1 00:00:02,122 --> 00:00:03,899 Eric, how are you this afternoon? 2 00:00:03,899 --> 00:00:05,920 Fantastic, thanks for having me, Ted. 3 00:00:05,970 --> 00:00:07,011 Awesome. 4 00:00:07,011 --> 00:00:12,436 So you and I got connected on LinkedIn as so many of my guests and I do. 5 00:00:12,436 --> 00:00:20,583 And I think we were riffing on fractional chief AI officer and legal and one thing led to another. 6 00:00:20,583 --> 00:00:27,749 And you and I chatted and I think it would, it's going to be a great conversation talking about AI and legal. 7 00:00:27,749 --> 00:00:31,202 But before we get into that, let's, let's get you introduced. 8 00:00:31,202 --> 00:00:33,826 So you, you have an interesting path. 9 00:00:33,826 --> 00:00:44,461 You started your career in construction and then later pivoted to legal and you've worked in a variety of capacities, including a fractional chief AI officer. 10 00:00:44,461 --> 00:00:52,734 And sounds like you spent some time in the startup world as well, but why don't you tell us a little bit about who you are, what you do and where you do it. 11 00:00:52,977 --> 00:00:57,361 Sure, cut me off if I start to run long, because it is a twisted tale. 12 00:00:57,361 --> 00:01:01,204 I like to stay entertained as I sort of learn and grow through my career. 13 00:01:01,305 --> 00:01:02,726 Absolutely correct. 14 00:01:03,007 --> 00:01:10,213 During the dot-com boom here in Chicago, tons of construction, so I kind of kicked off my career general contractor, property developer. 15 00:01:10,995 --> 00:01:18,147 Had a good run of it, but definitely always kind of had that thirst for just truly intellectually engaging. 16 00:01:18,147 --> 00:01:29,310 work and I pivoted my career about 10 or more years ago to focus more on, you know, technology enabled business process, legal process, just really was feeling some 17 00:01:29,310 --> 00:01:36,512 opportunities as I kind of moved through my life and through the world for, you know, so many different things that could be done better. 18 00:01:36,512 --> 00:01:46,765 Uh, you know, in, the meantime, I've, uh, you know, spent a little bit of time on boats, on Navy pier, you know, got my captain's license to sort of like continuing the learning 19 00:01:46,765 --> 00:01:47,601 adventure. 20 00:01:47,601 --> 00:02:00,401 as it were, but went to law school for a brief master's program where I, you know, really drilled down on legal process, process improvement, tech enabled law and all everything 21 00:02:00,401 --> 00:02:02,021 sort of related there. 22 00:02:02,021 --> 00:02:12,601 And yeah, spent some time at a startup focusing on basically transactional law, legal, legal departments at very large organizations and how is it that they manage their 23 00:02:12,601 --> 00:02:13,381 language? 24 00:02:13,381 --> 00:02:17,513 And then I've been consulting for a while and then 25 00:02:17,565 --> 00:02:26,141 President Biden put out, you know, the executive order for, you know, every department, every administration should have an AI officer. 26 00:02:26,141 --> 00:02:33,076 And that kind of kicked off my market exploration of, who else is looking for, you know, an AI officer? 27 00:02:33,076 --> 00:02:44,994 And as I might've mentioned earlier, the most popular response that I get when I talk about that type of role, which everyone recognizes as being brand new is, wow, that's an 28 00:02:44,994 --> 00:02:46,009 expensive hire. 29 00:02:46,009 --> 00:02:55,586 Like that exact quote I've heard multiple times where people realize, you know, you're going to need someone that speaks business, but also definitely if everything that's true 30 00:02:55,586 --> 00:03:01,559 about the possibilities of improving legal process are true with AI, they're going to need to speak legal. 31 00:03:01,559 --> 00:03:03,101 They're going to need to speak technology. 32 00:03:03,101 --> 00:03:05,332 So, you know, kind of acting as as a translator. 33 00:03:05,332 --> 00:03:08,534 that's, that's kind of the kickoff to, where I got to now. 34 00:03:08,952 --> 00:03:10,923 You know, it's interesting. 35 00:03:11,484 --> 00:03:16,908 actually, don't LOL often, but I laughed out loud. 36 00:03:16,908 --> 00:03:24,313 So if I remember correctly, there are 62 federal agencies that now need a chief AI officer. 37 00:03:25,514 --> 00:03:28,296 think about the skillset that you've just described. 38 00:03:28,296 --> 00:03:36,322 How many people who are on the cutting edge of technology want to go work for the government who's 39 00:03:36,334 --> 00:03:39,518 20 years behind, they're still running mainframes, right? 40 00:03:39,518 --> 00:03:44,685 So it's like, I thought that was a really interesting mandate. 41 00:03:44,685 --> 00:03:47,469 And I laughed out loud. 42 00:03:47,469 --> 00:03:54,155 I pictured somebody like, you know, the recruiter on a rotary phone calling up to recruit. 43 00:03:54,155 --> 00:03:54,758 upgrade. 44 00:03:54,758 --> 00:03:57,413 Yeah, our eight inch floppy disks don't work so well anymore. 45 00:03:57,413 --> 00:03:58,165 Let's upgrade. 46 00:03:58,165 --> 00:03:58,935 Yeah. 47 00:03:59,094 --> 00:04:01,525 Yeah, that I got a, I got a big kick out of that. 48 00:04:01,525 --> 00:04:06,678 That was a few months ago and I don't laugh out loud often, but I really did on that one. 49 00:04:06,678 --> 00:04:14,483 Um, well, you and I talked about something that I have talked about at conferences. 50 00:04:14,483 --> 00:04:23,748 I was actually on a panel with a very esteemed group talking about AI, where it should live in the organizational chart in a law firm. 51 00:04:23,948 --> 00:04:27,090 And we are seeing more and more 52 00:04:27,272 --> 00:04:30,824 C-suite titles with AI in it. 53 00:04:31,165 --> 00:04:33,117 Very few chief AI officers. 54 00:04:33,117 --> 00:04:41,673 fact, I only know of one, but there are lots of chief innovation in AI officer, chief knowledge in AI officers. 55 00:04:41,674 --> 00:04:50,190 So, you know, what are your thoughts on C-suite roles with AI in the actual title? 56 00:04:50,190 --> 00:04:52,862 Do you think this is going to be a thing going forward? 57 00:04:53,713 --> 00:04:55,253 I think it can be a thing. 58 00:04:55,253 --> 00:04:59,735 I'm going to be make a general statement and say that it's probably good bet. 59 00:04:59,735 --> 00:05:04,446 It'll be a little bit more show than substance depending on the organization. 60 00:05:04,446 --> 00:05:17,219 Of course, I've seen law firms and other types of organizations where this is the final kick that they truly needed to really start, you know, leveraging all the data they have 61 00:05:17,219 --> 00:05:21,760 in the back end to improve and accelerate their processes. 62 00:05:22,301 --> 00:05:23,681 And so it's like, well, 63 00:05:23,681 --> 00:05:24,692 AI is the biggest thing. 64 00:05:24,692 --> 00:05:29,645 We're going to look great having AI at some level of our organization. 65 00:05:29,645 --> 00:05:39,302 So let's just make the first sort of 21st century technology step be, we've got a person we can point to that says that's our AI person. 66 00:05:39,302 --> 00:05:49,139 And they might be someone who's just literally playing around with ChatGPT GPT on weekends, but is coming from like being a full-time lawyer background all the way to, you 67 00:05:49,139 --> 00:05:51,831 know, somebody sort of picked up from a different. 68 00:05:52,399 --> 00:05:53,120 different industry. 69 00:05:53,120 --> 00:06:06,139 But if, if, we're talking about the business world in general, C level, uh, I've, I've worked with organizations that really understand the power of the technology, but also how 70 00:06:06,139 --> 00:06:08,860 their business works, how their organization works. 71 00:06:08,881 --> 00:06:17,907 And it's, it's pretty similar to a conversation I had earlier today, which is if I get the quote, right, the gentleman I was speaking to who has, you know, straddled the worlds 72 00:06:17,907 --> 00:06:19,469 between technology and law. 73 00:06:19,469 --> 00:06:20,497 He said, 74 00:06:20,497 --> 00:06:30,337 As soon as something gets operationalized, take it out of the hands of legal, know, it's like HR exists for a reason, all these different departments that are sort of legal 75 00:06:30,337 --> 00:06:31,617 adjacent. 76 00:06:32,197 --> 00:06:38,337 The other way that he phrased it was legal is by default bespoke. 77 00:06:38,337 --> 00:06:40,817 That's kind of the two sides of the coin. 78 00:06:40,877 --> 00:06:47,367 So as soon as it can be operationalized, you sort of build it out into a department. 79 00:06:47,367 --> 00:06:50,093 And that's that's kind of the mindset. 80 00:06:50,093 --> 00:06:53,194 is like, nothing is actually AI specific. 81 00:06:53,194 --> 00:06:56,385 I mean, we're talking about, you know, dot com level transformations. 82 00:06:56,385 --> 00:07:02,066 It goes across the entire organization, but you know, how many chief internet officers did, did we see? 83 00:07:02,066 --> 00:07:12,009 Like if you sort of do a comparison about how organizations transformed, uh, and then let's be honest, there's some of them that realized, you know, we're, so much in the real 84 00:07:12,009 --> 00:07:18,953 world that there's some dot com transformation, but as opposed to say a retailer that's, you know, moving most of their business online. 85 00:07:18,953 --> 00:07:19,715 So. 86 00:07:19,715 --> 00:07:21,306 Long way of answering. 87 00:07:21,566 --> 00:07:26,949 I don't think we're going to see that many pure chief AI officer. 88 00:07:27,690 --> 00:07:37,075 I think the folks that really know what they're doing oftentimes will have a chief data officer who touches almost no technology, which is, uh, I've had several conversations 89 00:07:37,075 --> 00:07:40,978 with CDOs over the years and their perception has not changed. 90 00:07:40,978 --> 00:07:47,201 And I've got, I got a lot of respect for CDOs, because they are just the absolute 91 00:07:47,287 --> 00:07:54,979 Wranglers ambassadors translators oftentimes in their organization because their organizations have data. 92 00:07:54,979 --> 00:08:01,451 It can be applied to AI processes, but you've got to extract it, transform it, get it loaded somewhere else. 93 00:08:01,451 --> 00:08:09,013 And that's, that's where the CDO CDO comes in is kind of, you know, intermeshing the data and the processes of an organization. 94 00:08:09,013 --> 00:08:11,347 So you can call it, call it AI. 95 00:08:11,347 --> 00:08:16,417 Um, but I think titles mean a lot and, uh, a lot of. 96 00:08:16,417 --> 00:08:23,464 Orgs won't necessarily take that as meaning, know, our department doesn't have much to do with AI go away, right? 97 00:08:23,464 --> 00:08:25,156 Like there's always that danger. 98 00:08:25,156 --> 00:08:27,969 So that's my read on the world. 99 00:08:28,118 --> 00:08:28,598 Yeah. 100 00:08:28,598 --> 00:08:37,085 Well, you know, your point about this not being, there not being a lot of substance, legal has already demonstrated this with innovation, right? 101 00:08:37,085 --> 00:08:49,075 So the chief innovation officer, chief knowledge and innovation officer, um, so that they, I did a, a quick survey back. 102 00:08:49,075 --> 00:08:53,519 Um, we've been around in legal for a long time and I'm a hoarder of data. 103 00:08:53,519 --> 00:08:57,826 So I had Ilta rosters going all the way back to like 2010. 104 00:08:57,902 --> 00:09:06,462 Um, I took the 2014 Ilta roster since that was 10 years ago, looked at how many, uh, how many titles had the word innovation in it? 105 00:09:06,462 --> 00:09:07,502 16. 106 00:09:07,502 --> 00:09:07,822 Okay. 107 00:09:07,822 --> 00:09:16,362 So legal did not exist materially in, in legal, in law firms, uh, 10 years ago, there were, there were zero C-suite. 108 00:09:16,362 --> 00:09:22,742 There were just one or two directors and some managers, analysts, so on and so forth. 109 00:09:22,782 --> 00:09:27,074 So in the 10 years since 2014, 110 00:09:27,074 --> 00:09:28,809 There are now over 300. 111 00:09:28,809 --> 00:09:36,866 It was almost a 2000 % increase and has the legal innovation work increased by 2000 %? 112 00:09:36,866 --> 00:09:38,359 Absolutely not. 113 00:09:38,460 --> 00:09:50,205 There are some, there is some really good innovation work happening in the legal space, but I can tell you that just from, you know, I have deep relationships with a lot of folks 114 00:09:50,205 --> 00:09:51,726 in KM who 115 00:09:51,822 --> 00:09:57,592 suddenly I saw an eye appear in their name and just asked like, how did your, how has your job changed? 116 00:09:57,592 --> 00:09:59,962 And the answer was not much. 117 00:09:59,962 --> 00:10:08,774 So, um, you know, the concept of innovation theater is real and, um, we may run into that same thing with, with AI. 118 00:10:10,073 --> 00:10:10,963 Agreed, agreed. 119 00:10:10,963 --> 00:10:15,333 It's interesting you say 10 years because I was really digging into this subject five years ago. 120 00:10:15,333 --> 00:10:22,413 Don't have any hard data to present, but 2019 and then at the beginning of COVID early 2020, I had the opportunity. 121 00:10:22,413 --> 00:10:24,423 mean, everybody was locked in their house, right? 122 00:10:24,423 --> 00:10:25,993 First couple of months. 123 00:10:26,013 --> 00:10:28,103 Anybody and everybody would take a call. 124 00:10:28,103 --> 00:10:36,431 It was a truly like extraordinary singular time, I think, in the world of business and professions was everybody was just trying to figure out what to do and 125 00:10:36,431 --> 00:10:38,702 was very sort of closed off from the world. 126 00:10:38,702 --> 00:10:49,605 So, I was relatively early in kind of my exposure to the world of very high end legal, but you could get on the phone with just about anybody, which was cool. 127 00:10:49,906 --> 00:10:59,997 And the feedback that I got was it felt like we were at the tail end of the first push for innovation where everyone I talked to was like, yeah. 128 00:11:00,015 --> 00:11:01,576 Innovation got really important. 129 00:11:01,576 --> 00:11:08,410 So everybody put that word in somewhere, but then you have the conversations and then the feedback is basically mirrors what, what you had. 130 00:11:08,410 --> 00:11:16,073 The interesting thing is, kind of the most recent five years, uh, it's been sort of that, that curve, that wave has started to grow again. 131 00:11:16,073 --> 00:11:18,175 There's, there's absolutely been more hires. 132 00:11:18,175 --> 00:11:19,796 There's been more work. 133 00:11:19,796 --> 00:11:29,957 Um, but I think it's very, very similar innovation and AI sit a hundred people down and have them define both terms and you'll have. 134 00:11:29,957 --> 00:11:38,589 you know, 400 terms, different definitions, because it's such an amorphous name, role, topic. 135 00:11:39,051 --> 00:11:40,834 People don't know what they're supposed to do. 136 00:11:40,834 --> 00:11:43,828 Nobody else can say clearly what it is that they do do. 137 00:11:43,828 --> 00:11:46,251 So it's it's it's very much the same thing. 138 00:11:46,328 --> 00:11:46,998 Yeah. 139 00:11:46,998 --> 00:11:50,941 mean, innovation exists in other industries. 140 00:11:50,941 --> 00:11:53,623 It's usually called something different. 141 00:11:53,783 --> 00:12:01,829 I spent 10 years at Bank of America and worked in a business transformation group, which was essentially innovation. 142 00:12:01,829 --> 00:12:14,882 I think that the reason the legal profession has latched on to the word is because clients have commented publicly of just how un-innovative law firms have been. 143 00:12:14,882 --> 00:12:20,574 You know, they still use, I mean, you know, when I started in this actually five years after I started. 144 00:12:20,574 --> 00:12:22,565 So this is 2013. 145 00:12:22,565 --> 00:12:27,447 I remember doing a series of road shows through Ilta that law firms would host us. 146 00:12:27,447 --> 00:12:31,089 So, I mean, I did like 24 in one year. 147 00:12:31,089 --> 00:12:41,613 So I was all over the place, ran myself into the ground, but, um, I saw the inside of 24 different law firms and I saw typewriters. 148 00:12:41,613 --> 00:12:43,246 This is 2013. 149 00:12:43,246 --> 00:12:45,367 2014 timeframe. 150 00:12:45,387 --> 00:12:55,151 So, you know, we've all heard, you know, the billable hour won't die and, um, you know, value-based billing. 151 00:12:55,151 --> 00:12:56,181 Yeah, it's a thing. 152 00:12:56,181 --> 00:13:01,833 It's like, you know, as Richard Susskind says, you know, tell room full of millionaires, they're doing it wrong. 153 00:13:02,794 --> 00:13:03,305 You know, 154 00:13:03,305 --> 00:13:05,245 know, yeah, the billable hour is going to die. 155 00:13:05,245 --> 00:13:14,005 That thing where you pay someone $300 an hour and charge $900 an hour for their service, you know, one to one ratio that won't work anymore. 156 00:13:14,005 --> 00:13:19,465 At some point we're telling you, but you know, until then you're making 600 an hour. 157 00:13:19,465 --> 00:13:21,864 So yeah, totally, totally, totally agree. 158 00:13:21,864 --> 00:13:23,805 Yeah, it is. 159 00:13:24,386 --> 00:13:31,853 You know, I think AI has kind of reignited like there's been waves of pressure. 160 00:13:31,853 --> 00:13:36,477 I feel like in legal, the last one was ALS PS, right? 161 00:13:36,477 --> 00:13:44,644 The alternative legal service providers that were going to put pressure on law firms to do value based billing. 