How Founders Build and Scale AI-Native Crypto Products

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

1. 0:10

   energized to be here today. Very excited to share with you some p perspectives from the front lines of VC investing and also product building uh around what we're seeing at the intersection of web3 and AI. inspirational talks this morning by Ilia and the team on this intersection.

2. 0:33

   Uh I'm excited to cover how founders build and scale AI native crypto products. Uh I like to share our takeaways up front uh for the next 10 minutes. Uh I'll make the case that we're in a filtration cycle followed by a dispersion cycle. It's neither bare or bull market but a cleansing of what's going on.

3. 0:56

   This opens up opportunities. It opens up opportunities for web 3 and AI to thrive structurally positioning everyone in this room to create that transition and to build valuable companies. And then thirdly, based on what we're building in our AI agentic systems at Dexic, I share a practical framework for building products that endure through those cycles.

4. 1:24

   Uh, I'll start with the elephant in the room. You know, there's fear. Fear in crypto markets. Crypto fear and greed index is at an all-time low. Until you look at this chart, you don't realize how low it is. It's at 12. Liquidity is tighter.

5. 1:39

   Speculation has cooled. Narratives are repricing. A lot of this is driven uh by this compression of narratives and a focus by investors around fundamentals. That's where we are today. Under the hood, there are talks in markets around SAS apocalyptus.

6. 2:04

   SA Bitcoin is tightly correlated with the NASDAQ downturn, especially in nonAI SAS. Capital is repricing the growth of these SAS companies, SAS companies that have questionable AI differentiation going forward. Bitcoin, does it have an AI story? Who knows?

7. 2:27

   Near obviously is helping us lean into that narrative, differentiating from the pack. And you could see dispersion over time. Again, not fact, not financial attacks, legal advice in the AI AI era. Markets are asking where is that measurable productivity gain? Where's the durable advantage? And that's the big shift happening right now.

8. 2:56

   Near-term we're seeing filtration. What I mean by filtration is filtration removes the excess. Business models are being challenged. Products are being battle tested in time. The the shifts of adoption is happening in real time similar to do when I was a VC here down the road in in Silicon Valley. medium to long term, I believe we're in a dispersion cycle. Dispersion rewards founders who are going to lean in and differentiate their products. Capital rewards these durable AIdriven use

9. 3:30

   cases. That's why we believe I believe this intersection web 3 and AI really matters. Now, thank you Near for having me. Near is well aligned to this dispersion cycle. very inspired by what I'm hearing around intents agents confidential compute a lot of the demos you've seen today that is a positioning that we all collectively need to lean into build these use cases collaborate and then we will get rewarded with this

10. 4:02

    differentiation so a bit more context about me I'm Paul Sue founder CEO and managing partner of Deciconic is a $50 million early stage venture firm focused on the Web 3 and AI super cycle. I was asked earlier today on the podcast, do you really believe this is a super cycle? Absolutely. You know, these are synergistic technologies.

11. 4:25

    When we build our operating systems, we get deep into understanding that synergy and we're bullish. We're bullish long term. Not everyone will survive, but we're bullish. We operate an AI operating system as you see on screen that orchestrates workflows for five humans with 190 191 AI teammates a ratio of 40 to1. We believe in orchestration multiplayer AI.

12. 4:51

    This is how we coordinate our workflows to scale my expertise across our investment workflows for investment alpha. I'm a solo GP. We've cloned a lot of experts in our space to draw on non-consensus perspectives. What you see though is an operating system that's just the surface.

13. 5:12

    Underneath is our expression of a technical architecture that represents the future of a gentic infrastructure. Our technical architecture reflects these various systems. Inputs come into memory. We have our structured memory. It feeds into open router for model switching. It applies the right models to the right context at hand.

14. 5:41

    Market timing matters in our business. So context really matters. It spits out a very fine-tune output that then we feed back with reinforcement learning to improve our future inputs. This architecture is our expression of durability and we believe this is how we collaborate with founders like you to drive that durability.

15. 6:03

    I've had three lessons from building our AIOS. It's been informed by a lot of my do investing, my web 2 experience, building social media games like Farmville Words of Friends. Three lessons really stand out. Memory compounds durable advantages. You know, wisdom in investing gives us separation versus other investment firms. Context creates that operating leverage to apply that wisdom to particular markets at hand. Uh whether it's financial markets or adoption markets. And then

16. 6:36

    reinforcement loops create that defensibility. We have roughly a two-year lead over other VCs. They're just waking up to Open Claw right now. We built our personal AI agents roughly two years ago. I've then applied these three lessons into a framework that I'm excited to share with you guys today. This is how we work with our portfolio companies.

17. 7:00

    Utility, economics, and compounding effects is a quick shorthand of the questions we're asking in our framework. Does AI improve measurable outcomes, augmentation, not just automation? Does the token enhance real economic coordination? SAS is dead. Service as software thrives because it aligns economic outcomes with outputs. And then three, does the system create compounding network effects? Do you

18. 7:31

    extend that lead over time or does your lead shrink? I'm excited that near fits all three of these questions. I'm going to go through each of these three tests for my concluding remarks. Again the first test measurable workflow outcomes. Uh on the left here you have the near intense technical architecture.

19. 7:55

    Near intense introduces this intent abstraction human and AI specify desired outcomes similar to our agentic architecture. solvers. Now th this middle blue box compete to fulfill those intents execution complexity the routing the bridging the signing is abstracted away user never has to think about that what this leads to the outcomes fewer steps lower fail probability higher execution

20. 8:31

    reliability this isn't just UI it's an architectural improvement over uh existing systems. Second test, economic coordination. I love this concept of token back solver economics. Solvers compete in this market structure. Again, it's that blue box. It gives market makers or solvers economic skin in the game, competitive price discovery, and then execution, quality differentiation. You remove these incentives and the system kind of collapse. you know you you're not you

21. 9:05

    you see coordination weaken you you you see execution quality deteriorate the token help secure this coordination layer I described this earlier today this is structure structural utility and then finally how do you compound network effects with usage it's a question we're driving at in our underwriting conversations with founders near just has this very elegant model look more agentic intense leads to performance data. Reputation signals routing intelligence.

22. 9:41

    That data creates this flywheel. Better solver selection, better pricing, routing intelligence. This isn't growth through marketing, through events. It's repeatable. It compounds. It's something investors are looking to underwrite today.

23. 10:00

    Fear compresses narratives. Filtration will separate the weak from the strong. We're in a K-shaped economy right now. Dispersion will reward the different uh the differentiated approaches to product and product market fit. AI will flourish humans deliver measurable outcomes. The next trillion dollars of agent commerce will be in this room. If you're building these AI durable use cases, please talk to me. Yeah, we we love to partner early. Find me on LinkedIn. I'm

24. 10:34

    generally active there. We we also blog a ton on vexic.com. Thank you for having me.