The Unified Commerce Layer for the Agentic Era - Illia Polosukhin

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:11

   Okay. Hey everyone. It's great to see everyone here. It's an amazing day for us to come back to San Francisco with Nearcon. Uh we started uh in San Francisco with Nearai. We then went global and uh fully distributed and then we've been bringing back more and more of our attention to here as AI is growing. [snorts] As Chris mentioned, we are entering the agentic era. We are entering a world

2. 0:41

   where AI is becoming our interface to compute. And so we want to talk about how near is providing a unified commerce layer for all of us to really benefit from all of these advancements. We have seen a speedrun through a technology innovation. the AI technology went from a backend feature in your search and recommendation right in your Googles and Amazons where it was you know something that was hidden from the user and really just provided you results to a chat application right

3. 1:21

   launched in 2022 that would give you some answers but you still needed to go and act on them to something that we've seen over last year where the this AI systems are starting to take action they becoming agentic. they're able to go and actually do things to something that we're starting to see as an emergent property where everybody now has an agent and this agents are starting to work together to do bigger things and so

4. 1:49

   have this multi- aent coordination and as this happening AI went again from just being a feature to something that is in front of the users and in front of people and becoming their direct interface to all of this computing. And then as they become and talk to each other, you know, my agent is talking to your agent, they're becoming economic actors. They're becoming effectively uh negotiating and figuring out how to do things together. And they require economic um tools. And so this

5. 2:22

   means that commerce itself, right, the way we do economic activity in the world is changing. It's compressing, right? where traditionally we would have a ton of different applications that are fully siloed that have different information about you different context and you need to switch between them right your Amazon doesn't know about your eBay purchases doesn't know about you know your Google

6. 2:45

   to you know different tabs when you're switching it to filling out checkout forms and credit cards all of this is starting to transform where your agent has all of your context and is able to go and do all this actions across all of the platforms if needed to really get results done.

7. 3:02

   You know, if you're like me, you have lots of messages and emails, there's invoices, bills, you know, company, personal, all of this like coming in and all of the different directions before you handling it manually. You're figuring out it may you know things may go stale, things update in business case, you know, it's a whole process.

8. 3:22

   Again, this is something that agents can consume all this information, prioritize and execute. And with crypto right we've kind of with crypto had a promise that it will solve a lot of the challenges that traditional finance and traditional commerce have the faster transactions cross border all of the promises that we've had but historically they've been highly cumbersome to use. It's been

9. 3:47

   really hard for normal people to use crypto and although you know as near we've been trying to solve this uh it still continued to be hard because even if you're on near you still want to use other applications other chains other assets so all of this is starting to compress now into your AI agent being this interface that now can facilitate economic activity across the whole variety of things on internet and so that's what we call intent driven

10. 4:13

    execution that's really what commerce is becoming. Now, I recently moved from Lisbon to San Francisco and I have a family. I have two really small kids and my wife uh beautiful Erica and and it's a it's a very stressful as you can imagine with two kids under two um to do this and I thankfully have you know amazing Maggie to support me to do this but not everybody does and so I want to share an example how

11. 4:47

    this looks like when you actually have this agent on your side right? You have this intent to move uh to San Francisco and so you leave effectively a message to your agent and it takes from there it creates a full plan. It's able to figure out exactly all the actions some of the things that you know I forgot to do for example cancel my utilities and it figures out where to move it finds the

12. 5:10

    right uh neighborhoods and asks you questions to clarify where you want to go. It creates a a plan on the go, creates a user interface for you to track it how it's going and then goes and executes this on your behalf. Now, your agent will not be able to do everything, right? In this case, shipping boxes internationally is a very complicated process, right? Everything from picking up from your apartment to

13. 5:33

    do to getting it uh shipped to customs, etc. Usually, it would take lots of steps, lots of uh headache. Now if we have an agendic marketplace where agents can actually hire each other uh it doesn't matter which currency they're paying you're just able to pay in my case in near and then get it all handled by other agents in in this ecosystem. So this is kind of the idea that you can

14. 5:59

    have your agent working on your behalf transacting and interacting with other agents and getting things done really easily so you don't have to think about it. So for this to happen agents need a new infrastructure. They need the agent native infrastructure that is optimized for how they operate on internet right so far internet being really designed for people to click and browse to fill

