The B2B Podcast Index
Index
All categories
MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
MethodologySubmit
Best of:MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
An independent project byFame
SearchBest episodesGuestsInsightsMethodologySubmit a podcast
Index/ABA Journal Podcasts
ABA Journal Podcasts artwork

What is vibe coding, and how could it help lawyers?

ABA Journal Podcasts · 2026-06-17 · 34 min

0:00--:--

Key moments - from our scoring

Substance score

52 / 100

Five dimensions, 20 points each

Insight Density10 / 20
Originality11 / 20
Guest Caliber12 / 20
Specificity & Evidence12 / 20
Conversational Craft7 / 20

Jamie Cho, formerly of Clifford Chance and founder of Legal Quants, explores vibe coding - using AI tools to generate software through natural language prompts - and its practical applications for lawyers. Cho argues that lawyers shouldn't attempt to build expensive platforms like Clio or Litera, but rather focus on small, high-leverage tools that solve specific practice pain points: redlining comparisons, contract simulation, and signature page collation. The episode covers Cho's journey from frustration with manual document review to building lightweight legal tech using platforms like Replit and Claude. Critically, Cho emphasizes that lawyers should accumulate proprietary assets - custom prompts, evaluation datasets, and workflows - that give them competitive advantage, not commoditized software. The discussion covers essential safeguards: never use client data, comply with professional obligations, leverage synthetic data for testing, and focus on building what only your practice requires rather than replicating existing tools.

Key takeaways

  • →Vibe coding allows lawyers without technical backgrounds to build custom tools by describing their needs in plain English, abstracted away from complex programming syntax.
  • →Lawyers should focus on building small, specialized tools that solve specific practice pain points rather than attempting to replicate expensive enterprise platforms like Clio or Litera.
  • →The real value for law firms is in owning customized prompts, evaluation datasets, and interfaces that capture their unique practice expertise, not in copying commoditized software.
  • →Lawyers must never use actual client or firm data when building vibe coding applications; synthetic data and public materials should be used for testing instead.
  • →Building tools should be part of a deliberate strategy to gain competitive advantage, similar to how Kirkland partnered with Palantir to build funds practice platforms, not simply to reduce costs.

In this episode

  1. 1Introduction to Vibe Coding and Jamie Cho's Background
  2. 2Jamie's Journey from Law Practice to AI-Assisted Development
  3. 3How Vibe Coding Works for Non-Technical Lawyers
  4. 4Building Practical Legal Tech Tools: Redlining, Contracts, and Signatures
  5. 5Strategic Approach to Vibe Coding: Focus on Leverage and Practice Differentiation
  6. 6Avoiding Common Pitfalls: When Not to Build and Data Protection Best Practices

Mentioned

American Bar AssociationLegal Talk NetworkABA JournalLegal Rebels PodcastThe Jury is OutSimon LawClifford ChanceLegal QuantsUpworkPower AutomateReplitJamie Cho

Guests

Jamie Cho

Topics in this episode

PalantirGeminiClaudeReplitVibe codingClifford ChanceLegal QuantsLiteraSpellbookMicrosoft Power Automate

Questions this episode answers

What is vibe coding and how does it work for lawyers?

Vibe coding is using AI tools to create software through verbal prompts, allowing lawyers to describe what tool they need in plain English and have the AI write the underlying code. Lawyers can create custom applications like contract reviewers, redlining tools, and document generators tailored to their specific practice needs without requiring traditional programming skills.

What types of tools has Jamie Cho built using vibe coding?

Cho created several legal tech tools including a plain text comparison tool that generates legal-grade redlines in milliseconds (faster than commercial tools like Litera), a contract simulation app that uses AI to identify edge cases through adversarial scenarios, and a signature page collation app that extracts and organizes signature pages by counterparty and signatory.

Should lawyers try to build their own versions of platforms like Clio or Litera using vibe coding?

Cho advises against it, arguing lawyers should only build tools that give them disproportionate leverage and competitive advantage. Instead of replicating expensive platforms, lawyers should focus on small productivity tools addressing specific pain points in their practice, and build proprietary assets like custom prompts and evaluation datasets that cannot be easily replicated.

What data safety precautions should lawyers take when building tools with vibe coding?

Lawyers should never use client or firm documents for testing; instead use publicly available materials or create synthetic data. All professional obligations regarding confidentiality and data protection apply when building custom tools, just as they do with any other practice technology.

What background or technical skills do lawyers need to start vibe coding?

No technical background is required - Cho emphasizes that lawyers' strong writing and drafting skills actually make them well-suited to vibe coding since they can write detailed prompts. Experimentation and learning through trial-and-error is the recommended approach rather than memorizing coding best practices.

