Latest / Crypto Coin Show / XYO Layer One just got 2–5x faster — and most crypto investors still haven't noticed Arie Trouw, XYO
Transcript
- 0:00Speaker 1: I'm Ashton Addison from the Crypto Coin Show and today
- 0:03Speaker 1: on Blockchain interviews with Ari trou co founder of x Yo,
- 0:07Speaker 1: CEO of x Yo Labs, building x yo Layer one
- 0:11Speaker 1: network in the intersection of deep in, decentralized physical infrastructure, networks,
- 0:18Speaker 1: data and AI and blockchain being at the forefront of
- 0:21Speaker 1: it since we first started following xyo in twenty eighteen,
- 0:25Speaker 1: before Deepen was even a thing, and they've been going
- 0:28Speaker 1: strong with the twenty twenty five release of the x
- 0:32Speaker 1: Yo Layer one network as well and the intersection of
- 0:36Speaker 1: AI as that sector continues to grow. There's a lot
- 0:39Speaker 1: going on and excited to dive into all that today, Ari,
- 0:43Speaker 1: thank you so much for taking the time.
- 0:45Speaker 3: Thank you, ash Yeah.
- 0:47Speaker 1: Excited to hear the latest on the growth of not
- 0:52Speaker 1: just Xyo Labs, but the Layer one and the recent
- 0:55Speaker 1: updates to the network that has made it even faster
- 0:58Speaker 1: and cheaper than it was. It was already good from
- 1:01Speaker 1: what I understand since the launch in Q three, twenty
- 1:05Speaker 1: twenty five, but the team continues to improve and AI
- 1:11Speaker 1: is only getting better and Deepen is only growing more
- 1:15Speaker 1: into tokenization data verification, which I think AI plays a
- 1:21Speaker 1: big role in as well, maybe we could start with
- 1:23Speaker 1: a high level on how exactly x YO ties in
- 1:29Speaker 1: AI with data verification and deepen, and you can sort
- 1:33Speaker 1: of break that down and Layman's terms a little bit more.
- 1:38Speaker 2: Xy other protocol and XCEL one the layer one that
- 1:41Speaker 2: we have. It's always been based on data storage and
- 1:45Speaker 2: providing products for data and verification for data, bringing on
- 1:50Speaker 2: change so that way basically immutable and unchangeable, and with
- 1:53Speaker 2: AI coming around, there's more and more need for this,
- 1:56Speaker 2: and also there's much more data being used and consumed.
- 2:00Speaker 2: There's a lot of questions as far as is it
- 2:02Speaker 2: actually consuming and using the correct data, is it using
- 2:05Speaker 2: get it as as opposed to or allowed to making
- 2:07Speaker 2: sure that data is supposed to stay private, state is
- 2:09Speaker 2: private and the data was supposed to public is public.
- 2:14Speaker 2: So we really focus on those different aspects all along,
- 2:17Speaker 2: and now with with AI coming along, we've become more
- 2:20Speaker 2: of a linch pen as far as the need goes
- 2:22Speaker 2: to be able to do those different things and AI
- 2:25Speaker 2: calls for those. So we've been very excited about the
- 2:28Speaker 2: AI growth and also the ability we have to to
- 2:31Speaker 2: put things on chain and to add providence to it
- 2:35Speaker 2: without having to put the entire idea and chain for example,
- 2:37Speaker 2: So our our efficiency for storage is extremely high compared
- 2:42Speaker 2: to most bloxings out there.
- 2:43Speaker 1: M hmm. Yeah, And you mentioned there the private side
- 2:47Speaker 1: and public side with data verification, especially with the increased
- 2:52Speaker 1: use of AI with everyday people in small businesses, more
- 2:56Speaker 1: people are starting to think about, uh, not just the datafication,
- 3:00Speaker 1: but the security of their data and the privacy of it.
- 3:03Speaker 1: Because as AI is working so well, we're just giving
- 3:06Speaker 1: more and more of our private and confidential data into
- 3:10Speaker 1: these AI systems. And there's permissionless blockchains, but there's also
- 3:15Speaker 1: permission blockchains or like public and private blockchains. Where does
- 3:19Speaker 1: that fit in where you have maybe data data verification
- 3:24Speaker 1: or location data that's public and then private and balancing
- 3:28Speaker 1: those on a blockchain network.
- 3:31Speaker 3: So what we do is we really separate the actual storage.
- 3:34Speaker 2: Of the data and then the metadata for the data,
- 3:36Speaker 2: and primarily the metadata goes onto the blockchain, so for example,
- 3:40Speaker 2: the fact that it was stored, it's hash in state
- 3:42Speaker 2: and time.
- 3:44Speaker 3: For example. You can also have.
- 3:48Speaker 2: A reduced version of the data on there, so say,
- 3:50Speaker 2: for example, the location can be very precise in the
- 3:53Speaker 2: store data that's in a data lake that's private, but
- 3:55Speaker 2: then on the blockchain you can actually start a much
- 3:59Speaker 2: less precise location social like you might not it might
- 4:02Speaker 2: be in California somewhere, but not war specifically. And the
- 4:05Speaker 2: thing was nice part our network specifically is that once
- 4:09Speaker 2: it gets put on chain and that that providence is
- 4:12Speaker 2: stored there and the permanence is stored there, you can
- 4:15Speaker 2: later on reveal it and show it, so it's you
- 4:19Speaker 2: don't have to to go improve again in the future.
- 4:22Speaker 2: That was correct, and so the actual data itself, the
- 4:25Speaker 2: full data was actually preserved as far as being being
- 4:30Speaker 2: persistent and made permanent on the chain earlier on without
- 4:33Speaker 2: having to reveal that entire thing. So it's basically it's
- 4:34Speaker 2: optional and privacy and it's it's revealed on demand privacy basically,
- 4:40Speaker 2: M can.
- 4:42Speaker 1: You elaborate on probably the ninety five to ninety nine
- 4:46Speaker 1: percent of businesses that are putting a lot of their
- 4:48Speaker 1: data into these AI systems and the lll ms and
- 4:52Speaker 1: ones that are owned by you know, the oligopy of
- 4:55Speaker 1: the fang of AI companies, why should they be thinking
- 4:59Speaker 1: about data verification on blockchain or the privacy aspect.