162 00:13:44,644 --> 00:13:45,525 And you know what? 163 00:13:45,525 --> 00:13:47,150 ALS PS, I mean, 164 00:13:47,150 --> 00:13:59,233 They're still around, but it wasn't nearly the movement that the industry really thought it was or was writing about in terms of transforming the industry. 165 00:13:59,233 --> 00:14:04,295 seems like clients kind of like it. 166 00:14:04,295 --> 00:14:17,186 They kind of like the billable hour because I think they're partly to blame as to why, why it hasn't hasn't died because if they were demanding AFAs, I think 167 00:14:17,186 --> 00:14:20,156 I think we'd have them, we'd have more of them. 168 00:14:20,782 --> 00:14:32,670 there's an insight that I gained about a month ago from a long established, I believe their role was, yes, CIO, chief innovation officer of a large like AmLaw 100 law firm. 169 00:14:32,670 --> 00:14:38,257 And we're having this discussion about data analysis and deriving 170 00:14:38,257 --> 00:14:40,317 Uh, you know, how do you come to a flat fee? 171 00:14:40,317 --> 00:14:49,757 How do you analyze a case and you matter compared to, mean, especially if you're, serving large corporate clients, you, you, have, you know, similar cases for that same client that 172 00:14:49,757 --> 00:14:51,637 you could run analysis for. 173 00:14:51,637 --> 00:14:56,117 Nobody tracks time better than, you know, law firms, six minute increments. 174 00:14:56,117 --> 00:15:04,461 What was astounding to me I, and I got to admit, you know, I had not heard this before was a very simple statement that said, yes, we did. 175 00:15:04,461 --> 00:15:05,692 Awesome data analytics. 176 00:15:05,692 --> 00:15:13,928 And we got a really good handle, we think on being able to predict how much it matter was going to cost, you know, flat fee, you know, percentage profit on top. 177 00:15:14,009 --> 00:15:20,734 We brought that to more than one client who said, we don't trust you. 178 00:15:20,994 --> 00:15:25,878 We don't trust the math that you've done in the background to come up with that flat fee. 179 00:15:25,878 --> 00:15:34,525 We'd rather still pay you by the hour because we're assuming as always that you're just building in so much. 180 00:15:34,543 --> 00:15:38,356 BS to that figure that we just don't we inherently don't trust it. 181 00:15:38,356 --> 00:15:49,273 And it's just fascinating because, you know, who are you supposed to really trust in this world, right, your accountant, your doctor, your lawyer, and there's sort of that, that 182 00:15:49,273 --> 00:15:55,968 that holiness of the profession, that's oftentimes put forth the whole lawyer, non lawyer argument, we won't get into it. 183 00:15:55,968 --> 00:16:03,025 But if you're going to, you know, elevate a profession to that level of, you know, reliability, 184 00:16:03,025 --> 00:16:12,448 you know, fiduciary duty and trustworthiness to find out that, you know, multimillion dollar customers are just saying, yeah, no, I don't believe it. 185 00:16:12,448 --> 00:16:16,169 And then you really realize, well, who's managing that relationship? 186 00:16:16,169 --> 00:16:21,030 You know, it's oftentimes one partner whose client that truly is. 187 00:16:21,030 --> 00:16:23,171 And that's been another interesting part of the conversation. 188 00:16:23,171 --> 00:16:31,953 And I'll, I'll, I'll end here, but it's, uh, when I've been talking to firms about innovating, sort of going through this entire process. 189 00:16:31,983 --> 00:16:34,844 They say, you know, there's quite a tap dance. 190 00:16:34,844 --> 00:16:40,610 It's just a field of landmines because is the partner going to trust you to talk to the end client? 191 00:16:40,610 --> 00:16:48,119 I that's like one of the biggest, biggest issues because it's their relationship they're going to want to take it with if they do a lateral transfer. 192 00:16:48,119 --> 00:16:52,452 So trust is, is, a surprisingly large issue. 193 00:16:52,452 --> 00:16:58,295 And I got to say that I didn't fully see that coming with his, I'll say as little time in this field as I've had. 194 00:16:58,478 --> 00:16:58,918 Yeah. 195 00:16:58,918 --> 00:17:02,858 You know, and it's not to say that value-based billing doesn't exist. 196 00:17:02,858 --> 00:17:03,458 It does. 197 00:17:03,458 --> 00:17:13,208 I've, I've actually experienced it when I 15 years ago, when I got married, uh, we did a will in a state that was a flat fee. 198 00:17:13,208 --> 00:17:18,638 Um, when I, when I created a, uh, revocable trust, that was a flat, flat fee. 199 00:17:18,638 --> 00:17:24,248 When I did a trademark registration, that was a flat fee, but it's still, it's still on the fringes. 200 00:17:24,248 --> 00:17:28,110 The, the meat and potatoes of legal work is still done. 201 00:17:28,110 --> 00:17:29,191 hourly. 202 00:17:29,191 --> 00:17:44,281 So, you know, I don't know how much AI is going to influence that, but it does feel like it has reignited the conversation because people are now thinking how, okay, as a client, 203 00:17:44,281 --> 00:17:50,726 I want my law firm to be efficient, but you know, and I do want to reduce my spend. 204 00:17:51,607 --> 00:17:56,310 They have personnel and margins and overhead and things that they have to 205 00:17:56,416 --> 00:17:57,587 accommodate for right? 206 00:17:57,587 --> 00:18:00,892 make a huge investment in a Harvey implementation, for example, right? 207 00:18:00,892 --> 00:18:02,213 That's not free. 208 00:18:02,213 --> 00:18:06,999 So I can't give you the time for free and just bill you less. 209 00:18:06,999 --> 00:18:09,042 Somehow there has to be a strategy in there. 210 00:18:09,042 --> 00:18:15,870 So I don't know if you share that perspective that it seems like we're talking about it again more. 211 00:18:16,539 --> 00:18:25,915 We certainly are and there's been similarities drawn between the US auto industry of the 20th century. 212 00:18:25,996 --> 00:18:29,328 We've got a lot of large law firms, mid-size, small. 213 00:18:29,328 --> 00:18:33,380 Same thing was true with, you can't buy a Hudson automobile anymore. 214 00:18:33,461 --> 00:18:36,383 It's only a small number, very, very large. 215 00:18:36,383 --> 00:18:39,225 There are now, of course, new electric upstarts. 216 00:18:39,823 --> 00:18:48,341 But I truly do believe that a lot of the sort of low value work is going to be so thoroughly commoditized. 217 00:18:48,341 --> 00:18:56,057 It's like any commodity, who would have thought that computers would have been a commodity in the 90s, right? 218 00:18:56,057 --> 00:18:58,200 They were all these specialized machines, super expensive. 219 00:18:58,200 --> 00:19:02,875 And now it's just like, whatever, go to the nearest store and buy a laptop and you'll be fine. 220 00:19:02,875 --> 00:19:04,916 So I think that... 221 00:19:05,329 --> 00:19:16,943 You know, with the example of General Electric from what it was, the eighties of bringing legal in-house, a lot of these technologies are going to be implemented by the clients and 222 00:19:16,943 --> 00:19:28,945 a lot of the business that especially large corps put towards their law firms is going to be very geographically specific and very matter specific and be for the really, really 223 00:19:28,945 --> 00:19:33,208 complex matters because this all ties into the conversation that we just had. 224 00:19:33,208 --> 00:19:35,107 If you bring in, you know, I've 225 00:19:35,107 --> 00:19:37,640 I founded the Society of Legal Engineers here in the US. 226 00:19:37,640 --> 00:19:48,091 Like if you bring in some legal engineers and some business experts and you start building systems inside a corporation that can afford to build them, you can automate, streamline a 227 00:19:48,091 --> 00:19:53,396 heck of a lot of legal processes as long as your lawyers inside the order are willing to do it. 228 00:19:53,397 --> 00:19:57,371 And you'll see significant savings and speed over time. 229 00:19:57,371 --> 00:20:07,768 Yeah, you know, and I know there's a lot of fear, uncertainty and doubt, and especially at law firm leadership level, you know, when like the Goldman report came out and, you know, 230 00:20:07,768 --> 00:20:15,923 I am, I am confident that there will not be one legal job net loss as a result of AI. 231 00:20:15,923 --> 00:20:25,730 I think there will be plenty that are impacted and there will be shifts, dramatic shifts in where, how legal work is done. 232 00:20:25,804 --> 00:20:29,236 But you know, a good parallel is the accounting industry. 233 00:20:29,236 --> 00:20:34,140 So when, when spreadsheets came out 1983, I went back and looked this up. 234 00:20:34,140 --> 00:20:37,782 I was going to write a post on it. 235 00:20:37,782 --> 00:20:39,653 1983 Lotus one, two, three came out. 236 00:20:39,653 --> 00:20:42,986 There were about 850,000 CPAs. 237 00:20:42,986 --> 00:20:52,372 I don't know if you remember, I was a kid, but I was really into computers and I happened to remember my mom owned a restaurant and she had me doing compute. 238 00:20:52,372 --> 00:20:55,874 Like I was in the back writing basic programs. 239 00:20:56,078 --> 00:21:06,977 to, I got a list of, prospects mailing list on a floppy disc, loaded it in and then created mailing labels so she could mail out menus. 240 00:21:06,977 --> 00:21:19,677 Like, you I was that kid in 10 years old with a TI 99 four a, but I remember she had an accountant, she had a CPA who was taught, who was asking me 10 year old, if I knew how to 241 00:21:19,677 --> 00:21:25,472 use the spreadsheets that were coming out and I didn't have a clue, but they were worried and 242 00:21:25,624 --> 00:21:29,345 There are now 1.5 million in the accounting industry. 243 00:21:29,345 --> 00:21:33,076 Again, these are rough numbers, but I just looked them up maybe two weeks ago. 244 00:21:33,076 --> 00:21:35,477 So the industry is more than doubled. 245 00:21:35,477 --> 00:21:48,580 And if you think about impact to an industry, it's not that different in how spreadsheets think about all of the work that accountants used to do by hand that is now done 246 00:21:48,580 --> 00:21:53,822 projections, pro formas, you know, even just basic P and L's. 247 00:21:53,822 --> 00:21:54,648 So 248 00:21:54,648 --> 00:21:56,733 I don't think there's going to be any job loss. 249 00:21:56,733 --> 00:21:59,811 think there's going to be a ton of impact, but no loss. 250 00:21:59,811 --> 00:22:00,102 I don't know. 251 00:22:00,102 --> 00:22:01,254 What's your perspective? 252 00:22:01,881 --> 00:22:11,810 I think that I can't find a good comparison for an industry that is as underserved as the legal services industry. 253 00:22:11,810 --> 00:22:25,432 So I don't really have any worry except for if you're hiding something, which a lot of, I'm going to be honest, professionals in the legal industry tend to do, which is, you 254 00:22:25,432 --> 00:22:29,711 know, hiding the fact that they don't know about a particular matter or are. 255 00:22:29,711 --> 00:22:41,359 Recycling content and billing, you know multiple hours for drafting a document that they just change a couple terms on stuff like that's gonna go away, but there are so many unmet 256 00:22:41,359 --> 00:22:53,778 legal needs and regulations Regardless of who's running the country or what direction the government goes There's always a very large and complex regulatory environment and you 257 00:22:53,778 --> 00:22:58,423 know, nobody here in the US can really affect You know GDPR 258 00:22:58,423 --> 00:23:00,484 or the new EU AI act. 259 00:23:00,484 --> 00:23:05,856 we're going to have to comply with regulations all over the world. 260 00:23:06,516 --> 00:23:09,617 there'd be no sort of paucity of work. 261 00:23:09,617 --> 00:23:16,710 It's just that people are going to need to change and learn and grow to adapt to the new environment. 262 00:23:16,710 --> 00:23:19,841 And if you don't like change, then yeah, you're scared right 263 00:23:20,034 --> 00:23:20,714 Yeah. 264 00:23:20,714 --> 00:23:31,429 So, know, econ 101 or 102 macroeconomics supply and demand, they meet at the price of equilibrium. 265 00:23:31,429 --> 00:23:44,124 If you look at the hourly rates of legal professionals relative to comparably trained professionals in other fields, legal enjoys a premium. 266 00:23:44,224 --> 00:23:48,856 And it's for the exact reason that you're talking, there is way more demand 267 00:23:48,856 --> 00:23:50,668 than there is supply. 268 00:23:50,668 --> 00:23:53,750 There's a ton of unmet needs out there, even in the business world. 269 00:23:53,750 --> 00:24:10,454 For example, I had the CKO from Wilson Sonsini on the podcast and they have a platform called Neuron where, yeah, yeah, for startups where, you know, they'll do automated tasks 270 00:24:10,454 --> 00:24:12,586 like offer letters, right? 271 00:24:12,586 --> 00:24:17,664 Like how great would it be to, but who can afford to have an attorney draft your 272 00:24:17,664 --> 00:24:28,571 Offer letters, but if it's automated and you know and they're checking all the boxes and like you know there there is real risk even to like you go putting the wrong like go make 273 00:24:28,571 --> 00:24:41,330 an offer letter to a 1099 you've now put yourself in a potential co-employment risk scenario right like there is all sorts of even for for entities that can afford legal 274 00:24:41,330 --> 00:24:42,230 work. 275 00:24:42,242 --> 00:24:44,104 let alone all the ones that can't. 276 00:24:44,104 --> 00:24:45,286 So I agree with you. 277 00:24:45,286 --> 00:24:52,684 think there's huge unmet needs out there and we'll, there will be shifts, but I don't think there will be any net loss. 278 00:24:52,783 --> 00:24:54,354 Yeah, I feel obliged. 279 00:24:54,354 --> 00:25:02,777 I'm going to quote my mentor and professor from law school, Bill Henderson, who's now working in Indiana for, you know, access to justice issues. 280 00:25:03,097 --> 00:25:08,063 But, you know, he's worked deeply in the, the business of law as it were. 281 00:25:08,063 --> 00:25:14,982 It doesn't really publish anymore on his blog, legalevolution.org, but definitely a shout out for tons of great content. 282 00:25:14,982 --> 00:25:22,353 But his quote is, how else can an English major with three additional years of education make a million dollars a year? 283 00:25:22,353 --> 00:25:26,133 That's kind of like, it's a lot of words, but it's a question that's posed. 284 00:25:26,133 --> 00:25:29,093 That's like, it's not that much more school. 285 00:25:29,093 --> 00:25:29,633 Yeah. 286 00:25:29,633 --> 00:25:32,663 You have to, you know, work for a firm as an associate. 287 00:25:32,663 --> 00:25:37,853 Uh, but you know, I, I went to a T 14 law school, you know, I didn't get a JD from there. 288 00:25:37,853 --> 00:25:43,473 So I'm not eligible for the, at the time, $235,000 a year. 289 00:25:43,713 --> 00:25:48,413 Initial salary that a, that a first year associate was going to make at big law. 290 00:25:48,413 --> 00:25:50,991 Uh, but it just goes to show like. 291 00:25:50,991 --> 00:25:58,970 you power through and you get through that sort of that gauntlet of, of, of, know, winnowing down of, of who gets to play and who doesn't. 292 00:25:58,970 --> 00:26:01,323 And the rewards are still quite high. 293 00:26:01,368 --> 00:26:03,039 Yeah, absolutely. 294 00:26:03,039 --> 00:26:14,378 Well, you and I talked in our last conversation about AI solutions and you know, one of I shared my opinion on Copilot, which is is not a good one really today. 295 00:26:14,378 --> 00:26:20,021 I have full confidence in Microsoft's ability to get this right in the long term, but they have rushed this out. 296 00:26:20,021 --> 00:26:27,236 I mean, it's you know, I mean, and there's just this sense of urgency to put Copilot everywhere it's on. 297 00:26:27,236 --> 00:26:28,256 even putting it on. 298 00:26:28,256 --> 00:26:28,898 They read it. 299 00:26:28,898 --> 00:26:31,022 There hasn't been a keyboard change in like. 300 00:26:31,022 --> 00:26:36,645 20 years and they've there's now going to be a Copilot key on new keyboards. 301 00:26:37,146 --> 00:26:41,028 So I think they've gotten a little bit over their skis on this. 302 00:26:41,028 --> 00:26:44,199 Like I have not had a good experience and I've talked about it. 303 00:26:44,199 --> 00:26:48,392 My listeners are probably tired of hearing me complain about it. 304 00:26:48,392 --> 00:26:52,874 But you know the more you know in the beginning I felt like they're Microsoft. 305 00:26:52,874 --> 00:26:56,597 One thing Microsoft is really good at is rewarding evangelists. 306 00:26:56,597 --> 00:26:59,668 So they have this MVP program, the Microsoft. 307 00:27:00,180 --> 00:27:00,713 Yeah. 308 00:27:00,713 --> 00:27:10,169 professional and they re and people like evangelize and drink the Kool-Aid and and spout it and they have influence. 309 00:27:10,169 --> 00:27:19,194 And I think that, you know, in the beginning, before people really got their hands on it, they just assumed that it was going to be great and it's not great. 310 00:27:19,194 --> 00:27:21,005 At least not yet from my perspective. 311 00:27:21,005 --> 00:27:22,296 I think there's a lot of limitations. 312 00:27:22,296 --> 00:27:22,657 I don't know. 313 00:27:22,657 --> 00:27:25,458 Where do you stand on current state of Copilot? 314 00:27:25,949 --> 00:27:32,111 Having worked intimately with knowledge professionals that work very heavily in documents. 315 00:27:32,451 --> 00:27:41,293 I can tell you that automating and creating magic for people whose bread and butter is what you're messing with. 316 00:27:41,373 --> 00:27:42,513 It's not. 317 00:27:43,014 --> 00:27:44,344 It's not difficult. 