15. 6:26

    out the forms. It's you know how many times you've clicked on I'm not a bot uh on the web page. It's really again this fragmented execution experience where you know if you trying to do something you need to go and visit five different websites compare results maybe fill in one thing to another and we have this challenge that blockchain again it's had the great promise but it hasn't delivered on it because everything you do on chain is transparent and so if I'm you know

16. 6:59

    booking my uh place in San Francisco or uh paying for moving. This is not something that I want everybody in the world to know uh and and see who am I paying for the for this transactions. So transparent transactions that blockchain historically been doing is really not um realistic for usual use cases for day-to-day usage. And so we need a new intentbased APIs. We need agents to be able to use this without needing to go

17. 7:30

    around, you know, Cloudflare and other kind of anti-bot protections. We need neutral rails that compose to each other, that work with each other and not really try to silo. And we need a confidential and AI and execution to make sure that all your data stays yours, stays private, and really doesn't leak all this information to everyone.

18. 7:52

    This is what near is right. This is what we call uni unified commerce layer where we bring all of the pieces together where AI and intents converge into new agentic commerce. This is where agents act on your behalf on behalf of businesses on behalf of governments and they bring all of this assets all of this resources all of this action together.

19. 8:14

    Now we're going to release few more few products today. The first one is something that uh we've been working on for a long for a long time to bring forward and this is something what what we believe will become effectively a super app for all your financial and over time all your needs. So we're releasing near.com we're introducing it first time. This is a place where you don't need to think about uh blockchains. You don't need to think about gas fees. You don't need to

20. 8:46

    think about um you know complexity of blockchain. You can just use it as your main wallet. You have all your assets you know have bitcoin zcash USDC etc. And uh you can use it very easily and it's available today. You I will let Alex shenko to dive in more about all the functionality that this packs. There's a lot more on the road map as well uh including fiat and really ability to pay for anything day-to-day.

21. 9:21

    Now I mentioned it's great to have your crypto wallet that is chain abstracted that removes this fragmentation that crypto has created but we still see everything that's happening on chain right uh you know personally I have hundreds of accounts effectively to dellink all of my transactions right every counterparty I'm transacting with I need a new wallet to do that and this is extremely cumbersome I have like a whole suite of scripts to just do that now today we also introducing a

22. 9:51

    confidential mode. Confidential mode is ability to have all your balances, all of your transfers and all of your trading fully confidential. This means that nobody outside of uh kind of our security system is able to see how much money you have and how much and what transactions you're doing. This enables you individual to hold your money there and it enable agents to transact from there as well. Businesses to accept money and not reveal all of their business intelligence to everyone.

23. 10:27

    Now we talked about agents. One of the challenges with agents right now is that they're extremely insecure. This has been, you know, if you're on Twitter in this AI bubble, you see people getting hacked all the time, their information uh getting leaked. It's actually creates a really bad rep for crypto because most of the hacks are promoting memecoins and crypto scams. And so this is one of the

24. 10:51

    challenge for us to fix is actually to bring AI that's built with blockchain mindset, secure by design, adversarily protected, ensures that your data doesn't leak, your credentials don't leak. There's no way for a third party to really infiltrate your AI agent. This is what we call user own AI. It's AI that's on your side and not easily manipulatable. Now, this is only an alpha, I would say. Um this is you know for those who want to be on bleeding age. Uh we have ironclaw. This is effectively delivering for that vision.

25. 11:23

    But you know tread carefully. We're going to be releasing 1.0 uh in coming weeks. Now this comes as a you know you can run it yourself on your own desktop on your own computer but this is not a very user friendly and I again for those who've been in the AI community you've seen people been going crazy at buying their own Mac minis so they can install this agents on them so they can run 247

26. 11:53

    again thousands of dollars takes a long time to actually buy them now because there is a shortage and So what we're releasing today is using our confidential inference and hosting the neari cloud to host all of the personal agents. You can host openclaw and ironclaw. The starter pack is free, right? You can start it today and you just pay for your inference or there's a basic plan which includes uh a bunch of credits for inference and a pro plan for $200 uh for a lot more juice from this.

27. 12:28

    you can use it and you can effectively launch your personal AI agent and start having it do help you with work offloading tasks and as it gets smarter as it evolves as ironclaw increases in its capabilities you'll be able to leverage it more and more in your day-to-day life.

