Latest / Crypto Coin Show / Ethereum and Solana weren't built for AI. 0G Labs co-founder explains WHY - Crypto Coin Show
Transcript
- 0:00Speaker 1: We're live at Consensus twenty twenty six in Miami with
- 0:04Speaker 1: Michael a, co founder of zero G Labs. The layer
- 0:07Speaker 1: one for AI as AI and grows deeper into blockchain,
- 0:12Speaker 1: or blockchain grows deeper into AI. I think AI surpassing
- 0:17Speaker 1: the scale of blockchain and just benefiting the whole world.
- 0:21Speaker 1: Right now, let's first of all, I would just love
- 0:25Speaker 1: to start with a high level on what do you
- 0:28Speaker 1: see as the best synergy of blockchain with AI, and
- 0:33Speaker 1: then we can dive into zero G in all the
- 0:34Speaker 1: latest updates.
- 0:35Speaker 2: Absolutely, So the best way to think about it is
- 0:38Speaker 2: that when you look at AI, you think about hyperscalers,
- 0:42Speaker 2: black box type of models. You get a bunch of
- 0:44Speaker 2: GPUs together in one location. Even you go thinking about
- 0:48Speaker 2: nuclear power plants to power all of these types of things.
- 0:51Speaker 2: You have no idea where the data came from, what
- 0:53Speaker 2: happened at the end. You just got an API and
- 0:55Speaker 2: an interface you interact with a model, and that's kind
- 0:58Speaker 2: of blackpox AI. Well, we're providing an alternative called decentralized AI,
- 1:04Speaker 2: effectively where everything's transparent, yet it can be private and
- 1:09Speaker 2: anybody can participate as part of that network.
- 1:11Speaker 3: It's a permissionless type of network.
- 1:13Speaker 2: You as a user can provide a graphics card, you
- 1:15Speaker 2: can participate in that network. It can be fully distributed,
- 1:19Speaker 2: so it could be many many kind of power sources
- 1:21Speaker 2: that are powering this type of thing. So it's very
- 1:24Speaker 2: a user owned type of future. And then more importantly,
- 1:28Speaker 2: there's also a truth layer element to AI. You can
- 1:31Speaker 2: keep AI accountable and add an AI safety level on
- 1:35Speaker 2: top of that as well, and so that's really the
- 1:38Speaker 2: big difference and what zoogy and decentralized AI provide.
- 1:42Speaker 3: To the aspens.
- 1:43Speaker 1: Definitely, it sounds like there's two parts to that. There's
- 1:47Speaker 1: how the AI is hosted, if you could say, and
- 1:50Speaker 1: then actually how you're interacting with AI and the data
- 1:53Speaker 1: behind it, which has been controversial with certain different you
- 1:58Speaker 1: know some of these that all agout of AI platforms
- 2:02Speaker 1: your data.
- 2:02Speaker 3: Where's your data going?
- 2:04Speaker 1: Yes, you're using it to your own advantage, but also
- 2:06Speaker 1: they're probably using it to their advantage.
- 2:09Speaker 2: Yeah. Absolutely, that's the issue with BLUCKBOXL. You can't individually
- 2:13Speaker 2: verify what happened. You may be able to click a
- 2:15Speaker 2: button to say like don't use my I for training
- 2:18Speaker 2: data or my data for training of AI, but how
- 2:20Speaker 2: can you verify that the moment it's on somebody else's server,
- 2:23Speaker 2: they can technically do whatever they want with that particular data,
- 2:26Speaker 2: and it's very hard to independently verify and say like, oh,
- 2:29Speaker 2: you didn't use it for training purposes. So if I
- 2:31Speaker 2: have very sensitive customer data, personally identifiable data, medical data,
- 2:36Speaker 2: how do I know.
- 2:36Speaker 3: That's not being used to train for example?
- 2:39Speaker 2: And so with the central IZEDI, you can keep things
- 2:41Speaker 2: completely private. For example, we use something called te verification
- 2:45Speaker 2: Trusted execution environments, and so there's a physical hardware device
- 2:49Speaker 2: onographics card that ensures that the privacies there, and then
- 2:53Speaker 2: once the computation has been done, there's an entry on
- 2:55Speaker 2: a blockchain that basically says, okay, this has been done
- 2:57Speaker 2: to these specifications, according to this level of privacy, and
- 3:02Speaker 2: so therefore you can independently verify that something has happened.
