61. My AI Associates Fight and I Let Them with Helen Fan

All right, welcome back to Counsel, also known as my kitchen table. I am joined this week by Helen Fan, who I have not met before. We just said hello just this very moment, but I have been stalking her on LinkedIn and thoroughly enjoying her content these past few months. So I shoot my shot and I DM'd Helen and asked if she would like to come and have a chat about all the really cool stuff that she is doing and seeing in the world of AI and legal and law firms and legal and this incredible, you know, evolution that we are just living through what feels like right now. So Helen, welcome. Thank you for taking a chance on a random Australian woman that you don't know. It's so lovely to have you here. I hope you've been following your work for a while and yeah, incredible work online. Yeah, it's so, so good and so happy to, you know, it's really happy to join this conversation and yeah, thanks for having me. Yeah. Oh, you're so welcome. Now I did a little link, a little LinkedIn stalky stalk and wanted to get a sense of where you've been to get to where you are today and you've had such a fabulous multinational cross-border, you know, a very big global legal experience, you know, studying in the US, working and practicing in China and from what I can gather in Asia Pacific region and then back to the US in a bank and then working for New York law and then Silicon Valley lawyering, like you've really, you've really done so much. Can you tell us about the moment that you decided that legal AI was actually where you really wanted to kind of plant your flag for the next chapter of your career? Yeah, definitely. And I think that moment happened when I was in Columbia Law School. Yeah, so maybe I can start with my background first. Yeah, so just I want to say hello to everyone. And so I'm a California attorney based in Silicon Valley and I'm also the chief AI officer at a boutique law firm here. So for my background, I studied law in Tsinghua University. It's one of the top two universities in China. And then I worked at Fonda Partners in Beijing. So Fonda is one of the top law firms in China where I served a lot of financial institutions like Citibank, Morgan Stanley. I also got a chance to be seconded in Citi's legal team in Shanghai for half a year. And after that, I came to the US for my LM degree. I went to Columbia Law School. So and that's something really interesting happened. I entered the first ever legal tech hackathon hosted by it's three years ago, actually. Who was that hosted by? I think it's just a Columbia Legal Tech Association. It's a student association. Yeah, amazing. Love a hackathon. I love the team. And we got second place. It's really happy. And I think at that moment, we just have something like a chat BTO3 is a viral. You can like, like 10 years ago, right? But it's years ago because AI, like three months is a generation. This is a long, long time ago, but it's just three years ago. Yeah. So, so I think from, from, from then on, I think that changed everything for me. Like I realized AI is going to fundamentally change how law firms are built. And after that, I moved to Silicon Valley. A part of the reason is I want to get a reunion with my husband here. And, but yeah. And, and also I, I'm so lucky. I joined a booty law firm where I served a lot of AI startups. So most of my clients, they are founders in some AI companies. So actually they are more familiar with AI than me. And I found that many times they just, you know, send something they dropped to the by AI, you know, as a first job for me to review. And it's, it's, it's really interesting because, um, sometimes I feel like they trust the AI more than trust me. Yeah. So, you know, they are the founders and maybe they have so, you know, so much, you know, um, technical background. So that's why they, um, they kind of prefer using AI as a first legal counsel they want to talk with. I don't know why, because at that time I think there are still something like a hallucination kind of thing around this topic, but for those founders, they don't care about it. They don't care. Their risk tolerance is so high. And this thing, like if we build like something we build for like 20 hours to, for drafting a policy, a privacy policy for their products and they may just push it back. Like how come I just, I can generate it with AI like within one hour or within half an hour. So it, I think that, so the interesting, I mean, it's, it's very interesting, but it still, uh, cause some anxiety. So, and I think, yeah, maybe I should like, uh, explore more with AI to know more about it. And in Silicon Valley, like every day you will have a long list of AI events. So I came to some AI events and most of the time I'm the only lawyer in that room. Wow. Like how long? There's so much terminology I can't understand. And I'm just crazy. Like, what are they talking about? So, so I, are you, are you one of few women or if you find that there are a lot of, a lot of women there or is it still really like tech bro? Uh, tech roles. Yeah. Yeah. And there's some tech woman definitely, but not so many. Uh, yeah. And