The Delphi Podcast
The Delphi Podcast
TAOFU's Mikel & Mitch: Enabling Capital Formation for Subnets on Bittensors $10B+ Network
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Join Tommy Shaughnessy as he hosts Mitch and Mikel, co-founders of Taofu and TPN, to discuss Taofu, a launchpad for Bittensor subnets. Learn about how they're democratizing subnet funding and their first subnet launch.
Taofu: https://www.Taofu.xyz/
Taofu Funding Form: https://form.typeform.com/to/KmvNfUg5
🎯 Key Highlights
▸ Mitch and Mikel's journey from DeFi to discovering Bittensor and Tao
▸ Taofu's subnet launchpad: selling future emissions upfront to avoid selling pressure
▸ How Dynamic TAO (dTAO) changed Bittensor's power dynamics from voting to market-based
▸ The challenge of subnet registration costs and predatory funding deals pre-dTAO
▸ Taofu's tokenization process: multisig custody and Subnet Seeds (SNS) tokens
▸ Tao Private Network (TPN): First subnet registered post-dTAO, powering decentralized VPN
▸ How miners earn rewards for providing VPN infrastructure across 80+ countries
▸ Vesting mechanics and governance voting for SNS token holders
▸ Future AI predictions: robotics integration and potential for extended human lifespan
💡 Subscribe for more crypto & AI insights! 🔔
🧠 Follow the Alpha
▸ Taofu's Twitter: @taofuxyz
▸ Mikel's Twitter: @0xDutchmikel
▸ Mitch's Twitter: @Axilo
🔗 Connect with Delphi
🌐 Portal: https://delphidigital.io/
🐦 Twitter: / delphi_digital
💼 LinkedIn: / delphi-digital
🎧 Listen on
Spotify: https://open.spotify.com/show/62PR1RigLG2YN5Pelq6UY9?si=18ac7ccf36ab4753&nd=1&dlsi=50105fd66e6c4124
Apple Podcasts: https://podcasts.apple.com/us/podcast/the-delphi-podcast/id1438148082
Youtube: https://www.youtube.com/channel/UC9Yy99ZlQIX9-PdG_xHj43Q
Timestamps
00:00 - Intro: Mitch, Mikel & Taofu
01:00 - Mikel's DeFi background and transition to Bittensor
02:00 - Mitch's growth background and getting "red-pilled" into Tao
03:00 - Taofu genesis: From liquid staking to subnet launchpad
05:00 - Dynamic TAO (dTAO) impact on Bittensor power dynamics
06:00 - How Taofu solves subnet funding and exit liquidity problems
08:00 - Pre-dTAO barriers: High registration costs and predatory deals
11:00 - Taofu's tokenization process and multisig custody
14:00 - Due diligence and curation model for subnet selection
17:00 - Balancing AI advancement vs. Bittensor ecosystem growth
22:00 - Dream projects: Creative AI and privacy-focused solutions
25:00 - Tao Private Network (TPN): Decentralized VPN subnet
28:00 - TPN's global infrastructure and miner incentives
33:00 - User experience and dynamic incentive adjustments
34:00 - Upcoming TPN sale mechanics and vesting structure
37:00 - Taofu platform evolution: Analytics, governance, and bridges
38:00 - Future of AI: 6-month predictions and longevity implications
42:00 - AGI timeline predictions and closing thoughts
Disclaimer
This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host and members at Delphi Ventures may personally own tokens or art that are mentioned on the podcast. Our current show features paid sponsorships which may be featured at the start, middle, and/or the end of the episode. These sponsorships are for informational purposes only and are not a solicitation to use any product, service or token.
You're now plugged into the Delphi podcast. Hey everyone, it's Tommy. Welcome back to the podcast. Today I'm thrilled to have on one of our portfolio companies, Taufu. I have both founders, Mitch and Mikhail. How's it going, guys?
SPEAKER_01Very good. Thanks for having us.
SPEAKER_00Oh good. Thank you. Mikhail, let's start with you. We'll get a bit of your background. We'll go to Mitch and we'll dive into the project.
