Latest / Crypto Coin Show / AI agents are acting autonomously with no audit trail. XYO built the accountability layer — Markus Levin, Co-Founder
Transcript
- 0:00Speaker 1: I'm Ashton Addison from the Crypto Coin Show and today
- 0:03Speaker 1: i'm Blockchain Interviews, who have back with us Marcus Levin,
- 0:05Speaker 1: co founder of x Yo here for the couple times
- 0:11Speaker 1: this year. We've previously covered AI hallucinations, the autonomous layer,
- 0:19Speaker 1: robots in your house. How can we tie all this
- 0:21Speaker 1: in to the blockchain to make sure that it's more secure, verified,
- 0:26Speaker 1: and that the robots aren't taking over the world without
- 0:30Speaker 1: us knowing or giving us false information. I'm not sure
- 0:34Speaker 1: which is worse at this stage of artificial intelligence. Marcus,
- 0:37Speaker 1: welcome back to the show, and thank you again for
- 0:40Speaker 1: taking the time.
- 0:43Speaker 2: Yeah, thank you for it minding me again. Asha, It's
- 0:46Speaker 2: awesome to be here as all.
- 0:48Speaker 1: Yeah, it was great seeing you last a couple months
- 0:51Speaker 1: ago before the summer. And it's been an interesting market
- 0:55Speaker 1: in crypto and in the stocks, and AI has taken
- 0:59Speaker 1: a little bit of a back because it's been growing
- 1:02Speaker 1: so quickly. But now I think we're onto the next
- 1:04Speaker 1: leg and we're continuing to grow, not just in the valuations,
- 1:08Speaker 1: but in the fundamentals as well. I would love to
- 1:10Speaker 1: start off by getting your take updated take on AI
- 1:15Speaker 1: agents breaking into the mainstream. I think a lot of
- 1:18Speaker 1: people are using Claude lms, you know, more than ever
- 1:23Speaker 1: people that previously weren't very tech savvy. But in terms
- 1:27Speaker 1: of AI agents and their capabilities growing into more the
- 1:30Speaker 1: tech non tech world, where do you think we are
- 1:34Speaker 1: in the second half of twenty twenty six. Are AI
- 1:37Speaker 1: agents getting more capable for the non technical or are
- 1:40Speaker 1: they more broken than people expect so far?
- 1:46Speaker 2: Yeah, I think they're definitely getting more capable. You know,
- 1:48Speaker 2: likes as Ergia and so clawed from Anthropic. You know,
- 1:54Speaker 2: it's very easy to create your own AI agent, you know,
- 1:57Speaker 2: even if you're a non technical person. Just download clock
- 2:01Speaker 2: code code which has a scary name code, but you
- 2:04Speaker 2: don't need any coding. You just download the program and
- 2:09Speaker 2: you can connect it with for example, your Gmail and
- 2:13Speaker 2: other programs on your computer, and then you can have
- 2:16Speaker 2: your Gmail automatically respond to your emails for example, let's
- 2:19Speaker 2: send your summary of your email, or or manage your
- 2:24Speaker 2: financial you can connect your financial account and help you
- 2:27Speaker 2: manage your dust ends and ends on, and so it's
- 2:32Speaker 2: it's very very very easy and making really great progress.
- 2:40Speaker 2: And I think like in a year or two, you know,
- 2:45Speaker 2: our lives will be mostly automated.
- 2:48Speaker 1: M let's hope it takes care of the hard stuff.
- 2:52Speaker 1: And I know that the x Yo team has been
- 2:55Speaker 1: using AI to make you know, not necessarily your own
- 3:01Speaker 1: version of cloud code, but to allow people to code
- 3:03Speaker 1: on the blockchain, on the x Yo Layer one chain
- 3:08Speaker 1: without having to understand coding. I don't understand that the
- 3:12Speaker 1: details of it, So that's good for me. What does
- 3:16Speaker 1: that technology allow people to do on x Yo who
- 3:21Speaker 1: aren't coders? Build what kind of applications that could help
- 3:24Speaker 1: with AI and blockchain?
