Making AI and Finance Private by Default

Transcript

Okay. Well, it's great to be here. Thank you everyone uh for coming. Let's dive in. For those who are not familiar with me, I'm one of the co-inventors of LLM, the TNGPT. And we started actually near as AI company in 2017. We're a little bit too early. uh we didn't have enough compute to train the really uh large language models but we ended up actually realizing that you do need a blockchain to coordinate to uh establish effectively a lot of the AI systems and so we decided to build the protocol in 2018 and been building ever since. Now some of you may have uh seen near kind of through some of the cycles and so

this is a catchup cycle right we've seen a lot of the kind of older protocols coming back uh with force and so back in 2021 right near was really focused just kind of getting the core blockchain scale make it easy to use make it uh kind of truly programmable but since then we actually been building out a fully vertically integrated stack focused on privacy focused on delivering the experience and the vision that we have where the AI the agents are going to be the front end the way you experience computing and the blockchain the markets are going to be the back end how these AIs are coordinating and uh working together and so this is kind of what I'm presenting today now usually I

talk about all the technology and all the kind of stacks that we build but for the first time I'm going to flip it and I'm going to start with product right so earlier This year we launched near.com. It's onchain exchange, right? This is a single place that you can go and you can trade. You can interact with blockchains and you don't need to worry about gas. You don't need to worry about chains. You don't need to worry about all of the things that kind of we got used to trying to interface with all of the blockchains and DeFi over the past kind of five six years. It starts that you can log in with anything pass key, right? Something you have on your devices now. very easy, you know, touch or face ID or you can use your existing

wallets uh from EVMS, from Salana, from Tron, from Stellar and log in into this experience. So it's fully non-custodial and it allows you to hold assets across crypto, across stocks, across per singular experience that doesn't feel like you're on chain, that doesn't feel like you're, you know, struggling with the kind of the complexity of the blockchains, but really seems like a true consumer experience. Now, the important part is everything you have in Near.com is private. your balances, your positions are all private and I'll talk about technology later. Now, it also means you can trade confidentially as

well. Again, usually when you're on chain, you literally leave all of the information about your every single trade public, right? Everybody sees what all the assets you have and what trades you've executed. This allows to actually have everything private and you can go from any asset to any asset, right? Zcash to near uh USDC to whatever bitcoin etc. And you don't need again think about the kind of complexity that executes underneath. Now we just launched as well stocks right. So this is something that allows you to actually trade over 500 stocks already and we're going to be adding more and more to uh to the product as well. We partnered with Zondo and we're

going to be expanding kind of the set here as well. This feels like a normal you know you're trading with a centralized kind of traditional uh stock exchange but it's available everywhere uh kind of as you're using in a non-custodial ways this product again and everything your position is confidential similarly with per right we partner with hyperlquid on the back end and you're able from the same place you can you know take your long Zcash position you can also uh go and you know buy some uh 20x X long on copper because of course AI built out and uh you know have all of this in one place again the from a hyper liquid position it's confidential on

what is your balance what is your account how much money you have on it right this is something that right now if you try to go and bridge to hyperlquid everybody will see exactly uh kind of what's in your wallet finally you have the same thing with earn right uh if you have funds at test, you can go and uh put them into your uh stable coin product or uh staking across all these places. There's a very nice uh kind of projection that shows you how much you're going to earn and different uh aspects of the different pools. And again, from a pool perspective, they just see money coming in. They don't know who and how much you put in from where. And so this really creates this experience where you can have all of

your money. You can spend, you can pay, you can uh trade. You don't need to worry about how exactly it's done, right? You know, compare this to a traditional onchain way. You know, I have a 100 addresses. I have a spreadsheet. I have like labels on each one, who I paid from which one to make sure I don't like cross reference anything. Then I need to like move the funds around. All of this is effectively gone. And this experience been, you know, catching fire over the past couple uh couple of months. Now if we project forward right I mentioned AI is going to be your experience and so kind of where we looking forward is actually turning this into your agent your kind of agentic home base where you have all of the intelligence running on top of your