- 3:06Speaker 1: And if there are all of these benefits to decentralized
- 3:10Speaker 1: AI's the.
- 3:12Speaker 3: What's holding people back? Is it the user interface?
- 3:15Speaker 1: The friendliness of choosing that, you know, the convenience. It
- 3:19Speaker 1: seems like Claude pops up and it's just like hey, Ashton,
- 3:22Speaker 1: let's let's get to work. Is it that can more
- 3:26Speaker 1: much more convenient. That's stopping people from switching, or just
- 3:29Speaker 1: they just aren't aware.
- 3:31Speaker 3: There's a couple of things.
- 3:32Speaker 2: Awareness is certainly one one challenge, and that's something that
- 3:35Speaker 2: we hope to address as we grow. Open the eyes
- 3:38Speaker 2: spin around for I think going to be a decade
- 3:40Speaker 2: pretty soon, and it took them seven plus years before
- 3:43Speaker 2: they collection released something that made the world really take
- 3:47Speaker 2: AI like seriously at that point, and so we're only
- 3:50Speaker 2: it's our third year into the journey at this month.
- 3:53Speaker 2: May kind of twenty twenty three is when we got founded,
- 3:56Speaker 2: and so it's a lot of room to go. So
- 3:59Speaker 2: there's the attention element, but then there's a user experience
- 4:03Speaker 2: element to it as well. Block Chains are not easy
- 4:06Speaker 2: to use because you have to have a ballet interface,
- 4:09Speaker 2: you have to top it up with specific type of
- 4:11Speaker 2: new currency, and so with AI agents and various other
- 4:16Speaker 2: things like a payment layer, we're trying to abstract that
- 4:18Speaker 2: user experience, so it's significantly easier.
- 4:21Speaker 3: And then the final thing.
- 4:22Speaker 2: Is that open source models still lag about three months
- 4:26Speaker 2: behind the state of the art of solder level models
- 4:29Speaker 2: like Opus four point seven or a GPT equivalent, so
- 4:33Speaker 2: there's always some type of performance trade off there, But
- 4:37Speaker 2: eventually we see that gap closing and models are going
- 4:41Speaker 2: to be intelligent enough where they would be good enough
- 4:43Speaker 2: for most tasks that you need.
- 4:46Speaker 1: Yeah, Chat GPT for example, you know they have that
- 4:50Speaker 1: ten year history, but there was an inflection point when
- 4:52Speaker 1: it came out on the app Store and it made
- 4:54Speaker 1: it available to the public. I think they broke a
- 4:56Speaker 1: record for like the fastest to a million U and
- 5:00Speaker 1: like the adoption just happened overnight as everyone started like
- 5:03Speaker 1: experimenting with it. You mentioned about abstracting that stuff away
- 5:07Speaker 1: that is involved with the decentralized version and the wallet,
- 5:10Speaker 1: things like that. So what's the killer app on zero
- 5:15Speaker 1: g that we can have that happen so people get
- 5:18Speaker 1: onboarded into crypto without them knowing it.
- 5:21Speaker 2: Yeah, there's a couple elements to it. Of course, you
- 5:23Speaker 2: can build a similar chat interface, but it'd be completely private,
- 5:26Speaker 2: and so there's a kind of killer app aspect to
- 5:29Speaker 2: that alone. But ultimately it's going to be about the
- 5:32Speaker 2: types of models that can be created. So, for example,
- 5:35Speaker 2: if I want to participate in AI and I have
- 5:37Speaker 2: some sensitive data that I don't want to share, but
- 5:39Speaker 2: I want to create a model for other people to
- 5:41Speaker 2: use that I can then utilize blockchains to effectively monetize
- 5:46Speaker 2: some of that data, that can work with other experts
- 5:49Speaker 2: to create a new model that others can use, and
- 5:51Speaker 2: I can benefit from the revenue streams. And so I
- 5:54Speaker 2: think in the future those types of things would become
- 5:56Speaker 2: more and more important and then.
- 5:58Speaker 3: Finally really becoming the truth.
- 6:01Speaker 2: Layer for AI and having an AI safety layer that
- 6:04Speaker 2: can keep AI accountable as well.
- 6:07Speaker 3: So I see those as a few kilary use cases. Definitely.
- 6:11Speaker 1: It's always nice when you're able to earn from your contributions.
- 6:15Speaker 1: You know, the fact that your information is staying private
- 6:18Speaker 1: in your own hands, it's nice. But if you want
- 6:21Speaker 1: to give it away for the purpose of training, at
- 6:25Speaker 1: least you should be rewarded for.
