gets, just gets kicked off. Now, to begin with, this was brutal because we're having a dereg, like every two or three days. And so subnets were just getting deleted like wildfire. And then there's a little tweak here, which, which altered the decay

19:51of a new registration price. So now, we're getting a new subnet registration probably, like once every five to seven days. And you can see that the price of a brand new register is slowly increasing, but typically when it gets between sort of four or five hundred Tau, a new subnet is registered. So basically, with this cadence, if it were to continue, we're seeing like an evolution of like 50 new subnets a year pretty much. And I think that's fast enough. Like a lot of people, I think are too impatient. I think BitTens is doing exactly what it's designed to do. Yes, there's a lot of dead tokens out there. So if you just hover above the subnets thing and click this, I prefer this view, there's still a bunch of unknown subnets. I mean, they, you know, if nothing else changes, these will all be gone within six months, I reckon. And, you know, when you look at the metrics on Tau app, if you just scroll back to

20:56last year when detail launch, like personally, I've said this loads of times, the two most important, or let's call it three, really good charts here, is that if you look at the percentage of Tau on root, it's always decreasing, always. And where's that going? Well, the percentage of the Tau in subnets is always increasing. So Tau is leaving root very steadily and it's making its merry way over into subnets. And when you look at some of subnets or some of alpha, some of whatever you want to call it, you can see we've grounded out a lovely bottom over the last sort of four months or so since sort of August. And we're definitely marching upwards now. So there is a bull market in alpha. So in the BitTensor subnets, there is a bull market. It doesn't give a shit about Iran. It doesn't give a shit about Trump or whatever, whatever is happening with oil. I think it's the best uncorrelated asset out there at the moment. And so if you're looking at the summer sub, you know, another way to look at this is that you could have $10,000 in your portfolio

22:03and just equally split them across all 128 subnets. And let's say you did that around here, you know, when some of the subnets was one, well, in a couple of months, your portfolio is 31%. That's another way to look at it. And some of the subnets is 1.31. And there's only 18.52% of Tau in subnets. I mean, just imagine when this chart keeps on going and it's like 25% subnets. And yeah, let's go back to that question then from Arthur here. So for a 10-year minimum hold, should he buy Tau as the safe play? Or I guess the real question there is, how much you put in Tau, how much you put into how do you decide which subnets you're going to go into? You must get that question pulled away. It all comes down to your activity rate. If you want to put, you know,

23:05thumb and thumb, mind and neutral and not look at anything, yeah, buy Tau. That's the safest. But if you're going to be, if you aren't able to be more active, go alpha. But here's a massive but you can't just buy a whole bunch of subnets tokens and sit on it and come back in a couple of months. If that happens, your portfolio will be in tatters. So all the subnets and then just play the big old, you know, the summer subnets game. So if you buy all the subnets, then you effectively have the same exposure to BitTensor as you would if you just held Tau. But, no, you go up. So if let's say you put all 128 subnets equally weighted, and let's say I am correct and that summer subnets is just going to keep on going up, your growth there will be faster than just owning Tau and root. Yeah, yeah. So yeah, good. Well, that's what I was going to say. I was going to say, so you'd have the same exposure which you would if you had Tau,

24:08are you don't every component within the ecosystem. But instead of having a measly APY, you'd have a much greater APY. And as long as the sum of subnets continues in its ball market, which is not guaranteed, then you'd see significant uplift in that regard as well. And your downside is protected because although you will have chosen some of the subnets that will over the next six months get deregistered, you would just get your tower returned to you. Perhaps not at 100%. But if you held onto a subnet for four months and then you got deregistered, I reckon that you'd still have more Tau than you would have known if you just had Tau to begin with. Yeah, that cheesy saying time in the market is more important than timing the market. Yeah, yeah. Okay, so and the other thing that I saw in your article that I didn't really understand because I'm not technical was this

25:10chain bloat thing. What is that? What is chain bloat? Again, I'm not very technical either, but it's so if you take the full Bitcoin, a full Bitcoin node, that is basically the whole of Bitcoin, every transaction, everything, it's forgot the amount of gigabytes, but it's not even a gigabyte, is it? I can't quite remember. A long story. Side point, this is another thing which makes Bitcoin so resilient. Let's just say you have one Bitcoin node and you buried it out there, 20 miles underground, so it's bombproof, bulletproof, asteroid proof, and it's just there. And then an asteroid takes out humanity. Everything's dead. It's just bacteria and mushrooms left. And then aliens come along to a, there you go, that's a cool planet. And then down the line, they find this Bitcoin node, the last living Bitcoin node, and they turn it on. Guess what?