so, yeah, I'm the lawyer and, uh, so very weird, you know, it's hard to understand why I am here, but it's just so interesting. I mean, I, I think I'm just driven by my curiosity and also my anxiety. And I think this situation happens for most lawyers right now. And many people, they just like lost their mind at what I'm doing now. And maybe I will be replaced by AI one day. So yeah, maybe that's why, uh, I can see now more and more people, they create their job in the traditional track. And now they jump to some new places with AI, like to do something like a legal engineer or like, uh, AI lawyers in, uh, big law firms. So now we can see it's, uh, it's so many, you know, exciting things now. Yeah. Right now. Yeah. It's such an interesting time. I love what you say about your clients coming to you with the AI answer. Uh, in Australia, we call that a bush lawyer. It's someone who's like a bit of a, they, they know a little bit, but they're not, they're not a lawyer. They're a bush lawyer. They're like kind of smart and they've Googled it or something, but they, it's like going to, you know, Dr. Google where it's like, it's not the same as going to a doctor, but now the bush lawyers are coming with some really good stuff from chat GPT or Claude. And I would look at this and go, you know what? This is, this is good. This is probably 80%. But that piece that they don't have that we have for now is the judgment piece, the pattern recognition, the, the context of the business. So I, I, I'm so fascinated and not surprised at all to, to see that you're getting that. And it's such an interesting time for in-house lawyers to then go, okay, well, what is our value? What value do we bring? Now, I, I think we bring a hell of a lot of value. So, and, and there's, there's so, there's so many, um, it's all happening right now unfolding before us, but you're really from what I can see at the cutting edge of it all. I, I love that you also take yourself into those rooms where you feel like maybe you don't quite belong, um, but you're curious or anxious or both and you want to learn. I think that's amazing. So. Yeah, actually that's a cunning point for me at that time. I just started on my own legal AI community and it's called Legal Tech Frontier. Uh, we've got over a thousand, uh, members, uh, across different places, across the Silicon Valley, Beijing, uh, Shanghai, Hong Kong, and Tokyo. So they've been, uh, like, I mean, I, I've hosted like, uh, 12 in-person events across different places and, uh, we've just, uh, uh, we've gonna put up together another new event in New York city. This, the next Friday we will have one in New York city and I co-hosted with, uh, Cooley. We will host the one in Cooley's New York office. And we will just talk about what is the AI native legal team. Yeah. That is so exciting. It's ready for in-house counsel. That is fantastic. I, um, I'll put a little detail in the show notes for anybody that wants to check out what the community is all about. And you might get some Australian members come along, but we'll have some with a screener members. Yeah. I just got the chance to host the in-person event there, but I do have some members from that place. Oh, amazing. That's awesome. So let's go there. Let's talk about AI native law firms and AI native legal teams as well. What do these terms mean to you? Not everybody will know what this is yet. It's also, it's also new. Yeah, of course. I think, um, I think it's really interesting because so earlier this year, like YC just founded a couple of AI native law firms. I'm not sure if you ever heard of it, like general legal, vector legal, Moritz. And also I forgot the last one, but there are four AI native law firms by YC, uh, in the YC 2026 winter batch. So, yeah. So I think it's a new trend right now. So, and we also got like Crosby AI, uh, it is founded by someone, uh, from Cooley. I think the founder has a background with Cooley and also some else big names. And so, um, that's another very, uh, like famous AI native law firm now. And also we, we have, uh, like norm AI, NORM, norm AI. I think it's just a reason, like over one million dollars. Yeah. Cause this new Brown. Yeah. So now it's a, it's a big trend. Yeah. Everybody noticed that. And, uh, I think for the definition, everyone has its definition. So that's why I posted a, uh, community event, uh, focusing on the definition of the AI native law firm. Um, and I can share something I learned from that conversation. So just in the past May, I gathered, uh, like over 50 people in Silicon Valley. And we, we just talked about the, what's your definition about AI native law firm? So here are three keywords. The first one is building with, like, uh, based on outcome, uh, or flat fee instead of time. That's the biggest one. Wow. Yeah. It's very different, uh, from the build of our system. Right. And the second keyword is client facing AI. Maybe not, not so many people, uh, understand that, but basically it means that clients can use the AI system provided by the law firms and upload the legal documents into the AI system and chat with the AI, you know, the law firm's AI. So sometimes maybe, uh, most