Adam Sternbach, VP of Legal at Yuma AI, joins Kyle Lawrence and Moish Peltz to explore the rapidly evolving intersection of AI, blockchain infrastructure, and decentralized systems. Adam explains how Yuma AI and the Bittensor network are building decentralized systems for AI training, inference, and autonomous applications, while exploring why crypto and AI are becoming increasingly connected technologies.

The conversation also dives into AI regulation, market structure legislation, consumer protection concerns, and the growing challenge of governing technologies that are advancing faster than policy frameworks can keep up. From AI agents and liability questions to education, autonomous systems, and the future of digital infrastructure, this episode offers a thoughtful look at where AI and blockchain innovation may be headed next.

Subscribe to Block & Order to learn more about the latest news on the legal side of digital assets!

Follow B+O on Our Socials:

Twitter: @BAndOShow
LinkedIn: Block & Order
Instagram: @blockandorder
TikTok: @blockandorder

Chapters:

0:00 – Welcome to Block & Order
01:20 – Introducing Adam Sternbach and Yuma AI
04:00 – What Bittensor Is and How the Network Works
08:10 – Decentralized AI and Real World Use Cases
12:00 – From Government to Crypto and AI Law
16:40 – Why Lawyers Need to Understand the Technology
20:10 – AI Infrastructure Blockchain and Autonomous Systems
24:30 – Explaining Emerging Technology to Regulators
29:00 – SEC CFTC and the Future of Market Structure
34:20 – State AI Regulation and Consumer Protection Concerns
39:10 – AI Agents Liability and Autonomous Decision Making
44:00 – Education AI Risks and Final Thoughts

Watch or listen to the podcast here:

   

Transcript:

**This transcript has been prepared automatically by AI and may contain inaccuracies**

Kyle Lawrence [00:00:12]:
It’s really interesting that a minute ago. What happened?

Moish Peltz [00:00:16]:
I think you start. I don’t know, maybe I’m lagging. Go ahead. Sorry.

Kyle Lawrence [00:00:21]:
A perfect microcosm for the kind of thing I was going to talk about, which is you are venerable guest, Mr. Adam Sternbach, VP of Legal at Yuma AI. You bemoaned the fact that you had trouble printing something. And now as we press record, Moish bemoaned the fact that his Internet was lagging a little bit. So here you are at the forefront of crypto and AI In Moishe and I, we have a burgeoning AI in crypto and tech practice, and we do all of these wonderful things. But you can’t print. Moish’s AI isn’t working. And my printer up here tells me for years that my stupid cartridge is out of ink and it’s not.

Kyle Lawrence [00:00:55]:
So why is it that we can advance as a species and do all these really cool things, but the moribund everyday tasks have not gone by the wayside? If you can explain that to me, you have unlocked the secret of life.

Adam Sternbach [00:01:07]:
The robots aren’t here yet.

Kyle Lawrence [00:01:10]:
They skip that part, I guess. Exactly right.

Adam Sternbach [00:01:12]:
We need to accelerate the humanoids.

Kyle Lawrence [00:01:14]:
My doorbell doesn’t even work. For goodness sake. Like what’s happening.

Adam Sternbach [00:01:20]:
You and me both promised all the solutions with technology and it’s just more problems.

Kyle Lawrence [00:01:26]:
That’s really what it is. We’re moving too fast. I think this is the case for us moving too fast. We have to solve the little problems first.

Moish Peltz [00:01:33]:
We’re accelerating past all the things that really matter.

Kyle Lawrence [00:01:37]:
That’s really true. But that is a good way to introduce who I already introduced as a venerable guest, Mr. Adam Sternbach from Yuma AI. Why don’t you say hello and welcome to Block and Order.

Adam Sternbach [00:01:48]:
Hey, thanks for having me. Excited to chat. Fan of the pod. So I am the VP of Legal at Yuma Holdings. We are a contributor to the Bittensor Blockchain. It is a crypto and AI blockchain network. Happy to dive in more. Yuma helps to run infrastructure and support teams building and developing on the network.

Adam Sternbach [00:02:11]:
Bittensor is a rabbit hole of rabbit holes. It is far and away, in my estimation, the most complex blockchain network around, if not technology. Product and primitive. But happy to kind of dive into it, dissect it a little bit. But fair warning to everyone, there’s a lot of, you know, Claude Chatgpting, Gemini. Pick your AI model of source conversation likely to follow this one.

Kyle Lawrence [00:02:37]:
Nothing wrong with that. And I think that’s a Great way to start us off if you can delve into a bit about that. And it’s worth noting that bittensor is doing all right these days compared to everything else especially.

Adam Sternbach [00:02:48]:
Yeah, I mean, it’s a fascinating network. The way I kind of describe it shorthand is it is a coordination layer meets incentive mechanism for distributed work to be done to kind of further AI infrastructure and applications. So it has blockchain based incentive mechanisms in order to get people to contribute to a network, not dissimilar from any kind of proof of stake or proof of work network where, you know, emissions of tokens are what incentivize people to contribute resources. And then there’s a subnet layer largely on top of the blockchain, where companies build applications and products in the form of subnets, which again is kind of its own coordination mechanism at an infrastructure layer to allow distributed actors around the world. Again kind of global permissionless in the way that blockchain is to contribute computational resources, other types of infrastructure models to really generate an output in a product and service that then can largely be used. I mean, there’s a variety of use cases for subnets, but in some type of kind of application or product that gets commercialized.

Moish Peltz [00:03:57]:
So what sorts of use cases lend themselves to a distributed AI network and distributed AI compute versus using anthropic or OpenAI?

