data.bythebay.io: Panel: Data and Society
Recording: data.bythebay.io: Panel: Data and Society
This is the closing panel of our first uh kind of data for society day, right? This is the second day of text by the bay chunk uh of the data by the bay conference. And last year we had two days uh mostly technical and but a lot of applications were already uh in in various verticals either law government as well and this year I specifically designated uh two tracks data for democracy and law by the bay and we had amazing people we had amazing talks today uh we had big law firms we had legal startups we had city of San Francisco uh we have uh open gov right we kind of government hackers uh and legal hackers and I think the uh the atmosphere as you know in famous saying of Mark and recent software is eating the world and these two areas are ripe for you know eating being eaten by software and uh but obviously there's a lot of people working there already it may seem for us from the outside like we want to hack the technology but the language is very different the areas are very uh established and intricate and and so we have uh real experts working in this areas and uh working with data. Uh and so the topic of this panel is essentially to understand how these areas work and uh where data is uh kind of most useful and how all of us can can help what opportunities are there. Uh so my name is Alexi Krov. Um I'm the founder of data by the bay and uh where is uh I run developer meetup. So I kind of represent the uh technology community and uh I'll let all of the panelists uh introduce themselves. uh uh it will take about 3 minutes and will tell us uh about the areas of expertise and how they use data in their daily lives and how they wish they could use data and we'll we'll pick up on that and uh if you have some questions uh we'll basically see uh if folks have questions maybe we'll you know switch to them about halfway through through the panel so uh I'll let uh Brad start from that end. Hi I'm Brad Newman
I work for LP, uh, which is located in 101, California, uh, where there's a there was a Bernie rally. So, I'm I'm I'm Schvitzen here because I ran over. I got locked down by the Secret Service, although I'd probably be sweating anyway, frankly. But, um, so I'm practice innovation manager uh, at Uh, what does that mean? So, I work in the knowledge management department, which is in the information services department. And um as far as it relates to data or data, I don't know how are we pronouncing it here. Um I'm also Canadian so I just chalk that up there too. Um uh you know part of my responsibilities include um making things um easier and more efficient for both our attorneys and our clients. Um so sourcing um you know getting tons of demos of implementing various you know vendor technologies or building technologies um in part to harness a lot of the data that we gather in u mostly unfortunately unstructured ways which is you know what really um attracted me to the position to this opportunity at Kulie because you know we serve for example you know my interest was in startups um initially I practiced in Toronto know where I'm from
Then did a masters at Stanford in in law and then started, you know, working in the startup community. Um, and what what was really interesting to me about the position at was uh the fact that we serve, you know, hundreds if not thousands of of startup companies at any given time. And a lot of that data when I first joined a couple years ago, two and a half years ago, um, was not being harnessed at all. Um, so yeah, that's been a challenge. And by data I mean deal terms um capitalization tables and that you know some things were in spreadsheets some things were on CAPMX like on online platforms um um and even you know things like offer letters like all the data that's going into offer letters like we have a ton of information at our fingertips but the focus until now has been on um you know just document processing and you know not so much on capturing data in a structured way for analytics. etc. So that's my one of my focuses. Hello, I'm uh Jason Lai and um I'm the open data program manager for the city and county of San Francisco
Um I uh you may know uh the program as data SF. You may have heard of it, you may have not. Um hopefully you have. Um uh I work in for uh the um the office of the chief data officer uh who's Joy Bonaguro. Um and she set sort of a policy and strategic direction for this for the city as it relates to um uh what started as open data um and still be is very important uh to our mission um but really is or the core thing that we focus on is empowering use of the city's data. Um, and what because one of the things that Joy learned very early on in in doing her listening tour um, at the city was there were actually a lot of internal users that were struggling with data use. It wasn't just the public. It wasn't just about getting access to data
And so really the the problem is broader than open data. Um and so one although I'm the open data program manager um focus really on uh this this broader effort around um making data access easier across the city uh both to internal and external users. Um often that's a ven diagram. Open data can serve both internal and external um but sometimes it can't be open. Um and um what we're kind of struggling with right now and and really we're not struggling with but we're it's very similar actually um we've had a opportunity to talk to some of our private sector um uh peers. Um and what we've realized is that a lot of the challenges that we have at the city are the are similar um that their challenges of large organi datarich organizations um just trying to grapple with how do you use tools, how do you deliver data to users who need it. And one of the sort of uh complexities we have is that we also have legal and and policy frameworks that are very very different um although not entirely. Um and so we have to grapple with that
But we also manage you know the city ha has literally departments from a toz airport to the zoo. So I mean you're talking about 50 different departments some human services criminal justice I mean very very very different environments as it relates to data. So uh taking a comprehensive and approach to that across the city is challenging. Um and departments often have to do it on on their own. Um, and so the office of the chief data officer, we're trying to figure out ways that we can strategically address those those concerns without telling people how to do something, right? Without demanding that they do it a certain way, but but rather empowering them with the tools that they need. So that's what I'm doing. Um, I'll pass it on. Hi, my name is Gary Sanga
I've been told I'm the first serial legal tech entrepreneur. Uh, Intellig, my first legal tech startup. It's the market leader for preparing SEC filings like IPO documents, annual reports, and quarterly reports. And we use data analytics to tell people what's market. Um, my second startup, it's called Lit IQ. And the problem we're trying to solve is frankly drafting errors and screw-ups. So, it's a big problem. So, just so you guys know, about a quarter of all commercial disputes are made possible because of a drafting oversight of one form or another