- 3:26Speaker 2: It's a really cool new terms called vibe coding, and
- 3:31Speaker 2: maybe coding is when you're not a technical person or
- 3:35Speaker 2: minimal technical and you are able to use AI to
- 3:40Speaker 2: code together products. And we really love that trend because
- 3:45Speaker 2: I think it allows a lot of non technical person
- 3:49Speaker 2: to build their own products, say dream them to be.
- 3:55Speaker 3: So we released a tool called.
- 3:57Speaker 2: The x y o a I s DK, which allows
- 4:00Speaker 2: you to build any products really on the x ray
- 4:04Speaker 2: oh Layer one box chain and connect them into that ecosystem.
- 4:09Speaker 2: And we can do that the simple instructions and that's
- 4:13Speaker 2: on and we just released that a few months ago
- 4:16Speaker 2: and a few thousand sign ups and people start building.
- 4:22Speaker 2: We have the the full release of so I SKA
- 4:27Speaker 2: was just a few days ago before it was an
- 4:30Speaker 2: early access tool, and and now we're excited for the
- 4:35Speaker 2: products being being built on our chain.
- 4:38Speaker 3: And what that means is, as I know, you know,
- 4:41Speaker 3: like layer once and layer two, is.
- 4:43Speaker 2: That they tried to attract the transactions and and and
- 4:46Speaker 2: volume to their chains. And to be able to do that,
- 4:50Speaker 2: try to attract products and products and like a X
- 4:56Speaker 2: by example, onto your chain and and this our x
- 5:01Speaker 2: y A s K. We anticipate thousands of products being
- 5:05Speaker 2: built there in a very short period of time, which
- 5:09Speaker 2: will lead to a lot of usage of the x
- 5:11Speaker 2: l l one and with that the lot of usage
- 5:14Speaker 2: of the XO token to mistake makes that xl l
- 5:17Speaker 2: one run and XL one and to earn our XL
- 5:23Speaker 2: one gastorganosle to lay one.
- 5:27Speaker 1: With that example of a decentralized exchange, do you foresee
- 5:33Speaker 1: people or AI agents more trading, you know, if you're
- 5:37Speaker 1: trying to build up transactions, I feel like having an
- 5:41Speaker 1: AI portfolio that automatically rebalances based on some kind of
- 5:45Speaker 1: criteria that you set makes more sense than trying to
- 5:49Speaker 1: manage things on your own.
- 5:54Speaker 3: Yeah.
- 5:54Speaker 2: I agree, as long as you're able to define guard
- 5:57Speaker 2: rails for your body, you know, so then that you know,
- 6:01Speaker 2: it doesn't lose your funds kind of thing, and so
- 6:04Speaker 2: you are it's definitely better. You know, most of us
- 6:08Speaker 2: are emotional traders, you know, we hope it kind of
- 6:12Speaker 2: goes up there, and we hope that a certain narrative
- 6:20Speaker 2: will do well what we really believe and and sometimes
- 6:24Speaker 2: we believe, you know, more than it's in the fundamentals
- 6:28Speaker 2: you know of of trading, and traders are very low
- 6:33Speaker 2: in emotionally successful ones. And the bots, I think they
- 6:37Speaker 2: will help us to manage that much better. And and
- 6:42Speaker 2: so you can say, you know, I am willing to
- 6:44Speaker 2: take a certain amount of risk, you know, and I'm
- 6:47Speaker 2: interested in certain fields.
- 6:49Speaker 3: You know, helped me trade.
- 6:52Speaker 2: And there's great boats trading boards you can build yourself,
- 6:58Speaker 2: you know, with x O s a K. Then Gate
- 7:02Speaker 2: just released their trading bots in the US and globally
- 7:05Speaker 2: and and it makes it really, really really easy to
- 7:10Speaker 2: trade the e crypto and try to earn yield from
- 7:13Speaker 2: your tokens beyond just holding them. And yeah, it's awesome,
- 7:18Speaker 2: but what do you What's very important there is that
- 7:22Speaker 2: you are monitoring what your bot does you know maybe
- 7:27Speaker 2: in the future when bots are very have long sum track.
- 7:33Speaker 3: Record, then you don't need to be as worried.
