 And we're live for the June 2021 MIT computational law idea flow. And today, we have an incredibly timely topic that we're going to dig into, which arises from the newly enacted state law in Wyoming creating a new legal entity type, a decentralized autonomous organization, a DAO, that is wrapped as a otherwise run-of-the-mill US LLC, Limited Liability Company, which is like a corporation. And so this is, in some sense, like the dream come true that we've been talking about in the MIT Computational Law Report, and that Brian Wilson, co-host and editor-in-chief of the report, has been writing about, and we've all been hacking on. And now here we have it. Are we going to be disappointed by getting what we wanted? Now let's look at some of the implications here in terms of what are the details of this statute? How does it fit in the constellation of thought for DAOs? And what are potential issues that may arise from this, in particular, what is the role, if any, for human beings in this new legal entity type? So with us to talk about this and to, first of all, bring us all up to speed and then to help crisp up some of the issues and opportunities, is Eric Hess, who is a genuine expert in the field and who has testified before the select committee in Wyoming on this very topic and who I might also add is a real thought leader in the kind of idea space of telegram in this group and who have followed and admired for some time. And I'm just so delighted to have you join the MIT computational law community today. I hope you'll stick with us. And I'm glad for everyone to have a chance to meet you and for you to meet us. And so with that, Eric, if you wouldn't mind introducing yourself a little bit and then setting the table for this conversation. Sure. Well, thank you all for inviting me here. I, you know, the DAO is certainly a very intriguing topic. So let me tell you a little bit about who I am. I've been a practicing attorney for 30 years, which makes me maybe just a little over 40 years old. Don't do the math. I have worked in law firms, in-house financial technology. I was a general counsel of a equity stock exchange, brought them straight through, built an OTC market for equities, founded a cybersecurity business. I have a variety of digital asset and cybersecurity, encryption, privacy, clients. I tend to go tech heavy. I've designed a lot of compliance programs with developers. So I'm sort of tech heavy, but I can't code. I'm a lousy coder. I've worked closely with the government when I was at the stock exchange. I lobbied. I've changed rules. I've been very active in even law reform. And so how did I get involved in digital assets and even specific, you know, the Wyoming DAO? I did a lot of work with the Bermuda. They have a digital asset regime there, and I work heavily on that. And as some may or may not know, Wyoming borrowed heavily from Bermuda statutes when creating their own statute. So that got me interested in Wyoming, and I even did a little review of it. So, but we're here today to talk about the DAO. So when I first read the proposed statute for the DAO, I reached out to my contacts in Wyoming and I actually used the word trainwreck in my email, which is probably a little incendiary, but I wasn't all that excited about it. I didn't really intend for it to go back to the committee, but it did, and so it caused some attention. And basically one of my primary concerns when I looked at the statute was this very issue of human control. So the Wyoming statute initially said that you could either be member managed or, and I underscore the word or algorithmically managed. And it was like, if you chose algorithmically managed, that meant that you had nobody, no human in control. And I got spun up about this just to a great extent, because I felt passionately that that was just a terrible idea. And why is fully algorithmically managed a terrible idea? It's not really because I've been watching too many Terminator movies, but it's really has a lot more to do with, things go wrong and code isn't perfect and code needs to be maintained. Code sometimes has errors in it that need to be fixed. Sometimes if you're creating a LLC an entity that has sort of it's the characteristics of an entity, you may also need to enforce IP. That's a need, right? So all the things that you might need to do as an entity, it doesn't necessarily have to be mandated. You could be algorithmically managed, you could argue like I may not need to protect my IP. I may not, maybe I don't need to have somebody involved. Maybe I've considered that it might need maintained and it automatically generates a response to somebody or some group and maybe I can code for almost every single contingency imaginable to man and then I'm fully algorithmically managed. That sounds phenomenal except for one small problem or one single circumstance when it's not, right? When all of your phenomenal plans that you've incorporated into this DAO finally reached the fact that the point where they're actually not, where it doesn't achieve its desired objective, you now have a problem where you have a situation where you can't progress. And if it impacts the liability of the members or it impacts, basically you're accorded a privilege when you form an entity, a limited liability entity. It's limited liability. And so if there is not some way to handle that unforeseen circumstance, should you be accorded that status? And the answer is no, right? At some point, somebody has to be held accountable if everything that you've developed doesn't work, right? And that accountability doesn't necessarily mean you need to be sued and all your, you know everything that you've earned needs to be put at risk, but it does mean that there has to be some point of responsibility at some point. And that responsibility may be I need to dissolve. This isn't doing what it needs to do anymore, but there ultimately needs to, you're ultimately moving that choke point to a point on the chain where action has to be taken. Otherwise the argument is, you should not be accorded that privilege of limited liability if you do not have any mechanism at all to hit those specific situations where the whole intent of the statute is being effectively asserted, right? So that's, you know and so it might be a challenging concept for some because we all like to think, and I'm gonna go a little philosophical and I'm sure you sort of had this philosophical debate yourself, centralized versus decentralized, right? A lot of people, there's a lot of talk about the importance of achieving decentralization almost as a philosophical component, right? Like, you know, the decentralization is