- 6:26Speaker 3: It, right, Yeah.
- 6:28Speaker 2: Yeah, the way I think about it, if you're contributing
- 6:30Speaker 2: to like a Reddit or an x or some other
- 6:33Speaker 2: social media, you don't really own that data yourself. What
- 6:36Speaker 2: if you not only own that data, but then can
- 6:38Speaker 2: actually monetize about giving it as training data to other apples?
- 6:42Speaker 1: Definitely, And you know it's it's you need to have
- 6:47Speaker 1: a perfect storm to be able to create an app
- 6:50Speaker 1: that all of a Sudden has millions of users in
- 6:53Speaker 1: a couple of days, and if you just make one app,
- 6:56Speaker 1: I know, you know you're going to be fighting a
- 6:58Speaker 1: big battle. But from what I understand with the zero
- 7:00Speaker 1: G network, there's like over one hundred different apps and
- 7:03Speaker 1: companies that are running on the ecosystem testing out different
- 7:06Speaker 1: things right now.
- 7:06Speaker 3: Can you talk about that a little bit? Yeah? Absolutely.
- 7:09Speaker 2: There's many different types of applications, anything from you know,
- 7:12Speaker 2: infrastructure based ones that make a gentic commerce work, so
- 7:16Speaker 2: implementations of export to payment protocol for example, but extending.
- 7:20Speaker 3: It to other areas.
- 7:23Speaker 2: Then there's more on the application level, where there's a
- 7:25Speaker 2: gas AI for example, that helps you run open claw
- 7:28Speaker 2: and the Chrome browsers a plug in or life for
- 7:32Speaker 2: a couple of weeks thousands of users already.
- 7:35Speaker 1: I was just gonna ask about that because I'm like
- 7:36Speaker 1: open claw is sort of when people think of open
- 7:40Speaker 1: source AI. That's what comes to mind now, And it
- 7:43Speaker 1: seems like the functions are have surpassed these regular llms,
- 7:49Speaker 1: you know, like open claus phoning you and telling you
- 7:52Speaker 1: to do stuff for you know, tapping into the cameras
- 7:55Speaker 1: in your house if you let it to be able
- 7:57Speaker 1: to like guide you on doing things. Can we tie
- 8:00Speaker 1: that into a blockchain ecosystem?
- 8:03Speaker 3: Yeah?
- 8:03Speaker 2: Absolutely, And that's where the truth and safety layer comes
- 8:05Speaker 2: into place, Like.
- 8:06Speaker 3: Do you really want your agent to hold your private.
- 8:09Speaker 2: Keys and the access all your wells, for example, not
- 8:12Speaker 2: at this point until we have got this full safety
- 8:14Speaker 2: layer established effectively.
- 8:16Speaker 1: Definitely, it's going to come to that where why would
- 8:19Speaker 1: I call up some old guy my financial advisor on
- 8:23Speaker 1: the phone when I have like the world's information at
- 8:26Speaker 1: my fingertips and an AI agent to be able to help.
- 8:30Speaker 3: Me manage that in my own way? Yeah, exactly.
- 8:33Speaker 1: With the X four h two is the sort of
- 8:36Speaker 1: paint from what I understand, it's the way that AI
- 8:39Speaker 1: agents will be able to handle payments correct. How far
- 8:43Speaker 1: along do you think that is in people that are
- 8:47Speaker 1: using open claw and then people that aren't as technical,
- 8:50Speaker 1: and then maybe just using regular lllms right now?
- 8:53Speaker 2: Yeah, So for X four two specifically, it's a very
- 8:55Speaker 2: early stage. So if you think about last year thirty
- 8:58Speaker 2: three trillion and payment volume on stable coins or just
- 9:02Speaker 2: in transaction volume with stable coins expert to today's process
- 9:05Speaker 2: maybe six ord million or so, So it's like a
- 9:07Speaker 2: fraction of a percent and so very early on the
- 9:10Speaker 2: agentic commerce side, but our belief is that next five
- 9:14Speaker 2: plus years that will be most of the transactions of
- 9:17Speaker 2: blockchains in the future. So we're building towards that future.
