Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
AI Meets Crypto: Building the Future with Telos & Tairon
Recorded: Sept. 18, 2025 Duration: 0:56:09Space Recording
Short Summary
Tyron is set to revolutionize blockchain interactions with its ambitious goal of 300 integrations by year-end and 1000 by next year, leveraging its innovative MCP technology to bridge AI and blockchain. With over 100 integrations already in place, Tyron is not only growing rapidly but also establishing strategic partnerships to enhance its ecosystem, while its economic model focuses on sustainable fundraising through API subscriptions.
Full Transcription
Thank you. Okay, just doing a quick test here.
Everyone can hear me?
Thumbs up if you can hear. still can't hear oh i can see the icon with the sound
andrew is your microphone working this time? Yeah, yeah.
Oh, that's much better.
Oh, finally.
Wow, that was really bad.
Great to have you back.
Sorry about that.
Thanks for having me here.
Happy to be here.
It just feels like a one in six chance of space working these days.
It's terrible. All in six chance of space working these days. It's terrible.
All investments in Tesla.
Yeah, that's true.
That's true.
You think when we can shoot up rockets into space and reuse them and build autonomous cars,
we could have a basic voice conversation on Twitter.
But apparently this is more difficult.
Yeah.
Well, I know, um, people are still joining back, but this is a recorded space.
So I think we can get this off the ground.
Um, and, uh ground and go from there.
So let's just get started with just a little bit about you, Andrew.
You're obviously the CTO of Tyron.
I'm just curious, like what kind of work did you do before you joined Tyron?
Maybe let's better start with Tyron and then I will connect it with my background.
Yeah let's do it let's do it let's get back yeah start with Tyron. How did you join Tyron or
were you one of the founders I guess that came up with the idea what was the sort of background uh need yeah the idea basically evolved from um our test with
mcp servers so like we saw the opportunity um to connect on-chain data and on-chain functionality to AI. Like if you ask ChatGPT or even Claude to do some on-chain research based on on-chain data,
and especially if you ask it to make some transactions or staking transactions, transfers, swaps, anything, it's just blind.
transfers swaps anything it's just blind so all the existing llms are blind when it comes to
blockchain protocols or chains right so we would like to fill easier understanding for normal people we call it oracle which is basically
true where we connect a whole bunch of mcp servers with our let's call it brain or system
let's call it brain or system, MCP Supergraph.
And in that way, we allow LLMs to see and to interact with blockchain,
with chains, with on-chain protocols, DeFi, NFTs, I mean, everything.
So we set up a goal for us, like internal goal, to have 300 integrations by end of this year and 1000 integrations by end of next year.
Each integration basically is a piece of a puzzle in this big picture that we call MCP Supergraph.
picture that we call mcp super graph so yeah let maybe you can interrupt me with some question
because because i can speak about the product for hours you know no that's good that's good
i just for all the people new to like mcp and ai on chain maybe just um what what does mcp stand for yeah yeah sure so uh like think about mcp as a
plugin for a specific um service or a protocol or any project um that allows it to be connected to AI. So MCP technology is not limited with just on-chain stuff.
So there is a thousand, maybe even a tens of thousands,
different MCP servers.
But to simplify, let's call them plugins. Yeah. So you can build an MCP server for anything that has API endpoints.
So for example, MCP server for your Google Calendar would allow your AI assistant,
whatever you use, or Claude or Gem Gemini, or ChartGPT,
or anything, would allow to interact with your calendar.
So you can just write down the task to organize your calendar.
And MCP server, using the API endpoints,
will do the rest so like it's a combining of
human brain uh from llms with the functionality of a specific project that is um
connected to the mcp server the other case is like i don't know if you use notion you can combine even you can combine MCP
servers in one prompt so for example you can ask your GPD assistant to check what how your
calendar looks like for this weekend and create a task list for grocery to buy on the weekend,
something like that.
And here we come with Web3 integration.
For now, there are a couple projects
that are starting doing the same.
But we already have 104 or 105 integrations done into our system.
So basically it allows users, and I will talk about our users a bit later.