the ultimate, the thing you want to achieve, but the truth of the matter is it's really a continuum. True decentralization, absolute decentralization potentially leads to chaos. Maybe one degree short of it doesn't, but absolute decentralization is not really the objective. It has to be a sensible approach achieving that in much the same way we look at, you know what a decentralized autonomous organization does if it's fully algorithmically managed, if it's completely decentralized, there's no human intervention whatsoever that perfection is actually could potentially lead to chaos and actually going directly opposed to what the statutes are trying to achieve which is this privilege of according this limited liability, right? To an entity and it's not something that should be taken lightly. There needs to be some mechanism to correct for that. And again, I'm not talking about heavy handed and as I talk about this today, I think you're gonna see that actually my views on it are much more libertarian. So anyway, so what's the problem with DAOs today? DAOs are decentralized autonomous organizations. A lot of times they don't even have a legal characteristic. So what's the problem? Well, the problem is, is that that effectively makes all the members liable. Like I don't know if anybody remembers BitConnect, BitConnect, right? That crazy thing, right? So BitConnect actually was sued. They were that the promoters were brought up by charges by the SEC. What you say promoters hold it. They were working for the company. Why are they personally liable? And they actually, the SEC went further and said they actually initiated a securities offering. Securities offering, like, how are you talking about? They're promoters. They didn't do a securities offering but the SEC alleges they did. And why? Because BitConnect was actually a decentralized autonomous organization end period. No LLC, no corp, nothing. It was just a simple decentralized autonomous organization. Everybody who worked on it sort of worked as a sort of, you know, an unincorporated, unprotected entity. So that was, so think about the implications of that kind of case brought by the SEC and what it potentially says to any DAO that is operating without a legal entity status that would protect them, right? So that's the problem with DAOs. And then within a DAO, now we create this entity structure and we create these smart contracts and everybody likes to say terms like algorithmically managed or smart contracts and those are appropriate terms. But as we all know, and I don't need to restate it on this call, smart contracts are neither smart nor contracts, right? They program what they need to do and the contractual components they could be incorporated into a contractual structure but they themselves are not contracts. Now, computational law would argue that you can create a smart contract that maybe is a contract, but it really has, the characteristic of a smart contract has to be an agreement fundamentally. So the smart contract itself there are often components within this broader agreement structure. So let's go back to the DAO LLC and what's the DAO LLC in Wyoming trying to accomplish? So I went back to the committee and I had three primary concerns with the statute that was presented. It did not address a mechanism for dispute resolution. It did not address any kind of QA or code audit which is something I struggled with a lot because while I thought it was important I struggled with the notion of making it statutorily mandated. And the other thing I went back with was of course this notion of algorithmically managed was a problem because ultimately you needed human in, well, I'm gonna use the term that Daza came up which is human in control, which is what he introduced. So this notion of at the end of the day you need the human in control when things don't go right. But ultimately in the kind of, as I started working with the legislators I started to see that what they were developing toward they wanted this to be an extension of the LLC statute. So the LLC statute today obviously limited liability and the question is how does that characteristic needs to be changed in order to accommodate the DAOs? And the fundamental characteristic of the DAOs is either less human managed or potentially not human managed at all. So basically what's one of the things we wanted to accomplish? We wanted to ensure that you didn't have to, you didn't have to have a president, secretary, managing member. You didn't wanna have that. So today under LLC law you can actually delegate responsibility for this away, right? And in fact, you could also a trust. What do you have? Anybody ever hear of a voting trust? A voting trust is basically, that's where you could put your governance procedures and you could conduct all the voting and there you would only need a trustee, a trustee that didn't have any officer capabilities but merely it basically was in charge of being an extension or representative of the process. So effectively you could take all these governance components of it. There'd be no human, you could completely make it algorithmically managed, put it into a trust and then have that be effectively the managing member of the LLC. By doing so, you completely escaped the need to create a Dow LLC to begin with. Now, what is the primary value of the Wyoming LLC? The primary value of the Wyoming LLC in my view of the, I'm sorry, of the Dow statute. Number one, good media, right? Positions you as a thought leader in the space. It's also maybe a good start but truthfully, do we need a Dow LLC to accomplish it? Whether it's algorithmically managed or not algorithmically managed. We actually don't. There are mechanisms to do it today. Sorry about my dog in the background. There are mechanisms to do this today and you don't actually need it but it does frame the debate. It does have us discussing things like human in control and it gives us a chance to say what is actually 100% necessary to have within this Dow LLC structure? What do we actually really have to put in? What do we want to put in? Because if we start putting all the things that we want to put in or would be a good idea to put in that might be better put into an operating agreement which is how you come to terms with what's in the agreement. The question is what actually has to be in the statute because if you make it too much with the wants or the shoulds, all of a sudden you start to position in the case of Wyoming as a jurisdiction you don't even want to be a part of. Now I talked about a little bit about and I'll see whether anybody wants to go down this rabbit hole with me but I talked a