Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
The Transformation: AI x DePIN
Recorded: May 7, 2025 Duration: 1:08:30Space Recording
Short Summary
In a dynamic discussion, leaders from Ontology and CryptPlex explored the transformative impact of AI on decentralized infrastructure, highlighting innovative projects, strategic partnerships, and emerging trends that are shaping the future of Web3. Key insights included the integration of AI in identity verification, the launch of new models for decentralized storage, and the growing importance of user-friendly solutions in the crypto space.
Full Transcription
Music Thank you. Music Thank you. Music I'm going to go to the next video. Thank you. so Thank you. Hello everyone, this is Mike Chase.
Can you hear me?
Yeah, I can hear you now.
Yeah, thank you so much.
From Ontario Z, can you hear me?
Oh good, yeah.
Yeah, thank you, thank you so much.
I'm gonna wait about two more minutes for the audience to join. Then I will start the panel discussion. Okay. Thank you so much. Thank you. Okay, okay. I think it's time to start the band edition today. So this is my check again from Anthology.
Can you hear me?
Yeah, I've been clear.
Yeah, thank you. Thank you so much.
From Crip Black, can you hear me?
Yeah, of course, I can hear you.
Yeah, thank you so much.
Let's get started.
So welcome back to everyone to another episode of Web3 plus AI, the Connection of Insights.
So I'm your host today, Mary Lynn from OpenBath.
And today we will dive into the powerful and fast-envolving space of AI and DEPIN or what
we call Linf in the Planet of Business today, this is the transformation. So what are the deals, the BIN decentralized
physical infrastructure network is already changing
how we think about the thing like connectivity, storage,
and even identify.
So thus with the AI stepping in,
we are seeing a new wave of automation, optimization,
and intelligence being later into the decentralized system.
So we got two incredible guests today who are leading the transformation in the old field.
So first off, we got Giorgio Richards, head of community at Ontology.
So the project bringing decentralized identity and privacy to live with the ONID, Ontology Talking and OntoWallace.
So first of all, Richard, can you make a little bit short introduction about your project as well by yourself?
Yeah, thanks. Thanks for the short intro. Yeah, I'm Geoff, head of community at Ontology. Ontology has been around for over seven years and building identity all that time.
We're really interested in deep end and AI, actually, in terms of identity, because people think about identifiers always about as being identifiers for an individual or a person.
But actually, identifiers are also important for
infrastructure and for AI agents, for instance. And these should also add identifiers attached to
them. So we're really interested in this space. And there's a lot of development on there and a
lot of interesting things going on. But essentially, we're about identifying things, giving people
privacy, and giving people a trustless way to be able to verify and understand that who and what they're dealing with are actually the real things they're supposed to be dealing with.
So, yeah, that's the Tontology.
Thank you. Thank you so much.
About seven years to develop, that's really impressive.
And so today we got from CryptPlex, we got Andrew.
So Andrew, head of business development of Griptlex. Andrew, please make a short interview about yourself and about Griptlex.
Yeah, sure. Hi, Maurice. Hi, Richard. And hi, everyone.
And thanks for joining today's space. I'm Anrui. He's a business development at Griptlex.
And I'm super excited to be here to talk about the intersection of AI and dating
with exactly where Griptlex lies in place. as Griplag. And I'm super excited to be here to talk about the intersection of AI and Debian,
which is exactly where Griplag lies and rests. Well, Griplag is decentralized AI infrastructure
and that's what we use to turn an user storage into value storage for AI training data set.
We are part of Debian Way and but with a sharp focus on so solving real protect of ai from scalable data stories
and to smart node operations and we lost machines and i'm looking forward to dipping deeper into
how quickly as a partner they've been in the ecosystem uh with is how AI gets built and developed.
Thank you guys so much for joining the panel.
Today we got about 400 people in the live session right now.
So I also thank the audience to join the session today.
So I think I'm going to move on to the part one and the PIN infrastructure just got an
upgrade. So as you know that the PIN or decentralized infrastructure network are already
transforming the way we interact with the car system, whether it's connectivity, compute,
or storage. But now AI is stepping into the level of
thinking up everything.
So it's making the network smarter,
more automated, and more rewarding for users.
So in this part, we are going to explore how AI
is reserving the infrastructure layer of Web3
from behind the scenes optimization to real world
in parts.
So to kick off, let's talk about the chip we are seeing.
DPN already could burst now that AI in the mix.
So it's getting a seriously glow up.
So what are the biggest chip you are seeing at AI get parked into decentralized infrastructures?
So begin with Richard.
Yeah, I think it's a really interesting time at the moment.
So we start to kick off with, like with any new technology, with anything new coming in,
we kick off with people wanting to provide the infrastructure for it to work.
So one of the big things we've seen, of course, is deep in projects,
trying to build the infrastructure for AI to work. So one of the big things we've seen, of course, is D-PIN projects trying to build the infrastructure
for AI agents and AI to be processed, whether that be processing power, GPU power, storage
space, all these different things.
And that's been the sort of big focus in the first wave of AI and D-PIN combinations.
I think there's much more interesting things happening behind the scenes
and much more interesting things coming through.