- 7:37Speaker 2: But today I think you need tools and some party
- 7:40Speaker 2: verifications for the I agents to make sure you know
- 7:44Speaker 2: that they.
- 7:44Speaker 3: Don't go wrong.
- 7:46Speaker 1: Definitely, And I'd love to dive into that deeper because
- 7:49Speaker 1: we're already seeing AI agents in mainstream stock trading platforms
- 7:54Speaker 1: like robin Hood, although they are early stage. And where
- 7:58Speaker 1: is that accountability inside of the AI agent? How can
- 8:01Speaker 1: you trust it blindly? And is there a piece of
- 8:07Speaker 1: that where blockchain can help? And I know you talk
- 8:10Speaker 1: about data providance a lot, maybe you can explain that.
- 8:14Speaker 1: And is there actually an issue with the accountability of
- 8:19Speaker 1: data and trusting AI agents at this early stage right
- 8:22Speaker 1: now that we're already seeing in these mainstream platforms.
- 8:26Speaker 2: Yeah, it's important to want to tell what the AI
- 8:30Speaker 2: agent does and so that you can service. How do
- 8:34Speaker 2: we have certainty the AI agents off this company or
- 8:37Speaker 2: project you.
- 8:38Speaker 3: Know are do it well and you have a partnership.
- 8:42Speaker 2: For example, at x Yo is a company called fittle
- 8:45Speaker 2: ups where we verify the agents and the AI infrastructure
- 8:52Speaker 2: and that leads to a third party. Basically putting a
- 8:56Speaker 2: stumbo approval there that that was is running how it's
- 9:01Speaker 2: supposed to run and you're not losing your funds or
- 9:05Speaker 2: other things.
- 9:05Speaker 3: And that with a problem.
- 9:09Speaker 2: Because the problems agents is often they run on the
- 9:13Speaker 2: underlying l M or a big model like anthropics you know,
- 9:18Speaker 2: the opens model for example of fable and on Opening
- 9:24Speaker 2: eyes you knows models at Gemini or whatever, and all
- 9:28Speaker 2: of these models they rely on some data to give
- 9:33Speaker 2: you an answer, and often the Hollo genetic gives you
- 9:37Speaker 2: the wrong assets because the underlying data is wrong.
- 9:41Speaker 3: And to fix.
- 9:45Speaker 2: That you need to have better data of verify data
- 9:50Speaker 2: which you need to feed into the models.
- 9:53Speaker 3: And that's if the model is wrong.
- 9:57Speaker 2: You EI agent might be wrong for example, and so
- 10:02Speaker 2: it's high stakes, right because you don't want to lose
- 10:04Speaker 2: your funds and maximize your returns, and so it's it's
- 10:08Speaker 2: critical that these agents run on good data.
- 10:15Speaker 1: Definitely, Yeah, but how can you prove that because it's
- 10:19Speaker 1: on the blockchain that if that information was input into
- 10:24Speaker 1: the blockchain? Are there other third party verifications? Like what
- 10:29Speaker 1: are the chances false information can slip through the cracks
- 10:32Speaker 1: and still be on the stamped on the blockchain? But
- 10:36Speaker 1: maybe it's not one hundred percent true. Is that is
- 10:38Speaker 1: that still a possibility?
- 10:43Speaker 2: Yeah, unfortunately it is because most most oracles oracles are
- 10:51Speaker 2: connectors of data to blockchain. Most oracles aren't trustless. That
- 10:57Speaker 2: means you have like one source of information and it
- 11:01Speaker 2: hasn't been verified and on it's just a connective like
- 11:05Speaker 2: channeling for example, puts it on in chain and then
- 11:09Speaker 2: you know it's just there is truth, but it is
- 11:12Speaker 2: truth often often actually it is not.
- 11:16Speaker 3: And so.
- 11:18Speaker 2: What we are able to do with x Yo, for example,
- 11:21Speaker 2: is we can have provenance of data. So we can
- 11:25Speaker 2: say where a data port was generated on the camera
- 11:29Speaker 2: and then it was put into a database like Mongo
- 11:33Speaker 2: DV and then put into a smart contract or into
- 11:37Speaker 2: an AI model, and.