What our scoring noted

Our reviewer’s read on each dimension, with quotes from the episode.

Insight Density

10 / 20

The episode contains a handful of genuinely useful, non-obvious ideas - particularly the adversarial contract simulation concept, the AGPL copyleft risk for vibe-coded tools, and the asset-accumulation framing - but they are diluted by multiple ad breaks, a meandering intro comedy bit, and generic AI-will-change-law platitudes. The density of actionable insight per minute is moderate at best.

adversarial review, essentially using AI to simulate different Personas in a contract and have them argue against each other and based on those simulations, find out the edge cases and gaps in a contract
knowing how to build is not as important as knowing what to build. That means software is becoming commoditized as well

Originality

11 / 20

A few ideas stand out as genuinely fresh for a legal audience - the adversarial multi-agent contract review, the securities-fraud auto-detection workflow example, and the 'unicorn lawyer' solo $10M practice thesis - but the broader framing of 'AI will disrupt law and lawyers should upskill' is thoroughly recycled discourse.

There's a special uh, counsel at a AM10 law firm who created this solution that monitors the stock prices every single day and uh, detects which one drops the most and then automatically dispatch AI agents to look for evidence of security fraud and then automatically draft a statement of claim
we will see unicorn lawyers that are with a, uh, one man ban create a $10 million practice

Guest Caliber

12 / 20

Jamie Cho is a credible practitioner who actually built and shipped legal-tech tools inside Clifford Chance and then publicly, making him a genuine operator rather than a thought-leader tourist; however, he is relatively junior (senior associate level) and his community is nascent (140 members), limiting the weight of his claims about firm-wide or market-wide dynamics.

I hired some engineers on UPWORK to build some small language models to help with uh, entity mapping
the tablet review tool it only took an afternoon to build. So that was the tool that got a lot of attention. Kind of like now people look back and say that as, you know, the moment where sparked the vibe holding movement among lawyers

Specificity & Evidence

12 / 20

The episode is reasonably well-evidenced for a podcast of this type - named tools (Replit, Litera, Spellbook, Power Automate), named firms (Clifford Chance, Kirkland), a concrete partnership (Kirkland + Palantir funds platform), specific open-source licenses (AGPL, MIT, Apache 2.0), and build-time estimates (afternoon, two days) - but lacks hard outcome metrics or rigorous data to back the bigger claims like the 80/20 value-concentration thesis.

Kirkland and Palantir, they uh, together built a funds platform that automate silatic drafting and LPA drafting
the tablet review tool it only took an afternoon to build

Conversational Craft

7 / 20

The host asks structurally reasonable questions covering use cases, risks, IP, and data safety, but consistently fails to probe or challenge: key assertions like the 80/20 thesis, the Kirkland/Palantir claim, and Jamie's non-expert IP commentary all go unchallenged, and 'gotcha' serves as a near-constant filler response that signals passive listening rather than genuine interrogation.

Yeah, but do you think lawyers might try and do what we were talking about, like either building new platforms or kind of trying to get an advantage that way
Well I think the fact that it's such a gray area just kind of Speaks for itself

Conversation analysis

Computed from the transcript - who did the talking, and the words that came up most.

Share of words spoken

  • Jamie Choguest71%
  • Victor Leehost22%
  • Narrator7%

Most-used words

lawyers41legal34build29coding23tools22building22practice20vibe20data17tool16code14create13help13codes13first12case10

Episode notes

Defined as the practice of using artificial intelligence tools to create software through verbal prompts and then having the AI write the underlying code based on those prompts, vibe coding has plenty of applications and implications for lawyers. Lawyers can create all sorts of programs and applications to help make them more productive. But of course, with any powerful technological tool, there are some risks and things you should be careful of.

Full transcript

34 min

Transcribed and scored by The B2B Podcast Index.

Narrator: The American Bar association provides access to career changing and life changing opportunities. Invest in your growth, deepen your knowledge and join us in our pursuit of making a positive impact for all. Uh, the American Bar association

Victor Lee: the strongest trial attorneys never stop learning. One of your most valuable resources is the Jury Is out, the podcast that brings together veteran litigators, constitutional scholars, patent experts and nationally recognized practitioners to sharpen your courtroom skills, deepen your strategy and strengthen your practice. The verdict's in. This is Must Listen on demand, game changing access all guided by high level attorneys John Simon, Tim Cronin and Eric Veith. The Jury is Out as a production of Simon Law and is found on Legal Talk Network and wherever you get your best podcast.

Narrator: Welcome to the AVA Journal Legal Rebels Podcast where we talk to men and women who are remaking the legal profession, changing the way the law is practiced and setting standards that will guide us into the future.

