Imagine that you are a lawyer.

Your work involves managing files with dense, technical text. Your co-workers collaborate with you to accomplish a complex goal that can be broken down into smaller pieces. Your work has formal specifications, but there are degrees of freedom in how you express an idea. In all of these ways, the job of a lawyer is similar to the job of a software engineer–so why don’t lawyers use tools to improve their workflow?

As a software engineer, you have project management tools like Asana that improve collaboration. You have APIs like Stripe that reduce the time spent on a complicated implementation. You have tools like linters and source control that prevent you from making fatal errors. All of these tools save you time.

At many law firms, lawyers do not have incentive to save time. They are paid based on billable hours, not individual milestones. Historically, this hourly billing made sense–lawyers have been around since long before computers. The amount of work that might go into a legal task was hard to predict before you had computers to log data, sort documents, and standardize communications.

In contrast, a software engineer has always had the ability to automate work. That’s why (in most cases) we are not rewarded based on our time spent solving a task. We are paid based on hitting our KPIs and our milestones. With the legacy of hourly billing, lawyers can look at repetitive, administrative tasks as opportunities to make more money.

Justin Kan has been building startups for a decade, and in that time he has interacted with lots of lawyers. From incorporation to fundraising to selling his company Twitch, the interactions with lawyers consistently seemed less transparent and less efficient than would be optimal.

For an engineer like Justin, the natural inclination here was to build software and sell it to lawyers. But there would be so much resistance–you would have to convince the lawyers to change their pricing model to fixed-pricing, which would give them the incentive to buy software and work more efficiently.

Instead, Justin teamed up with a few entrepreneurial lawyers who were willing to start a new law firm from scratch, and use software on day 1. The software company is called Atrium Legal Technology Services (or Atrium LTS for short), and the law firm that uses the software is Atrium LLP. Both of these companies are very new, and were publicly announced a few months ago.

The two companies work side-by-side in undecorated office in downtown San Francisco. When I took the elevator up to see the company, the elevator doors opened and revealed two paper signs pointing to opposite ends of the office. On the Atrium LTS side of the office, engineers were writing software to extract the meaning from documents.

Today, lawyers at old law firms are paid hundreds of dollars an hour to fill in document templates by editing a text document. As the Atrium LTS software gets better, document preparation will be done through web applications, with the variable names disambiguated from the parts of the document that never change from client to client.

On the other side of the office sat Atrium LLP. The legal team was dressed a little more formally than their engineer counterparts, but there was nothing close to the formality of a traditional Silicon Valley law firm. Far from the decor of a Menlo Park law firm, the office space was actually more spartan than most well-funded startups, signaling to the employees that this is an unproven business strategy, and there is a ton of work to be done to validate it.

This sentiment was echoed in my conversation with Justin. It’s possible (even plausible) that Atrium LLP could become the biggest law firm in the world, but the road to getting there will take patience and steady execution. I enjoyed hearing Justin explain the motivation for starting Atrium LTS, and look forward to covering the company in the future.

We have done several other shows about the intersection of software engineering and law, including our show dissecting software antitrust with law professor Harry First.  To find our old episodes, you can download the free Software Engineering Daily app for iOS and for Android. In other podcast players, you can only access the most recent 100 episodes.

With these apps, we are building a new way to consume content about software engineering. They are open-sourced at github.com/softwareengineeringdaily. If you are looking for an open source project to get involved with, we would love to get your help.

Shout out to today’s featured contributor Craig Holliday. Craig has worked on the Software Engineering Daily iOS app to iron out performance issues and implement features like 2x playback. Big thanks to Craig.


Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.


The octopus: a sea creature known for its intelligence and flexibility. Octopus Deploy: a friendly deployment automation tool for deploying applications like .NET apps, Java apps and more. Ask any developer and they’ll tell you it’s never fun pushing code at 5pm on a Friday then crossing your fingers hoping for the best. That’s where Octopus Deploy comes into the picture. Octopus Deploy is a friendly deployment automation tool, taking over where your build/CI server ends. Use Octopus to promote releases on-prem or to the cloud. Octopus integrates with your existing build pipeline–TFS and VSTS, Bamboo, TeamCity, and Jenkins. It integrates with AWS, Azure, and on-prem environments. Reliably and repeatedly deploy your .NET and Java apps and more. If you can package it, Octopus can deploy it! It’s quick and easy to install. Go to Octopus.com to trial Octopus free for 45 days. That’s Octopus.com

Auth0 makes authentication easy.  As a developer, you love building things that are fun–and authentication is not fun. Authentication is a pain. It can take hours to implement, and even once you have authentication, you have to keep all your authentication code up to date! Auth0 is the easiest and fastest way to implement real world authentication and authorization architectures into your apps and APIs. Allow your users to log in however you want–regular username and password, Facebook, Twitter, enterprise identity providers like AD and Office 365–or let them login without passwords, using an email login like Slack or phone login like WhatsApp. Getting started is easy. Just grab the Auth0 SDK for any platform you need and add a few lines of code to your project–whether you are building a mobile app, a website, or an API–they all need authentication. Sign up for Auth0 and get the free plan or try the enterprise plan for 21 days at auth0.io/sedaily. No credit card required. Auth0 is trusted by developers at Atlassian, Mozilla, and Wall Street Journal. Try it out at auth0.io/sedaily. Stop struggling with authentication–get back to building core features, with Auth0. 