missing language, conflicting language, ambiguous language. Um, and it's not because people are sloppy. Uh, for all you people that are founders of startups and hope you have, you know, an M&A exit sometime, uh, you appreciate how big these documents are. The average merger agreement has doubled in size in the last 15 years and folks still only have 24 hours in a day, right? So, this stuff is not easy and people are human. They make mistakes. So, we're using computational linguistics to develop a diagnostics tool that can scan documents and find bugs that the lawyer may have missed. The type of documents that we use, we use contracts for training data. U we annotate them for ambiguities and whatnot
And we also analyze case law so that we can then apply the lessons and rules from that into the software to help prevent needless litigation in the future. I'm also a professor at the University of Pennsylvania, which is a great school, right, Alexi? Pretty good. And um I'm partnering with the computational linguistics faculty. Another great group, right, Alexi? Very awesome group. Uh he got his PhD there. uh with this project uh lit IQ. So that's my story. Hi, I'm Lisa Malashenko
Uh by way of background um I used to work at IBM in the energy and utilities practice and I specialized in uh smart grid and smart cities and things like that. And then about five six years ago I moved over to the government sector and now work for the state of California at the California Public Utilities Commission which is a place that uh regulates all of the infrastructure in California. So all essential utilities, gas, electricity, water, uh all the railroads in the states and uh transportation including Uber. And one of the stories that I like to tell is when I was still at IBM and I was looking for for a job and I would came in to interview at the PUC, one of the first steps is you have to take a state service exam. And that exam entails reading an article and then basically summarizing it. and they gave me a floppy disc onto which to save um my you know replies to the to the questions and um I wasn't I wasn't really sure if it was a joke or not but it wasn't and I was given very specific instructions of how to write my name on the on the floppy disc cover. Remember those stickies that it cames with? Uhhuh. Um so that and that's five years ago and um so that's the state of technology in the government space and as I started working there um I learned of one of the reasons why we're still using floppy discs and that was when I tried to upgrade one of the databases that we use um and it's a it's a database that I have for gas and electric audits and it's an access it was programmed by some guy in literally like in the early 90s and I was like okay this isn't very good um let's do something more creative and I put in a project request through a state system and then I learned that it is a two and a half to threeyear approval cycle so not a project cycle approval cycle that has about four different steps of submitting proposals of different detail to the uh California department of technology that then you know reviews it for six months and then gives you a green light to go to the next step of the analysis
Um and that's just to start implementation. Um from there it basically proceeds at pretty much the same uh pace and one of the catches is that you can't substitute your um original request for a different technology down the line. So when I was requesting for this upgrade project um three years ago now uh whatever tool or solution I selected as um as what I was requesting that's what has to get implemented. So um I'm not going to go into more details as to why that's a problem but that is how u the state system works. So one of my messages to um to to everyone is it's not impossible to sell solutions to the government but it is extremely complicated and you need to know what you're doing and how to navigate uh the government system. There are areas where it's easier such as defense and security. uh some types of agencies or government entities are easier than others such as cities are a lot easier and local um government entities are a lot easier than state or federal. Uh but what I want to focus on today is really opportunities for government and p and private sector partnership to solve big problems
And I'll give you one example of a problem that we are uh looking to solve right now. Uh we've literally been having workshops on this and that's utility polls and it might not sound very exciting but we have about 4.2 to utility polls and that's what you think it is. You know the polls you see outside in the state of California and they actually present a 10 over10 billion dollar problem because they're old and they're continuing to age and to replace and upgrade those poles it costs billions and billions of dollars and if a pole fails it can potentially lead to very serious consequences including uh massive wildfires which has happened in the past. Um so what do you do? Do you you know the utilities right now um have um systems that are very basic in terms of trying to predict of way poles to replace. They're mostly based on age. Um but it's very inefficient especially because the utility infrastructure go built out in in sort of in in phases and now everything is failing all you know the average age of a poll is about 50 years is what we're expecting and over half of polls are already over 50 years old in the next decade 80% of them will be over um 50 years old. So what do you do? Um so that's uh kind of the issue that we're trying to solve as a state and there's different uh potential uh technological solutions in terms of predictive analytics uh for asset failure and also data collection in the field of building apps and really being able to connect different data sources to manage this issue better. Um so I kind of give you a mixed message today
Um, working with government is really hard, but I think there's some exciting problems that if we all come together that we can figure out how to solve. Thanks. Um, hi, I'm Nicole Shanahan. I wear two hats. Um, one is the founder and CEO of Clear Access IP. It's a startup that's trying to fix the patent system. Um the second hat I wear is as a residential fellow at Stanford's Codeex and Codeex is a joint center between the law school and the computer science department and there um almost all of my work right now is dedicated to reforming the criminal justice system. Um so two big problems and two problems that are really great um for a technological solution
Um both are I've tried to draw comparisons between the two and one is very uh impersonal. Um patents are are you know paper and and invention and um business. Uh and then you know the criminal justice system really is um it's it's it's all about kind of you know humans and welfare and fairness. Um, but I think the grand promise that data offers us is the ability to see these systems um for really what they are. Um, with patents, it's about um really understanding the pace of innovation and understanding transactional levers. Um, and so there's a legal theory called transaction co cost theory. And this was um spearheaded by a man named Ronald Coos and his theory was that where transaction cost um equals zero, there's no place for lawyers. And that's because lawyers are exchangers of information