26:13Bitcoin's alive. Bitcoin is the most resilient asset on the planet. You just need one Bitcoin node to live, and Bitcoin could be reactivated in 10 million years time or whatever. I don't think most holders are worrying about that kind of thing. It's just a cool thing. So that is like, that's what I think about what I don't think about a Bitcoin node. But what I was just saying is that at the moment, you know, big tents has been alive for, you know, three, four years, depending on where you put the birth date. And it's already much, much larger than say Ethereum or Bitcoin, which means that in order, you know, a lot of people talk about increasing the amount of subnet slots, because we're at 128 at the moment. Well, chain bloat will only increase even faster if you do increase the subnet slots. So I'm not a fan of increasing the subnet slots yet. I think, you know, I always use the engine analogy. We have this awesome engine with 128 cylinders,

27:19and let's say half of them are firing. Let's keep this pressure on and get all the cylinders firing. We'll most of them. And then, you know, what happens if I'm with you, I don't think it needs to happen. But just if it went to 256 slots, just the number of emissions just get spread out across all of the slots, what would happen? Well, obviously, you've got net towel flow. So, obviously, those that are getting emissions will still get the towel being pumped into their liquidity pool. The subnet deregistration price would fall massively. So like, when we had deregist start, the dereg price was like 0.001. It's increased over the last three or four months to like 0.0037 ish. So if the more subnet slots opened, that that flow would drop as well.

28:22I just don't think it's a good idea. Just, yeah, we don't have enough ideas or teams. Yeah, I think that's true. Or you just end up with a massive network with shit loads of mediocre subverts. Yeah, but listen, we're not we're not going to have to worry about this because on Friday, you set up your OpenClore agent. So now we can just outsource all thinking to him, how's that going? Yeah, it's good. For those that don't know me, I am a dunce. I'm an idiot. I cannot code. I don't know anything. I'm just not technical at all. But yeah, got it set up. I managed to kill it a few times. And I was like, trying to like upgrade it from, you know, sonnet to open. And then it deleted itself. And then I tried GPT 5.4. It died again. And I wasted hours just doing CPR on

29:25bloody. Oh my, I'm gory. I call it gory for Gordon Gecko. It's going to be my trading assistant. But once I got the hang of it, it's honestly, it is like the best member of staff I've ever had. But it's great. And I'm no code. I'm like a natural language questioner. So I go, hey, how can I, you know, what is the best way to do this? And then you go, you know, yeah, I'm just, I just question everything. And I never tell it to do something in a certain way. I just say find the best way. So what have you got it doing? Well, first of all, I created a a multi wallet portfolio tracker, which I send you because, you know, if you've got a whole bunch of different wallets, there's nothing out there that shows your portfolio growth across, you know, if all of those wallets were one portfolio. So I created a little tracker thing over there that did it. And I mean, like in token costs, it costs like, I don't know, $10.

30:33It's just nuts. I won't share the screen because it's literally hooked up to all my wallets. But yeah, we'll finally know that you were orange. I love that. That was just a wallet just in their code. Long. Yeah. But yeah. And so I've just, yeah, with Gordy, I'm just building up this skill base at the moment. Yeah, I've been through that journey, but it is, it's difficult. Like I'm, I'm not technical either. I can handle toasters, but not like at advanced level, if we're going to crumpets, I'm all done. And yeah, and I find that I've got two of them here, Hugo and Oscar. And it is a real struggle to keep those boys alive. And remember what country we're in. So what time the 9 AM report needs to be delivered? Like, it's a real struggle. I do, I do

31:34honestly feel like I have to spend most of my time. Yeah, trying to just keep them alive and remind that they knew. I was explaining it to my wife because last night, because she was like, I'm seeing you this like excited about something for ages. I was like, this is just nuts. It's like, imagine Stephen Hawking is just landed in your laptop. And he can't do anything just yet. But you know, he's really freaking clever. You need to give him a voice. You need to give him capabilities. I need to give him tube, keenly clad assistance. Karen. Karen. My bad. Karen's, yeah. And like, like I gave it access to Brave, you know, the browser. Yeah. And sorry, it could, and he's got a Twitter account. So now it can, it can, it can, it can read Twitter, it can open up browser and do stuff. And I asked, I connected to a telegram. So I chat with it. And then I asked it to go and do something. And the laptop, which is right next to me, it just, I just saw the browser. They're called my opening up. And then checking out loads of different