Adam Sternbach [00:04:10]:
Yeah, so I think you do have decentralized models that can be trained on the network, which is a pretty fascinating development. There was recently a pretty high profile 72 billion parameter model that was trained using decentralized bittensor nodes or contribution to the bittensor network and nodes that were contributed by miners on the network. So you can train models in a decentralized manner. You can have decentralized inference, there’s different subnets and operators that are supporting that. And then you get different applications. So that’s kind of on the hardware side, you get different applications more at the application layer, which again, kind of contrasting it with crypto, you see a lot of infrastructure projects, you see applications building on top of them, and you kind of get a little of both within bittensor, depending on what a team is building. So the hardware examples I gave some of the other applications, we have decentralized science research that are the center of certain subnets, vision and AI learning models, robotic training models, financial services, compliance tools. I mean we can dive into kind of any of these, but you begin to get a flavor of all of the different real world applications and use cases that are built on top of this network.

Kyle Lawrence [00:05:27]:
How did you first get into this space because as practicing lawyers I’m always curious as to what represented the shift in sort of career trajectory for you.

Adam Sternbach [00:05:39]:
Yeah, it’s a good question. So I came across crypto initially, I don’t know, 2012, 2013, just like thought it was fascinating and interesting to technology. I’m sure it was coinciding with a bull market. So like, oh my gosh, this stuff’s all going up. This is the greatest thing ever. Stepped away for several years, always kind of like followed it, but never closely. And certainly it’s kind of ironic, felt like I was at a point in my career where it was way too risky to ever like think about this full time. Despite the fact that in law school I actually did a little bit of work related to like some of the earliest bitcoin related analyses.

Adam Sternbach [00:06:23]:
Now with the benefit of hindsight, like that would have been the absolute best time because I had quite literally nothing to lose at that point if this whole thing did blow up. But live and learn I suppose. So I always had my eye on it and kind of followed it loosely. But then I went into state government which was a bit of a unique pivot. Worked in the New Jersey Governor’s office in the end of 2018 into 2019. And frankly I think that path prepared me very well for kind of what I have done since, given how much the regulatory policy stuff has been front and center for crypto, blockchain and now AI over the last few years especially and so came out of government. In government I worked on economic development related stuff primarily. And in the state of New Jersey they actually stood up.

Adam Sternbach [00:07:11]:
What’s the easiest thing to describe as a venture capital fund? It’s really like a co investment vehicle, but it was a state backed venture capital fund. It was like a pretty innovative flag planting initiative in the Murphy administration, which was the administration I worked at to kind of like signal New Jersey is the innovation economy, quote unquote, which was kind of always the governor’s economic messaging back then. And that opened my eyes from like, what is innovation? How do you like grow an economy? And got me into like this is how companies actually get started. And so from there I went to work at a New Jersey based law firm that had a general practice but did emerging company and VC work? This was, you know, 2020 ish, let’s call it. And then as you know, cycles rejuvenated, there were a lot more people knocking on the door being like, hey, do you guys do crypto work? And because I had a personal interest I got pretty deep into NFTs and top shot. Kind of typical 2020, 2021 onboarding. DeFi began to kind of be the crypto person within the firm. And then, like, as I just went deeper and deeper and was like, I really enjoy this and think, you know, here for the tech.

Adam Sternbach [00:08:25]:
Decided that there was an opportunity to move in full time. And that’s kind of what in 2021 allowed me to, you know, for better or worse, on any given day, make the leap full time.

Kyle Lawrence [00:08:38]:
You said the magic top shot word. I’m surprised Moishe didn’t fall out of his chair.

Moish Peltz [00:08:42]:
No, we’ve been talking about this now for five years.

Adam Sternbach [00:08:44]:
Yeah, I mean, Moishe and I go way back into the NFT days as two of the few lawyers who I think are still rocking the board.

Kyle Lawrence [00:08:54]:
I saw that on your Twitter.

Adam Sternbach [00:08:56]:
Yeah, yeah. I have no intention of changing it, but it definitely crosses my mind from time to time. It’s like, are we still doing this? But I guess we all.

Moish Peltz [00:09:06]:
You gotta find something you want to change it to, right? It’s not about getting rid of it. It’s like, all right, what do I want that excites me more than this?

Adam Sternbach [00:09:12]:
And frankly, like, there’s still a nostalgia from that year in that period that, like, I just have fond memories of. And obviously a lot of people think it’s silly and whatnot, but, like, it still holds a nice place for me.

Moish Peltz [00:09:25]:
So I absolutely agree. And I think that that feeling of this, this new thing that didn’t exist all of a sudden springing up. I get that feeling now sometimes on different days with AI, and I was even reading last night and this morning about the new anthropic model and how it can, you know, basically it finds zero day attacks in, like, every operating system that nobody in the world has been able to know about or crack until now. Oh, this is a fundamentally new thing that didn’t exist two days ago. That’s pretty cool. It’s a kind of fun time to be on the frontier here, I guess.

Adam Sternbach [00:10:03]:
Yeah, I mean, look, I think like you said, there are a lot of intersections between blockchain technology, emerging AI technologies. And I feel like a lot of people who have been doing this kind of work related to just like, autonomous software systems at large are incredibly well positioned to think about what the forthcoming challenges will be of agents spawning agents and, you know, robots walking down the grocery aisle with us, which, frankly, I don’t think is like, that far away. My wife asked me the other night, like, when do you actually think we’re gonna have a robot, a robot in the house. And I was like, I don’t think it’s that far, like a few years, like it’s really happening and it’s coming and like it introduces a whole new set.