They're the ones that bring party A and B together where they otherwise wouldn't come together. Um, and this really matters in the patent system and in a system of high innovation such as our society. Um, we have people that are coming up with things incredibly quickly and we need to find a way to pair these ideas and make the ownership of these ideas fair. Um, our current patent system does not do that. Uh, the only way to really monetize um these assets is incredibly aggressive um and has led to an entire industry called patent trolling. So, you know, it's it's not a simple problem um to solve, but I certainly think that in in the next 5 to 10 years, we're going to have some very unique approaches, including applying, for example, the blockchain. Um what does a blockchain patent system look like? It it could it could happen in the next 10 years in which people have ideas. Um they use Ethereum's platform and they can hash those ideas
um and and it's a record, it's a ledger of of creation, and that's what the US patent system is um effectively. Um and then on the criminal justice side, what what I really care about is understanding um you know, what is actually happening. I mean, we see anecdotally things on the news and it's pretty atrocious. Um but but deeply, when we really really dig deeply, what do we see? And so using big data approaches, things like data mining and machine learning, um, you know, we're going through some, um, government databases right now at Stanford that you just wouldn't believe how how little, um, our government knows about what what's actually happening. So, so yeah, my goal is to bring parties together to work on uh, these two um, subject areas. Thank you. uh I think there are multiple themes emerging from this you know I was kind of wondering we need to tie together two areas and I think there are clearly points of intersection right so for instance criminal justice systems government and and law right there right and also IP I think IP is very crucial to this space but also in the whole patent law is hinging on on on government uh idea to have patent law right so in Japan they didn't have patent law until very recently in 19th century and so when one kind it's introduction in you know major era kind of led to may maybe one of the things which led to to rapid development so so obviously government has a hand in in uh in this so uh so it's I think uh and but for whatever reason right these two areas are very conservative so we do not see much adoption of all these kind of you know we can call this Netflix- like technology so we have a talk uh on Friday in the life sciences day which is another slowly moving area you know using Netflix-like approach to to healthcare here. So obviously you know everybody knows Netflix and or Amazon right? So the the question everybody was posing is you know why can't we bring these things to to law and government and there are multiple reasons lawyers like to bill by the hour right they don't want to improve efficiency but I mean that cannot be uh all the all there is to it and government obviously has a lot of fret but that cannot be all the the there is to it so I I'd like to pose to each of you the following question uh what is the best example of innovation you've seen in your area so so you uh you lead startups to kind of uh uh uh have positions in in these big corporations uh where you can affect change where uh what are the examples what are who are your role models? What are the shining examples of digital transformation in law and government you know about and what is it that prevents these kind of examples from taking hold spreading faster in any order? Shining examples
Yes. Um there's a I mean there there are a number of really great things happening in this space and there it's actually moving very fast. Um in some ways it might I mean government often seems slow but um I I don't want to risk offending some of some of my peers in this but um I'll highlight a couple that I I think are really important uh to pay attention to. Um, and they're probably things that most of you know about, but um, I I'll mention them anyway. Um, I think what 18F, um, which is at the federal level has been doing and then is now evolving. Um, uh, is really important to watch. You I mean, they're you're talking about essentially a consulting model inside federal government that is, um, not just I mean, they're doing open source, they're doing it well. They're building applications
They're they're rethinking the way that they publish data. Um uh but they're also thinking about procurement and um working with private vendors. And so they're we we are I mean the city um we're constantly watching 18F very closely. Um what's nice is they have offices in both DC and San Francisco. um which is also kind of an innovation in in federal government that um that these folks work remotely um they um uh they're they're in this sort of center of innovation um and uh it's diverse group of folks that are really doing some really important work I think um there are other great I mean USDS is another um variation it's it's slightly different from the ATF model but also at the federal level um coming out of the the White House um put essentially embedding people in other departments which I think is an interesting model um and something that again we're we're watching very closely um and then there's always gov.uk which was the one that kind of started them all and I think um for us we at the city we did sort of a comprehensive analysis of different folks doing different doing sort of digital transformation and gov.uk UK, we really there's a lot to be inspired by. We wouldn't apply it exactly the same way in this in the San Francisco context because uh it's just different. Um but definitely there's some a lot of um inspiration there. Um so what's innovative in the law firm world? Um well there are a couple of interesting developments mainly around um again trying to structure this largely unstructured data um you that we have in our document management system
Um so again we have I mean let's see between January and and I think March we produced 5 million documents. Um, so Kulie is um a 900 lawyer law firm, which which is pretty big. I mean, there's much bigger. Um, so I can't even imagine what what that's like at at like uh like the the DA Pipers of the world, for example, where there are like 300 3,000 lawyers, but um you know, so so there are a lot of strides being made in u machine learning technologies and how they're being applied to um sorting data, classifying data. um this unstructured data whether it's in the documents or even in um you know so whenever a lawyer bills you uh they typically just say um you know billing to your general corporate matter and you know here's a free text narrative and that narrative can can change uh wildly between people uh between different lawyers. Um obviously if they were able to put it into specific tasks um and use codes for those tasks and there are some efforts underway to create a a national classification. Um but frankly on the uh lawyer adoption side I feel like that's kind of unlikely. Uh lawyers are are notoriously difficult to um modify their behaviors