little bit about the model Dow Law that was just put out by the, I guess Koala, Koala. So this has much more prescriptions than exists in the Wyoming Dow statute. They are thoughtful, they're good prescriptions like as a lawyer, if I take the call I say, this is great. I'm advising a client that wants to open a, that wants to form a Dow. I'll look at this and I'll think about putting addressing these concerns in a thoughtful way in the operating agreement, right? Question, do I actually want to go and form in a jurisdiction that has adopted this as a legislative mandate or even a legislative default? So if I go into a jurisdiction that has adopted all these, this more prescriptive regulation do I want to actually submit to that? I think unless I believe that that jurisdiction is much more sophisticated in managing Dow's like if I believe that the judiciary or whatever has handled these things, honestly, you look at jurisdictions like Delaware. I mean, Wyoming's got a long way to go before it becomes a Delaware in terms of its pure body of case law but Wyoming relies on a lot of precedent. But the point being is do as an attorney, would I advise my client to form in a jurisdiction with more prescriptive regulation? And this goes directly to human control, right? I would not, shocker. I would not, why wouldn't I? Why wouldn't I do that? Well, okay. First of all, I have the model law set in front of me. I can choose whatever I want to do anyway. Now you could say, well, hold it. It could be, you could change it. Like the operating agreement could provide otherwise these would just be the defaults. There's an argument there, but certainly out the gate, not something I would likely do. Because I'd want to ensure that particularly a jurisdiction that had adopted it I wouldn't want to be tripped up by whether it was a default or otherwise or some sort of potential legislative leaning that these are the ways that we should be acting, right? Like I don't want to have that predisposition to a prescriptive legislative model. I'd rather provide for it in the operating agreement in a jurisdiction that provides a flexibility for the interpretation of what I put into the operating agreement. So as it relates to things like human control, I can, that is probably the one area where we have to think hard. And I'm going right back to what has to be in the LLC statute, what has to be there? So I had said dispute resolution. Ultimately what I told Wyoming, as I said, Wyoming and I had the committee agree on this, at least address it. Say you have to say what you do for dispute resolution. If you want to say, we do, you know, we ask our dog and if he scratches once it's good. If he scratches, no, not at all, it's bad. And that's how we decide. Okay, fine, put it there. I don't care, that's not the point. Hopefully people run away from that project, but at least address it. Don't mandate any kind of requirement, just address it. You can provide for it in the operating agreement. Second, this notion of code audit also kind of goes to human control because who's doing an audit of the code? Presumably a human. Now it could also be another Dow, blah, blah, blah. But the point being is that there has to be, in my view, a good project should have some sort of QA to say, hey, what I am anticipating is contained in this Dow, how this protocol is being governed, what I think I'm investing in or being a part of is actually the case. How do I even know that? How do I, you know, and it goes to sort of like that anti-fraud component of the statutes, right? Like that's really what a lot of these statutes are designed to protect against investors or people being engaged from being defrauded, right? And there's other disputes and components but very fundamentally it's not securities law but it's still meant to sort of protect, you know, there's some baseline protections that, you know, that prevent, you know, the fraud or bad acts or willful misconduct or what have you, right? And the question becomes is should that even be something that they have to address, you know, in their, you know, as part of their filing, as part of their articles of formation, should we mandate that they address that single fact? There was a lot of debate. It wasn't 100% clear when we brought it up and Dazel was there, he'll tell you, it was, you know, there was a struggle, like, and they said, well, you know, what about if you're relying on ERC, you know, and, you know, it's already sort of gone through it and you're just adopting it? Well, and my point was, well, who cares? You could say, again, you could say, we have a lousy process, we have a lousy process. We rely on ERC, in fact, you might get that but the point is it's just a checklist and saying, okay, we addressed it and somebody looking at it will say, okay, that either works or not. In some ways over disclosure can be, you know, can be transparency by obfuscation too. So this was just simply a requirement. And now we get to this notion of a continuum which they actually addressed, you know, and they said, you didn't have to say that you're completely algorithmically managed. And then the last point I brought up at the end was this notion of petition. Again, what happens when things don't go the way that everybody wants it to? It's like the basis of every single agreement, right? Things will go wrong. And I said to myself, how do you manage things going wrong in a DAO that's an entity? And I looked at the dissolution provisions that were there and I think it could be done by Secretary of State if there's no member, it could be done by the one remaining member of the LLC but what about if all that doesn't really work? Like, how do we close this thing down? Like if it's a problem. So then there's also sort of two components to it. When I say this thing, there's two things I'm referring to. The limited liability protections that this entity is getting from the state of Wyoming in this case. And in the other case, it's the smart contract operation itself. So the fact of the matter is, is whether or not we dissolve the entity, the smart contract, it doesn't necessarily impact the smart contract. I'm not, you know, again, in the interest of not being overly prescriptive here, we'd be extending beyond what a statute normally does if we say, well, now they have to, they have to terminate the smart contract operation or we're going to enforce against the DAO. That's ludicrous, it doesn't happen today. There may be other avenues in the judicial system, you know, that the members of the DAO could avail themselves. But the question is, what do we really wanna do in this LLC statute, right? What do we