- 5:05Speaker 2: Well, on the verification side of things, you know, obviously
- 5:08Speaker 2: when you're creating models and you're learning, you know, learning
- 5:11Speaker 2: about bad things is going to make something inaccurate, so
- 5:15Speaker 2: you want to have purity of data when it comes in.
- 5:17Speaker 2: It's one of the concerns by AI is like, well,
- 5:19Speaker 2: it's really good at reading the Internet, and as everyone knows,
- 5:22Speaker 2: not everything on the Internet is true necessarily, so there's
- 5:25Speaker 2: some verification that goes on there because if it just
- 5:28Speaker 2: regurgitates what's being reported on or what's being set on
- 5:31Speaker 2: X or on different social media networks, it's going to
- 5:35Speaker 2: correctly reflect what the Internet is and says, but not
- 5:39Speaker 2: necessarily correct quickly validate or say what is in the
- 5:42Speaker 2: real world. So I think that's one of the concerns.
- 5:44Speaker 2: But there's also the providence concern, where's what did it
- 5:46Speaker 2: learn from and did learn from something which was actually
- 5:49Speaker 2: copyrighted or pattern that should it be using that data
- 5:53Speaker 2: and where it comes from, And there's a lot of
- 5:54Speaker 2: difficulties with that because once it goes into a model,
- 5:56Speaker 2: it becomes blurred and it becomes kind of lost that providence.
- 5:59Speaker 2: So I think for their own concerns as far as
- 6:04Speaker 2: defending where they came got stuffed from, they they have
- 6:06Speaker 2: to at some point start having audit trails and more
- 6:11Speaker 2: of a proof of what did you learn on, where
- 6:13Speaker 2: did it come from? And how do we know where
- 6:15Speaker 2: the answer came from? For verification purposes.
- 6:19Speaker 1: Definitely, I think with AI, yeah, it's just pulling data
- 6:24Speaker 1: from you know, some of the top Google results, But
- 6:27Speaker 1: where did that information come from? How can we use
- 6:31Speaker 1: the data verification through through x yo to prove or
- 6:37Speaker 1: double check that the sources that these ar are pulling
- 6:41Speaker 1: from are verified? You know, I can imagine if you're
- 6:44Speaker 1: looking at location data, specifically having a drone or a
- 6:49Speaker 1: weather station, having that location verified might be easier than
- 6:53Speaker 1: when you expanded out to everyday common knowledge. And in
- 6:57Speaker 1: the social context, it's hard to verify every source. And
- 7:02Speaker 1: how does the network on the blockchain pull that and
- 7:06Speaker 1: verify things in a way that the centralized AI can't.
- 7:11Speaker 2: Well, what we do is we do what we call
- 7:13Speaker 2: witness based system. So you have these different witnesses that
- 7:16Speaker 2: are witnessing things like temperature. For example, Let's just say
- 7:19Speaker 2: if you know a thermometer, they just get them the temperature.
- 7:21Speaker 2: What do you want to do is you normally want
- 7:23Speaker 2: to sample reality for lack of a better term, with multiple
- 7:27Speaker 2: anonymous sources. So if you have multiple different parties actually
- 7:30Speaker 2: checking the same thing, but they don't know about these others,
- 7:32Speaker 2: so they can really collaborate on making false data, they
- 7:37Speaker 2: can then basically be trusted because if four people or
- 7:40Speaker 2: five people that can up with people. But four different
- 7:43Speaker 2: or five different sources get asked for the same piece
- 7:45Speaker 2: of data and four out of the five actually matched
- 7:48Speaker 2: and the fifth ones in outlier, you can you assume
- 7:51Speaker 2: that probably the fifth one is is is problematic or safe?
- 7:55Speaker 2: For example, two out of three are correct. You can
- 7:57Speaker 2: always just pre sample and see, well, you know, not
- 8:00Speaker 2: lot of the consensus. So it is it's almost like
- 8:02Speaker 2: having micro consensus on de deflection mm hm.
- 8:06Speaker 1: And you mentioned earlier bout the provenance knowing where the
- 8:09Speaker 1: data came from. It's almost more important than the data itself.
- 8:13Speaker 1: And I always take news headlines, uh, you know, newspaper
- 8:18Speaker 1: headlines or digital media as a great example. You know,
- 8:21Speaker 1: it's there's a headline, but like who's writing it is
- 8:24Speaker 1: almost more important to know.
- 8:27Speaker 2: Yeah, I've always been kind of fascinated with that, where
- 8:29Speaker 2: like literally I've written, I read like different articles are
- 8:32Speaker 2: written by people and factually they're all accurate. And you know,
- 8:37Speaker 2: you can't fall to any of them, but you can
- 8:38Speaker 2: tell from the tone, or you can tell from the
- 8:40Speaker 2: way it's told, or what's emphasized and what's not emphasised.
- 8:43Speaker 2: Like there's a different perspective on those things. And people
- 8:46Speaker 2: have gotten very good at those those techniques to be
- 8:49Speaker 2: able to to communicate a story in a certain way
- 8:52Speaker 2: which it's part of their their their added becazazz to
- 8:57Speaker 2: that story. They have to to have some personality in there.
- 9:01Speaker 2: But at the same time, if it goes so far
- 9:02Speaker 2: where it's actually untrue, it becomes problematic potentially, or if
- 9:07Speaker 2: it omits omission to some degree. Also can be a
- 9:10Speaker 2: case where where it's factually untrue because you admitted a
- 9:14Speaker 2: specific thing. So omission and untruth are not that far
- 9:19Speaker 2: away from each learn but also one of the most
- 9:21Speaker 2: important things. I'm actually an anti privacy advocate for many things.
- 9:25Speaker 2: Obviously privacy is important for security and for all those things.
- 9:29Speaker 2: But the idea of having people who are anonymous reporting news,
- 9:34Speaker 2: for example on AX and saying, oh, yeah, you're believing
- 9:37Speaker 2: with this account is saying we don't even know who
- 9:39Speaker 2: the account is or where it's coming from. It's you know,
- 9:42Speaker 2: it's titled whatever they want to. The idea of having
- 9:45Speaker 2: some sort of of providence where it's like, you know
- 9:47Speaker 2: this this account has been reporting this, you can't take
- 9:50Speaker 2: it down. You see this software with people who use
- 9:52Speaker 2: APIs for XL screenshots some things like hey, you know
- 9:55Speaker 2: this person posted this thing and then thirty seconds later
- 9:57Speaker 2: they took it down, but forever they have the screenshot
- 9:59Speaker 2: of it.