318 00:27:44,344 --> 00:27:50,046 It's not extremely difficult, but it is much more specific than most people would like to imagine that it is. 319 00:27:50,046 --> 00:27:55,177 And it's much more time consuming and a grind than most people would like to accept. 320 00:27:55,403 --> 00:28:07,836 Um, so, uh, I keep on quoting myself and others, but, uh, another thing that I often say these days is, you know, one of the, the, the worst tricks, the ChatGPT GPT, you know, 321 00:28:07,836 --> 00:28:17,019 pulled on the world was, you know, it convinced that AI was, was good at everything because it was just such, you know, I've seen so many lectures talking about, know, the 322 00:28:17,019 --> 00:28:21,490 most rapidly adopted platform in history, ChatGPT GPT, and, know, why is that? 323 00:28:21,490 --> 00:28:23,971 And it's like all of the technology already existed. 324 00:28:23,971 --> 00:28:25,263 It just wasn't as fun. 325 00:28:25,263 --> 00:28:25,613 Right. 326 00:28:25,613 --> 00:28:33,335 So getting back to Copilot, it's just not that it's not that much fun because how often are you really pleasantly surprised? 327 00:28:33,335 --> 00:28:34,116 Right. 328 00:28:34,116 --> 00:28:48,129 And I think you really have to spend time with your end users, understanding specifically what it is exactly that they do beginning to end and really focus on where can I make 329 00:28:48,129 --> 00:28:48,730 magic? 330 00:28:48,730 --> 00:28:53,689 And you know, the whole conversation about like, you know, do you, are you making, you know, and 331 00:28:53,689 --> 00:29:01,134 an aspirin or a vitamin, you, helping people do something that's sort of tangential or are you really solving a huge pain point? 332 00:29:01,134 --> 00:29:03,856 And, um, I think formatting is a big deal. 333 00:29:03,856 --> 00:29:08,409 haven't really seen that much like actual, I mean, let's call it like voice activated. 334 00:29:08,409 --> 00:29:11,555 mean, co-pilot is like, I want to be able to communicate with it. 335 00:29:11,555 --> 00:29:15,384 I want to be able to tell it like, fix this thing about my document. 336 00:29:15,384 --> 00:29:16,245 Right. 337 00:29:16,245 --> 00:29:22,329 Um, and the, the software company I was at, Xmentium, one of, one of the things about automatically building documents. 338 00:29:22,477 --> 00:29:26,679 was that the numbering was always rock solid, the outline numbering. 339 00:29:26,679 --> 00:29:33,132 And I don't think anyone suffers more from outline shenanigans than lawyers, right? 340 00:29:33,132 --> 00:29:41,106 The number of times I've heard folks say, I would be doing a pitch, I'd be saying, hey, yeah, this platform does this and it automatically builds your document and the outline, 341 00:29:41,106 --> 00:29:45,008 no matter what you do, no matter what you change is going to be completely accurate. 342 00:29:45,008 --> 00:29:47,409 Roman numeral ads will be in the right place. 343 00:29:47,601 --> 00:29:54,865 And people will say, you know, I was crying over my tab key last night because I was trying to like manually build. 344 00:29:54,865 --> 00:30:01,738 just gave up on styles, gave up on the outline control and was, you know, had to go through and then, you know, last minute somebody wants to make a change. 345 00:30:01,738 --> 00:30:07,891 that's, that's where I think we're going to start to see making people truly happy, but we're not there yet. 346 00:30:07,891 --> 00:30:13,173 And it's a system that's trying to do magic and everything and not really being that magic. 347 00:30:13,518 --> 00:30:13,898 Yeah. 348 00:30:13,898 --> 00:30:16,838 Speaking of the formatting issue, that's one of the areas. 349 00:30:16,838 --> 00:30:25,448 So I, I, I've tried so many times to, use Copilot Um, I paid for a year's license of it at $30 a month. 350 00:30:25,448 --> 00:30:27,198 wanted to get value. 351 00:30:27,198 --> 00:30:28,978 So I have tried. 352 00:30:29,158 --> 00:30:38,458 One of the things I try is like when I go to ChatGPT GPT and I get back a bulleted list and I've tried multiple times copying, pasting it into a word document, the bullets don't 353 00:30:38,458 --> 00:30:40,688 come over like in normal rich texts. 354 00:30:40,688 --> 00:30:41,874 So I prompted. 355 00:30:41,874 --> 00:30:43,656 But it's clear what they are, right? 356 00:30:43,656 --> 00:30:44,337 They're tabbed. 357 00:30:44,337 --> 00:30:48,661 There's a little emoji, you know, dot. 358 00:30:48,661 --> 00:30:55,088 I've asked Copilot to clean it up, put, these real bullets, and it spit out the instructions to do it manually. 359 00:30:55,088 --> 00:30:56,389 And I'm like, man. 360 00:30:56,643 --> 00:31:05,462 Or even better, let me ask you this question, given the tight relationship and interoperability, why are you using ChatGPT GPT and not just using Copilot? 361 00:31:05,462 --> 00:31:12,638 Yeah, exactly, because I have had too many bumps on my head trying to use Copilot. 362 00:31:13,070 --> 00:31:14,111 Yep. 363 00:31:14,111 --> 00:31:14,832 Plain and simple. 364 00:31:14,832 --> 00:31:18,374 It's nerfed, guard railed, whatever term you want to use. 365 00:31:18,614 --> 00:31:21,426 It's insane how many times I have done the exact same thing. 366 00:31:21,426 --> 00:31:25,214 And I think that's, I'm not here to trash talk anybody's, hard work. 367 00:31:25,214 --> 00:31:26,400 mean, co-pilot. 368 00:31:26,400 --> 00:31:28,431 It's amazing to exist. 369 00:31:28,431 --> 00:31:32,224 It's a lot of work, but it's not doing what people want it to do. 370 00:31:32,224 --> 00:31:35,386 And you you know, there are certain things that it refuses to do. 371 00:31:35,386 --> 00:31:36,277 You're like, forget it. 372 00:31:36,277 --> 00:31:37,738 I'm just going to ChatGPT GPT. 373 00:31:37,738 --> 00:31:39,489 Cause I know what's going to go for me. 374 00:31:39,583 --> 00:31:40,203 exactly. 375 00:31:40,203 --> 00:31:40,487 Yeah. 376 00:31:40,487 --> 00:31:46,006 had a, actually that Ilta, that Ilta roster study that I was telling you about. 377 00:31:46,006 --> 00:31:51,459 I, I put, um, all of the titles, the job titles with innovation. 378 00:31:51,459 --> 00:32:02,415 I cat, I bucketed them into, um, C-suite direct, senior director, director, attorney manager, analyst, and I gave it examples. 379 00:32:02,415 --> 00:32:08,010 And then I uploaded the spreadsheet and told, or now I'm sorry in 380 00:32:08,010 --> 00:32:14,133 Excel, the in-app Copilot, asked it to mimic what I did completely choked. 381 00:32:14,133 --> 00:32:20,315 So I took the exact same prompt, paste it into ChatGPT GPT, uploaded the spreadsheet and it did it. 382 00:32:20,315 --> 00:32:22,116 Now it didn't do it perfectly. 383 00:32:22,116 --> 00:32:27,978 There were a ton of other, like it bucketed way too many in other. 384 00:32:28,599 --> 00:32:31,430 and like it made some silly mistakes. 385 00:32:31,430 --> 00:32:37,430 Like, uh, there were titles with the word chief in it that it didn't 386 00:32:37,430 --> 00:32:49,629 recognized as C suite and I prompted it like I I kind of scolded it like why did you not put titles with chief in the C suite and he was like, oh sorry, I should have done that 387 00:32:49,629 --> 00:32:50,699 and then it would fix it. 388 00:32:50,699 --> 00:32:56,023 But yeah, I mean, uh, ChatGPT GPT and Claude aren't perfect either. 389 00:32:56,023 --> 00:33:06,210 But what, is your perspective on like purpose built, you know, solutions for legal versus the, general J uh, gen AI. 390 00:33:06,398 --> 00:33:08,137 applications that are out there. 391 00:33:09,377 --> 00:33:22,186 I some of them work, but referring to recent conversations I've had, you can't even think about it honestly as legal work. 392 00:33:22,186 --> 00:33:27,660 There's just so many different tech stacks that so many different types of lawyers use. 393 00:33:27,660 --> 00:33:37,837 think, I'll say we, sort of like greater legal technology ecosystem really has to start talking about like, what's the practice area? 394 00:33:38,226 --> 00:33:42,928 that you're applying your technology to, not just like, it's for lawyers, right? 395 00:33:43,149 --> 00:33:45,870 Because it doesn't work right. 396 00:33:46,071 --> 00:33:49,193 There's an early example that I cited. 397 00:33:49,193 --> 00:34:00,320 mean, it was fascinating to have a conversation with someone who was an established patent professional in a reasonably sized IP boutique. 398 00:34:00,341 --> 00:34:06,577 Him and his team filing 300 or more patents per year, and they were working with a platform that would take 399 00:34:06,577 --> 00:34:17,297 Uh, the initial inventors disclosure, their description of what, you know, the novel idea was in their invention and turn it into a set of claims, which is, know, the very core 400 00:34:17,297 --> 00:34:20,257 meat of, a very large patent application. 401 00:34:20,257 --> 00:34:31,617 I want a professional that's doing something 300 or more times per year is telling you that a platform like out of the gate is saving them 50 % of like the time spent on each 402 00:34:31,617 --> 00:34:32,297 project. 403 00:34:32,297 --> 00:34:33,697 That's that's enormous. 404 00:34:33,697 --> 00:34:34,991 And just to 405 00:34:34,991 --> 00:34:42,267 circle back to our earlier conversation, patent prosecution, patent drafting, very much a place where you can flat-fee quote upfront. 406 00:34:42,267 --> 00:34:44,338 Cause you know, what are the unknowns? 407 00:34:44,338 --> 00:34:49,092 I see your technology, I will draft, we will push it through the USPTO as best we can. 408 00:34:49,092 --> 00:34:52,475 So that's a specific use case. 409 00:34:52,635 --> 00:35:03,714 And you know, one quick other one that I looked at and I've had, I've spoken to local governments and law enforcement about is taking police body cam footage, the transcript 410 00:35:03,714 --> 00:35:05,145 just from the audio. 411 00:35:05,149 --> 00:35:08,791 and running that through a platform and generating a police report. 412 00:35:08,791 --> 00:35:11,733 And we're talking at least 50 % time savings. 413 00:35:11,733 --> 00:35:16,365 So this is kind of my view on purpose built versus general. 414 00:35:16,365 --> 00:35:19,777 There's things that are supposed to help people in business. 415 00:35:19,777 --> 00:35:24,079 There's, know, gen AI and technical tools that are supposed to help people in legal. 416 00:35:24,079 --> 00:35:34,767 And then there's like somewhat boring, but very, very specific purpose built tools where the model has been detuned, hallucinations trained correctly. 417 00:35:34,767 --> 00:35:38,622 and weighted properly that it like does magic. 418 00:35:38,622 --> 00:35:41,015 And that's really what you're expecting the AI to do. 419 00:35:41,388 --> 00:35:41,708 Yeah. 420 00:35:41,708 --> 00:35:48,101 You and I talked a little bit about small language models and ones that, that you can run locally. 421 00:35:48,101 --> 00:35:55,364 mean, you know, Meta puts out llama, which you know, you can download and run locally. 422 00:35:55,364 --> 00:36:01,407 I haven't done it, but I think I've downloaded it, but really just kind of played around. 423 00:36:01,407 --> 00:36:08,770 I haven't done anything serious with it, but I think, I think you were talking about how you could 424 00:36:09,122 --> 00:36:21,166 leverage a small language model for something that is really sensitive, like a patent application or transactional documents and work on them locally, blow the thing away and 425 00:36:21,166 --> 00:36:28,384 then re-download it to make sure that there's no overlap in customer data. 426 00:36:28,384 --> 00:36:30,826 mean, is that, people really doing that? 427 00:36:31,561 --> 00:36:34,003 That's the exact story that I got. 428 00:36:34,003 --> 00:36:41,169 So for full clarity, the police body cam app that was a private instance run on the cloud, I believe. 429 00:36:41,850 --> 00:36:52,699 But specifically the patent drafting application, the claim drafting was a local downloaded model that didn't learn or remember anything. 430 00:36:52,699 --> 00:36:59,717 So it was interesting talking to the creators of this platform and they're saying, now there's absolutely no way that if you're working on a 431 00:36:59,717 --> 00:37:08,101 patent for Google and then you need to work on a patent for Apple, it is totally impossible for there to be any sort of cross-infection. 432 00:37:08,101 --> 00:37:18,425 But because it's a downloadable model and we only updated every whatever it is quarter, delete it like you said, redownload it and then use it for your Apple work. 433 00:37:18,425 --> 00:37:21,486 And then you are like so air gapped, it's not even funny. 434 00:37:21,486 --> 00:37:27,649 But overall, local models, I think we spoke about sort of... 435 00:37:28,141 --> 00:37:29,772 It's not American exceptionalism. 436 00:37:29,772 --> 00:37:39,390 It's a passion for what I'll say is avoiding the true sort of energy costs or implications of, you know, what it is that we do. 437 00:37:39,511 --> 00:37:40,181 Giant homes. 438 00:37:40,181 --> 00:37:42,913 I used to work in heating and cooling, high efficiency analysis. 439 00:37:42,913 --> 00:37:47,297 You know, it's just like three air conditioners on a house, gigantic cars. 440 00:37:47,297 --> 00:37:56,284 And I'm not going to go on a bender about save the trees, but I'm just saying that, you know, our nation is so incredibly wealthy that we have the absolute privilege of 441 00:37:57,049 --> 00:38:00,530 almost completely ignoring how much energy we consume to do something. 442 00:38:00,530 --> 00:38:04,251 And I really think that's taken hold in AI. 443 00:38:04,251 --> 00:38:19,355 again, not going into all of the environmental hazards of AI, but like the real fallback to that is we're not building the tools that we could that have way better privacy 444 00:38:19,355 --> 00:38:25,367 controls and way better security because we're just saying, well, there's going to be a new model. 445 00:38:25,367 --> 00:38:27,137 Claude puts out something new. 446 00:38:27,365 --> 00:38:31,376 Uh, you know, open AI makes their adjustments and it's just going to be awesome. 447 00:38:31,376 --> 00:38:39,909 Like at what other time in industry were we, were we so passionate about using like the brand new, most new thing. 448 00:38:39,909 --> 00:38:44,910 It's like a quick comparison is, you know, I'll go shopping for a car. 449 00:38:44,910 --> 00:38:47,561 They do the sort of model refreshes, right? 450 00:38:47,561 --> 00:38:51,302 You don't get the first year of the new model refresh. 451 00:38:51,302 --> 00:38:55,781 You get sort of the, let the company learn how they're building that exact. 452 00:38:55,781 --> 00:39:01,803 form and feature set and then buy in, you know, before they sort of switch over again. 453 00:39:01,803 --> 00:39:03,854 And I really think that mindset's gonna come out. 454 00:39:03,854 --> 00:39:19,220 I think we're gonna see a lot more small compact local models running and possibly some steps taken back from a pure SaaS model to the point of like, go out and get this kind of 455 00:39:19,220 --> 00:39:25,777 laptop that has this level or more of processor and it will run the tool that we're gonna give you. 456 00:39:25,777 --> 00:39:32,393 And it's going to do magic for you, but it won't be cloud based because, you know, that's not the way it should work. 457 00:39:32,942 --> 00:39:49,142 Yeah, I'm one of those suckers who, uh, I bought a 2022 Tundra after it was completely overhauled and, all of the 2022s, the non-hybrid 2022s and half of the 2023s have to have 458 00:39:49,142 --> 00:39:51,222 complete engine replacement. 459 00:39:51,562 --> 00:39:52,902 Yeah. 460 00:39:52,902 --> 00:39:54,698 So they're great. 461 00:39:54,698 --> 00:40:00,442 Yeah, you know, and the problem is like you could just you just I could just rack up miles. 462 00:40:00,442 --> 00:40:05,086 They have announced the recall, but they have not established the process yet. 463 00:40:05,086 --> 00:40:08,007 I mean, imagine how it's hundreds of thousands of vehicles. 464 00:40:08,068 --> 00:40:17,995 So, you know, I could just drive mine forever and I know I'm to get a new engine, but you're going to be out of a vehicle for eight to 10 weeks and they give you like 40 bucks 465 00:40:17,995 --> 00:40:18,565 a day. 466 00:40:18,565 --> 00:40:23,228 I'll be driving the clown car and I'm six foot five, 270 like. 467 00:40:24,066 --> 00:40:26,629 You know, um, so I'm not looking forward to it. 468 00:40:26,629 --> 00:40:27,000 You know what? 469 00:40:27,000 --> 00:40:29,193 And I know better because you're right. 470 00:40:29,193 --> 00:40:36,532 You do want to buy it a year or two, uh, after the remake and then, and then lay your money down. 471 00:40:36,546 --> 00:40:43,456 At least I actually go for the post, the midlife refresh where you get a couple of upgrades. 472 00:40:43,456 --> 00:40:44,834 That's my strategy. 473 00:40:44,834 --> 00:40:50,257 Yeah, I wish I would have followed that advice because now I'm stuck with a view. 474 00:40:50,257 --> 00:40:56,951 mean, I, I'm to take a beating because when you look up the VIN in Kelly blue book, it shows the recall. 475 00:40:56,951 --> 00:40:58,822 So anywhere I traded in, they're going to know. 476 00:40:58,822 --> 00:41:08,678 And if I were to sell it to a private seller who wants to get a new engine, you know, maybe, but, um, all right, well switching gears, uh, you and I talked a little bit about 477 00:41:08,678 --> 00:41:12,170 Harvey and, um, Harvey's a really interesting case. 478 00:41:12,170 --> 00:41:12,970 Like, 479 00:41:13,038 --> 00:41:21,968 You know, they have their series B, which was, think like last December, November, something like that. 480 00:41:21,968 --> 00:41:28,678 It was a 700 million ish valuation, 715. 