Latest / Crypto Coin Show / AI creates amazing Prediction Markets w/ PredX | Blockchain Interviews
Transcript
- 0:00I'm Ashton Addison from the Crypto Coin Show, and today on Blockchain Interviews we
- 0:04have doctor Rain Wu, founder and CEO of pread X. Rain Welcome to
- 0:08the show and thanks for taking the time. Thank you, thank you for
- 0:12having me. Very excited to dive into AI in blockchain prediction markets, what
- 0:18the future may look like, and what you and your team have been working
- 0:21on at pread X. I would love to start off our conversation from a
- 0:25high level for those who aren't familiar with the project, if you could talk
- 0:29a little bit about what is pred x, what have you and your team
- 0:31been building, then we can dive into all the details afterwards. Yeah,
- 0:36great to do that. Yeah. So PREDA it's a generative prediction market ecosystem.
- 0:41So we are trying to build the most engaging and informative prediction market ecosystem
- 0:46in the world. So we as actually coming from the angle that AI actually
- 0:51empowered the Web three technology, especially the generative AI parts. So most of
- 0:57the current prediction market they are actually based on the platforms generated events. It's
- 1:03not actually generated by ars, it's not committee based. So we want to
- 1:07use an air technology to enable the community members as well as managers to creating
- 1:15their own asset so that they can list in them but as a treatable asset
- 1:19in the platform as events. And so that is actually the general idea,
- 1:23but we also figure out that the current prediction market we lack of two importancing
- 1:29One is a relevancy and second is liquidity. Internal relevancy is that most of
- 1:34the users, they're actually coming to the website but seeing exactly the same screen.
- 1:41The events is the same. So no matter I'm a political interest in
- 1:45the political events, or I am a kind of like a hagical guy who
- 1:49are actually looking for some Microsoft A media events, or am I just a
- 1:53personal digion who want to trade cryptocurrencies, They're all looking at the same screen,
- 1:59actually all even the same, so that one of the big problems.
- 2:02The second thing is that most of the market don't have good liquiity, so
- 2:07specially that when you put the order, either it's hard to fund someone to
- 2:12be a counter party, or there's a bid a bid as a spread between
- 2:17your orders. So that actually makes the liquity of the market pretty bad,
- 2:23which also actually give people bad costomic experience. So we want to have this
- 2:28kind of take out layer to promoting the events inside the community so that the
- 2:36people will find out the most interesting event by using our app or our website.
- 2:42The recommendation fully automatic via processed. We also have the liquidity providing engine
- 2:47which are injecting the liquidity to different market so that we will not be a
- 2:53seriously curity problem if I think like an AI market maker, automatically integrating into
- 2:58our systems. And also we want to motivate everyone to promoting our events,
- 3:05so Induescence we basically have a very well designed revenue shared economics, so that
- 3:10will allows the community managers for the KOLs to creating but also promoting these events
- 3:16because most parts of the transitional fee generated by our platforms will be shared to
- 3:23the people who creating these events or promoting these events. The Indicescence, you
- 3:28can think like everyone will have their own critical exchange but at the same time
- 3:32also sharing or have the benefits of receiving huge amount of revenues from our platforms.
- 3:39Our Idea is trying to build up the ecosystems to enable individual persons and
- 3:44community members have their own prediction market, but we're using our ai plass website
- 3:49technology to enable them. So that is what we do. Very cool and
- 3:53thank you for that overview, Rain And I've seen some prediction markets before,
- 3:59but yeah, there definitely have been issues with either irrelevancy or liquidity, which
- 4:03I'd love to dive into a little deeper. But I want to just first
- 4:08of all, explain a little bit more about those those for those who haven't
- 4:12used prediction markets, you're talking about events, you know. One some of
- 4:16the things that come to the top of my head when I think about prediction
- 4:20markets are, for example, predicting the who's going to be the President of
- 4:26the United States in the next election. But you can have almost any type
- 4:30of event, right like whether it's a sports event or betting on if this
- 4:33coin will be listed on ginbase or anything else. And I guess that's why
- 4:40there's so many options that it's hard to make it relevant for everybody on the
- 4:44front of the page. Is that correct? Doubtinely doubting. There's so many
- 4:47different topics and coverage and the basically you can feed the universe intel potational markets.
- 4:51Yeah, and now introducing AI into the equation sounds like an amazing way
- 5:00to make it relevant for each person. Can you talk a little bit about
- 5:02the merge of blockchain and AI finding a suitable use case and how exactly you
- 5:09can use artificial intelligence to bring more relevancy to the prediction market when a user
- 5:15goes there for the first time. Yeah, of course certainly from general perspectives,
- 5:19right, So the current prediction markets are not defined by the community members,
- 5:25so they're actually generated by the platforms players. What do you think like
- 5:30your majority of the audience coming from and then basically we we actually feed the
- 5:34events towards are kind of theman or you can think average the user standard.