The one I'm really interested in particularly
is the predictive capabilities of AI
in terms of if you want to optimize the system
and deep in systems do want to be optimized,
they need to be optimized,
then actually being able to predict
when busy periods are,
predict when people need extra periods are, predict when people
need extra storage space, predict when people need a decentralized Ethernet, Wi-Fi connection,
a decentralized mobile phone signal, all these different things. That predictive capability
of AI to gather all those large data sets and start predicting demand, and as such can then
maybe reward supply at a certain time incentivize supply at a certain
time i think he's going to be a big play in in deep in ai coming together so that that ability
for ai to take large data sets and predict trends and things that are happening i think it's really
exciting and i'm really looking forward to seeing more projects so to bring that in and and use that within their deep in infrastructures yeah 100% agree with you that the possibility of the the tradition predictive
of the ai uh i think that a lot of projects already developed themselves to to the market
analysis uh market prediction to to to the current crypto market.
So a lot of potential in the field when integrating the AI into the product.
So. And so from Andrew, let's bring with you,
CryptPlex is taking a different layer of state decentralized storage.
So how are you seeing the AI push that from T forward?
How are you seeing the AI push that from T forward?
Yeah, so absolutely.
It's lovely way to reverse it.
And you're right, Devin was already getting movement with AI step in to gain a chance
on my perspective at QuickLax.
The biggest shift we are seeing is that Devin is no longer just about centralized physical infrastructure.
It's now become intelligent
infrastructures well when i enter the mix everything gets smarter storage node and just
store data anymore they optimizing what store how it works and how it is reliable for ai changing for AI-changing datasets, routing and resort of closing, even user rewards.
Well, this thing can be dynamic turned by AI, what makes the whole system more efficient and scalable.
Well, another piece is the expectation.
AI demand performance is not just to be decentralized,
the network has to handle large data sets,
support real-time tasks, and offer reliability. Thus, we will centralize cloud giants that put
Debian projects, including us, at Krivelet to level up everything from hardware optimization
for software intelligence. So, QShot AI is just a boost work in Debian anymore. It's become more a bracing system for the next generation of decentralized infrastructure.
What makes this moment really exciting?
Well, that's Ladan.
Yeah, yeah.
I think that when AI entered to the platform of technology,
everything got smarter, better, and stronger every day, every night.
So not just bringing the AI into the data storage and rewards to them, but also for the
identified ship. So when integrating AI into technology into to other products, that a lot
of potential into the product to develop themselves.
What we are hearing is that AI isn't just an as-on.
It's becoming the core engine behind the smarter, more efficiency,
decentralized infrastructure whenever it's identified data or
resource resources allocation.
So it's clear that we entering a new phase where automation and intelligence are
plugged into the architecture.
So we are talking about the big picture, but let's get into what exactly
working today.
So AI and Web3 isn't just a future idea.
It's already powering the real use case across the deepening ecosystem.
So from the intelligent routine to identify verification,
we are seeing smarter infrastructure unfold right now.
So what are one AI private use case in your vertical that feel like a game changer
or something that actually leads and delivering the value today.
So Andrew, let's start with you this time.
So what one quick-like use case where AI already adding real value to your user?
AI already adding real value to your user?
Yes, so a real AI in Web3 is way past the white paper phrase
and is already creating real, commercial values for us.
Like the standard use of AI,
optimized decentralized storage for changing data is for this new game changer.
Here's the view.
Well, traditional storage uh traditional store
network in red3 are rich for decentralized but when it comes to a hangover large scale the
hyper-formed ai workflow like changing multi multi-model model will with video or audio
and complex data sets uh they start to buckle.
Well, that's where we come in with like we build a team.
That's where AI at all reasons, means data allocation
and the content is set, category and encrypted and distributed well and the final is reward based on the actual utility and performance of their storage
it's not just uptime and the reason is we are already seeing ai team tapping into crypt like for things like computer vision model is global distributed data that's useful to
switch idle on tuple device.
Well, it's not just a cherry is happening right now and it's unlocking AI
capability that's weren't possible with classic WebP in place.
Yeah, so basically the CrabLab PUDP networks
to where the data allocation,
you can allocate the data and content and distribute it to the correct space.
Right.
And and also utilize the ability of this data and
make more
use it to make more potential for the data, right?
I think I think
as my own data on my computer or anything, I think without the AI, I think my data in my laptop or anything is a mess.
A lot of information, a lot of data, I just put it. And I don't know when to take it down or, or, or do anything.
So I think that the data involved in the, the, the AI
involved in the data allocation that that gonna be good and potential in the future.
So Richard over to you.
So what's something in the ontology ecosystem that whenever it's on ID or on login or on wireless or something else where AI is already proving this out to the real world on ontology ecosystem?
Yeah, and this is important because, you know, I think, as I said earlier, I think a lot of what we see is actually Web3 ecosystems enabling AI.
And I don't think we see enough of AI enabling Web3 ecosystems at the moment.