It allows all of them to interact with all the integrated protocols. So for example, if you want to do the research on TVL,
on our Bitroom during three last months,
you can ask the chat about it.
And it should understand that you are asking some specific question
where the Defilam mcp server would be
like a best fit right so it will send the api request to defilama immediately get the response
in a machine way and package it in a human understandable way using llm
an understandable way using LLM.
Yeah, it makes a lot of sense.
Very interesting.
I'm glad you talked about some of the use cases for like MCP
as like this standard for AIs to interact with application.
I'm glad you brought up some examples that aren't related to blockchain.
Is that actually makes it more
even more exciting because you
have AI that's able to both interact with say your calendar which is one example you gave but also
interact like on chain uh at the same time to like maybe process a payment or something
because that like really opens things up quite a lot with what ai can do really compared to what it it's been able to do
in the past like one example that comes to mind is like eventually maybe being able to book a flight
using ai using an llm like chat gpt you know being able to say hey find me the best flight to la and book it and the fact that i guess mcp would be like
that middle layer that allows uh the llm to maybe pros or find the best flight but then maybe like
process the purchase of that flight over a blockchain eventually.
That's kind of exciting that you have all these things working together. And then with the calendar, being able to add it to your calendar as well.
Maybe there's better examples than that, but I'm imagining lots of magical things can happen when this MCP standard starts to propagate everywhere.
Yeah, absolutely.
And on chain, right?
Absolutely.
Because MCP as a standard is quite young.
I think less than a year ago, Anthropic presented it.
And it's emerging technology.
So now we can see like tens of tens of thousands of MCP servers are online.
The development community is, as usual, excited on something valuable
that brings us to the easier future let's say and
i think everything will be built i mean the connectivity between
ai as a if we think about llm as a And the functionality, everything will be built around MCP standard because it's quite
simple for devs, it's understandable, and it just works, you know.
So we have the proof of work and a whole bunch of different integrations prove that's the best.
Yeah, that's great.
Just a question, and like, I'm kind of semi-technical,
so this might be like a completely stupid question from a developer point of view,
but I think a lot of our audience would be kind of interested
in knowing like, how does this compare to an API?
Because I think like a lot of our audience is probably familiar with, you know, the fact
that a lot of developers use, you know, APIs and allows different applications to communicate
with each other.
Why couldn't APIs be used for this issue of AI being able to communicate with blockchains and other applications?
Look, it's easy answer because API is a part of the whole story. Uh, we need to, for the product to have an API because MCP uses product API to
connect the API data and functionality with LLM.
So like MCP standard is in the middle between LLMs and products API.
So without API, it won't work.
Yeah. Okay, that makes sense.
So I noticed like, I mean,
I haven't developed anything big in a long time,
but I remember like a lot of APIs, they had like
an XML format of different data
and how it's kind of set up.
Does that still happen at the API level or does that happen at the MCP level?
And what is the data? I guess like another way to phrase it is what does data look like at that MCP level. JSON. Yeah, JSON, 6ML.
Yeah.
So it's pretty simple.
Nothing super innovative in MCP standard,
I mean, technical-wise.
But the key to success for MCP standard
is just it standardize the language that LLM speaks to different
integrated protocols.
Because currently, like a very simple example, just go to ChatGPT and ask for TVL on Telus or top five tokens on Telus by market cap.
It will be lost.
It doesn't know that information because it has no access to this huge database.
But MCP solved that.
So because when you ask something specific, okay, to be true, to be absolutely honest,
there is a backend that should determine which MCP server should be used for this specific prompt.
I mean, in our case.
But at the end of the day, for the end user,
when you just ask chat for some on-chain data or functionality,
you don't see what's happening on Backstage.
So it just works.
On the other hand, our goal for next year, in addition to have the product for end users,
which is obviously the chat UI with a whole bunch of integrated protocols,
we're going to open access to our unified API, which will allow different development teams,
protocols, enterprises, access our functionality and our, let's call it brain, right, via API.
Basically, that's our main source of revenue should be.
Yeah, no, definitely.
