IMR337: Legal AI Agents, Micro-Products, Decreasing Token Cost, Law Firm USPs of the Future

0:00Teaser Marc hier, moin, kurzer Hinweis, das hier ist eine Episode im Rahmen der zweiten Staffel im Rechtsmarkt, das Special bei Irgendwas mit Recht, wo wir gemeinsam mit Expertinnen und Experten die aktuellen Entwicklungen bezüglich des juristischen Berufsbilds, aber auch allgemeiner Trends beleuchten. If you0027re interested in the show notes, there are the links to the roadshow of the Center on the Legal Profession of the Bucerius Law School, seven German cities, jeweils two hours over the afternoon.

0:30Teaser It0027s just about it, if you0027re still in the training, you0027ll find out more information about the future, including in the future of the Bucerius CLP. We hope to see you again, until then.

0:55Marc Herzlich willkommen zu einer neuen Episode Irgendwas mit Recht. Heute haben wir eine Premiere. Das ist die erste Episode, die wir auf Englisch machen. Wir haben nämlich einen sehr spannenden Gast hier. Den stelle ich euch jetzt gleich aber vor, after I switch to English. My name is still Mark Orendorf. Very nice to meet you and very nice to welcome you here in the studio. Dan Katz. Hello, Dan.

1:19Daniel Katz Hello. It0027s good to be here. It0027s a great honor to be here. the first English Language International version of the show.

1:26Marc Yeah, that0027s very nice. Maybe introduce yourself a little bit. What do you do?

1:30Daniel Katz Well, I wear a few hats in life, but one is I0027m a professor at the Illinois Institute of Technology, in particular, the Law School of the Illinois Institute of Technology, which is called the Chicago-Kent College of Law. So that0027s my sort of primary academic affiliation, but I0027m also work fractionally here at the Bucerius Center for Legal Technology and Data Science.

1:54Daniel Katz And I0027ve been working with Bucerius for a number of years now in that capacity. And so we0027ve run a number of things, including Bucerius Legal Tech Essentials. We have a summer program in legal operations and artificial intelligence and law, and in a variety of other kind of capacities. So we0027ve done a few things over the years also with like the CLP, the Center for the Legal.

2:20Marc And that0027s obviously how we met because I0027m a director on the Bucerius CLP, apart from some other hats, as you just said. So that0027s kind of the disclaimer. This is not Bucerius podcast. It0027s just coincidence that we met over that connection. Let0027s start a little bit. You studied law. You0027re a lawyer by training, right? How did you end up where you are today? Okay.

2:41Daniel Katz Well, I got very focused in law school, I would say, on the notion that people would make statements about the world and about the operation of the law, and they had no evidence or scientific backing for what they were saying. So they might be right, they might be wrong, but we basically were settling everything as an essay contest.

3:04Daniel Katz And that seemed very at odds with the rest of the world, which was mostly run by evidence and scientific progress. And we didn0027t really see that in law. In the American Academy, particularly starting at the beginning of this century, we started to see empiricism come into the field. So people would say, hey, you know, if we0027re going to select, say, this legal rule versus that legal rule, we ought to sort of evaluate whether the claims people made about how that rule would work really play out in reality that way.

3:37Daniel Katz And so, this sort of rise of empiricism in law, I was very interested in. And so, I pursued that with getting a PhD in political science, but it was at the University of Michigan. It was a very quantitative program. And so, that sort of put me on a journey, which I can fill in a little bit more if you0027d like.

3:58Daniel Katz But that was kind of the, I would say, the original impetus that kind of put me down this path.

4:04Marc So if I understand you correctly, that also gave you a little bit of the research toolkit that you need for what you do, and you brought that back into the world of law, so to speak.

4:13Daniel Katz Yeah, that0027s right. That0027s right. I mean, so in political science, there is a tradition, particularly, you know, where I was in studying courts quantitatively. So trying to understand how and why kind of the decisions are made. And that0027s at odds with the traditional way that courts were studied in the legal academy.

