Decentralized AI is Not a Crypto Issue… It’s a Business Issue

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

1. 0:10

   Hello again. I am here today to talk about a war. In a way, it's a war in this space. And honestly, if you'll indulge in a little therapy here, an internal war within myself. as the war of the suits verse the hoodies. And don't let the fact that I'm wearing a suit like tip the scales here. I thought about doing some dramatic wardrobe change, but that would be both cringe and I'd like mess it up a weird wardrobe malfunction. So, I'm not doing that. But

2. 0:43

   there's a bit of bio to kind of illustrate this point of my my struggle of the suit and the hoodie. I kind of live two different lives. And in many ways the AI space lives two different lives. And one of my life is hoodie life. The world of web 3 AI. You can call it crypto AI decentralized AI. What many of us are building. I love this world. To give you some of my credentials here. I for many years have been a writer for Coindesk.

3. 1:17

   Years ago I was writing about oh could crypto and AI work together. At the time it was a very like vague notion of like let's take the chocolate of crypto, the peanut butter of AI and like voila, it'll taste great. We didn't really think through it more than that. I produce and host the AI summit at Consensus.

4. 1:34

   And I also am a host of the people's AI, the decentralized AI podcast. I interview folks like that guy there. You might recognize our first guest on that podcast. And I look deeply at the principles of decentralized AI. We all care about things such as one, a focus on principles. I think most folks here are trying to build AI the right way, right?

5. 2:05

   Do it in an ethical way. Do it in a fair way. And there's often a concern about kind of the little guy versus like big tech AI, right? There's a concern of if centralized AI is built on its current path that could likely lead to oligarchies and a concentration of power. Not that we're not already there today, but even that on steroids. It's a reasonable concern. There's a lack of trust in institutions

6. 2:38

   also reasonable. Again, doing things the right way, a focus on privacy, a focus on data sovereignty. These are all really good concerns. And the language we use is often language like that about principles, decentralization. And my other life though, I'll call it suit life, is much more focused on mainstream AI, talking to business leaders and AI strategy.

7. 3:10

   And it kind of puts me in a unique position to see both these worlds. So again, in a business life, I speak at Davos, at web summits. I host the podcast AI curious where I interview folks like LinkedIn's head of AI, IBM chief scientists, other business leaders.

8. 3:31

   I also lead AI workshops and help companies figure out their AI strategy, help figure out their AI roadmap, do full day workshop, things like that. I only mention this to share with you. I spend much of my life speaking to leaders and enterprise and also business leaders and like super basic and I don't mean that derogatorally. I mean like normcore companies like construction companies, marketing companies, what we would all call quote

9. 4:06

   web 2 companies, they all have a lot of concerns and questions about AI. They have pain points. They have problems. Their problems though, they're more focused on, hey, what's the ROI? Is this worth it? They do have concerns about safety. They're worried about compliance.

10. 4:29

    Their CTO is asking questions like, "Oh, damn it. Will I lose my job if I deploy this? Will I get fired?" Governance, their customer data, IP protection. They want to make sure their customers data is not leeched by and gobbled up by the open AIs of the worlds.

11. 4:53

    I would say there's broadly speaking three main buckets of pain points and concerns of enterprise AI. I'm kind of making this up, but I think it more or less works, right? Bucket one is concerns over leakage and liability, right? Cisco has some stats showing that employees actually are entering in sensitive information about internal processes into LMS that perhaps they shouldn't be.

12. 5:25

    That's the kind of thing that CTO's have nightmares about. Policy violations are occurring. source code is being sent into these large large mainstream centralized AI platforms. This really concerns CTO's. Bucket two, governance, auditability.

13. 5:51

    Can we figure out where this data is coming from? What data can and should be entered? There's plenty of stats that they are seeing negative consequences. Again, they are worried and frankly they should be worried. AI is exciting. AI is also terrifying to many of these CTOs.

14. 6:13

    Finally, bucket three, vendor lock in concern over, okay, even if we do this system, even it's working, how much will we get charged in six months or a year? Are we stuck with these guys? And these guys could be contractors, could be your SAS pipelines, data storage, or open AI. So again, I hear these two different worlds, the hoodies and the suits, and it dawned on me that these are all the same thing.

15. 6:49

    Virtually all of the problems of business suitland AI are similar problems that the world of decentralized AI is racing to solve. It's not a space of web 3 AI over here, a space of Fortune 100 AI over here. It's all the same thing. We're just speaking different language.

16. 7:19

    We won't go through this exhaustively, but just to give you some examples, business AI thinking about, okay, how do we not use how do we not leak data? That's the data sovereignty problem that so many projects in the space are working to solve, right? Web 3 might say userowned data or data sovereignty and business is talking customer IP, audits, compliance. How do we make sure that what this says it is actually is?

17. 7:51

    Guess what's really good at that? Cryptography, blockchain. These are the problems that web 3 is racing to solve. It's different language. There's a great concern in the web 3 space about decentralization. I promise you, this might be unpopular take. I promise you, no business leader in a mainstream world is out there talking about decentralization.

18. 8:23

    They just don't care. They don't. What they do care about though is making sure something is trusted and making sure they're not locked in and relying upon one entity. Again, that could be OpenAI, could be SAS, could be cloud. So any instead of focusing on the language of the principles, we can instead focus more on the outcome, the solutions, the problems we're solving. And again, I know I'm

19. 8:55

    preaching the choir to some extent. I think it's great that NAR is focused in many ways on helping enterprise solve problems. I love the chief commercial officer at Near. That's fantastic. And I'm hoping we see more energy in the web 3 AI space that's speaking enterprise language. So why does this matter? Like what's the point? Why am I why am I talking about hoodies and suits and kumbaya? Can't we all get along and sing the same tunes?

20. 9:26

    I think for perspective, it is worth noting that the risk of being very unpopular here. We use the word web three and web two almost as like a neg using calling the rest of the world web two is like wow it's kind of kind of suggests you guys in your web two land of your you know your heads buried in the sands you're doing web 2 AI like okay fair but the reality is web two is the world and it's not the most perfect metric here but looking at just a market cap.

21. 10:03

    There's the hoodie market cap and there's the suit. If we want to change the world with decentralized AI, it might help us to start talking to the suits. Thank you. This topic is so important to me that I would love now to transform into a moderator to interview our next two guests to drill down in this deeper and to talk about the issue of legitimacy on how for AI and decentralized AI to truly scaled to what extent does it need to be seen as legitimate?