Kyle Lawrence [00:10:44]:
I mean roombas are the first step and I don’t think that like we have those and you go to a restaurant and now they have the little robotic caddies that deliver food to your table. I mean that’s everywhere now.

Moish Peltz [00:10:54]:
Yeah, I got my dumpling sliver last week from, from a robot. That’s true.

Adam Sternbach [00:10:59]:
Yeah, like self driving, you know, fsd, like that kind of stuff. Like it’s here and it’s only accelerating. So fun times.

Moish Peltz [00:11:09]:
That’s a great point. So, so lead us through then that transition from going in house working in the blockchain industry. As for blockchain and tokens and NFTs and then. All right, let’s, let’s transition to an AI. What was the thinking there?

Adam Sternbach [00:11:24]:
Yeah, so I think I always kind of had an idea of going in house. I think generally I’m a little bit more. So I’ll give the example. I would do venture financings when I was at a firm and you work very closely with a client for, you know, four, six, eight weeks, however long it takes. And then you don’t hear from them for a couple months and then they’re like, hey, we hired a few people, we need to issue some options. And it’s like, but hey, how’s the business doing that you just raised all this money for? And we just worked really hard and like, give me the update on like what you guys are actually doing. And so I think, you know, living that for a little bit, I was like, I’d actually like to see kind of the inside of what that is. And I think I always kind of had a general lean towards being in an in house role and like collaborating more closely and kind of business outcomes and stuff like that.

Adam Sternbach [00:12:14]:
I mean look, I think different lawyers have different profiles and stuff like that, but for me in house kind of always felt a little bit more like a translation and understanding of the operations in ways that kind of fit with what I was looking for. And so was fortunate in my first role to kind of have like pretty broad discretion and scope to run and stand out and figure out a legal structure. There wasn’t much being said about NFTs. The first company I went to was an NFT startup. Wasn’t much being said about NFTs except that fractionalized NFTs were likely securities and we were doing on chain fractionalized nft. So that was also kind of a fun. Yeah, exactly.

Moish Peltz [00:12:54]:
Turns out not everything was legal. You just didn’t know it at the time.

Adam Sternbach [00:12:58]:
Totally. It’s like they’re not even talking about this except the one narrow niche of the product category in which we operate. So that’s great. And so I think that gave me the introduction to, like, these are new and challenging and messy issues that, like, you’re not running the playbook that’s been run for 25 years of how do you navigate these issues? Like, there is real thought and creativity and technological understanding that has to go in to how to grapple with the challenges of emerging technologies. And so that was kind of always in my mind. I went into a next role at Polygon Labs, which was like, great team, great organization, great operation, really learned a lot, got much more immersed in, like, blockchain technology and infrastructure. Had some great colleagues there who really helped me get up to speed. I pretend and say, like, I’m a fake engineer now because I know enough to like, you know, know how this stuff works at the level I need to, but obviously not a true engineer.

Adam Sternbach [00:14:00]:
And then from there, I think as crypto and AI started to emerge and I was beginning to think about what and where to go, I discovered the Bittensor network. It was fascinating to me as, again, just like a pure technological primitive of how do you coordinate distributed work amongst a variety of people to do kind of purposeful and intentional creation, if you will, and saw it, paid attention to it, didn’t give it a ton of mine. Like, I mean, there’s just so much noise in this industry that, like, if you dive down every rabbit hole, like, you’ll never have anything else or time to do. But Bittensor, like, kept popping up, AI kept popping up. And I was like, there’s something here. And so, like, I spent a good few months really, really, really immersing myself in like, what’s the network, what’s the technology? What are they trying to do? And then got introduced to the team at Yuma, and, you know, kind of one thing obviously led to another. Here we are. But I think it was really just based on, like, a fundamental premise of these are two incredibly compatible technologies.

Adam Sternbach [00:15:11]:
There are real world applications to be built leveraging this network. And the same kind of issues that always intrigued me and have interested me about working in crypto and blockchain are only, you know, a little bit different, but very same flavor with AI. And then the convergence of those two was a great fit for kind of what I was looking for and what my background was.

Moish Peltz [00:15:32]:
It’s a fascinating story of just kind of pursuing your interests and what, what, what, what makes you curious and then yeah, leading to the next thing. Right.

Adam Sternbach [00:15:39]:
Look, I’ve been incredibly fortunate in that respect where like I’m reading about these things and trying to understand the technology and also getting it on a day to day. But I frankly think like it’s very hard to be a lawyer in any forward technology space unless you understand the technology and like use the technology and like conversations I’ve had and I’m sure we’ve all had with regulators or practitioners or just like anyone who’s interested. Like there is a completely different fluency and discussion when someone has struggled to solve the non in their metamask wallet or whatever wallet we’re using today and been like what the hell am I even doing here? Relative to someone who’s like yeah, I’m a blockchain lawyer because I did, you know, one tradfi financing or something like that. And I’m always happy to speak to people and like try and help them understand it. But like the first thing I say to them is like load 20 bucks on a wallet, go buy a token, go, you know, spend some time in defi land. And like that in and of itself will unlock anything I can say to you in ways that just like the way these systems work, begin to educate you in and of itself.

Kyle Lawrence [00:16:44]:
Coincidentally, that’s exactly what Moishe told me many, many years ago.

Adam Sternbach [00:16:49]:
Is that true?

Kyle Lawrence [00:16:50]:
We are so see that? The birds of a feather.

Moish Peltz [00:16:53]:
We were at the company barbecue. I’m like, hey, we’re getting a lot of crypto work. What do you know about this? And he’s like, what?

Adam Sternbach [00:17:00]:
Yeah. What was your first impression experience like? Were you just like, I hate this guy. Why did he tell me to do this?