really feel the need that we have to protect against? So the bankruptcy laws basically has this ability as everybody knows, you have an involuntary petition, what happened? Well, you know, you have this creditor has the benefit of limited liability, but they're like, they can't pay anybody. And, you know, they're getting this protection, but they really don't deserve it. That's really what the bankruptcy laws are. They don't deserve the protection anymore, right? They, you know, we, you know, in order for them to continue to operate, either they have to reorganize, they have to basically fess up. They can do chapter seven, they can do chapter 11. We give them all these different possibilities, right? You can reorganize or you can just liquidate because we're saying to you, you've kind of screwed up. You know, when we get it, we're forgiving, we love you, but you gotta like take a stand. You can't just sit here and watch the dissipation of your assets, not pay your creditors and still get the protections that we're giving you, right? And so the same concept in my view seemed to parallel what we were talking about with regard to human and control. We're giving you a privilege. What are the circumstances where that privilege needs to be revoked, right? And so basically one of them is if nobody, if we say, you know, if we think that this thing is being operated, that there's some notion of a human in control and this thing is out of control and perhaps I'm trying to read some of these questions, what I might do is, Daz, if you think it makes sense for me to sort of stop and answer some questions or I could do that. Yeah, I think if you feel comfortable that you've sort of set the table and I think you- Yeah, I'm actually just about to be done. So I'll finish my little thought here and then I'll take questions because I've gone on for- I think that would be good. And then Brian and I can help sort of synthesize them a bit and help catalyze discussion. And then once this group gets started, they're very self-generating. Okay, awesome. So I'll just finish this thought. So I saw the parallel to bankruptcy laws. I raised it. There was a lot of discussion as to who could petition. You know, like, you know, do we want anybody who's got a beef with a Dow? No, all you say, okay, it's a party in interest. Like in an involuntary petition, sometimes there has to be, you know, some number of creditors that petition. Now also remember that the Secretary of State doesn't necessarily just rubber stamp them. They either grant the petition or they don't. And I basically, you know, Daz and I were kicking and something around and I said, listen, they could, what they could do is they could petition and the Secretary of State could go back to, it could be the registered agent in this case, somebody. And if the registered agent does not respond, then they could dissolve it. I mean, on some level, you know, we went to the person you indicated was best positioned to respond to this and they didn't respond. And so by virtue of that, then you don't get the protections anymore. So there has to be some mechanism to protect this sort of notion of runaway. But the runaway doesn't really end. All that ends is the privilege of the limited liability protection. It's not like the smart contract shuts down the next day. It's not like anything happens the next day. It just means that if you're, that that limited liability protection is no longer there. And that could have consequences for some of the people, some of the members within it, they have limited liability. But, you know, that's something that has to be sort of addressed at the outset. Like, you know, the law can't sit there and say we're protecting these people no matter what, even when they're acting irresponsibly. And that also means that they should be indicating who the registered agent is. Doesn't I went back and sometimes you say, okay, this guy's in charge. Well, what about if that guy's really not in charge, right? What about if he's relying on somebody else and he can't get any answers? So this is something by raising this in the statute, it sort of addresses what happens when, right? Like in other words, this default mechanism, they should be indicating who they, today they indicate who the registered agent is, but a lot of places they, the registered agent will be like a CSE, a Harvard business. Somebody who's, it's basically a process more than it is actually an agent, right? It's like the registered agent is Harvard. Like all the corporations I form, rarely do I actually have an individual standing in. It's always like somebody who has no power or whatsoever. So we also have to be thoughtful as to there has to be some mechanism, but that notice process could be the way that it's communicated to the group and that group could respond within a certain period of time. So that's my, that was my notion of the petition. There's probably a lot of questions and I'm happy to even talk about COLA and some of my issues with COLA, although, you know, because I think it relates, but you know, at this point I'll just take questions. Great, thank you so much for that, Eric. I think it's really valuable. And I hope that the remarks that you've just made would also be valuable for those that see this on YouTube in the coming days and months and perhaps beyond. Most people of course are not yet aware of this kind of radical new legal entity type. And so just to kind of get us in here, I think there's an underlying thread to some of the comments that have come through. And it's almost like the first brick to lay in the foundation I think with respect to assumptions underlying what you've been saying, Eric, and what we wanted to focus on primarily today, which is there's an assertion that you're making. And well, I made it too, but it shouldn't be untested. And it's that there should be a human in the loop, in fact, in control of an LLC. And then we get to questions like what are mechanisms to ensure that? How do you discover if it's not the case? But let's go back to this assertion. Some of the ideology almost, if you will, or the principles or axioms underlying the DAO initiative is that things could run a much better, more efficiently and with better outcomes and more certainty if they were coded in smart contracts. And so why are we assuming it would be worse? And then when asked by Brian or somebody, what do you mean by like what would be not working? He mentioned, would it create an ex, it's not achieving the goals, but more to the point legally, is it creating an existential threat to humanity or otherwise large negative externalities for society like in that way. And