• Predicting judges’ decisions using personality insights and past data. 0:03
  • Speaker 1 describes a product liability case in front of a judge in Alabama, where they tried to predict the judge’s decision by analyzing past decisions.
  • Speaker 1 explains how they used past experiences with the judge to make predictions, but found it challenging to divine the judge’s true intentions.
  • The speaker discusses the limitations of judges’ past opinions in predicting their future rulings.
  • The speaker proposes using personality traits and life experiences to better understand judges’ perspectives.
• Predicting judges’ rulings based on case characteristics and judge profiles. 4:19
  • Analyzed judge characteristics, including gender, education, work experience, and more, to predict case outcomes.
  • Speaker 1 explains how their software analyzes case details and judge reputation to predict motion to dismiss outcomes (85% accuracy)
  • Speaker 2 questions the accuracy of predicting judge rulings based on case details and litigant information
• Using AI to predict federal judge’s rulings based on their past decisions and demographic information. 9:15
  • Speaker 2 questions the judges’ management of their docket and plaintiff-oriented approach.
  • Analyzing judges’ appointments reveals counterintuitive results, with female Democrats being as pro-corporate as Republicans.
  • Speaker 1 discusses the limitations of predicting judge outcomes based on their background and experience.
  • Speaker 2 suggests that using AI to analyze judges’ past decisions and predict their future rulings can improve litigants’ chances of success.
• Using AI to predict court outcomes, with a focus on motions to dismiss. 15:01
  • Company developing AI software to predict judges’ decisions in motions to dismiss, with a focus on state court data.
  • The speaker correctly anticipated 20 out of 25 case outcomes using their model.
  • The speaker identified 11 cases that were dismissed, which could be turned into wins with proper jurisdictional analysis.
• Using AI to predict judge’s decisions in patent cases. 19:23
  • Speaker 1: Judges’ decisions are most important in predicting case outcomes.
  • Speaker 2: Analytics can help lawyers make informed decisions about jurisdictions and venues.
  • Speaker 1: Briefs don’t improve prediction, despite AI analysis (15% gap)
  • Speaker 2: Recurring legal issues, such as federal preemption, persist (judge variability)
• Predicting judge decisions using AI and big data. 23:38
  • Speaker 1 explains how they developed an AI capability to predict judge decisions by analyzing data from past cases.
  • Speaker 1 describes how their approach involves breaking down judges into component parts and finding commonality between them.
  • Speaker 1 proposes using data analytics to predict judges’ decisions by analyzing their personal characteristics and external factors.
  • Dan Rabinowitz, CEO of Predictive PR, discusses his company’s predictive software that can accurately predict federal judges’ rulings to 85% accuracy.
  • John Quinn hosts the podcast “Law Disrupted” and interviews Dan Rabinowitz about the potential applications of Predictive PR’s technology.

SUMMARY KEYWORDS

judge, dismiss, motion, case, analysis, decisions, data, briefs, approach, lawyers, predictions, fascinating, simply, recurring, models, litigator, predict, plaintiff, ultimately, arguments

SPEAKERS

Dan Rabinowitz, John Quinn

John Quinn  00:03

This is John Quinn and this is law disrupted. And today we’re talking with Dan Rabinowitz, who has a business Pre/Dicta, which is really fascinating. I think it’ll be fascinating to all litigators. mean every litigator when you have something going before a court a motion, you’re trying a case you’re going to a court courts gonna make a decision. You’re always thinking about, of course, what might motivate that court? How is the court likely to, to rule? What things might make a difference to that particular judge? And Dan’s business, put some science around that helps predict how judges are likely to rule? And is that a fair summary? Or correct me if I’ve got that wrong? No, that’s

Dan Rabinowitz  00:49

a very good summary.

John Quinn  00:50

So tell us about this business. It sounds so fascinating. Sure.