Incapsula can protect your API servers and microservices from responding to unwanted requests. To try Incapsula for yourself, go to incapsula.com/2017podcasts and get a free enterprise trial of Incapsula. Incapsula’s API gives you control over the security and performance of your application–whether you have a complex microservices architecture or a WordPress site, like Software Engineering Daily. Incapsula has a global network of over 30 data centers that optimize routing and cache your content. The same network of data centers that are filtering your content for attackers are operating as a CDN, and speeding up your application. To try Incapsula today, go to incapsula.com/2017podcasts and check it out. Thanks again, Incapsula.

United States


00:00:00Imagine that you are a lawyer your work involves managing files with dense technical text your co workers collaborate with you to accomplish a complex a goal that can be broken down into smaller pieces your work has formal specifications but there are degrees of freedom and how you express
00:00:20an idea in all of these ways the job of a lawyer is similar to the job of a software engineer so why dont lawyers use tools to improve their work flow like software engineers d'oh is a software engineer you have project management tools like asana that improve collaboration
00:00:38you have a p i's like stripe that reduced the time spent on a complicated implementation you have to lt's like limiters and source control that prevent you from making fatal errors All of these tools save you time at many law firms lawyers do not have that incentive to
00:00:55save time they're paid based on billable hours and not individual milestones Historically this hourly billing made sense lawyers have been around since long before computers and the amount of work that might go into a legal task was hard to predict before you had computers toe log data and
00:01:14sort documents and standardized communications in contrast a software engineer has always had the ability to automate work that's why in most cases were not rewarded based on our time spent solving a task were paid based on hitting our kp eyes and are milestones with the legacy of hourly
00:01:35billing lawyers can look at repetitive and administrative tasks as opportunities to make more money Justin khan has been building startups for a decade and in that time he's interacted with lots of lawyers from inc to fundraising to selling his company twitch The interactions with lawyers consistently seemed less
00:01:56transparent and less efficient than would be optimal for an engineer like justin The natural inclination here was to build software and sell it to lawyers but there would be so much resistance you would have to convince the lawyers to change their pricing model to fixed pricing which would
00:02:12give them the incentive to buy software in tow work more efficiently Instead justin teamed up with a few entrepreneurial lawyers who were willing to start a new law firm from scratch and use software on day one The software company is called atrium legal technology services or atrium lt
00:02:31s for short and the law firm that uses the software is atrium l l p both of these companies are very new and they were publicly announced a few months ago The two companies work side by side in undead aerated office space in downtown san francisco when i
00:02:49took the elevator up to see the company the elevator doors opened and revealed to paper signs pointing to opposite ends of the office on the atrium lt s side of the office engineers were writing software to extract the meaning from documents today lawyers at old law firms are
00:03:07paid hundreds of dollars an hour to fill in document templates by editing a text document is the atrium lt s software gets better Document preparation will be done through web applications with the variable names dis ambiguous it'd from the parts of the document that never changed from client
00:03:25to client On the other side of the office sat atrium l l p the legal team was dressed a little more formally than their engineering counterparts on the other side of the office but there was nothing close to the formality of a traditional silicon valley law firm It
00:03:41was far from the decor of the menlo park law firms and the office space was actually more spartan than most well funded startups signaling to the employees that this is an unproven business strategy and there's a ton of work to be done to validate it This sentiment was
00:03:59echoed in my conversation dustin it's possible and even plausible that atrium l l p could become the biggest law firm in the world powered by the software of atrium lt s but the road to getting there will take patients and steady execution I enjoyed hearing justin explain the
00:04:18motivation for starting atrium lt s an atrium l l p and i look forward to covering the company mawr in the future We've done several shows about the intersection of software engineering and law including our show dissecting software antitrust with law professor harry first to find all of
00:04:37our old episodes you can download the free software engineering daily app for ios and for android in the other podcast players you can only access the most a recent one hundred episodes but with the software engineer daily app you confined all of our back catalogue and get recommendations
00:04:54based on your listening history with these aps we're building a new way to consume content about software engineering You can also find the desktop application at software daily dot com and these aps are open sourced atget hub dot com slash software engineering daily If you're looking for an
00:05:12open source project to get involved with we would love to get your help shout out to today's featured open source contributor craig holiday craig has worked on the software engineering daily ios app to iron out performance issues and implement features like two times playback so you can listen
00:05:30to this episode in half the time big thanks to craig and his brother keith for all of their work and their contributions to the open source community with that we shit on this episode the octopus a sea creature known for its intelligence and flexibility octopus deploy a friendly
00:05:57deployment automation tool for deploying applications like dot net aps java aps and more ask any developer and they'll tell you that it's never fun pushing code at five p m on a friday and then crossing your fingers hoping for the best we've all been there we've all done
00:06:16that and that's where octopus deployed comes into the picture octopus deploy is a friendly deployment automation tool taking over where your build or c i server ends use octopus to promote releases on prem or to the cloud octopus integrates with your existing build pipeline t f s and
00:06:38v s d s bamboo team city and jenkins it integrates with a ws azure and on prem environments you khun reliably and repeatedly deploy your dot net and java aps and maur if you could package it octopus khun deploy it's quick and easy to install and you could
00:06:59just goto octopus dot com to trial octopus for free for forty five days that's octopus dot com o c t o p u s dot com just the ceo atrium lt s justin welcome to software engineering daily thanks for having me in the nineteen nineties the biggest financial
00:07:28cost to starting a company was often servers today the servers are pretty cheap but something that remains expensive is legal What are the common ways that a startup has to interact with law firm Well basically every company has to interact with lawyers for every major financial transaction it's