481 00:41:28,918 --> 00:41:37,498 Seven months later, July, 2024, 1.5 billion was the valuation of their series C round. 482 00:41:37,498 --> 00:41:39,756 So they doubled in value in seven months. 483 00:41:39,756 --> 00:41:44,910 Now, what could have taken place in so think about how difficult it is to double enterprise value, right? 484 00:41:44,910 --> 00:41:48,473 Like they didn't double, they didn't even double revenue. 485 00:41:48,754 --> 00:41:52,017 I think they went from 20 to 30 and those are unofficial numbers. 486 00:41:52,017 --> 00:42:03,416 They don't disclose revenue, but you know, so, um, you know, they've had a couple of interesting pivots, you know, they were trying to raise 600 million to buy VLex and then 487 00:42:03,416 --> 00:42:04,808 that fell through. 488 00:42:04,808 --> 00:42:07,758 They've been running in stealth mode and now they're 489 00:42:07,758 --> 00:42:14,238 I think starting to realize that there's been a little bit of, you know, skepticism about the platform. 490 00:42:14,238 --> 00:42:14,978 I don't know. 491 00:42:14,978 --> 00:42:18,342 What's your take on, on all the, these happenings. 492 00:42:18,353 --> 00:42:30,160 Sure, I've been thinking about it a little bit and I don't know if we can call it the boy band model of entrepreneurship and tech startups, but I'm gonna be honest, I think they 493 00:42:30,160 --> 00:42:40,809 did a truly stellar job of reading what was going to look good to investors, to the market. 494 00:42:40,809 --> 00:42:43,347 I really been thinking about this. 495 00:42:43,347 --> 00:42:47,023 I wanna say, if you've seen the way that 496 00:42:47,023 --> 00:42:53,515 Ford Motor Company started assembly line and has grown and it became incredibly valuable, of course, as a company. 497 00:42:54,475 --> 00:43:11,900 Imagine if like it happened again and you saw in a similar industry that a company was sort of situating itself like just like a previous successful sort of like business model. 498 00:43:11,900 --> 00:43:14,160 And I think that's what Harvey did very, very well. 499 00:43:14,160 --> 00:43:16,501 They knew what was going to appeal to the market. 500 00:43:17,361 --> 00:43:27,521 They knew, I think, with decent confidence that they didn't need to have a fully fleshed out magical product that made everyone happy. 501 00:43:27,521 --> 00:43:30,731 They needed to look really good coming out of the gate. 502 00:43:30,731 --> 00:43:33,361 And they did really well on that. 503 00:43:33,461 --> 00:43:46,277 And because we have these, the industrial revolution has enough sort of back history, we can look at it, how the different companies, American Automotive now is only a few. 504 00:43:46,627 --> 00:43:57,672 I think a lot of investors are prognosticating that Harvey is going to win, that they just went above a couple of critical thresholds, know, started off with not that much, but 505 00:43:57,672 --> 00:44:07,266 looked really good doing it and then got enough traction, got enough customers where it's just like, whether they bought Villex or not, they look like they're going to be a winner 506 00:44:07,266 --> 00:44:09,487 in more than one space. 507 00:44:09,487 --> 00:44:14,009 And so you might as well invest early, the earlier, the better. 508 00:44:14,189 --> 00:44:18,151 because they're probably going to win more likely than many others. 509 00:44:19,292 --> 00:44:21,704 And that's where I think a lot of the bets are being laid. 510 00:44:21,704 --> 00:44:26,436 So I think, you know, if you look at our, do the revenues justify the valuation? 511 00:44:26,436 --> 00:44:30,309 You know, no, but it's, you know, human optimism, right? 512 00:44:30,309 --> 00:44:34,761 But also just people really gaming out, like how is this going to work? 513 00:44:34,761 --> 00:44:40,755 You know, how many different tech platforms is a law firm going to be hiring to do X, Y and Z? 514 00:44:40,755 --> 00:44:41,615 And if... 515 00:44:41,709 --> 00:44:43,730 you can get to that critical mass. 516 00:44:43,730 --> 00:44:52,443 You can be Harvey, you know, they didn't buy VLikes this time, but they've definitely signaled to the market their willingness to, you know, work on very large acquisition 517 00:44:52,443 --> 00:44:53,233 deals. 518 00:44:53,233 --> 00:45:02,957 And I think that's very much just a snowball that keeps happening because, you know, they'll be the one platform to rule them all if everything goes their way. 519 00:45:03,010 --> 00:45:03,790 Yeah. 520 00:45:03,790 --> 00:45:04,060 Yeah. 521 00:45:04,060 --> 00:45:05,171 There's real value there. 522 00:45:05,171 --> 00:45:08,762 It's not like these are bored ape NFTs, right? 523 00:45:08,762 --> 00:45:14,315 what you, they're, they are building a real business. 524 00:45:14,315 --> 00:45:19,997 Um, it's just the, um, their, their go-to-market approach is, interesting. 525 00:45:19,997 --> 00:45:30,892 Um, you know, in my experience, having, having relationships and experience in 526 00:45:30,892 --> 00:45:44,529 your go-to-market team in legal tech, not just having lawyers on staff who've practiced in big law, but actually having your business, your go-to-market team understand the dynamics 527 00:45:44,529 --> 00:45:51,313 in legal tech, which is unique, is a really important ingredient. 528 00:45:51,313 --> 00:45:55,095 When I look at their go-to-market team, I don't see a ton of that. 529 00:45:55,475 --> 00:46:00,780 I see a lot of folks, really smart looking folks, but 530 00:46:00,780 --> 00:46:09,346 Yeah, it's been, and again, the stealth mode and then now coming out of it, it's been, it's been interesting as an outsider who's had very little exposure. 531 00:46:09,346 --> 00:46:12,739 I've seen, I've seen, you know, the stuff they published on the web. 532 00:46:12,739 --> 00:46:15,740 I've actually seen a little bit more than that. 533 00:46:16,702 --> 00:46:18,153 You know, it looks interesting. 534 00:46:18,153 --> 00:46:21,645 It looks like, it looks like it's a UI play. 535 00:46:21,885 --> 00:46:28,800 You know, they're really crafting a user experience for legal that's 536 00:46:28,802 --> 00:46:32,016 going to always probably leverage open AI, right? 537 00:46:32,016 --> 00:46:33,667 They're one of their biggest investors. 538 00:46:33,667 --> 00:46:36,340 So, it's interesting. 539 00:46:36,369 --> 00:46:52,569 I'd have to imagine, I feel obliged to point out that the specific strategy is familiar to the private equity, venture capital space in the pros that I've talked to in that region, 540 00:46:52,569 --> 00:47:05,409 the ones that vet and invest in, put a lot of dollars into tech, AI tech, legal tech, see that it is an emerging way to grow a large business in. 541 00:47:05,717 --> 00:47:16,533 You you're you're building the plane as it flies and that's much more acceptable than it used to be because the Technology is happening so quickly and because there are so many 542 00:47:16,533 --> 00:47:19,034 unknowns about what is AI gonna do? 543 00:47:19,034 --> 00:47:20,355 What is it going to do next? 544 00:47:20,355 --> 00:47:29,900 know, they say You know if you've got a market opportunity you realize that it's your TAM is your total addressable market is is big enough and it's not being satisfied or you can 545 00:47:29,900 --> 00:47:35,397 do it cheaper You've got a business opportunity great or you can be yeti coolers and you can convince people 546 00:47:35,397 --> 00:47:42,421 there's this new thing that you never spent a lot of money on, but now we have a new awesome thing, new market, you do that. 547 00:47:43,121 --> 00:47:52,187 I think it would be sort of, it was impossible to ascertain like what was the big successful market play for legal, partly because there's so many different practice areas, 548 00:47:52,187 --> 00:48:03,113 but partly because I think it was especially hard to even come up with that magical thing that you could convince legal professionals to sort of buy into that, you there is no Yeti 549 00:48:03,113 --> 00:48:03,919 cooler. 550 00:48:03,919 --> 00:48:05,219 for lawyers, look, it does this thing. 551 00:48:05,219 --> 00:48:06,640 It's like, don't want to change. 552 00:48:06,640 --> 00:48:07,450 I don't like that. 553 00:48:07,450 --> 00:48:10,731 Or that makes me more efficient and I don't want to the billable model. 554 00:48:10,731 --> 00:48:11,271 Right. 555 00:48:11,271 --> 00:48:19,393 So I've talked to a decent number of big law customers of Harvey who very much have input into the product. 556 00:48:19,393 --> 00:48:23,174 So they are, is still sort of like a development developing mindset. 557 00:48:24,075 --> 00:48:27,035 I would love to hear some of those product conversations. 558 00:48:27,035 --> 00:48:30,416 You know, they said this, they said that, how do we rectify? 559 00:48:30,416 --> 00:48:33,233 It's kind of how I describe what a lot of my job is, is that 560 00:48:33,233 --> 00:48:42,233 sort of Venn diagram process of finding enough to be successful without, you know, making anybody too upset about not satisfying all their needs. 561 00:48:42,274 --> 00:48:43,096 Yeah. 562 00:48:43,096 --> 00:48:45,152 I mean, it's going to be, it's going to be an interesting journey. 563 00:48:45,152 --> 00:48:47,464 I know we're running out of time here. 564 00:48:47,503 --> 00:48:48,861 Yeah, I'm pretty good. 565 00:48:48,861 --> 00:48:49,887 So this is your show. 566 00:48:49,887 --> 00:48:50,430 You're running it. 567 00:48:50,430 --> 00:48:51,926 Happy to do whatever you need. 568 00:48:51,970 --> 00:48:58,992 Yeah, the well, guess just kind of wrapping up one final question. 569 00:48:58,992 --> 00:49:11,295 So what would you say to a firm who is starting to map out their gen AI strategy in terms of where to get started? 570 00:49:11,295 --> 00:49:19,718 Because there's, you know, there's a lot of confusion in that space and there's a lot of different approaches I see being thrown around. 571 00:49:19,718 --> 00:49:21,638 I've always advocated for 572 00:49:21,696 --> 00:49:25,928 start with the lowest risk, highest reward use cases possible. 573 00:49:25,928 --> 00:49:29,689 And a lot of those exist on business, on the business of law side, right? 574 00:49:29,689 --> 00:49:32,901 Finance, marketing, HR, right? 575 00:49:32,901 --> 00:49:34,851 You don't have client data. 576 00:49:35,652 --> 00:49:43,936 know, it's, it's, if something goes wrong, it's not, you're not putting your, your client and your fiduciary obligations at risk. 577 00:49:43,936 --> 00:49:44,386 I don't know. 578 00:49:44,386 --> 00:49:50,228 What would you say to a firm who's starting to think about this and as to where to start? 579 00:49:50,587 --> 00:49:52,739 Sure, well, it's gonna take resources. 580 00:49:52,739 --> 00:49:55,641 So you're gonna have to make that hire. 581 00:49:55,701 --> 00:49:59,197 And I think you can find unicorns are out there. 582 00:49:59,197 --> 00:50:02,727 I provide certain services, consulting and advisory. 583 00:50:02,727 --> 00:50:12,095 know a lot of other folks that do about, you're not gonna hire someone in-house right away, but at least talk to someone who knows this process and understands the psychology 584 00:50:12,095 --> 00:50:14,417 specifically about working with law firms. 585 00:50:14,417 --> 00:50:16,198 That's number one. 586 00:50:16,398 --> 00:50:17,839 Number two is, 587 00:50:20,133 --> 00:50:29,358 Don't ever, don't expect anything whiz bang from any type of new provider, especially if it's an actual firm. 588 00:50:30,179 --> 00:50:39,864 The most disappointing conversations I have with people are like, given sort of your privacy requirements, and this is firms, but also other organizations, given your privacy 589 00:50:39,864 --> 00:50:48,089 requirements, given your budget, given your appetite for risk and innovation, you're going to have to stick with whatever it is. 590 00:50:48,933 --> 00:50:52,234 that your current software vendors are willing to provide you. 591 00:50:52,234 --> 00:50:54,265 Like what new platform should you get? 592 00:50:54,265 --> 00:50:56,965 What's your first step in AI or gen AI? 593 00:50:56,965 --> 00:51:02,187 Like talk to every one of your current relationships and see what it is that they're doing, right? 594 00:51:02,187 --> 00:51:13,060 Because it's just like, you know, I talked to, you know, the CISOs of the world and the IT managers and they're like, there's no way we're gonna keep our employees from going off 595 00:51:13,060 --> 00:51:14,240 the rails. 596 00:51:14,481 --> 00:51:16,229 lot of these platforms are just not. 597 00:51:16,229 --> 00:51:20,830 well established enough to have the proper guard rails, the SOC 2, whatever it might be. 598 00:51:20,830 --> 00:51:23,591 And so that's, that can be disappointing. 599 00:51:24,031 --> 00:51:29,953 But the flip side of that is, you know, go play in co-pilot is not the best advice either. 600 00:51:29,953 --> 00:51:38,415 So I would say starting out is, you know, be intentional about your resources, be small in your aspirations, but also in your, your wins. 601 00:51:38,415 --> 00:51:39,995 Don't shoot for a big win. 602 00:51:39,995 --> 00:51:45,905 And also, you know, talk to your core providers, which in the case of law firms, they've got the huge advantage. 603 00:51:45,905 --> 00:51:56,125 I think of having that small cohort of, you know, Thomson Reuters and all their, you know, that, that, that small cadre of legal information service providers that have been buying 604 00:51:56,125 --> 00:52:04,605 these other platforms and have been connecting them and have the budget to build them into robust solutions and to just, you know, I mean, how many people were using co-counsel, 605 00:52:04,605 --> 00:52:05,005 right? 606 00:52:05,005 --> 00:52:07,285 Like here's this thing that helps you draft. 607 00:52:07,285 --> 00:52:08,885 Like, you know, I don't want to use it. 608 00:52:08,885 --> 00:52:10,015 know how to draft a document. 609 00:52:10,015 --> 00:52:12,675 Like, well, give it to an associate. 610 00:52:12,675 --> 00:52:15,005 And if it makes them faster, it makes them faster. 611 00:52:15,005 --> 00:52:15,913 And then 612 00:52:15,931 --> 00:52:19,904 figure out your business of law and your profit margins later. 613 00:52:20,024 --> 00:52:21,355 Yeah, that's good advice. 614 00:52:21,355 --> 00:52:34,235 mean, I think it's, that's a prudent path is to, you know, maximizing that risk reward equation, talk to existing vendors that you are already established in the firm and who 615 00:52:34,235 --> 00:52:39,349 you've already made investments and commitments in and, uh, start plotting your journey there. 616 00:52:39,349 --> 00:52:40,430 think that's, 617 00:52:40,593 --> 00:52:49,537 I do a lot of talking people down and I've got decades of experience in it, know, working as a contractor, walking into somebody's building and they've got, they know exactly 618 00:52:49,537 --> 00:52:57,620 what's wrong and you're like, okay, let's talk about, you know, what are the symptoms before you try to pay me for a cure that won't work, right? 619 00:52:58,101 --> 00:52:59,007 Small wind. 620 00:52:59,007 --> 00:53:12,386 I heard this quote, I can't remember where recently about Albert Einstein said that if he had a problem that were to have the potential to end humanity and he had one hour to solve 621 00:53:12,386 --> 00:53:19,500 it, he would spend 59 minutes trying to understand the problem in one minute coming up with a solution. 622 00:53:20,581 --> 00:53:22,062 I mean, I think that's... 623 00:53:22,242 --> 00:53:24,283 That's, that's prudent advice. 624 00:53:24,343 --> 00:53:25,973 you gotta understand what you're solving for. 625 00:53:25,973 --> 00:53:30,925 And with technology, it's so easy to get, I call it ADSO instead of ADD. 626 00:53:30,925 --> 00:53:32,976 It's a 10 attention deficit. 627 00:53:32,976 --> 00:53:34,546 Ooh, shiny object. 628 00:53:34,547 --> 00:53:40,649 You know, it's a, it's, it's one step above, um, ADD. 629 00:53:40,649 --> 00:53:50,773 So, well, um, how, how to wrap it up here, how do folks get in touch with you, learn more about the, the, the work that you do or the writing that you do, how, to, how to folks 630 00:53:50,773 --> 00:53:51,515 find you. 631 00:53:51,515 --> 00:53:54,988 Sure, absolutely on LinkedIn, definitely not on Twitter. 632 00:53:54,988 --> 00:54:02,454 I'll start with that website, gammawave.ai, know, email, eric at gammawave.ai. 633 00:54:02,454 --> 00:54:08,349 But I'm findable on LinkedIn and I think there's only two people in the world that have my first and last name. 634 00:54:08,349 --> 00:54:11,191 So honestly, you can Google me and find me. 635 00:54:11,406 --> 00:54:14,930 That's amazing because I have a much more unique first name and last name. 636 00:54:14,930 --> 00:54:19,475 It feels like Theodore Theodoropoulos and there's several. 637 00:54:19,475 --> 00:54:21,935 So that's surprising. 638 00:54:21,935 --> 00:54:23,820 Greece is a big country, if I'm guessing right. 639 00:54:23,820 --> 00:54:24,792 Yeah. 640 00:54:24,835 --> 00:54:25,206 All right. 641 00:54:25,206 --> 00:54:26,010 Well, good stuff, man. 642 00:54:26,010 --> 00:54:29,425 I really appreciate you spending some time and hope to see you soon. 643 00:54:29,425 --> 00:54:29,887 You bet. 644 00:54:29,887 --> 00:54:30,482 Thanks for having me. 645 00:54:30,482 --> 00:54:31,414 Really appreciate it. 646 00:54:31,414 --> 00:54:32,736 All right, take care. 00:00:03,899 Eric, how are you this afternoon? 