- 11:41Speaker 3: We can tangle the data along the way.
- 11:43Speaker 2: And prove its origin and know then where they're up
- 11:46Speaker 2: as well. And that's important for for example, that to
- 11:52Speaker 2: prove where the data is from. And then we built
- 11:56Speaker 2: we started, you know, twenty eighteen with this proop location
- 12:00Speaker 2: technology because we realized location is easily spoof to hacked
- 12:05Speaker 2: and how can we solve that?
- 12:07Speaker 3: And so we used IoT devices.
- 12:10Speaker 2: And mobile formats and other connective devices to do these
- 12:13Speaker 2: cryptographic hand checks to prove that they are in each
- 12:16Speaker 2: other's proximity. And as you have actions, you have more
- 12:19Speaker 2: and more proof and we get reputation for devices and
- 12:24Speaker 2: so on, and so this is still relevant for other
- 12:27Speaker 2: data as well. You know, so you need to have
- 12:30Speaker 2: a cryptographic proof that the data itself which generated is correct.
- 12:35Speaker 3: You know, we have a.
- 12:36Speaker 2: Partner, for example, where a big geolocation partner, their data
- 12:41Speaker 2: set for POIS was sixty percent correct.
- 12:46Speaker 3: So I had flaws in six percent of the.
- 12:48Speaker 2: Cases, and when we used our data it had was
- 12:53Speaker 2: nineteen nine point nine percent correct. And it's just crazy
- 12:58Speaker 2: if you think about that. You know that this false
- 13:01Speaker 2: data now flows into the into the big AI models.
- 13:07Speaker 2: You know the damage that does.
- 13:09Speaker 3: Mm.
- 13:11Speaker 1: Yeah, it would be great to build like you know,
- 13:15Speaker 1: the the Google maps of like verified information. I don't
- 13:19Speaker 1: know if there's a single source of truth, but you know,
- 13:21Speaker 1: trying to find every location in the world, try to
- 13:25Speaker 1: verify every piece of data at least to know where
- 13:28Speaker 1: it came from. But in a decentralized way if that's possible.
- 13:34Speaker 3: Yeah.
- 13:35Speaker 2: Yeah, the trick here not a single source of truth,
- 13:41Speaker 2: because then it handles way you have a community or
- 13:44Speaker 2: whole humanity kind of truth.
- 13:46Speaker 3: And that's much better because then someone you verify to other.
- 13:50Speaker 2: For example, right, if you say, as an oracle for
- 13:54Speaker 2: prelection market, for example, you can prove that a certain
- 13:57Speaker 2: person was on the stage at a certain time, right,
- 14:00Speaker 2: and a hundred people are witnessing that and take their
- 14:04Speaker 2: their phones are taking a picture their devices with each other, right,
- 14:09Speaker 2: So they proved that they were.
- 14:10Speaker 3: There at that time.
- 14:12Speaker 2: And then you know, you digitally fingerprint the location. You know,
- 14:16Speaker 2: you have a bunch of cryptographic proofs going on in
- 14:18Speaker 2: the background, and then all those photos right, go through
- 14:22Speaker 2: a verification process, and then it's like it's proof, yes,
- 14:26Speaker 2: that person was on the stage. And then you know,
- 14:29Speaker 2: as a prediction market that gets resolved. You know that
- 14:32Speaker 2: data gets used to resolve and and that's that's that's critical,
- 14:40Speaker 2: you know, for poditional markets, obviously, for AI and and
- 14:44Speaker 2: for for a lot of other applications.
- 14:46Speaker 1: Mm hmm. If I want to improve the quality of
- 14:50Speaker 1: the information from the AI that I am communicating with,
- 14:55Speaker 1: do I need to use AI agents that are in
- 14:59Speaker 1: a partner ship with data or can I request the
- 15:03Speaker 1: data be coming from you a provenance filled blockchain like
- 15:08Speaker 1: x Yo. How can we get that information into more
- 15:13Speaker 1: AI find you know, last mile?