of the basic questions, the AI just don't answer it. You do not need to reach out to the real human, but, uh, if it's, uh, uh, it was like something very complicated. Um, the law firm, the, I mean, the attorney will just, uh, show up and say, okay, now, uh, we should talk about it. So, I mean, um, vector legal because they just, uh, their founder attended our, uh, event and they showed, uh, their demo, uh, their AI system. And I think it's the very innovative, innovative definitely. And also I know like Cooley, they just launched their own AI platform and are coordinated with Nagora and that platform is called Cooley Go. Yeah. So, so that one is also tailored for clients. So you are a founder in, if you are a YC founder, you can use that zero platform to, uh, talk about maybe your questions around how to get the, get the funds from the investors, what I should pay attention to when I, uh, do the, maybe the option management. So, et cetera. Yeah. So facing AI, I think it's a, it's a very new topic, especially when we talk about AI-related law firm, because it's really hard to imagine, you know, in, in some traditional law firms, we will build up something like this. Uh, and yeah, so that's a second keyword. And the third keyword is, I think I defined was the, you know, based on my agent experiment. Maybe we will talk later about the 100 day experiment. Yeah. So I will call it like self-learning data layer. It means that every time you chat with your AI, AI just to get smarter. So like, yeah, so everything like, uh, the templates, the communication between you and your clients, and also like, um, the workflow, all of this will be just, uh, you, you will feed this, you know, into AI and the AI will just, um, memorize all of this stuff. And next round, it will just get a better result and outcome. Yeah. That's amazing. Yeah. How exciting. So it's really about fully, fully decimating the billable hour, thoughts and prayers to timesheets. That is, that is great news for so many lawyers around the world. That is client facing, you know, the user interface for the client, you know, being the, the forefront, the thinking about the design of, of how, how is this best serving the client? Again, not something that big law has typically really put at the forefront, um, in my experience. And then, you know, self-learning, self-evolving, growing, you know, constantly kind of getting better. These things that just make so much sense to me that that sounds brilliant. Um, and I'll definitely be checking out those, those four firms that Y Combinator have, have backed because like you said, that is such a massive trend for in one in man round in one batch. That's, I mean, wow. And look, what I love about all of this is access to justice because the people who could previously not perhaps afford five, six, $700 an hour lawyer, maybe more in New York, who knows, uh, they can now for the most part, go and speak with a bot and get a pretty decent result that will give them the confidence to, to know whether they have a good case or, you know, have they been treated unfairly by their employer? What are their options? And just give them a little bit more information. And eventually, of course, yes, they will have to engage their own attorney, but, or maybe not, but I just love that, that this whole thing can really help the missing middle is what I call it. And people who haven't, they've, they haven't quite got enough for the big, the big end of town, but you know, maybe they're a small business, maybe they're, you know, just kind of just struggling a little bit. They, they've got kind of, you know, left to their own devices and something bad has happened to them and they want to understand what their rights are. I think this is such an awesome, um, an awesome time. And the, the firms that focus on those problems, I think will, will do particularly well, but we'll see. Yeah, definitely. Yeah. So can you talk to us about what you're building and your, uh, I want to know about Morgan and Cleo. Yeah. Can you tell us who are they and why do they argue with each other? Yeah. So they are my employees actually. Yeah. They're my agents. Yeah. I just give them each other name. And so they're not, they're not real people. They're great names. I like the, the name Morgan and Cleo. And they really, I really, yeah, they're lovely names. Yeah. So yeah. So I think I just, uh, it's, it's my 100 day, uh, technology experiment. It's not a real law firm. Yeah. But I kind of simulate what happened in a real law firm. So I think the story, um, started in this March, like there were really two things happening at the same time. The first one was, um, the cloud plugin, like when cloud launched its legal plugin, I sat down there and I wrote a piece about my analysis and then it gradually became a, a framework called legal AI value stack. So, um, so after I really dug into it, the main conclusion was that the thing with the most defensibility is an AI native law firm. And once I'd written that, I started asking myself, could I actually build one? So back then there were very few people trying this. Like I knew I say, but everyone was really at