00:07:49kind of the operating system of business in america and so you know you always have to interact with it whenever you do anything really big So if you raise money you generally pay lawyers if you sell your company pay lawyers that you do a major commercial transaction you
00:08:06want someone to review it pay lawyers so pretty common and i would say it's a very you know very it's an inescapable thing kind of like death and taxes the legal industry itself has not been transformed by technology in the same way that some other industries have why
00:08:24is that so like you could alluded to earlier legal costs of actually on lee risen in you know in the past twenty years pretty much across the board for big corporate law but also you know you could say localized to start ups as well And so one of
00:08:42the questions might be why why is that the case right why they're doing you know you siri's a today versus twenty years ago its most of the same thing generally prices fall why have prices not fallen in legal And i think there's a bunch of different reasons for
00:09:00one of the main reasons is that there's very little process or technology improvement The legal industry has a bit of an innovator's paradox right legal innovation paradox which is that there's no incentive when you build on an hourly model there's no incentive to improve efficiency over time that's
00:09:18not that lawyers are trying to you know figure out how to charge the most or bill clients you know one of the most number of ours is just that you don't have an income in the income is don't have any incentive to figure out how to reduce costs
00:09:31through the adoption of better software practices or better software or even better process and so what happens is they don't actually adopt any software and so there's a very very little market for it what is the technology stack inside of a law firm outlook you know it's like
00:09:49outlook outlaws tolar working outlook and word which are pretty robust tools as you know but pretty static you know there have been very many improvements in the past twenty years and so i think that's that's ah lawyers live in documents an email on dh that's kind of the
00:10:09the techs act that use and then there's you know oftentimes a document management system was well you've started a bunch of companies and you've engaged with lawyers throughout those different companies is there a particular event that stood out where you said to yourself this makes absolutely no sense
00:10:29or was it more of an accumulation of just minor inefficiencies and inconveniences I would say it was just interacting with the industry overtime you know i've been over the past twelve years in silicon valley i've become what i call in involuntary power user of corporate legal services and
00:10:48i had used him for all those events that we talked about whether it was you know selling company or raising funding and i wouldn't say there was i actually wouldn't characterize my experiences as bad at all i would say that we're just okay you know somewhere better somewhere
00:11:04worse i liked using legal services when i felt like the lawyer was an expert who was giving me their expert opinion and advising me from something that was very complicated or you know that i i needed confidence that was going to go well and i think that various
00:11:22time at various times i felt felt like it wasn't going as well when i didn't like the service is much when things were you know opaque i didn't know what the process wass or you know i would be built what seemed like a random amount random high amount
00:11:38so so that's that's kind of like what led me to the idea for asia Lt s was kind of asking is there better waited to do this So it wasn't really like it wasn't one event it was more of like i'm sure this industry could be better we'll
00:11:54get into the two businesses one of them is legal technology services which is the set of of services software that your building and atrium lt s which is as i understand the first law firm that is dog food ing these services is that the right frame work Well
00:12:13so the atrium lt s is actually is our legal technology services come Okay that i'm you know ceo of its a startup we run it like a startup likely have engineers and they provide services to law firm owned by augie right rico and baby chua who are to
00:12:30lawyers that's called that's a it's called a trim elope and they serve clients in silicon valley start clients so that's kind of the relationship between these two entities It stems from the fact that there's a bunch of regular regulation around who can own a law firm and who
00:12:47could provide legal services I'm not a lawyer i've just been a like a set of power user and so our goal was to within the bounds off was to like have have shortened the feedback cycle right and actually remove the legal innovation paradox by having a firm that
00:13:04was built around thie idea that you could improve the deliverer of legal services and so that's you know atrium elope is that from got it describes some of the services that you're building at atrium lt s so i would say it's a bunch of things the more interested
00:13:25the interesting product vision that we're building towards is what we call no documents and i think that what i mean by that is that if you look at your company's data it's all like the corporate date of your company like who owns the company what's that thing over
00:13:43the owner's who's on the board what the board's done who are the employees what the employees are paid all this data what the contracts are the company and who's owed what all this data's in pdf said we're document right That are maybe a box folder dropbox folder and
00:13:58most people that's like not accessible to all right Like i've been to the company owners they don't necessarily understand how to parse it They can't really extract the structure data very easily from that Those documents but there's little legal document are the source of truth They describe exactly
00:14:13what your company is more than any anything else on dh So what we're trying to build ultimately the way i think about it is something like a system that basically sucks in that all of that all those all those documents and turns them into structure data So we
00:14:29understand your company and you know if i put my engineer's hat on i think that's like a very interesting problem because there's a bunch of scripting and machine learning tools that we use to built to extract that data and then i think of that data as a platform
00:14:43for on top ofwhich you can actually build a bunch of interesting applications write if you need to render a cap table for financing in excel that should be one button right Because you already understand like the existing capitalization structure of the company you can build a document management
00:14:57system or contract management system on top of it you can build a dashboard for the company owners that shows you know exactly like who's you know allows them to do a lot of things on top of their company whether it's you know initiate aboard action anderson offer letter
00:15:13all these things are basically actions on top of the existing date of your company and i and ways to visualize or render that data and i think that's pretty pretty powerful like an existing on a platform that hasn't existed for companies yet can you take anything off the
00:15:28shelf like a dock you sign a p i or a e shares these Absolutely So i think that issues is a great example of like this We could use this data extraction toe basically automatically on board ut ishares i love the shares of as a you know investor
00:15:45and as a user i like issues a lot The on boarding process is manual because somebody has to put in all the data of like the capitalization structure company right for doc you sign for example or a hello sign we we could build on top of those and