so just to crisp up the question, if that's part of the expected potential harms, if there's no human at the steering wheel with these autonomous legal entities, pulling it back, maybe you can also address like, why do we say to form an enforceable contract, you need not only a human, but an adult? So why can't a five-year-old form an enforceable contract? And what is the level of intelligence and cognitive capacity that we would describe to a smart contract? Is it even the level of a five-year-old? So if you could just address some of these assumptions in the law, but what it means to have a company that up to now has always been driven by humans and contracts and transactions and decisions where there has been humans that are making the judgments and are accountable. And how is that different with a Dow LLC when there's no human and do we care? So I think, I'm gonna start with the smart contracts because it's a very good question. Could a smart contract or an automated process engage with another automated process without the intervention of humans whereby a transaction was occurred with funds for a computational service? The answer is yes. Could that relationship then be embedded within another relationship that also facilitates the same thing? The answer is yes. But fundamentally, this is all about moving the choke point or that decision point or that human in the process further up the chain because you could have a multitude of smart contracts like today, perhaps we have, let's say we have 200 license agreements. Tomorrow, like license agreements between an entity like where I signed, I send over an agreement, you sign and come back, I'm gonna deliver it to you or we do it automated, right? Or I just click, click through. Let me click through is sort of like the same thing. We could do all of that, right? So there could be 200 smart contracts, no click through, no human's involved, but that whole structure, everything that exists within that structure funnels up. Doesn't mean that there's a level of engagement, that there's a level of agreement, that there's a level of specificity that necessarily trickles down to every single one of those contracts. But what it does mean that as an underlying framework, there was an agreement by the parties that participated in the system that this is how it would occur, that we would rely on these smart contracts to do it. So fundamentally, at some level, somebody is agreeing somewhere to have done this. You can't get around it, right? It's sort of like trying to get to the beginning of the universe, right? Who is the maker? How was it natural evolution? You keep asking that question, well, how was the world designed, right? If it wasn't for somebody, you go back and back and back and back and you can make all these arguments and trust me, I'm not getting there. But the point being is that this infinity of smart contracts that we're talking about potentially, it all has to be within an underlying structure. And those participants arguably need to consent to it in some way, shape, or form, at some level. They need to opt in as participants in the system and to the consequences of whatever occurs within that system. So fundamentally, we can't get around the fact that at some level, there's gotta be some agreement somewhere and not maybe with the specificity that you would normally have with these peer-to-peer contracts. But the underlying framework, we've just moved up. So in fact, you may have like, think about like sometimes, I don't master contracts, like you have a master contract. These days, when I was a practicing attorney 20 years ago, oh boy, when I was a practicing attorney 20 years ago, like this concept of master contracts wasn't as popular. We just did contracts for every single little thing. It was like a contract, a contract, a contract. Now I signed a ton of these master services agreements where all these little, you know, addendum and contracts all come underneath it, particularly in the technology space. And but the point is, I don't even necessarily have to get involved in all these old trickle-downs, but I do have to get involved in a master level contract. So why is this even relevant here? It's sort of the same principle. You may not need the same master full-blown agreement because it's the nature of whatever it is you're agreeing to. It may not necessitate that, but at some fundamental level, there has to be an agreement by individuals that they're opting into this underlying system. And the same parallel also applies with regards to the Dow LLC, right? Because there is a liability component, right? With all these questions, one of the big questions we also often come away with is, how do we actually, what happens when thing goes wrong and who's liable? Right, who's responsible? Who has to be sued if things don't go right? And you may have basically, as part of the smart contract process, you may have consented in this overall framework to say, it's gonna be, you know, I think it was like Claros, it's gonna be a smart contract for automating my smart contract dispute, you know, or I'm basically, you know, waving it, or I don't care, I just wanna ever do business with them again, there's all these different things you could do. But fundamentally, there still has to be sort of this acknowledgement of what happens when things don't go right. And so this ability to sue or to be liable in the context that they all see is a privilege. And so why do we care about a human in control? Why can't it just be all computers in control? Because a computer, that fundamental responsibility at the very end, it's sort of like trying to figure out the beginning of creationism or, you know, evolution. There has to be, at some point, there has to be a decision that has been made to opt into that system. And in return for that decision, you've been granted limited liability. But the only way that you can, but that privilege, it ultimately needs to have some mechanism that exists outside the system because things may not go right. I don't know if I answered the question because some people could say, well, philosophically, Eric, I don't know if I agree with you, but that philosophical debate ignores the fact that that limited liability characteristic of an entity is a privilege. It's not a right, it's a privilege and it's revocable. And so that's why you need a human in control. And when we use a term control, control doesn't mean that they actually control everything, but you need to have a touch point for when things don't go right. Because guess what? Like with a Dr. Seuss book, when things don't go well and sometimes they won't, you