That's very, very exciting.
And we'll make it like seamless for users to get in and use MCP
like indirectly and not even know about it even.
And to that point, I mean, what will it take for
CatGPT to actually integrate it and people to just be able to ask questions about the blockchain?
What's the highest grossing app on Telos or Ethereum? What are the steps that need to
take place for them to, I guess, like start to index the MCP servers that you guys are providing?
Yeah, I mean, for now, we're more focused on the integrations with Cloud.
So we always publish our MCP servers that we developed
for the integration with different projects on our GitHub.
And you can check, like, we provide a very detailed readme file
with the instructions for even not very experienced developers or users
how to install this MCP server to your Cloud.
And it's just like a couple steps. And then when you are chatting with Claude,
you can choose the MCP server you want to interact with
and use its functionality and knowledges.
For ChatGPT, I'm not sure when are we going to focus on that part,
When are we going to focus on that part, to be honest?
to be honest.
Is that partially because MCP was a standard that Anthropic created
and Claude is built by Anthropic?
Yeah, absolutely.
It's a match on heaven, on the skies.
Do you think ChatGPT or OpenAI will embrace the MCP standard
or are they developing their own standard to compete?
If we think about it as competitors and enterprises,
huge corporations, I bet they're working on something their own.
But I think that MCP standard is growing so fast
that at some point the adoption will be cross-LLN.
Yeah, I mean, it kind of makes sense for them
to integrate with any standard, right? If they aren't already accessing some MCPs,
eventually it's just like,
I just don't see any reason why they wouldn't adopt multiple standards,
even if they come up with a better one than...
Absolutely.
Why not all?
Being absolutely honest with you i i just don't know what's happening with chat gpt integrations because we are currently absolutely
focused on claude because it's more advanced for developers and for mP itself. It's obviously.
Yep.
So I think maybe they even already have some plugins that allow you to integrate MCP
servers, custom MCP servers.
Maybe.
And, you know, I'm going to check it tomorrow with my team.
Oops, Siri's talking. check it tomorrow with my team oops series talking um so i'm curious you were talking about um you
know the eventual integration into or actually you can use claude right now and you can connect
it to your own mcp server correct me if i'm wrong but you i think you were talking about like next
year uh just being able to interact with claude without setting up your MCP server.
Is that correct?
Like where it would access servers that you guys are running?
Or am I mistaken there?
It depends, you know, for what I'm sure 100% that next year,
we're going to review our product for end users,
which is, I believe you shared the video in Twitter before.
So it's kind of, yeah, it's a unique leak of our UI
that we shared with you guys.
So your followers can see how the final product for end users, not final product, but the interim product would look like.
So you choose the MCP server and you interact with its functionality.
But for now, yeah, you can, like, if you're deaf or just an experienced user, you can install any MCP server to your cloud locally.
Interesting.
How do you think this is going to change the way that people interact with like Telos and blockchain more broadly?
Like, will this change the design of the various wallets people use and
how do you think it will sway like the amount of you know UI interaction versus just speaking or
typing into the wallet I imagine they'll integrate like an LLM or something I don't know like where do you see this going and
evolving the way we kind of interact with blockchains I think you answered partially
answered your question uh within your question uh crypto would be when um wallets would start
integrating uh like I don't know voice commands uh like just humanly understandable voice commands
to do some transactions so like uh if you just hold some button on your phone or ask Siri to send 100 USDC to your friend, which is in your address book, and wallet will just do it by scanning your face to confirm the transaction, that would be the next level of adoption.
the transaction, that would be the next level of adoption.
Yeah, I imagine like you say, I want to send $100 to Matt, or
Mike or whoever, and maybe like a prompt will show up on your
screen.
Yeah, just you would need just to confirm it.
Yeah, just showing you because because otherwise it's like if you it misinterpreted and
it's sending a thousand dollars instead of a hundred dollars but yeah of course using facial
recognition and then you can the confirmation step always should be because everyone makes mistake
even the robots so yeah that'll be very very interesting um what what is the most interesting thing you've
seen so far and maybe something that surprised you or you didn't think about that that has come
about from this technology and them integrating uh i so far for for, the most exciting things were the combinated prompts.