SPEAKER_02Yeah, my background is mostly in the DeFi part, uh, specifically on EVM. Um, just over three and a half years ago, me and my partner set up a regulated DeFi native yield farming fund here in Amsterdam, which to my best knowledge was also uh the first one to do so in the Netherlands. And after uh doing DeFi um full-time for like one and a half years as a business, um, I also fell in love with the concept of uh building a DeFi protocol myself, which led to the first um builder endeavor of uh deploying lending and borrowing market um called Lendel on the Mental Network. Um, and after doing that for a while, um Mitch red built me into Tau. We got together to see if we could somehow find a good marriage between my background and his. Um, from that Taufu was born, but I'll leave it to Mitch first for his background.
SPEAKER_01Cool. Yeah, my name is Mitch. Um, of course, one of the co-founders of Taufu. My background in crypto is more on the growth side of things, where I've helped exchanges and other platforms doing big growth campaigns. And one of the main reasons for ever coming into crypto was sort of to gain access to high-quality deal flow, um, which is kind of aligned with what say Echo or Legion or Taufu are doing.
SPEAKER_00That's awesome. So, Mitch, you were the one that uh red pilled Mikhail into Tau. What was that like? Was that like a late-night talk? Was that over time? It's always interesting to hear those stories because we all remember getting red-pilled into crypto.
SPEAKER_01Yeah, so for Tao at the time, it was like around the GPT moment where I really realized that AI would be the next big thing in tech in general. And then there was this guy in a shared WhatsApp group um that we were in together, and he kept on rambling on about Tao and why it's special. And at some point I just decided to look into the white paper. Um, and when reading that, I was like, oh wow, this actually might be the uh yeah, it might be the next big thing. And that's like pre-subnets, pre-dTAW, pre-everything, when everything was really premature and early. And then after doing some digging into like the people that were following it on the sidelines, I saw a lot of Eve Maxis um basically following this, and that made me realize that this might be uh the next big thing. And then I reached out to Mikhail and a bunch of other friends, like, hey, it's rating really low right now. We might uh yeah, we might need to buy some.
SPEAKER_00That's awesome. When was that? How long ago are we uh are we talking?
SPEAKER_01I think this is like end of 2022.
SPEAKER_00Nice. So that was like so Tau's at 10 bil FTV now. That was pretty early in the the life cycle.
SPEAKER_01Very, yeah. I think that's also one of the parts which made um Tao interesting at the time. I think we got in at like uh one uh 180 million market cap.
SPEAKER_00So give me the genesis for for Taufu. So you guys are you find BitTensor quite early, you're spending a lot of time, you're seeing you know the early subnets, the early community, you've watched it for the past, you know, let's call it, I think, you know, several years now. How did you come up with the idea for Taufu? And and let's just dive into exactly what Taufu is doing too.
SPEAKER_01Yeah, well, it's a long story. But like the start of Taufu, we saw that there wasn't really an easy way to buy Petenser at the time. So we figured why not make like a liquid staking protocol uh protocol for TAO um and do it in a way that it's uh institutionally friendly, um, so that people or institutions with a lot of capital would have a safe way of holding and acquiring TAO, as it wasn't really listed on any exchanges. And there was like one um custodian here in the Netherlands, actually in Amsterdam, that offered these services for TAO, and they were the only ones doing that. Um so we set out to build like the liquid saking protocol um and STAW for that. And then at some point when the Micar regulation got announced, it actually became impossible for us to continue the concept as they backed out. Um so we had to build it ourselves, and then basically after uh taking six months to build the whole custodial setup ourselves and make it institutionally friendly and all that stuff, uh DTAW got announced. And basically DTAW then uh yeah, destroyed the concept of building a liquid staking protocol, at least for us.
SPEAKER_00What is the the main innovation behind behind DTau for you guys? Like how do you view that changing the BitTensor network? I mean, plenty of people, including myself, have talked about it on Twitter, but just curious how you guys view it and how you view that helping Tau long term.
SPEAKER_02Yeah, I think the largest shift um that that DTAO basically enforced onto the network is a change of power dynamics from a vote-based system where uh a number of or a limited number of large validators get to decide um where the incentives are allocated. It is now fully transformed to um an economical and market-based system where price dictates um how much incentives all of the subnet gets. And so the largest change is basically just um from a num a limited amount of people um setting weights in a risk-free way to uh market dynamics, and uh you're gonna have to put your money on the line if you believe something and put the money where your mouth is.
SPEAKER_00That's helpful. So let's dive into let's dive into Tafu a bit more from from first principles. Like, I don't know if we should describe how you guys are helping subnet owners or we should describe like end-to-end how it works. How do you guys want to take like the uh the intro to Tafu?