One of the things we have in the ontology ecosystem is on Onto Wallet, we have our own AI agent in the Onto Wallets doing things on there. And I think one of the things this can do really
well is similar to sort of data and storage on a computer. It's that ability to be able to have a
look at your positions, have a look at where you're at in your investments or your NFTs or
whatever you're working on and very quickly understand what's happening and very quickly
have trends and patterns pointed out to
you. Something that's really interesting is humans are not great at pattern recognition. We're not
very good at it. It's why we invent statistics. We're not very good at understanding trends. We
don't just look at something and understand a pattern. We don't look at something and understand
a trend very well. AI is much better at it. So in OntoWallet, you can download OntoWallet,
you can set up your
identity, you can use an AI agent to do various things in the Web3 world to understand positions
and where you're at much better. In the future, I think that gets more interesting when you can
actually start using it to potentially issue proofs of who you are, check on proofs of who
somebody else is, all without revealing anything outside privacy. And I think, you know, on the data thing, I want to go back to that as well.
I think having AI be able to help organize your data, organize your storage is really,
really useful and really important. We're very bad, actually, at data organization and storage.
People tend to be sloppy. We keep things we don't need. We don't optimize storage facilities and all these things. And I think that becomes really
important. One of the things I'd like to see added to that is a sense of privacy.
So do you know which AI is organizing your data? Where are they moving it? What are they doing with
it? How are they programmed to do it? What training set are they set up on? And this is where I think something like decentralized identity with AI, with D-PIN comes really useful,
because then you can actually have an understanding of which model, when, how, who they are, is it the
right one, and all those different things. And I think that's really important as well. So I think
we're really, this is, again, the catchphrase of Web3, right? I think we're
really early in the deep in AI revolution. I think we've got a lot to come. I don't think we're even
touching the start of what we're going to see. It's going to be quite exciting. But yeah, people
are starting to build things and they're pretty cool. Yeah, the AI is really important in nowadays.
Yeah, the AI is really important nowadays.
Web3 ecosystem and applying the AI into the ecosystem,
that's going to make everything stronger and smarter.
Also, the UntoWallet integrate AI to support users,
that is a good case for everyone because add a new user to the new app,
to the Farmer Web2 to former web 2 to web 3.
So when integrate the AI into the wallet to support them, that's going to have a lot for the new user.
And so I'm sorry about the trend, production market.
I think AI is going to have a loss.
So basically integrate the AI agent into every product for the production in the trend of the market.
It's a really good case.
So it's clear that AI already doing the work, whether it's improving the data flow and
enabling seamless identifying or optimizing infrastructures is delivering real outcome and lift environments.
So let's keep things going and zoom into trends that are shaping the future of AI and
Depean. So now that we have seen where AI already delivering in Depean, so this space is moving
really, really fast from the AI to Depean. So every month, there are new trends, new tools, or completely new ways of thinking about decentralized systems appear in the market right now.
So let's talk about the standing now.
What are one AI plus the pin trends that are completely flipping in the script in your world right now?
your world right now so richard let's start with you what train in identified space that uh
So Richard, let's start with you.
cost your attention something that could could fundamentally reset how user interact with the
centralized ecosystem yeah one of the things we're talking to a couple of deep in projects about
is how we can have a position where we understand what the infrastructure is, who owns it,
what is it doing, how well do we understand it.
And I think this is a trend we're going to see come out of the AI deep in space in that,
let's say you're doing decentralized, let's say you're doing healthcare,
you've got some sort of deep in health care
where you've got these uh systems set up around the world and people can take photos do biometrics
and various other things i think you know this becomes really important people without access
to health care people without direct access to things and or just can't afford health care and
this can make it more affordable and ai is is helping that, AI medication, AI doctors working with this deep in infrastructure
to do those things.
I think this is going to change across a lot of the way we access different facilities.
It could be education, it could be healthcare, it could be storage, it could be all sorts
of things.
This is already starting to move and this is starting to change not just what Web3 looks
like, but what infrastructure looks like for people, particularly important in areas that
don't have robust health systems or education systems in place. And they could be accessing
healthcare or education from anywhere in the world, anywhere in the country. So if you're
living rurally, this could be really important. And I think this sort of deep in AI combination
is really important.
I am going to just reiterate,
I'm at an identity privacy project with Ontology.
So of course, I'm going to reiterate this.
One of the things we have to have underpinning that
is robust privacy and identity and data storage.
It has to be robust.
And I think in all of these things,
a trend I'm really looking
forward to seeing develop even further is how AI or AI agents can talk to each other without
revealing private data. So you imagine you've got an AI agent as a healthcare provider in some deep
in healthcare infrastructure, being able to talk to your AI agent, which is your
identity provider, authorize who you are, prove who you are, share some health data and so on.
All of that happening without it actually ever being seen by a human, without it actually being
put out into the wide world for people to be able to see your private data. Again, we're at very early
stages, but I think this is a trend we're seeing develop
that I'm really excited to see how far we can go with it and what we can actually pull
off with it. So yeah, quite a big one that, but I think it's quite exciting.
Yeah, I think that a lot of potential of when you talk about the AI agent, talk to the real
people to check about their condition and got into analyze about their disease and take medicine and something else like that.
That's got a lot of potential in the health field.
So, and also, I think it's going to be a trend for the AI in the near future that the AI
can talk to each other to develop themselves.
For the example of the healthcare ecosystem, a lot of disease every day, a lot of patients every day.