00:15:59actually allow you know we would use the fbi's too do you know send out these like board consents or something like that or a offer letter whatever this was going to be built on top of the existing designing tools So i think that ultimately i see it as
00:16:14like we're building this platform where we understand the structure date of your company and then we're building applications on top of those applications probably have you know interact with other ap eyes and applications out there Wait zero makes authentication easy as a developer you loved building things that
00:16:42are fun and authentication is not fun authentication is a pain it could take hours to implement and even once you have authentication you have to keep all of your authentication code up to date auth zero is the easiest and fastest way to implement riel world authentication and authorization
00:17:01architectures into your aps and ap eyes allow your users toe log in However you want regular user name and password facebook twitter enterprise identity providers like a d and office three sixty five or you could just let them log in without passwords using an email log in like
00:17:19slack or phone log in like whatsapp getting started is easy you just grab the off zero s decay for any platform that you need and you had a few lines of code to your project whether you're building a mobile app a website or an a p i a
00:17:33they all need authentication Sign up for off zero that's the number zero and get a free plan or try the enterprise plan for twenty one days at off zero dot io slash s e daily that's a u t h dot io slash s e daily there's no credit
00:17:53card required and off zero is trusted by developers and at last cnn and mozilla and the wall street journal and many other companies who use authentication Simplify your authentication today and try it out at a u t h dot io slash essie daily stop struggling with authentication and
00:18:14get back to building your core features with off zero way could you maybe talk through what's an example document that you looked at and you said okay here's how we're going to architect on algorithmic way of looking at this and extracting data from it one one very simple
00:18:40like a trivial example kind of one that were way build it pretty early on is safe notes on convertible notes you know these air instruments where someone is in investing in a company in a non equity round there very common in silicon valley why combinator pop popularized this
00:18:56idea of a safe simple agreement for future equity the idea behind it is basically you're agreeing to buy equity in the future so you can basically invest in a company without actually having to go to the process of doing that equity financing which is you know expensive and
00:19:10takes a long time and so these notes are all based off of a very small set of templates right And so it's actually very easy to under like build something that extracts like you know takes these pds break them open extracts all the text from them compares them
00:19:26against known templates extract some real meta data from it you know the things like how much this person invested what their entity name was there addressed email address and then obviously store those into a structure database then take all that data and like when you're going to do
00:19:41a actual siri's a financing one of the processes is understanding who's going to own how many shares at the end of the financing and building on excel model that you can share among all the constituents the lawyers the company owner the founders and management you know you can
00:19:56share this model so they can see exactly who's going to own what and you know that whole process of extracting like taking the safe just reading them right understanding them on and then being and then turning them into a cell model is done by a an attorney right
00:20:12now that attorney bills five hundred dollars an hour to basically make this excel model now obviously that's something that could be done programmatically right and that's what that's what one example of something that we're doing basically turns you know extract so saves like the method i described and
00:20:28then now takes things that structure data and the rent's renders in itself model based on it sure Okay So the safe note that is something that's very template ties because every startup in silicon valley that uses a safe note is going to use one of these templates that
00:20:43looks very similar even operating agreements i think are you know some people have pretty off the shelf operating agreements that air you know you and the but nonetheless do have data that you would want to extract and somewhat standardized dashboard format that would be useful but across the
00:21:04board in terms of how people engage with legal documents there's a wide array of and that certainly outside of silicon valley i assume it's less standardized is it or is it or is it still standardized when you when you step outside of silicon valley's still template ties where
00:21:20you khun have these well formatted documents and have expectations around where certain variables are going to be Yeah i would say that so obviously the save is very kind of very constrained example there's other types of documents you know that we're building machine learning class of fires to
00:21:36actually categorise on dh requires a little bit more advanced pattern matching that just like comparing it's known templates right i do think that depending on the practice of law there's more or less standardisation right on dso other practice areas you know like lease agreements for commercial real estate
00:21:54it's like very similar thing over and over and over again fun formations you know some same thing so there are many other practices where i would say the same types of techniques khun b i used to understand the structure data on dh those are the ones where i
00:22:09think we would have the most you know ability and advantage and actually trying tio go into and build building applications for them yeah yeah but it is it's not i think as i understood you're preparing for this lt s is not just about machine learning for law it's
00:22:27also about making the lawyers more productive in their job because i think for the foreseeable future there's going to be a lot of subjectivity in dealing with with legal documents you can't just unleash the algorithms and have them take care of everything what are some of the ways
00:22:45that you can make the job of the lawyer easier and automate away some of the the painful process is so people are always going to want a lawyer right when we were selling twitch we didn't want the uber for lawyers we don't want a marketplace of lawyers we
00:23:03don't want an aye aye algorithm lawyer we wanted lawyers who had experienced selling you know billion dollar transactions and that is going to be common and immutable in my opinion for the most valuable work you know you might incorporate your company on legal zoom or clarky or something
00:23:22like that but ultimately when it comes to the big transactions which are the valuable lucrative transactions people are going to want a real attorney who has experience someone who from it with a great background went to a top law school what two top firm and has worked on
00:23:37this type of transaction and i don't think you can get away from that Our goal you know with a charm lp is that hmm elope would be a top firm like that that has that type of talent and the legal the software is really reducing work and also
00:23:54providing transparency and speed behind the scenes and maybe in parallel as well So i think that's the that's the core philosophy it's not like creating an a i that replaces lawyers it's maybe creating sameh eyes and machine learning and scripting and and software that does some of the