Um they're you know welloiled machines. they have their processes to disrupt those processes and add something like that in very difficult now if you have a system that can automatically classify those things based on the narrative. I mean, now we're getting somewhere. The um and you may have heard there's like the first AI lawyer that's come out, uh Ross uh intelligence. Um and I will put that in quotes, big quotes, big fat quotes. Um first AI lawyer. Um and you know in the research world that you know those kind of solutions are a bit more feasible um because they're not really dealing with um and again Ross um is a Watson powered legal research tool. Um legal research doesn't really contain any sensitive client information
You just you know it's like running a Google search right? you don't necessarily know who you're running it on behalf of um for what reason. So anyway, that leads me to the biggest, you know, what I view as the biggest um barrier to legal tech adoption um at law firms at least is the cloud. It's a problem. Um, we have clients that won't let us use it for anything. Um, you know, for for storing and processing their information and which makes it very difficult to silo clients that do, you know, that are okay with it and their information versus clients that aren't. And usually the clients that aren't are the ones that are, you know, regularly paying their very, you know, significant bills. So um it's really those technologies that have developed hybrid deployment um technologies or or approaches um or just fully on premise deployments. Those are the ones that um you know we we tend to find greater adoption for but of course also tend to be less effective than the purely cloud-based ones
Um anyway move on to someone else. I mean for me uh in the legal services sector the biggest change I've seen uh especially from 10 years ago when I used to practice law is the advent of quantitative analytics. I mean when I practice 10 years ago you couldn't find a lawyer that knew how to use MX Excel. We literally just use Microsoft Word, right? And you got to look at the profile of the lawyer, right? The modern lawyer is typically person that probably didn't like the sciences or the maths, right? They're more into the liberal arts or the social sciences. And I think that created a barrier for quantitative analytics. People just didn't didn't gravitate to it naturally. Uh these days I feel like virtually every startup probably does some in the legal services sector does some form of quantitative analytics. uh it's a long time coming especially if you compare it to say the financial services sector right where it came in the 70s and 80s right the bloombergs and all that um so that's the biggest change I've seen where the lawyers are finally becoming getting realizing the power of the quants and how they're now beginning to adopt technology that can give them insights through data um I I think the examples of um successes that I've seen are where uh government acted to liberate data
So uh in my industry the example of that is the deployment of uh smart meters that you may have heard about um and probably have at your home uh that allowed to uh collect interval data about energy usage. And then what uh my agency in particular worked on is making this data available to uh to everybody really who wanted to build apps and uh offer services uh to people's and businesses of how to use their energy or um all kinds of things really. And um so what the uh so what the California did was basically develop a uh platform for people to register to be able to download uh this data in a common format and then also develop data privacy rules that um basically stripped any personal and identifiable information and rolled up the data to a certain level that it was um anonymized. So in my experience of dealing with uh big government is that it's a lot faster when the government allows for something to happen and kind of gets out of the way. I think the big exceptions to that are uh really defense and cyber security where government does actually collect the the information and and builds the tools to solve um uh problems. But in other areas, I think yeah, it works well when the government just allows uh the access to the data. Yeah. Um along that those lines, uh someone from the California Department of Justice recently told me that there's a 1930 penal code law that gives the DOJ access to every single database, government database in the state of California for the purpose of policymaking
This means that, you know, if you're the DMV or the health department or any government agency, um that data is subject uh for review and no one's really called on this code. It was it was um the founder of California's Open Justice Initiative found it and he saw it and he was like, "Oh my god, this is this is groundbreaking. This is real access for real change. Well, so hopefully can somebody like call the uh DMV line waiting time into question and analyze what's going on there? Uh so but should you be a government body to request this? Is it open to the So it's not uh it's not open to the public. So it's for the Department of Justice to access um on behalf of the public who can call like can the public act and compel Department of Justice to request the review of certain piece of data? Um yeah actually uh it's I mean it's not like a a FOA request um which is Freedom of Information Act request. Um that's a federal government request but this is slightly different. Uh it's you know they're hosting these open justice hackathons and they're open to the public and these are great place to come to put in your requests and to get in touch with people within the DOJ that um can really put in to yeah put into process something interesting. So we should look this up
Uh I want to pick up on something you mentioned. I think you all mentioned that you have all different kinds of documents, right? So I think that's another area where government and and law are very uh much related is that uh I wonder how much of your day uh do you spend in front of documents? What do you do with them? Uh what people in your profession do with them? And how do they exchange these documents? How do they uh iterate on them? Uh and I have another hat. I'm a chief scientist at Nitro which is a PDF uh provider and so you know it's an established company which basically competes with Adobe so we have a reader called Nitro PDF so so it's it's a pretty wellestablished legacy technology but what I see a lot of people are sending files to each other right and uh so I wonder uh basically how much of your life are these documents and and essentially uh what are the productivity improvements or deficiencies you see happening around documents. I love this question. Um, I wish you let me use my PowerPoint presentation because my first slide was a closeup of something that's called a uh rubber finger, which is a little rubber thing that you put on top of your finger um to help you flick uh through paper faster and also to avoid getting ink on your hands, which is a problem that most of you, I'm assuming, don't have. But in my office, we actually have a supply um cabinet full of those damn things. Um because we are still very heavily actually fully paper um agency. There are a lot of processes that are completely lack any kind of automation to the point where we have um like one of the things that my group does is process these applications of uh like limousine companies and drivers to get your TCP number and you have to go and fill out this form