So the AI talk to the other AI to develop themselves, to analyze the world ecosystem, health ecosystem.
That's going to be great.
So, Andrew, what are one trend that's changing the game for Qriplex or decentralized infrastructure more broadly?
Yeah, that's a good question. And yes, moving fast, like script, right?
And one trend of really failing with script work is the rise of performance based on been incentive powered by ai and let me break it out the old tevin
another reverse contribute pretty even this if you if your notes were online you would pay
doesn't matter if your hardware was high performance from just bearing scrapping by
that's fine at first but with ai workload they don't play like us you need fast reliable and
intelligent infrastructure that means we need fast, reliable and intelligent infrastructure.
That means we need a smarter way to reward contributors.
So now we are seeing AI step in to elevate performance in real time.
How fast does the nodes handle AI changing data?
It is developing slow-lensity results for inference.
Can it support video, audio and coding task efficient
while quickless use data agent from small meter on a dash and then might
demand the code adjust reward better performance equal more cdx points that creating a new kind of competitive economy where contributors
are incentivized to optimize not yet exist. Well, that's it is massive and it's transforming
dating from passive network to living in Boeing AI optimizing infrastructure layer. We are not just store data anymore. We are building the compute
type form for the next step of AI. Yeah, really, really not.
Interest in delivering the data into the AI and have them improve quicker. So that's a powerful shift when infrastructure adapt in real
time and scale smarter with the AI. So it really
opens up the ecosystem. So
there you have AI
is just adding power into the business.
It's introducing the new playbook whenever it's
rethinking how users connect and how resources get allocated.
So the trends are pointing forward to a radically more adaptive and intelligent in Web3.
So let's drill down a bit and get specific.
So Andrew ClickBlight is doing something that feels futuristic, but already happening
in the market right now. So you are connecting the unused disk space and AI training and
rewarding users in the process. So that's really a big deal. So turning unused disk space
into rewards while training the AI. That's the next level. So how does AI make
Crisitlite smarter and optimizing the storage and keep users happy at the same time?
So it's not just cool, it's efficient. At Crisitlite we don't use AI as a music work. It's
spread into how our storage network works and involves the the real talk and well decentralized stores are
awesome but it can get messy fast you've got a model on a kind difference with different reliability
different locations and if you choose all stories the same you either with potential and get
from next where that's where i am starting well i am a quickly smarter in three ways smart
allocation instead of applying distributing data our data our ai and dynamics aside by based on
model performance you got a model that's better at handling lag video data set or high speed request
a handling lag video dataset or high speed request.
Well, that model will get priority tasks.
That way, users can get fast except contributors get more targeted rewards.
Second is self-hailing redundancy.
If a node goes offline and underperforming, the AI doesn't punish and it will be it will be router data intelligently restore
redundancy and keep everything running smooth with manual intervention well
send up sorry final is performing personal optimization on the user side
personal optimization on the user side,
AI predict demand and adjust storage distribution
for smoother AI training or inference workflow.
Think of it as content-aware storage management
so training model as quick-like is seamless and fast.
Personal-like quick-like is learning
just not store data. It learns how to store better, faster and more efficient and reward everyone who makes that possible.
This will make our users happy.
Yeah, I think that's Kriplik infrastructure is going to be great because if you got an AI
integrated into your machine, your laptop or anything,
the AI will know which are the priority, which is the one that you have
to get all the user space, all the disk space, the RAM or anything to put the performance.
That's going to be great.
Yeah, 100% impressive with the quick play infrastructure.
So that really clear the picture that is so how AI isn't just improving the system performance,
but it directly shaping user experience and engage at the scale.
So let's shift the gears from storage to
the identifier. So one of the biggest blockers in
Web3 right now is onboarding and user control, especially
around the identifier. So most people still struggling
with the wallet key,
login and verification.
So where AI can really shine in that field.
So Web3 identifier needs to be seamless and secure, right?
And how do AI help ontology eliminate the friction
friction while keeping people in control of their data?
while keeping people in control of their data?
I think this really cuts to one of the big issues with AI.
You know, everybody's very positive about AI, including myself, and rightly so.
But I think we have issues.
And this is sort of around how identity and AI can work together.
And the problem we have is they don't necessarily seem natural
helpers to each other. What you don't want to do is hand over your identity and data to an AI agent
without full knowledge of how that identity and data is going to be used. You don't want to do it
and lose control of it. Web3, decentralized identity is all about sovereignty, right? It's all about owning your identity, owning that control.
And so it becomes a challenge, a technical challenge,
but also a philosophical challenge of how you bring those things together.
How do you bring together identity, AI, and control?
Now, in terms of onboarding, I think there's lots of things already happening with onboarding.
If I'm perfectly honest, I don't think AI is necessarily the solution for onboarding, I think there's lots of things already happening with onboarding. If I'm perfectly honest,
I don't think AI is necessarily the solution for onboarding. I think we have account abstraction,
OntoWallet and OntID just announced the move towards account abstraction on things like that.