Victor Lee: I'm so old I remember a time when people associated AI with Allen Iverson or a Steven Spielberg movie. I'm so old I actually used those AOL CDs that you used to get in the mail to access the Internet. Heck, I'm so old I know what CDs are. Hold for applause. Or not. What was the point of that bad stand up act? Well, when I first heard of Vibe coding, I put on my old man hat and figured it was something the kids were doing that would pass in time. Like some TikTok dance craze or that weird 67 thing that I still don't get. Turns out Vibe coding is a pretty big deal, especially for lawyers. Defined as the practice of using AI tools to create software through verbal prompts and then having the AI write the underlying code based on those prompts. Vibe coding has plenty of applications and implications for lawyers. Lawyers can create all sorts of programs and applications to help make them more productive and custom build them to fit their practices and needs. But of course, with any powerful tech tool there are some risks and things that you should be aware of, and that's especially the case when it comes to Vibe coding. My name is Victor Lee and I'm Assistant Managing Editor for the ABA Journal. I am the host of the Legal Rebels podcast and my guest today is Jamie Cho, formerly a Senior Associate at Clifford Chance. Jamie helped popularize Vibe coding in the legal sphere, using AI to create several programs to help lawyers be more productive, including contract evaluation, redlining and document generation tools. Jamie, who founded Legal Quants in March, is here to talk about Vibe coding and how it can help lawyers and what Legal issues they should be aware of. Welcome to the show, Jamie.

Jamie Cho: Thank you for having me.

Victor Lee: Yeah, so obviously I just gave the very quick elevator version of your bio. Can you tell me a little bit more about yourself and your background?

Jamie Cho: Sure. So I'm born and raised in Hong Kong. My first job was in a local law firm called Deakins. And I remember when I was a second year training there I got so frustrated, uh, by the work that I was doing. Uh, it was annotating the funded prospectuses against the content requirements in appendix D of the UT code. In Hong Kong. I became so frustrated that I hired some engineers on UPWORK to build some small language models to help with uh, entity mapping so that I can learn of define those paragraphs. It's almost like a, um, advanced control F. That was my first AI experience when it was still called machine learning. Uh, it was still using something called TensorFlow, Bert Whatever to build them. Uh, luckily I didn't turn this into a product, otherwise I'll get absolutely killed and steamrolled by GPT3. But then I switched to Clifford Chant around four to five years ago and I think two to three years in again I felt super, super bored because the learning curve flattened and so on. But you know, uh, funny enough at that time, uh, I started building some AI enabled workflows, uh, within my team. I shared some of the tools I've built, uh, using Power Automate, which is a no code platform offered by Microsoft. And surprisingly my colleagues really love them. So that was my first time actually building something that uh, Lloyds, you know, use. So I learned that all by myself, you know, and it was a no code platform so this one's pretty easy to pick up. But of course the, with all these no code platforms they're pretty limited, you know, their workflows is almost like Lego. And with Lego you basically have to follow the instructions, right? Uh, there are only so many holes, there's only so many pieces that you can piece together. So I quickly found out that, well, I can't really do much with that platform. So that's when I first started to explore vibe coding. And um, with 5 coding it's basically like there's no rules and you can piece anything together and the building blocks are essentially really low level codes and not some modules that are uh, packaged, uh, and offered on these no code platforms. Which means you've got so much more flexibility to build stuff. And basically the programming language would be English. And I remember the first tool that built using this technique was my wife's Clinic, uh, website. So I started building on a platform called Replit, which is now, I think the largest web coding platform. And it was really good at designing the front end of the website. But soon enough I found that, well, I can actually use that to also build legal tech tools. And I think at one point Gemini 3 came out and I saw on threads that a lot of people used it to replicate a lot of the classic legacy software like Windows 97, you know, creating your own YouTube and, uh, almost like as a benchmark on how good these models are, or I kind of ran with that idea and then, uh, built some lightweight replica of these other, uh, most famous legal AI tech tools out there, including Spellbook, including offering the Gora's tablet review tool was really easy to build, especially the tablet review tool it only took an afternoon to build. So that was the tool that got a lot of attention. Kind of like now people look back and say that as, you know, the moment where sparked the vibe holding movement among lawyers. And I went on to, in rapid succession, also, uh, created quite a number of tools, posted them online. So that's how I first got, you know, I got started and how, how it went. Uh, and these days I, I'm running this community, a network of lawyers that are essentially building their own tools to practice laws. So that's, uh, how. That's my journey.

Victor Lee: So take me through sort of like what that entails. Is it just you type in, hey, give me a version of Microsoft Word and then it creates it? Or are there more specific prompts that you have to put in, like, I want this, I want this, I want this, and then do this kind of thing.