- 10:00Speaker 1: Uh.
- 10:00Speaker 2: But really what you want is you want to have
- 10:03Speaker 2: things go on their permanent record, where it's you have
- 10:05Speaker 2: to stand behind what you say, and if you're not
- 10:07Speaker 2: willing to actually put your identity on that, then it
- 10:11Speaker 2: should have much less weights because you know you're going
- 10:13Speaker 2: to stand behind it. And also you can't edit the past.
- 10:16Speaker 2: You can't say well, I never said that, where if
- 10:18Speaker 2: it says well, obviously you know on the blockchain this
- 10:21Speaker 2: was the case. Stereo was witnessed by different people and
- 10:23Speaker 2: it was stored. You can't go back from that. So
- 10:26Speaker 2: having transparency into the providence of something, I think it
- 10:30Speaker 2: is a big step towards improving the quality of data
- 10:33Speaker 2: and the validity of it because then you can actually
- 10:37Speaker 2: use that either as a feedback loop for determining which
- 10:39Speaker 2: ones to trust because after a certain number of times
- 10:42Speaker 2: that if something's found false, you stop listening to it.
- 10:45Speaker 2: Or also it could be even used as a way
- 10:47Speaker 2: to prosecute somebody if they were intentionally you know, putting
- 10:51Speaker 2: false things on there or attacking somebody you know, a
- 10:55Speaker 2: certain way. So I think that transparency is important. Then
- 10:58Speaker 2: we can really provide the transparency and permanence of posts
- 11:02Speaker 2: or things that AI could use as well.
- 11:05Speaker 1: Definitely, I think that's really important, and that's been a
- 11:07Speaker 1: narrative that's been growing since around the time of the
- 11:12Speaker 1: x Yo Layer one network launch, there was a shift
- 11:16Speaker 1: in the narrative towards more privacy in Ethereum and like
- 11:21Speaker 1: Midnight Network we just had Charles Hoskinson on. But it's
- 11:23Speaker 1: more of a compliant privacy. It's not like a monarow privacy.
- 11:28Speaker 1: It's like privacy that for the things that are important
- 11:32Speaker 1: you can keep private or you can use zero knowledge
- 11:35Speaker 1: proofs to prove something, but you still are tying your identity.
- 11:40Speaker 1: There's KYC there. You're not going wild West against the
- 11:44Speaker 1: government or something. But having the perfect amount of privacy
- 11:48Speaker 1: where it's needed optionally, I feel like can be important,
- 11:52Speaker 1: especially if institutions are coming in and they're trying to
- 11:54Speaker 1: use verified data to make trades or to get a
- 11:57Speaker 1: leg up. They don't want all their information public, but
- 12:00Speaker 1: they don't they can't necessarily have all their information.
- 12:02Speaker 2: Private, correct, right, So a privacy for the stake of security,
- 12:07Speaker 2: where your passwords or your your key phrases, obviously you
- 12:11Speaker 2: want those sorts of things. There's also privacy from the
- 12:13Speaker 2: standpoint of you know, things people going to judge you on,
- 12:17Speaker 2: or there's social impacts from privacy in many cases. But
- 12:21Speaker 2: when it comes to things like like accounting, for example,
- 12:24Speaker 2: transparent accounting is why they see as filing requirements for
- 12:27Speaker 2: public companies and so on and so forth to have
- 12:29Speaker 2: as much transparency as possible. So to me that a
- 12:32Speaker 2: lot people feel that this single bullet of privacy and
- 12:36Speaker 2: privacy equal security. It's like, in certain lights, privacy definitely
- 12:40Speaker 2: equal security, and in certain lights privacy actually is the
- 12:43Speaker 2: antithesis of security because it gives a shroud to somebody
- 12:46Speaker 2: who wants to do something.
- 12:47Speaker 3: That's not good or it's inappropriate.
- 12:50Speaker 2: So knowing well when privacy is actually security tool and
- 12:55Speaker 2: when it's actually a security fault is important, And I
- 12:57Speaker 2: think people don't often think about the fact that you
- 13:00Speaker 2: can have both sides of that edge.
- 13:04Speaker 1: Definitely, as blockchain overall integrates AI more deeper and x
- 13:09Speaker 1: Yo is more getting involved with AI, especially on the
- 13:13Speaker 1: data integrity side. It seems like there's a big discussion
- 13:17Speaker 1: around there's a bottleneck of AI, whether it's in the
- 13:20Speaker 1: models or the compute or the data integrity side. Where
- 13:24Speaker 1: do you land on that?
- 13:26Speaker 2: Well, I think one of the most important things with
- 13:28Speaker 2: AI is to learn that the tools around it. That now,
- 13:31Speaker 2: the tooling is what makes it more efficient. So for example,
- 13:33Speaker 2: you know, if you have a code base and you
- 13:34Speaker 2: to say a giant repdeing and say we'll go and
- 13:37Speaker 2: find every bug in here for being every missing STEMI
- 13:40Speaker 2: call it, it's gonna be a very intense AI exercise
- 13:43Speaker 2: where if you're say we'll just run yes Lind and
- 13:46Speaker 2: look at the results and go fix those things, or
- 13:48Speaker 2: even say yes, lint stash facts so it'll fix it
- 13:50Speaker 2: for you with minimal effort. And so the way to
- 13:53Speaker 2: make AI more efficient isn't necessarily to throw more h
- 13:56Speaker 2: two hundreds at it. Yeah, that's obviously one way to
- 13:58Speaker 2: give it more horsepower, but to give the tools so
- 14:00Speaker 2: it can be more efficient. So It's almost equivalent of
- 14:03Speaker 2: you take the caveman and you say, well, just run
- 14:06Speaker 2: the Caveman's stronger and faster and harder, versus giving them
- 14:10Speaker 2: a hammer or a saw is a tool to be
- 14:12Speaker 2: able to make things faster. So just because AI is
- 14:14Speaker 2: really smart, it doesn't mean that you can do things
- 14:17Speaker 2: as efficiently without tools. And so I think the tooling
- 14:19Speaker 2: around there and XYO definitely or an x X one
- 14:21Speaker 2: doesn't falls in that tooling category where it gives AI
- 14:25Speaker 2: the ability to do things in a much more efficient way.