And I think usernames, like name services, like Jeff at Ontid, for instance, would be better than all our long
keys and all our confusing things and all our security codes. I think we have other solutions
in place. I think where AI does have a role to play is around when you prove things and how you
prove things. Risk assessment, for instance. If know, if you're going to sign something with your identity, can your AI do a risk assessment of what you're about to sign? Can it check to see
whether you're going to sign something that's really secure? It is what it says it is. It's
doing what it says it's doing. I think this is one of the big problems we see in Web3 is we don't
know what we're signing half the time, you know, unless you're a developer or you've studied it, we're all signing these smart contracts. And this can be true of
identity as well. Smart contracts signed to share your identity and data. And we don't always know
what we're signing. And so I think this is something we're really interested in with
ontology and ont ID is how do you use AI to help that security, to make sure what you're signing
is what you think you're signing,
what you're sharing is with somebody legitimate,
and all those sorts of things.
As with everything that's new,
there's a real temptation to think it's the solution to everything.
In terms of onboarding, I think personally we have better solutions,
but I do think it has a role to play within risk assessment
and how we share data and how we do things from there.
Yeah, that a critical insight i'm pretty sure that that we don't know about what we are signing
uh especially in my real user case uh basically if i i go to to some new website or some new product, I will use a brand new wallet to connect to
them, but not the one that I own the asset inside because I'm really scared of hacking
or anything like that.
So especially when you think about the mass adoption in the market, that's going to be a big, big problem.
So I think that AI making the identified system feel more natural and secure, and that's really
a real unlock in the web through ecosystem. So let's wrap up the part two and look into the
future a little bit. So we talked about the trend and use scale that are real today in the AI and
the pins, but now let's get real a little bit.
So imagine that in the 2013, the AI and the pins space has evolved far
beyond what we see today. So in the next five
years, we will see something bold, something unspecialty or even a little bit crazy. So what
in your mind, what do you think that will become the real in the next five years, Richard?
in the next five years, Richard?
Yeah, I think we could see some really interesting things.
Let's imagine this, right?
It's 2030.
You are applying for a new job.
You apply for a new job.
Your digital identity, your aunt ID, the AI agent,
proves who you are to your new employer,
shares your education certificates, do all these things.
But your employer never gets all this data. It just goes to their identifier as well, their aunt ID, and everything gets approved.
Somewhere along the line, this job is going to involve you using extra storage at home.
Maybe you need some extra GPU usage. You need some more file storage.
you need some extra GPU usage, you need some more file storage. And your AI takes your aunt ID and
talks to a file storage provider on your behalf and predicts that you're going to need more file
storage. And so when you come to use it, you've never done anything. You've authorized your AI
agent to act on your behalf. They've ordered you some file storage. They've proved who you are.
They've checked the ID of the file storage provider to make sure they're legitimate and
they're doing a good deal. And all of this happens in the background. You then want to order
some beer. And actually, you go online, you order some beer to be delivered because you're going to
celebrate. And your AI agents just check for the agent for the store and that beer gets delivered.
And, you know, at the same time, you think I'm drinking too much beer. I best have a health check
because I'm not looking after myself. And you log on to your decentralized healthcare provider who's got some deep in biometric scanners and things like this,
and you do your scans and you do these things. And again, your AI talks to their AI, making sure
that the data's stored okay, share some private data without sharing it clearly using zero
knowledge proofs, things like this. And all of a sudden, you get your health check, they give you your results and tell you actually you drank too much
beer, don't drink as much beer next time. And you get those results back. And all of this happens
under the scenes, you've got children, and your children want to log online, and they try to log
on to watch a program. But this, they're logging on to a decentralized entertainment provider who's
got deep in around the world for people to upload and store movies and do things like this. They're logging on to a decentralized entertainment provider who's got deep in around the world for people to upload and store movies and do things like this. And AI predicts
the child's age, because actually what we don't want to do is share lots of information about what
that child is, who they are, how old they are, images of them. Very difficult to do age verification
for minors, for children. It's actually a very
sensitive subject. AI can make those predictions based on their behavior, based on what else
they've done on their machines and so on, and actually make a prediction if they're age
appropriate for those services. And so my 2030 vision, I guess, is this idea that we have AI
agents doing the hard work for us in the background. They're organizing new file storage when we need it.
They're proving who we are to get that file storage and then making sure the file storage provider is legitimate and who they say they are.
It could be decentralized and physical infrastructure for energy.
And we could be getting our energy provision from Bob down the road, who's got some solar panels installed and doing
these things. But actually, we know because our AI agent has spoken to their AI agent,
and we've proved their identity and our identity. And there's a little verification tag on OntID
that says, yeah, this is green energy. This has been taken from solar. They've taken so many
watts per hour for the last year, and it's all done in the background. And so we know, and we get a little report
maybe from our AI agent at the end of the month
to tell us how much carbon we've saved,
how healthy we are, what our job structure is.
Have we used too much storage?
Is our GPU usage too high or too low?
So on and so on and so on.
And I would love to see that level of automation
where we can trust the agents to go
away. But to achieve that, we need to make sure, one, that we underpin that with privacy, and two,
that we know who these agents are. And I keep stressing that. But at the moment, even if you
use something like ChatGPT, right, the same model changes. Sometimes it changes in the background.