know, that was the one that they give all the graduate students. And everything goes wrong at a certain point or maybe it doesn't, but sometimes it does. And what do you do when it does? And that's really what this notion of human in control. We haven't reached a point of perfection in dows for sure or smart contracts, no matter how advanced we think we are, humanity hasn't really ever achieved a perfection level in terms of operating. Oftentimes what happens in humanity is we just, the problems get bigger. We find a way to make the problems, we put all the problems for longer and then when there is a problem, it's bigger. So anyway, I'm getting philosophical again, but that's essentially fundamentally the point. I hope I answered the question and I hope it was valuable. Let's open it up a little bit now. Brian, I think that you're thinking along the same lines. Yeah, so I wanna build on this a little bit. And I think it's, you know, I think the ultimate notion that there needs to be some human in the loop for now is the right notion. So I just disclaimer, I agree. And one of the ways that I arrived at this point of agreeing is a few years ago, I got to work with some of the people in Malta who were developing some of their regulatory framework for how they deal with dows. And they came up with this new role that each dow had to have which was called the Technological Operator or something like that. And so if you wanted to set up a dow that was legally wrapped and compliant with the Maltese regulatory framework, you had to have somebody in Malta who could go in and kind of revert the code back to like the last known working moment before it got messed up. So kind of like hard for, or not necessarily a hard fork, but like, you know, similar sort of concept. And one of the things that we kind of did as an exercise there was we looked at, you know, all of these instances of legal personality containers that had been issued throughout history, you know, issuing legal personality for, you know, bodies of land, like rivers, issuing legal personality for, you know, buildings, ships, you name it, you can come up with these various containers and look at, you know, why was that, why was legal personality issued there? And I think when we look at the digital transformation of organizations from, you know, kind of like an analog paper-based organization to an organization that runs on code, there's a lot of institutional knowledge that has been built up throughout, you know, these centuries of doing things this way. And so, you know, I think because we don't know what so much of that future is, it seems like there's a chance to, you know, think about this in terms of not as like, you know, we're trying to solve all of the problems and get it perfect, but we're thinking about this in phases. So the question that I have, and that was a very long-winded buildup, but the question that I have is, are you thinking about this in terms of like, how do we get to like some ultimate goal of being able to maybe, you know, come up with less and less human involvement, or do you think that forever kind of like in perpetuity that it'll be this way? I think the, as I was kind of noting before, I think you can, you know, much of the, you know, smart contracts are almost easier to answer the response for because as we get better at this, and it's really just gonna be a natural evolution, it doesn't even ask to be legislatively dictated, and it shouldn't be, you know, and there's gonna be blow-ups and so be it. You know, sometimes the perfect is the enemy of the good or what we think is perfect is the enemy of the good. You know, but I think it ultimately, you know, you'll see more and more smart contracts or the structure or the complexity of what we do under smart contracts increase. And you know, cause listen, humans love complexity. We are as a, you know, as an organism, we're driven to constantly create a world that's more complex, particularly under the guise of simplicity, right? So it will be more complex and, but it doesn't like, even in the smart, as it gets more complex and as more is accomplished under it, that decision point may be much more minute, but there still has to be an opt-in. And I think the same is true with regards to the Dow LLC, it's a privilege. And no contract is going to be able to, you know, I guess it would almost be like, like, you know, IBM Watson, like a test, like we find the perfect Dow and it'd be the first ever that we accord that treatment to the perfect Dow and that'll be experimental and it'll probably be heavily scrutinized and it'll be a lot of big money put into it and, you know, and it will probably, you know, have so much focus on it, it'll probably go, it may go fall asleep because it'll be such a minute step. And then there may be certain circumstances where it actually does make sense, because, you know, as a society, we'll get more comfortable with these things and maybe in certain instances we can, but it's gonna be hard to say, oh, it's always the case. And that's where I think the challenge is. And to some extent, like, that's where like a registered agent, like there may be liability. So I think fundamentally, what I actually think is that you get to a point where there is like a trustee-like or a registered agent-like person who is, can act for the Dow. And the need, that person doesn't need to be an officer because they won't have the powers of an officer. They're gonna be had, they're gonna have very limited delegated duties. Like a trustee, you could oversee a thousand trusts, but there's still a trustee there. So even with the IBM Watson situation I gave, what would be the downside of having a trustee? If I could get paid to be the trustee and like give that additional comfort that if something goes wrong, that I can take an action, particularly around dissolution, why not? So it's more, I think it is almost, it is a mental exercise to think about this world where we take the human out of it completely. We can minimize that involvement, particularly in the smart contract level. On the LLC front, on an entity front, I would say never. Having said that, there may be experiments, but why do I say never? Because I actually think it becomes almost meaningless as a subject point, right? Like at some level people are like, oh my God, do you think the trustee is saying that we have to have a trustee always? Eric, what the hell's wrong with you? We're trying to get to a world where it's fully autonomous. It just doesn't matter. Like you have a registered agent for service of process for every single company that's ever been created in the United States. There's always