2 00:00:03,899 --> 00:00:05,920 Fantastic, thanks for having me, Ted. 3 00:00:05,970 --> 00:00:07,011 Awesome. 4 00:00:07,011 --> 00:00:12,436 So you and I got connected on LinkedIn as so many of my guests and I do. 5 00:00:12,436 --> 00:00:20,583 And I think we were riffing on fractional chief AI officer and legal and one thing led to another. 6 00:00:20,583 --> 00:00:27,749 And you and I chatted and I think it would, it's going to be a great conversation talking about AI and legal. 7 00:00:27,749 --> 00:00:31,202 But before we get into that, let's, let's get you introduced. 8 00:00:31,202 --> 00:00:33,826 So you, you have an interesting path. 9 00:00:33,826 --> 00:00:44,461 You started your career in construction and then later pivoted to legal and you've worked in a variety of capacities, including a fractional chief AI officer. 10 00:00:44,461 --> 00:00:52,734 And sounds like you spent some time in the startup world as well, but why don't you tell us a little bit about who you are, what you do and where you do it. 11 00:00:52,977 --> 00:00:57,361 Sure, cut me off if I start to run long, because it is a twisted tale. 12 00:00:57,361 --> 00:01:01,204 I like to stay entertained as I sort of learn and grow through my career. 13 00:01:01,305 --> 00:01:02,726 Absolutely correct. 14 00:01:03,007 --> 00:01:10,213 During the dot-com boom here in Chicago, tons of construction, so I kind of kicked off my career general contractor, property developer. 15 00:01:10,995 --> 00:01:18,147 Had a good run of it, but definitely always kind of had that thirst for just truly intellectually engaging. 16 00:01:18,147 --> 00:01:29,310 work and I pivoted my career about 10 or more years ago to focus more on, you know, technology enabled business process, legal process, just really was feeling some 17 00:01:29,310 --> 00:01:36,512 opportunities as I kind of moved through my life and through the world for, you know, so many different things that could be done better. 18 00:01:36,512 --> 00:01:46,765 Uh, you know, in, the meantime, I've, uh, you know, spent a little bit of time on boats, on Navy pier, you know, got my captain's license to sort of like continuing the learning 19 00:01:46,765 --> 00:01:47,601 adventure. 20 00:01:47,601 --> 00:02:00,401 as it were, but went to law school for a brief master's program where I, you know, really drilled down on legal process, process improvement, tech enabled law and all everything 21 00:02:00,401 --> 00:02:02,021 sort of related there. 22 00:02:02,021 --> 00:02:12,601 And yeah, spent some time at a startup focusing on basically transactional law, legal, legal departments at very large organizations and how is it that they manage their 23 00:02:12,601 --> 00:02:13,381 language? 24 00:02:13,381 --> 00:02:17,513 And then I've been consulting for a while and then 25 00:02:17,565 --> 00:02:26,141 President Biden put out, you know, the executive order for, you know, every department, every administration should have an AI officer. 26 00:02:26,141 --> 00:02:33,076 And that kind of kicked off my market exploration of, who else is looking for, you know, an AI officer? 27 00:02:33,076 --> 00:02:44,994 And as I might've mentioned earlier, the most popular response that I get when I talk about that type of role, which everyone recognizes as being brand new is, wow, that's an 28 00:02:44,994 --> 00:02:46,009 expensive hire. 29 00:02:46,009 --> 00:02:55,586 Like that exact quote I've heard multiple times where people realize, you know, you're going to need someone that speaks business, but also definitely if everything that's true 30 00:02:55,586 --> 00:03:01,559 about the possibilities of improving legal process are true with AI, they're going to need to speak legal. 31 00:03:01,559 --> 00:03:03,101 They're going to need to speak technology. 32 00:03:03,101 --> 00:03:05,332 So, you know, kind of acting as as a translator. 33 00:03:05,332 --> 00:03:08,534 that's, that's kind of the kickoff to, where I got to now. 34 00:03:08,952 --> 00:03:10,923 You know, it's interesting. 35 00:03:11,484 --> 00:03:16,908 actually, don't LOL often, but I laughed out loud. 36 00:03:16,908 --> 00:03:24,313 So if I remember correctly, there are 62 federal agencies that now need a chief AI officer. 37 00:03:25,514 --> 00:03:28,296 think about the skillset that you've just described. 38 00:03:28,296 --> 00:03:36,322 How many people who are on the cutting edge of technology want to go work for the government who's 39 00:03:36,334 --> 00:03:39,518 20 years behind, they're still running mainframes, right? 40 00:03:39,518 --> 00:03:44,685 So it's like, I thought that was a really interesting mandate. 41 00:03:44,685 --> 00:03:47,469 And I laughed out loud. 42 00:03:47,469 --> 00:03:54,155 I pictured somebody like, you know, the recruiter on a rotary phone calling up to recruit. 43 00:03:54,155 --> 00:03:54,758 upgrade. 44 00:03:54,758 --> 00:03:57,413 Yeah, our eight inch floppy disks don't work so well anymore. 45 00:03:57,413 --> 00:03:58,165 Let's upgrade. 46 00:03:58,165 --> 00:03:58,935 Yeah. 47 00:03:59,094 --> 00:04:01,525 Yeah, that I got a, I got a big kick out of that. 48 00:04:01,525 --> 00:04:06,678 That was a few months ago and I don't laugh out loud often, but I really did on that one. 49 00:04:06,678 --> 00:04:14,483 Um, well, you and I talked about something that I have talked about at conferences. 50 00:04:14,483 --> 00:04:23,748 I was actually on a panel with a very esteemed group talking about AI, where it should live in the organizational chart in a law firm. 51 00:04:23,948 --> 00:04:27,090 And we are seeing more and more 52 00:04:27,272 --> 00:04:30,824 C-suite titles with AI in it. 53 00:04:31,165 --> 00:04:33,117 Very few chief AI officers. 54 00:04:33,117 --> 00:04:41,673 fact, I only know of one, but there are lots of chief innovation in AI officer, chief knowledge in AI officers. 55 00:04:41,674 --> 00:04:50,190 So, you know, what are your thoughts on C-suite roles with AI in the actual title? 56 00:04:50,190 --> 00:04:52,862 Do you think this is going to be a thing going forward? 57 00:04:53,713 --> 00:04:55,253 I think it can be a thing. 58 00:04:55,253 --> 00:04:59,735 I'm going to be make a general statement and say that it's probably good bet. 59 00:04:59,735 --> 00:05:04,446 It'll be a little bit more show than substance depending on the organization. 60 00:05:04,446 --> 00:05:17,219 Of course, I've seen law firms and other types of organizations where this is the final kick that they truly needed to really start, you know, leveraging all the data they have 61 00:05:17,219 --> 00:05:21,760 in the back end to improve and accelerate their processes. 62 00:05:22,301 --> 00:05:23,681 And so it's like, well, 63 00:05:23,681 --> 00:05:24,692 AI is the biggest thing. 64 00:05:24,692 --> 00:05:29,645 We're going to look great having AI at some level of our organization. 65 00:05:29,645 --> 00:05:39,302 So let's just make the first sort of 21st century technology step be, we've got a person we can point to that says that's our AI person. 66 00:05:39,302 --> 00:05:49,139 And they might be someone who's just literally playing around with ChatGPT GPT on weekends, but is coming from like being a full-time lawyer background all the way to, you 67 00:05:49,139 --> 00:05:51,831 know, somebody sort of picked up from a different. 68 00:05:52,399 --> 00:05:53,120 different industry. 69 00:05:53,120 --> 00:06:06,139 But if, if, we're talking about the business world in general, C level, uh, I've, I've worked with organizations that really understand the power of the technology, but also how 70 00:06:06,139 --> 00:06:08,860 their business works, how their organization works. 71 00:06:08,881 --> 00:06:17,907 And it's, it's pretty similar to a conversation I had earlier today, which is if I get the quote, right, the gentleman I was speaking to who has, you know, straddled the worlds 72 00:06:17,907 --> 00:06:19,469 between technology and law. 73 00:06:19,469 --> 00:06:20,497 He said, 74 00:06:20,497 --> 00:06:30,337 As soon as something gets operationalized, take it out of the hands of legal, know, it's like HR exists for a reason, all these different departments that are sort of legal 75 00:06:30,337 --> 00:06:31,617 adjacent. 76 00:06:32,197 --> 00:06:38,337 The other way that he phrased it was legal is by default bespoke. 77 00:06:38,337 --> 00:06:40,817 That's kind of the two sides of the coin. 78 00:06:40,877 --> 00:06:47,367 So as soon as it can be operationalized, you sort of build it out into a department. 79 00:06:47,367 --> 00:06:50,093 And that's that's kind of the mindset. 80 00:06:50,093 --> 00:06:53,194 is like, nothing is actually AI specific. 81 00:06:53,194 --> 00:06:56,385 I mean, we're talking about, you know, dot com level transformations. 82 00:06:56,385 --> 00:07:02,066 It goes across the entire organization, but you know, how many chief internet officers did, did we see? 83 00:07:02,066 --> 00:07:12,009 Like if you sort of do a comparison about how organizations transformed, uh, and then let's be honest, there's some of them that realized, you know, we're, so much in the real 84 00:07:12,009 --> 00:07:18,953 world that there's some dot com transformation, but as opposed to say a retailer that's, you know, moving most of their business online. 85 00:07:18,953 --> 00:07:19,715 So. 86 00:07:19,715 --> 00:07:21,306 Long way of answering. 87 00:07:21,566 --> 00:07:26,949 I don't think we're going to see that many pure chief AI officer. 88 00:07:27,690 --> 00:07:37,075 I think the folks that really know what they're doing oftentimes will have a chief data officer who touches almost no technology, which is, uh, I've had several conversations 89 00:07:37,075 --> 00:07:40,978 with CDOs over the years and their perception has not changed. 90 00:07:40,978 --> 00:07:47,201 And I've got, I got a lot of respect for CDOs, because they are just the absolute 91 00:07:47,287 --> 00:07:54,979 Wranglers ambassadors translators oftentimes in their organization because their organizations have data. 92 00:07:54,979 --> 00:08:01,451 It can be applied to AI processes, but you've got to extract it, transform it, get it loaded somewhere else. 93 00:08:01,451 --> 00:08:09,013 And that's, that's where the CDO CDO comes in is kind of, you know, intermeshing the data and the processes of an organization. 94 00:08:09,013 --> 00:08:11,347 So you can call it, call it AI. 95 00:08:11,347 --> 00:08:16,417 Um, but I think titles mean a lot and, uh, a lot of. 96 00:08:16,417 --> 00:08:23,464 Orgs won't necessarily take that as meaning, know, our department doesn't have much to do with AI go away, right? 97 00:08:23,464 --> 00:08:25,156 Like there's always that danger. 98 00:08:25,156 --> 00:08:27,969 So that's my read on the world. 99 00:08:28,118 --> 00:08:28,598 Yeah. 100 00:08:28,598 --> 00:08:37,085 Well, you know, your point about this not being, there not being a lot of substance, legal has already demonstrated this with innovation, right? 101 00:08:37,085 --> 00:08:49,075 So the chief innovation officer, chief knowledge and innovation officer, um, so that they, I did a, a quick survey back. 102 00:08:49,075 --> 00:08:53,519 Um, we've been around in legal for a long time and I'm a hoarder of data. 103 00:08:53,519 --> 00:08:57,826 So I had Ilta rosters going all the way back to like 2010. 104 00:08:57,902 --> 00:09:06,462 Um, I took the 2014 Ilta roster since that was 10 years ago, looked at how many, uh, how many titles had the word innovation in it? 105 00:09:06,462 --> 00:09:07,502 16. 106 00:09:07,502 --> 00:09:07,822 Okay. 107 00:09:07,822 --> 00:09:16,362 So legal did not exist materially in, in legal, in law firms, uh, 10 years ago, there were, there were zero C-suite. 108 00:09:16,362 --> 00:09:22,742 There were just one or two directors and some managers, analysts, so on and so forth. 109 00:09:22,782 --> 00:09:27,074 So in the 10 years since 2014, 110 00:09:27,074 --> 00:09:28,809 There are now over 300. 111 00:09:28,809 --> 00:09:36,866 It was almost a 2000 % increase and has the legal innovation work increased by 2000 %? 112 00:09:36,866 --> 00:09:38,359 Absolutely not. 113 00:09:38,460 --> 00:09:50,205 There are some, there is some really good innovation work happening in the legal space, but I can tell you that just from, you know, I have deep relationships with a lot of folks 114 00:09:50,205 --> 00:09:51,726 in KM who 115 00:09:51,822 --> 00:09:57,592 suddenly I saw an eye appear in their name and just asked like, how did your, how has your job changed? 116 00:09:57,592 --> 00:09:59,962 And the answer was not much. 117 00:09:59,962 --> 00:10:08,774 So, um, you know, the concept of innovation theater is real and, um, we may run into that same thing with, with AI. 118 00:10:10,073 --> 00:10:10,963 Agreed, agreed. 119 00:10:10,963 --> 00:10:15,333 It's interesting you say 10 years because I was really digging into this subject five years ago. 120 00:10:15,333 --> 00:10:22,413 Don't have any hard data to present, but 2019 and then at the beginning of COVID early 2020, I had the opportunity. 121 00:10:22,413 --> 00:10:24,423 mean, everybody was locked in their house, right? 122 00:10:24,423 --> 00:10:25,993 First couple of months. 123 00:10:26,013 --> 00:10:28,103 Anybody and everybody would take a call. 124 00:10:28,103 --> 00:10:36,431 It was a truly like extraordinary singular time, I think, in the world of business and professions was everybody was just trying to figure out what to do and 125 00:10:36,431 --> 00:10:38,702 was very sort of closed off from the world. 126 00:10:38,702 --> 00:10:49,605 So, I was relatively early in kind of my exposure to the world of very high end legal, but you could get on the phone with just about anybody, which was cool. 127 00:10:49,906 --> 00:10:59,997 And the feedback that I got was it felt like we were at the tail end of the first push for innovation where everyone I talked to was like, yeah. 128 00:11:00,015 --> 00:11:01,576 Innovation got really important. 129 00:11:01,576 --> 00:11:08,410 So everybody put that word in somewhere, but then you have the conversations and then the feedback is basically mirrors what, what you had. 130 00:11:08,410 --> 00:11:16,073 The interesting thing is, kind of the most recent five years, uh, it's been sort of that, that curve, that wave has started to grow again. 131 00:11:16,073 --> 00:11:18,175 There's, there's absolutely been more hires. 132 00:11:18,175 --> 00:11:19,796 There's been more work. 133 00:11:19,796 --> 00:11:29,957 Um, but I think it's very, very similar innovation and AI sit a hundred people down and have them define both terms and you'll have. 134 00:11:29,957 --> 00:11:38,589 you know, 400 terms, different definitions, because it's such an amorphous name, role, topic. 135 00:11:39,051 --> 00:11:40,834 People don't know what they're supposed to do. 136 00:11:40,834 --> 00:11:43,828 Nobody else can say clearly what it is that they do do. 137 00:11:43,828 --> 00:11:46,251 So it's it's it's very much the same thing. 138 00:11:46,328 --> 00:11:46,998 Yeah. 139 00:11:46,998 --> 00:11:50,941 mean, innovation exists in other industries. 140 00:11:50,941 --> 00:11:53,623 It's usually called something different. 141 00:11:53,783 --> 00:12:01,829 I spent 10 years at Bank of America and worked in a business transformation group, which was essentially innovation. 142 00:12:01,829 --> 00:12:14,882 I think that the reason the legal profession has latched on to the word is because clients have commented publicly of just how un-innovative law firms have been. 143 00:12:14,882 --> 00:12:20,574 You know, they still use, I mean, you know, when I started in this actually five years after I started. 144 00:12:20,574 --> 00:12:22,565 So this is 2013. 145 00:12:22,565 --> 00:12:27,447 I remember doing a series of road shows through Ilta that law firms would host us. 146 00:12:27,447 --> 00:12:31,089 So, I mean, I did like 24 in one year. 147 00:12:31,089 --> 00:12:41,613 So I was all over the place, ran myself into the ground, but, um, I saw the inside of 24 different law firms and I saw typewriters. 148 00:12:41,613 --> 00:12:43,246 This is 2013. 149 00:12:43,246 --> 00:12:45,367 2014 timeframe. 150 00:12:45,387 --> 00:12:55,151 So, you know, we've all heard, you know, the billable hour won't die and, um, you know, value-based billing. 151 00:12:55,151 --> 00:12:56,181 Yeah, it's a thing. 152 00:12:56,181 --> 00:13:01,833 It's like, you know, as Richard Susskind says, you know, tell room full of millionaires, they're doing it wrong. 153 00:13:02,794 --> 00:13:03,305 You know, 154 00:13:03,305 --> 00:13:05,245 know, yeah, the billable hour is going to die. 155 00:13:05,245 --> 00:13:14,005 That thing where you pay someone $300 an hour and charge $900 an hour for their service, you know, one to one ratio that won't work anymore. 156 00:13:14,005 --> 00:13:19,465 At some point we're telling you, but you know, until then you're making 600 an hour. 157 00:13:19,465 --> 00:13:21,864 So yeah, totally, totally, totally agree. 158 00:13:21,864 --> 00:13:23,805 Yeah, it is. 159 00:13:24,386 --> 00:13:31,853 You know, I think AI has kind of reignited like there's been waves of pressure. 160 00:13:31,853 --> 00:13:36,477 I feel like in legal, the last one was ALS PS, right? 161 00:13:36,477 --> 00:13:44,644 The alternative legal service providers that were going to put pressure on law firms to do value based billing. 