- 14:28Speaker 2: Like there's especially in the next couple of years when
- 14:31Speaker 2: it gets to the point where there's nothing AI can't do.
- 14:34Speaker 2: It's just can't do it in an efficient way, and
- 14:37Speaker 2: it becomes even more of a case where you don't
- 14:39Speaker 2: want to run things in the cloud. So if I
- 14:40Speaker 2: want to say, well, I want to run my AI
- 14:42Speaker 2: model at home on a reasonable PC or reasonable set
- 14:46Speaker 2: up that doesn't cost fifty thousand dollars for a bunch
- 14:48Speaker 2: of GPUs, how do I do that? Well, potentially, if
- 14:52Speaker 2: you have the right tools and you just don't have
- 14:53Speaker 2: to use as main tokens, you know, I think there's
- 14:55Speaker 2: many iterations you can actually do that and I think
- 14:57Speaker 2: looking at optimists and some of these new robots are
- 15:01Speaker 2: coming out, they have to have onboard computers or on
- 15:04Speaker 2: board uh GPUs to be able to run their models
- 15:07Speaker 2: because otherwise, just as the Internet goes out to your
- 15:09Speaker 2: robot freezes and that would be kind of problem. So
- 15:12Speaker 2: there's another case there where how do you actually make
- 15:14Speaker 2: those you know, you use less tokens and you use
- 15:17Speaker 2: less horsepower. So it's kind of we've seen with the computers.
- 15:21Speaker 2: It's the one side is you you know, you make
- 15:24Speaker 2: them faster and faster. On the other side, you make
- 15:25Speaker 2: them more efficient so they don't use them with as
- 15:26Speaker 2: much battery for example. So I think the place where
- 15:31Speaker 2: there's a lot of opportunity for the non fang group,
- 15:34Speaker 2: as you called of AI is in that tooling, because
- 15:37Speaker 2: that tooling can be very specific to different use cases
- 15:39Speaker 2: or different domains, and it can really make AI more
- 15:42Speaker 2: efficient for people to use and even in some cases
- 15:45Speaker 2: easier for them to use, because many people who are
- 15:47Speaker 2: using AI don't understand how do you make skills or
- 15:50Speaker 2: how do you make make your AI more efficient to
- 15:53Speaker 2: just ask you to do things and you know and
- 15:56Speaker 2: expect it to be you know, self optimizing if that
- 15:58Speaker 2: makes sense.
- 16:00Speaker 1: Definitely, And it's a really interesting point about the efficiency.
- 16:04Speaker 1: Of course, we want AI to be as efficient as
- 16:06Speaker 1: possible to not have to use as much compute processing
- 16:09Speaker 1: power cost. But I feel like a lot of people
- 16:13Speaker 1: think that integrating blockchain might mean that it becomes less efficient,
- 16:18Speaker 1: at least when you compare traditional server systems, or people
- 16:22Speaker 1: are looking for storing things and they feel like they're
- 16:26Speaker 1: sacrificing by adding in a blockchain network or decentralization. Is
- 16:30Speaker 1: that not the case that it actually can make it
- 16:33Speaker 1: even more efficient.
- 16:35Speaker 3: Well, I would saying in generally you're right.
- 16:37Speaker 2: As far as anything which has cryptographic signing or hashes
- 16:42Speaker 2: and those sorts of things, there's all overhead that there's
- 16:44Speaker 2: a cost to that. That's one of the reasons why
- 16:45Speaker 2: we store a lot of our data off chain and
- 16:48Speaker 2: then link it on chain and we have the provenance
- 16:51Speaker 2: of the metadata on chain, because it would be an
- 16:53Speaker 2: efficient for us store a twenty gigabyte video, for example,
- 16:56Speaker 2: on chain, and so we want to minimize the amount
- 16:59Speaker 2: of times that we pay that price. You know, that
- 17:01Speaker 2: overhead for doing that. But I would definitely agree if
- 17:04Speaker 2: there's no need for something like provenance or permanence. Then
- 17:09Speaker 2: I wouldn't use blockchain for it necessarily to solve the problem.
- 17:11Speaker 2: So back to the if one hammer and everything looks
- 17:15Speaker 2: like a nail sort of a thing. If he said, well,
- 17:17Speaker 2: I want everything to work run on blockchain, that's not
- 17:19Speaker 2: necessarily a solution. The blockchain should be doing what it's
- 17:22Speaker 2: good at, and you know, use Web two or traditional
- 17:25Speaker 2: technologies for what it's good at. And so I would
- 17:27Speaker 2: not solve every problem with that AI has with blockchain necessarily.
- 17:31Speaker 2: But there are cases where having a shared ledger, for example,
- 17:35Speaker 2: makes a lot of sense. You know, when it comes
- 17:36Speaker 2: to money, for example with bitcoin, doing that without a
- 17:40Speaker 2: shared ledger in a decentralized or a way is not
- 17:42Speaker 2: really very practical, and so the benefits you get out
- 17:45Speaker 2: of it is worth that cost and that provenance and permanence.
- 17:50Speaker 2: It's definitely beneficial to have that. But at the same time,
- 17:53Speaker 2: if you were to ask AI to generate a SHO
- 17:56Speaker 2: fifty two hash or especially say a show to fifty
- 17:59Speaker 2: six hash, your code have way less efficient than just
- 18:02Speaker 2: having a SHAW library I actually run it for you
- 18:05Speaker 2: and get that. So that is an example of a
- 18:08Speaker 2: tool where having it just use a CLI tool that
- 18:11Speaker 2: does a hash for it is way more efficient than
- 18:13Speaker 2: it figuring out the hash itself, because it probably can
- 18:16Speaker 2: do I try. That kind of interesting is like, hey,
- 18:18Speaker 2: here's an hash algorithm, just you know, in your mind
- 18:21Speaker 2: without using external tools. See if you can actually generate
- 18:24Speaker 2: the hash I want. I bet it can, but it'll
- 18:27Speaker 2: probably be a very expensive test.