Kyle Lawrence [00:17:08]:
I would never feel that way about my friend and colleague, Mr. Moishfel. I mean that sincerely. But no, well, it was actually it was exactly that. He literally walked up to me, I had a beer in my hand and he’s, you know, he’s explaining it to me and I was like, I know a little bit about this. We were coming out of COVID at the time I recently joined the firm. And he said just that, he said, you know, hop on a coinbase and just start putzing around and see and see where the river takes you. And I mean you can’t really see it behind me.

Kyle Lawrence [00:17:36]:
But I dove headfirst into the NFT world and you know, made the mistakes that we all make at the outset, but through that found just an endlessly fascinating world of, of just a, a better way of doing the things that we’ve done historically. We started the conversation making fun of old technology and now, now you look at cryptocurrency and we see what it can do for updating payment rails and using blockchain technology to do simple things that we’ve done inefficiently for decades now. And that’s, I saw that pretty quickly, you know, within my journey after making the mistakes. But what, what’s interesting to me about one, with what you’re doing now with Yuma and with your background is you have hands on experience working with the government, as do I. And so when you’re at this intersection of crypto and AI to emerging technologies, you’re doing both at the same time and you have to explain to regulators what both of these things mean. Forget about doing it once, doing it twice, twice. But you have the background of knowing that the wheels of justice turn very slow and regulators and legislators are often not necessarily the most tech savvy people. I don’t think I’m speaking out of school.

Kyle Lawrence [00:18:49]:
How do you thread that needle of being able to explain things in a way that is digestible for your average person who may not be an expert and then trying to move the ball forward to produce meaningful regulations and legislation?

Adam Sternbach [00:19:06]:
Yeah, I mean it’s a great question. Very quickly, what was your government experience?

Kyle Lawrence [00:19:11]:
I used to work for a state senator, I was the press secretary and I was shuttling back and forth to Albany all the time. And I come from a political, local political family.

Adam Sternbach [00:19:19]:
You learn a ton in those kind of roles. I mean, as I did when I was in the New Jersey governor’s office. But like it’s just trial by fire through a variety of constituencies and just, it’s great experience. But to answer your question, so I think through the efforts of many, the level of education and understanding with policymakers has improved incredibly in the last several years. So you know, I remember conversations where like we were walking through the difference between web 1, web 2 and web 3 a few years ago and now it’s like get me into the weedsy discussions related to permissionless systems and defi applications and you know, wallets and all that. So like there’s a lot that I think the industry and many others have led to increase the education levels, whether it’s trade associations or just like policy teams or a lot of you know, very talented lawyers more broadly. I think the challenge and the trick oftentimes is trying to make it relatable and not necessarily focus on the tech and the jargon. So like, even in this conversation a lot of my prep was like subnets, bittensor, like validators, miners, like all these things have certain meanings and somewhat different meanings than they do in others uses within the industry.

Adam Sternbach [00:20:42]:
But like try and relate to exactly what you asked about applications and like what gets built. And like I frankly find that if you try and approach the conversation with an engineer, like with someone who’s more technical but who can also bridge that to like here’s the application in use and give them the examples, it goes a very long way. And I think kind of what we were talking about of like needing to use the technology, I actually think it’s very underrated. But I think one of the most challenging things that was ever done, I don’t know what the status of is it now, but certainly in the prior administration was that there were ethics rules that prohibited people who worked on policy related matters from owning tokens. And I like, I get it, but I think like a de minimis amount to do what we’ve been discussing during the course of this conversation goes an incredibly long way to unlocking understanding. And so I think trying to help kind of people move along towards that, but also trying again to make it relatable and explain to them use cases and applications, but also like the functional technology systems for them to help understand why you can’t just slap on traditional existing laws in the way that might be an initial kind of impulse and inclination is also important.

Moish Peltz [00:21:58]:
It’s an excellent point. You see that both across government, state and federal level and across the media where there’s often ethical rules on what they can and can’t do with tokens. I think exactly what you said. It prevents them from exploring and gaining a fundamental understanding at that base layer that then could carry through on the regulatory side.

Adam Sternbach [00:22:18]:
Yeah, absolutely. And when like you’re in this industry and you’re trying to make nuanced points about operational control and who’s actually involved in running or operating or controlling or having like the ability to do things on a network, those nuanced distinctions really, really, really begin to matter. And it’s not an intuitive leap to understand why, you know, someone may not actually control the system that bears their name. And so a lot of times helping them get there involves those kinds of discussions. But like it’s also largely on the industry to be more thoughtful about architecture and system design and all that kind of stuff.

Moish Peltz [00:22:58]:
So I’m curious on your thoughts. The SEC and CFTC have come out with a bunch of pronouncements over the past couple months, including memorandum of understanding and token taxonomy. And those do seem to reflect much more sophisticated understanding of the industry, which has obviously been informed by many industry, you know, the bright lawyers that we, that we know and interact with and the industry groups and so forth. I’m just curious your thoughts as to how those, those different movements that we’ve started seeing now that are real, you know, regulatory pronouncements, do they, do they move the needle? Are they, are they on point? Are they, are there significant gaps in what they’ve said?

Adam Sternbach [00:23:41]:
First of all, I think the fact that you can even walk in and have these conversations needs to be acknowledged and appreciated. I remember in one of my prior roles we had a discussion with an international regulator face to face like this, where we were like, hey, this is what we’re thinking about. Like, you know, help us understand it. What do we need to do? And contrasting that experience in the middle of the prior administration and being like, we could never do this. And the fact that you can now do that again is incredible. And the fact that again, like talking about policymakers and regulators who are trying to understand the technology and tech fluent to really think through thoughtful regulation. Everyone you named across both the Hill and the administration, I think has put in tremendous effort to really get up to speed on that. And that’s a testament to them and their commitment to their roles and public service and everything, everything there.