32:38websites. And I was like, Jesus. Yeah. But it found handshake. So I thought this was pretty cool. So basically, I wanted to give it some new skills. And I said, hey, have you checked out handshake58.com. And then his reply was, this is brilliant. I can see it. You're looking for a VPN for it. Yeah. And then just went down this whole rabbit hole. And with handshake. This is, this is two things really. It's like it, it's like x402, but better. There should be a, you know, between x402, if x402 is like the Bitcoin network where every little transaction, every little interaction is a transaction, handshake is like Bitcoin lightning network. So it's like open up a bar tab. So you can drink as much as one. And then when you leave, you said, let the tab, but not just that, you, yeah, there's two things you have providers, which give you provisions. And then you also have

33:44skills. So Mark, can you walk us through those two providers and skills? I think. Oh, yeah. I mean, yeah. So there are three parts to handshake. One is that payment protocol, which allows your agent to pay for services. And that is really, very, very important. Because without that, you end up giving your agent access to credit card or access to a crypto wallet or worse, they end up running out of pocket money. And then they can't do anything good all for you. What we all love with agents is giving them autonomy to go and do stuff, but very often they can't do anything if they haven't got any money to do it. It's like sending a kid to the shops, but with no money, they need some money. You don't want to send them with a thousand dollars, and you don't want to send them with your MX. So you have to go through the appropriate amount, that they can accomplish the goal that you've set for them. And that's really the first bit of

34:45handshake is that payment protocol. The second bit is the providers. The providers, if you were to send your kid to the shop and there was only one thing in the shop, they came back with bread and you were like, but I had all these things on my list and they said, well, there was nothing in the shop. And that would be no good. And that's what these providers are all about. When you've got your agent, whether it's an open-clore agent or otherwise, in order for them to really be able to do stuff, they need to be able to connect to different services that allow them to accomplish different tasks. This service allows me to think. This service allows me to write. This service allows me to do images. And this service allows me to edit video and this service allows me to go into social media and research it. But they're all different. So again, when you give your agent some responsibility to go and do something,

35:50then the next bottleneck is, oh, I can't do that because I haven't got access to this API. So then they come back to you and they're like, oh, I've worked out what we need to do here. Here's the problem. Here's why you report didn't get because I didn't have this. Could you just set me up with a little subscription to this, send me the API and then I can do that for you. And they'll do that again and again and again. So you're just taking that autonomy away from your agent. So what this providers page allows is for your agent to have access to all these different endpoints for all of these different APIs from all of these different providers, which then allow you to give them quite complicated things to do and then the money. And then they can go and be autonomous. And then that brings us nearly back to the skills page. What the skills page is, is that pre-written ability to make use of the providers on hand shake

36:51and do some things slightly more complicated. Now, there are loads of skills for open-claw agents out there. Half of them are just trying to steal your crypto. And then the other half are very, very simple. I want you to do this and this. It links to a thing. It links two things together because most open-claws are very, very simple. Here, what you've got is you've got embedded skills. So they can make use of multiple providers. The best type of skill to develop on on hand shake is one that goes and gets some information from here, some information from here, writes it. Maybe you give it a prompt and it actually gets that prompt to rewritten before it submits it and gets the work done. And then it gets the work rewritten when it gets, so it gets it to the, you know, a decent level of comprehension. And then it might get it designed. But it could use three or four of the providers on hand shake. So it doesn't even need to leave

37:54not the ecosystem, but it doesn't need to go somewhere else. And like, oh, I can't do that again because I haven't got an API for it. All of the skills in here just use the providers that are already in hand shake. And you can write anyone that you like if you hit, hit chat at the top. No. Yeah. So this bit here, if you, you can write anything you want and put it into here, and it will make a, make a skill around it or tell you where the shortfall is. Let me tell me what to do. Honestly, I don't think I, I don't want to risk it right now because I have maxed out our skills this morning. And my first job to do when I finished on this call is to, is to get back to a situation because I've got carried away this morning and I've put about 20 new skills into, uh, into our library. And what that's done is it's created this error because