- 18:29Speaker 1: It's very interesting, but of course we always like it
- 18:33Speaker 1: when AI gets more efficient. It seems like the models
- 18:36Speaker 1: are coming out every month, it's like getting better and blockchains,
- 18:42Speaker 1: some of them haven't really improved their speeding efficiency over
- 18:47Speaker 1: many years. But the x Yo layer one, which you
- 18:51Speaker 1: launched less than a year ago now just had an
- 18:54Speaker 1: upgrade that improved the efficiency I read two to five times,
- 19:00Speaker 1: which is significant. Can you talk about those upgrades and
- 19:04Speaker 1: you know, was that with the help of AI or
- 19:06Speaker 1: what does that really enable?
- 19:09Speaker 2: Well, I definitely say it was with the help of
- 19:10Speaker 2: AI because we're using claud code for example, and codex
- 19:13Speaker 2: as tools. So the pace at which we can develop
- 19:16Speaker 2: and improve our blockchain is a lot faster because we're
- 19:19Speaker 2: using AI tools. I think everybody who's developing anything at
- 19:23Speaker 2: this point has to use those to keep up. So
- 19:26Speaker 2: we would have been able to make these improvements over
- 19:27Speaker 2: say six months, where it took us maybe a one
- 19:29Speaker 2: month to do now. So I'm really excited from the
- 19:31Speaker 2: standpoint that our ability to actually iterate and improve our
- 19:34Speaker 2: blockchain using AIS is really good. And there's a lot
- 19:37Speaker 2: of little things like, for example, one of the improvements
- 19:40Speaker 2: is you know, getting the current balance, for example, of
- 19:45Speaker 2: an account, and so when we do the verification and
- 19:47Speaker 2: the validation for that, you know, if you can get
- 19:49Speaker 2: that balance faster with indexing, then of course the validation
- 19:53Speaker 2: of the block that you're making goes faster, so you know,
- 19:55Speaker 2: it kind of flows up there. And so using AI
- 19:58Speaker 2: for example, also to do things like constant profiling where
- 20:02Speaker 2: you know, every once in a while besides like running
- 20:04Speaker 2: unit tests, which it can generate as well. You can say, well,
- 20:08Speaker 2: run the profile on this and make a blockchain with
- 20:12Speaker 2: our protocol with a million items on it, see how
- 20:15Speaker 2: fast it runs, and compare that with the previous version
- 20:18Speaker 2: and doesn't run faster, run slower. So being able to
- 20:21Speaker 2: do brute force checks like that, A building those things
- 20:25Speaker 2: and b having it go and frequently go and run
- 20:28Speaker 2: it and check it to see if it's there makes
- 20:30Speaker 2: it also a lot easier to both improve performance and
- 20:35Speaker 2: efficiency but also make sure we don't regress, because that's
- 20:38Speaker 2: always the worst is what you have improved something and
- 20:40Speaker 2: then you know, a build comes out that you know
- 20:42Speaker 2: it's ten percent slower, but you don't even realize it
- 20:44Speaker 2: because it's only ten percent. And then you know, Nobile
- 20:47Speaker 2: comes out and almost ten percent slower again, but you
- 20:49Speaker 2: don't really realize it.
- 20:49Speaker 3: And so on and so forth.
- 20:50Speaker 2: So so making sure we don't slide backwards is something
- 20:53Speaker 2: which AI really helps us with as well. And so
- 20:55Speaker 2: I'm really excited about the fact that our efficiency and
- 20:58Speaker 2: our stability and those things that we've really improved the
- 21:00Speaker 2: last three to four months and we've gotten all those
- 21:04Speaker 2: things out that the t's crossed and the ies dotted
- 21:07Speaker 2: in many ways.
- 21:08Speaker 1: Definitely, it's super exciting when you're able to take something
- 21:12Speaker 1: that would take six months and do it in a month.
- 21:15Speaker 1: AI is just amazing at that. And the fact that
- 21:20Speaker 1: the chain is more of a I don't know if
- 21:24Speaker 1: we're in generation three or four of these chains where
- 21:28Speaker 1: you're not sort of bloated from technology from twenty thirteen
- 21:31Speaker 1: like Ethereum is and having to change things, you know,
- 21:35Speaker 1: when you've already spent ten years is building the old
- 21:37Speaker 1: way I've done it in traditional SaaS, it's no fun
- 21:42Speaker 1: because you know, you're sort of built on old technology.
- 21:46Speaker 2: One of the things that's nice also about the way
- 21:47Speaker 2: that we develop our code is that since we try
- 21:50Speaker 2: to minimize how much actually goes on chain and what
- 21:52Speaker 2: really happens on chain, and a lot of the things
- 21:54Speaker 2: that are done are off chain, like for example indexing
- 21:57Speaker 2: and so etherere is kind of that way where like
- 21:59Speaker 2: there's certain things you can't get from Ethereum from the
- 22:02Speaker 2: normal client or a normal provider, like the same for example,
- 22:05Speaker 2: if you want to say, well, how many n fts
- 22:06Speaker 2: or what all the n FT is a person has
- 22:07Speaker 2: in their wallet, you can get that from ether scan
- 22:09Speaker 2: because ether scan goes in indexes the entire chain and
- 22:12Speaker 2: will know what that is. But there's no easy way
- 22:13Speaker 2: to get that from a provider, for example, because it's
- 22:15Speaker 2: not built into the provider architecture. And so we kind
- 22:19Speaker 2: of do the same thing where there's we try to
- 22:21Speaker 2: minimize how much the true provider actually provides and have
- 22:24Speaker 2: as much as possible that's provided by external sources, kind
- 22:27Speaker 2: of like an ether scan type thing where we have
- 22:29Speaker 2: our API where it can go index these things and
- 22:31Speaker 2: final for you. The key though is to make sure
- 22:33Speaker 2: that when it does provide that answer, you can still
- 22:34Speaker 2: go verify it. And it's kind of how like the
- 22:36Speaker 2: NFT thing works for ether scan, where you'll say, well,
- 22:39Speaker 2: you know, here's the twenty different things that this person has.
- 22:42Speaker 2: You can go and check it those are actually true
- 22:44Speaker 2: if they actually have them in their account pretty easily.
- 22:46Speaker 2: The one thing that's hard there is well do they
- 22:48Speaker 2: maybe have thirty? You know it is a full complete answer.