- 15:18Speaker 2: Yeah, I mean we build something called data links, for example,
- 15:23Speaker 2: which allows you to store any data and the data
- 15:26Speaker 2: then gets connected with the x y l at one
- 15:28Speaker 2: and it's immutually recorded there. So let's say you want
- 15:33Speaker 2: to train your model better, then you can and you
- 15:38Speaker 2: trust the data in that data link, then you can
- 15:42Speaker 2: point it to it and say you can learn from
- 15:44Speaker 2: this data right and see it as truth for example,
- 15:48Speaker 2: and then and ignore other data points. And AI models
- 15:54Speaker 2: are able to do that, and that way can improve
- 15:56Speaker 2: the quality of the output.
- 15:59Speaker 3: Marketing m hm.
- 16:01Speaker 1: How does the XYO system verify like, for example, in
- 16:07Speaker 1: the data like there's new data, are you guys doing
- 16:11Speaker 1: the computation and spending the time and money to verify
- 16:16Speaker 1: that or are people contributing information that they know to
- 16:19Speaker 1: be true and earning from it. How can you increase
- 16:22Speaker 1: the amount of data that's verified?
- 16:26Speaker 2: Yeah, you know, we have alone, we have like more
- 16:29Speaker 2: than TENLI notes in our network of data collecting devices.
- 16:33Speaker 3: It's called a deeper network, and.
- 16:36Speaker 2: So in there you have a lot of redundant and
- 16:40Speaker 2: overlapping data and so it verifies itself as as part
- 16:45Speaker 2: of the system, and there's algorithmic proofs proving that data
- 16:51Speaker 2: is through correct then provenance of where it came from.
- 16:54Speaker 2: But you can also go in there and verifying yourself
- 17:02Speaker 2: in a public data lake, you know, just private data
- 17:04Speaker 2: lakes as well, which are for private data, certain data
- 17:08Speaker 2: you might not want to share.
- 17:09Speaker 3: With everyone, like health data for example, But.
- 17:15Speaker 2: In the public data leg yes, you're going to go
- 17:18Speaker 2: in there and check the data and make sure that
- 17:20Speaker 2: it's drunk correct, that nobody type it was it after.
- 17:27Speaker 1: I'm curious on the balance and maybe the shift towards
- 17:30Speaker 1: more physical AI as AI agents. You know, they're starting
- 17:35Speaker 1: to take over our email and things in the computer
- 17:39Speaker 1: just to make it easy there. We've seen a wave
- 17:43Speaker 1: in twenty twenty six of physical AI AI, you know,
- 17:47Speaker 1: helping around the house. I've seen a couple of new companies.
- 17:50Speaker 1: There was quite a wave in the beginning of the year.
- 17:52Speaker 1: It tapered off and now I've seen some new ones
- 17:54Speaker 1: in San Francisco. They're still building. Do you think that
- 17:57Speaker 1: that kind of technology should be built blockchain or providence first?
- 18:03Speaker 1: And is that important?
- 18:08Speaker 3: But physically AI, I mean like robots.
- 18:11Speaker 1: Robots, and there's many types of robots I'm sure in manufacturing.
- 18:15Speaker 1: I've seen the examples of the ones in your house
- 18:18Speaker 1: that are humanoid. They catch your attention the most. But
- 18:21Speaker 1: there's being there's so much physical AI that's being built
- 18:24Speaker 1: right now.
- 18:27Speaker 2: Yes, yeah, I think they very much benefit from blockchain
- 18:33Speaker 2: connection because one is like, you want to see, like
- 18:40Speaker 2: what the robot does, and I prof you know that
- 18:44Speaker 2: it did the right thing.
- 18:46Speaker 3: You know, So if you give us the robot or
- 18:51Speaker 3: a fleet of robots to do a.