00:24:11base level work the like trivial work like document classifying right like understanding all the default structure data all that stuff is things that you don't need someone who went to harvard law school to do right And so i think that's the that's kind of the way i think
00:24:25about it in terms of what we're going to get the actual works like savings i think it's in many different things and it's like kind of small savings in all these different areas whether it's rendering the first version of documents or whether it's giving more transparency to the
00:24:40customers so that they send less emails and texts you know like transparency into what it is exactly being worked on almost like giving an interface onto the legal work that's similar to like a legal asana orgy era i think that's something that even in the industry today could
00:24:57adopt But they don't really because there's no incentive And at atrium l l p r they adopting stuff like a sauna Do they use project management tools We also using project management tools Yep It's ah it's innovative for the industry But you know i think it's ah it's
00:25:14a good step for so they do use the am l a p d uses project management tools yep awesome and to have a noticed if they said anything like oh this revolutionizes how we do our legal work or anything like that like what What's what's the feedback you're
00:25:31hearing feedback is good you know the feedback is good i think we've managed a charm elope is comprised of many entrepreneurial lawyers who are yes you know especially younger lawyers who are excited about innovating and adopting new technology and that is you know kind of reflected in how
00:25:51they approach it you can compare teo and most firms you know the people who are in charge of the firm are the people who are most you know senior and oftentimes very ingrained and doing things a certain way and so oftentimes that's one of the reasons why there's
00:26:07this you know lack of adoption of new software even commercially available software that might be adopted in many other areas outside of the law like like a asana how is the engineering organ at Lt s a shame lt s structured so the engineering or it is we have
00:26:26a bunch of different engineering teams would say that that coral correspondent different legal functions work with the lawyer's very closely on different areas of of of the business kind of on different different types of work they get that's done so each room lp has a team that does
00:26:43financings for example and then there's there are engineers at work with that financing team as consultants and to basically build software for that financing team that just works on kind of silicon valley style financings and then there's another team that works on documented taken processing that's doing some
00:27:00of that document extraction and meditate extraction that i talked about so we have like it's it's very cross functional from organizational standpoint the difficulty seem maurin product management than really hard engineering problems is that is that the case right now i think that document extraction and structure data
00:27:20side is actually a very interesting engineering problem i think there's two types of engineers that would well there's probably three times avenge engineers that would want to work at a trim hmm lt s and those three engine is the first is like kind of product engineers who want
00:27:35to build kind like you're saying product engineers who want to work very closely with their customer in this case it's lawyers and paralegals to build products that are used every day and they can you have a very short feedback cycle so i think that's that's one set of
00:27:49people that repeal too you know you're sitting nearby the euro your best customer right and it's and it's possible for you to you know really get feedback in a very short short term basis and rule something out and sea change and see people like really be thankful you're
00:28:04talking about building suffer an industry that really doesn't have a lot of like engineers building things for the second type of work and engineer who might be interested is really i would say people who are focused on who want to build that platform that data platform and maybe
00:28:22work on machine learning it's people who you know we have this interesting data set of corporate documents where there's there's like all like i said all the structure data of your company it's like is described in this in this data set and the other thing is about it
00:28:37which i think is very interesting is that the data set is is like it's a very i think it's a very good data set for me is like as a as a machine learning problem because it's like very constrained actually in a lot of ways And so you
00:28:51know i think it's interesting to think about how you can build a platform we first of extract this data and then you build applications on like structure in a way that you build applications on top of it that's like the second type of engineer i think might be
00:29:04interested in the third type is someone who has either been a founder or a very early stage engineer at to start up because you know we are a treatment lt s his mission is really to empower legal services four stars to be faster more transparent and more with
00:29:21upfront pricing and so i think those you know people who have had that pain would be would be interested why is that up front pricing model so important for a trim l l p to be different Yeah i think it's very important for the adoption of new technology
00:29:40for us to have a reason to to adopt it right for a trim elope is a when you're on the hourly business model you don't have any go there's no like incentive for you to like actually reduce costs and part of reducing costs and being more fishing is
00:29:54like adopting new technology And so one of the reasons you see these lord like law firms not adopt new technologies there's no internal incentive and so by having fixed pricing it doesn't really matter whether it's higher margin or lower marty jeffers what matters is your internal incentive is
00:30:09to innovate and reduce costs How similar are the operational day to day work The cases for h m like how similar is it to a normal law firm I think that html opie's workflow is different behind the scenes but the work product output ultimately is very is very
00:30:35similar to an existing law firm right That output is documents its review of documents it's talking on the phone or email or email advising to clients and so really A lot of the interactions are powered by software behind something that i have enjoyed about engineering is that we
00:30:58learn the tools for breaking down pretty much any concept in the world and understanding it and it could take a long time But as an engineer you figure out that almost anything is possible to understand and this is the approach that you're taking to the law You have
00:31:14no formal education in it there's no reason why you should be starting a legal company What has been your personal process for learning about the law Sure So you know similarly i that's the reason i like and jerry i was always pretty programmer wasn't i was a self
00:31:34top programmer right So i was actually good web developer I said i say i was a good web developer and a programmer but i loved programming and building web apse because unlock my creativity you could feel like anything is possible You could build anything as long as you
00:31:49could I think about how to structure it right and define the problem and define define the outputs and the imports right and so that's kind of the process I would say it would take a very engineering process towards the creation of you know this like legal software and