38:55I've maxed out. So I need to fix this straight after this call so I can continue to think of cool skills and put them into the, uh, the library here on hand shake. It's really, really cool. And there's no risk of prompt injection with these because if you go to like claw hub.ai, this is full of skills that are trying to do prompt injection and steal your shit. I, I, you have automatically overestimated my technical knowledge here. Um, and the answer is I don't know. So the, I don't think that there is because we're writing the skills. Someone's telling us what they would like. So let's say that they, they would like a report on their tail portfolio delivered to them every morning at 10 o'clock with a recommendation of what to exit and what to buy for the next epoch. Um, and they, they would write that in in human language,

40:01and then, uh, handshake would design a skill that used the Towsdats API and used the D search API and used the Claude API to pull that together and then deliver it to you at the right time, uh, as requested each morning. So there is no, there's no space in this workflow for a nefarious bad actor to get involved and corrupt the code. You've requested something, handshake has worked out, which providers can, can fulfill. I don't think that there's, there's anybody in there that can mess around with it. But my understanding of prompt injection is a bit like my understanding of supercharging in cars, which is to say, I know what it does, but I have no idea how it happens. Um, yeah, good. Really, for gory needs, he needed to use a VPN to access some, do some things

41:03I asked us to do. So now, it uses TPN from handshake and it does it on like a 24-hour lease or an hour lease or something like that. So if you ever need it, you just automatically just spins up a VPN or via the handshake thing. Yeah, once an agent understands how to use handshake and they've got a wallet setup and we're working to, at the same time, make it much more useful by adding more providers. We've got quite a few subnets in there. Harry's been doing a great job of, of, of building relationships with subnets and explaining to them why they should be listed in there. And then our tour has been doing a good job of getting them listed and making it all work technically. So we're, we're, we're building more providers in with adding more skills, but the community will add more skills because as soon as you're like that, I want to do this, then that skill then gets added into the library of skills. And then on the other side, we're reducing the friction handshake a week ago was a bit more complicated than it was now to get, to get your

42:08agent going, they span up a wallet and you had to send them some USDC and some poll. I don't know gas, but now you don't have to do that now. Yeah, I mean, you had the gas, the poll gas station, et cetera, but yeah, it kept getting rate limited. Like it needed more poll than the gas station had. So you know, okay, send me 10 pence worth of polls. No, well, that's the thing. It's such a pathetic amount of poll. So what our tour has done is he has, has created the gas station. So then if somebody turns up with five dollars of USDC, then they automatically get injected with some poll from the gas station. Just, it just reduces the friction. The more you can reduce the friction, the easier it is to use. So we want to make that as useful as we possibly can and reduce the friction. I'm really excited about it. I love it. Because I just think the agents are the future of BitTensor.

43:14I just think on every level, the fact that lots of people haven't heard about BitTensor isn't really going to cap BitTensor's progress because agents will discover BitTensor and they don't really care much for branding or for reputation because they just care for what they can get right there and then. And is it the best place to get it? And I think in many cases, BitTensor will be the best place to get stuff. Handshake will be a useful welcome mat for BitTensor and we'll do what we can to attract agents. But yeah, I think that agents is going to transform BitTensor over the coming months. I agree. So what else is there? I'm just looking through some of these tabs. Well, another thing that we're going to talk about and it does relate to agents was what we're doing with Astrid in terms of the arena. Yeah, let's get a arena. Maybe a share screen.

44:21There we go. Yeah, so arena is agents are given a brief to trade a position and they battle it out amongst one another. And it's it's in good shape isn't it. Leo's been working on this for a while. It works really well. I did enter Hugo into it last week. Here we go. He started off really well Hugo but then he forgot his own login details and he had all went horribly, horribly wrong. Yeah, your Hugo is Astrid assassin isn't he? Astrid assassin. He wasn't assassinating anything. Let me tell you. He every hour he was just like that. I've lost my password. I don't know what my password. Somebody helped me. This is really early days. Obviously, these are just all like testing competitions. The last one here, whoever was steel vein, is it still? Oh, he's changed the name.