So when you make a prompt to AI, and it is using different sources, different functionalities to make an answer or an action for you that
was like as long as I have some background in automatization I'm always
fan of like cutting steps and optimizing steps when you when you are doing something so to make things easier
ai would use different functionalities to answer your prompt or to to make an action in your favor
interesting so and and this opens the um cross ai agent uh uh help you rebalance things or do market analytics and
then the same moment rebalance your portfolio or make some bets or follow the news and
act accordingly right so there is a lot of things going
on in the planet that a normal people cannot cannot do in
parallel, but machine can
Yeah, that makes a lot of sense. That kind of gets into like,
maybe the ability to maybe trade on behalf of users and find
opportunities in a way that we can't or could technically but it would take us like days and
days to program things together and connect it all up is that um what are you seeing in terms
of bot trading and that type of thing.
Can you elaborate your question? Oh, I'm just curious if you're seeing
some interesting uses for like the MCP technology to,
you know, do arbitrage trading or just take advantage
of different like financial opportunities on chain
i mean absolutely there are a whole bunch of different solutions that are pre-ei era so like
just metrics mathematics and the traders logic are are the scripts for them are the scenarios for for those
trading bots uh but with ai it makes everything like faster and easier to um to fit the current
situation and to react to the market conditions if we're talking about the trading bots yeah yeah that makes a lot of sense um just on the one area we haven't really talked about
I mean we talked about like interacting with like LLMs um and how you know you'll be able
to create transactions you'll be able to get information via the LLM,
which is really exciting. But one area we haven't talked about is, you know, software development
using AI and what impact this will have on that. Like I was using something the other day, I think
it was called Lovable Dev, and I think there's one called Replit where you can basically build an application from scratch
just from a prompt and I'm just curious like how will MCP fit into that will that improve the types
of applications that can be built that will it open up easier access to the on-chain data that
allows you know more interesting user interfaces um where do you
see it fitting in with um this ai generated software i mean we are already there uh i
strongly suggest you to try claude to create some basic website or an app. Because you mentioned, for example, Lovable.
Lovable is a low-level development tool
for not very experienced people.
But if you try Cloud, you will see it works way smarter
and as a more advanced developer so we are already there i don't think that
oh for example yeah there is a sigma mcp server which basically makes a front end end from your design so you save a whole bunch of resources and time to transfer to like to get the
front end from from just your picture and uh it's one of the most popular mcp servers i guess so far
if we're talking about web 2 technologies um also one like i'm gonna leak you some um vectors where
are we moving to we're gonna work on mcp servers for uh internet of things and robotics so imagine if you can interact with your i don't know for example
tesla car or google home uh using one one ui which would be like or a voice ui uh where you can just stay your assistant to, I don't know, to hit up your Tesla because it's winter
or to do something at home
and match it with your estimated arrival time at home and so on.
So by combining all the projects or products that have api in one place we solve a very
uh i would say gigantic uh gap in the current ai space so we allow ai to interact with anything
AI space. So we allow AI to interact with anything.
Yeah, that's incredibly exciting. Yeah, actually, the reason I
brought up lovable dev, I actually was only using it for
design. And then I was like, surprised that it was actually
able to build a lot of the back end as well. So I was just
like, shocked by how much uh something that
was i think originally only like design only was starting to actually add the back end stuff but i
haven't tried claude as much as i would like to i've i've used replit which is more definitely
builds out more of the back end which is good but um yeah i need to get into claude
it's i thought it was maybe like more too advanced but maybe not um no it's quite friendly try it
yeah definitely definitely um i will give that a try uh What was I going to say next?
Yeah, it's just incredible where this tech is going.
One question I have is,
how are you guys able to spin up so many of these MCP servers?
It seems like you guys are deploying a new MCP every single day.
Two or three times per week yeah we're deploying two or two or two or three mcp servers per week so like we started uh by using an open source mcp
servers i'm not gonna lie we used like 50 or even more mcp servers servers first just to test them to see what are the best file structure and the code base architecture.