162 00:13:44,644 --> 00:13:45,525 And you know what? 163 00:13:45,525 --> 00:13:47,150 ALS PS, I mean, 164 00:13:47,150 --> 00:13:59,233 They're still around, but it wasn't nearly the movement that the industry really thought it was or was writing about in terms of transforming the industry. 165 00:13:59,233 --> 00:14:04,295 seems like clients kind of like it. 166 00:14:04,295 --> 00:14:17,186 They kind of like the billable hour because I think they're partly to blame as to why, why it hasn't hasn't died because if they were demanding AFAs, I think 167 00:14:17,186 --> 00:14:20,156 I think we'd have them, we'd have more of them. 168 00:14:20,782 --> 00:14:32,670 there's an insight that I gained about a month ago from a long established, I believe their role was, yes, CIO, chief innovation officer of a large like AmLaw 100 law firm. 169 00:14:32,670 --> 00:14:38,257 And we're having this discussion about data analysis and deriving 170 00:14:38,257 --> 00:14:40,317 Uh, you know, how do you come to a flat fee? 171 00:14:40,317 --> 00:14:49,757 How do you analyze a case and you matter compared to, mean, especially if you're, serving large corporate clients, you, you, have, you know, similar cases for that same client that 172 00:14:49,757 --> 00:14:51,637 you could run analysis for. 173 00:14:51,637 --> 00:14:56,117 Nobody tracks time better than, you know, law firms, six minute increments. 174 00:14:56,117 --> 00:15:04,461 What was astounding to me I, and I got to admit, you know, I had not heard this before was a very simple statement that said, yes, we did. 175 00:15:04,461 --> 00:15:05,692 Awesome data analytics. 176 00:15:05,692 --> 00:15:13,928 And we got a really good handle, we think on being able to predict how much it matter was going to cost, you know, flat fee, you know, percentage profit on top. 177 00:15:14,009 --> 00:15:20,734 We brought that to more than one client who said, we don't trust you. 178 00:15:20,994 --> 00:15:25,878 We don't trust the math that you've done in the background to come up with that flat fee. 179 00:15:25,878 --> 00:15:34,525 We'd rather still pay you by the hour because we're assuming as always that you're just building in so much. 180 00:15:34,543 --> 00:15:38,356 BS to that figure that we just don't we inherently don't trust it. 181 00:15:38,356 --> 00:15:49,273 And it's just fascinating because, you know, who are you supposed to really trust in this world, right, your accountant, your doctor, your lawyer, and there's sort of that, that 182 00:15:49,273 --> 00:15:55,968 that holiness of the profession, that's oftentimes put forth the whole lawyer, non lawyer argument, we won't get into it. 183 00:15:55,968 --> 00:16:03,025 But if you're going to, you know, elevate a profession to that level of, you know, reliability, 184 00:16:03,025 --> 00:16:12,448 you know, fiduciary duty and trustworthiness to find out that, you know, multimillion dollar customers are just saying, yeah, no, I don't believe it. 185 00:16:12,448 --> 00:16:16,169 And then you really realize, well, who's managing that relationship? 186 00:16:16,169 --> 00:16:21,030 You know, it's oftentimes one partner whose client that truly is. 187 00:16:21,030 --> 00:16:23,171 And that's been another interesting part of the conversation. 188 00:16:23,171 --> 00:16:31,953 And I'll, I'll, I'll end here, but it's, uh, when I've been talking to firms about innovating, sort of going through this entire process. 189 00:16:31,983 --> 00:16:34,844 They say, you know, there's quite a tap dance. 190 00:16:34,844 --> 00:16:40,610 It's just a field of landmines because is the partner going to trust you to talk to the end client? 191 00:16:40,610 --> 00:16:48,119 I that's like one of the biggest, biggest issues because it's their relationship they're going to want to take it with if they do a lateral transfer. 192 00:16:48,119 --> 00:16:52,452 So trust is, is, a surprisingly large issue. 193 00:16:52,452 --> 00:16:58,295 And I got to say that I didn't fully see that coming with his, I'll say as little time in this field as I've had. 194 00:16:58,478 --> 00:16:58,918 Yeah. 195 00:16:58,918 --> 00:17:02,858 You know, and it's not to say that value-based billing doesn't exist. 196 00:17:02,858 --> 00:17:03,458 It does. 197 00:17:03,458 --> 00:17:13,208 I've, I've actually experienced it when I 15 years ago, when I got married, uh, we did a will in a state that was a flat fee. 198 00:17:13,208 --> 00:17:18,638 Um, when I, when I created a, uh, revocable trust, that was a flat, flat fee. 199 00:17:18,638 --> 00:17:24,248 When I did a trademark registration, that was a flat fee, but it's still, it's still on the fringes. 200 00:17:24,248 --> 00:17:28,110 The, the meat and potatoes of legal work is still done. 201 00:17:28,110 --> 00:17:29,191 hourly. 202 00:17:29,191 --> 00:17:44,281 So, you know, I don't know how much AI is going to influence that, but it does feel like it has reignited the conversation because people are now thinking how, okay, as a client, 203 00:17:44,281 --> 00:17:50,726 I want my law firm to be efficient, but you know, and I do want to reduce my spend. 204 00:17:51,607 --> 00:17:56,310 They have personnel and margins and overhead and things that they have to 205 00:17:56,416 --> 00:17:57,587 accommodate for right? 206 00:17:57,587 --> 00:18:00,892 make a huge investment in a Harvey implementation, for example, right? 207 00:18:00,892 --> 00:18:02,213 That's not free. 208 00:18:02,213 --> 00:18:06,999 So I can't give you the time for free and just bill you less. 209 00:18:06,999 --> 00:18:09,042 Somehow there has to be a strategy in there. 210 00:18:09,042 --> 00:18:15,870 So I don't know if you share that perspective that it seems like we're talking about it again more. 211 00:18:16,539 --> 00:18:25,915 We certainly are and there's been similarities drawn between the US auto industry of the 20th century. 212 00:18:25,996 --> 00:18:29,328 We've got a lot of large law firms, mid-size, small. 213 00:18:29,328 --> 00:18:33,380 Same thing was true with, you can't buy a Hudson automobile anymore. 214 00:18:33,461 --> 00:18:36,383 It's only a small number, very, very large. 215 00:18:36,383 --> 00:18:39,225 There are now, of course, new electric upstarts. 216 00:18:39,823 --> 00:18:48,341 But I truly do believe that a lot of the sort of low value work is going to be so thoroughly commoditized. 217 00:18:48,341 --> 00:18:56,057 It's like any commodity, who would have thought that computers would have been a commodity in the 90s, right? 218 00:18:56,057 --> 00:18:58,200 They were all these specialized machines, super expensive. 219 00:18:58,200 --> 00:19:02,875 And now it's just like, whatever, go to the nearest store and buy a laptop and you'll be fine. 220 00:19:02,875 --> 00:19:04,916 So I think that... 221 00:19:05,329 --> 00:19:16,943 You know, with the example of General Electric from what it was, the eighties of bringing legal in-house, a lot of these technologies are going to be implemented by the clients and 222 00:19:16,943 --> 00:19:28,945 a lot of the business that especially large corps put towards their law firms is going to be very geographically specific and very matter specific and be for the really, really 223 00:19:28,945 --> 00:19:33,208 complex matters because this all ties into the conversation that we just had. 224 00:19:33,208 --> 00:19:35,107 If you bring in, you know, I've 225 00:19:35,107 --> 00:19:37,640 I founded the Society of Legal Engineers here in the US. 226 00:19:37,640 --> 00:19:48,091 Like if you bring in some legal engineers and some business experts and you start building systems inside a corporation that can afford to build them, you can automate, streamline a 227 00:19:48,091 --> 00:19:53,396 heck of a lot of legal processes as long as your lawyers inside the order are willing to do it. 228 00:19:53,397 --> 00:19:57,371 And you'll see significant savings and speed over time. 229 00:19:57,371 --> 00:20:07,768 Yeah, you know, and I know there's a lot of fear, uncertainty and doubt, and especially at law firm leadership level, you know, when like the Goldman report came out and, you know, 230 00:20:07,768 --> 00:20:15,923 I am, I am confident that there will not be one legal job net loss as a result of AI. 231 00:20:15,923 --> 00:20:25,730 I think there will be plenty that are impacted and there will be shifts, dramatic shifts in where, how legal work is done. 232 00:20:25,804 --> 00:20:29,236 But you know, a good parallel is the accounting industry. 233 00:20:29,236 --> 00:20:34,140 So when, when spreadsheets came out 1983, I went back and looked this up. 234 00:20:34,140 --> 00:20:37,782 I was going to write a post on it. 235 00:20:37,782 --> 00:20:39,653 1983 Lotus one, two, three came out. 236 00:20:39,653 --> 00:20:42,986 There were about 850,000 CPAs. 237 00:20:42,986 --> 00:20:52,372 I don't know if you remember, I was a kid, but I was really into computers and I happened to remember my mom owned a restaurant and she had me doing compute. 238 00:20:52,372 --> 00:20:55,874 Like I was in the back writing basic programs. 239 00:20:56,078 --> 00:21:06,977 to, I got a list of, prospects mailing list on a floppy disc, loaded it in and then created mailing labels so she could mail out menus. 240 00:21:06,977 --> 00:21:19,677 Like, you I was that kid in 10 years old with a TI 99 four a, but I remember she had an accountant, she had a CPA who was taught, who was asking me 10 year old, if I knew how to 241 00:21:19,677 --> 00:21:25,472 use the spreadsheets that were coming out and I didn't have a clue, but they were worried and 242 00:21:25,624 --> 00:21:29,345 There are now 1.5 million in the accounting industry. 243 00:21:29,345 --> 00:21:33,076 Again, these are rough numbers, but I just looked them up maybe two weeks ago. 244 00:21:33,076 --> 00:21:35,477 So the industry is more than doubled. 245 00:21:35,477 --> 00:21:48,580 And if you think about impact to an industry, it's not that different in how spreadsheets think about all of the work that accountants used to do by hand that is now done 246 00:21:48,580 --> 00:21:53,822 projections, pro formas, you know, even just basic P and L's. 247 00:21:53,822 --> 00:21:54,648 So 248 00:21:54,648 --> 00:21:56,733 I don't think there's going to be any job loss. 249 00:21:56,733 --> 00:21:59,811 think there's going to be a ton of impact, but no loss. 250 00:21:59,811 --> 00:22:00,102 I don't know. 251 00:22:00,102 --> 00:22:01,254 What's your perspective? 252 00:22:01,881 --> 00:22:11,810 I think that I can't find a good comparison for an industry that is as underserved as the legal services industry. 253 00:22:11,810 --> 00:22:25,432 So I don't really have any worry except for if you're hiding something, which a lot of, I'm going to be honest, professionals in the legal industry tend to do, which is, you 254 00:22:25,432 --> 00:22:29,711 know, hiding the fact that they don't know about a particular matter or are. 255 00:22:29,711 --> 00:22:41,359 Recycling content and billing, you know multiple hours for drafting a document that they just change a couple terms on stuff like that's gonna go away, but there are so many unmet 256 00:22:41,359 --> 00:22:53,778 legal needs and regulations Regardless of who's running the country or what direction the government goes There's always a very large and complex regulatory environment and you 257 00:22:53,778 --> 00:22:58,423 know, nobody here in the US can really affect You know GDPR 258 00:22:58,423 --> 00:23:00,484 or the new EU AI act. 259 00:23:00,484 --> 00:23:05,856 we're going to have to comply with regulations all over the world. 260 00:23:06,516 --> 00:23:09,617 there'd be no sort of paucity of work. 261 00:23:09,617 --> 00:23:16,710 It's just that people are going to need to change and learn and grow to adapt to the new environment. 262 00:23:16,710 --> 00:23:19,841 And if you don't like change, then yeah, you're scared right 263 00:23:20,034 --> 00:23:20,714 Yeah. 264 00:23:20,714 --> 00:23:31,429 So, know, econ 101 or 102 macroeconomics supply and demand, they meet at the price of equilibrium. 265 00:23:31,429 --> 00:23:44,124 If you look at the hourly rates of legal professionals relative to comparably trained professionals in other fields, legal enjoys a premium. 266 00:23:44,224 --> 00:23:48,856 And it's for the exact reason that you're talking, there is way more demand 267 00:23:48,856 --> 00:23:50,668 than there is supply. 268 00:23:50,668 --> 00:23:53,750 There's a ton of unmet needs out there, even in the business world. 269 00:23:53,750 --> 00:24:10,454 For example, I had the CKO from Wilson Sonsini on the podcast and they have a platform called Neuron where, yeah, yeah, for startups where, you know, they'll do automated tasks 270 00:24:10,454 --> 00:24:12,586 like offer letters, right? 271 00:24:12,586 --> 00:24:17,664 Like how great would it be to, but who can afford to have an attorney draft your 272 00:24:17,664 --> 00:24:28,571 Offer letters, but if it's automated and you know and they're checking all the boxes and like you know there there is real risk even to like you go putting the wrong like go make 273 00:24:28,571 --> 00:24:41,330 an offer letter to a 1099 you've now put yourself in a potential co-employment risk scenario right like there is all sorts of even for for entities that can afford legal 274 00:24:41,330 --> 00:24:42,230 work. 275 00:24:42,242 --> 00:24:44,104 let alone all the ones that can't. 276 00:24:44,104 --> 00:24:45,286 So I agree with you. 277 00:24:45,286 --> 00:24:52,684 think there's huge unmet needs out there and we'll, there will be shifts, but I don't think there will be any net loss. 278 00:24:52,783 --> 00:24:54,354 Yeah, I feel obliged. 279 00:24:54,354 --> 00:25:02,777 I'm going to quote my mentor and professor from law school, Bill Henderson, who's now working in Indiana for, you know, access to justice issues. 280 00:25:03,097 --> 00:25:08,063 But, you know, he's worked deeply in the, the business of law as it were. 281 00:25:08,063 --> 00:25:14,982 It doesn't really publish anymore on his blog, legalevolution.org, but definitely a shout out for tons of great content. 282 00:25:14,982 --> 00:25:22,353 But his quote is, how else can an English major with three additional years of education make a million dollars a year? 283 00:25:22,353 --> 00:25:26,133 That's kind of like, it's a lot of words, but it's a question that's posed. 284 00:25:26,133 --> 00:25:29,093 That's like, it's not that much more school. 285 00:25:29,093 --> 00:25:29,633 Yeah. 286 00:25:29,633 --> 00:25:32,663 You have to, you know, work for a firm as an associate. 287 00:25:32,663 --> 00:25:37,853 Uh, but you know, I, I went to a T 14 law school, you know, I didn't get a JD from there. 288 00:25:37,853 --> 00:25:43,473 So I'm not eligible for the, at the time, $235,000 a year. 289 00:25:43,713 --> 00:25:48,413 Initial salary that a, that a first year associate was going to make at big law. 290 00:25:48,413 --> 00:25:50,991 Uh, but it just goes to show like. 291 00:25:50,991 --> 00:25:58,970 you power through and you get through that sort of that gauntlet of, of, of, know, winnowing down of, of who gets to play and who doesn't. 292 00:25:58,970 --> 00:26:01,323 And the rewards are still quite high. 293 00:26:01,368 --> 00:26:03,039 Yeah, absolutely. 294 00:26:03,039 --> 00:26:14,378 Well, you and I talked in our last conversation about AI solutions and you know, one of I shared my opinion on Copilot, which is is not a good one really today. 295 00:26:14,378 --> 00:26:20,021 I have full confidence in Microsoft's ability to get this right in the long term, but they have rushed this out. 296 00:26:20,021 --> 00:26:27,236 I mean, it's you know, I mean, and there's just this sense of urgency to put Copilot everywhere it's on. 297 00:26:27,236 --> 00:26:28,256 even putting it on. 298 00:26:28,256 --> 00:26:28,898 They read it. 299 00:26:28,898 --> 00:26:31,022 There hasn't been a keyboard change in like. 300 00:26:31,022 --> 00:26:36,645 20 years and they've there's now going to be a Copilot key on new keyboards. 301 00:26:37,146 --> 00:26:41,028 So I think they've gotten a little bit over their skis on this. 302 00:26:41,028 --> 00:26:44,199 Like I have not had a good experience and I've talked about it. 303 00:26:44,199 --> 00:26:48,392 My listeners are probably tired of hearing me complain about it. 304 00:26:48,392 --> 00:26:52,874 But you know the more you know in the beginning I felt like they're Microsoft. 305 00:26:52,874 --> 00:26:56,597 One thing Microsoft is really good at is rewarding evangelists. 306 00:26:56,597 --> 00:26:59,668 So they have this MVP program, the Microsoft. 307 00:27:00,180 --> 00:27:00,713 Yeah. 308 00:27:00,713 --> 00:27:10,169 professional and they re and people like evangelize and drink the Kool-Aid and and spout it and they have influence. 309 00:27:10,169 --> 00:27:19,194 And I think that, you know, in the beginning, before people really got their hands on it, they just assumed that it was going to be great and it's not great. 310 00:27:19,194 --> 00:27:21,005 At least not yet from my perspective. 311 00:27:21,005 --> 00:27:22,296 I think there's a lot of limitations. 312 00:27:22,296 --> 00:27:22,657 I don't know. 313 00:27:22,657 --> 00:27:25,458 Where do you stand on current state of Copilot? 314 00:27:25,949 --> 00:27:32,111 Having worked intimately with knowledge professionals that work very heavily in documents. 315 00:27:32,451 --> 00:27:41,293 I can tell you that automating and creating magic for people whose bread and butter is what you're messing with. 316 00:27:41,373 --> 00:27:42,513 It's not. 317 00:27:43,014 --> 00:27:44,344 It's not difficult. 