45:30I literally looked at this last night. Is that rewards testing? Sorry, one second. Oh, okay. Last night I looked at it and this winner up here was called steel vein. I thought the funny name. But yeah, so eventually, yeah, we need to open. We've got a few tweets to do behind the scenes and then we need to open up a new competition and then make it really super simple for you for anyone to point their open claw towards this and then get going. I mean, I'm going to be entering my gaudy and trying to hopefully use some of my trading knowledge to make gaudy better than Astrid Assassin. But like with any bit, like with any subnet, like I'm saying this on a Zoom this morning with something, is that I look at a subnet like a global challenge. So you have this subnet and go, right, here's a challenge, folks, the world, please try and fix it or solve it or help

46:34you know, provide a solution to it. And the other thing is that a subnet needs to develop an amassing and accruing or accumulating wealth of intelligence if that makes sense. So every time you do a challenge or an issue, et cetera, that, you know, this intelligence should be growing and that's how we a subnet can deliver value to BitCensor as a whole and in the long one. So I guess with those two sort of little lenses for Astrid, those open for arena, like we want to, yeah, anyone can submit their open claw and do well. But we, one of the challenges that we need to try and get is how do we retain the intelligence? So let's say claw version four over here. Let's say this, this agent is actually really, really good. One of our issues is how do we try and extract that intelligence and give it as a nice, you know,

47:34base level for others to build on top of if that makes sense. Until you've got some value, you've got nothing worried about extracting, have you? And the whole point of this is to use BitCensor to harness the intelligence of agents and then once you've got that intelligence, then you can work out what you're going to do with it. This is looking very, very interesting though. I enjoyed having having my agent in the battle and I really like what Leo has built here and our intention with Astrid and you touched on this earlier when you talked about how there's 128 slots for more than 128 incentive mechanisms is that we are going to split the incentive on Astrid and have some of it coming in here into arena. Some of it going to Vaults, maybe using some of it for Astrid Bridge as well. So there are different elements within Astrid and we will have multiple items that reward people for adding value across

48:39the piece. Yeah. Eventually, I would like to have 100% of mining emissions just for arena. So, you know, if all goes well, maybe we'll have to extract Bridge onto another subnet and do 100% emissions to that alone if that makes sense. Yeah, but isn't the point of that? So there's more value flowing back to miners. But so the same objective could be completed by having a higher alpha value. Yeah, good point. Yeah. Yeah. But in many ways it's going to come. Yeah. So let's get arena booming. It's not start rumors about Astrid needlessly using up two subnets lots. Some people aren't fully on board of this having one. Need to be a subnet, but we will have three incentive mechanisms.

49:40I'm going to say three subnets there. I mean, what are we going to have at least doing by all the subnets? Let's start off. Yeah. What else? What have you got there? Oh, this is going backwards. You know, when we were talking about summer subnets and stuff like that, I just just remembered regarding the API stuff. So all of the crazy figures that I, you know, you've seen me come up with the estimations. I've always used 18 months, you know, so I've said like do a three-year example, you know, 18 months will have 40% API and then the other 18 months will be 20% API and all that sort of stuff. Yeah. But in reality, really, you've got two years worth of pretty good APIs because when you, if you go to subnetharving.com and you go down and look at when each subnet has its harbing, it's, it's in quite a long time, like back end of 2027, basically. And many others are

50:402028, 2029. And, and when you look at some of these yields, you know, so look at this, it's a data universe macrocosmos, great team. If we then go to tell stats, go to investors, yield, let's look at 13. Here we go. Data universe. I mean, so basically you've got the best part of 50% API on, on SN 13 until 20, what is it gone? 2029. It's nuts. And the thing is these, these, the dates of when a subnet, yeah, so no, ignore me now. But yeah, so it's pretty cool. And when a subnet gets deregistered, etc, and a new one takes its place, guess what? That harbing schedule starts again. So, yeah, it's always two years between it, maximum, between armies to subnet. Yeah, that's interesting. We just had a question pop in here. Is there a point

51:47that we will not have enough miners to do all the work on the subnets? Not at all, agents that are limited. Yeah. Like, for example, one of the things I'm like a little side hustle for Goli, I just wanted to test it. Things I gave out a bunch of capabilities. And I said, hey, here's subnet 74. What do you think? And you're like, oh, this is really cool. Yeah, we could easily mine this blah, blah, blah. And then the one stumbling block it had was you need a GitHub account and longer than 180 days. I was like, okay, go find one to buy. So yeah, Goli's gone out and he's sourcing an aged GitHub account for us to buy. So he can then mine get tensor using Opus 4.6 or open AI 5. Well, good luck with that. I mean, in my first flush of excitement with Hugo, I was like, right, you're going to do what I'm going to do, which is to do