Jamie Cho: So I would say you should start with something that you. It's almost like seeing your wish, you know, before five coding. I think lawyers are, uh, in this, in the backseat of legal tech because they're the users but not the creators of legal tech. But they will always have these moments where they say, oh, uh, I wish there is a tool that does xyz. And now I would say the easiest way to start vibe coding is to type out that XYZ and then put it into the tool, the Vibe coding tool, and see what happens. So, interestingly, because lawyers, they are really good at drafting, they can usually come up with detailed enough prompts, uh, to describe what they need. They're good at writing instructions to their clients to their juniors. So they usually have no problem trying to describe what they want. And I wouldn't even complicate how to Vibe code because the whole idea behind it is to let your creativity kind of run wild. And uh, if you don't know what you don't know, usually you're able to boil the ocean in the sense that you can easily experiment limits of these tools. And as a result lawyers, uh, usually surprise, uh, technical people, including engineers, as to what they're capable of making. So I wouldn't necessarily try to say, well, uh, these are the best practices for vibe coding. In fact, I would say try to experiment with whatever to find out the limits of these tools as soon as possible. And only if you're stuck or you encounter some bottlenecks, then you just discover how to actually get around them. So yeah, that's how I would go about it.

Victor Lee: And do you have a tech background or um, were you always just very law oriented?

Jamie Cho: No, I tried to learn Python when I was in uni, but then I gave up when it asked me to do some data frame stuff where I need to learn how to merge tables. And I thought, well, I'm a lawyer, why I'm studying law, why do I need to merge tables? Why do I need to do big data analysis and learning SQL? Because back then when I was in uni, uh, everything is about big data and people, uh, were learning Python to crunch numbers. So it didn't really interest me. But AI is really different as a language model, right? So it's very, very different. I would say I picked up a lot of tenable skills along the way because of coding or AI assisted coding because you simply have to debug when these things break, right? So you tend to pick up a lot of um, new skills along the way. So right now I would say I'm half technical, but the frontier moves really quickly. So what a good Engineer looks like 6 months ago is different from what a good engineer looks like now. So even now you need to be extremely talented to become uh, an AI engineer. So to be hired by the big labs, right? So I would say just because you can five go doesn't mean that you are technical. So, um, I would still say that. And you know, non technical in that sense.

Victor Lee: I definitely see what you're saying because I mean, I think I took like a very low level computer science class when I was in college and just like seeing like some of the more advanced, I mean it seems very inaccessible, especially for people if they don't have that background. It can be very intimidating, especially if you see like all those lines of code and you have no idea what any of it means. It's not even in real, in Plain English most of the time. So vibe coding kind of cuts that out. It makes it more accessible for people, especially lawyers, who might not have that kind of technical experience.

Jamie Cho: Right, Absolutely. Yeah. I think that, um, abstracts away a lot of the syntaxes, so you don't need to remember a lot of the syntaxes that you otherwise need to remember.

Victor Lee: So for your apps, what gave you the idea to create them was just things that you saw from your practice that you was, oh, I wish I had that tool. Or, oh, um, uh, I could have used this when I practiced. Or was it more just like, okay, well, a lot of people want, like a document generation tool. A lot of people want a redlining tool, so I'm just going to do that. Like, what was sort of the impetus behind your apps?

Jamie Cho: There are a few small tools that I feel are just solving some really small and minor pain points that no big deal AI companies are solving. Uh, and that includes something, uh, like a plain text comparison tool where you put two strings of text into text boxes and it will generate a red line. And that red line is basically the legal grade red line that we usually see. But the thing about, using, uh, some commercial tools like Litera, because of how it's designed, it takes forever to generate just a, uh, compare between two strings of text. The solution I've created, although it's not using the same algorithm, this generates it within milliseconds. So that was extremely popular among lawyers. It turned out to be quite popular because it's a small tool. They use it all the time, every single day. The other tool that I built is more sophisticated. It's a contract simulation app. Uh, because I was thinking when we use AI, we should be using it, using it in a way that reimagines how we practice law, but not just using it to make our workflows faster. So I kept thinking, well, what is that thing, uh, that is now possible, that wasn't before AI, and that is basically adversarial review, essentially using AI to simulate different Personas in a contract and have them argue against each other and based on those simulations, find out the edge cases and gaps in a contract. So I got that idea actually from practice. You know, a hedge fund client actually asked me, but right before he signed a SAT letter with an institutional investor. So he asked me, well, can you confirm that this contract is commercially watertight? It's in my mind that, uh, means I need to look into like 14 million possible futures to ensure that the whole thing is bottom type. And I just couldn't give them that advice. So that was the kind of use case that I think AI would be able to help. So I put a really lightweight version of that then kind of uh, will generate different scenarios from best case to worst case and then have the agents argue against each other until they come up with these issues and gaps that could be addressed. So yeah, so these are just two examples. I also created a signature collation app. So it's a well known pain point. So just need to, especially in the banking team, they always need to extract these signature uh, pages out of documents and sort them by counterparty by signatory and then ah, sip them. Right. So it was really simple to be honest. But then I kind of created a mini app that extract all these signature pages uh, and package them into one PDF. So it was really simple to build. It just was built within two days. So that was really popular as well.