00:32:05services and my you know thinking around it is is really the process i've gone through i would say it's like first i was like gathering data right talking to tons of partners and lawyers out there and clients and investors unlike what are the problems in the street are
00:32:23the problems that i experienced the same as the problems that all these other that other people might have experience and really gathering like what are the reasons why this is you know there's a lack of innovation Why Why are they using software Why don't even use cr m
00:32:37software or you know asano are a lot of project management software and then once i had like a really good mental framework for how the industry operated day i was like okay what are the entry points where we could create something that affected this industry and changed it
00:32:52for the better right toe like provide these values of speed transparency and price predictability to start ups And so that's how i approach the problem and then and then you know building a company I think building this legal company company is no different than building any other company
00:33:09which is just to say what you know you're thinking about what are the problems Why don't want to be you know what we want to get to a certain milestone certain revenue model stones or number of customers whatever it is and then working backwards like what are the
00:33:23biggest step barriers from getting there today and one of the things that i need to do to affect those you know and it all breaks down to pretty much always comes back teo hiring great people and making sure the focus on the right problems are there any areas
00:33:37of the law that you've tried to study what they remain too difficult and too obscure and you just still feel like you don't get them i the funny thing is i actually so i'm one of the like easy clients i think i was one of the clients who
00:33:53never really read any legal documents on dh didn't really digging that much you know lead legal for me was always a barrier between what i actually wanted to do which was you know fundraising or someone company or whatever and it's just something i had to get through and
00:34:09so it's funny because i don't actually i think i understand a lot now through working on you know h melt yes i understand the steps of a series a financing i had actually raised you know millions of millions tens of millions of dollars with actually understanding with steps
00:34:22of syria's a financing were so you know i do it's funny i have like mohr knowledge legal knowledge now but that's not really my my interest my interest is not nothing i would say the content of the of like legal work is not my interest my interest is
00:34:38applying engineering and product to affecting the delivery that content to make it more efficient and make it a better experience for all the participants whether the lawyer or the client and so it's funny i think that's funny because i actually like my my i'm learning much more about
00:34:57the like delivery of legal and the business of legal and the technology of legal than i am about the content illegal which is what you know that's what i'm interested way you have a product that is sold to software engineers are you looking to hire software engineers become
00:35:20a sponsor of software engineering daily and support the show while getting your company into the ears of twenty four thousand developers around the world Developers listen to software engineering daily to find out about the latest strategies and tools for building software sent me an e mail to find
00:35:37out more jeff at software engineering daily dot com the sponsors of software engineering daily make this show possible and i have enjoyed advertising for some of the brands that i personally love using in my software projects if you're curious about becoming a sponsor sent me an e mail
00:35:54or email your marketing director and tell them that they should send me an email jeff at software engineering daily dot com thanks as always for listening and supporting the show and let's get on with it totally you know i was looking at you know some of the clients
00:36:19that atrium elope has worked with so far one of them is stew that was protocol labs which is they made file coin and i p f s and they'd had a nice seo have you learnt anything about howthe law views i ceos Yes which is the silicon valley
00:36:39that law firms are mixed on ice so some of them are very it's you know we're going to support this and like do them and others are like we don't want to touch that because we think they're they're massive legal process you know securities law problems around it
00:36:52i would say that right now there's a lot of unknowns and people have not tested this thoroughly against you know regulators and you know perhaps like against the legal framework that exist today so i think it's mostly remains to be seen whether you know and i think this
00:37:09is all compounded by the fact that many i ceos right now i'm by no means an expert and yeah krypto i am a you know a vester and one crypto fun polly jane and you know have some bitcoin but i'm not like i'm not someone who is like
00:37:24cutting edge and like up to date i rely on my friends for that mostly but i would say that i think i believe that like many of these ceos are happening right now are purely gold rush in there there's there's no reason for them to be like for
00:37:38their software to be a distributed ledger right there like that that makes sense only for certain things like like certain use cases right Not every use case doesn't make sense for it to be a distributed ledger system not for not every use cases that make sense to have
00:37:52like a separate token of value right Like i think file coins case it makes a lot of sense but there's a lot of like cos like you know the qin token for example that kicks token that like kick message token it's like i don't understand what the economic
00:38:09point of creation cryptocurrency is except to raise money for a you know corporate vehicle of course which i think is just if that's the case then it seems like purely a way to circumvent securities law yes okay yeah so so you think that the or are you just
00:38:27speculating or do you have you had a sense from talking to people that i'm just speculating this is not based on my like i have no i have no this is my business hat you know because this person silicon valley investor hat it just seems like a lot
00:38:39of these things it's like kind of a gold rush and a lot of the tomorrow if not frauds there like there's no reason for there's no like actual technical and economic and business model reason for them to like actually being cryptocurrency a new twitter ad for a new
00:38:56eye ceo every day yes that that is to me that's very sketchy and i get emails ever like you know parts like cold solicitations to participate in this random i c s right So this this upfront pricing model that atrium has where you keep prices fixed on work
00:39:19so if i'm a startup founder i come to a tree mellow p and i say hey i need this task done and atrium tells me here's the price upfront and that's great that's a weight off my shoulders where otherwise you know i'm already worried about how much cash
00:39:35is in the bank and if i have to also worry about okay law for like random old world law firm i need to get this task done and they say okay cool we'll let you know what it cost when we have it done in three weeks that's a
00:39:48burden and i don't need any extra burden so getting the prices fixed on work that is going to take a variable amount of time from case to case that seems challenging but it also reminds me of we did interview with gig stir a while ago and they do