318 00:27:44,344 --> 00:27:50,046 It's not extremely difficult, but it is much more specific than most people would like to imagine that it is. 319 00:27:50,046 --> 00:27:55,177 And it's much more time consuming and a grind than most people would like to accept. 320 00:27:55,403 --> 00:28:07,836 Um, so, uh, I keep on quoting myself and others, but, uh, another thing that I often say these days is, you know, one of the, the, the worst tricks, the ChatGPT GPT, you know, 321 00:28:07,836 --> 00:28:17,019 pulled on the world was, you know, it convinced that AI was, was good at everything because it was just such, you know, I've seen so many lectures talking about, know, the 322 00:28:17,019 --> 00:28:21,490 most rapidly adopted platform in history, ChatGPT GPT, and, know, why is that? 323 00:28:21,490 --> 00:28:23,971 And it's like all of the technology already existed. 324 00:28:23,971 --> 00:28:25,263 It just wasn't as fun. 325 00:28:25,263 --> 00:28:25,613 Right. 326 00:28:25,613 --> 00:28:33,335 So getting back to Copilot, it's just not that it's not that much fun because how often are you really pleasantly surprised? 327 00:28:33,335 --> 00:28:34,116 Right. 328 00:28:34,116 --> 00:28:48,129 And I think you really have to spend time with your end users, understanding specifically what it is exactly that they do beginning to end and really focus on where can I make 329 00:28:48,129 --> 00:28:48,730 magic? 330 00:28:48,730 --> 00:28:53,689 And you know, the whole conversation about like, you know, do you, are you making, you know, and 331 00:28:53,689 --> 00:29:01,134 an aspirin or a vitamin, you, helping people do something that's sort of tangential or are you really solving a huge pain point? 332 00:29:01,134 --> 00:29:03,856 And, um, I think formatting is a big deal. 333 00:29:03,856 --> 00:29:08,409 haven't really seen that much like actual, I mean, let's call it like voice activated. 334 00:29:08,409 --> 00:29:11,555 mean, co-pilot is like, I want to be able to communicate with it. 335 00:29:11,555 --> 00:29:15,384 I want to be able to tell it like, fix this thing about my document. 336 00:29:15,384 --> 00:29:16,245 Right. 337 00:29:16,245 --> 00:29:22,329 Um, and the, the software company I was at, Xmentium, one of, one of the things about automatically building documents. 338 00:29:22,477 --> 00:29:26,679 was that the numbering was always rock solid, the outline numbering. 339 00:29:26,679 --> 00:29:33,132 And I don't think anyone suffers more from outline shenanigans than lawyers, right? 340 00:29:33,132 --> 00:29:41,106 The number of times I've heard folks say, I would be doing a pitch, I'd be saying, hey, yeah, this platform does this and it automatically builds your document and the outline, 341 00:29:41,106 --> 00:29:45,008 no matter what you do, no matter what you change is going to be completely accurate. 342 00:29:45,008 --> 00:29:47,409 Roman numeral ads will be in the right place. 343 00:29:47,601 --> 00:29:54,865 And people will say, you know, I was crying over my tab key last night because I was trying to like manually build. 344 00:29:54,865 --> 00:30:01,738 just gave up on styles, gave up on the outline control and was, you know, had to go through and then, you know, last minute somebody wants to make a change. 345 00:30:01,738 --> 00:30:07,891 that's, that's where I think we're going to start to see making people truly happy, but we're not there yet. 346 00:30:07,891 --> 00:30:13,173 And it's a system that's trying to do magic and everything and not really being that magic. 347 00:30:13,518 --> 00:30:13,898 Yeah. 348 00:30:13,898 --> 00:30:16,838 Speaking of the formatting issue, that's one of the areas. 349 00:30:16,838 --> 00:30:25,448 So I, I, I've tried so many times to, use Copilot Um, I paid for a year's license of it at $30 a month. 350 00:30:25,448 --> 00:30:27,198 wanted to get value. 351 00:30:27,198 --> 00:30:28,978 So I have tried. 352 00:30:29,158 --> 00:30:38,458 One of the things I try is like when I go to ChatGPT GPT and I get back a bulleted list and I've tried multiple times copying, pasting it into a word document, the bullets don't 353 00:30:38,458 --> 00:30:40,688 come over like in normal rich texts. 354 00:30:40,688 --> 00:30:41,874 So I prompted. 355 00:30:41,874 --> 00:30:43,656 But it's clear what they are, right? 356 00:30:43,656 --> 00:30:44,337 They're tabbed. 357 00:30:44,337 --> 00:30:48,661 There's a little emoji, you know, dot. 358 00:30:48,661 --> 00:30:55,088 I've asked Copilot to clean it up, put, these real bullets, and it spit out the instructions to do it manually. 359 00:30:55,088 --> 00:30:56,389 And I'm like, man. 360 00:30:56,643 --> 00:31:05,462 Or even better, let me ask you this question, given the tight relationship and interoperability, why are you using ChatGPT GPT and not just using Copilot? 361 00:31:05,462 --> 00:31:12,638 Yeah, exactly, because I have had too many bumps on my head trying to use Copilot. 362 00:31:13,070 --> 00:31:14,111 Yep. 363 00:31:14,111 --> 00:31:14,832 Plain and simple. 364 00:31:14,832 --> 00:31:18,374 It's nerfed, guard railed, whatever term you want to use. 365 00:31:18,614 --> 00:31:21,426 It's insane how many times I have done the exact same thing. 366 00:31:21,426 --> 00:31:25,214 And I think that's, I'm not here to trash talk anybody's, hard work. 367 00:31:25,214 --> 00:31:26,400 mean, co-pilot. 368 00:31:26,400 --> 00:31:28,431 It's amazing to exist. 369 00:31:28,431 --> 00:31:32,224 It's a lot of work, but it's not doing what people want it to do. 370 00:31:32,224 --> 00:31:35,386 And you you know, there are certain things that it refuses to do. 371 00:31:35,386 --> 00:31:36,277 You're like, forget it. 372 00:31:36,277 --> 00:31:37,738 I'm just going to ChatGPT GPT. 373 00:31:37,738 --> 00:31:39,489 Cause I know what's going to go for me. 374 00:31:39,583 --> 00:31:40,203 exactly. 375 00:31:40,203 --> 00:31:40,487 Yeah. 376 00:31:40,487 --> 00:31:46,006 had a, actually that Ilta, that Ilta roster study that I was telling you about. 377 00:31:46,006 --> 00:31:51,459 I, I put, um, all of the titles, the job titles with innovation. 378 00:31:51,459 --> 00:32:02,415 I cat, I bucketed them into, um, C-suite direct, senior director, director, attorney manager, analyst, and I gave it examples. 379 00:32:02,415 --> 00:32:08,010 And then I uploaded the spreadsheet and told, or now I'm sorry in 380 00:32:08,010 --> 00:32:14,133 Excel, the in-app Copilot, asked it to mimic what I did completely choked. 381 00:32:14,133 --> 00:32:20,315 So I took the exact same prompt, paste it into ChatGPT GPT, uploaded the spreadsheet and it did it. 382 00:32:20,315 --> 00:32:22,116 Now it didn't do it perfectly. 383 00:32:22,116 --> 00:32:27,978 There were a ton of other, like it bucketed way too many in other. 384 00:32:28,599 --> 00:32:31,430 and like it made some silly mistakes. 385 00:32:31,430 --> 00:32:37,430 Like, uh, there were titles with the word chief in it that it didn't 386 00:32:37,430 --> 00:32:49,629 recognized as C suite and I prompted it like I I kind of scolded it like why did you not put titles with chief in the C suite and he was like, oh sorry, I should have done that 387 00:32:49,629 --> 00:32:50,699 and then it would fix it. 388 00:32:50,699 --> 00:32:56,023 But yeah, I mean, uh, ChatGPT GPT and Claude aren't perfect either. 389 00:32:56,023 --> 00:33:06,210 But what, is your perspective on like purpose built, you know, solutions for legal versus the, general J uh, gen AI. 390 00:33:06,398 --> 00:33:08,137 applications that are out there. 391 00:33:09,377 --> 00:33:22,186 I some of them work, but referring to recent conversations I've had, you can't even think about it honestly as legal work. 392 00:33:22,186 --> 00:33:27,660 There's just so many different tech stacks that so many different types of lawyers use. 393 00:33:27,660 --> 00:33:37,837 think, I'll say we, sort of like greater legal technology ecosystem really has to start talking about like, what's the practice area? 394 00:33:38,226 --> 00:33:42,928 that you're applying your technology to, not just like, it's for lawyers, right? 395 00:33:43,149 --> 00:33:45,870 Because it doesn't work right. 396 00:33:46,071 --> 00:33:49,193 There's an early example that I cited. 397 00:33:49,193 --> 00:34:00,320 mean, it was fascinating to have a conversation with someone who was an established patent professional in a reasonably sized IP boutique. 398 00:34:00,341 --> 00:34:06,577 Him and his team filing 300 or more patents per year, and they were working with a platform that would take 399 00:34:06,577 --> 00:34:17,297 Uh, the initial inventors disclosure, their description of what, you know, the novel idea was in their invention and turn it into a set of claims, which is, know, the very core 400 00:34:17,297 --> 00:34:20,257 meat of, a very large patent application. 401 00:34:20,257 --> 00:34:31,617 I want a professional that's doing something 300 or more times per year is telling you that a platform like out of the gate is saving them 50 % of like the time spent on each 402 00:34:31,617 --> 00:34:32,297 project. 403 00:34:32,297 --> 00:34:33,697 That's that's enormous. 404 00:34:33,697 --> 00:34:34,991 And just to 405 00:34:34,991 --> 00:34:42,267 circle back to our earlier conversation, patent prosecution, patent drafting, very much a place where you can flat-fee quote upfront. 406 00:34:42,267 --> 00:34:44,338 Cause you know, what are the unknowns? 407 00:34:44,338 --> 00:34:49,092 I see your technology, I will draft, we will push it through the USPTO as best we can. 408 00:34:49,092 --> 00:34:52,475 So that's a specific use case. 409 00:34:52,635 --> 00:35:03,714 And you know, one quick other one that I looked at and I've had, I've spoken to local governments and law enforcement about is taking police body cam footage, the transcript 410 00:35:03,714 --> 00:35:05,145 just from the audio. 411 00:35:05,149 --> 00:35:08,791 and running that through a platform and generating a police report. 412 00:35:08,791 --> 00:35:11,733 And we're talking at least 50 % time savings. 413 00:35:11,733 --> 00:35:16,365 So this is kind of my view on purpose built versus general. 414 00:35:16,365 --> 00:35:19,777 There's things that are supposed to help people in business. 415 00:35:19,777 --> 00:35:24,079 There's, know, gen AI and technical tools that are supposed to help people in legal. 416 00:35:24,079 --> 00:35:34,767 And then there's like somewhat boring, but very, very specific purpose built tools where the model has been detuned, hallucinations trained correctly. 417 00:35:34,767 --> 00:35:38,622 and weighted properly that it like does magic. 418 00:35:38,622 --> 00:35:41,015 And that's really what you're expecting the AI to do. 419 00:35:41,388 --> 00:35:41,708 Yeah. 420 00:35:41,708 --> 00:35:48,101 You and I talked a little bit about small language models and ones that, that you can run locally. 421 00:35:48,101 --> 00:35:55,364 mean, you know, Meta puts out llama, which you know, you can download and run locally. 422 00:35:55,364 --> 00:36:01,407 I haven't done it, but I think I've downloaded it, but really just kind of played around. 423 00:36:01,407 --> 00:36:08,770 I haven't done anything serious with it, but I think, I think you were talking about how you could 424 00:36:09,122 --> 00:36:21,166 leverage a small language model for something that is really sensitive, like a patent application or transactional documents and work on them locally, blow the thing away and 425 00:36:21,166 --> 00:36:28,384 then re-download it to make sure that there's no overlap in customer data. 426 00:36:28,384 --> 00:36:30,826 mean, is that, people really doing that? 427 00:36:31,561 --> 00:36:34,003 That's the exact story that I got. 428 00:36:34,003 --> 00:36:41,169 So for full clarity, the police body cam app that was a private instance run on the cloud, I believe. 429 00:36:41,850 --> 00:36:52,699 But specifically the patent drafting application, the claim drafting was a local downloaded model that didn't learn or remember anything. 430 00:36:52,699 --> 00:36:59,717 So it was interesting talking to the creators of this platform and they're saying, now there's absolutely no way that if you're working on a 431 00:36:59,717 --> 00:37:08,101 patent for Google and then you need to work on a patent for Apple, it is totally impossible for there to be any sort of cross-infection. 432 00:37:08,101 --> 00:37:18,425 But because it's a downloadable model and we only updated every whatever it is quarter, delete it like you said, redownload it and then use it for your Apple work. 433 00:37:18,425 --> 00:37:21,486 And then you are like so air gapped, it's not even funny. 434 00:37:21,486 --> 00:37:27,649 But overall, local models, I think we spoke about sort of... 435 00:37:28,141 --> 00:37:29,772 It's not American exceptionalism. 436 00:37:29,772 --> 00:37:39,390 It's a passion for what I'll say is avoiding the true sort of energy costs or implications of, you know, what it is that we do. 437 00:37:39,511 --> 00:37:40,181 Giant homes. 438 00:37:40,181 --> 00:37:42,913 I used to work in heating and cooling, high efficiency analysis. 439 00:37:42,913 --> 00:37:47,297 You know, it's just like three air conditioners on a house, gigantic cars. 440 00:37:47,297 --> 00:37:56,284 And I'm not going to go on a bender about save the trees, but I'm just saying that, you know, our nation is so incredibly wealthy that we have the absolute privilege of 441 00:37:57,049 --> 00:38:00,530 almost completely ignoring how much energy we consume to do something. 442 00:38:00,530 --> 00:38:04,251 And I really think that's taken hold in AI. 443 00:38:04,251 --> 00:38:19,355 again, not going into all of the environmental hazards of AI, but like the real fallback to that is we're not building the tools that we could that have way better privacy 444 00:38:19,355 --> 00:38:25,367 controls and way better security because we're just saying, well, there's going to be a new model. 445 00:38:25,367 --> 00:38:27,137 Claude puts out something new. 446 00:38:27,365 --> 00:38:31,376 Uh, you know, open AI makes their adjustments and it's just going to be awesome. 447 00:38:31,376 --> 00:38:39,909 Like at what other time in industry were we, were we so passionate about using like the brand new, most new thing. 448 00:38:39,909 --> 00:38:44,910 It's like a quick comparison is, you know, I'll go shopping for a car. 449 00:38:44,910 --> 00:38:47,561 They do the sort of model refreshes, right? 450 00:38:47,561 --> 00:38:51,302 You don't get the first year of the new model refresh. 451 00:38:51,302 --> 00:38:55,781 You get sort of the, let the company learn how they're building that exact. 452 00:38:55,781 --> 00:39:01,803 form and feature set and then buy in, you know, before they sort of switch over again. 453 00:39:01,803 --> 00:39:03,854 And I really think that mindset's gonna come out. 454 00:39:03,854 --> 00:39:19,220 I think we're gonna see a lot more small compact local models running and possibly some steps taken back from a pure SaaS model to the point of like, go out and get this kind of 455 00:39:19,220 --> 00:39:25,777 laptop that has this level or more of processor and it will run the tool that we're gonna give you. 456 00:39:25,777 --> 00:39:32,393 And it's going to do magic for you, but it won't be cloud based because, you know, that's not the way it should work. 457 00:39:32,942 --> 00:39:49,142 Yeah, I'm one of those suckers who, uh, I bought a 2022 Tundra after it was completely overhauled and, all of the 2022s, the non-hybrid 2022s and half of the 2023s have to have 458 00:39:49,142 --> 00:39:51,222 complete engine replacement. 459 00:39:51,562 --> 00:39:52,902 Yeah. 460 00:39:52,902 --> 00:39:54,698 So they're great. 461 00:39:54,698 --> 00:40:00,442 Yeah, you know, and the problem is like you could just you just I could just rack up miles. 462 00:40:00,442 --> 00:40:05,086 They have announced the recall, but they have not established the process yet. 463 00:40:05,086 --> 00:40:08,007 I mean, imagine how it's hundreds of thousands of vehicles. 464 00:40:08,068 --> 00:40:17,995 So, you know, I could just drive mine forever and I know I'm to get a new engine, but you're going to be out of a vehicle for eight to 10 weeks and they give you like 40 bucks 465 00:40:17,995 --> 00:40:18,565 a day. 466 00:40:18,565 --> 00:40:23,228 I'll be driving the clown car and I'm six foot five, 270 like. 467 00:40:24,066 --> 00:40:26,629 You know, um, so I'm not looking forward to it. 468 00:40:26,629 --> 00:40:27,000 You know what? 469 00:40:27,000 --> 00:40:29,193 And I know better because you're right. 470 00:40:29,193 --> 00:40:36,532 You do want to buy it a year or two, uh, after the remake and then, and then lay your money down. 471 00:40:36,546 --> 00:40:43,456 At least I actually go for the post, the midlife refresh where you get a couple of upgrades. 472 00:40:43,456 --> 00:40:44,834 That's my strategy. 473 00:40:44,834 --> 00:40:50,257 Yeah, I wish I would have followed that advice because now I'm stuck with a view. 474 00:40:50,257 --> 00:40:56,951 mean, I, I'm to take a beating because when you look up the VIN in Kelly blue book, it shows the recall. 475 00:40:56,951 --> 00:40:58,822 So anywhere I traded in, they're going to know. 476 00:40:58,822 --> 00:41:08,678 And if I were to sell it to a private seller who wants to get a new engine, you know, maybe, but, um, all right, well switching gears, uh, you and I talked a little bit about 477 00:41:08,678 --> 00:41:12,170 Harvey and, um, Harvey's a really interesting case. 478 00:41:12,170 --> 00:41:12,970 Like, 479 00:41:13,038 --> 00:41:21,968 You know, they have their series B, which was, think like last December, November, something like that. 480 00:41:21,968 --> 00:41:28,678 It was a 700 million ish valuation, 715. 