- 5:40Right. But the problem is that if you using this way, it's really
- 5:44hard to actually creating something that most of people enjoy it. So you really
- 5:47need to have a coupled AI engine from both the generative side but also the
- 5:53predictive side to let them work together. So you only have the generary plats,
- 5:57it's really hard to actually to recommend this to the right users. Right.
- 6:01Everyone see a lot of huge amount of events. But if you only
- 6:04have recommendation side, probably that many colored events. They are very very kind
- 6:11of a diverse and really decentralized. For example, with the Indian elections as
- 6:15a kind of like the US city and you even don't know the names so
- 6:19it's really hard for like an operation company, right, so actually to fit
- 6:24each people appetite or the personal interest. So you really need to have a
- 6:29decentralized creation capability as well as this kind of decentralized recommendations or like the central
- 6:35topic modeling capability. So that is the two things you really need to have.
- 6:40So AI actually help you creating something, but really the decentralization of people
- 6:46to promoting this inside the community. Yeah, So my personal understanding is that
- 6:49the webstery is really about the people business, but AA is about automation,
- 6:55right, so you need to like feed these two things well into the new
- 6:59game. So the AI actually really targeting about how you can make the events
- 7:05generate more tranditional volume. From from perspectives, but from a person perspective,
- 7:10is that how you can make them want to put orders. It's kind of
- 7:13like attention motizations, right, so you need to first catch people attention and
- 7:16then monetize them. So how to keep people attention? That is the thing.
- 7:20Yeah, The next level question is that how you can really make the
- 7:25topic more interesting enough to let people start actually training it some of the topic
- 7:30actually Tomorrow, whether New York will ruin or not probably it does not matter
- 7:33too much, right, So it's like I feel that people feel that interesting,
- 7:38but not too much. People want to really train it as a backside,
- 7:41you need to have some conflicts or counter party underlying on the topic side,
- 7:45so that actually the larger language model can help us to actually make the
- 7:49topic a little bit arguable or basically figure out like scripting the internet right,
- 7:54figure out some of the interesting topics that may kind of crossing the conflicts between
- 7:59different foursurces, and then we can put there as a topic. So the
- 8:03reason like advance of the model really helps a lot, that could say,
- 8:07because without them, it's really hard to model all these topics the sentiments.
- 8:11But the thing is that like I don't think there's too much people have ever
- 8:13thought about using that to creating the treatible assets. So uually the general AI
- 8:20was used for like a general image or generally like an article messages, but
- 8:26if generally treating assets and people to treat on it, we don't think we
- 8:30see so many product out there, probably all of the first people never trying
- 8:33to do that, and especially in the battery field, the battery you know,
- 8:39people actually more degent, like they're actually looking more for the tokens rather
- 8:43than the productselves. So I think that probably if we can create the programs
- 8:46as good as your token system and using the token reay to promoting them to
- 8:52make it more engaging, like people to actually broadcund in the community, you
- 8:54really need to like have this aim westery combined with each other for that to
- 9:00create something already for the users. It's a great point Rain I find when
- 9:05you're in an earlier stage company and it's in the Web three space, you
- 9:11really do have to focus on those crypto native users, those early adopters what
- 9:16they want, and then as it develops out to more of a general betting
- 9:22and gambling or other people that aren't in crypto as well. Then I'm sure
- 9:26you know the political or sports betting, which is a huge industry in itself.
- 9:33Combining that into Web three, they're going to see the benefits of the
- 9:37AI and the blockchain technology, but I think the token first, just because
- 9:41they're so eager to use this kind of technology that makes sense, and you
- 9:46mentioned about the decentralized nature of it and the community driven part is very important.
- 9:50I wanted to ask you a little bit about the blockchain side of pred
- 9:56X. From what I looked into the documentation, you're actually utilizing the SAY
- 10:01network blockchain, which I've used a little bit, but it is a newer
- 10:05blockchain. I'm sure it has tons of advantages. Could you talk about why
- 10:11SAY blockchain is the best for products and how it fits into the ecosystem nicely?
- 10:16Yeah, definitely. So one advantage of SAY is based on their parallel
- 10:22VN set up. So you know, the ev N is actually a general
- 10:28compatible like underlying infrastructure right or empower a lot of different kind of like blockchains,
- 10:35including these leader tools and as well as them layer one I deserve it.
- 10:41So in that sense, the VN that means the general compatibility, which
- 10:46means that system can be really easily go multi chain. And also it provides
- 10:52a good number of engineering base which can help us implement the high level like
- 10:58a complexity in terms of data structure as all the functionality. Yeah. So
- 11:03the one choice for the SAY is actually based on the v M narrative.