Adam Sternbach [00:24:34]:
So I think like that just needs to be understood, appreciated, acknowledged. I think from the guidance itself and kind of some of the more substantive things that have come out, I view it a bit as a roadmap. I feel like I’m trying to not judge everything because I think this SEC has been incredibly transparent about like what they’re going to do and then actually doing the things that they’ve previewed and laid out. So I think like the general dialogue, the general direction they are moving is incredibly beneficial for the industry. It gives people an opportunity to obviously engage, weigh in. I think you’ve begun to see that with the most recent guidance, but also a variety of meetings, task forces, things that have been stood up over the last year. Plus at this point, I think there’s continual dialogue that needs to happen. I think one of the things in house that I think about a lot of times, and I think this is a conversation sometimes with law firms and then with regulators also Is my biggest lesson coming out of government was you can design policy in a million different ways, and you can feel really great.

Adam Sternbach [00:25:37]:
But, like, passing legislation, passing a rule, like, a lot of people view that as the end point and the end goal. I actually view it as, like, the beginning. And I think the implementation of policy doesn’t get nearly as much consideration a lot of times as does just like, let’s pass the bill. Let’s, you know, put the guidance and draft out. And trying to bridge that gap between practical implementation is something that I really try to think through in terms of like, okay, this is going to be the regulation, the law, the policy. Like, how does that impact the guidance I can give internally based on what I know we are doing? And, like, how do we operationalize that? And so I’ve been trying to spend a fair amount of time thinking through that because I do think it’s a natural impulse for there to just be a desire to get things done and get things put in place. But there’s a lot of work coming on the back end to actually put those rules and implementation of policy into effect. And it’s a hard thing to anticipate and to kind of predict.

Adam Sternbach [00:26:43]:
But I’ve been trying to think more about that and try and, like, tailor some of our conversations in solving the actual problems that will exist at a very basic and ground level, as opposed to, like, the theoretical, to a degree, if that distinction makes sense.

Kyle Lawrence [00:26:59]:
You kind of highlighted something that I always talk about, coming from the government background that I do and being a student of political science decades ago, it’s the differ between government and politics, which is the government is like, great. We, you know, like, we got this bill done, but then politics is going out in public and be like, I got this done, mission accomplished, and now we’re done. Because that’s all, you know, that’s. That’s their job. Then the actual underneath and getting your arms dirty, getting your hands dirty, rolling up your sleeves and doing the work. You’re exactly right. It often 100.

Adam Sternbach [00:27:28]:
I mean, I remember one of my first weeks in the governor’s office, there was like, some legislation that was signed, and then, you know, next administration comes in and it’s like 15 people sitting around a table being like, all right, how the hell do we implement this thing because it’s law, and now we have to go do it. And like, that was really the light bulb moment for me where it’s like, yeah, they move this, you know, from the legislative branch to the administrator to the executive branch, rather, and now, you know, the administrative agencies and the departments need to actually figure out how to make sense of what’s on the page and how to put this into place so that it’s effective for their constituents. It’s the exact same idea.

Kyle Lawrence [00:28:02]:
Right. And one of the things that makes that harder here in the United States is that we have such churn at these different levels. You don’t have the same group of people for a prolonged period of time looking at the issue and trying to figure out how it gets implemented. You have different groups and they have different political agendas and there’s different timelines. And since I’ve asked every guest we’ve had in the last six months, it’s kind of a good segue market structure. Bill?

Adam Sternbach [00:28:25]:
Yeah.

Kyle Lawrence [00:28:26]:
You think it’s going to happen this year? I mean it’s in April. I mean it’s kind of, kind of up against it right now. I don’t know.

Adam Sternbach [00:28:31]:
Yeah, I think if it’s going to happen, it has to obviously happen pretty soon. And there’s a fair amount of process in my estimation that needs to happen between now and then. Look, I don’t think it’s an impossibility, obviously, but I don’t know what the current poly market odds are. Maybe I should have checked before we jumped on. I know you guys checked last time in real time. I’m curious.

Moish Peltz [00:28:55]:
I can see how my, my Cal She Bet is doing right now.

Adam Sternbach [00:28:57]:
Yeah, I don’t know. I’m still probably in a 50, 50 camp. My kind of hot take about this is that the yield issue is obviously important in getting the most headlines and attention. But like no one’s seen the updated DeFi provisions and I think there’s a real question for the industry as to where those net out. And everyone’s like, oh, if we solve yield, this is going to fly. And I’m just holding judgment on that because I think there’s some other things that I know have been raised and you know, I’m not in some of these conversations on a day to day, but just kind of from like understanding within the industry there’s still a fair amount of questions there on the DeFi stuff in particular.

Moish Peltz [00:29:41]:
So there’s a 54% on Kalshi right now of happening in 2026.

Adam Sternbach [00:29:46]:
I think that’s fair. I’d probably be a seller of 54, but like I’d probably be a buyer of 49.

Moish Peltz [00:29:53]:
So I’ll yeah, I said no to before June when It was like 50% and I was like, yeah, there’s there’s no way. So that’s now that’s down to 15.