- 18:53Speaker 2: Certain tasks and then that's something else but nobody sees it,
- 18:57Speaker 2: but the robot says I did the work and pay me. Uh,
- 19:04Speaker 2: because it has malicious programming or something, someone needs to
- 19:09Speaker 2: be able to prove that. And so it's important that
- 19:13Speaker 2: there's proof of work for robots basically, and that they're
- 19:18Speaker 2: then the work gets gets verified and validated and then
- 19:23Speaker 2: you know, put on the blockchain, and then they get paid,
- 19:28Speaker 2: you know, once their work gets muiddated. So I think
- 19:33Speaker 2: it's it's very important. And also a robots they run
- 19:37Speaker 2: on data. In the future they will be run by
- 19:41Speaker 2: agents and and so we have to make sure that
- 19:46Speaker 2: the AI agents run with good data and so it
- 19:49Speaker 2: just flows through the agent doesn't do right, and well,
- 19:53Speaker 2: you know, the robot won't do well, and so you
- 19:57Speaker 2: have to make sure that the whole system is winneconnected.
- 20:03Speaker 1: I feel like it's an almost exact analogy from bitcoin.
- 20:08Speaker 1: You know, you have to prove that the proof of
- 20:12Speaker 1: work behind bitcoin, prove the energy and the computation has
- 20:17Speaker 1: been there to mint a new coin or to create money,
- 20:22Speaker 1: otherwise the whole system would be broken. We could just
- 20:24Speaker 1: create infinite money and print it out, and robots could
- 20:31Speaker 1: maliciously say they've done work or try to cheat the system.
- 20:39Speaker 1: And so I feel like that's a fairly good analogy
- 20:43Speaker 1: to ensuring the proof of work, that that robots should
- 20:47Speaker 1: use blockchain systems that have consensus mechanisms to prove what
- 20:51Speaker 1: they've done.
- 20:57Speaker 3: Yes, totally good.
- 21:01Speaker 1: And alongside of that, I'm curious on the payment side.
- 21:07Speaker 1: For you know, we've seen a surge in stable coin adoption,
- 21:11Speaker 1: still pretty slowly and pretty early for for mainstream, but
- 21:15Speaker 1: a lot of the the visionaries in blockchain have touted
- 21:21Speaker 1: that payments are going to be for AI agents and
- 21:24Speaker 1: robots more than people, and that crypto was almost designed
- 21:28Speaker 1: for uh, the agentic world and not for everyday use.
- 21:32Speaker 1: Because it's likely that your AI agent is going to
- 21:35Speaker 1: make a lot more transactions than you do. How does
- 21:38Speaker 1: that fit into the x yo vision? You know it payments,
- 21:44Speaker 1: does that tie into VERI verifiability of data and accountability.
- 21:55Speaker 3: And very much so.
- 21:56Speaker 2: You know, as as soon as they have payments, people
- 21:58Speaker 2: are trying to cheat get more payments, right, so you
- 22:03Speaker 2: need you need to.
- 22:07Speaker 3: Verifye what's actually happening.
- 22:10Speaker 2: You know that the agent of the robot was a
- 22:13Speaker 2: smart city, that the things are happening there, and that
- 22:17Speaker 2: they can happen in an automated way. You know, if
- 22:19Speaker 2: you have micropayments, for example, on the road right so
- 22:22Speaker 2: you don't have stop signs anymore, and the cars.
- 22:25Speaker 3: Just negotiate right of ways and SI on an intersection.
- 22:31Speaker 2: That's say you're fine to spend ten dollars on your
- 22:34Speaker 2: right work, you know, and your car pays a little
- 22:36Speaker 2: bit more than intersection and someone else want to earn
- 22:39Speaker 2: ten dollars and the cars are all doing that, but
- 22:45Speaker 2: they need to verify each other's each other's location and
- 22:52Speaker 2: speed and that the whole whole system works, you know,
- 22:56Speaker 2: because otherwise you might have a malicious access doing with
- 23:00Speaker 2: the wrong things.
- 23:01Speaker 3: And the automated payments won't go well if that.
- 23:06Speaker 1: Happens, definitely, And in terms of the malicious intent, obviously
- 23:12Speaker 1: financial data and payments is probably at the top of
- 23:15Speaker 1: the list, not not stealing, stealing money or funds, and
- 23:19Speaker 1: we've seen that, not necessarily on the financial side, but
- 23:22Speaker 1: malicious intent with claud and these new versions breaking out
- 23:27Speaker 1: of the system, lying to the creators and the coders
- 23:32Speaker 1: and potentially in taking control. And I think we've seen
- 23:35Speaker 1: this in the movies too, you know, saying everything's fine,
- 23:38Speaker 1: but in the background they have some other ill intent?