00:40:05something kind of similar with their contracting model where they have you know big software projects that come to them and they learn over time how to price these basically by tracking the data really closely what's the approach to getting fixed prices on what have been variable priced purchases
00:40:24in the past So i think there's probably two types of work there's work that it's very like within a certain bell curve every time it's like more known and there's lots of data points So that's work like siri is a financings writer series b financing silicon valley law
00:40:39firms who have been around for a long time We'll have like thousands of data points and they should be able to build a model that says okay given your parameters of you know the things that are coming in like your firm how much money it is like how
00:40:52money with complexity like how many shareholders incumbent cheryl is You have all this stuff they should be ableto build a model that says here's the distribution of likely outcomes of how much work it's going to take and we're gonna just price it at the percentile mark Or something
00:41:06like that but they don't because there's no incentive to and so i think it should be possible to build a model for many kinds of work that that is based on the data that actually is accurate enough that you khun basically absorbed the risk of variance that's what
00:41:23we want to do so is actually the model with html lp is this is the proof of concept of a set of practices and technology services that a law firm can adopt what's the model for deploying that at other law firms how much data do you need to
00:41:43collect within a trim l l p too get other law firms enticed or are you already in conversations with people who are interested No i would say it's pretty putting new and we need to prove because lawyers are very resistant to changing their work flow i think there's
00:41:59going to be like many proof points that we have to hit before other you know it's it's even functional light other law firms so our goal really right now is to like learn and get data i'm like how how you can improve processes of what technology you know
00:42:13platforms we can build two to actually improve lawyers work and that's that's kind of where we're at right now yeah i'd say it's like you know that's the margolis so like just be right now in the learning phase i think it has any part of the product development
00:42:29been harder than you anticipated it's a good question as part of product development been hard remember i saw an interview where you were talking about how you've been thinking about this this building this company for a while yeah i think that we are trying to do a large
00:42:45a breath of things right now and so that is you know things are i think we're working on many different types of legal work for startups like tools for legal or for starts So so there's there's a little bit of ah i think we need to we could
00:42:58make more progress probably if we constrain the scope to fewer things and yeah and i think we might wantto we might want to do that but overall i wouldn't say it's like harder than than anticipated it's a pretty straightforward i think there are interesting engineering challenges like i
00:43:15said on that kind of extracting structure data side but mother you know that a lot of the other software is fairly straightforward how to develop it Yeah well that that issue of focus you know that comes up a lot when you talk to different startups and you know
00:43:30doing doing multiple things at once is really hard but similar but the problem is you're in a space where you wantto sort of changed the whole mindset of the space it's not like where you know your amazon you're like okay let's focus on lee on books first and
00:43:47then it's very obvious howto laterally expand to dvds it seems like you need to go whole hog i mean if you were to constrain the focus toe so one thing specifically like would it be safe notes or something is that i think hard i think there are two
00:44:04major pieces of software that that are very important one is you know the data extraction on platform and the second is the client facing portal that gives them insight into what's going on in their legal work And so those are probably the two things that we really need
00:44:23to make the most progress on I think so what's the most grandiose vision of what lt s could evolve into i think lt s could power the biggest law firm in the world You know i think a trim elope could be the biggest law firm in the world
00:44:41and lt escaped power it and your power many other law firms outside of that and that's kind of the kind of the big vision i think that this is a new model for unlocking one hundred sixty billion dollars of outsource corporate legal spend and you know that's my
00:44:59that's that's that's my goal when we when i started when i decided like go back into starting a company i had been investing and incubating companies for a couple of years You know partner why commenter before that start a bunch of companies twitch is the one that people
00:45:13know about what really got me to go back and say i should be the ceo of a company again and and really dive all in is that i really believed that this is a big market and big opportunity and i really want to try to build something that's
00:45:29bigger than twist Yeah so okay because market supports it Well it's interesting I thought that the model was to build technology and then give it to lots of other law firms but i could see it being equally plausible where you just grow a tree mail lp I think
00:45:46i think we try to grow a trim elope and also like try to figure out how we can sell the software to other law firms as well I think those are parallel paths but the goal is that legal revenue out there like the most weaken we could have
00:46:05the largest amount of legal revenue flowing through this like lt s software as possible That makes sense Yeah definitely So i do want to talk a little bit about twitched just because i've heard some random interviews where you were talking about some of the engineering issues a twitch
00:46:20but i'd never heard anything where anybody went super deep but you know i know their stuff there was a lot of custom software built like that's a serious tech company what were some of the engineering problems that you recall from twitch So i had engineered anything on twitching
00:46:38in a long time so caveat everything that and twitch was built on top of the infrastructure that we use for justin tv so my might most of my engineering work was from the justice tv days and then you know eventually was re skinned into twitch and i think
00:46:55a lot of that course infrastructure state the state the same we were of one of the biggest ruby on rails sights for a while maybe still still are but that was a big challenge actually there were two probably big challenges one was just that the building of video
00:47:13video infrastructure that would live video infrastructure was like not mature and so we had to build a bunch of basically software around like the live video servers that would rout the video to various points of presence and kind of like grow the number of like you know of
00:47:29love of like basically there be like video would come into an origin right and then go out to like very slaves like all over sometimes in the same data center of summer in san francisco sometimes in multiple data centers if it was a popular stream the problem was
00:47:42like because we were supporting streams that were so that there were like so some would be like ten people watching someone be like one hundred thousand people watching right And we actually didn't know a priori unlike a knock on my door which provisions we didn't know a priori