4:37Daniel Katz And again, those two roads have come together more over the last 25 years. But that historically, there was a big divide between the fields. And, again, I think overall, if you take the big picture, the legal academy has become more like the rest of, say, the universities and less of a kind of exceptionalism is taking place.

5:03Daniel Katz But it0027s a slow process because, you know, obviously people don0027t generally switch their paradigms. That0027s a rare idea. I think Max Planck had a famous quote about that idea, about the pace of scientific progress.

5:16Marc Maybe we can get into a little bit on what you researched and what you did in the last many years already. Maybe let0027s jump directly to one of the most famous papers that you were part of. I0027m getting at a GPT-Past-The-Bar exam.

5:33Daniel Katz Yeah, GPT-4 came out in March of 2023.

5:36Marc Yeah.

5:36Daniel Katz And obviously, there0027s been subsequent models that have come out. And so, at the end of 22, two, my colleague, Mike Bomarito and I evaluated what was then the leading model, which was GPT three and a half on just a multiple choice portion of the U.S. bar exam. And there0027s not really a U.S. bar exam, each state, but 37 of the states basically follow this version of the exam.

6:04Daniel Katz So you can, you know, approximately say that0027s the. US bar exam. So we evaluated just the multiple choice portion and it got about half the questions right, which is not enough to pass, but it was well in excess of anything a computer had ever been capable of doing by a long shot. And so it was a sort of harbinger of what was to come.

6:23Daniel Katz So kind of thereafter, I ended up kind of connecting up with the folks at OpenAI and some folks we knew at Case Text who were working pretty closely with OpenAI at that time and were given, you know, sort of special access in advance of the public to GPT-4, and we did the evaluation. And it was, you know, it was really something that sort of, there0027s, it0027s like a magic moment when you first see the capability increase, and it was significant.

6:52Daniel Katz And it was obvious that, you know, when you just saw the quality, the results, and of course, it couldn0027t do everything, and it didn0027t get everything right. But it was a serious, serious step up the hill there from what anything we had seen before. And actually, three and a half was pretty incredible in its own right.

7:10Daniel Katz So we live in amazing times.

7:12Marc Did you already have an idea that this would happen when you wrote the paper in fall of 22?

7:21Daniel Katz Yes.

7:21Marc Yeah.

7:22Daniel Katz We wrote it right in there, zero to 18 months. We said it right in the paper.

7:26Marc Yeah, yeah.

7:26Daniel Katz You know, you see the capability increase that had already happened. The multiple choice seemed like an absolute certainty. And then the more incredible part was the essays. I think somebody actually just recently evaluated the German bar exam, which, of course, I don0027t know. Until recently, you had to handwrite the whole thing. So I don0027t know if they got a robot to handwrite the exam to make it realistic.

7:53Daniel Katz But the robot0027s hand fell off because it was trying to... I feel like people who took the historic German bar exam, they needed to lift weights or something to get their arm to be strong enough to just keep writing. I know they don0027t do that now. It feels like, you know, they say that the legal profession is slow to adopt new technology.

8:16Daniel Katz I think that handwriting the exam is sort of an exemplar of that idea.

8:20Marc Yeah, that0027s true. Okay, so let0027s go to that intersection again of LLMs and legal knowledge and work of lawyers and judges and so on. From your perspective, where do we stand today? How can we use LLMs in legal practice?

8:39Daniel Katz Well, I think you have to think about the progression as when these models really hit the mainstream for people, They, first of all, they weren0027t enterprise ready. So people0027s typical first experience was ChatGPT. ChatGPT was literally taking people0027s queries and putting them back in the model and was very explicit about that.

9:03Daniel Katz And so obviously a law firm, client confidentiality is the hallmark of the legal profession. And so the notion that you could use these, I mean, they weren0027t even really ready to be used. And certainly, by the way, beyond the legal profession, lots of organizations are not going to be very hot on the notion of having their data used and fed back into a model.