Dan Rabinowitz  00:54

So you know, I remember when I was a young associate, and we were practicing, and we had a, we had a product’s life liability case in front of a judge, the judge was in Alabama, and we had a motion to dismiss pending. And my task was to go ahead and collect all the cases or all the decisions the judge had, or motion to dismiss and tried to determine whether our, whether the party that we represented, like what effect that would have on the judge, what impact would it have we we had, you know, all of us had sat down, we had written a great brief, we had great arguments. But the question is, is what is the judge going to do and whether or not we can divine, any information about that from past decisions. Now, as you would imagine, there were only a handful of decisions, none of which were actually relevant in the sense that they didn’t address the specifics of our case, which was we had a very large pharmaceutical client. And this was a products case, and it was an individual on the other side, we didn’t know whether the judge would be sympathetic to any of that, or whether or not the judge would be more skeptical of our firm, which was, you know, a large international firm, this was an Alabama, you know, what, how would the judge view that, and then, you know, sometimes people go ahead to their past experience in front of any given judge and try to make predictions based on that. But the limitation of that, of course, is judges here, you know, dozens of cases, cases, potentially a day. And even if you have, if you will, you know, consistent track record in front of the judge, think even even, you know, a litigator his practice over 20 years, it’s 2530 appearances via for the judge and and you think that the judge actually had, you know, 4000 cases or motions in front of them, or 10,000. So it’s really hard to use that small dataset to extrapolate what the judge and how the judge is going to perceive the specifics of your case. So bunch of years ago, I had an idea that perhaps one approach is to step aside from the law and the facts. And instead, look at the person, look at the personality, look at who the judges, because we all understand that the way that human beings are programmed, there are different life experiences, that we have different ways that different perspectives that we have on a variety of of different aspects of life. And in order to determine how those might influence us, you really have to know us as a person. And so what we decided to do was to take two different pieces of data that we had available. The first is to go ahead and take a look at what the judge has done in the past. But, of course, as I mentioned, the limitation is that their opinions don’t really provide a lot of insight into how they’re going to react to the particulars. And also the limits that the severe limitation of the opinions is that when it comes to motions, judges write opinions, and around 2% of all motions that that they hear, but 98% There is no evidence, you know, there is no record,

John Quinn  03:53

they’ll just be a documentary granted or denied, basically,

03:56

exactly. And there are some, you know, that try to extrapolate from that. But of course, you know, even basics, basic knowledge of statistics tells you that, you know, pass history is not it really indicative of how the judge is gonna rule here 75% In the past, does, that means you actually have a 75% chance. And the reason for that is because we don’t know all the variables that put you in a 25 or put you in the in the 75. So we got went ahead and looked at all those decisions, but we decided the way to look at those was to try to classify, categorize and understand the parties and the attorneys that were involved in those decisions. So once we did that, we then took that human piece of it and we looked at who the judges we started modeling, who the judges you know, politics, gender, net worth, work experience, educational experience, geography, around a couple of dozen others. So

John Quinn  04:52

how many how many total datasets would you be able to pull together typically on an average on a judge?

04:58

So in terms of Total datasets, in other words, picking particular datasets, we probably looked at, you know, ultimately, I would say it’s six large datasets that that relate to judges. And from those we we call, because not all that is necessarily relevant or, or even, you know, even tangentially applicable, but then ultimately that that created a model where were those characteristics were included. And we have around, as I said, it’s, it’s probably six or seven dozen different characteristics, some of them matter. In some instances, some of them don’t matter. But that’s all part of the analytics. So

John Quinn  05:34

these would be for example, gender, locality, political affiliation, ethnicity, things like that.

05:46

Yeah, education, work experience, were they in private practice? Are they you know, in public service, where they went to undergrad, you know, where they’re born, where they live, all of that comes together. Because if you think about how Google is so successful in placing those ads that we almost assume that they’re listening in on our conversations, it’s because they have all that information. They know where we live, they, they know our income, they know who our friends are, they know where we went to school, and so on. And that’s how you’re able to create highly accurate predictions. And then the question is, though, how those high how you then understand the judge with regard to the characteristics of the case. So that’s where those historic decisions and the classifications and categorizations come in. And then from there, because all the work is done on the back end, that is in creating those models, and creating both the judicial models and analyzing and processing the data, in order to see patterns and trends that are otherwise invisible to attorneys. Our software is incredibly straightforward. We only require the case number. And to be clear, not only do you not have to upload your briefs or any other information for us. When you do that, we don’t look at your briefs, we’re analyzing a motion. Instead, it’s those other factors that ultimately will determine how the judge will rule. And through our algorithmic models, and through the API that we use, based on our back testing with just the case number, we are correct, 85% of the time. That’s

John Quinn  07:23

amazing. is astounding, you know nothing about the substance of the case, you know nothing about what you have, I assume that one of the factors is what’s the nature of the motion is this a motion for summary judgment, that would be one factor that you would include in, in your model.