So you don't know which data set
it's trained on. You don't know what it's doing. I think it's going to be essential that we get
some sort of date and timestamp on the models we use so we know what data has influenced those
decisions, what data has influenced those reports. And to do that, AI agents in reality need a decentralized identity.
They're going to act as agents on our behalf.
They need to be identifiable.
And so I think 2030 could be exciting, but we still need to build out the security and the infrastructure to make sure that users are protected in terms of privacy and data.
Yeah, I think that your imagination and the vision in the NetFigy
is not too crazy, right?
So especially I think that's going to be real for everyone,
for the user, for the AI.
So everything is going to be real, I think about that.
But the pet, the important thing that we have to develop is the identify
and the identify and the importance of leverage the AI into the identify ecosystem
to get more involved to the real people.
So I love it. I love your insight.
I love your vision about the next five years.
So let's get a little bit to Enryl.
How about you?
Yeah, all right.
So the platform is quite right.
But it's like a closer look. Well, by 2020, I think AI with data will completely be decentralized in the region itself, if not just data or compute.
But the thinking layer where talent of AI Asia across the dual team network involves autonomously owners and directors by community, not corporations.
Well, which are these?
You've got an open-source DBT-line model living on physics,
if not just change, runs, and develop,
but constantly improving using decentralized compute, storage,
and real-world feedback from users across the world.
Well, it learns from different environments tasks and cultural
and it's not single entity can shut it down and on monoprolitment it now layer or tokenization
with two patient systems where every contributor that ai wrote is longer and rewards your problems, your data, your training power, but
they want you to value and you share it that upside. So yeah, by 2013, we might see community
hours super intelligent running on the central line as well, changing 24 seven on real-world tasks and open set to powerful
productive AI and that's no big tech in case.
Yeah, that kind of prediction, right?
So we like the ambitious, right?
We like the encouragement or all the totally everything
about the prison is totally possible in the near future.
In the next five years, right?
So whenever it's a aware infrastructure or real time identified,
guardian, so it's clear in this space that we have no intention of stating what is in the box, right?
It's got a lot of potential in the market, a lot of things to do, a lot of potential for the next five years.
So really, really happy to see your ontology and quick vision about an FIG,
about the imagination, that's really, really good.
Now we have talked about trend, prediction,
so getting to the actual system and the impact they are making.
So let's start with the ontology.
They identify one of the biggest challenges in Web3,
especially when it comes to making it secure and user-friendly.
That way, AI can quietly put the powerful transform to experience.
Richard, check us inside the process.
How is AI having streamlined verification, detect a bad actor or even
personalized onboarding for new user? Yeah, and I kind of touched on the new user
thing already and I'm not fully convinced by it. I think one of the things that can be done and we are looking at is whether
you can use AI to analyze people's sign up for identities, right? So can you use it to detect
false signups? If somebody's trying to create an identity, can AI be used to actually detect
telltale signs that where somebody's trying to spoof your identity or create something that isn't theirs that's quite difficult i think it's quite an interesting one um but it is a difficult
one as well um in in terms of what it can do for the user experience actually i i think
one of the things i think can be done really well is ontID. So OntID has got its own Twitter account. They just
announced today that they're using account abstraction and changing the name in things
to make it easier for everybody. But I think one of the things you can do once you've done that
is combine your identifiers with AI. So rather than making it easier to log into and create an
OntID, which I'll be honest is really simple anyway, it makes it easier to log into and create an ont ID, which I'll be honest, is really simple anyway.
It makes it easier to onboard into other platforms as well.
I think that's part of the issue we have.
Identity is complicated, but it's quite easy to do.
It's easy to create your identity.
It's easy to link it to your passport, your driver's license.
It's easy to do those things.
We have a system of
their verification already um verifiable credentials this is an age-old system it's
been around for a long time we do it with passports driving licenses and so on bringing it on chain is
slightly different challenge but it's happening anyway i mean the european union are building it
the usa are building it and decentralized providers are building it. This is also to be built in quite an accessible way.
I think if we do it right though, and combine it with AI,
what it can do is all the things I've been talking about.
So when you decide to look into a project,
when you decide to put your trust in a project,
when you decide to share data,
AI can do a lot of the checks around
how legitimate do we think this is?
How good do we think this is? How good do we think this is?
How safe do we think this is?
And I think that's what we'll see coming through in the next few years on there.
And I think it can also measure, be good at predicting fraud.
One of the interesting things about fraud is you can almost predict who's going to be
the victims of fraud.
One of the interesting things I found out on that is if you've already been a victim of fraud, you are
more likely to be targeted by fraudsters in the future because they expect to have more success
again. And so AI can also predict who is going to be vulnerable to fraud attacks and can actually
tell you to be vigilant and start warning you to be vigilant. It can take data from around the net. It can take data from around the Web3 networks and say,
these are some typical frauds going on at the moment. Please be aware of them. And so I think
there's lots of things it can do. And I think linking that up with OntID will be very powerful.
How that will work, I think is going to be a huge challenge for not just on-ID, but for identity providers around the world because of all the questions and issues around privacy and sovereignty that I spoke about earlier.
your thought about the prediction fraud because when Identify comes to the real world, so who
will want to do the fraud and reporting it to the user and the web user about this, that's going to
be huge potential in this field. So that's really a strong user case. Basically, Identify needs to
feel monitor, not just look like security checkpoints.