481 00:41:28,918 --> 00:41:37,498 Seven months later, July, 2024, 1.5 billion was the valuation of their series C round. 482 00:41:37,498 --> 00:41:39,756 So they doubled in value in seven months. 483 00:41:39,756 --> 00:41:44,910 Now, what could have taken place in so think about how difficult it is to double enterprise value, right? 484 00:41:44,910 --> 00:41:48,473 Like they didn't double, they didn't even double revenue. 485 00:41:48,754 --> 00:41:52,017 I think they went from 20 to 30 and those are unofficial numbers. 486 00:41:52,017 --> 00:42:03,416 They don't disclose revenue, but you know, so, um, you know, they've had a couple of interesting pivots, you know, they were trying to raise 600 million to buy VLex and then 487 00:42:03,416 --> 00:42:04,808 that fell through. 488 00:42:04,808 --> 00:42:07,758 They've been running in stealth mode and now they're 489 00:42:07,758 --> 00:42:14,238 I think starting to realize that there's been a little bit of, you know, skepticism about the platform. 490 00:42:14,238 --> 00:42:14,978 I don't know. 491 00:42:14,978 --> 00:42:18,342 What's your take on, on all the, these happenings. 492 00:42:18,353 --> 00:42:30,160 Sure, I've been thinking about it a little bit and I don't know if we can call it the boy band model of entrepreneurship and tech startups, but I'm gonna be honest, I think they 493 00:42:30,160 --> 00:42:40,809 did a truly stellar job of reading what was going to look good to investors, to the market. 494 00:42:40,809 --> 00:42:43,347 I really been thinking about this. 495 00:42:43,347 --> 00:42:47,023 I wanna say, if you've seen the way that 496 00:42:47,023 --> 00:42:53,515 Ford Motor Company started assembly line and has grown and it became incredibly valuable, of course, as a company. 497 00:42:54,475 --> 00:43:11,900 Imagine if like it happened again and you saw in a similar industry that a company was sort of situating itself like just like a previous successful sort of like business model. 498 00:43:11,900 --> 00:43:14,160 And I think that's what Harvey did very, very well. 499 00:43:14,160 --> 00:43:16,501 They knew what was going to appeal to the market. 500 00:43:17,361 --> 00:43:27,521 They knew, I think, with decent confidence that they didn't need to have a fully fleshed out magical product that made everyone happy. 501 00:43:27,521 --> 00:43:30,731 They needed to look really good coming out of the gate. 502 00:43:30,731 --> 00:43:33,361 And they did really well on that. 503 00:43:33,461 --> 00:43:46,277 And because we have these, the industrial revolution has enough sort of back history, we can look at it, how the different companies, American Automotive now is only a few. 504 00:43:46,627 --> 00:43:57,672 I think a lot of investors are prognosticating that Harvey is going to win, that they just went above a couple of critical thresholds, know, started off with not that much, but 505 00:43:57,672 --> 00:44:07,266 looked really good doing it and then got enough traction, got enough customers where it's just like, whether they bought Villex or not, they look like they're going to be a winner 506 00:44:07,266 --> 00:44:09,487 in more than one space. 507 00:44:09,487 --> 00:44:14,009 And so you might as well invest early, the earlier, the better. 508 00:44:14,189 --> 00:44:18,151 because they're probably going to win more likely than many others. 509 00:44:19,292 --> 00:44:21,704 And that's where I think a lot of the bets are being laid. 510 00:44:21,704 --> 00:44:26,436 So I think, you know, if you look at our, do the revenues justify the valuation? 511 00:44:26,436 --> 00:44:30,309 You know, no, but it's, you know, human optimism, right? 512 00:44:30,309 --> 00:44:34,761 But also just people really gaming out, like how is this going to work? 513 00:44:34,761 --> 00:44:40,755 You know, how many different tech platforms is a law firm going to be hiring to do X, Y and Z? 514 00:44:40,755 --> 00:44:41,615 And if... 515 00:44:41,709 --> 00:44:43,730 you can get to that critical mass. 516 00:44:43,730 --> 00:44:52,443 You can be Harvey, you know, they didn't buy VLikes this time, but they've definitely signaled to the market their willingness to, you know, work on very large acquisition 517 00:44:52,443 --> 00:44:53,233 deals. 518 00:44:53,233 --> 00:45:02,957 And I think that's very much just a snowball that keeps happening because, you know, they'll be the one platform to rule them all if everything goes their way. 519 00:45:03,010 --> 00:45:03,790 Yeah. 520 00:45:03,790 --> 00:45:04,060 Yeah. 521 00:45:04,060 --> 00:45:05,171 There's real value there. 522 00:45:05,171 --> 00:45:08,762 It's not like these are bored ape NFTs, right? 523 00:45:08,762 --> 00:45:14,315 what you, they're, they are building a real business. 524 00:45:14,315 --> 00:45:19,997 Um, it's just the, um, their, their go-to-market approach is, interesting. 525 00:45:19,997 --> 00:45:30,892 Um, you know, in my experience, having, having relationships and experience in 526 00:45:30,892 --> 00:45:44,529 your go-to-market team in legal tech, not just having lawyers on staff who've practiced in big law, but actually having your business, your go-to-market team understand the dynamics 527 00:45:44,529 --> 00:45:51,313 in legal tech, which is unique, is a really important ingredient. 528 00:45:51,313 --> 00:45:55,095 When I look at their go-to-market team, I don't see a ton of that. 529 00:45:55,475 --> 00:46:00,780 I see a lot of folks, really smart looking folks, but 530 00:46:00,780 --> 00:46:09,346 Yeah, it's been, and again, the stealth mode and then now coming out of it, it's been, it's been interesting as an outsider who's had very little exposure. 531 00:46:09,346 --> 00:46:12,739 I've seen, I've seen, you know, the stuff they published on the web. 532 00:46:12,739 --> 00:46:15,740 I've actually seen a little bit more than that. 533 00:46:16,702 --> 00:46:18,153 You know, it looks interesting. 534 00:46:18,153 --> 00:46:21,645 It looks like, it looks like it's a UI play. 535 00:46:21,885 --> 00:46:28,800 You know, they're really crafting a user experience for legal that's 536 00:46:28,802 --> 00:46:32,016 going to always probably leverage open AI, right? 537 00:46:32,016 --> 00:46:33,667 They're one of their biggest investors. 538 00:46:33,667 --> 00:46:36,340 So, it's interesting. 539 00:46:36,369 --> 00:46:52,569 I'd have to imagine, I feel obliged to point out that the specific strategy is familiar to the private equity, venture capital space in the pros that I've talked to in that region, 540 00:46:52,569 --> 00:47:05,409 the ones that vet and invest in, put a lot of dollars into tech, AI tech, legal tech, see that it is an emerging way to grow a large business in. 541 00:47:05,717 --> 00:47:16,533 You you're you're building the plane as it flies and that's much more acceptable than it used to be because the Technology is happening so quickly and because there are so many 542 00:47:16,533 --> 00:47:19,034 unknowns about what is AI gonna do? 543 00:47:19,034 --> 00:47:20,355 What is it going to do next? 544 00:47:20,355 --> 00:47:29,900 know, they say You know if you've got a market opportunity you realize that it's your TAM is your total addressable market is is big enough and it's not being satisfied or you can 545 00:47:29,900 --> 00:47:35,397 do it cheaper You've got a business opportunity great or you can be yeti coolers and you can convince people 546 00:47:35,397 --> 00:47:42,421 there's this new thing that you never spent a lot of money on, but now we have a new awesome thing, new market, you do that. 547 00:47:43,121 --> 00:47:52,187 I think it would be sort of, it was impossible to ascertain like what was the big successful market play for legal, partly because there's so many different practice areas, 548 00:47:52,187 --> 00:48:03,113 but partly because I think it was especially hard to even come up with that magical thing that you could convince legal professionals to sort of buy into that, you there is no Yeti 549 00:48:03,113 --> 00:48:03,919 cooler. 550 00:48:03,919 --> 00:48:05,219 for lawyers, look, it does this thing. 551 00:48:05,219 --> 00:48:06,640 It's like, don't want to change. 552 00:48:06,640 --> 00:48:07,450 I don't like that. 553 00:48:07,450 --> 00:48:10,731 Or that makes me more efficient and I don't want to the billable model. 554 00:48:10,731 --> 00:48:11,271 Right. 555 00:48:11,271 --> 00:48:19,393 So I've talked to a decent number of big law customers of Harvey who very much have input into the product. 556 00:48:19,393 --> 00:48:23,174 So they are, is still sort of like a development developing mindset. 557 00:48:24,075 --> 00:48:27,035 I would love to hear some of those product conversations. 558 00:48:27,035 --> 00:48:30,416 You know, they said this, they said that, how do we rectify? 559 00:48:30,416 --> 00:48:33,233 It's kind of how I describe what a lot of my job is, is that 560 00:48:33,233 --> 00:48:42,233 sort of Venn diagram process of finding enough to be successful without, you know, making anybody too upset about not satisfying all their needs. 561 00:48:42,274 --> 00:48:43,096 Yeah. 562 00:48:43,096 --> 00:48:45,152 I mean, it's going to be, it's going to be an interesting journey. 563 00:48:45,152 --> 00:48:47,464 I know we're running out of time here. 564 00:48:47,503 --> 00:48:48,861 Yeah, I'm pretty good. 565 00:48:48,861 --> 00:48:49,887 So this is your show. 566 00:48:49,887 --> 00:48:50,430 You're running it. 567 00:48:50,430 --> 00:48:51,926 Happy to do whatever you need. 568 00:48:51,970 --> 00:48:58,992 Yeah, the well, guess just kind of wrapping up one final question. 569 00:48:58,992 --> 00:49:11,295 So what would you say to a firm who is starting to map out their gen AI strategy in terms of where to get started? 570 00:49:11,295 --> 00:49:19,718 Because there's, you know, there's a lot of confusion in that space and there's a lot of different approaches I see being thrown around. 571 00:49:19,718 --> 00:49:21,638 I've always advocated for 572 00:49:21,696 --> 00:49:25,928 start with the lowest risk, highest reward use cases possible. 573 00:49:25,928 --> 00:49:29,689 And a lot of those exist on business, on the business of law side, right? 574 00:49:29,689 --> 00:49:32,901 Finance, marketing, HR, right? 575 00:49:32,901 --> 00:49:34,851 You don't have client data. 576 00:49:35,652 --> 00:49:43,936 know, it's, it's, if something goes wrong, it's not, you're not putting your, your client and your fiduciary obligations at risk. 577 00:49:43,936 --> 00:49:44,386 I don't know. 578 00:49:44,386 --> 00:49:50,228 What would you say to a firm who's starting to think about this and as to where to start? 579 00:49:50,587 --> 00:49:52,739 Sure, well, it's gonna take resources. 580 00:49:52,739 --> 00:49:55,641 So you're gonna have to make that hire. 581 00:49:55,701 --> 00:49:59,197 And I think you can find unicorns are out there. 582 00:49:59,197 --> 00:50:02,727 I provide certain services, consulting and advisory. 583 00:50:02,727 --> 00:50:12,095 know a lot of other folks that do about, you're not gonna hire someone in-house right away, but at least talk to someone who knows this process and understands the psychology 584 00:50:12,095 --> 00:50:14,417 specifically about working with law firms. 585 00:50:14,417 --> 00:50:16,198 That's number one. 586 00:50:16,398 --> 00:50:17,839 Number two is, 587 00:50:20,133 --> 00:50:29,358 Don't ever, don't expect anything whiz bang from any type of new provider, especially if it's an actual firm. 588 00:50:30,179 --> 00:50:39,864 The most disappointing conversations I have with people are like, given sort of your privacy requirements, and this is firms, but also other organizations, given your privacy 589 00:50:39,864 --> 00:50:48,089 requirements, given your budget, given your appetite for risk and innovation, you're going to have to stick with whatever it is. 590 00:50:48,933 --> 00:50:52,234 that your current software vendors are willing to provide you. 591 00:50:52,234 --> 00:50:54,265 Like what new platform should you get? 592 00:50:54,265 --> 00:50:56,965 What's your first step in AI or gen AI? 593 00:50:56,965 --> 00:51:02,187 Like talk to every one of your current relationships and see what it is that they're doing, right? 594 00:51:02,187 --> 00:51:13,060 Because it's just like, you know, I talked to, you know, the CISOs of the world and the IT managers and they're like, there's no way we're gonna keep our employees from going off 595 00:51:13,060 --> 00:51:14,240 the rails. 596 00:51:14,481 --> 00:51:16,229 lot of these platforms are just not. 597 00:51:16,229 --> 00:51:20,830 well established enough to have the proper guard rails, the SOC 2, whatever it might be. 598 00:51:20,830 --> 00:51:23,591 And so that's, that can be disappointing. 599 00:51:24,031 --> 00:51:29,953 But the flip side of that is, you know, go play in co-pilot is not the best advice either. 600 00:51:29,953 --> 00:51:38,415 So I would say starting out is, you know, be intentional about your resources, be small in your aspirations, but also in your, your wins. 601 00:51:38,415 --> 00:51:39,995 Don't shoot for a big win. 602 00:51:39,995 --> 00:51:45,905 And also, you know, talk to your core providers, which in the case of law firms, they've got the huge advantage. 603 00:51:45,905 --> 00:51:56,125 I think of having that small cohort of, you know, Thomson Reuters and all their, you know, that, that, that small cadre of legal information service providers that have been buying 604 00:51:56,125 --> 00:52:04,605 these other platforms and have been connecting them and have the budget to build them into robust solutions and to just, you know, I mean, how many people were using co-counsel, 605 00:52:04,605 --> 00:52:05,005 right? 606 00:52:05,005 --> 00:52:07,285 Like here's this thing that helps you draft. 607 00:52:07,285 --> 00:52:08,885 Like, you know, I don't want to use it. 608 00:52:08,885 --> 00:52:10,015 know how to draft a document. 609 00:52:10,015 --> 00:52:12,675 Like, well, give it to an associate. 610 00:52:12,675 --> 00:52:15,005 And if it makes them faster, it makes them faster. 611 00:52:15,005 --> 00:52:15,913 And then 612 00:52:15,931 --> 00:52:19,904 figure out your business of law and your profit margins later. 613 00:52:20,024 --> 00:52:21,355 Yeah, that's good advice. 614 00:52:21,355 --> 00:52:34,235 mean, I think it's, that's a prudent path is to, you know, maximizing that risk reward equation, talk to existing vendors that you are already established in the firm and who 615 00:52:34,235 --> 00:52:39,349 you've already made investments and commitments in and, uh, start plotting your journey there. 616 00:52:39,349 --> 00:52:40,430 think that's, 617 00:52:40,593 --> 00:52:49,537 I do a lot of talking people down and I've got decades of experience in it, know, working as a contractor, walking into somebody's building and they've got, they know exactly 618 00:52:49,537 --> 00:52:57,620 what's wrong and you're like, okay, let's talk about, you know, what are the symptoms before you try to pay me for a cure that won't work, right? 619 00:52:58,101 --> 00:52:59,007 Small wind. 620 00:52:59,007 --> 00:53:12,386 I heard this quote, I can't remember where recently about Albert Einstein said that if he had a problem that were to have the potential to end humanity and he had one hour to solve 621 00:53:12,386 --> 00:53:19,500 it, he would spend 59 minutes trying to understand the problem in one minute coming up with a solution. 622 00:53:20,581 --> 00:53:22,062 I mean, I think that's... 623 00:53:22,242 --> 00:53:24,283 That's, that's prudent advice. 624 00:53:24,343 --> 00:53:25,973 you gotta understand what you're solving for. 625 00:53:25,973 --> 00:53:30,925 And with technology, it's so easy to get, I call it ADSO instead of ADD. 626 00:53:30,925 --> 00:53:32,976 It's a 10 attention deficit. 627 00:53:32,976 --> 00:53:34,546 Ooh, shiny object. 628 00:53:34,547 --> 00:53:40,649 You know, it's a, it's, it's one step above, um, ADD. 629 00:53:40,649 --> 00:53:50,773 So, well, um, how, how to wrap it up here, how do folks get in touch with you, learn more about the, the, the work that you do or the writing that you do, how, to, how to folks 630 00:53:50,773 --> 00:53:51,515 find you. 631 00:53:51,515 --> 00:53:54,988 Sure, absolutely on LinkedIn, definitely not on Twitter. 632 00:53:54,988 --> 00:54:02,454 I'll start with that website, gammawave.ai, know, email, eric at gammawave.ai. 633 00:54:02,454 --> 00:54:08,349 But I'm findable on LinkedIn and I think there's only two people in the world that have my first and last name. 634 00:54:08,349 --> 00:54:11,191 So honestly, you can Google me and find me. 635 00:54:11,406 --> 00:54:14,930 That's amazing because I have a much more unique first name and last name. 636 00:54:14,930 --> 00:54:19,475 It feels like Theodore Theodoropoulos and there's several. 637 00:54:19,475 --> 00:54:21,935 So that's surprising. 638 00:54:21,935 --> 00:54:23,820 Greece is a big country, if I'm guessing right. 639 00:54:23,820 --> 00:54:24,792 Yeah. 640 00:54:24,835 --> 00:54:25,206 All right. 641 00:54:25,206 --> 00:54:26,010 Well, good stuff, man. 642 00:54:26,010 --> 00:54:29,425 I really appreciate you spending some time and hope to see you soon. 643 00:54:29,425 --> 00:54:29,887 You bet. 644 00:54:29,887 --> 00:54:30,482 Thanks for having me. 645 00:54:30,482 --> 00:54:31,414 Really appreciate it. 646 00:54:31,414 --> 00:54:32,736 All right, take care. -->