a registered agent. Now it's a process. There's a person, somebody who gets it. We've named somebody, we've mandated it. You have to know where to send that thing. Nobody gets upset about it. It's sort of like that decentralized nirvana will be in a state of complete decentralization. Really? Do you really want that 100%? Really? How are we ever gonna progress even if we get into this decentralized environment? Once we're in a fully decentralized environment, I think people often identify, maybe we need just a little bit of decision-making. Even DAOs, they like to build in these administrative backdoors to do things. Well, what about if it breaks while you're a DAO? Yeah, but just killing me. Like I just have to have a little bit. No, it could break. And that's sort of like you see that now when with entities that are trying to go fully decentralized, and then they're faced with what it actually means to be fully decentralized. And they're like, whoa, that's just a little too much. Even the people who are the biggest proponents for decentralization hit that point. Now, maybe some of you say, well, oh, I don't know if that's true. Probably is to a very large extent, right? Because I see it now. When I talk to DAOs, it's decentralized, decentralized, decentralized. But when you really start drilling down to the full extent of what that means, there's that breaking point where it's like, oh, do I really, no administrative act, like 100%? Really? That's what it means? It's a tough one. And I think decentralized entities are the same way, so sorry. Oh, not at all. Just because we're starting to grow a little bit short on time, I wanted to just throw out something that it arguably is arguing against my own point of view, but it's idea flow. So you asserted that when is the right time to have algorithms completely and holistically in charge? Probably never. And I just wanna throw out there that at MIT, I hear a lot of things that I would have thought when I used to practice law were almost crazy sounding and that's partly because they were based on how do we build, how do we engineer systems that make possible things that were thought to be impossible today? And so I think the judgment that you said and that I share for today is probably accurate. However, maybe a finer way to put it is under what circumstances might it be appropriate for algorithms to take the place of humans and one could postulate things like Alan Turing postulated with the Turing test that wasn't possible at the time he postulated it, but it basically was a set of criteria of when will we deem computer program to be a functional equivalent of a human for some purpose? In his case, the purpose was just that it could fool somebody in conversation, but could you imagine a test suite like a test harness for software that would throw like law school fact patterns against it to see in simulation, does it know how to deal with being sued? Does it know how to deal with a labor crisis? Does it know how to deal with typical supply chain issue and having to change its contracts or a natural disaster? Like, can it deal with the completely anticipated range of situations that we know humans have to exercise judgment to preserve the capital and to not cause needless harm to innocent third parties and then start coming up with test patterns, almost like a bar exam where you can't just code for the last 10 years bar exams because they're always inventing new plausible hypotheticals every year to see if you understand the rule. So could we imagine if there was a test set that the secretary of state or whomever some quasi-public entity could run against algorithms so that they could prove their fitness for certain purposes? Could they be licensed or permitted to have dows instead of humans to the extent that they've proven it? And if they haven't proved it, you need humans in charge for that. I'm talking about, I would estimate the distant future because these dows are not very smart right now, but could you at least imagine that in the future? I think as a idea flow thought exercise, it's wonderful to think beyond the bounds of infinity. It's not really a legal premise. It's really just how do we open our eyes and think of things that we never thought about before and creative and open ways. And so I commend that spirit because why not have that as a goal saying, could we actually do that? Why not? But I'm gonna take a step back now. So in my cybersecurity practice, oftentimes I look at threats in four boxes. And you've probably all heard these before. There's, we have known known threats, things we can identify. We have unknown known threats. Somebody might do something, we don't really know who they are, but we know that these kinds of things exist, known unknowns. We have unknown, so we have unknown knowns, we have known unknowns. And all these things we can properly mitigate. What we can't mitigate against is the unknown unknowns. And there will always be unknown unknowns. And that is why perfection is impossible in this. And that's why, that's why I don't think you ever 100% nor should, I don't know if I'd say you shouldn't aim for 100%. I just don't think that that's the end state that is optimal. Having said that, could they, could, is that a 100% statement on my part? No, why? Well, because something could be simple enough, could be kind of discrete enough that it's just, everybody looks at it, it's like, okay, I put $10 into the smart contract. I give it, that's all it's ever gonna do, that's all it's ever authorized to do, and it blows up, they lost $10. Okay, maybe that's the worst thing that can happen. Maybe that's fine. We know everything that could possibly, it all fits into the known known category. Or maybe it fits into the known unknown and the unknown known category as well. Maybe we can qualify all of those. How do we actually qualify the unknown unknowns? I can't tell you. And I think when you have the greater complexity you have, that's a challenge. And how do you identify that you have unknown unknowns? Well, good luck with that. Unless you make things so incredibly discrete that maybe you have a separate, somebody has to grant that license to you and says, okay, we've made a determination that your project has no unknown unknowns or that the mitigating risk is so minimal. This is, that we don't have an issue with granting this. But it almost seems to me like an academic exercise, right? Because at the end of the day, if you have somebody who's just sort of like, all you're saying is, yeah, this thing blew up, I better send this note in. And that really is a definition of that human in control or I'm just gonna make