Victor Lee: Gotcha. Before we continue, let's take a quick break for a word from our sponsor. We asked more than 1,000 legal clients if they thought their lawyer was competent

Jamie Cho: and got the unfiltered truth.

Victor Lee: Find out what your clients really think. See the data on what really impacts the client experience in the ACE report from InfoTrack.

Jamie Cho: Download the full report for free at,

Victor Lee: uh, InfoTrack.com/LTN that's InfoTrack.com/LTN

Jamie Cho: Imagine getting a free marketing consultation or workshop for your law firm every month without ever booking a meeting or paying a dime. Sounds too good to be true, right?

Victor Lee: They don't have to imagine it, Conrad,

Jamie Cho: because they can actually experience it.

Victor Lee: Our podcast, Lunch Hour Legal Marketing, which you can watch right now on YouTube, has dozens of episodes explaining how to update your Google business profile, what kinds of campaigns are working for lawyers, and what SEO strategies you should implement.

Jamie Cho: And that's just scratching the surface. Now, there are a lot of people out there who think they know the ins and outs of SEO, social media and paid marketing, but in reality they're just greedy or grifters or faking it most of the time. As agency owners, my co host Guy and I know what we're talking about and would love for you to be part of our growing community.

Victor Lee: Check out our library of episodes and subscribe to the Lunch hour legal marketing YouTube channel or wherever podcasts are found. And we're back. So let's talk about what vibe coding could do for lawyers. So how do you think lawyers should use it? If they want to create their own apps for their own practice, should they rely on them or are there things they should be careful of when they're creating these tools.

Jamie Cho: The way I see firecoding is simply an advanced way to use AI and I think that is a symptom of becoming extremely fluid in AI. I don't think that lawyers need to build tools to do everything. I think that they should only build tools that will give them disproportionate amounts of leverage. The reason I say that is because as coding agents become, uh, better and better, knowing how to build is not as important as knowing what to build. That means software is becoming commoditized as well. So I would say because software is getting commoditized, you can basically use something like Claude, just build something up for you on the fly to complete your tasks. And even that on its own is some sort of vibe coding done by Vlad mid flight. So I wouldn't necessarily learn how to code something that is easily replicable by uh, the frontier models. But I would however look at my current practice and see what is it that you know, you need to build to give you leverage that will make you a 10x or a hydrated better lawyer. And I think that starts from trying to, you know, productize, uh, your own practice. I keep coming Back to how 20 years ago lawyers scale their practice by creating templates by creating a standardized precedent, uh, where every single deal that they do, they find the edge cases and decide whether to incorporate a new clause back into the template. And ah, that becomes their asset. Right. To my mind, I think lawyers should be in the business of accumulating assets and uh, a small tool here and there that grants a little bit of productivity that might help. But to really harness the power of software, uh, I think that they should be owning the layer that will really scale their practice. And uh, using today's terminology, it would be owning for example sets of prompts, uh, scrapers, evaluation, benchmark, data set and the interface that uh, brings everything together. And that is something that could be hosted on these LLMs. We think that, well, these templates 20 years ago, they live on Microsoft Word, right? And these days our assets will also need to live on a service. And it cannot just be a service that is hard to use, right? Microsoft Word is a really popular software. So anything that if you send a document that the other party cannot open, you can't really uh, scale, uh, that way. So I think that the same goes to tools. So I would say my honest take is that a lawyer would need to trace things that uh, are differentiable from other lawyers in terms of actual practice and host it on a Surface that allow them to easily scale those assets. And you know, to my mind, I think that, you know, even if you're building on Copilot, building on Claude, those are the services that allow you to customize a lot of your stuff. But what you own essentially is the, uh, prompts, right? The evaluation data set or, uh, whatever. And those are the stuff that are worth building. So I would say vibe coding is just a sim M or a sign that you should start building. But what they should build should be something related to their practice.

Victor Lee: Okay. Because obviously one of the things that people keep harping on is that this idea of, okay, well, what if lawyers were going to create their own versions of popular software like Clio, or you mentioned later, or things like that, and then kind of run their practices based on that, so that way they save money and blah, blah, blah. Is that really doable though? Because it seems like that would potentially open yourself up to like, potential problems. You know, if something goes wrong with the code or something goes wrong with one of the prompts or whatever like that, and then you're relying on it to, to run your practice. That seems problematic to me. But is that the case or am I kind of misinterpreting that?