9:24Daniel Katz So there was that famous example of the engineers at Samsung putting trade secret code into the model. And then obviously, it would autocomplete it back out because it was being fed back inside of ChatGPT. So yeah, overall, the stuff wasn0027t ready. And so a bunch of things had to be built out. And they have been.

9:44Daniel Katz And so that0027s not the capabilities, but it0027s the deployment and accessibility. So that0027s kind of one dimension. And then there0027s the pure capabilities. I mean, the generation of models that came out when GPT-4 came out, they weren0027t reasoning models. They are text completion models. And so there0027s just pure limitations there.

10:10Daniel Katz So now we have these reasoning models. We have the ability to connect agentic capabilities, basically tools and planning. And when you connect tools and planning, that gets closer to the type of knowledge work that, especially at the top end of the profession.

10:26Marc Yeah, and also I think one thing that was misunderstood by many in the beginning was that these models were not necessarily trained A, on very specific legal knowledge. It was more or less the legal knowledge that was on the web and in some books maybe. At least for Germany, we know that it wasn0027t trained on legal databases and so on and so forth. Right.

10:48Marc Hopefully is never trained on very specific client data, but it can, of course, be fed with client data if you get that whole setup, the whole tech stack right within your firm nowadays. So I think the capabilities are not even fully discovered yet today. What do you think about that?

11:09Daniel Katz No, I agree. I very much agree. We don0027t have the information. Of course, if you0027re building a model that0027s a general form model to solve lots of different problems, law is not going to be your focus. And so it0027s almost by, I wouldn0027t say by accident, but it0027s incidental that the notion is incidental that the models could perform on any of these legal tasks.

11:31Daniel Katz German legal sources are obviously mostly behind paywalls. And so the ability to consume, not that that0027s necessarily seem to stop some of the AI companies. But, you know, I saw recently that there was a deal between a startup and the Beck.

11:51Marc Yeah, Noctua and Beck deal for $87 million.

11:54Daniel Katz Yeah, yeah. And, you know, obviously that0027s the type of information one would need to build a higher fidelity model for German legal sources. And there0027s parallel examples around the world, obviously, of these knowledge repositories that are... And you0027re right to say, of course, there0027s another source of information, which is all the legal work product that0027s been produced and is sitting in a document management system of a law firm or the legal department. Yeah.

12:20Marc True. Speaking from a law firm0027s perspective, should they build or buy? Of course, it depends. But what is criteria?

12:28Daniel Katz I would execute a mixed strategy over those two things. It0027s sort of hard to tell a story of differentiation if you0027re buying the exact tool that the next firm down the street also has.

12:39Marc It0027s all a level playing field if you do.

12:41Daniel Katz Yeah, but there0027s also other things. It0027s not clear why... You need differentiation. For example, I wouldn0027t advise a law firm to start building their own legal research product or something like that. That exists as a standalone for a reason. But then there0027s specialty work, which the firm, particularly a firm that has focus expertise or demonstrable expertise in a particular area, the extension of those expertise is a software layer, So we0027re seeing that. We0027re seeing law firms beginning to try to build things.

13:18Daniel Katz I think one of the most important things that sits behind that is that the unit economics of building have come down substantially. And there0027s often this discussion of the jagged frontier of capabilities. So that0027s the notion that, hey, a model can do something incredible like the bar exam and then can0027t do a basic task and then can do another incredible task and then fails on, I don0027t know, a math problem or something like this.

13:42Daniel Katz And that0027s a jagged frontier. But that frontier also includes, should be thought of as to include the economics too, which is you could say, look, a machine0027s capable of this if we do a bunch of engineering in particular, but it0027s the problem0027s just not worth the effort. And it0027s when those two gating criteria are cleared, then it starts to be an attractive candidate.

14:01Daniel Katz And so probably the most important thing that0027s happened is just the total cost decline. So when GPT-4 came out, it cost roughly 200 US dollars to process a million tokens. And now it0027s something like $2 to process a million tokens. And so 100x decline in 30 months or 24 months or whatever is incredible. It0027s just incredible.