SPEAKER_01Yeah, maybe as like uh jet like a bit of general background information, the way the emissions work on BitENSER. So on TAO, uh miners can secure 41% of the rewards, validator secure 41%, and then subnet secure 18% at this current stage. And then the amount of emissions a subnet would get, the the emissions I just described are per subnet. The amount of emissions in Tau that they would then receive really depends on the EMA of the subnet. So the higher the token price a subnet is trading at, um, the more emissions in TAW the team behind that or the subnet owner would earn. So I think that's like um handy to offer like some background um insights. And then right now, what you have on Petencer and the subnets that are there, there's basically no easy way of access for most people to invest in these things, not even for VCEs. Everything's happening behind those closed doors, um, and it's kind of being gay cap. So you don't really know, okay, who are they funded by, what is the exact deal, um, what is the market structure. And if the investors or um other partners have like some sort of stake in a subnet, there's no real way for them to have like an exit strategy. There's no real exit other than dumping the alpha tokens that are tied to the to the subnet into the pool. And if you would basically sell your alpha tokens into the pool, it would then also hurt the future emissions of a subnet, um, basically making it a less interesting case for them business-wise to continue or be incentivized to continue development of the subnet itself. And that's basically what we're looking to solve uh with Taufu. And then we can go like a little bit more in detail on that.
SPEAKER_00That is really helpful. I mean, I mean, so if you want to launch a subnet on BitTensor, which has dozens and dozens of subnets, the only way to do so without Taufu would be like what are your options before you guys exist? Like, is it raise VC money and and fund yourselves? Is it be personally rich? Is it convince the community? Like, what was the path before Taufu?
SPEAKER_02Yeah, that's quite an interesting and um sort of a double, double asseted question where um before the Dita migration, the registration cost for Subnet was really high, um, in the orders of like uh one mil plus in in in dollar terms, which is not something that many teams uh can just pay out of pocket or bootstrap. And furthermore, um it is it was sort of a lockup. So then you register to subnet, pay the fee, and if you deregister the subnet, you get back the registration cost. So you basically needed a bank to provide set money, but now post-DTAO registering a subnet um burns the TAO, and the registration cost has come down by quite a bit. So I think on average, subnets register for around 150 tau right now, which is a lot cheaper than the thousand plus tau that it used to be. So registering a subnet became more accessible now, in regards to um before DTAW. So hypothetically, you could you could now do it out of pocket, but the more important part for Bitensor um as an ecosystem and also for Taufu's role into this ecosystem is that now there is no longer the need to find um uh an entity with very deep pockets in terms of tau willing to um basically lend you the tau for registering the subnet because now the registration costs come down. Taufu democratizes the process of funding um for new and also for existing subnets without having the need to lock up too much of the TAO for registration. So most of the um most of the funds from the race can actually be good, um can be put to uh work operationally um to fund the business or to validate on their network.
SPEAKER_00That is really helpful. Yeah, getting a trying to get a million-dollar loan per subnet must have been difficult. You guys probably saw some like pretty good projects not launched given that that barrier, I'm assuming.
SPEAKER_02Yeah, and also unfortunately, um some of the deals for that were provided um where people did help uh co-fund the um the subnet slot were quite predatory, meaning that the subnet owner would have to renounce quite a large chunk um of their ownership or of their alpha tokens because there was basically this power position where you didn't really have any other option, and then when it's like uh a supplier's market, so to speak, they dictate the terms.
SPEAKER_00Yeah, that is that is quite annoying. You just lose so many shots on goal with when the barrier is is that high. So turning to Taufu, so if I want to start my own subnet, I have this phenomenal idea. I go to Tau Fu and could we just run through like a sell process end-to-end, like mint, price and curve, vest, like the custody. It would just be awesome to hear for potential people who want to launch a subnet on Taufu.