Jamie Cho: I think that lawyers should focus on building something that only they could build. For example, to build another Clio. Unless there is a really pressing need to build, then, uh, I wouldn't necessarily build it that way. The reason why small tools like the signature page collation app, um, test converter app, I think the reason why they're popular, the reason why they should be built is because those are the small point tools that if you don't build, no one else would be interested in building. Because everyone out there is getting feasy money to eat, uh, the entire vertical. At least that's what they lead people to believe. But I think that for small tools, small productivity tools, because the barrier to entry is so low, I think that's worth building. But if you were to build a large platform from scratch just so that you can match it, unless there is passing reason, uh, to do that, then I would not necessarily, for example, if I were to build another litera, I would only do it if the cost stops making sense or essentially becomes so the coding agents become so easy that actually building it gives me so much, so, uh, much more benefits then I would do it. But otherwise I, uh, would just focus on becoming the architect. Like creating new workflows, thinking about the system, thinking from the client's perspective, what will be the best User experience. When they're interfacing the lawyers, should they be interfacing the tools? Should they be interfacing the lawyer? At what point? That is what lawyers actually add value. That's where we come in to say that, well, uh, based on my experience, this should be the best possible client experience and only I can build that. And I would focus on building some uh, the things that uh, requires a lot of domain expertise to build.

Victor Lee: Yeah, but do you think lawyers might try and do what we were talking about, like either building new platforms or kind of trying to get an advantage that way so that, you know, again, so I mean, I mean the money savings would be, would be a big deal. Do you think they might just try to do that anyway? Even if it wouldn't necessarily, you know, help them scale the practice? They might just do it just to, just to save money or to uh, you know, help them, you know. Yeah, like kind of save costs that way.

Jamie Cho: I think that law firms, at least the ones I've talked to, they are in the business of creating an advantage for their clients or creating an advantage for themselves. And they can only create advantage if for example, uh, their advice is data driven faster or they are able to deliver super intelligence. And to my mind as of today, I think a good example of what building professionally looks like for a law firm would be how Kirkland and Palantir, they uh, together built a funds platform that automate silatic drafting and LPA drafting and so on. And everyone knows that in Kirkland has a uh, world leading funds practice. Uh, they have a lot of data and that's. And Palantir is actually really good at products called ontology. Right. So really good at data labeling. So that is a really good case to build stuff because from uh, Kirkland's point of view it kind of prolongs their uh, monopoly of the funds market. Right. So I would say that's a good use case. It's a good example of the build strategy which is used to augment your capabilities. But I will not necessarily. If the 500 million was basically used to build something that does not really give them more advantage and they can basically rebuild something that puts them at the same level as other law firms. I don't think that would, that's a good use case.

Victor Lee: Gotcha. And so I mean you mentioned data. So what are some things that lawyers should be careful of if they are going to vibe code? Because obviously you got to safeguard your data, you got to protect your information. What are some things that um, you would recommend that lawyers do to kind of safeguard that stuff as they're creating their apps.

Jamie Cho: So I would say first and foremost, do not use fine data. Um, I think that's a given. I would say, uh, all the professional obligations to apply when you're building your own thing. I would say there are plenty of materials online that you can use to test your apps. You don't have to take client documents, you don't have to take firm documents. You can even create your own synthetic data. So we can see from the Harvey Lab benchmark, there's a lot of synthetic data and documents that they have created. You take them. So that's the first thing. The second thing is hosting, right? I would say there's a case for building your own stuff because if the data lives within your own private cloud, then, uh, within your own environment on your private cloud, then arguably it's actually safer. And if you were to use a third party app, which is essentially a uh, black box, unless they uh, make it transparent and you have little control or physicality over how the data is being managed. Whereas if you build in house, then you actually have got more control. You can control where your data lives, live locally, which is safer than if, uh, it's living in a public cloud. You can host it internally in your own private cloud. You can integrate with your own dms. So I think there's definitely a case of building internally that gives you more visibility, more transparency and more control.

Victor Lee: Uh, let's take another quick break for a word from our sponsor and we'll be back soon.

Narrator: I'm Kat Moon, co host of Talk Justice. Did you know that low income Americans don't receive meaningful legal help for 92% of their civil legal problems? This crisis is called the Justice Gap. Whether it's employment, consumer or medical debt, divorce and child custody, or landlord tenant disputes, these are serious problems. But there simply aren't enough legal aid lawyers to help everyone. So we must innovate. The Leaders Council of the Legal Services Corporation invites you to get in the know about all things access to justice by listening to Talk justice, where leaders and innovators come to discuss life changing projects, new technology and big ideas. Hear how lawyers, technologists and other professionals are helping address medical debt, neighborhood blight, natural disasters, domestic violence, and so much more by subscribing to Talk Justice.