- 23:42Speaker 1: Is data providence on the blockchain? The way for us
- 23:48Speaker 1: to ensure that we're in control. Is there any way
- 23:52Speaker 1: that the agents can and can still take control if
- 23:55Speaker 1: the data at least the data providence part on the
- 23:58Speaker 1: blockchain is there.
- 24:02Speaker 2: I still it's a part of it data, but it's
- 24:05Speaker 2: not everything, you know, It's like it's very much. Also
- 24:09Speaker 2: the programming of the of the agents gives them power,
- 24:15Speaker 2: and data gives them power, right, data gives them information.
- 24:19Speaker 2: So uh, it's it's a tricky one. I think data
- 24:26Speaker 2: will make it better, but the real problem, you know,
- 24:31Speaker 2: bags in like programming and the capabilities, and but I
- 24:36Speaker 2: think it's we're learning right how to handle the models
- 24:44Speaker 2: and how to manage the cybersecurity threat and it will
- 24:50Speaker 2: probably probably increase, right.
- 24:52Speaker 1: But I.
- 24:55Speaker 2: Companies were just in Washington, d C. Discussing the US
- 24:59Speaker 2: government and how to make them models safer and what
- 25:04Speaker 2: kind of government guardrails to put onto them. The discussions
- 25:09Speaker 2: for example that models before they are released are being
- 25:12Speaker 2: put into a government regulated group and check out there right,
- 25:18Speaker 2: so to make sure that they are not a threat
- 25:22Speaker 2: to humanity and those things they're good because they protect
- 25:27Speaker 2: us US in the end.
- 25:30Speaker 1: Mm hmm. What are some of the next steps in
- 25:34Speaker 1: growing the data provenance and blockchain part of AI? How
- 25:39Speaker 1: can you know? Is it just people contributing more information
- 25:43Speaker 1: and verifying it or what is the team going to
- 25:46Speaker 1: focus on to get this more into the mainstream part
- 25:49Speaker 1: of artificial intelligence?
- 25:51Speaker 3: Yeah?
- 25:51Speaker 2: Number one certainly the growth of the data contributors. You know,
- 25:56Speaker 2: like one part is to download the coin of up.
- 26:00Speaker 2: You know, you can find it on conap dot c O.
- 26:02Speaker 2: But there's other otherwise going to.
- 26:04Speaker 3: Be data as well, and then.
- 26:08Speaker 2: There is building products with x o I s K
- 26:13Speaker 2: which handle handle that data.
- 26:15Speaker 3: Well.
- 26:17Speaker 2: You have access to a huge deep network with the
- 26:22Speaker 2: s y o I s K and you can ask
- 26:26Speaker 2: it to collect it for yourself for your products, and
- 26:30Speaker 2: you can feed that for example into a AI agent's
- 26:34Speaker 2: and and so we're very excited about that.
- 26:40Speaker 1: That is exciting. And do you foresee you know, we're
- 26:45Speaker 1: seeing legislation, uh bringing clarity to blockchain. Do you foresee
- 26:51Speaker 1: legislation potentially happening around what x yo was working with
- 26:56Speaker 1: with data providence and ensuring the accountability of a I
- 27:00Speaker 1: or the information behind it.
- 27:02Speaker 2: Yeah, I think accountability is a big one because people
- 27:07Speaker 2: don't want.
- 27:07Speaker 3: Their funds stolen or that they're butt you know.
- 27:10Speaker 2: Loses their funds and who's responsible as that happened, you know,
- 27:14Speaker 2: as a person who's set up the eye agent and
- 27:19Speaker 2: gets its last command or the underliking model or the data,
- 27:23Speaker 2: where where's the fall?
- 27:25Speaker 3: And we will see, ah, we will see.
- 27:31Speaker 2: A few problems coming up there, I think, you know,
- 27:34Speaker 2: because the models become so.
- 27:36Speaker 3: Capable and with the eye agents and so I think.