14:25Marc And we0027ve never seen that really before.

14:26Daniel Katz Yeah. I mean, technology products tend to have a decline, like a decay cycle, but not on that timeline and not at that rate. And so I guess it0027s starting to pencil out to be a reasonable decision to think about doing building, at least in certain areas. And so I would expect to see quite a bit more of these, I would call them micro products.

14:49Daniel Katz They0027re not infinitely scalable, but they solve a real problem and they have a more addressable market. And particularly when they0027re connected to services that the firm is also providing, it0027s a bundle of product and service put together.

15:05Marc Yeah, and I think that is really where the future is going to be, where we have very standardized nowadays services, then products, and then really services on top. And also, it doesn0027t necessarily mean the end of the billable hour, eh? It could even mean that the billables are even a bit more expensive in the future, because they are that little rest of 15-20 % of work that is really so highly skilled that you can even charge more for that.

15:32Daniel Katz Yeah, I think, well, first of all, I mean, people have called for the death of the billable hour. My position is people, it will go away the second somebody figures out how to make money a different way, more effectively. And until that time, you can expect. And one of the other problems we have is that it0027s hard to measure the quality of things.

15:52Daniel Katz So we use time, we use inputs as the measuring stick instead of outputs. Now that0027s not true everywhere, but that is what we0027ve fallen back on as a kind of proxy for quality. But it don0027t.

16:08Marc Yeah, value-based billing is really one of the toughest challenges to solve.

16:13Daniel Katz Now, we see it in some areas around the world where there0027s things like contingency fees, and so a percentage of the recovery. Now, there0027s a lot of alignment there between the client and their lawyer when they0027re both essentially have a stake in the outcome directly. It can produce other issues, but that0027s a place, but that0027s harder to, it0027s not even always allowed everywhere. but even when it is, it0027s hard to define criteria sometimes in advance.

16:41Daniel Katz So for example, if you work a transaction, maybe walking away from the transaction is the right business decision, but is that viewed as a success or a failure? I mean, getting into a bad transaction is a bad, you know, avoiding that is success, but defining that in advance is difficult. So we fall back on inputs measured in time.

17:02Marc Well, what you could define in a scenario like that is to enable the business leaders to make a decision right so if the outcome of the legal analysis would be the that the business maker can make a sound decision well then by definition it could always be a success sure sure yeah it doesn0027t it0027s.

17:22Daniel Katz Not quite the as clean as you know we have a claim and you get a percentage.

17:27Marc Of the recovery that0027s a very.

17:29Daniel Katz Easy thing to measure and execute and calculate. And so, again, there0027s a reason that the billable hour exists. And now, if we get into the world where we0027re selling bundles of software and services, then I think if a firm builds specialty software, they should be paid for it. It should not be... This is going to fall into... Is it CapEx or OpEx?

17:52Daniel Katz Is it a capital expenditure or an operating expenditure? Now, there0027s a tendency in the clients to want to make everything an operating expenditure in other words they say well listen, I0027m not going to pay your light bill. I0027m not going to pay for your administrative assistance and your lease of your office or whatever it is. And I understand that.

18:11Daniel Katz But if a person builds a specialty software product that enables services, they should be able to have a license fee or a utilization fee or something to recoup the investment. Because if you0027re the client, you want your firms to try to build these things. You should want that. But there has to be an economic model that0027s going to work for that.

18:30Daniel Katz And I think forcing it all into an operating expense is not, it0027s actually going to restrict the amount of innovation and lower the quality of your service. So you should be creating reasonable incentives. Now, obviously what the exact price ought to be, that0027s to be determined, but above zero.

18:48Marc Obviously.

18:49Daniel Katz Yeah.