- 22:51Speaker 2: It's hard to verify that with a provider, but you
- 22:53Speaker 2: can at least verify the positive. So our system is
- 22:57Speaker 2: very solilarly where we try to push as much because
- 23:00Speaker 2: you can off of chain, and when it's off chain,
- 23:02Speaker 2: you can iterate on the index thing or the performance
- 23:04Speaker 2: of those things as much as you want to, because
- 23:06Speaker 2: as long as the truth is the truth and you
- 23:08Speaker 2: can actually prove that that's correct, getting that answer faster
- 23:12Speaker 2: is something which you don't have to worry about altering
- 23:14Speaker 2: the primary protocol every single time.
- 23:16Speaker 3: That you did.
- 23:18Speaker 1: Yeah it is. It's nice to not have to put
- 23:22Speaker 1: everything on chain, and there are like on chain maxis
- 23:25Speaker 1: where it's like we need everything there. But when you're
- 23:28Speaker 1: intersecting with AI and you're especially dealing with real world data,
- 23:33Speaker 1: you know, whether it's from the weather station example, is good.
- 23:37Speaker 1: You know that it's hard to have the source at
- 23:40Speaker 1: some point that information has to come from off chain
- 23:42Speaker 1: to on chain and maybe not necessarily every single thing
- 23:46Speaker 1: needs to come because it bloats it down and it
- 23:49Speaker 1: isn't as efficient.
- 23:51Speaker 2: Yeah, and also you have to distill things like I
- 23:53Speaker 2: always used video as an example, where if you have
- 23:55Speaker 2: a car, for example, with a video camera that uses
- 23:57Speaker 2: for auto driving, you can't just stream that ENTIREDEO over
- 24:00Speaker 2: the internet to a server somewhere and say, hey, AYI
- 24:02Speaker 2: you know, here's a live stream of my ten ADP
- 24:05Speaker 2: camera in the front of my car.
- 24:06Speaker 3: Drive me.
- 24:07Speaker 2: What do you have to do is you have to
- 24:08Speaker 2: have an edge device there of some sort which takes
- 24:10Speaker 2: that translates into little boxes and arrows and metadata, which is,
- 24:14Speaker 2: you know, it finds the different items on there, it
- 24:17Speaker 2: finds people, finds all the cars, and then it actually
- 24:19Speaker 2: drives using that metadata instead. So distilling and reducing data
- 24:23Speaker 2: at the edge or outside of the system is something
- 24:26Speaker 2: which is very important. But you still want to have
- 24:27Speaker 2: a hash potentially of that the frames that video where
- 24:30Speaker 2: it came from. So if you want to go back
- 24:31Speaker 2: and say, well, you know, do this metadata get correct
- 24:34Speaker 2: correctly generated, we can go and verify it by potentially
- 24:38Speaker 2: pulling that video off of the hard drivers in the
- 24:41Speaker 2: car for example, and comparing it so you can still
- 24:44Speaker 2: have the provenance of the original video without having the
- 24:46Speaker 2: video go all the way upstream. And so I think
- 24:49Speaker 2: data reduction is something which is going to be a
- 24:51Speaker 2: big thing and big challenge also ongoing basis for AI
- 24:55Speaker 2: and Xyo and xby one is very good at that
- 24:57Speaker 2: where you keep the provenance of the original data and
- 25:00Speaker 2: the source data, but you also have the ability to
- 25:02Speaker 2: reduce it into useful forms that are very efficient.
- 25:06Speaker 1: Definitely, And one of the first types of data back
- 25:10Speaker 1: when I first spoke to Xyo in twenty eighteen was
- 25:13Speaker 1: location data. And now we're talking about the vehicles, the
- 25:17Speaker 1: rise in autonomous vehicles and robot vehicles, and as soon
- 25:21Speaker 1: they're going to have wallets to pay with crypto as well.
- 25:23Speaker 1: I'm sure is there a special focus on autonomous vehicles
- 25:29Speaker 1: or just physical AI with x yo. As AI continues
- 25:34Speaker 1: to grow and move more into the real world.
- 25:37Speaker 3: Well, we've always been an oracle.
- 25:38Speaker 2: So onboarding data onto our chain and into our data
- 25:41Speaker 2: leaks is one of the important things. And so you
- 25:43Speaker 2: can either onboard it from a website or something virtual
- 25:47Speaker 2: or digital, but you're also can to onboard from sensors
- 25:49Speaker 2: or from cameras or from IoT devices, and so there's
- 25:55Speaker 2: tons and tons of data. I think the IoT ones
- 25:57Speaker 2: are more interesting in many cases because that reduction is ness,
- 26:00Speaker 2: because you can't just you know, you can't store a
- 26:02Speaker 2: snapshot of the entire data set of reality constant basis.
- 26:06Speaker 2: It's just there's just way too much data there. And
- 26:08Speaker 2: so how do you actually take make senses eyes or
- 26:12Speaker 2: ears or or whatever is these robots or these cars need.
- 26:15Speaker 2: How do how do you take that reduce it to
- 26:17Speaker 2: the important parts that you need to use to be
- 26:19Speaker 2: able to steer it and store that somewhere. That's why
- 26:23Speaker 2: to me, io T and the edge devices are are
- 26:27Speaker 2: very interesting, and I think an important part of x
- 26:29Speaker 2: YO and XL one going forward and location is important,
- 26:32Speaker 2: but times also very important thing. People also forget that
- 26:35Speaker 2: time is actually a component of location. So because it's
- 26:38Speaker 2: x y z T really is what we live in
- 26:39Speaker 2: we live in those four dimensions. The only difference between
- 26:41Speaker 2: time and the other three is that time marches in
- 26:44Speaker 2: a certain direction at a certain speed and you can't
- 26:46Speaker 2: change it. Your movement in the other x, y Z
- 26:50Speaker 2: coordinates can change, you know, the velossity can change, you
- 26:52Speaker 2: can move different speeds or you also you can return
- 26:55Speaker 2: to the same spot, but you can never return to
- 26:56Speaker 2: the same spot.
- 26:57Speaker 3: And T.
- 26:58Speaker 2: But T to me is not fourth dimension as far
- 27:01Speaker 2: as like, it's only different. It's just really the fourth
- 27:03Speaker 2: dimension of location. But it's it's an uncontrollable dimension of location,
- 27:09Speaker 2: which is interesting. So we we think times are a
- 27:12Speaker 2: very important part of our data storage as well, and
- 27:15Speaker 2: we include that in the bucket of location.