SPEAKER_02Yeah, so there's there's two parts um of this. Um, first, it's good to provide a bit of background where um Bitensor is a substrate-based chain, meaning that it's built on the Polkadot um SDK. And then we have a frontier integration um for the EVM, which basically means it's not an L2, but it's uh an EVM sort of way to communicate with the base chain. But the subnet logic um mostly lives on the substrate part. So the owner is registered to a key pair consisting of like a code key and the and a hotkey. And upon um wanting to tokenize on Tafu, the ownership, the full ownership of the owner key is transferred to a joint multisig wallet where TaoFu is a cosigner, subnet owner is a cosigner, and two uh trusted third entities are cosigners. The subnet owner is then granted um proxy uh access to the admin ballot, meaning that they can run operational upgrades to the subnet, but they cannot touch the financial part, so the alpha. Um for that they we need full cosigning of the multisig. Once that is uh done and secure on the Bitensor EVM side, we then uh generate a uh contract specifically for each subnet where the contract mints 21 million fungible tokens, um, which we call SNSs, subnet seeds. And the subnet owner then gets to decide how much of their subnet seeds um they want to sell on the launch pad. Um, it can be 10, it can be 20, it can be 30%, um, anything they'd like to go with. The Taufu funder base can then fund the subnet. Once the sale concludes, the participants can claim their SNSs. We automatically add liquidity to uh Taufu's decks consisting of a part of the SNSs and a part of the fund raised. And then the sale, the tokenization and sale process concludes after the add liquidity event.
SPEAKER_00That's really helpful. Would it be would it be simplifying too much to say that Taufu is a launch pad for future subnets where the owners can sell a percentage of their future emissions up front so that they don't have to sell their subnet tokens, which impacts the price negatively?
SPEAKER_01Yeah, that's a perfect description.
SPEAKER_00That is really cool. So if I'm if I'm if I want to launch a subnet and I go to the community, how do I decide? I guess being on the other side of the table, if I want to buy some of the the subnet itself, I need to make my own guess as to the amount of emissions that that subnet will receive from the network, right? I guess that's the biggest the biggest what-if question.
SPEAKER_01Yeah. Um, of course, it's really hard to expect something um prior to to prior to having the subnet live on how good it will perform. Um, but just like anything else in crypto, um, you have like an aspect of speculation, you have an aspect of like a team being really skilled and executing on what they're setting out to do. So if the team really delivers on what they're supposed to be building, and then also they have their alpha locked into a multi-sig through Taufu, then that's also really interesting because that would mean less selling pressure on the alpha token itself, alongside with like a good skilled team, then that would you would expect that that would drive positive uh price sentiment to their subnet and therefore then also acquiring more emissions.
SPEAKER_00One other question I had for you guys is it always felt like there are a small amount of potential subnet slots, right? Um on Tau attracting emissions, but for Tau Fu to be successful, you sort of want to launch uh many subnets, right? And hopefully some very successful ones. How are you guys thinking through you know launching many subnets versus you know the fight for emissions, right? It becomes a it becomes quite hard.
SPEAKER_02Yep. I mean the classical analogy, right, is that we shouldn't try to take pieces from the pie, but that we have to grow the pie collectively, which in in in this case sort of means yes, uh there will be many, many, many subnets, uh, hopefully in the future. And no, the number of tau emitted per block is not going to increase. That is hard-coded. So indeed, um, every time a new subnet is registered, sort of the pool of people or the pool of subnets wanting to receive and tap into the uh protocol native incentives is diluted, which also makes for the good point that in order for Pitensor as a collective, meaning the all of the subnets together to um to grow, the most important factor is not to cannibalize from within the system and by launching subnets drain value from other subnets. The um better goal would be to collectively strive to attract as much value capture from the outside world, be it um crypto or be it Web2. But eventually, the whole goal, the whole point is to have as much value capture into the system. And then once the full value capture into the system grows, that automatically trickles down to things like tau price and therefore total incentive value, and thereby also per subnets an increased incentive budget. And that is something that uh Tafu and good Subnets launching can definitely contribute to.
SPEAKER_00That is helpful. I mean, two of the things I'm most excited on in crypto is is obviously capital formation is is a killer use case, and we're we're talking about that here specifically for Tau, but lowering the barrier to to get that capital is just leads to more shots on goal and more creative ideas. So it is it's an exciting concept. One of the harder things, I'm I'm curious to how you both approach this is you know, it it is up to the market to decide what's valuable and what isn't, but you're both in a you know a privileged seat as to, you know, do we think that this specific subnet will be successful or not, right? I'm I'm curious, like how will you go about you know talking to founders, interviewing them, and making that decision of, you know, hey, we want to do, we want to help this founder launch their subnet on Telfu.