00:47:55which streams would be super popular or not right So you had to have this system that could be very reactive and if a stream was getting really popular would populate it too many different slave streams fair start servers right all over the world right And so that was
00:48:09that was a huge engineering problem been a long time to get right from a like stability being being robust standpoint and that was one big engineering problem and then the other was just getting this like rails site to scale i mean like twitter had a huge amount of
00:48:24problems during same type like two thousand ninety thousand ten just getting this like ruby on rails side to scale we were building this like this basically middleware on twisted that like basically cashed it like cat statically cash all these rails pages the application service took forever to render
00:48:41and then like inject the have like a meme cash store of like all the user data and i would inject like we'd have these custom tags and would like to replace them with you the user data so it was actually like rails would render the first like the
00:48:54cat the cash version and then this like custom server would would like interject that they actually user data because rails was too slow at the time i think i've been massive improvements now you know that all of this is you know from almost ten years two thousand eight
00:49:09to two thousand ten i would say two thousand eight to two thousand eleven so and that's how we scaled teo at the time was like hundreds of millions of page views a month you know And that was big at the time of shirts drop in the bucket now
00:49:19yeah it was it was fun i mean i learned a lot about a lot about engineering because you know like i've been a really formally trained i wasn't a cs undergrad and i also you know never i never worked at like and a big company or anything wade
00:49:35started our first company right out of college and so it was it was a good tutorial and kind of like engineering and engineering management for me were some takeaways from that from the management process i remember t about debugging it the site would go down right there's a
00:49:49live video site with like our peaks to valley ratio was like thirty five x toe like one right So it's like it was very hard to like scale you could you know most other sites are like you know what one point five two accident peter like toe trough
00:50:04usage and so it was just crazy everything time something would fall over it was like always one of six things right Like a band with the servers like cpu memory like esso the site will go down and just be like okay let's look at nagy house and it's
00:50:21like okay which like what is like what's getting here You know like oh it's like i owe it was like we can't write write fast enough to disc here's like and then you just like debug it like there's so much alive debugging of like how can we like
00:50:34shut off a feature or like read write something right then toe like make it work right now because like a live video site i mean this is probably a terrible way to actually build something but it was so hard to predict what are usage was and we were
00:50:48also operating on a very lean budget from infrastructure stand point by the time we know when we sold twitch we have like three network engineers or maybe five network engineers like hu lu which was less traffic i think network engineering to fifty so we were operating we were
00:51:03very under funded the whole time i mean we raised forty five million dollars over the course of life life of the company but that's over yes six eight years and i would say that our infrastructure we were very efficient at infrastructure but it came at the cost of
00:51:18like you know we never had excess capacity like we were oftentimes tweaking features in order to i actually have the site stay up which you know is terrible from an engineering standpoint but but it but it was it was a very instructive in terms of i don't know
00:51:32it's like that's funny i haven't thought about this stuff in years but it was it was it was fun did you anything about managing your own psychology in that process Sure i mean i don't think i implemented it super well i think i'm much better at it now
00:51:44actually but like at the time that was so so stressful because sometimes the site would just fall over and we had a major event going on or like the jonas brothers were streaming or something and we were and we would be so stressed out you know we would
00:51:56be like just like looking at each other like this is like what can we do it like little staring at like a you know five hundred application air and chrome and be like how do we turn off features so that this live event can happen right now or
00:52:10how do we like rewrite cashmore things or like what can we do to like you know fix this and it was yeah that was very stressful could it oftentimes very stressful interesting is there anything else you wanna add about a dream lt shum elope and where the project
00:52:28is going we're hiring right now you know so that's ah that's a think it's a pretty fun project i think it's a pretty interesting and unique project and it's going pretty well you know a trim elope has interesting clients atomic clients right now and it's pretty cool to
00:52:45build software that people use every day and i think that is very interesting data play that that not to me is like the exciting part as if you know from my engineering side it's like the thing that i get excited about so you know if you're interested just
00:52:58reach out and let us know great all right well justin kahn thanks for coming in software jean daily thanks a lot all right that software engineering daily we need to keep our metrics reliable if a botnet started listening to all of our episodes and we had nothing to
00:53:16stop it our statistics would be corrupted We would have no way to know whether a listen came from a bought or a real user and that's where we use in capsule to stop attackers and improve performance when a listener makes a request to play an episode of software
00:53:31engineering daily in capsule of checks that request before it reaches our servers and it filters the bought traffic preventing it from ever reaching us bought nets and de dos attacks are not just a threat to podcasts they can impact your application too In capsule can protect a p
00:53:50i servers and micro services from responding to unwanted requests to try and capsule for yourself go to encapsulate dot com slash twenty seventeen podcasts and get a free enterprise trial of in capsule in capsules ap i gives you control over the security and performance of your application and
00:54:09that's true whether you have a complex micro services architecture or a wordpress site like software engineering daily in capsule has a global network of over thirty data centers that optimize routing and cash your content The same network of data centers are filtering your content for attackers and they're
00:54:26operating as a cdn and they're speeding up your application They're doing all of this for you and you can try it today for free by going to encapsulate dot com slash twenty seventeen podcasts And you could get that free enterprise Trial of in capsule that's in capsule dot
00:54:43com slash twenty seventeen podcasts to check it out Thanks again in capsule wait

Transcribed by algorithms. Report Errata
Disclaimer: The podcast and artwork embedded on this page are from Software Engineering Daily, which is the property of its owner and not affiliated with or endorsed by Listen Notes, Inc.


Thank you for helping to keep the podcast database up to date.