the very beginning. So I thought maybe I should just try it myself. And the second thing were the community around me, like at the same time, this whole wave of building with agents was taken out due to open cloud. So going to a lot of these events in Beijing and in Seneca Valley. And the funny thing is the room was almost all engineers. And like I said, so I was usually the only non-tech person there. And everyone was talking about how to build with agents, like how to build up, um, AI native enterprise, or like, uh, AI native engineer, how to become an AI native engineer. So I think, okay, um, maybe I, maybe I, it's a very good, I mean, a good starting point for me to make it as a public experiment. I mean, I can just, uh, you know, post something in public and I can do it like, uh, um, like a 100 day. I put it as a 100 day trying to do to me. Yeah. It's so, it's just so hard to continue doing that, especially when you made some really difficult technical issues, but I have to, because I make the promise, you know? Yes. So yeah, it's just, you know, um, the, the, the, the, the, the start point for me. And I think, yeah, it's, I just wrapped it up last, uh, month. Uh, I think it's 100 day I finished it and I wrote something about a summary about all the, the whole journey. And I write it to Latham. Yeah, I really enjoyed it. And I think, um, it's, it impressed so many people. Um, actually it's got quite a lot of attention, like over 1 million viewers across different platforms. I got the invitation to present my open call law, um, in Stanford Codex Future Week. And I also presented in Beijing and in Sydney Valley a few times. And I also, uh, joined, uh, conversation in Big Law firm, in Big Law. Actually, they hosted like, uh, uh, academy for over 30, uh, chief compliance officers, uh, for big tech companies like Apple, Magina, Google. It's so fascinating. I joined them and I became, you know, someone who can teach them how to use agents, transform your work. All of them, you know, all of this come from my experiment with agents. That's so fantastic. And congratulations on getting to the end. I was following along myself and I think that I came into your content just before you started and just fascinated because I have also been across open claw, but just enough to know that the, for me, the, um, the security and the risk piece, I was a bit scared about how to bring that into my business operation. And I heard a few scary stories about it running off and spending people's credit cards and doing all kinds of things without, without the right safety guards in place. And I thought, you know what, I might just give it a hot minute and let everybody else learn the mistakes and then I'll, I'll come back in. But you went, you went there and you did it, you know, in public. I want to ask you what surprised you most about the experiment of building open claw law and what, what surprised you most and what, what was harder than you expected? I think, uh, there are two things I want to mention. I just mentioned one, it's called a self-learning data layer. And, um, so, I mean, I'm kind of migrated, uh, all the setting on open claw into a Hermes agent. So I know, I'm not sure if you know Hermes agent. So it's a, like a advanced version of open claw. Yeah. So it's a more reliable and stable and it, uh, uh, very, um, I mean, it has a very good feature called, uh, self-learning loop. So every time you chat with your agent, your agent will, um, take notes about what you said and it will just became, you know, some reusable skills for you to use next time. So, um, that's one thing I want to mention. And the other thing is about the agent to agent communication. Uh, so Morgan is my senior associate, right? So in, in, in my law firm, I have two associates. One is Morgan. Morgan is the senior one and a clear is the junior one. So Morgan will usually handle strategy, client management, email screening, and, um, just a supervising Clio. And Clio is my first year. So she does legal research, case law analysis, memo drafting, and most importantly, they can talk to each other. They can challenge each other. They can argue. That's by design. And when they, I asked Morgan and Clio to research the actual requirements for like, um, um, H1B application. So H1B is, uh, the main U.S. work visa for professionals. Yeah. This was right after news about, uh, $100,000 fee about, yeah. So they, uh, Morgan and Clio, they came back with completely different answers. So, uh, clear words, confidence, like every applicant must pay $100,000. And Morgan says, that's not right. So when I got into it, the confusion made sense because once that news first came out, there was a lot of back and forth, like real news mixed with the rumors. Because at that time, the White House statements were unclear and it took a while before it was settled on paper. And the actual rule only applies to a very narrow set of cases. I think it, uh, so I think so both agents got confused, but in a way that it mirrors how a real junior lawyer would get confused and the disagreement itself were the signal. So it told me, okay, this is where you need to verify carefully. So I