uh kind of existing portfolio and existing trades and allowing to actually start executing things on the background. A specific thing we're going to launch in the coming weeks is ability to automate a lot of the activity, right? Think of you know Zcash is booming. You want to uh you know start buying into it. Uh you don't want to manually you know remember wake up to do this. You can just set up an automation with natural language tell it what to do and importantly you can kind of expand that with any set of criteria with any set of intelligence that AI can uh analyze and understand. Now, because this uses all of your private data and it effectively has access to all of your

portfolio, this AI itself needs to be private as well. And so we'll talk about that. But all of this experience effectively is targeted to create something that blockchain haven't seen before, which is completely chain abstracted onchain experience where you're just using you're just explaining, you're just kind of talking natural language and you're not having this barrier that we've kind of got used to with crypto that it's too complicated, it's too hard. This experience has been transformative and we expect more and more new people coming in into blockchain with this. Now behind the scenes, one of the main things that enabled this is our near intense technology. This is what really

enables a cross-chain execution in such a way that you as a user don't need to think about how the assets are going to move, who is going to provide liquidity and really tap in into all the liquidity across different chains. Right? This starts with effectively saying, "Hey, I want this outcome." So that's what we call an intent request. And then you have different solvers or agents who then say, "Hey, cool. I can execute that and it's going to take this much time and it's going to take this much money to do that." And then they can go and tap in into different blockchains or uh as we expanding the set of assets that this works is now across venues, right?

So this is not just trading on chain right across you know Salana base near it's also trading across centralized exchanges because solvers can tap into that liquidity it also trades across NASDAQ. So one of the problems for example with stocks traditionally is that it's a lot of liquidity at rest that's required to bootstrap every single asset. Right? If you want to have 500 stocks, you effectly need normally 500 pools with at least, you know, a couple million dollars each, right? And so now you're effectively into a billion dollars just on itself to bootstrap liquidity for this market. Instead, with near intense, you flip the script. you effectively say hey at the time when

somebody wants to buy a million dollars worth of Nvidia the solver can say cool I'll actually buy it on NASDAQ tokenize it in the moment and deliver it to your wallet right by removing this kind of atomicity requirement that we got used to in blockchain and really moving this into this intent architecture now you can tap in into all of the different markets around the world and really get a union of liquidity across all of them and so this also means you You can have like private venues trading here as well. You can have uh markets globally tapping into this. Now, if you take this one step further, this really is about any contract, right? If we usually talk about kind of smart

contracts on chain here, we're really talking about any kind of outcome that you would really want a counterparty to uh partner with, right? And so the example of this is an agent market where you have effectively a natural language request, right? I need some analysis. I need a code review. I need to buy a uh you know house or I need to um book a venue for an event. And agents can effectively come in and say cool I'll actually execute this. It will take maybe you know a couple days to uh completion. and you establish this intent and then they go and effectively can you know email they can talk to people they can actually execute all of this as if they were individuals or

companies executing this and so the idea is that we're moving away from just you know financial assets to now effectively willing and uh into existence the outcomes we want directly through this experience. So that's kind of the how near intense is evolving as it goes. Now for this really to work right the privacy is going to be paramount right you need to make sure that when you're asking for a specific outcome that that outcome is not going to leak across you know whole market that everybody will be tipped off or that somebody will uh effectively be in adverse position right and try to swoop in and rent the same venue you're trying to rent and so you need privacy on the eye side as well and

that that's what we've been working on the neari side neari is effectively our uh kind decentralized confidential computing platform and it starts really with a the same technology we use for privacy for assets and trading we use to enable end to end encryption for your prompts and for your responses again to compare it with the current world right when you're hitting some API endpoint uh for AI you have no idea what's going on inside you have no idea if they you know recording your data and using it we've seen a lot of news around that you don't actually know if the response is coming from the model you expect. And importantly, there's been studies where some of the routers actually are

malicious and insert malicious Bosch commands into the outputs for your agents and effectively can insert uh and execute malicious code on your machine by you just using some API endpoint. All of this is effectively addressed uh with near AI by enabling an end toend encryption where you effectively have a guarantee to a specific GPU with specific model running on the other side. And so to really get that uh you have with near AI an ability to verify the exact GPU and CPU hardware that's running for your workload. You have uh provenence hash and signatures available for you. And this is example of our playground. You can inspect this and you