And we did build um a a website where you can go and fill that in while on the back end it just prints it into paper and then we have to go and file it and do everything else and then you know send a physical paper confirmation back um to the person and the and those forms don't actually even have a field for email. So even when we have made attempts to digitize parts of this process, there's again limitations that we never collected emails for anybody. So you still have to print the invoices and you know physically stuff them in envelopes and mail them out. Um so I I personally see a great um range of um everything in government. you know, you have pockets where um like when I work with the office of emergency services and DHS on certain infrastructure security issues where you can see great sophistication and then you see other parts of the process where you know you're using the rubber finger to you know flip through uh hundreds and hundreds of documents that are not digitized in any shape or form like that example of the front end of the web. And it gets people so frustrated because they think they filled in the web form and they're like, "Why the heck is it taking you so long?" Right? But on the back end, it's literally getting printed and we don't even have a digital record of it because as it's printing it, it's not even saving a PDF um version of it in the database. I mean, there's literally just a script that says, "Okay, somebody fill this thing in, hit the print button, and you know, that's that's how it goes." So, um, yes, there's a lot of opportunity for improvement. Let's just end it on this positive note
Um, I'll just add real quickly. I am personally incredibly lucky to never I've never had to use one of those rubber fingers. Um, in in our office, I think what what's great is we've been a we're leaprogging. We're we're really thinking looking at, you know, how to make data accessible. There's lots of laws and rules around making documents and records retention, all that sorts of stuff. So I in my day-to-day I don't have to interact with it but in working with departments um I' I'd say I was actually pleasantly surprised it came up less than I thought it still comes up. Um there are I think we have a lot of departments that are in this sort of in between stage where there there's like their records like really good data structured data around something but not all the elements were put in the database. So you still got to go and pull that permit record or you So there that that gets a little that's always a little frustrating because you you'll get a data request and then you you realize oh that that that never got modeled
It's it now it's it is maybe as of 2013 but before that you still got to go and pull those records. So there's still work to be done I think going forward there you know some departments are leaprogging they are putting forms online there and those are actually going into databases and I think there's an opportunity still to um the the tools have gotten easier and easier and I think uh I I hope that um my peers in in government uh start to leverage those more and more because those workflows are becoming even easier. I I'll be quick on this one, but um we're one of my major projects is actually uh coding up all of the legal forms um to use in a document automation system. The one we use is called contract express. So, you know, I take a word document and I, you know, put conditional spans around things and create variables and computable variables, etc., etc., and then I um I create a questionnaire based on, you know, those that that markup. Um what we're going to be doing soon I hope is actually making it so that um this questionnaire which is currently only um really completed by the lawyer internally now um is going to be actually completed by the the client in part. So an example would be an incorporation. So currently our attorneys when they onboard a new client they send over at best a PDF a fillable PDF um which then comes back in and previously you know they would take that PDF and then you know create the you know separate legal documents opening up previous ones and then saving as etc
Now they take that PDF information and you know they put it into the internal document automation system. Going forward what I'm hoping is that the client will fill in much of the information. it'll go right into the document automation system and documents are created like right then and there. Um, the benefit of all this as well is that we're starting to capture all of this data in a structured actionable way, not just residing in disparate PDF files at best. Sadly, unlike Brad's firm, most law firms I've seen are still super old school. Uh you mentioned PDF being a legacy document. Most attorneys I deal with still use vanilla MSWord. I mean the same product you use to prepare your high school essays
I mean these lawyers are using to prepare incredibly technical documents, right? Which is just maddening when you think about how technical the subject matter is when how you know fallible human beings are. And yeah, it's just really surprising how old school most lawyers still are, which is great for folks like us, right? Cuz it creates an opportunity where you can use tech to just make people's lives easier, make their work product better, help their reputation out. So that's been the biggest surprise for me, just how old school it all still is. I just have a quick comment. Um, Codeex has a project right now called Computable Contracts. Um and the idea is is to create a computable format with declarative language to you know basically execute on on the terms. And so as you write the contract it actually goes into a database and then the database has all these rules and and you know some logic that can um you know help you work through all the terms of the contract. that is an it's fairly new and I think that you know if there's people in the crowd that are interested in being connected to that um it would uh we would love love more minds on it
Interesting. Uh so I think at this point maybe we'll ask uh if the audience have have questions for any of the panelists or for the panel as a whole. Okay. So, uh, we can bring the mic to you. Um, if we have a volunteer. Yes. So, Matt will bring the mic to the question. Yes