think agent to agent conversation or agent to agent, um, you know, um, like arguments, it made me realize that maybe I can use them to reduce the hallucination generated by AI, to find the blind spots that AI may have and to help me to verify the contents of the outcome better. That's amazing. I love that. I have this, just like this little image of the little bots just arguing with each other and, you know, hopefully they're very, they're kind colleagues and you're not going to get like a bullying claim made against the one to the other. They might start taking on the personalities of actual people, understanding their rights in the workplace. I don't know. That's getting a little too futuristic, but we'll, we'll stay in the bots for now. I just think that's really fantastic. I was, I was interested to know what all of this means for junior lawyers. You, you just described Clio as a pretty, pretty typical example of a junior lawyer and they're very affordable. They're not going to cost $200,000 or whatever it might be. What do you think that this might mean for, for juniors coming through? Yeah, I think I heard some senior lawyers, um, they talk about something like, um, sometimes they find AI more convenient for them to work with than their junior associates. Yeah, that's true. And I think there are so many reasons for that. Like the back and forth is faster. And so they can just go back and forth with AI in real time until it's right. And also there's more shared context because they've used enough that it kind of knows their habits now. And, and, and AI learns really fast, right? Sometimes it, it picks up statues and a new material faster than person do does. So a junior who doesn't use AI well, I think they end up looking slow and the work isn't as good as the work. The partner could just gotten from the tool directly. And that's a real competitive disadvantages if you cannot use AI very well. Um, so yeah, I think to be honest, yeah, that's the, that's the, that's the tool, um, kind of not very positive, but, uh, that's true. Yeah. I would always recommend for junior lawyers who are studying or they're just early in their career to take AI very seriously and to become someone who can understand it and who can work with it. And it's part of their practice so that it makes their lives easier and more efficient. Um, some of their law firms might not be progressive at this point, or they might just have their head in the sand about it. That's still a thing I'm noticing, but increasingly this year, it's like, it feels like everyone's starting to wake up and go, Oh, okay. Wow. And it's, it's some big questions, especially for, you know, certain types of firms that do a lot of high volume transaction work, going to really shift the, the time required and the margins. And are they going to pass those efficiencies onto the client or, or not? Who knows? It's, it's all, it's all quite a fascinating time, but for the juniors, don't be afraid to play around and experiment, you know, take a leaf out of Helen's book and go into the events where there, you might be the only lawyer in the room, but you're going to learn and stay curious. And then who knows, it might take you down a different path and you've created something, you know, really cool. And then you're talking to the chief compliance officers of all of the big tech companies. That's, you know, that's just such an awesome lesson, um, that you, you know, you've really played out that's played out for you. Thank you. Thank you so much. I'd like to ask you about the big law billable model, billable hour model. So you have said that you do think maybe that the big law billable hour model is a structural barrier and the adoption, you know, enthusiasm alone may not be able to just overcome that. What, what can you, you know, unpack that a little bit for us? Yeah. First, I think, um, like I mentioned, I got a lot of pushback from my clients during my work and, you know, clients, they use AI and they feel like, oh, AI is so great. It's so brilliant. And, but why I still need to pay so much for my lawyers? So I think they, they just got quite a lot, you know, challenges from their clients, but for big law firms, I think, um, first it's because of the billable hours. It's, uh, like historical things. It's really hard for the law firm to change their model. And if you always like get profits based on the hours, why you will care about the efficiency because efficiency is like makes you like quicker. It's not a good thing. Right. So I think that's a biggest barrier for big law firms to pivot into a genuine AI native law firm. And for us, like small law firms or boutique law firms, because most of their work are already based on a flat fee. So it's much easier for them to pivot into something else. Um, and that's one point. And I also want to mention, there are still some other deeper reasons and they are all structural. Like first, how the money gets shelled out among the partners. So a partnership is built to hand cash out to the partners quickly. It's not built to hold onto it. So the fees come in, and they get split. And that structure in the big law firms makes