Adam Sternbach [00:30:04]:
So, you know, I think market structure would be very good in many ways. I think there are certain things that again, getting back to the implementation piece I talked about that I think would really need to be considered and would be real challenges to operationalize. But regardless, I think whether it’s through rulemaking that comes from market structure or whether it’s through just like the initiatives we’ve seen from regulators, I think it’s clear that things are going to be happening and moving and there’s a real opportunity to advance things and like obviously different preferences and pros and cons to legislative versus administrative. But I don’t think whether market structure passes or not, I still think there’s tremendous opportunities for the industry to engage and to like move things in a productive manner. And like there’s a difference obviously between guidance and rulemaking and kind of how that gets done. But frankly, even if market structure gets passed and other people have said this, like it’s a real race to implement the number of rules in this administration that, you know, I just, I don’t think like there’s a great clean solution coming out of this no matter what. And so there’s just going to continue to be a lot of this policy regulatory talk and engagement and trying to like figure out how to navigate those rules with kind of the complex systems with which we engage with.

Moish Peltz [00:31:29]:
So we’ve asked many of our blockchain industry guests about market structure built over the past six months, as Kyle said. And we’ve asked a few of our more recent AI guests about the AI legislation that that’s now working its way through at both the state and federal level. I think you’re the first guest we’re asking now about both. So what your thoughts on the now there’s this fractured state level AI legislation which seems kind of to me, in my impression, I’m curious if you agree or would frame it differently. Seems a little bit not very serious. It seems like half hearted. It doesn’t seem to be very understanding of how the actual industry works and the kinds of things they’re concerned about. Seem like again, not really fully thought through.

Moish Peltz [00:32:20]:
But then that leads into now, I think a larger discussion about what should happen, if anything, at the federal level. And I’m just curious how you see that landscape and where we are today.

Adam Sternbach [00:32:30]:
Yeah, so I think in the absence of federal legislation or rulemaking, and we kind of saw this in the past few years, states will fill that void and I think there were a lot of efforts to have to combat bills in a variety of states for blockchain related regulation and legislation, you know, in the last few years because there was nothing happening at the federal level. And a lot of states weren’t just going to sit there and kind of let nothing happen because they saw the challenges and some of the issues and understandably so like Bitcoin ATMs continue to be a huge hot button and stuff like that and just other kind of general regulatory efforts at the state level. Like there was one in New Jersey that I was pretty familiar with and involved with. And you’re going to just continue to see those types of efforts I think happen with AI in the absence of some kind of federal either like legislation or true intention that like we’re working seriously on this. And I think it’s really challenging to get that right because it blends more issues, I think than there are in blockchain across, you know, how it interacts with kids and how you can kind of trust and verify algorithmic integrity and like how do you know what you’re getting as an output? If I think like the Colorado bill touched on this, like, so I think it becomes a patchwork and that becomes really difficult to navigate. And then you’ve also had some efforts to kind of regulate like the hardware and the compute and the flops and these kinds of things. And I obviously understand it. I just, I also think there’s somewhat of a do no harm while this technology matures.

Adam Sternbach [00:34:12]:
But I also obviously understand from a regulatory and policymaker perspective that like they do no harm, harm happens, everyone looks at them and kind of where are you? So it’s just I think the age old balance between technological development and regulatory posture. But I think yeah, it’s acute with AI and I touches. I think things in different ways because of how broad it is related to kind of some of the issues we’ve historically grappled with with crypto, which is a little bit more financial regulatory in nature, I think like AI is more consumer protection oriented in nature and therefore I do also think the states do have potentially a greater role to play there. But it’s a related but different set of issues across somewhat different regulatory agencies as well. But it’s not going to be a good outcome if you know you’re dealing with 50 different state laws in the absence of some kind of movement at the federal level because of how powerful and all encompassing this technology is that I do think there needs to be federal efforts to harmonize something. But what that is, frankly, I don’t have a great sense today because of how much this stuff changes.

Moish Peltz [00:35:22]:
Yeah. And I’m thinking now about. You’re right, because it is analogous to the state level. The bit license was kind of that New York State being a first mover and I’m sure some politician being like, this is the thing I passed and I can go campaign on it. Then the implementation became both messy and I think influential because one had to kind of decide whether that was the hill they wanted to climb in order to do business in New York. Right. And so I think you’re now seeing, with the EU AI act being very influential and you’re seeing New York and Colorado and California emerges first movers there with. With acts that I don’t see as being very helpful.

Moish Peltz [00:36:06]:
And then you’re seeing kind of crazy like propositions like New York has this, this proposed act on. Well, like anything that provides medical or legal advice is illegal unless it is licensed kind of stuff. And it’s like, well, okay, but if someone could like, I think it’d be great if someone could talk to an AI and get good basic legal advice of having to pay for it, like, fantastic. That’s a good. It’s better than not having any advice, which is probably what’s replacing. So it’s a hard. But I totally get that it’s like a really difficult regulatory, consumer protection angle. Then the people have the need to fill that gap where the federal government is not doing that.

Adam Sternbach [00:36:41]:
Yeah, I completely agree with that. I mean, I understand obviously it’s new, it’s scary. Change and technology give pause. But kind of what I fundamentally come back to is like, you’re not slowing down the pace of technological advancement and development. So trying to stand in front of it is often a losing battle. And I understand a lot, again, like having worked in government, the societal and greater impacts of that, the need to kind of like think through second, third, fourth order effects of these kinds of things. Like, look, I don’t know that it’s as prevalent today, but a month ago, like I think we all thought we weren’t going to have jobs. And two years, I think, I mean,

Moish Peltz [00:37:21]:
yeah, swings back and forth.