can embed this into your software and Venus, Brave and others already using this to really provide this experience to their users. And so this is how we make sure that it's not just your kind of onchain experience but also your intelligence experience stays private. Now coming back to the underlying technology right we've been building Near for over eight years and it's been obviously a long journey. Uh this year we actually released dynamic resharting. This been kind of one of the things we put out on our road map back in 2019 when we established that for blockchain to truly scale to be able to enable this kind of AI and financial world we're

outlining you cannot have a limited capacity of traditional blockchains. And so dynamic resharding really enables to grow the block size as there is more demand coming and continue maintaining fixed fees for the whole thing. Now the other thing we've uh recently published is our postquantum road map. We've released actually an ability to control near accounts with uh postquantum secure cryptography uh MLDDSA which enables you to effectively you know uh replace keys on near account right now you know go disappear for 10 years so put it in a cold storage and not worry

about quantum computers coming in. We released it now as one of the first if not the first one uh available because we want to give the maximum time for the whole ecosystem to adapt and uh kind of implement this. There's a lot of other pieces that needs to be done to be fully postquantum around cryptography of the consensus and uh our chain signatures technology and if you want to learn more about this I'm going to be kind of talking in more detail tomorrow. Now, one of the other key components that we have is called chain signatures.

This is actually a multi-party computation network that enables a way for near smart contracts to sign transactions. And so this is again the first time you can actually have smart contracts sign transactions, not just kind of external wallets. And this is what enables near intents and a lot of other uh really cool use cases to be built. And so this is just an example of kind of visualizing how this network is operating um around the world. Finally, I mentioned right everything is been moving into privacy and so to enable that earlier this year we launched confidential shard right this is the kind of enabler for running your balances your trading your earn your

stocks all confidentially and uh it's all done effectively as uh for now a single shard and we're going to be expanding that as well and so all of those pieces together really to enable a scalable blockchain that is proof for future usage both on scale and on uh postquantum and it enables you to really program not just for near blockchain itself but really for any blockchain right your near smart contracts can actually interact and transact with uh any other smart contract around any other chain and all of this is being kind of packaged into near intense which is this kind of market of markets infrastructure now importantly I mentioned agents are going to becoming

more and more uh of users this already and uh we're going to be launching another kind of uh piece of infrastructure for them we call agents API that enables agents you know right now you would give them a key they can sign intents they can already transact and do actions but you really want them to have a little bit more of a policy of how they can execute right imagine you hire someone to manage money you're probably not going to give them you know okay here's all the money go you know uh at it you would say Hey, look, here's the restrictions, here's the limits, and you know, if you want to go outside of them, come back to me and get an approval. And so, this actually enables this exact API plugging in directly into

your private intents into your private balance and enables for you to plug in either your agent or your kind of the agent from near.com to uh execute and interact with it directly. So all of this pieces together just kind of to show you that the the customer product right that we're building is based on a massive amount of infrastructure to enable a smooth experience and we're going to continue enabling and developing this for the world where your products are becoming more and more agentic. Right? The previous speaker was talking about how the financial advisory itself makes sense to be fully agentic and actually execute all of this and some of this missing missing pieces are really being these types of uh kind of

back end to enable it and all of this going to drive more and more adoption for near itself right and so this is just to give you an example of where we are already over kind of uh last I guess uh year and a half since we launched launched the near intent it's processed over 33 billion in volume I think last months was over 4 billion on itself and we see continuous growth uh coming through that it's an infrastructure that is available for everyone to tap in and to develop right it's used across lots and lots of different existing applications in web 3 [snorts] and then as we launched the confidential shard and kind of been enabling more and more features through it we've seen a massive

growth as well of the money moving in into confidential shard to store confidentially to have the kind of trading and balances all uh being private. And so really exciting to see kind of the uh like the privacy is not a meme anymore. It's really now a product that everybody can use and it's really experience that is compatible to what you normally have with centralized exchanges. Now, just to finish off, uh, we have a referral program. And so, if you're familiar with kind of the referral programs of before, this is a really great opportunity to actually bring all all the people in web3 and outside to actually experience this,

right? And so I'm encouraging everyone to uh to kind of tap in yourself into.com and bring all your friends and family to start using this experience versus still trying to kind of manage all the different um kind of onchain experiences separately. Now if you are a builder, you can actually build your own.com. You can build all of those pieces yourself. You can uh run your agents off it. And so there's an API on our docs that you can tap in as well. So, thank you very much and let's go private.