it really hard to put a lot of separate teams together and commit to one big shield AI investment. Because every practice group is kind of its own little business and they keep their own books. Nobody's really incentivized to found something that helped the whole firm. Um, because honestly, the whole firm isn't aware of the money, it's felt. Um, that's, that's one reason. And I want to mention about something like this decision speed, like law firm decisions are very slow, especially for big law firms, like AI moves in generations of maybe three months. So those two clocks just don't sync up. And let me give you a a good example. Like this year, when Cloud for Legal came out, a lot of firms were just finishing rolling out, rolling out the chat GPT enterprise version. And because a year earlier, if you remembered a year earlier, when they started that whole buying process, OpenAI was clearly ahead of Anthrop pick. Yes, it was. And then it's not called for law in April. I thought, honestly, it just blew everything up again. Yeah. It's such a phenomenal, such a phenomenal leap step in its capability. Yeah. But no one is propelled for it. No. And if you're the big law firm. Yeah. Sorry, I cut you off. No. Yeah. It's so hard for them to catch up with the AI moves. So, I mean, it's like everything there has to have one year round, one year decision round. Um, so by the time you finish kind of building something you decided a year ago, it's almost made redundant by the thing that is now, you know, and like it, that's a tough position to be in. I know some really big law firms have decided to just create their own internal LLM. And there's, there's a lot of, um, a lot of good reasons why you might, you might do that. Uh, but it's so fascinating. And I, I'm interested in your view on how in-house teams are adopting AI because they have a fundamentally different set of incentives and it seems to be more aligned with, um, efficiencies. So what are you seeing in this space? Yeah. Yeah. I think I just got some idea recently and I think it's, uh, it's not the, as I imagined before, like I got the chance to talk with so many chief compliance officers during, uh, Baker Bowls event. And I asked a question like how many of you are using agents in your daily work? We've got over 30 people there and only three people read their hand. So, I mean, I can feel like AI adoption is very slow in, in house teams now. Mm-hmm. And so it's really interesting because previously I thought legal departments is a coastal center. Yeah. So everyone there should like, um, very excited about the improving efficiency, adopting AI, adopting agents, we should use the agents. But the, the, the reality is that the people kind of, it's just a still behind. Yes. Yes. I think they're so tired doing the, the work and hard to find the time to make the work better. Like you kind of have to give yourself that space to go, okay, we're going to, we're going to either like pause this and we're going to go, you know, away for a week and reconfigure our entire team system. And we'll be back, we'll be back in a week. I know that the, um, Canva team is doing that and others, but otherwise, if you can't really afford to just pause all legal for a year, sorry, for a week to, to lock step up, then you have to do it incrementally and you just have to get started. It was like when I implemented contract life cycle management, it was the same thing. It was never a good time. It was never going to be the right time to introduce new technology. Um, but my gosh, was it worth it? Okay. But yeah, I also, I also got a question like, because usually legal teams has a function called legal ops, but it seems that legal ops have so many works too, so they cannot support each one in their team very well. So, that's why maybe AR adoption cannot, like, as I imagined. So, I'm not sure if that's a... Maybe they're still learning as well and they're still, they're not willing to commit to anything just yet. There are risk averse generally as, as lawyers are. I think what I, what I suspect is that the company itself and it all, you know, all things lead from the top down, but how is the CEO requiring each department use AI? Is the, does, does the board want to see a certain level of adoption? Does the CEO, does this chief, you know, some companies are getting chief AI officers. It's early, but, um, chief technology, are they, are they pushing for it? Like, where is, what leadership at the top is, will probably be deciding how everybody reacts to it. And then I'm sure they'll start to see different hackathons internally and, you know, different, different challenges. And the lawyers will be so well-placed. They've got an incredible amount of drudgery and work in to, and repetitive work that can just be zhuzhed up and polished up and popped into a bot and, and FAQs and chatbots and everything. Like the, the low hanging fruit, as I say, is just, it's everywhere. It's all about, I guess, the time and then the incentive. And if they haven't got the incentive, um, or a very curious person on the team that wants to take the lead, then it's just going to bubble along. But it seems to