- 11:07But the difference between its v M other v M is its parallel u M,
- 11:11which means that they have the potential to RND the system new ultra fast
- 11:18mode. So why is important is because when we allow everyone they're creating a
- 11:22defend on assets, they will be a huge number of ages happened in the
- 11:28network. So you assume they're like an traditional kind of like a exchange,
- 11:33right, so the exchange is a center, no, and all these others
- 11:37kind of like traders are surrounding this centerde So the complexity of the problem is
- 11:41all in its number of traders. Right now, if you enable everyone to
- 11:46create on assets, the complexity Oden square, which is actually we can magnitude
- 11:54higher. So basically the idea is that like there's a lot of the different
- 11:58kind of things happening, and because the creation capability, because of the like,
- 12:05we have a new dimension out there. So if you want to make
- 12:09it decentralized and empower everyone, you really need to have a high speed network
- 12:13that can potentially the capacity to handle this level of workload. Because it's not
- 12:18only about asset creation, but also the urticles. You need to upload the
- 12:22data to supporting the outcome of events, right, so that is one big
- 12:28thing. Another is we need to record everything on the chair. So if
- 12:31everyone creating their own asset, you can create everything interaction between asset and also
- 12:37user themselves. Right, So there are two kind of like type of like
- 12:41ages, you know, like the type of notes basically the asset nodes as
- 12:45well as the users you're connecting them. Yeah, that creates a huge level
- 12:48of complexity, so we need to based on U Actually one of the art
- 12:54is about chain, so based on some numbers that have the same efforts six
- 12:58times a Salona previous consider but the best actually speed up like block chance.
- 13:07So in thatscense, actually we feel that they have a big future. Another
- 13:11thing that like we gotta say it's EVM compatible, so that means we have
- 13:16a lot of people who had coding on it now. So actually the people
- 13:18side as well as like the a infrastructure side, we actually make this consideration.
- 13:22But to be also open to others blockchains and in the future we definitely
- 13:28able to go mutag chain. Yeah, m hm, very cool. I
- 13:31I didn't know that about the speed of say being six times faster and Solana
- 13:37and I'm going to look into also the parallel EVM, which sounds like,
- 13:41yeah, you want the best of both worlds because there is a huge network
- 13:45effect on the EVM chains, uh, and they don't want to have to
- 13:48bridge over an extra steps to have to ux if they want to try it.
- 13:54So that's really interesting. I'm looking forward to seeing how that develops,
- 13:58and also keep an open mind for multichain or cross chain compatibility I think is
- 14:05a good idea as well. Rain. Can we talk more about the application
- 14:09itself, n You guys have been building it for a while. From what
- 14:11I understand, the current stage is in the test net right now. So
- 14:16what does that mean and what can you do on the test net currently?
- 14:20Yeah, So in the test net, actually the general idea is to figure
- 14:26out some of the potential problem bags or limitation of the system. So the
- 14:30general idea be basically trying to implement everything we have and see what the user
- 14:35feedback. So thatta actually the purpose. So what we do is that we
- 14:39actually built out the online structures everything on the test at and also our platforms
- 14:45surrounding it because like we still have the other book of chain, but we
- 14:48have the constandian basically the users can think about it on chain. So that
- 14:54is actually a combination of hybrid mode ye in that's basically trying to test this
- 15:00infrastructure whether it's good or not. Yeah, because like other tests that will
- 15:03be manut we're using the real money. But if there's some problem with the
- 15:07real money, it's really hard for you know, further blood chains actually to
- 15:11revert back or everything. So that's the thing we need to have everything fully,
- 15:16i mean laid out and tested. Yeah, so you know, the
- 15:20on the test net, the thing that like right now, say deploys VNM
- 15:24starting from the test acup dem net. So in that sense, basically we
- 15:28have a very good platform to test out all the features of say, and
- 15:31that is how we why be starting from the test net and trying to make
- 15:35this more ready and then we can go with the manet, which might happen
- 15:39in the next selver or two months is good to know. And what I've
- 15:45seen with other test nets is there's some kind of incentivized test net or you're
- 15:50encouraging early adopters to try out the test net for whether it's future rewards or
- 15:54some kind of status in the application upon the main net launch. Is there's
- 16:00something like that happening on products tests and as well, yes, be looking
- 16:04forward to it. Yeah. So the thing that like protests that another good
- 16:08thing will led people to accumulate some of the scores and based on their interactions
- 16:14with the platform, right so accumulating some of the footprints, because yeah,
- 16:19if they interact with smart contract, if they actually deposit or withsdraw some money
- 16:26from our contract, and if they're won the events or the ranks are high
- 16:30in the big or too was secure consecutively or the logging like in five can
- 16:37stay in a row, so that will give them different number of like rewards
- 16:41and letter round. That actually will transform you to some air drop potential or
- 16:47air job expectations. So that is actually would be going to rundo game basically
- 16:52in cinema people to be small. Yeah, that's nice to know, especially
- 16:56when you're on the test net and you don't have to use any real money.