18:50Marc We are recording this show alongside the Hamburg Legal Hackathon. And we had a little lecture this morning where one of the guest speakers presented a solution of searching through 1 million a4 pages of documents and that cost them roughly coming back to the decrease of cost cost them roughly 200 euros and someone asked the question about data security and data protection and so on and so forth and the bottom line was well of course law firms in their operations even if they don0027t build these specific solutions that you just mentioned and they just use it as internal you call it micro products they will have their own server farm and they will also really be really need to be tech savvy even more than

19:38Marc they are today do you agree with that kind of vision.

19:42Daniel Katz Yes i do obviously i i agree with that it0027s a question of i think, It0027s a little bit of a challenge. Just the one part that0027s a challenge to me is who exactly in the organization and to what extent that is true is really kind of the question. So I don0027t know if every single person has to go to the maximum extent on this question.

20:06Daniel Katz I think people have to recognize the value in it, enable it to occur. But if you0027re running a law firm, it0027s probably not rational to have... I mean, everybody needs a minimum amount of competency.

20:20Marc Yeah.

20:21Daniel Katz I do think you have to grade it out a little bit.

20:25Marc But does that mean law firms will have product teams who work together with their tech teams in the future?

20:33Daniel Katz I would think so. So if you say, look, okay, I accept the proposition that we need to have our own micro products, and you say, okay, who exactly is going to build them? Well, I think the idea that a pure product team isn0027t going to be in the position to understand the nuances of the legal problems.

20:52Daniel Katz And so the subject matter expert is going to have to participate in it. I mean, the firms I personally work with, I make this very clear to them. You have to have respect for the notion that they have real expertise, and it has to be manifest in any of these micro products. And so they have to have a seat at the table. so we have two cultures kind of putting together and everybody thinks they0027re smarter than the other person so the lawyers obviously think they0027re you know unassailable you know tech the people in tech kind of like tend to dismiss the details sometimes of some of these problems and why they0027re hard and so there just has to be a healthy respect for both the value of each the expertise and that each party brings

21:38Daniel Katz to the table that that is a successful quote that0027ll be a successful story if both parties have that mindset.

21:44Marc And I think that0027s quite hard to establish because that kind of comes down to the current pyramid scheme we see in law firms where we have partners and associates and then of course leverage of one to two, one to four, whatever. This picture of the pyramid if you just look at it at this very basic model first of all it doesn0027t even include other stuff that is not an associate or partner then you can include them but then in the future if you think of what does the law firm sell nowadays it0027s just in quotation quote unquote just legal services if that becomes products and services that also means the pyramid needs to change somehow, right?

22:29Daniel Katz Yeah, it0027s going to have to change. It has to change. Now, if one of the issues has been the sharing of profits with so-called non-lawyers, and that0027s been a hotly debated issue for the entire time I0027ve been on the scene and for many, many, many years, in all, basically everywhere. So obviously, it0027s a unique thing. If you look at the rest of the economy, it0027s a unique idea that there0027s this class of folks who can0027t be part of the party.

23:01Daniel Katz They can0027t reap the gains of the upside of things. So one way would be to change that. That0027s difficult. The other way is to change the skill set of people such that they can participate meaningfully. The cost and the difficulty of building MVP products, it0027s like falling off a cliff. I mean, obviously, we0027re all vibe coders now.

23:26Daniel Katz And you0027re not going to vibe code your way to an enterprise product, but you can build a pretty effective MVP relatively quickly. And then the thing is like, what do I need to take this into production? Well, you0027re going to need some other things. But... But let0027s just start with that notion. And I think with these products, by the way, there is what they would call a land grab.

23:46Daniel Katz In other words, if you can get to market with a product that is also reasonably connected to your services and acquire clients on that basis, you0027re sitting in a very good position. The person who goes second, third, fourth, fifth, it0027s not that they could never displace whoever0027s inferred, but it0027s harder to do that than to just be there kind of at the front.

24:14Daniel Katz And so I think that impulse is going to hopefully lead to more folks going down this path. I0027m seeing it in other markets around the world, not just like the US, the UK, and other markets, you0027re seeing, again, firms building these capabilities.