- 9:21Speaker 2: And then as far as the adoption of open claw
- 9:26Speaker 2: versus other llms, open claws is slightly different because an
- 9:30Speaker 2: LLM is just the base model that you can prompt
- 9:32Speaker 2: an open claw takes it a step further to then say,
- 9:35Speaker 2: not only is there an underlying model, but I'm going
- 9:37Speaker 2: to give an agent the ability to use different types
- 9:40Speaker 2: of tools and do work for me. And so open
- 9:44Speaker 2: call would then use these kind of payment rails to
- 9:46Speaker 2: actually then pay for things through blockchain, and they it's
- 9:50Speaker 2: a natural synergy because on the blockchain based thing, as
- 9:54Speaker 2: I mentioned, it's again about the trust.
- 9:56Speaker 1: And true player definitely, And with the interp probability of blockchains,
- 10:02Speaker 1: you know, tying into the current infrastructure of DeFi on
- 10:06Speaker 1: Ethereum or Solana, are there any issues with zero g
- 10:11Speaker 1: being able to help all parts of blockchain.
- 10:15Speaker 2: So zero g has a full AI stack, so it's
- 10:18Speaker 2: not just a blockchain itself. It's got a storage network,
- 10:21Speaker 2: it's got a compute network, it's got an AI safety layer,
- 10:24Speaker 2: Service Marketplace, and so these other components are built on
- 10:27Speaker 2: a modular fashion, so any other blockchain can utilize those components.
- 10:31Speaker 2: Even the agents that we built on zero G can
- 10:35Speaker 2: utilize some of the bridging elements.
- 10:36Speaker 3: And then support on other networks. Yeah, very cool.
- 10:40Speaker 1: And with the second half of twenty twenty six, you know, AI,
- 10:45Speaker 1: at least traditional AI models are just continuing to get stronger.
- 10:48Speaker 1: There are so many faucets of this vertical stack that
- 10:51Speaker 1: you have. What are some of the top priorities for
- 10:54Speaker 1: you and the team to focus on in the second
- 10:57Speaker 1: half of this year.
- 10:58Speaker 2: Yeah, it comes down to basically all the AI goals
- 11:02Speaker 2: that we have, so things like deep research. A couple
- 11:05Speaker 2: of our papers just got accepted to ICML, for example.
- 11:09Speaker 2: We're doing research on anything from agent agent interactions to
- 11:14Speaker 2: decentralized compute to AI safety type of issues. So a
- 11:18Speaker 2: lot of those aspects we're working on. We're working on
- 11:21Speaker 2: building our own models that provide new capabilities. So we're
- 11:25Speaker 2: just releasing a model that is able to have a
- 11:27Speaker 2: twenty to thirty percent increase in benchmark results versus the
- 11:31Speaker 2: underlying QUIN three point six model because we were able
- 11:34Speaker 2: to refine and kind of retrain it and so that's
- 11:39Speaker 2: those very exciting. So that's the AI side. Then there's
- 11:41Speaker 2: a blockchain based side, which is all about ecosystem building.
- 11:44Speaker 2: So we have an accelerator called Apollo that's in combination
- 11:48Speaker 2: with Stanford and the Blockchain Builders Association there and building
- 11:55Speaker 2: anything from kind of the defile liquidity layer to further
- 11:58Speaker 2: AI applications. We've even built AI applications just to demonstrate
- 12:03Speaker 2: the centralized network capacity. So we've built a vibe cooding
- 12:07Speaker 2: app app dot zero g dot ai and within three
- 12:10Speaker 2: weeks we've already.
- 12:10Speaker 3: Burned about four billion tokens and so a very good start,
- 12:14Speaker 3: very fun.
- 12:15Speaker 1: Okay, I'll put that link to that vibe coding app
- 12:17Speaker 1: in the show notes. What's the best way for being
- 12:20Speaker 1: able to follow along with the other apps, the Apollo
- 12:23Speaker 1: and the updates as they come out.
- 12:25Speaker 2: So we will update our houb page hub dot zerg
- 12:29Speaker 2: dot ai so that you can follow along some of
- 12:31Speaker 2: the those latest news but also follow our x zero
- 12:35Speaker 2: g underscore labs for a lot of these updates.
- 12:38Speaker 1: Sounds great, awesome, Thank you Michael, all the best in
- 12:42Speaker 1: keeping growing AI in the decentralized way. Tying in blockchain
- 12:46Speaker 1: and making it open source, private and secure in the
- 12:50Speaker 1: battle of AI.
- 12:51Speaker 3: So thank you so much.
- 12:52Speaker 2: Other love what we're doing, our missions to make AI
- 12:55Speaker 2: public good and we're not stopping anytime soon. Yi