SPEAKER_01Um, so one of the things that we look at when we're talking to teams ourselves right now is trying to find out if they can deliver on what they're promising to do. And also looking at it from a more traditional business perspective, where say a team is building something that would be relevant in Web 2, and if they're able to leverage TAO and basically subsidize part of their costs uh through the incentives that TAO offers, um, if they then decide to utilize some of that revenue for example for buybacks, because you have less costs on the the normal side where you would have like hardware and infra and all sorts of expenses that you don't have on TAO, that could be a really interesting case for them to capture more emissions in the network and also um to to have like a sort of easier pathway to building a profitable business. Yeah. So I think that's really interesting about teams building on Tau, and that's something that we try to look at ourselves. And then instead of always trying to do it ourselves, we also want to work in a curator model where, for example, we can work with the guys from Contango or um, you know, Yuma, uh, where they would then take charge of doing the due diligence on the teams and companies that they're working with before that they're actually going on Taufu. Because we we want the best for the people investing through Taufu. Um and yeah, really filter out the stuff in the weeds.
SPEAKER_00That is really cool. Mikhail, I'm also curious your take. Like, what are the like key things that you look for with a subnet or sorry, potential subnet and and the founders of which to launch on Taufu?
SPEAKER_02Well, Taufu is a permissioned launch bed, meaning that you have to go through basically through us in order to launch on Taufu, which makes us sort of responsible uh in a way for protecting the community. So what um in in addition to what Mitch said, what I think is very important is that um we don't allow crazy valuation sale subnets, and setting a valuation um eventually the decision is up to the subnet owner, but we do have constraints as to what what we deem to be reasonable barriers for um for setting the valuation. But then, um, like you also said, eventually it is an open market, it is a free market, and people have to make their own investment decisions. That's not something that we want to or can do for them. But in the grand scheme of things, to me, what's really important is that um it's very important that there is a very clear understanding of what the external value capture is from the efforts of the team and how those value captured are funneled back into the subnet. Because as a team, you have your operational business and then you have your subnet, and for some teams, the subnet is the product, but for other teams, the subnet might be a subcomponent of a larger business that they run, and specifically for the uh teams and companies that have the subnet as a part of the business, um, it's very important to protect SNS holders that there's sort of like a guarantee how that subnet functions in the grander scheme of things, um, because we don't want to fall in a situation where then the company uh takes off, so to speak, and they cut the uh the the profit share or the buybacks with the subnet, and then all of a sudden the whole subnet dilutes, they move their business elsewhere. Those like high-level conceptual blueprint decisions, um, they they are very important to me.
SPEAKER_00It's awesome. It's it's great to hear that you guys have such emphasis and effort on finding the best founders. Maybe a second order question for you, but you you both spend so much time in in crypto AI, and I guess there's like two hats, right? It's you know, we want to find the best subnets to push the Tau ecosystem forward, but on the other hand. We may want to support a project that just generally pushes AI forward for humanity, right? And I'm not sure if both of these conflict, but you might have to make a decision on do we want to support just AI broadly or something that you know fills a niche within the BitTensor eco. I'm curious if you've you've ever had to deal with this or if this is not an issue.
SPEAKER_02We we actually have where one of the subnets that we spoke to, they are very much on the side of wanting to build uh sort of the Linux for um global AI, um, in a way where they have open source first. They want to build like global, open, open, openly accessible systems for ML, for AI, for models, for everything, which is a very, very long journey, uh, a very cost-intensive journey, although a very important journey. But it's not always easy to find the right funder type for these types of endeavors, especially not in crypto, where many people are uh still just also looking to grow their portfolio and maybe don't have the capacity to make 5, 10, or maybe even 20, 20 year long horizon deep tech investments. So you could almost say that they're sort of public good or they're leaning towards public good type of um subnets, which I think are very important. But like I said, it's not trivial to find the correct funder base for those.
SPEAKER_00Before we get into the the first subnet you're doing the cell forum, I'm just curious on longer aspirations and maybe a reach. I'm curious if you both had to pick a project type that you would love to see launched on Taufu, like dream project, like feel free to, you know, obviously not open AI, but you know, would it be a decentralized training project with a with an ex OpenAI founder? Would it would it be an inference project? Would it be maybe a hardware play of some kind? I'm not sure. Like, what would be the absolute dream project that you see launch on Taufu in six months? Maybe we'll we'll start with Mitch and then go over to Mikhail.