- 27:18Speaker 3: Mmmm. That's very cool.
- 27:20Speaker 1: I love that. And you know, with blockchain, from the beginning,
- 27:25Speaker 1: time has been a major piece of that, you know,
- 27:29Speaker 1: understanding not just the chain of transactions. From the beginning
- 27:33Speaker 1: of time, there's times right right in there. It's all
- 27:36Speaker 1: time stamped on and that can't be changed when when
- 27:39Speaker 1: it's on chain.
- 27:41Speaker 2: Yeah, blockchains are they themselves are time vectors if you
- 27:44Speaker 2: think about it, because basically, you know, you may not
- 27:46Speaker 2: know how how long compared to like physical time those were,
- 27:49Speaker 2: but there there's uni directional. So you can't go back
- 27:52Speaker 2: in time in the blockchain. That's whole point of it.
- 27:54Speaker 2: I can't go and say, well'm gonna go change block
- 27:55Speaker 2: four on block on the Bitcoin blockchain.
- 27:58Speaker 3: That's not practical at all.
- 28:00Speaker 2: In theory, you can if you have enough hash power,
- 28:02Speaker 2: but that's kind of like you know, theory can go
- 28:03Speaker 2: back in time also if you have a big enough
- 28:05Speaker 2: quantum computer. Perhaps, but it's not practical at all. And
- 28:08Speaker 2: so time is uh. It's always been a fascinating thing
- 28:10Speaker 2: for me because it provides security, it provides permanence, and
- 28:15Speaker 2: not being able to manipulate time is what makes time interesting.
- 28:17Speaker 2: If time was manipulable, manipulatable, just like x Y and
- 28:21Speaker 2: ZI was, then the whole world would be you think about.
- 28:24Speaker 3: It differently, very cool.
- 28:28Speaker 1: And with the new upgrade to the x YO layer
- 28:32Speaker 1: one network, does that affect the efficiency of the xl
- 28:37Speaker 1: one governance and how the token works within the ecosystem?
- 28:44Speaker 2: Well, if anything, what ends up happening is we can
- 28:46Speaker 2: make more blocks a faster paced so people who are
- 28:49Speaker 2: staking xyo, because the way our system works is you
- 28:52Speaker 2: stake xy O and then you receive xl one as
- 28:54Speaker 2: the reward as being one of the stakers of the
- 28:56Speaker 2: network and that sort of thing. So you'll in theory
- 29:01Speaker 2: or receive more xcel one for that because more blocks
- 29:04Speaker 2: being made, but also more excel one gets burned because
- 29:07Speaker 2: the xcel one is used for gas. It gets burned
- 29:11Speaker 2: as part of the components that are there, so it
- 29:14Speaker 2: does accelerate the time frame basically of our Our goal
- 29:18Speaker 2: is to get to the point where we because right
- 29:22Speaker 2: now we're a slightly inflationary currency because we get block
- 29:25Speaker 2: rewards kind of same way that bitcoin does r at
- 29:27Speaker 2: their investment, and the goal of course for all blockchains
- 29:31Speaker 2: is to get to the point where your consumption of
- 29:33Speaker 2: of tokens equals or is greater than the production of that,
- 29:37Speaker 2: so you you kind of switch from being the fact
- 29:41Speaker 2: of inflationary currency to a deflationary currency, even though you
- 29:44Speaker 2: kind of still you're always inflationary because you're always producing
- 29:47Speaker 2: slightly more of the token. But if you as long
- 29:50Speaker 2: as you burn burn faster than you produce, you're a
- 29:53Speaker 2: net negative as opposed a net positive growth token.
- 29:57Speaker 3: So our goal and our path to.
- 29:59Speaker 2: That think is getting shorter and faster because of the
- 30:02Speaker 2: fact that our efficiency and our usage is going up definitely.
- 30:05Speaker 1: And with having more transactions and blocks, is it more
- 30:15Speaker 1: if different IoT devices and physical AI coming into the network,
- 30:19Speaker 1: or the same amount of devices being able to compute
- 30:23Speaker 1: more things and store more more efficiently, or combination of both.
- 30:28Speaker 2: I think it's a combination of both. But what if anything,
- 30:32Speaker 2: we have probably more partners who want to put things
- 30:35Speaker 2: on there at a faster rate, and so in some cases,
- 30:38Speaker 2: what it is, it's your sampling size changes. So you know,
- 30:41Speaker 2: if I say, for example, I have my private data
- 30:43Speaker 2: Lake and I'm storing you know, a thousand items a second,
- 30:45Speaker 2: but I only want to go and roll that up
- 30:47Speaker 2: and hash it and put it on the main chain,
- 30:50Speaker 2: you know, once every minute or once every ten minutes,
- 30:54Speaker 2: I could not put it on there once or once
- 30:56Speaker 2: every ten seconds, and that sort of thing. So basically,
- 30:58Speaker 2: your your resolution or your your frame rate to a
- 31:02Speaker 2: large degree is getting getting faster, which is good for
- 31:05Speaker 2: them because they can have a a better frame rate
- 31:09Speaker 2: as far as being able to have permanence to this
- 31:10Speaker 2: on there, and they're can have more data that story
- 31:12Speaker 2: on there because the amount of of hashes that can
- 31:14Speaker 2: put inside of one transaction gets bigger. So I don't
- 31:18Speaker 2: think if we've ever had a shortage of data that
- 31:21Speaker 2: we want to have on there. It's just how much
- 31:24Speaker 2: of reality you want to capture and put on a blockchain.
- 31:26Speaker 2: And if we can capture more and more, you can
- 31:28Speaker 2: analyze more and you can do more, and you can
- 31:30Speaker 2: you know, and learn more. And one of the things
- 31:32Speaker 2: work then also is like for people to capture their
- 31:34Speaker 2: own queries and responses for AI. And if you can
- 31:38Speaker 2: do that, then you can actually use those to build
- 31:40Speaker 2: contexts and train your own models, and that's sort of thing,
- 31:44Speaker 2: but that becomes very voluminous as well. So you want
- 31:46Speaker 2: to have these data sets of your your own data,
- 31:50Speaker 2: so you can propagate those into a learning feedback loop
- 31:55Speaker 2: basically for your AI.