And the AI can really have to create the balance between security
and simplicity into the product.
Let's keep going with the ontology for a moment.
One of the two URL role now is onto loginLogin, OntoWallet and OntoID, right?
So it gives users more control over their data than over before.
So we all have the login
features, so OntoLogin and OntoWallet are changing that.
What the AI layer doing behind the scenes
to make Identify Flow like second nature?
Yeah, not a lot.
I know that's not very exciting.
Yeah, yeah, yeah.
That's not an exciting answer.
Yeah, you know, I kind of covered this,
but I will take this chance to plug it a little bit, right?
I mean, Onto ID, follow the Twitter account.
There's always updates coming out on that
with what's happening in, not just in Ant ID, but in the identity space in
general. And, you know, I think account abstraction and I think replacing addresses with Jeff at
ant.id is going to be much better, much easier. Passwordless login is really important and all
those things. I don't think we need AI for those things. I think there are other things we can do AI in.
I spoke about that a lot,
so I'll not take up everybody's time just going over it.
But check out AntID.
Lots of information there and lots going on.
Yeah, basically, this is all about how to stay for an easier, right?
Yeah, I think, you know, it's one of those things.
Sometimes there are better solutions.
When we get some shiny new toy like AI, there's a temptation to try and use it for everything.
And I don't think we should.
I think we should use it for what it's really good at.
Yeah, yeah, 100% agree with you that really showed the power of AI behind the curtain, making something that feel more
visible, but deeply improved user experience.
So, all right, last week geared to decentralized storage.
So, Andrew, over to you.
Group Lake is going fast, which is exciting, but also mean complexity, right?
So when you are dealing with decentralized storage,
according to many nodes, according to global,
we are real-time challenging in performance,
conditional, and security.
So how the AI help keeps it all working seamlessly.
So Henry works us through how are support real-time data routine
not reality or even anomaly detection at preplex
yes so it's a great question and honestly it's one of the biggest things we've focused on
as critical scale when you are managing millions of decentralized nodes across the global
thing that childhood class and performance of time security you can manually record the data
that's a scale and that's where ai become mission typical and it's not just a cool feature of public AI at the core of our occupation layer. We use AI driving routing and load latency to intelligently matching storage tasks to the best node based on their speed, uptime, historical performance and even energy efficiency well so instead of a static system you get a dynamite self-optimizing network
that can get better the more it runs on the security side ai how to detect a nominal
like part actual time to select storage or many many toolets reward well way faster and that's occasional
system it's far unusual behavior button and text action outcome
automatically the way before is become a threat so in short AI is the brain that
keeps everything humming smoothly behind the screen, making sure we are decentralized and reliable without needings to
clarify performance or expose ourselves to bad actors. Is the secret hot last let's quit the scale without breaking?
That's a powerful application.
So you're not just storing data,
you're making sure that that is secure,
accessibility and intelligence
is a scale, right?
So talk about the reward.
In the BIM model incentive or everything so
they can also be a little bit tricky.
We got to reward good behavior, discover bad actor
and do it fairly, so ultimately and transparently.
AI can be a huge asset here.
So let's dig in a little bit.
Andrew, how is Prudbeck using AI to evaluate performance, distribute the reward to the user and protect the integrity of the ecosystem?
Yeah, absolutely.
Decentralized storage sounds amazing until you start
rewarding people for this.
Then suddenly it's the GM who real, who cheat things,
and who just farming rewards.
That's why AI Comrade just changed the GM for us at
Cripp.
We use AI to run continuous evaluation and storage proof.
It's not just let your just um let your store with five
and how fast how often is what assess how reliable your service and how relevant your stories with
high priority data like ai data that has a large training that assess that way reward goes to the node that actually adds value if not just feels like.
We also use AI to detect sufficient buttons like node spinning up to fake graphics,
capitating content, or playing a gaming game.
Instead of getting the reverse for all these hours sitting back and limit those actors in real time,
it's predictive, not realistic.
And the end goal is a reward system can contribute who are actually helpful, earn more, and free lower is all the broader gets payouts.
It keeps ecosystem fair competitive and most importantly scalable without turning in child.
At the end of the day AI makes sure Privilege is rewarding contribute and if not many play stations.
Yeah so basically it's not just how fast out and the data the user got involved into the product
how fast the data the user got involved into the product.
It is about the leverage the AI and have been it's to improve the ecosystem
and reward for the contributor, right?
So that's a key piece of the puzzle.
So with the AI, it sounds like Krip Lake can stay alive, scalable and
align with user incentive or as one, right?
So from the CQ identified to smart storage, it's clear that the AI isn't just optimizing the layer.
It's becoming a core operating system
for how the SYNC line infrastructure works.
So we have covered a lot of information today.
So from how AI research in the future,
research in the infrastructure to real-world use case,
and about the prediction in the next five years.
What happened in the pandemic today is a lot of
information and a lot of things really insightful.
We know that both of your projects are building fast,
sitting options and influencing where the space is going.
So let's hear what's next into your core product.