SPEAKER_01Um, I think for me it's more on the creative side. Um, there's so much foundational stuff being built right now by extremely talented teams. Um, but what what I would love to see is music generation really leveling up or repopulating old games through through agents, more more stuff in in that sense. So, not really traditionally on what you would expect as an answer. But for me, the creative part, I think there there will be so much amazing stuff happening on that front where you could basically extend your favorite movie, um, have like an alternate ending. I think yeah, that direction is something that I'm I'm really excited about for the future.
SPEAKER_02Yeah, and for me, there's I I hope that um sometime in the near future uh a large base of um builders, a large base of founders finds out that you can creatively uh deploy the Bitensor infrastructure for many, many, many, let's say, questions where you know what you want the answer to be, but you don't know uh or you don't even care how people get to the answer, which can be like very broad. And if you think of like the network in such a way that basically enables miners to run their own um type of solution providing for your specific problem without you knowing how they do it, and if you have this like radically open system solely focused on the result without um bothering about the methodology, um by default, the system will be more efficient than if you put constraints on uh on the solution providers, and so we are seeing uh an increased number of subnets which are sort of trying to focus on more niche use cases, um, which might even not need to be tied directly to either crypto or AI. And I think that will be a very interesting wave to witness on Bitensor. And on the other part, I think one quite like large issue what people aren't really realizing enough yet is the privacy issue when using mostly LLMs. People really like they don't care what they tell their LLMs in in like chat GPTs or clouds or whatnot. And so if people were to realize what they're actually sharing and what they're actually doing, and then then hopefully they realize that on Bitensor there's a subnet that's specific or a combination of subnets most likely that's specifically built to have your own trained model with your own uh secure storage, with your own um preferences, I think that would be highly needed in the world.
SPEAKER_00This is a good a good segue to the TPN as a test pilot. Like let's go into the the VPN subnet. Like let's talk ambitions there, let's talk what the plans are and why you guys are so focused on that as the first go.
SPEAKER_02Yeah, for for context, for um for the people is we we indeed do Taufu, uh which is uh a service specifically tailored to subnet owners. And then we also have a subnet ourselves, which is called Tau Private Network, and it's subnet 65. Fun fact it is the first subnet that's registered both DTAO and um, I think to date also the cheapest subnet registered so far, which is a fun gimmick. Nonetheless, the the purpose of um TPN is to have a developer-friendly and open infrastructure for any dev to tie into for building um IP tunnels globally. Um, the first use case that comes to mind then naturally is a VPN application um for consumers, which is also what um the team of Dow Private Network is shipping as a flagship product to showcase to the world look, this is what the decentralized VPN service powered by Bitensor can do. But since it's an open system for developers to use, we don't know what the future might hold and what people might be building. But the important part, what it comes down to, is that TPN enables for secure and private IP connections across the globe. I think we already have over 80 unique um countries where you can build an IP connection. And why this matters is it is quite important on the internet that you can hide yourself or uh mask your IP location for security purposes, but also for censorship purposes. One very fun example, for example, is that in um Spain, the the guys that streamed the football there they uh wanted to stop uh people from illegally streaming, and so then they thought that they found out a server from which um people were trying to stream. They ended up blocking Cloudflare, which led to a situation where if a football match was being streamed in Spain, half of the internet in Spain collapsed because uh it had nothing to do with um people trying to illegally stream, but they just blacklisted one big uh Cloudflare server. That that's a really weird situation to me, and that that should definitely be uh people should definitely be able to avoid such situations.
SPEAKER_00So the I think one of the interesting parts too about offering a VPN subnet, and I think everyone in crypto is is definitely up to speed on VPNs, um, is that you're not only using Taufu for the coming sale, but you're using the miners on the BitTensor network to power the VPN, right? Is would that be fair to say?
SPEAKER_01Yeah, yeah. So the the miners are basically powering the entire VPN from start to finish, and that's saving us a lot of cost on stuff that we would normally have to set up ourselves. And then also on the crypto side, if you would try and make um a VPN protocol, you would have to really get the incentives right, and it needs to be worth it for people. And we're just lifting on top of the the network effects that DAO currently has um to do so, and what we've built is really open, so anybody can build like their own VPN app on top of this, or any developer on Bitensor can actually leverage the API and utilize it for themselves for whatever they're doing. So whether they're data scraping or prompting different LLMs from say Russia or China, you would get like completely different outputs, and that's really interesting because that's all like incentivized and powered through TAU. And there are so many apps or so many things that you could think of that you could do like a similar approach uh for this.