Jamie Cho: You like legal podcasts because you're curious and want to be the best attorney you can be? I'm Dave Scriven young host of Litigation radio, produced by ABA's litigation section with legal Talk Network Search in your favorite podcast player for Litigation Radio to join me and my guests as we examine hot topics in litigation and topics that ah, will help you to develop your litigation skills and build your practice. I hope you'll check out Litigation Radio and, and join the ABA Litigation Section for access to all of the resources, relationships and referrals you need to thrive as a litigator.

Victor Lee: And we're back. So can you talk a little bit about the IP issues relating to programs that are coded through vibes like anytime. You know, obviously with AI generated stuff there's question as to who really owns it. So what are some of the IP issues that lawyers should be careful of?

Jamie Cho: I need to first of all say that I am not an IP lawyer.

Victor Lee: Right.

Jamie Cho: But I can. So please do fact check my statement. But according to my based, uh, on what I understand, there seems to be cases saying that AI generated codes are not copyright protected and only the prompt history can be protected. So that's the first thing. It doesn't really change the picture because I think lawyers, they can still own the output as long as the prompts that are original, you know, the idea is original, uh, and so on. At least that's how I understand it. I think that there, there's a lot of debate in that area, so I don't want to go there. But let me come back to something that is more ah, obvious. For example like copy left, um, and essentially codes that are actually not fully open source. So there's different licenses out there. There's MIT, there's AGPL, Apache, uh, 2.0 and for example licenses like AGPL, uh, sometimes it's ran as the poison pill because if you adopt those codes, which is HTPL protected, whatever enhancements, modifications that you do to those codes, you will need to contribute back upstream to the original repository. So what happens is if you use HTPL codes, for example, within the frame environment and kind of customize it and make it better, uh, then it becomes problematic because you would be legally required to essentially contribute those codes back to the original repo, which may m or may not be what you actually want because uh, you might have dedicated a lot of resources in developing those codes or even those codes may be proprietary. So that is something that I would watch out for. But on the AI generated, with the AI generated codes copyright protected, I am unfortunately not educated enough to comment on that. And I think that that is uh, itself really controversial as well. Yeah, yeah.

Victor Lee: Well I think the fact that it's such a gray area just kind of Speaks for itself. And it's probably an issue that we will see as it develops. Sort of like how, how things go and like, you know, and you know the law has, has yet to like really kind of. Yeah, definitively I guess.

Jamie Cho: One comment I would make though is a lot of the codes these days are AI generated anyway. Like anthropic. They say that like their engineers are using uh, cloud to write cloud itself. Right. The Harvey engineers, they say that, well we're no longer writing codes manually. You know, we're using AI to uh, code. So the argument is always that to the extent you are using your own original ideas to prompt, then you also own those codes that you know, the ideas generate. So that's how I would think about it.

Victor Lee: Gotcha. So where do you kind of see the whole Vibe code thing going? I mean, do you think more lawyers will start utilizing it?

Jamie Cho: I think that that's for sure. I would encourage everyone to even or try out building things. I would say. Like I said, Vibe coding is a symptom of a lawyer becoming very, very fluent in AI. And I think that the more interesting shift is not really about the tools, about the vendors, it's about the people. It's about first time in history lawyers themselves have harnessed the power of software as an extreme leverage. Law firms have always scaled through people and more headcount, creating templates and hiring juniors to actually execute on those templates. The more interesting thing is for lawyers that are actually able to code, they can easily own the full stack from being the architect, to building out an AI native practice, to commanding their own agents. And that will lead to a very interesting development where we will see unicorn lawyers that are with a, uh, one man ban create a $10 million practice. And I'm really looking forward to that future. So I founded this community called Leapquants. Uh, that community was founded on a really simple thesis that we think that 80% of the economic value generated in the legal market will come from the top 20%. Most AI fluent lawyers because they have leverage and know how to use software to scale their practice. And um, the community exists to kind of create and find those new markets for these type of lawyers. And I can show you, give uh, you a few examples. There's a special uh, counsel at a AM10 law firm who created this solution that monitors the stock prices every single day and uh, detects which one drops the most and then automatically dispatch AI agents to look for evidence of security fraud and then automatically draft a statement of claim and send it to potential plaintiffs that means a new market that's being monopolized and dominated by this lawyer who happens to create a system that automate the work there. Right. That is more interesting than asking a question. Well, can I create a 5 quota solution and use it internally in a big law firm? Of course not. Right. Imagine the hooks that you jump through. But the more interesting thing is what if that lawyer is good enough to create something that he owns and then, uh, he not only owns the product, he also owns the distribution as well. So I would say vibe coding is simply a representation of how the most advanced lawyers operate these days. So I think that if we look two years out, three years out, I can see that lawyers, uh, that actually own their own solutions. The lawyers will become more and more like product owners. They will have their own distribution, they will have clients that will use their tools autonomously. They will have solutions where maybe thousands of hundreds of thousands of people use, but then wrap services around that tool. So I think that's the whole premise of AI native law firms. But I don't even think that AI, uh, native law firms need to exist at a firm so we can potentially see so practitioners actually reaching that fast on their own.