24:34Marc Okay, so after we talked a bit about technical development, about the evolution of business models, let0027s focus, as the third part of this show, on the talent and the young lawyers that come into the firms in the future. Nowadays, they learn law, in law school at least. Most likely, they will also learn the same things, at least from a German perspective, in the next five to ten years.

24:57Marc There won0027t be that much change. what would you advise a young lawyer either a law student either wanting to become a judge or a lawyer or work for a public administration to learn as well aside from law.

25:13Daniel Katz Well i do think technical competency is going to be an important differentiator out there i mean one of the things that if you want a shortcut you0027re well how do i say this the right way if you show up on the first day at a job, you0027re never going to have more experience than the people who you work for.

25:31Daniel Katz But you may have a dimension of knowledge and expertise that they don0027t have. And it0027s a sort of cheat code to be at the big table. Yeah.

25:41Marc And don0027t rely on the fact that you are Gen Z or Alpha, because this is just a stereotype.

25:48Daniel Katz No, it is. It0027s a stereotype. You have to really know what you0027re doing. And the point is, you have time to make these investments. You0027ll have less time in the future when you0027re working a job than you do now to make these intellectual investments. If you just sort of zoom out a little from wherever you are in your world and look at the broader world, you0027re going to say, this is kind of the direction of travel.

26:11Daniel Katz And it0027s kind of, if you look, I think seven of the nine most valuable companies in the world are in and around the. Either on the hardware or software side are tech companies. And a decent percentage of the top 500 companies are involved in this. And so if that0027s where the commercial world is, especially if you0027re working at a law firm, a bigger law firm in particular, that0027s who your clients are.

26:39Daniel Katz So if that0027s what the world becomes, that0027s what your world becomes by extension. And more broadly, you know, this is just a, just come back, like, what is your differentiation? It can0027t be that you know more, I mean, it can be that you know more law than somebody, but these are fine-grained distinctions, and they0027re hard to acquire, and they often take a lot of time.

27:00Daniel Katz And just some abilities in this area, when most people in the field are pretty, to be direct about it, fairly weak on these topics, you0027re immediately could be in the top X percent of people in the field with just a reasonable amount of time invested. And so I would encourage people to consider that because it could make all the difference for you.

27:23Marc Do you sometimes hear that people are afraid of being replaced by AI in the field of law?

27:30Daniel Katz I think that generalizes to a broader societal concern. But yes, that is a subset of the problem is... Now, listen, let0027s just be very direct about it. Nobody in the public is going to care if we have less lawyers. That0027s the last thing on the agenda, probably, of concerns in society. It0027s pretty low on the list.

27:50Daniel Katz So let0027s just start with that. No one0027s going to have a funeral for on this. In the time since I finished law school, the present, I look at electronic discovery is less of an issue, but investigations are still an area that people used to look through documents with eyes on review. They0027d look at every document from the first to the last in an investigation or in a discovery matter.

28:18Daniel Katz And that0027s basically gone almost exclusively the way of machines now, at least for first pass review. And yet the field has persisted. People still sue one another. People still have disputes. And so if a person said back years ago, hey, what I want to do, no one would have ever said this, is I want to do document review for the rest of my career.

28:41Daniel Katz That wasn0027t going to really work as a point of view. Now, again, remember, nobody wanted to do that. Nobody liked it in the first place or nobody. Most people didn0027t like it. It was considered a fairly mind-numbing thing to do. And so I think people have to have the flexibility to recognize that there is going to be automation of certain tasks and you0027re going to have to remain flexible.

29:03Daniel Katz And that0027s just kind of how it0027s going to be. So that0027s kind of the background. Now, it0027s hard for people to see the things that are being created by these same capabilities. Now, just as an easy way to see the point, every bad actor out there also has access to these capabilities and they0027re going to create mayhem with it.