Um, well, introduce yourself, please. Vlad. Um, for the legal kind of profession. I thought I hate to name your competitor's names but uh Wilson Sancini and Perkins already have like automated like contract filling for incorporation and kind of percolator and the and the open contracts. So I thought that was already a trend that so yeah um we do have something like that. It's called go cool.com which I do manage actually. Um and yes you can go in there and but that's not for for client use. Um the version that we have internally is is you know much more robust I I'll call it
Um yeah but thank you for giving me the opportunity to plug go. Yeah. And a followup also, do you see also the first cases where like human against AI cases and maybe somebody starting to specialize in that like AI made the fold and then the humans suing AI systems? Um I I I have not but I look forward to the day when I when I see the the um hey the you know well so again there there's a there's a distinction between um I don't know if this is what you're driving at but for example legal zoom right um they escape liability as a lawyer right because they're not providing legal advice right they are you know you'll see see it everywhere like these documents were prepared at your direction I think it is right and that's because we're just saying to you how many shares do you want you know to purchase right um it you know we just ask you basically you know questions we don't tell you what those answers should be in your particular situation we may say generally x or y now where we you know where that could get into trouble in terms of AI versus man um is where we start relying on vendors maybe like Gary over here um whose products are you know maybe making decisions right I mean it's a really interesting area right at what point does a software product become a lawyer and I mean I've been you know I've been I've been racking my brain in particular about like Ethereum right and smart contracts like at what point does someone who's coding a smart contract become a lawyer like are they giving legal legal advice when you're drafting a smart contract, you know, at what point does that happen? Um, but no, to answer your question, I haven't seen it yet. So, uh, this is question for, uh, the codeex project that you were just talking about. Um I I I was kind of wondering if this contract system is going to be something similar to like if this then that or zap year of like oh this thing happens and you can declaratively say oh this amount of money should be dispersed or just introduce yourself. Sorry. Uh introduce yourself. Oh I'm sorry
I'm Samir. I work at Mattermark. I lead the machine learning team there. And like what's the I'm actually also a fellow at Codeex. not not as deeply involved as Nicole, but um I I I would also be curious to know how that computational contract project reacts to things like Ethereum that are coming at you. Yeah. Um so there's a few ways to approach it. So one of it is um standardized forms for certain contracts, right? You can have a standardized prenup, you could have a st, you know, there's a number of standardized contracts um out there
And then once you agree that these are the key terms, you can do a similar if then then if if this then that type of declarative form. Um the the project is you know it's at a point where does it make sense to borrow another language? um you know do is it you know an XML type of a markup or is it you know like what like what should it be and how do you plug in the more complex contingencies into it. Um that's still being worked out and uh in regards to Brad your point of bringing up Ethereum Ethereum is really interesting but it's still a little hard to use. Um so there was uh we had a we we have lots of presentations um at Codex. Every week on Thursday we have a new legal technology project that comes and presents and there was recently one called Doc Assemble and Doc Assemble put together an entire library um that was trying to coach lawyers to set up their own computable contracts. So it was like, well, here are kind of the key things you need to get and then here are the things you can add in. Um, and he thought that, you know, anyone could figure out how to read the instructions on on how to use this library. The reality is is that lawyers are not even capable of doing really basic um deployment of code
So it that's the understatement of the day right there. Yeah. But should that really be necessary? Like shouldn't the abstraction not really need people to do that? Exactly. So that's the design layer. Um and there's a there's a another fellow in my group um group Margaret Hagen and that's all she does. she is is like, "Okay, let's think about the people and who they are when they go home and wake up and after they have their coffee, and then let's think about, you know, when they sit down at their computer, like how much time are they really going to spend, you know, trying to figure out something new?" Um, and what colors do they like? [Laughter] Is is there potentially a chance that like something like lit IQ or even with lit IQ there's like unsupervised learning of what the contract should be doing and it just turns it into the computational contract. Yeah, I've seen some stuff like that. Um well then you go into kind of the NLP uh area
You should work with Gary and and that's a Gary Yeah, that seems like a Gary question Stanford Codex. Oh, cool. We've got them all. This is already happening, huh? Uh, we have a question back there. Hi. Uh, my name is Matthew Seal. I work at OpenGV, which is a company down here um in Redwood City. But my question was more you you brought up the kind of hackathons and the idea of uh you know the Justice Department open um kind of access to try and get data and get uh code to you know do something with that data
Have you found in in the government with the hackathons and the government sponsored coding events, have they actually contributed to improving some of these problems you've described or have they so far just kind of been interesting ideas that haven't uh come to full futation? Do you do you want to go first? Uh yeah, sure. I'll go for it. Um, you know, I I I personally find that um, you know, while those can be useful to jump start some initial ideas, but we're not getting enough traction to actually, you know, turn these into really serious projects. So you know I've gone to some energy related ones and most of them result in somewhat superficial um results because you know even you know where whichever industry we're talking about whether it's energy or you know you're really talking about government a lot of the problems are really not because there aren't you know some obvious solutions that haven't really been thought. It's that uh the systems are extremely complicated whether they're regulatory systems or u just the industry dynamics or people components and um there you know and there is a a barrier to entry when you're starting to talk about um some of those things and that requires real engagement. So I get a lot of uh