sure that we predicted our IP or I'm gonna make sure we do this. Or I'm like a trustee, I manage a bazillion of these things and I go through my checklist every time. And maybe I even put that into a smart contract myself so I can relax and drink my margaritas all day and not have to do any work. And if something blows up, then I jump in. But so even at the trustee level, he might find this falling into place and he's gonna have his own mechanisms to do this. But even then at the end of the day, so the trustee could be in theory, I guess smart contract or smart contract driven, but at the end of the day, that person is still responsible for carrying or ensuring that it's carrying it out. It's like the check. And then even with the check, you say, well, hold it, couldn't we make a smart contract for the check? Yes, but then who's gonna check that check? Well, we'll get another smart contract to check that check. Okay, well, who's gonna check that check? Does a human make it better? But at least a human, you can say you didn't do your job, right? You can, you know, that person can be discredited. Is that worth it? Is that really, I mean, on some level it's, I think the interesting part of the exercise is seeing how far we can progress to that. I don't think the interesting part of the exercise is trying to identify that single point where that human in control is actually all that big of a deal where we need to take them out. You know, they're, you know, the question is, is do we really want that as a society? Is this something that's, you know, potentially gonna have a boomerang effect? I just think there's so many, you know, listen, as a society, we could probably get around it, particularly given specific use cases, whether or not as a legislative mandate, we should have a provision that allows for it, maybe in discreet circumstances, but the notion of saying, you know, any Dow could form and have that limited, you know, could have that limited liability privilege without actually having some accountability at some level. To me, it doesn't make sense. And a lot of this, you know, listen, better than a human control is what kind of, you know, code review processes do you have? What kind of QA processes do you have? What kind of transparency do you have? These are things that are actually meaningful. You know, the person ultimately who's in control, who actually really isn't have all that much control because they have no operational control, it's sort of, it's almost the wrong question. And I think trying to achieve that risks a lot more. And honestly, that risk could be detrimental to the whole ecosystem. Why? Because God forbid we do this and we have this algorithmically control and the bad thing does happen, then you're gonna get this whole populace that's gonna be like, oh my God, the machine took over and look what happened and we're gonna lose the whole dialogue. So there's also a political implication to this. And why? For what? Just have that dude up there or the woman up there who's pulling the lever or whatever, just have somebody up there. It just doesn't matter. Just have, you know, a registered agent. I don't know. So that's my view. I think it's, I'm not gonna say it's much to do about nothing because that would diminish the importance of the debate or the pathway to try to achieve it. I think the pathway is a hell of a lot more important than the endpoint on this. We're here. So yeah, it's so just by way of kind of starting to wrap, we're going to save the chat. It's an amazing discussion there that we didn't get to much, some of it in the live feed. But I guess just two echoes from the chatter. One is really is the right time to tighten up on requirements of human control in the beginning, you know, as part of the enabling legislation or is the right time after there's, I'll use my own words, but like demonstrated market failures or demonstrated legal failures and then to tighten up based on that. And then that raises the question you just posed in a sense, which is for technology policy, if one of the risks of making it too open in the beginning and going through a hype cycle and then having disasters, is that it may end up being decades before people are ready to trust again algorithms and we could overall set the economy back. The other thing is compared to what? And so if you actually look at the decision control of run of the mill LLCs now that are run by humans, it's not like it's some kind of like award winning, you know, Pulitzer panel of Dean geniuses that are running out in LLCs, you know, and so it's like maybe our expectations are in a sense too high or maybe they're just different in kind. So we shall see, it does, I'll just say my own opinion as hosts and hopefully not abusing my responsibility, it does seem appropriate to at least have a transition path to complete algorithmic control of legal entities. But one of the great things about Wyoming is they're trying new stuff. And so, you know, we've got laboratories of innovation that in state legislatures and so now we'll have a data point to see what happens and we can take it from there. So with that, I wanna thank you so much, Eric, for joining us today and sharing your views and kind of joining the scrum and thank you so much to the MIT computational walk community for showing up as well and sharing your thoughts and ideas. And for all of you who are new to IdeaFlow this month, please do join us. It's the last Friday of every month of 12 p.m. Eastern and feel free to let us know what ideas you would like to discuss in the future. So with that, thank you everybody. So I just wanna be before everybody breaks, I wanna thank you all for listening in and also I also have a podcast, The Encrypted Economy. So if anybody's interested, it's on YouTube, it's on all the podcasts. I tend to go deeper. I don't talk as much as I do here because I'm not the subject, I'm the person asking the questions, but we cover a lot of things related to privacy enhancing technologies, some, a lot of these challenging issues. Just check it out, you know, all of it may not appeal to you, but run through it, I'm sure some of it will. So, and if you look at it and you're like, hey, this is, I know somebody who knows about ZKP, zero knowledge proofs, that's one area I'm looking to really bolster on it and I haven't gotten any takers to sort of start building on that, so. Raising my head. All right, there you go. So... You pay snarks, baby, all day long. Starks, snarks, bulletproofs, all that stuff, so all good stuff. So thank you all, I appreciate it and hope what I said had some value to the discussion. Indeed it did. Okay, thank you.