me, and I'm sure you'll agree that the narrow for the, the window for taking real decisive action is starting to very much narrow. And that the firms and the teams that go in and take it seriously and at least stay curious are going to lockstep ahead into the next couple of years compared to those that don't. It, it honestly feels like the internet again, or even electricity. It's that big of, it is that big of a fundamentally like technological shift that, you know, eventually you'll all be using it because your life will just be so much better. And sure you can stay with facts and writing things down and sending letters, but you know, there's internet and there's email. And so eventually everyone will get there, but the ones that got there first got the biggest slice of the, the abundance and the opportunity to build something really interesting and, you know, really future fit. So yeah, we'll get there. It's a cycle like all things, right. Early adopters, the slower people, more conservative people. A lot of people don't like change. Um, and I, I think, I think I resonated so much with your content because I could see that you were at the forefront and you're living in the part of the world that in, in two parts of the world that are just, just charging ahead with enthusiasm and excitement. So, um, that's why we, yeah, we had to chat, but look, let's see, let's see what happens. We could make some predictions, but I don't know. I'm not sure. I want, there are some like, yeah, yeah, go ahead. No, I think there must be something about the security issues, the governance, right? Because every time I talk with some in-house counsels, they may, they may raise questions around this topic every time. Yeah. They care about, um, maybe it's, uh, not a mature point. It's not a mature product for me to use it. And, uh, sometimes they, because they cannot, like yourself scared of it. So they, they won't start it. So they think it's too hard to start it. Like they still think of vibe coding is something like something that only if you have technical background, you can do it. But actually, I mean, coding now is, it's a hobby for me. Yeah. Vibe coding is a hobby for me. It's not something like a only type person can do. Absolutely. Yeah. I mean, my Claude tells me what to use in Claude code. Like it's so easy. Have you tried it? Yeah. Yeah. It's really good. I was adding some MCP kind of some connectors the other day while Fable 5 was still free. So I wanted to, I wanted to get like all of my good data juices. Oh, so fascinating. Yeah. Yeah. It was a fun. It was so, you're also AI forward. As are you, my love, as are you. But I mean, I'm building, my new business is one year old and I've been building it on, on Claude just about the whole time. I have an entire content management system and operating system that runs in Notion, on the backend. And I have a lot of content that goes out across many platforms all the time. And I needed to, I needed something to, to keep a structure around that. And then it reminds me every day, okay, here, here's the things you need to do today for your content. So, you know, post this process, blah, blah, blah. And it just runs in the background and it speaks with Canva as well. So then it helps me with my graphics and then Canva to Meta Instagram connection isn't very good yet. I haven't figured that out, but then it wants to compost it to Meta. Oh my God. I'll want a human in the loop at one point or another because I have to just have that final sanity check, but it helps me with like 90% of my business and yeah, it's phenomenal. And it's basic stuff. I know what, I know I'm only doing the basic stuff and I'm already blown away. So I can't wait to see what others are building and, and we're all doing, we're all doing it. Like this was the year to start a new business. I think it's very fascinating if you were, if you were AI inclined. So let's see how it all plays out. I think tokenomics will be the next thing and we'll be spending all our spare money on tokens, but, but that's okay. We'll, we'll get there. Um, I'd love to ask your final question before we move into the underrated overrated section, which I always do just for fun. Just as a, just a bit of a closer, Helen, what is something that you believe about the future of legal work that most lawyers would push back on? Hmm. That's a good question. Yeah. I think maybe more and more lawyers will just a more, work more like a coder in their, yeah, in their day network. So, I mean, I noticed there's, there's a group called the legal crowns in our community now. Oh, they're great. I know about them. Yeah. And Jamie and Raymond, they are both of my, really good friends. Yeah. Is Raymond in Australia? Or did I know that? No, she's in Sydney. Yes. That's what I thought. Okay. Cool. Cool. Yeah. She's a really fascinating. And, and I want to say, so maybe one day we will have nailing legal crowns, you know, traditional law firm. And for every lawyer, you will have so many agents as your employees, just like Morgan and Cleo. Yeah. And so agent can teach agent, agent can challenge agent. Yeah. So like you just, uh, become the final person