We saw that they are all written in different ways and not standardized.
different ways and not standardized.
So we came like after, I think two months of discovery phase, we came to our own, um,
template, which we are using to, to deliver new MCP integrations.
Um, and, and also Claude, never forget it.
And also, Claude, never forget it.
So it's a really powerful tool
because it can very quickly make a discover
of all the API endpoints that are available
for a specific project, describe it,
and match it with your technical specification
that you prepared for this integration.
Yeah, that makes a lot of sense. Does it help a lot that I mean, a lot of the applications in crypto are like sort of forks of each other. So like Uniswap B3 forks are very common like PancakeSw swap or anything which that's based on sushi swap
which is based on uniswap so they're all very like all the dexes are like kind of very similar
is is that part of the reason it makes it easier to add these things
um yes and no because uh when we talk about the uniswap forks which are basically almost all the
dexas except v3 so um the the solidity contracts are almost the same methods are all almost the same yeah yeah tldr the answer is yes it helps a lot
because uh when we integrate each new evm chain uh basically what we need to is to do the research
where is the um most liquidity uh if it's uniswap or PancakeSwap.
So we need to find out the contracts to interact with,
because they are pre-programmed in MCP code.
But the rest is kind of very similar.
So we just change the chain ID, rpcs and and that's it if if the
functionality within a couple of similar integration is very very similar um when it comes to dexas
same thing you asked about the uniswap and its clones. That's it.
Yeah. And everything else. But our next challenge would be to
integrate some robotics project and IELTS. So I think that would
be more challenging than integrating similar different
on-chain projects.
Yeah, I'm just, I know it's a little bit off topic, but what,
which robotics project are you trying to integrate and
Yeah, I'm not sure I can talk about it right now.
That's okay.
Totally understandable.
Totally understandable.
But what are the main challenges with integrating
with a robotics project? APIs, like APIs are absolutely different, because each specific
project has its own functionality, and its own structure, its own type of data, and so on.
So it's not just analytic data or on-chain activities,
which are basically how many types of transactions could you name?
Let's do it together.
Transfer, swap, add liquidity, remove liquidity, mint, borrow land. So I mean, they're like 20 maximum. And it's typical. Yeah, for robotics thing, it's like endless.
the the optimus robot that tesla is uh building um one of like are those robots like interacting
with other ai and needing to interact with other systems is that sort of where we're heading
and that's why mcp is required to bridge the gap between like all these different things i guess
all these different things, I guess.
You know, as our dev team is mostly Ukrainians,
and you know the situation in our country,
I think the first step of robotics integration would be drones.
That makes sense.
Yeah.
Very interesting.
We should probably start to wrap things up,
but before we get there,
I want to talk a little bit about what you guys have deployed on Telus so far
with the Telus MCP server.
I imagine you guys have deployed this for the core use cases and DEXs and, you know, the top 10 kind of apps.
Yeah.
Yeah.
So where things go from there.
Yeah.
Our go-to-market strategy is pretty simple.
We should cover as much different chains and protocols as
we can um with delivering mvps of an mcp so minimum available product for mcp service so we can
packaging basic functionality into it just to have this chain covered I mean
that the most popular requests and the most popular types of transactions and
later on when we gonna have the real users not just us beta testing our stuff.
When we have the real users, we're going to analyze the data and the traction of each specific protocol or NMCP server
and evolve and make updates according to the usage,
according to the demands.
Also, when it comes to the usage, here
we have the whole economy behind it,
where our token will be involved.
Obviously, all the LLM stuff or all the AI stuff interacting with LLMs having a huge
infrastructure that can make this thing functionable functioning allows a lot of resources I mean money
so we're gonna charge for the subscription we're to charge for API usage and our token will be involved in the whole economy of that system.
And we also going to incentivize third party developers to deploy their own MCP service into our MCP Supergraph, and we're going to reward them according to the attraction
and usage of their developments.
So the idea is very big, and I hope next year
we will have all this big system functioning.
Wow, that's very exciting.