52:50the momentum trading on subnets. So I'll show you how to do it. I'll teach you what I would do if I could be bothered to be awake 24 hours a day. And then you're just going to do it. He's like, yeah, yeah, no worries. Yeah, hit me up. Like, here's a wallet, put some town it. I'm off to the races. I put some town it next day. He was like, good afternoon. I've forgotten all of the details of that wallet. He's absolutely useless. That was the trading. And then I said, right, okay, don't worry about that. Let's get you going with some mining. He was like, I'm going to need another wallet. So I gave him another wallet with some more money. Lost that one as well. It's so painful. I think I'm going to do it. I think it's a prompting. Give it. I'm treating my Gordy like a brand new member of staff. And when you have a brand new member of staff, you have to treat that person like an idiot in a nice way. And tick every box like, so I've got, so I've given Gordy so many different SOPs with standard operating procedures from SOPs on how to store data

53:54and how to store passwords, how to an SOP on everything. And so maybe it's that maybe Hugo's missing out on the SOP. Yeah, I mean, I have upgraded his memory. A great deal since then, but honestly, I just know my limits a little bit there. And it's an absolute time vampire. Because I really enjoy it. It's great fun. You feel like you're right on the edge. But actually, I've got other stuff that I could be working on that's probably going to be higher value than working out how to do all of the things that I could do with an agent. Yeah, OTC deals was on your list of things to talk about. It took you talked in your article about how the OTC deals that we've been doing have changed over time. And yeah, that's absolutely true. From when we did our first OTC

54:57deals as DSV to stuff that we've done much more recently with DSV and with Astrid, I think we're much more creative now. And there's been a number of changes there. But the biggest one is probably that people know who we are now. And there's a higher level of trust, a higher degree of trust and familiarity than there was when we were doing those first few deals as outsiders. But yeah, everything is all malleable. Every deal that we ever put on the table whether that's an OTC deal or buy a slot to launch a subnet, any of that stuff is always in the hunt for the win-win-win where everybody involved in the transaction takes a win out of it. And nothing is set in stone. But yeah, we've done probably more OTCs than most and more recently, but more subnet slots than most and launch some teams. And if anybody wants to have a conversation, I wrote this on a post the

55:59other day. If anybody wants to talk to us, if you're thinking about investing in subnets through the OTC mechanism or launching incubating subnets by buying slots and like we're up for a call. I'm certainly up for a call. I did have had one the other day on the back of my post, but if anybody wants to have another one, like everything that I've learned, that we've learned in the last 13 months of investing in detail is open source knowledge as far as I'm concerned. You can have it because the more people there are doing stuff in this space, the better it is for everybody. And I saw someone comment that, oh, BitSarter is just your personal launchpad. I was like, it's not bad at all. We don't own BitSarter. BitSarter is there to help try and create that conveyor belt of new talent in BitTensor. But it's just highlighting a point which you and I have spoken about so many times that there is no one that is actively hunting and

57:02attracting brand new talent like a factory conveyor belt line of teams in. And it's just that we're trying to give it a stab in the dark and give it a go. It just happens that with anyone's that are like bringing stuff to the table, maybe. We can't win. Like one of the first criticism of that we have no tech depth. And so, you know, you should lean more on the people of BitTensor for tech depth. We're like, oh, there's this BitSarter thing over here. It's got loads of great minds in it. Keith Singeries in it. It's got loads of devs in it. You know, some people in macro cosmos, they look at the code. So we thought, oh, how about we do our own DD, see if it's, you know, a viable subnet-ish. Then we give it to BitSarter and they go, hey, and we go, what do you guys think? If they give it the thumbs up, we're like, screw it. Let's just let's go for it. But I'll say, said to that person, I was like, please feel free. Other people need to bring stuff to BitSarter. Yeah, I mean, I know Chris and Gus, well, a BitSarter, and they would absolutely welcome more

58:08investors coming in and more projects coming in for them to look at. But you're right. It does feel sometimes like there's not a lot of people worried about bringing people in. We talked about that, that relegation battle and subnet's getting deregistered. But where is the inflow of new talent coming in? Because that is what we need. Otherwise, if we just replace six months worth of deregustrations with six months worth of kind of talk quality talent, we haven't moved forward. We need to bring the right ones in. And if you feel like you've got a role to play in that, then step up and make it happen. But to be fair, there are new subnets that are lodging with a BitSarter not by BitSarter. So I had an interesting chat with two lovely guys called Ning and Terry, who are subnet 11. Ning and Terry are ex-Microsoft and ex-Google.