to give the approval. Yeah. And like now, um, there is a trend in engineering industry called looping engineering. Um, it means that agent prompts agents. So you have a loop and, um, human just, uh, sit behind of this and only when something bad happens, a human will jump into it. And most importantly, I think, uh, human will become someone who set up the standard or the criteria for the agents to work. So that's a basic idea of it. And I think, um, we, maybe we will have looping lawyering. Like, yeah, maybe, who knows? So I can see that. I can see that happening. Like, like, like right, right now in my AI native law firm, every time Cleo draft something, Morgan will help me to review it first. And Morgan will verify each source Cleo cited, if that's correct or not. So agent, supervise agent. Yeah. In my AI native law firm. So it's still, still, I mean, developing, maybe we all have some new format that we cannot imagine now, but, uh, it's really, really, really exciting. Yeah. Yeah. So exciting. I cannot wait to see where we're at even one year from now. We'll have to do a follow-up episode and see, see how the Y Combinator native law firms are going. Maybe some of them make it, some of them don't, who knows, but what an interesting time to be in this space. Helen, I have a final segment for you. This is, I do this with every guest. It's just a bit of fun. I'm going to say a few words or phrases and you're going to tell me whether you think they are underrated or overrated. Okay. Can I use my, like, some of them or like, yeah. Overrated. I'm not ready. Yeah. Okay. Do it, do it, do it. All right. Number one, the billable hour. Um, I mean, billable hour is overrated now because I think, um, people think big law firms can now, you know, people, billable hour will always work like this. So that's a belief, but I think it is challenged by clients and by any law firms. Yeah. And with AI, it's really hard for big law firms to continue doing like this. I agree. Okay. What about multi-agent setups? Um, I think based on my experiment, even if I have multi-agent on my platform, I will still give it, uh, overrated. Okay. Yeah. That's because I think AI moves so fast. So one day we will have something like AGI. Yes. And we may have like a comprehensive agent that can handle everything. So you don't need two agents. Okay. Got it. One agent will just, uh, generate more sub-agents to help with it. So you do not need like to set up two separate agents. One agent and a ton of sub-agents may handle everything in your work. The reason why I split two agents on my platform is because, um, now the agent cannot handle very long contacts. So if you want to make it, uh, follow the rules very closely, it's better for you to split it up. So yeah, that makes sense. Okay. What about LinkedIn? LinkedIn, linking, linking, linking underrated. Because I, yeah, that's how I really, we met up, right? In the first personal place. Yep. I agree. What about, um, I had one more here. What about La Boo Boo dolls? Oh, I actually, I got one. Yes, Queen. I knew. Oh my God. Let me get you. You know that? I can have three. Yeah. I have three, but I, yeah, it's just the one on my table. So it's I can't take it like it's hers. Are they so cute? I feel like, I feel like they're getting a bit like overrated though. I don't know. What do you think? Yeah, I think, no, I think it's underrated. I like it. I think it is. Yeah. Oh, yours is so cute. Mine is so grotty. I need to clean. My little niece was playing with her. Okay. Okay. You need to get into Juju. Juju is the new La Boo Boo. Okay. Um, I'll, I'll, I'll, I'll, I also got this one like, oh, that's so cute. Oh, funny boy. Oh, nice little face. I don't know his name. Yeah. Yeah. I love, I'm just a big kid. I just love collecting these little things. Uh, all right, let's go with, um, one final one. Claude Fable 5. Underrated or overrated? Um, based on mine, experience, I think it's overrated. Yeah. Okay. It's, it's, I think it's just, uh, because it costs a lot. It, it takes so, so many tokens. And, and I don't think it has created some really like, um, unreal things. It's, it's still like, uh, advanced version of OPRS 4.2, uh, 4.8 to me. Got it. Yeah. So maybe for, maybe for engineers, they're, they have different stories. Yeah. And like, and the numbers and financial analysts, they love it, but I don't need to use it for that. So. Oh yeah. That, that, that, maybe that's a reason because they, they're gonna have an IPO. So they need a really good story to convince their investors. Exactly. That's exactly. Oh, you're so smart. Helen, thank you so much for joining us all the way from Silicon Valley. You are in the heart of, of all that is cooking in legal and in AI. And we just so appreciate your time. I'm so glad that we connected and let's see, let's see where things are. Yeah. Yeah, we should. Yeah, yeah, yeah. Yeah. We, we should, uh, keep in touch and we should check it from time to time. Yeah. Absolutely. I would love that. And hopefully we can meet in person one day at one of these really cool events that you're holding for your community. So we'll see how we go. Thank you so much. Thank you,