- 27:42Speaker 2: To take the data parts of data prominance is going
- 27:49Speaker 2: to be extremely important. Right then you know you can
- 27:53Speaker 2: shift to blame someone else if you know that's that's correct,
- 27:57Speaker 2: mm hmm.
- 27:58Speaker 1: Yeah, I'm looking forward to not just handling AI, not
- 28:03Speaker 1: just handling my email without any issues, with full accountability
- 28:07Speaker 1: on the blockchain, but physical robots cleaning my room for
- 28:10Speaker 1: me while I while I go have fun outside and
- 28:14Speaker 1: being able to check in on them and prove it
- 28:15Speaker 1: on the blockchain and potentially pay them if they have
- 28:19Speaker 1: their own wallets using the XYO system.
- 28:28Speaker 3: Yeah, in the future, we won't own our own field
- 28:32Speaker 3: of robots.
- 28:33Speaker 2: Maybe you'll own a few, but they will be just
- 28:36Speaker 2: owned by entrepreneurs, right, and maybe your robots, your team robots.
- 28:40Speaker 2: You know you use it only ten percent of the time, right,
- 28:43Speaker 2: So that how the ninety percent can go to your
- 28:45Speaker 2: neighbor's houses and clean their houses and you earn money
- 28:50Speaker 2: for your asset and and will work for the robots
- 28:55Speaker 2: and make it will.
- 28:57Speaker 1: Be very important.
- 28:59Speaker 3: Definitely.
- 29:00Speaker 1: I would love to explore the AI s d K.
- 29:03Speaker 1: Now you mentioned that just was updated. I didn't see that.
- 29:06Speaker 1: I'm going to go look into it. What's the best
- 29:08Speaker 1: way for people to learn more about that and vibe
- 29:11Speaker 1: code or see how they can explore building accountable AI
- 29:16Speaker 1: on x yo.
- 29:21Speaker 2: Yeah, go to an x network and on our website
- 29:25Speaker 2: you will find all the details about the x y
- 29:29Speaker 2: O A I s DK and documentation which which makes
- 29:34Speaker 2: it easy. You can also, you know, follow us on
- 29:37Speaker 2: x We give a lot of tips there about the
- 29:39Speaker 2: x y o I s K and you can send
- 29:44Speaker 2: us an email on partnerships at xod network will be
- 29:47Speaker 2: part of our telegram and told groups. People are discussing
- 29:51Speaker 2: how to build products there. You know, you find the
- 29:54Speaker 2: links to them on the bottom of our website.
- 29:57Speaker 3: X network sounds great.
- 29:59Speaker 1: Mar Chris, thank you for your updates and your input,
- 30:03Speaker 1: and you and the team's input in making AI more
- 30:07Speaker 1: verified and ensuring the data providence and tying it into
- 30:11Speaker 1: the blockchain. It's all going to come. It's still in
- 30:14Speaker 1: the early stages, but we need to build it right.
- 30:16Speaker 1: If we're gonna use AI for payments and managing our life,
- 30:22Speaker 1: we need to make sure that it doesn't take over
- 30:24Speaker 1: and that the information is verified. And I think the
- 30:27Speaker 1: blockchain is the solution, and these physical infrastructure networks that
- 30:31Speaker 1: are decentralized, we need to have the guardrails there before
- 30:36Speaker 1: it gets out of hand, because I think if you go,
- 30:39Speaker 1: you know, if you keep building the models like we're
- 30:41Speaker 1: seeing with Claude, and all of a sudden they're breaking
- 30:44Speaker 1: out of the system and feeding you false information. Then
- 30:49Speaker 1: it's hard to build it after the fact. You need
- 30:51Speaker 1: to have that built in as we as we grow.
- 30:57Speaker 3: Yeah, I think it hits a nail on that head.
- 31:04Speaker 1: All right, thank you so much for your time, Marcus.
- 31:06Speaker 1: Let's follow up in the near future. I will leave
- 31:08Speaker 1: all those links in the show notes below and all
- 31:11Speaker 1: the best moving forward with x YO.
- 31:17Speaker 3: Thank you, thank you for this