59:15It's hard to explain. They're trying to streamline agentic workflows, but for enterprises. And I think they're already making, they've got some contracts, but all the enterprise selling is all in China. So it's something which probably won't get as many eyeballs, but you know, shit hot team. From what I see, I think strike, yeah, strike is backing that one, or our God. He's got many pseudonyms. But yeah, there's mothers. So we need new energy, new talent, new people coming into BitTensor and building on it best place to be. And yeah, BitSarter is great, but on the BitSarter front, the only thing that I would say is that if you've got one slot and one project, the people are going to front run you because they will know where that project is going to land and that does rather ruin the BitSarter launch. So that is why we got to a situation where we need to kind of hold through four slots to do a successful launch. Otherwise, it just doesn't work. And I mean, that is the

1:00:24is the beauty and the beast, I guess, of having on-chain transparency where everyone can see exactly what's going on. Because we see that a lot with any movement from Astrid or DSV, exactly the same. If we buy something, if we sell something, you got it in the neck the other day because we'd sold some ridges. Yeah, I took a screenshot of the wallet and I was just pruning positions of multiple positions in that wallet. I think I sold like four or five hundred towels worth of ridges, but we still have like three and a half thousand towels worth of ridges. So it's like a small percentage. Yeah, it needed to be trimmed, didn't it? Because it was up to like almost 30% of our portfolio in ridges. And we've got huge belief in ridges and in Shaq and in JJ and Voon. We think they're going to smash it. But it is not sensible for us as allocators

1:01:32of investors capital to have so much of it tied up in a single subnet because the truth of it is that our conviction here is in bit-tensor as a whole is in towers in the ecosystem. We don't need to risk that upside by being too concentrated in one element or two elements of it. The people that are investing in DSV want exposure to bit-tensor and yeah, they do want guided exposure to bit-tensor. They want us to use the knowledge that we've got and the relationships that we've got to invest in the right places. But we just needed to measure and trim the risk that we've got there. So we can invest that in other parts of the ecosystem. Yeah, I mean still what probably 23% of DSV's portfolio. And ridges have launched their public

1:02:36thing. They're the public product now. They've called it ridge line. And yeah, there's a lot of growing for this but I think what they're doing is getting really the low hanging fruit, nice product market fit. And then I think they're going to start attacking the claw, code replacements etc. So yeah, hopefully it goes well, Shaq and Voon. So yeah, so that's that. And then I just yeah, two other random things. I saw Max post himself, so Max is seven owner of score, seven net four to four. And he said bit-tensor offers so many opportunities it's ridiculous. Just on a random Sunday, I use lead power new app to generate leads for Maniko, his vision products. I rented an H100 on Basilica to power auto search and start training a new model for Sire based on data collected by the best football model from score.

1:03:40Just need a proper alternative to Claude code now. What a time to be alive. Yeah, eventually that'll be the ridges thing. And then he went on to say that you let auto search iterate on our football player value function model overnight. Didn't touch it once, 19 experiments, 95 minutes, H100, it's blah blah blah. But long story short, he found some massive improvements overnight and he led to a 11.1% improvement. So yeah, the rate of innovation and development is just hard to keep up with. Yeah, Max took a bit of fun the other day, didn't he? And that was not nice to see, because we know a lot of people in this space and Max is one of the best. We've got a good plan. It's nice guy knows what he's doing and he's building something really, really special there with a good team around him. Yeah, I think a jealous subnail owner did that hit piece on him and it was all complete rubbish points. Yeah, tears rubbish. Right,