SPEAKER_00That is pretty cool. How do you convince the miners on Tau to run and support the VPN? Like, do they earn token rewards for doing so? Because I'm assuming this increases the the hardware requirements and complexity of running a miner.
SPEAKER_02The miners um actually get paid for offering up their services um in our alpha token, uh, in the alpha 65 token. And right now, since we're still bootstrapping the network, the usage of the subnet is free, uh, meaning that anybody with uh basic knowledge of how APIs work um can use the subnet and build uh build an IP connection somewhere. And um, in the current sprint we're working on, um we're doing um a payment payment rails deeply embedded into the subnet structure, where for um anybody wanting to request an IP connection over uh TPM has to uh pay a microtransaction, which will then be used to buy back our alpha token. And so once we have that completed, it is um sort of like an L1 analogy where it doesn't really matter what is deployed on the chain, so long as there's transactions, um, the chain can have growth. And then our flagship consumer app is also just a client of its own network, which I think is a pretty cool setup.
SPEAKER_00Yeah, that does sound pretty cool. I'm curious, what is the user experience like versus a popular VPN today? Like there are, you know, I'm out of my depth here, but I know that there's just like a lot of metrics around DNS leaking and and how hard it is to track and things like that, and and obviously just the user experience on latency and yeah, just curious how the uh how that shapes up.
SPEAKER_02Yeah, that's that's actually where where the real fun kicks in um as a subnet owner, because the the parameters that you reward miners for in your subnets are adjustable, meaning that um depending on how the subnet um is used, we can steer the incentive mechanism in a way that incentivizes miners to do better at that time at the thing that we need or at the thing that what the subnet needs. So until now we've mostly been incentivizing uh uniqueness of geographical IP location. Um but depending on the type of users and the type of traction that the subnet gets through our app and third-party applications, um, we can basically monitor what those applications need. And if it's lower latency, we can incentivize lower latency for miners. If it's more bandwidth, we can uh incentivize more bandwidth. If it's more more granular uniqueness of IP locations, then we can steer the system that way. So it's it's sort of like this ever-evolving, moving river meandering to um where it needs to go.
SPEAKER_00That is helpful. Let's walk through this the coming sale. What are the what are the dynamics there? When does that sale go live? Um, if I'm a user and I visit the site, like what sort of information am I getting? And how do I, you know, how would I take part in that in that sale?
SPEAKER_02So the the first SNS sale um for TPN will take place um in the upcoming uh weeks. The exact confirmation of the dates uh we will confirm on the official um X account of both Tao Fu and also TPN. So keeping a lookout for any of those two channels, you'll definitely be kept up to date. Um for now we are collecting a funder base, so to speak, where everybody that's interested in participating in the sale um they can just um hop in the Discord request for the link and then they can sort of already pre-register into the system. But um for the sale period itself, um, we will have a whitelisting period where um which will last a couple of days where you can register and basically show interest for the sale with your wallet address. And um the sale itself will also last a couple of days. And if you have registered your wallet for the sale, you're free to participate at any given time in the sale so long as it hasn't sold out yet, because it is a sale with a fixed price and a soft cap and a hard cap. So once the hard cap's reached, then basically the sale is done. Um yeah, and for the rest, actually, it's quite quite a straightforward process.
SPEAKER_00And Mikhail, how does vesting work for sales on Taufu? Like, because there might be a world where a subnet becomes extremely popular and gets a lot of emissions from Tau, which is different from the vesting schedule that I guess Taufu would offer? Or yeah, it's just curious because the value flows might be might be different on both sides.
SPEAKER_02Yeah, to to basically ensure um that it's a good sale for the people that participated in the sale, we have a mandatory um vesting for the subnet owner part, for Taufu's commission part, and for the potential curator commission part, which by default is a linear vest over two years for the sale participants. Since the price of the sale is identical to the price at which we add liquidity, um, there is no vesting whatsoever. So if you have participated um and bought SNS in a sale, when the sale concludes, everybody can claim 100% of their tokens immediately. It is just that the subnet owner uh and the Taufu commission and the curator commission they are vested over two years.
SPEAKER_00That is that is helpful. And I know you guys have ambitions for the site itself to go beyond uh just the sale, right? You want to do analytics, you want to do much more. Like, how do you envision like the homepage itself for Taufu uh growing from here?