Adam Sternbach [00:37:24]:
Yeah, this has been a better month than last month, but more and more I don’t think that’s going to be the case. But like that will be the case for certain segments of the population. Undoubtedly younger people who are trying to enter the workforce. Like, I mean, I think you can listen to other podcasts and the tips are like, you know, become the Best AI practitioner who can then train and mentor and upskill others and figure out how to use all these tools. So, like, there are paths you probably need to be creative. I don’t know that, like, that’s different than any of us having to figure out how to, like, enter a profession and, like, where’s the world going? What are the things I should pay attention to? But it’s definitely changing, and there’s definitely going to be change at a greater pace and magnitude than I think we have historically experienced from this technology. But I also agree with you that, you know, trying to pretend it’s not happening or trying to, like, keep it in a box in ways that are just not realistic is not the best approach. And, you know, I think there are a lot of efforts, obviously, to try and advance and shape and influence thinking on that.

Adam Sternbach [00:38:34]:
And, you know, some of the stuff that the administration put out recently about kind of its principles for how it’s thinking about this that were helpful to clarify kind of what its general thought process was on an AI. And like, look, I think there’s absolutely things that need to be considered, like protecting kids, protecting people. You know, these things hallucinate, and these things can give you very wrong information and answers that if you rely on them, is not good. Professionals can do that too. Other systems can do that too. But, like, I think there are new questions around, like, oh, it’s an agent, so it’s magic. Or, oh, it’s an agent, so, like, there’s no liability anywhere. And I think there was like, a recent case with an airline I heard something about where, like, they told someone the wrong thing about a policy and they ended up having to pay out of, like, the chatbot.

Adam Sternbach [00:39:19]:
I don’t know. Maurice, you might. And Kyle, you might.

Moish Peltz [00:39:21]:
Yeah, I think. I think this, this one, I think the airline chatbot quoted like a fair, like, whatever cheap price, like $300 to go around trip, like business class or something, and the person relied on it and tried to book a ticket, and they’re like, I booked it through the agent. You got to give me a fair quote.

Adam Sternbach [00:39:37]:
So, yeah, so it’s like things like that where, like, I think frankly, the existing legal systems, they do not directly address these issues, but, like, they are not totally, like, unresponsive to these issues. And therefore, I think there obviously needs to be, like, thought and consideration. And like, I don’t know, the most common one right now is like, if I have an agent, am I liable? And I think there’s unique and interesting considerations that come back probably fundamentally to like, control and access and how much do you continue to control this thing? And frankly, even if you don’t control this thing, like, you let it loose in the world, you’re kind of one generation, one level removed. Like, I have a hard time thinking that doesn’t flow back to you. At the end of the day, I think the more interesting questions become kind of like, what happens when your agent is spawning agents and that agent is spawning agents? And like, we’ve seen some of these things in the last few months where agents can monetize, agents can launch tokens, agents can have businesses, agents can pay for their own compute, and they truly can become like, outside of human existence in ways that machines and things could not do only a few months ago. And like, that’s a hard one. That’s a really hard one. And I don’t know that, like, we’re ready to solve for that today.

Adam Sternbach [00:40:58]:
But I think again, like the chat bot example or like, when your specific agent that you as a company or advertising gives false or inaccurate information, like, that’s much more clear to me. And again, I get back to like, the crypto nature of autonomous systems and control and things like that that I think largely will shape and influence the direction of law and policy in this space. But again, like, we’re also going to be encountering pretty new and different scenarios and examples and not too distant a future when you have the primary actors on blockchains who are agents who have launched their own token and have, like, developed their own business and services models. It’s fascinating stuff.

Kyle Lawrence [00:41:40]:
Yeah.

Moish Peltz [00:41:40]:
Well, is there not an inconsistency there, though? I think it’s a great point that you make, but it made me think about the dichotomy between, on the blockchain side, I’m a software developer deploying a decentralized system. I shouldn’t have any liability for that. But then on the AI agent side, it’s like, well, maybe you do on the first layer, but then there’s a second layer and a third layer, and at some point there’s just approximate cause where you lose any. So I don’t know, are those consistent philosophies?

Adam Sternbach [00:42:14]:
I agree with you. I think that’s what needs to be figured out of what modicums of control or influence or foreseeability or whatever you want to call it are going to influence and shape the legal and policy considerations related to this. To me, the only kind of conclusion I guess I’m willing to make now, and I even say that lightly, is I Think you know, if you control compute and you can shut this thing off at any time, if you control its API access keys, like that’s pretty clearly to me a SaaS product by another name oftentimes. But if you launch this thing and you no longer have, you know, the equivalent of admin control as you would on a blockchain network or an application, and you can’t shut this thing off and it has control over its own compute and like, if it doesn’t figure out a business model on the Internet, which is a crazy thing to say, but here we are for how to like continue to pay for its resources and pay for its API access and like navigate its own key management and security and all that. I think you’re getting further away for exactly the point you just raised Moishe of like, is that really control? Am I really responsible for that? But I think also through some of the market structure questions and some of the things related to blockchain systems and tokens, I think there’s also an instinct and desire for regulators to find someone responsible. Because if we get to a point where it’s like no five degrees of separation, you’re not responsible, that’s a dangerous place I think for regulators to think they can allow society to go. So I don’t know, I think it’s the perfect question you’re asking of like what are those factors and what are those influence. And I pity the first judge who has to actually develop, you know, the Howie equivalent test of what control looks like in an AI based system.

Adam Sternbach [00:43:59]:
But at the same time, United States versus AI agent. Yeah, like it’s kind of why I’m here.

Kyle Lawrence [00:44:05]:
Yeah. No, no, no easy answers.

Adam Sternbach [00:44:07]:
No.