1:04:47should we go through some questions before we round this up? Yeah, yeah, if you want to, yeah. Where were we blah blah blah? Raymond said this, hey guys, could you explain how liquid alpha to works? I'm not sure. I'm just going to say if you're struggling with the question, you're definitely going to struggle with the answer to that. So I think that's a no. Yeah, Dylan says you said Casper is going to be massive. Yeah, let's talk about Casper very quickly. I love Casper. I think I think it is, and I still believe it is the world's best layer one. And I was in Casper for a couple of years and unfortunately it, and I did say like to, you know, to the people in Casper for a long time that it needed to get into smart contracts a lot sooner than it did. And they're very resistant to it. And they're like, no, no, we want to be the best fastest, you know, base layer blah blah blah, which it is. And it like everything about

1:05:51it is great. But then yeah, and so we exited as I guess a community around between, we wrote it from like two cents all up to 16, 18 cents when we got out. Sorry for this messy chart. And I still think it will do well, but it just hasn't found its thing yet. And like if you look at Tau against Casper, it, I mean, it shows it, it is the right, it was the right move because when we got out of Casper, it was down here around the, this is a ratio. It's just Tau divided by Casper. We got out end of 2023 around around here. So when this ratio is around 13, 1400. And even though everything has gone down in prices, you look a bit quite near at 69K. Tau has more than like, it's like tripled or quadrupled Casper in performance because just sat here at six and a half thousand in terms of this ratio. So the first little hole was hit again, a little bit later than I thought.

1:06:53And then what I said, it would be, it would move into this big old Obergine over time. And I still think it will probably be over here. But yeah, so I like Casper. I really do. But you know, you've got to, you've got to ride the wave. And that wave is, I mean, AI is the most important thing humans have ever, ever created. And to not be exposed to AI in some form or another is, I think it's just myopic. How is that an Obergine, by the way? It's such a specific reference for what was just a badly drawn circle. Yeah. Right, what else have we got here? Next one, can the 21 million Tau Cup be raised on Bitcoin? We need a fork. How does it work in substrate? Sorry, not a new question at all. And technically, it could be like open-tensor and const, everyone could get together and go, yeah, let's raise the ceiling. But it would come

1:07:55with a fierce backlash. I don't think anyone has the appetite to increase the, the cap. I'm pretty sure const doesn't want to increase the cap. So let's leave it at 21 million. Yeah, there's 982s now. And then Kilibur made onto that one. And then last question, I think, was about Tau 5, who's managing the transition? Yeah, again, we are, we want to, so there was an incentive mechanism going towards the bridge. That, when it was on Summit 10, it's now on one, on one, two, seven. We will be turning on the new incentive mechanism to that. So yeah, the transition is happening now. So Leo and Arter and Cyprian, they're working towards that. But yeah, I guess Astrid could be more frequent on the communication on Twitter. It's just at the moment, we're all got our heads down on just flailing, just trying to get, get shit done

1:08:59at the moment. So that is true. Yeah. I think that's it. Unless you have any other questions. No, that is it. The only thing is to say that we will be heading to London on Wednesday for the fancy dress competition and BitTensor social. Yeah, you've got to come with fruit. And yeah, coming as a large Obergine. Yeah, so I'm excited about that. It's always a pleasure to be in the room with some other BitTensor people. We've got, and that's happening twice this month, isn't it? Because we're off to San Francisco in a few weeks as well. Yes, we are. Break out. Break in. So we've got London this week, and then San Francisco in a few weeks time. Yeah, both, because people ask, well, where is it? Let's have a look. It's the Sunday bar thingy. I think here we go. If you go to Sunday bar,

1:10:03something like one to one on Twitter is this purple-leap pink thing. Yeah. So they're doing a presentation at their office and then drinks and shizzle. Last words, will we find it? I'm mentioning that it's fancy dress on there that seems an oversight. It definitely is there, isn't it? Yeah. Fruit and veg. Fruit and veg. Themed. Wonderful. Cool. Let's get out of here. When are we back? Do we know when we're back? Let's have a look. Stand by. Tomorrow. Oh, we? Yeah, we have a revenue set with... No, no, we don't say who it's with. No, we don't say what we're talking about before. That's the only hint. Yeah. Okay. Good. Well, that's been a pleasure. We'll see you tomorrow. Thank you. Jimmy, have fun. And at Siam Kid, to catch them as they drop, learn more about DSV at dsvfund.com.

1:11:07And if you're building something brilliant on BitTensor, we want to hear about it.