Victor Lee: Gotcha. And so what are things that you're working on right now that you would like to talk about?

Jamie Cho: Sure. As I said, I'm currently running this community called, uh, Legal Quant. We find we vet, connect and uh, deploy these nodes that we think are the most AI native. The reason why they come to us is because they have been building in silo. We give them, um, the feeling that they have found their people. We currently have around 140 people in the community. We have around 6 to 700 in the waitlist. Because we're trying to define the standard of what it means to be a legal quant. And every week the community is extremely active with 2000 messages per week. And then we distill a lot of these community knowledge into a curriculum that will soon enough define the standard of what it means to be a native and AI forward. And uh, yeah, we work with different companies who might want access to this new category of talent. So that's something that we have been

Victor Lee: working on to wrap up. If our listeners wish to get in touch with you, what's the best way to do so?

Jamie Cho: They can follow me on LinkedIn. They can also reach out to me at j.tsohlegalquants.com all right, well, hey, thank

Victor Lee: you for joining us today, Jamie. I really appreciate it.

Jamie Cho: Thank you so much. Really appreciate it. Too.

Victor Lee: If you enjoyed this podcast and want to hear more, please go to your favorite app and check out some other titles from Legal Talk Network. In the meantime, I'm Victor Lee and I'll see you next time on the ABA Journal Legal Rebels Podcast.

Narrator: If you'd like more information about today's show, please visit legalrebels.com, legaltalknetwork.com subscribe via iTunes and RSS. Find, um, both the ABA Journal and Legal Talk Network on Twitter, Facebook and LinkedIn. Or download the free apps from ABA Journal and Legal Talk Network in Google Play and itunes. The views expressed by the participants of this program are their own and do not represent the views of nor are they endorsed by Legal Talk Network, its officers, directors, employees, agents, representatives, shareholders and subsidiaries. None of the content should be considered legal advice. As always, consult a lawyer. I'm Kat Moon, co host of Talk Justice. Did you know that low income Americans don't receive meaningful legal help for 92% of their civil legal problems? This crisis is called the Justice Gap. Whether it's employment, consumer or medical debt, divorce and child custody, or landlord tenant disputes, these are serious problems. But there simply aren't enough legal aid lawyers to help everyone. So we must innovate. The Leaders Council of the Legal Services Corporation invites you to get in the know about all things access to justice by listening Listening to Talk justice, where leaders and innovators come to discuss life changing projects, new technology and big ideas. Hear how lawyers, technologists and other professionals are helping address medical debt, neighborhood blight, natural disasters, domestic violence, and so much more by subscribing to Talk Justice.

Related episodes across the Index

Other episodes covering the same guests and topics, from across The B2B Podcast Index.

  • The AI-Native Law Firm, with Ryan Walker of General LegalMeeting of the Minds · on Claude88 / 100
  • How SSW turned AI into ½ their pipeline - Ulysses Maclaren, COO of SSWSaaS Stories · on Claude86 / 100
  • 657. Waziri Garuba, CEO of Harlem Labs, Introducing G.R.I.O.TUnleashed · on Claude80 / 100
  • Unresolved.cx - When Feedback Has To Matter - Paul TuckerUnresolved.cx · on Claude80 / 100
  • 321 | How to Build an AI App Without Coding: From Idea to App Store with Replit & Claude by Bryce Rattner KeithleyLeveraging AI · on Claude80 / 100
  • John Suarez on How Businesses Should Actually Use AI Without Losing Human JudgementMarketing for SMEs · on Claude78 / 100

More from ABA Journal Podcasts

All episodes →
  • Why some states are introducing pathway to bar that doesn't include bar exam95 / 100
  • How AI has supercharged dispute resolution and arbitration80 / 100
  • Quantum Leap? Development of new tech gives lawyers plenty to think about65 / 100
  • Techshow Attendees Assemble: A preview of this year's show
  • Generative AI is now capable of grading law school exams; what's next?
All ABA Journal Podcasts episodes →