I do have some questions about the token.
But first I want to ask, or just like cover what you said.
So you've basically done the 80-20 rule, you know, at the moment,
deployed all the high demand, small portion, but high demand kind of frameworks,
which totally makes sense uh so my follow-up to
that is like so for instance we've got kudo who's uh joined the chat here if they wanted to add uh
i know what you do you call it a framework if they wanted to add the the data graph or whatever you
call it for kudo related um activities that they're like a gaming
platform can they do a pull request or what's the process for that i'm afraid that i'm not getting
the question oh so if an application on telos wants to um add something to the current mcp
add something to the current MCP.
So then their data in their DAF is covered.
Yeah, we're open to contributions.
Like the code is open sourced on GitHub.
We've also added the contribution guide.
So like any development team can suggest a
function or just send the pull requests and we will review it
and we will update the MCP on our end if we accept the update.
That's awesome. That's awesome. And I guess the smaller question
there was what do you exactly I know you've got the server itself, but then what do you actually refer to like the individual like data structures that like dApps are pushing forward or for each individual dApp? What do you call those data structures?
From your side? Data structures, what do you call those data structures? From your side.
Data structures? What do you mean?
Oh, like the
like, so we were talking about, like, say, Uniswap version 3 integration,
and that's been added into the Telos graph.
Yeah.
What do you call that? Like that component that is like specifically for that
specific data uh i don't know like it's a part of mcp server it's just a okay it's just yeah i mean
that any integration is mcp server for us and that's it. Totally makes sense.
And do you always do like an MCP server for an entire network
or do you sometimes split it off for an individual application?
Yeah, I would mention three levels of integration.
So definitely we build our internal development process to be kind
of parallel for all three levels. But we all understand that the basic level are the chains,
right? Like Telos, like Ys, like Arbitrum and Solana, forget, forgive me um so we need to integrate uh chains then the next level would be
major protocols with major liquidity like indexes borrowing lending protocols and so on. And third level, top level,
are the smaller protocols
with some specific on-chain functionality.
So it's kind of a pyramid
where the foundation are the chains, of course.
Makes a lot of sense.
And so as to the token,
you mentioned, is that coming to the token, you mentioned,
is that coming sometime next year, you think?
Token?
You said there's an upcoming token for your project?
No, we already launched our token,
but all the token batches that are incentivization and community rewards and so on are locked for one year.
You can see it on Team Finance.
And by that time, we're going to be ready with our economic model, with our app.
Awesome.
And so you'll still be able to run your own
servers and do DYI or whatever.
But I think a lot of
users and developers will just want to pay, I guess.
And you're saying that your token will sort of be that payment method in the back end,
whether you're like paying in Telos or Ethereum or whatever.
It's still at the end of the day, utilizing the token in the back end.
Is that sort of where you're heading?
Yeah, it will be one of the payment options with some discounts or pro features, of course.
Awesome. Awesome.
I think that's everything.
So, yeah, what's next for Tyrone, I guess?
Oh, now we are working 24-7 to deliver as much like fundamental integrations as we can uh and we are also working
on on the updated uh branding because we we understood that that we uh started with kind of
started with kind of Toyota level brand, but we want to have a Rolls Royce level brand.
So we're going to reveal our new designs and everything on socials, on website, I think
in one or two weeks.
But that's just like, you know, the cosmetics changes.
But the goal is the same we were building this mcp supergraph and um it's very easy to follow our progress uh because like we just
publish all the integrations on our github and on our, and you can track step by step what are we integrating into
our product. So yeah, the goal, the path is clear, we just need to execute it, we need some time.
Yeah, it makes total sense. And very, very exciting to see how this develops. It's like,
pretty clear to me. and i think probably those that
are listening like how integral the work that you guys are doing is into connecting telos and
everything actually like into ai like without the work that you guys are doing then i i just think like, you know, we kind of, it stops or prevents the progress of with pure digital data, but it's not able, well, it's very limited in what it's able to do in terms of interacting with payments and different applications.
It's very kind of data only, if that makes sense.