- 17:00You just experimenting with the application and then you get rewarded in the future.
- 17:03And with that, I'm guessing that leads to a pread x token of
- 17:08some sort. Can you talk about if there's a token launch, how does
- 17:12that fit into the ecosystem of products? You know, what's the value of
- 17:15the token? Does it create a sustainable ecosystem for all of the users if
- 17:21they're using the products token in the platform? Yeah? Yeah, sure,
- 17:25So right now we are in and a lot of like token functionality. We
- 17:27have not deployed or testa yet we're actually trying to roll over a step as
- 17:32that I could give a general high level like description of what we have.
- 17:37So we indeed have a token system and it plays a very central role in
- 17:41our role map. So in order to have a good token system, right
- 17:45from a gender personective, the token need to have real value. So you
- 17:48need to have like some like a real benefit we can keep to the users
- 17:52instead like just to actually make a like a token and you just share some
- 17:56future you know, incompetition after you do the tge and all like that.
- 18:03So we really want to have let the token holders to share a big chunk
- 18:07of revenue platform internal transaction fee. So if the we have like a two
- 18:11tokens analysism, one token is called prdx, so basically you can think that
- 18:17government's token and like a normal utility token, right, so you basically using
- 18:21that token to do a lot of things in our platforms like upgrading the parameters,
- 18:26or you can use in the tokens to to vote or actually works on
- 18:30the events actually especially which actually even up going like listening or not. So
- 18:37that's basic capability and as well as the system can actually one important features share
- 18:45twenty percent of the transaction fee with all the prdx tokenholders, so that they
- 18:49actually have the motivation to hold it more for the longer as well as actually
- 18:55it may have some real values if the perfect become like a really profitable another
- 19:00eighty percentage, we go with the pr LP token in the second token.
- 19:03You can think it's like more T or ARC one fifty five. This is
- 19:10kind of like a kind of policy. The idea is that each event will
- 19:15have their own pr LP token, so which means that like if I am
- 19:18even the creator of this event, I only can share the revenue from the
- 19:23event I created. So in essence, like only the people who most or
- 19:29create the most interesting events can get eighty percent of the revenue of that event
- 19:33from our platforms. So that actually, dramatically, we thought will incentivize people
- 19:40creating greatly e when are promoting it inside the community because of the big chunk
- 19:45of our revenue, just give to the event creators. Another kind of pr
- 19:48LP holders are the people who injecting equated to our platforms. So the idea
- 19:53is that we will let people to deposit your USDC or this kind of like
- 20:00other podcasts, and we have AI liquative providing engines to distribute this kind of
- 20:04like liquidity two different events based on the need. So user can just deplod
- 20:10it to our system and we will decide which kind of like events will advocate
- 20:15more token term of marketmaking. Yeah, so that is actually fully automatic intelligence.
- 20:19But the people who deploded pr uh, the who deployed your DC will
- 20:25actually get the real commission fees sharing because firstly they may take some risk because
- 20:30they deploded, and secondly because like they put the money there, so they
- 20:34need to be uh kind of like GUS listenerators. So that is a second
- 20:40ye of people who holding the pr lp to and basically who count built the
- 20:44creating or raining these events. So this category the users of the type two
- 20:49token basically prr pole. So that is actually in general to economics, and
- 20:56we are trying to get this tokenomically after run, after we have the men
- 21:00net available and the TV where we will be on in July or basically second
- 21:06or third quarter of this year. We got trying our best. Yeah,
- 21:10very exciting. I didn't know that about the two token system. That's very
- 21:14intricate and I'm looking forward to seeing how that plays out. So moving forward
- 21:18with you know, your potential roadmap. Of course, Uh, there are
- 21:22so many factors and variables. But you mentioned that the main net could be
- 21:26around a few months away and then the token launch afterwards after the main net
- 21:33has been running for products for a while as well. Yes, yes,
- 21:37awesome. And what is the best way for people to follow along with those
- 21:42updates and to try out the incentivized test net and prepare for the main net
- 21:47launch? Yeah so, actually pretty easy. So please follow our Twitter and
- 21:52also our discord so that's our two men channels. You can also joined our
- 21:56tigram so tigram is security for the discussion of our communities and the Twitter and
- 22:03discorder you already we have the offictional announcement. Yeah so, and also you
- 22:07can visit our website pr E d X dot ai and the data. Also
- 22:11we have the newest update from the news side. Sounds great, Rain,
- 22:15thank you so much for the information. I can leave those social links and
- 22:19the website in the show notes below as well. All the best with preducts
- 22:25moving forward and let's definitely follow up in the near future. Definitely, thank
- 22:29you so much for having me. Yeah, thank you. Question