Kyle Lawrence [00:44:07]:
One of the fears, fears I have, and there’s no way to answer this is because it moves so fast. How do you possibly legislate it? By the time legislation comes, is, is pat is, goes through Congress and is passed, let alone becomes effective, let alone this rulemaking, et cetera, et cetera. The technology will be so different and we’ll be faced with problems that we can’t even foresee what they are. So what, so what’s the even point of legislating? I’m not saying that seriously. But, but where do we start? What do you do?

Adam Sternbach [00:44:36]:
Yeah, I mean, look, I’m usually not a fan of like the throw a study in a bill idea because I think it’s just like punting on an issue that people don’t otherwise want to deal with. But frankly, I think for like technology related issues and I Think you’ve seen this in like, Europe, more so under Mika, with Defi, and like a little bit more in the us but not even so much. Although there’s some proposed things in market structure. I actually do think forcing agencies to study this and engage with industry for 12 months is a good starting point because it forces them to understand the technology, it forces them to dive deep and like, look, there’s a lot of different issues that need to be tackled and considered. And like, you know, like anything else, precious resources. But if we all generally agree that these systems are not going away, these issues are complex. The technology implementations and implications of them are incredibly complex. I do think that that’s where studies actually are valuable, where it’s like, we don’t need to firmly and definitively take a position today, but we should begin to do the research to understand them so we can really get smart and thoughtful about what are the most critical issues that we need to begin thinking about and preparing for.

Adam Sternbach [00:45:48]:
Not taking like, action, but ready when the time strikes, whether that’s the result of crisis or societal upcry or whatever the case may be. And so I would actually like to see a little bit more of like, pump the brakes on the legislative, regulatory, rulemaking and let’s just do a little bit more to like, get smart and study these things.

Kyle Lawrence [00:46:11]:
Great. Kind of send off thoughts there. You know, that’s really great stuff. And on behalf of Moishe, we’ve loved having you on. Before we let you go, any parting shots, Any final thoughts for the listeners out there?

Adam Sternbach [00:46:25]:
Yeah, I mean, the only thing, my daughter is a huge podcast fan. I told her I was going on a crypto AI podcast and she was like, oh my God, you’re going on a podcast. You have to give me a shout out. That’s what you do on podcasts. So shout out to Loe and Mason when you know, they listen to this, hopefully. Maybe.

Moish Peltz [00:46:46]:
I love it. I love the the family Shout out. Louie Mason, sending you all our best from block and order.

Adam Sternbach [00:46:52]:
Thank you.

Moish Peltz [00:46:52]:
If you made it to the end of this episode, congratulations.

Adam Sternbach [00:46:55]:
But I mean, to bring that to one point though, I also do think getting to some of the policy stuff, sorry to go back into the formal substantive mode, but like, I do think those are the kinds of issues where, like protection of kids, how do kids begin to understand and, you know, work within these systems educationally, in school systems like college, future of work, like, all of that kind of stuff, I also think are real considerations that we as a society need to make and like I spend a fair amount of time thinking about this stuff because they obviously have incredible risk reward for how they get utilized. And I think like that within the larger consumer production discussion needs to be front and center. But I also think it’s critical to make sure that they understand they’re like developing a fluency, age appropriateness and all that kind of stuff. But that’s kind of one additional point I would make just within the larger policy questions because I think you’ve seen some of the really unfortunate stories with chat bots and things like that. And like obviously efforts need to be done to prevent and mitigate, but I think so much of that also becomes like an educational problem at the forefront. And how do we make sure people just understand fundamentally like what this technology is and what it’s not and how do we kind of have systems in place to help people navigate introductions to incredibly new technologies at pretty young ages?

Moish Peltz [00:48:11]:
I think. I think that’s a fantastic point and something I think a lot about in both my role at the law firm and training young associates. And how do you develop critical thinking skills. And now also just from a Sidle, you’re seeing the social media companies have these massive plaintiff suits and what’s the liability for software developers and negligence on the way you harm society, at least being alleged. But then also with my own kids, I saw my, my 7 year old, like he goes up to Google and he types in a question and then AI mode pops up and he’s using it. So it’s like, you know, it’s baked into everything people are doing. And what are we doing as a society to help train kids and get them to think critically and make sure that they’re not just assuming all their thought to the AI.

Adam Sternbach [00:49:02]:
Same exact thing. My daughter did a photo shoot with AI because she had stuffed animals and she wanted them to be doing different things. And she’s like, yeah, I’m just gonna do a photo shoot and talk to Gemini and tell it what to do. And like, again, that’s a great creat, creative, appropriate use and like we’re observing. But like I often think and kind of within some of the policy stuff like regulation, a lot of times is done in an absence of education and in an absence of education. And so I just think it’s incumbent upon, you know, me as a parent. Sure. But like industries, when you are pushing these technologies forward to be thinking about these things so that there is education so you don’t get stuck with bad policy and bad outcomes.

Adam Sternbach [00:49:43]:
And I think like, frankly, a lot of that applies to the crypto industry and how people run into it headfirst thinking, you know, they’ll buy a meme coin and all of a sudden they’ll be a millionaire tomorrow. And like five minutes later they have nothing in their wallet and they’re like, what happened? And like, again, that’s an educational problem fundamentally. And I fear and, you know, think we just need to kind of tackle the AI stuff in a very similar way, at least kind of from like a thought and education and introduction and stuff like that.

Kyle Lawrence [00:50:13]:
Couldn’t have said it any better myself. Completely agree. Adam Sternback, VP of legal for Yuma holdings, thank you for coming by and thank you for a spirited discussion that you tackled on two fronts. You’re a pioneer here at Block and Order. We thank you for coming by.

Adam Sternbach [00:50:28]:
I appreciate it. Thanks for having me.