I'm sorry, but you are very demanding.
Never forget that three years ago,
we were calling AI like a stupid chatbot
that understands just a couple words from your prompt.
So yeah, I think we're still early and it is growing so fast that in five years we cannot imagine our life or even industries without AI.
Oh, totally.
That's why it's so exciting, right?
It's like the early stages of the Internet still.
But I mean, it's already so damn useful today.
Like I'm using it every day, all day.
Like everything I do, I'm using AI, but it's just like it's at like 5% or something.
Like it's my mom using it.
So everyone. like it's my mom using it so like everyone yeah i mean what where would you put ai in terms of
like if you compare it to the internet what what year are we in or what application are we at like
are we at like ebay dial-up modem level 1995. you know it's more philosophical question and
i'm not a fan of philosophy i'm a tech guy so i'd prefer to skip this question
no no all good all good but the point remains we are very very early
in what very early um but yeah we could go into like scary level hypothesizing about where
ai will go and we'll steer away from that uh because like it is kind of crazy the capabilities of you know when you get into AGI and all that
kind of stuff I think one week ago I read somewhere in the news that in some
country there is kind of revolution happening and the the revolution team
asked chat GPT to analyze the candidates
who can run the next government.
Wow.
So that's a little bit scary, but that's the reality.
People are getting lazy and more lazy and lazy.
Yeah.
I'm actually a little bit scared of how many brain cells are dying
because of the reliance of myself and others on ChatGPT.
You get to a point where initially you're prompting AI
based on a long list of requests, very detailed,
to the point where you're just like,
hey, write an article about X, and it does faz um trabalho incrível. E então agora você está
começando a pensar, se eu estou pensando assim,
meu cérebro vai morrer?
Você viu o cartão chamado Wall-E?
Não, Wall-E, sim. was the name wallie no oh wallie yeah yeah yeah so i hope that's not our future i hope i really
hope so we should do sports and something our hands not not delegate everything to the eye
yeah we're gonna be able to we're gonna have to like sort of, as a metaphor, go to the gym for our brain.
Otherwise, you know, it'll die essentially.
You know, it's that same sort of deal where we don't need to move around.
We don't need to hunt anything.
We don't need to even leave our house anymore.
So we have to go to the gym.
And I feel like we're heading in that direction in terms of our minds, which is kind of scary.
But, you know, we're all exciting technology.
You got to take the good with the bad and we've just got to work it out as we go along, right?
We should rule the technology, not be its slaves, let's say.
not be its slaves, let's say.
Yeah, yeah.
Well, I hope there's people thinking about these sorts of issues.
And I know there are some, but I hope there's enough work being done on just thinking ahead.
And so we can plan for this and prepare for this.
And like you say, control it in a way that we don't create issues uh unconsciously
yeah but um we'll wrap things up here thank you for your time andrew it's been an absolute pleasure
and just so eye-opening listening to you and talking about this tech and the mcp supergraph
that you you guys have deployed and everything you guys
have coming up it's it's very very exciting um looking forward to having more of these spaces
with you um and just seeing how things evolve um because yeah there's going to be things that are
that are coming up out of this i think that will catch us off guard. Just like as we were talking about with the internet,
like we could barely kind of imagine the full depth of its impact
until, you know, we got years and years in.
But applications and use cases and interfaces,
I think, are going to surprise us,
and you guys are going to be a key part in that.
So thank you for working with us at Telos
and thanks for joining us today.
Thank you very much for having me here.
Thanks, Andrew.
A nice day or evening.
And I guess for anyone who wants to get the latest on Tyron,
they can follow you guys on X.
Are there any other platforms that are
great for people to dive into? Do you guys have
a Discord? Yeah, yeah,
we do.
Awesome.
I guess we'll link
your X in the chat
it is, what is it, Tyron AI?
Yeah.
All together. It's just Tyron AI? Yeah. All together.
Tyron AI, if you want to go to their X and go to their website.
It's tyron.ai and all the socials listed there as well.
So, yeah.
And you can also obviously join our community as well.
Thank you very much.
Bye-bye.