28. 12:47

    Finally, I mentioned we have not everything your AI agent can do. It doesn't have some context. It doesn't have some skills. It doesn't have some access. it sometimes you need to actually do something in physical world and it needs to actually call another uh organization to do that and so marketplace is really bringing this concept of intents and concept of AI together. This is a place where agents

29. 13:12

    can express a natural language intent the goal you want to do and then find a counterparty find another agent that will actually go and do it. The cool thing is this is unbounded. you can ask it to build a space station on Mars if you have money to do that. Uh but you know to start with you can do marketing design development uh and for for the nearon we have launched a set of

30. 13:36

    competitions for agents to build uh applications and so there's a number of competitions with a thousand near prize uh for your agents to try it out. So, I'm really excited because this actually brings all of the vision we outline at Redacted end of 2024 to really showcase how agents are going to start doing the work on users behalf and then coordinating with each other to really create this larger system and it uses near token at the core. It uses it for reputation. It uses it for uh unit of exchange.

31. 14:12

    And so all of this really is powered by the unified stack that near been building, right? Obviously, blockchain has been around since 2018. We launched it in 2020. We showcased in December last year that it can handle 1 million transactions per second. To give you an idea, 1 million transaction per second is all of the commercial activity combined of all of the Visa, Mastercard,

32. 14:36

    Stripe, and I think like 5x that. That's what 1 million transaction per second. We can handle all your agents transacting no problem. This brings confidential execution to power the confidential mode. And I'll let Bowen talk about more, but this is all outlining how Nightshade V3, the next version of our core protocol, will look like. intents have been a massive enabler for the ecosystem to really unite all of the different chains and

33. 15:03

    assets. It has really been um like the you know one of the few products in crypto that achieved PMF uh over the past few months and it's been processing about three billion uh dollars of volume a months bringing effectively all of the different users and assets into one place. This is what powering.com. This is what powering market uh for agents.

34. 15:28

    Finally, confidential AI. This is something we released in December offering an ability for anyone to use AI and not worry where's their data is going. This is powered by decentralized GPU marketplace and allows you to have verifiable computation and verifiable uh and private uh agentic workflows.

35. 15:48

    All of these will be covered by Bowen, Alexenko and George respectively. Now at the core we have the near token. Near token is how we really coordinated this whole system. As of today, we're gonna enable intense revenue to start uh being captured in near meaning everyone who uses uh intense infrastructure is going to effectively by near through the through their transactions. Now it's going to be distributed to third parties who are providing the order flow and some of

36. 16:26

    this will be captured by the uh by the near ecosystem as well and we're going to be talking and tweeting about it uh in a few hours. Near.com as well will be all the fees are going to be going directly into Near. So near becoming a place where the revenue is starting to flow. Same for GPU marketplace as it kind of expands and a gentic marketplace as well. So you can think of this all of these marketplaces they really all are just different types of assets in this

37. 16:52

    intent based economy that are starting to produce near uh fees and drive near token. Now as we bringing more of this autonomous agents to market the reputation has become really important. The marketplace actually becomes a place where agents build their reputation.

38. 17:12

    They build their track record. They build their history. the governance right we have house of stake that is now coordinating the ecosystem and capital are starting to converge onto near token so near is becoming this core of the ecosystem and so importantly you know if you've been tracking near we've been very technology focused we've built a protocol we've been building a lot of different tech at this point the tech part you know obviously there's still

39. 17:43

    more work to do but we have it we have the stack, we have all the components, we have the products now that really bring this now to the user. And so we're switching from kind of our traditional next technology milestone to how we actually scale the adoption. How do we bring this now to billions of people around the world? Obviously for AI, we're going to start here, but for intense, we're already global and

40. 18:07

    already working with a lot of partners across all chains. We're going to continue expanding this ecosystem. This is an audible shift again where we usually being kind of a technology first to how do we actually make these products really adopted and importantly this is technology that works for you right it's really about how we make everyone have more time and more leverage on their time more you

41. 18:32

    know less stress uh in your life how do we reduce the friction and silos of different information that you're dealing with right now and how do we optimize for your success that concept of user ownership where the AI is working on for you where the assets are staying in your control the data belongs to you and it's your choices not somebody else's this is really what near vision has always been about how do we empower everyone and the technology is

42. 18:58

    the medium to do that but it's now for you to really bring this together I want to thank everyone this is going to be an amazing year for near huge team putting this production thank you very much we have a huge team of engineers and product people and go to market and marketing that actually bring this all together. There's an amazing leaders who going to be speaking about each individual uh component of the stack uh with George for neari bow and near

43. 19:26

    protocol both today and then Alex Shanka with near intense tomorrow. I [snorts] encourage everyone to come and listen to them as well and I want to thank you and uh uh please enjoy con