requests for example from startups or you know data scientists who have a bright idea and they come in all excited and they try to sell me on it and I try to explain to them well it's not that I don't buy into your idea it actually sounds great but in order to actually materialize it you're going to have to engage in our process you're going to need to understand how you know state procurement system works if you're trying to sell to an agency if you're trying to to solve the pole example that I'm talking about it's not going to be enough for you to go sit and code the solution. That's not the complicated part. The complicated part is integrating the into the ecosystem of governments and research labs and pole owners and pole attachers and really you know solving it as at that level of all the parties bring coming together and that's where I kind of see that gap that the the hackathon type model it really focuses heavily on the coding part but a lot of the solutions um really would require multiple people coming together that's not, you know, just focused on the on the code part of it
Ditto. Um, so I I um real quick, uh, yeah, I I think there needs to be um a a maturation of the way we we do hackathons, particularly when it when it relates to the public sector for all the the reasons mentioned. Um, I think there there a shift is coming. Um, I think as data becomes more accessible, it becomes easier to have ongoing engagements, to have conversations about data, uh, Kaggle and and other platforms that allow you to explore. And so you have these sort of ongoing engagements that are it's not about a single event. It's not about it doesn't put the pressure to like produce, right? Like we're it's not Facebook. I mean, I I love the fact that they can build products so quick. We can't do that
Like we we do have to deal with all those those other things. And you know, some people may think that's a flaw, but I I think that's just a feature. Um it there there's just stuff that you have to deal with. And I I think couple models that I really like or one that I really like is um what Caravan Studios does out of Tech Soup and they do these generators um and they actually get with subject matter experts and they have a whole process to really generate things that that can actually get get adoption. Um and then National Day of Civic Hacking, uh civic makers, I don't know if Lawrence is here, um but uh National Day Civic Hacking is despite the name going to be focused more on um this time around, uh doing that that part like getting to understand the issues. We might not even write a piece of code that weekend. Um and then code for San Francisco meets every week. So we'll continue that engagement
So that that feels more promising to me than these one-off kind of events. Thank you. And so uh I think we're kind of uh coming to uh uh to our um uh ending time. I want to ask uh each of the panelists uh uh a wrap-up question and and kind of Brad inspired me when we talked with Brad about this panel and I mentioned various legal startups uh which we have. It looked like all of them tried to sell to him. So he knew all of technology because obviously legal startups that they coming to big firms and offer their their uh products. So uh I kind of want to pick up on this and basically ask you guys so we have a lot of startups and innovators in the community. So they want to disrupt either legal or government spheres
So uh what would be your advice to startups which want to sell to legal firms or engage in this market? And for the government folks, I would probably kind of pose two options. Either, you know, if you want to affect real change with digital transformation, what should you do? And if you're a startup hoping to sell to the government, you know, should you hope so or what should you do, right? Like I will kind of pause it two ways, right? Because you can either try to, you know, enter the market as a provider of government services or you can try to kind of join the government and and do something from the inside. Um, one, um, don't build cloud only if you plan to process or store client confidential information. For us, I'm I think that there I'm sure there are some firms out there, perhaps ones that, you know, that are our size two or bigger that may be more amendable to that. I'm sure there are solos and mediumsiz law firms that may not have an issue with it. We do. Um, and I you from what I've heard a lot do. Um, uh, two, um, shy away from that VC money
Um, try and bootstrap as much as you can and have patience and have like, you know, I'd say a 12-month runway at least from the time you start selling to a law firm. Um, we aren't as bad it sounds as as part as the government perhaps. I wouldn't say we have a three-year sales cycle, which is crazy. Um I mean, like I thought we were bad, but um you know, I'm being asked soon to put together a budget for next year, right? Um there are five things that I still want to do this year that came up this year. And you know, so I'm I'm working hard internally to try and create like an R&D department with a budget that we can more quickly stand up, you know, prototypes or, you know, um pilots. But you know in the meantime I mean we have a hundred things to do and we have you know one guy that does the security evaluations right of of prospective vendors and um you know you just have to understand that even though a law firm is really big it is does not necessarily have the kind of resources that like a large multi multinational corporation has to vet deploy test products and we um you have to understand that to for us to do a proof of concept or a pilot there's a high bar because um we don't have like a group of lawyers that are dedicated to trying new technologies like you know part of my job is bringing that lawyer mentality and vetting something before it gets to you know the associates which you know which before that gets to the partners and I mean u putting a new technology in front of lawyers even a small pilot group is takes a massive amount of resources so thank you for your empathy sympathy and I think that that's it. Um, so over my two startups uh selling to law firms, learn a couple lessons. So first off, if you're in the B2B space and you're selling a software product, look, the most compelling thing you can do is help your customer make revenue, right? This is why the financial service sector is the largest purchaser of tech, right? They can help them do quicker trades, whatnot
Um, you're never going to be able to do that in legal, right? You're never going to help a lawyer make more money unfortunately, right? Uh so what are you selling then? What what's your pitch point? You know, I try to mentor a lot of legal tech entrepreneurs when they come through. Most folks when they come in, you know, they come in from a different sector and they want to sell efficiency. They want to sell efficiency. And lawyers, I guess they conceptually get efficiency, but it's not their top-of- mind pain point. Geez, I need to be more efficient, right? um more powerful is if you can solve a painoint, you know, whether it's just current workflow that's crappy or tedious or whatnot. Solving a painoint, I think, is the best you can do because you're not going to help them make more money. Uh and don't try to sell efficiency. It's just not going to work