- 31:57Speaker 3: That's very cool.
- 31:58Speaker 1: Are there developers that are working on the layer one
- 32:04Speaker 1: right now or x Yo you know, in the combination
- 32:07Speaker 1: with AI to be able to do things like that
- 32:09Speaker 1: to build their own models or to capture more data,
- 32:13Speaker 1: Like can anybody that's a developer go to x Yo
- 32:15Speaker 1: right now and start tinkering with that?
- 32:18Speaker 2: Well, you can definitely. Well, the simplest thing to do
- 32:20Speaker 2: is if you just install our wallet for example and browser,
- 32:23Speaker 2: then kind of like Ethereum. You can go and grab
- 32:25Speaker 2: the gateway that it injects into the page and you
- 32:28Speaker 2: can connect to that right there. We also have an
- 32:31Speaker 2: STK where you can do it like from note for example,
- 32:33Speaker 2: or from other things, or if a person's really industrious,
- 32:37Speaker 2: if they want to do from State Go or those things,
- 32:39Speaker 2: they can just call our the RPC end points for
- 32:43Speaker 2: submitting transactions. But we're working on stks to be able
- 32:46Speaker 2: to make it easier for people to be able to
- 32:50Speaker 2: to roll the transactions and submit transactions from other languages
- 32:55Speaker 2: besides JavaScript at this point. But again, with AI really
- 32:59Speaker 2: helps us with that because to say, hey, we have
- 33:01Speaker 2: a really perfectly working jobs cond version of this, make
- 33:03Speaker 2: me a complan version of this, or make me a
- 33:05Speaker 2: GO version of this is pretty easy with AI, Wherever
- 33:09Speaker 2: the human going through and porting to a different language
- 33:12Speaker 2: is a lot harder. And so I think the ability
- 33:14Speaker 2: for us to actually make unlimited SDKs that all work
- 33:18Speaker 2: as well as the primary ones we've developed is going
- 33:22Speaker 2: to be much more rapid than it would have been
- 33:25Speaker 2: even a year ago.
- 33:26Speaker 1: Definitely. Well, it's great to see the updates happening quickly
- 33:32Speaker 1: with x Yo and you guys are still building strong,
- 33:36Speaker 1: whether it's bull market or bear in the crypto industry,
- 33:38Speaker 1: which is a small piece. You know, AI and physical
- 33:42Speaker 1: physical AI and IoT is going to encompass the entire
- 33:45Speaker 1: world outside of blockchain. Looking forward, how do you see
- 33:52Speaker 1: the pace of these different devices moving on chain and
- 33:57Speaker 1: into x yo compared to the rate that's happened so far.
- 34:02Speaker 2: Well, I think everything in the world's going to be
- 34:03Speaker 2: moving faster. I think it's an extremely exciting time to
- 34:06Speaker 2: be alive as far as developing software and developing technology
- 34:09Speaker 2: and products. It's also a scary time because you know,
- 34:11Speaker 2: like sometimes I'll be a wake up in the morning,
- 34:13Speaker 2: Hey have this really cool idea write down. I'm like,
- 34:16Speaker 2: maybe we should do this, and then a week later
- 34:18Speaker 2: somebody else strips it. So it's like the the race
- 34:21Speaker 2: to be the first, first develop or deliverable of that
- 34:25Speaker 2: that product is really hard now because everyone has this
- 34:28Speaker 2: ability to develop things fast. So knowing what's novel and
- 34:32Speaker 2: knowing what's interesting and building tools that are are slightly
- 34:35Speaker 2: better perhaps than somebody else is because of usability and
- 34:38Speaker 2: those sorts of things are becoming more and more important.
- 34:41Speaker 3: But I think.
- 34:41Speaker 2: Everything's gonna move a lot faster once we have I
- 34:44Speaker 2: think Optimist is supposed to strip later this year. They
- 34:46Speaker 2: are their first robots right now, there's like a few
- 34:48Speaker 2: Chinese companies that can actually order robots from the point where,
- 34:52Speaker 2: you know, if people have robots in their homes and
- 34:54Speaker 2: they want to store the data from that, they want
- 34:55Speaker 2: to have an audit log also where if something bad happens,
- 34:58Speaker 2: they can go back and look at what do you
- 35:01Speaker 2: want to know what they did? Because you never know
- 35:03Speaker 2: like why did they do this or why did they
- 35:07Speaker 2: do that? So I think that the need for storing
- 35:10Speaker 2: providenced data on online or in a blockchain is going
- 35:13Speaker 2: to just explode over the next two to three years.
- 35:16Speaker 2: With things like, you know, any sort of self self
- 35:20Speaker 2: motivated items like robots or cars or any of those
- 35:23Speaker 2: things where people need something to help them as a
- 35:27Speaker 2: human have some sort off I would say, defense against that,
- 35:30Speaker 2: but some sort of of layer of understanding and comfort
- 35:36Speaker 2: of using those because for a lot of people it's
- 35:38Speaker 2: very uncomfortable to think that this thing you can do things,
- 35:40Speaker 2: it doesn't even you know, you don't even understand how
- 35:42Speaker 2: it does it.
- 35:44Speaker 1: Yeah, it's going to be a crazy world. I would
- 35:47Speaker 1: love to see my car being driven in sync with
- 35:50Speaker 1: x Yo or the robots in my house. When when
- 35:54Speaker 1: they get better, which you know, we've seen some of
- 35:56Speaker 1: the early videos and it's not just remote controlled by
- 36:00Speaker 1: by a person, but quickly it's going to be getting
- 36:02Speaker 1: really good, and I'm sure there'll be the top surgeons
- 36:04Speaker 1: in the world and top janitors and everything have that
- 36:07Speaker 1: all on chain is i think every blockchain fanboys dream.
- 36:12Speaker 1: So I'm excited to see that and wishing you and
- 36:16Speaker 1: the team the best on playing a big part and
- 36:18Speaker 1: making that happen.
- 36:20Speaker 2: You know, thank you very much, thanks for having me
- 36:22Speaker 2: on today, and I'm very excited about the space and
- 36:25Speaker 2: it continues to be a great journey.
- 36:27Speaker 1: Yeah, thank you so much.
- 36:28Speaker 2: Ri