29:23Daniel Katz There0027s going to be tons of cyber incidents around using these capabilities because you can automate penetration, kind of exercises. There already is a degree of that. There0027ll be even more. There0027s just going to be a bunch of... Again, it0027s hard to see these things. People see the doors closing and not the doors that are opening.

29:43Daniel Katz It0027s hard to say what the net labor market effect is going to be. But I think there0027s just more of certain things going to be created. In the end, if there0027s more or less lawyers, again, no one0027s going to really care if there are fewer lawyers in society. Arguably, that might even be a better thing for society. But I know that0027s a statement against interests, obviously.

30:01Daniel Katz So yeah, I think that0027s something that people need to focus on their own situation. And in some sense, you control your own situation. You have the opportunity to make investments in knowledge and expertise. And that knowledge and expertise has to be related to what the world is and what the world is becoming.

30:21Marc Well, we are a bit lucky, quote unquote, in Germany, because after we finished law school, the debt that you collect isn0027t as high as in other parts of the world.

30:31Daniel Katz True.

30:31Marc So maybe people are even more flexible in terms of changing their career and, you know, studying for one more year on top of that and so on and so forth. So that0027s also quite good news. But maybe finishing with that question in today0027s podcast, what is the mindset that you would advise people to train or look into or find in themselves, depending on how you would see it, how this works, that you need to be successful in 2035, 2040 in our realm?

31:05Daniel Katz Well, I mean, it0027s a cliche, but I mean, if you think of yourself as an entrepreneur and that you0027re looking for market opportunities, I mean, there0027s a book that was written a number of years ago. It0027s called To Sell is Human. And I think that you0027re always selling, in some sense, yourself or what have you.

31:26Daniel Katz And you have to look for the opportunities that are being created and not get stuck or static or sit in stasis because otherwise you0027re going to lose the... Again, just coming back, I0027ve said it now, this will be the third time. No one cares if a lawyer doesn0027t have a job. That0027s just not a major concern of any society. and so it0027s up to you to basically stay relevant.

31:52Daniel Katz You0027re right to say that people don0027t have the debt to the same extent as they would say in America, but you have opportunity costs in your life. So you should be putting your time... Investments of time and efforts towards things that are most rewarding or most fruitful for you. And technology is always a double-edged sword because we0027ve had, obviously, if you take the last 50 years, we0027ve had amazing technical developments.

32:20Daniel Katz Not all of them have been positive, uniformly positive, let0027s say. There0027s usually two sides to the coin. But yeah, I think your mindset has to be flexible to the changes of the world. You need to get yourself on an information diet of diverse sources of information so that you can do trend spotting and understand where the world is heading.

32:43Daniel Katz And obviously, people live in these filter bubbles, or they have these feelings like, what I0027m going to do is put my head down, and I0027m just going to... And this is a classic kind of thing you see with law. I0027m just going to do law stuff. I0027m going to put my head done, and I0027m just going to lean into that.

32:59Daniel Katz And I understand that impulse completely. The only issue is like, if we really zoom out, even if you really do law and you really get it down, all of this is related to what the world is. What is law even for? Law is for managing societies, solving disputes, kind of creating a functional world. And so, it sometimes feels like this, but it does not operate atomistically from the rest of society. It0027s supposed to be part of all of that.

33:29Daniel Katz So if you0027re going to play a role in shaping all of that, you have to have some information, which kind of brings me back to where I started here. The thing that I didn0027t like when I was in law school was people talked about the world as if there was no ability to learn about it. We would solve things as an essay contest and not using the scientific method.

33:49Daniel Katz It felt like pre-enlightenment thinking.

33:51Marc Mm-hmm.

33:52Daniel Katz And I think it0027s a great development that the field of law has begun to look more like the science is. I know not everybody agrees with that, but let me just put it to you this way. If you went to your doctor and you found out that medicine was run on the basis of who wrote the best essay, find another doctor.

34:10Marc Professor Dan Katz, thank you very much for taking the time today.

34:13Daniel Katz Thanks for having me.