Uh otherwise, I concur with everything Brad says, don't take venture in legal. Uh you should have a pathway to break even, ideally with 24 to 36 months. Um, but once you do sell, it's a great market to sell to because they're incredibly loyal customers. They don't really turn out. It's a beautiful market if you can just figure out the customer acquisition process. Um, I guess switching gears to uh to government. Um, you know, I would say that there are three primary models of selling or engaging with government. And first is directly to government, right? and you the government agency is your client and there are firms who specialize in that and are very successful
You do have to be extremely patient and there is a a path to that that I won't go into details but you can reach out to me if you're interested. The second path is uh public private partnerships and um that's really when you're not selling to government but you're trying to get on the ground floor of solving a problem. So the example of that is polls. How do you do that? You have to be an industry expert. You have to understand what the problem is. You have to understand who the players are. Follow these conversations and start kind of getting in on the ground floor to become part of of of solving that. And the exciting part there is that it's a lot of work, but I mean there's a ton of money because like with polls, it's a 10 billion plus dollar problem
So there's a big potential in those kinds of issues. And then I think the third kind of sales path is is following the R&D uh model and so knowing what kind of issues the government is interested in solving and then learning the grants um paths and how to get those government uh money. So there's always you know you can apply for a grant for your idea and you know there's different phases of it. that can be a very early stage R&D all the way through to proof of concept in real life. Um, and that's its own path that requires knowledge of how to apply for those grants and uh get the R&D funding. So I would say figure out which of those three models uh suits you best and then learn how to work uh the area and go forward. Um I just have a quick anecdote about uh bodywn cameras. So there are very few good vendors in the body worn camera space and it is a huge need and the government is putting a lot of money in but um a lot of these bodywn camera companies sell us uh data storage with the hardware
The problem is it's really crappy and really expensive to the point where there are some police departments in the United States that are having to decide between hiring a new police officer or buying more data storage for camera footage. And this is not even reviewed camera, you know, it's just raw film. They haven't even gotten use out of it yet. So, I would um you know, encourage people to really think about this space. I think it's going to be an important one. Um it would be hard for me to give I mean it's there's a lot of uh challenges in in selling to government in general. Um I I think it it it depends on the scale. Um so at the local level um if you're selling to a single government and that's it, it's actually not that hard
But the problem is that that's not what you want, right? Um uh in San Francisco for example, yeah, the vendor process is a pain in the butt and but um the information is all there. I actually talked to a startup in San Francisco that said, "Oh, I w I walked down and I went to the office and I turned out I had everything I needed to be a a certified vendor with the city, you know, and you know, but that doesn't you like doing that for every city, right? Like that's that's not the business you want to be in." That's why folks like these larger firms um they have big sales teams right because they can they can do those cycles they can get on those um so it is hard um I think there are a couple of things to watch out for um I think coming down the line that is um I think government is going to get a little bit smarter about procurement um particularly in the technology space because there is a broad acknowledgement that the way that technology is being purchased it it cannot be purchased the same way that we're we purchase pen pencils and other other commodities and um and there's this blurring between services and technology and that it's starting to get a little bit more cleanly defined and so that's happening I think um that will continue to happen and so my broader advice is um uh you know find find your community a little bit um there's a lot of there there's emerging interest in this right civic GVtech. Um, it goes by a couple different names, but there are there are there are others in this space that are trying to figure it out. And I I would band together, understand that command how governments work and and figure that out together rather than on your own. Um, because it it's it's hard to navigate by yourself. Um and uh and to echo point made earlier about pain points. I mean I I've seen a lot of activity or behavior where someone has a product and they're just trying to make it work for S city of San Francisco and I'm it doesn't it doesn't work. Um uh don't this seeking behavior is you like if you really really want to like it needs to actually address an actual painoint
the body cameras issue is a very good one like and that's a sort of a a hidden from view issue um that that you you've brought up but like these are the kinds of things like um you really need to do your research to really understand like what what what does government really need? Do we need yet another data platform that does blah blah you know standard things that we are commodity or you know what what do we need? So I think really asking yourself those questions and also one last thing one one thing this is just a personal bias I have unless unless the solution is very specific to government's problems I um I I always have an allergic reaction to any vendor that says I know the government um and I know that that that goes against what I just said but someone like particularly if you're selling a a product that is really really is a commodity is is something that's a platform that lots and lots of people are selling. I've seen a lot of people say I I've this is specific to government and it's not um and be careful of that because if if you are siloing yourself into trying to sell only to government but your your platform actually solves a lot of different problems at scale go there too. Um so don't you know I think it again this broad this is all broad advice. Um, it just depends on the product. It depends on where you're at. Um, I don't know enough about funding cycles to give you any kind of advice on what what to do there. But, um, in general, understand the space as much as you can. Thank you very much
I think, uh, that's probably all the time we had and I want to thank the panelists for deep insights into their fields. So, uh,