SPEAKER_02In in the current form, Taufu sort of has three uh actually four components to what we do. Uh one is the tokenization, one is the sale, and one is the uh the decks. Then we also have um a bridge from for going from BitTensor EVM uh to substrate and and from substrate back, which is not a bridge that we uh build ourselves. It is using the uh uh the official bridge from the OpenTensor Foundation, but we've put it in a nice uh user experience in a nice jacket so that people can just use it. But more importantly, we have uh a governance module where SNS holders they get to vote where the alpha that the owner key has accrued needs to be delegated to, to which validator. And so we have a good interface for actively participating in the governance part post sale. And then we're working indeed on detailed analytics per subnet that we incubate for SNS holders to keep track of.
SPEAKER_00That is cool. It'll become like you know, way more than just a sales site. You'll have all analytics, you'll have stats, it'll be be cool to see. Guys, let's switch to the let's switch to some of the fun stuff as we close out. I want to know what you both think about the future of AI. And the reason I'm asking you both specifically is because you've spent so much time, you know, five plus years in the weeds of the first crypto AI project. And on the other hand, you're you know, fully tuned into what's coming on the pipeline for AI subnets who wants to launch while watching everything go on with AI from OpenAI to Gemini to Anthropic, et cetera. Yeah, I I don't know the best question to ask you both, but maybe just the future of AI. I don't know, maybe we'll say six months, because anything longer than that might be uh might be totally wrong.
SPEAKER_01Yeah. Um well, if we look at Google's presentation, of course, last week, I think the future is extremely exciting. Um, there's gonna be so much stuff that we can do from like repopulating old games or generating music or having like an agentic team of coworkers that you can just steer and have them do whatever you want them to do, and they'll work 24 hours a day, forever, infinitely. And I think it will go as far as you know, drug discovery, making rapid advancements, and and basically any area in life that we really care about that will yeah, that will improve for the better, and that's like super exciting. And then as for the teams building on BitTensor, I think BitTensor will just evolve alongside AI, and whatever new trends emerge or new breakthroughs uh come through, there will be open source teams testing out their concepts on BitTensor, bootstrapping themselves in the process, and then by generating enough uh you know runway through Tao, they can then move along and start like a bigger business if they don't really get swallowed by the by the market of AI, which is like that's happening so fast right now for many teams. They're working on something, and then a year later, uh Google basically does whatever they did for free, and that's crazy. Um, so yeah, all in all, I think life is gonna be much better for anybody. Maybe we'll live longer, and that's like uh the part where I'm most excited about. And it really feels like Ray Girls Wilde's uh vision for your future is coming much faster than we uh would have thought maybe a couple years ago.
SPEAKER_00Yeah, that is helpful. The the Google release uh just this week of May 22nd recording, so two days ago, was just all encompassing. It felt like they hit every single target market you could with an AI product. Um Mikhail, I'm curious your take. Like, how do you uh do you differ from Mitch? Or do you guys have are you in the same boat?
SPEAKER_02No, sort of in the same boat, sort of in the same boat. Um me, myself, I'm also very interested to see how the how the robotics uh meets AI is going to go in in the maybe not the next six months, but definitely um the upcoming couple of years, where I look I I don't really see a world where um people from uh let's say our age right now don't at least age into like 120, 150, maybe even 200. Which if I tell people for the first time, like, yo, I think I'm gonna be at least 120, or maybe I'll kill myself in like a motorcycle accident or something. But I don't see any reasonable world where with the current acceleration of AI, there should be a reason for us to die um from natural causes, given the the advancements both on the computational side, but then also in translating that into hardware, both um biological hardware, but also like mechanical hardware. So yeah, I'm I'm really looking forward to that.
SPEAKER_00Do you guys think we get uh AGI in the timeline that's been shared? I think it's some people are saying 2027, some go to 2029, but it yeah, what what's your personal date on on getting that?
SPEAKER_01Uh yeah, hard to say, but maybe sooner than we think. Maybe 2027 is uh is a good guess. Yeah.
SPEAKER_00That is awesome. Well, well, Mitch and Mikkel, thank you so much for coming on. It is awesome to learn about Taufu. Like just capital formation is just a massive use case for crypto and a proven one. Um and lowering the barrier to to just create that capital is obviously a massive boom. So really excited for for Taufu, the coming sale you guys have on that Thursday, and uh and appreciate you both coming on. This has been awesome. Thanks. It was a pleasure.
SPEAKER_01Thanks for having us, and um, yeah, see you soon on the first sale.
SPEAKER_00Yeah, absolutely.