So further on, Gio and Andrew, Richard and Andrew,
so what are the nets for Ontology and Kripp Lake?
Anything about the major of that, about the product lens or about the milestone
that community should be working for?
So Richard, let's start with you. What we can expect from Ontology in the coming month
on the identify side or the integration or the broader ecosystem pro?
Yeah, so so many things at the moment.
So at the moment, some things.
And actually, I noticed one of the users on Twitter
had asked a question about zero-knowledge proofs.
So I can touch on that as well.
So Mr. Bullish, he's a honeybee on his Twitter profile.
And so zero-knowledge proofs are really important.
So people who don't know what zero knowledge proofs are, it's the ability to prove something
without giving away all the knowledge about that, which is really important for privacy.
So we have a, and I was really interested to hear the idea of reputation almost to keep an ecosystem
safe and to reward the correct users based on their reputation for improving an
ecosystem. And so we have a protocol called Orange Protocol that is decentralized reputation built on
top of ONT ID. And ZK proofs are coming anytime now on there. So the idea behind that is that
you'll be able to prove sort of reputational things without giving away all the proof and
all the data.
So I hopefully not stole their thunder on that,
but that's coming out very soon,
built on top of Ont ID,
and will be our first foray out into the field
of ZK proofs in a usable way.
And I think that's really important.
I did mention earlier Ont ID is as launched
or updated the roadmap progress.
And so what's coming is account abstraction.
So you can just have a username.
Also things such as passwordless logins used to that as well,
but also private chat.
So the ability to chat using your decentralized identity
on-chain in a private way.
So we've got that coming in,
expanding the AI systems on Onto Wallet.
And then looking at partnerships with people like we've got a GameFi series coming up.
So looking at how we can support gaming projects using identity and reputation, talking to
healthcare providers to see how we can support healthcare providers, talking to a deep-in infrastructure provider to see how we can
use verifiable credentials, identity onto a wallet reputation to mean that people can make better
choices around infrastructure, around energy, and things like that, and make sure that everything's
verifiable. Lots in the pipeline, lots happening, lots going on. So, you know, we run conversations every
Tuesday on Telegram. So join us on there for a chat. We also have just launched working groups
on Discord. So if you're interested in Deepin, there'll be a Deepin working group, you can join
the working group, and you can help us better understand how identity and Deepin can work
together. We have an AI one as well. So you can join it if you're interested in our AI.
So you can join our working group
on how AI and decentralized identity
can better work together.
So that's all on Discord.
Come and say hello,
either on Discord or on Telegram.
We also have a starter package,
which gives you everything you need
to get started with Ontology,
including some token rewards to get started,
plus some personalized guidance from our Harbingers.
I can see at least one, two, three, four of our Harbingers in the chat today listening in.
So thank you for being here, guys.
Really appreciate that.
And you can reach out to them,
and they will literally give you personalized guidance into the ecosystem.
So yeah, lots happening, lots of support, lots of help.
And just to finish off saying thank you for having us,
really enjoyed being here.
Yeah.
Thank you, thank you so much for joining the
benefits of today's information, a lot of activity
that's gonna happen in the anthology in the near future.
So glad you are here with us and sharing about anything about the AI,
about the DPS, about what happened in anthology.
Thank you so much.
And so about the Crueb Lake.
Andrew, over to you.
Crueb Lake has been scaling fast.
So what are the roadmap?
Are there any new future partnership or ecosystem integration coming up?
Yes, absolutely. And this is one exciting. So one thing for crypto is honestly it's already
happening. We just dropped our new model, this and info and that they are serious next ever like you.
If you thought running the node was cool before
wait till you see how much faster and more powerful these are this for the speed thing
real-time performance snappy processing and inferno that's one of the best well it's built for
heavy lighting advanced bi training and just straight up to maximizing your earning just cool fast well these models
i just play on your performance and yeah you can earn up five time more city toys and plus it if
you can move fast they are still early birthright for you and so if you are past a quick farm
you are part of Qpublic Farm or you will be thinking about joining NavidadChamp and start
talking rewards while borrowing through AI in phrase.
So go check it out before the pick run out and see you on Qpublicverse guys.
Thanks for everyone and good bye.
Yeah thank you so much Andrew. For everyone, Preplex just
integrated their new model into the ecosystem, into the product so if you're the one that's
interested in Preplex, please come to Preplex website, come to try their product and see what
happened, what is going on in group like.
So again, thank you guys so much for joining the panelists today.
And so for the user, whenever you are a builder, an investor or just a
deep-end process user, so we have today the question gave you something
vulnerable to think about and maybe we will even inspire you to to dive deeper into the AI and the PIN movement.
on Preplex, you are both set in the future of Web3 and it's really exciting to talk about
Again, thank you to to the Ritter from Ontology and to the Andrew Davis from Coldplay. You are both step in the future of Web3 and
the AI and the PIM with you both.
So for the audience, be sure that you follow Anthology and Preplex for more updates about
their product.
Stay locked in for the next episode of the Consonero Insight.
We have got more panel and more voice and more story into the Web3 and AI.
Thank you guys so much.
Thank you, Richard.
Thank you, Andrew.
And goodbye.
Host
OpenPad AI
from
OpenPad