07:40

So So currently, we focus on motions to dismiss right now we have other capabilities as it relates to summary judgment. But in terms of the actual predictive capabilities, right now, where this where the technology is, is, is that the motion to dismiss, but certainly, the motion to dismiss one factor that we consider is the nature of suit. But even that is somewhat amorphous. Because, number one, you can only enter one nature of suit in, in in the federal system, and even in most state systems. So that’s number one, or number two, there, there’s no one that’s actually checking the accuracy of that nature of suit. So that’s certainly one component. That is, if you will get to the facts of the law. But then all the rest that we’re looking at are those other unique characteristics, those other influencers that are involved in the case? You

John Quinn  08:25

also look at characteristics of the litigants, the parties and the lawyers, in addition to all the things you described about the judge?

08:31

Yes, exactly. We are trying to match those, again, to see the patterns and trends in the judges past decisions, right, how they dealt with that individual, how they dealt with a particular lawyer type, or particular lawyer firm type. And the combination of all those together, and that is, that is specific to the case at hand. And then how that’s influenced that judge in the past.

John Quinn  08:57

I guess the idea that you could predict how judges could rule on a motion to dismiss May, to some degree, would not be surprising to a lot of litigators, because judges get reputations about, you know, what’s their belief and jury trials? how aggressively do they manage their docket? What’s, you know, how plaintiff oriented? Are they do they believe, liberal that they have sort of liberal ideas when it comes to pleadings? I mean, that strikes me is one motion that is particularly susceptible to this type of analysis.

09:34

Well, it isn’t, isn’t because if you simply look at, for example, liberal or conservative, there’s only one factor that that ultimately forms us as human beings and highly reductive and then if you start overlaying other aspects, obvious aspects of of humans. So if you have a liberal that for example, went to a conservative college or conservative law school, how does you know they went to and these are broad generalities University Chicago versus, you know, Berkeley, but they’re a Republican, and maybe they live in a particular neighborhood that you wouldn’t necessarily associate necessarily associated with conservative values. And how do they deal with truthfully deal with, you know, individuals? Are they totally sympathetic individuals? I mean, one of the analyses that that we actually did sort of as a secondary analysis, frankly, almost driven by my own curiosity was what is the effect of presidential appointment on judges specifically, as it relates to a motion to dismiss and whether or not we can detect corporate favoritism. And we detect that particular judges based on who appointed them are more favorable towards corporations. And the results were, in some ways, highly surprising. And then sort of when you try to walk it back a little bit less so. So as it comes out, female Democrats are about as sympathetic to corporate parties, as Republicans overall. Yes, the worst, if you will, the least favorable towards corporations, are females appointed by President Trump by almost 10 percentage points. Now, how would you that’s counterintuitive, counterintuitive in some ways, although if you sort of try to roll that back of it, and you think about how, you know, and this is not a, you know, this is not to get into politics, I think everyone would agree that President Trump is a different type of Republican that was that preceded him. And his judicial appointees, many of them were, let’s just say that they didn’t take the same mold as other judicial nominees, whether it be Republicans or Democrats. So the fact that they are, if you will, outliers, or that they fall to one extreme or the other, isn’t necessarily all that surprising, once you start thinking about that, where they went to law school, right? What associations all those judges were affiliated, you know, and when you start breaking it down that way, you can see how that totally counter intuitively right? You would arrive at something an entirely different assumption that that, in fact, is it’s not the case at all. So that’s really the limitation that attorneys have, that they have such a limited lens into who the judge is, you know, maybe they know which judge appointed that which which President appointed them. Are they really liberal, when it comes to a contract matter between two large, large corporations? I mean, doesn’t matter to a judge that went to Harvard, that the lawyers that are appearing before them also went to Harvard? Does it matter in the sense that they’re more skeptical of their former colleagues? Or are they more willing to accept what they say? And give them more credence? And when when you start sort of playing this out, and you realize, no one element, can you really use if you want to get an accurate prediction that you can go to your client with? No, and say, I know how this is going to come out?

John Quinn  12:53

Well, I mean, this raises the question, um, this, based on this capability, are there things that a litigant a lawyer, a party can do to improve their chances of success? I mean, don’t you, you typically can’t change the judge, you know, the judge, at least in federal court, you’re assigned to a judge, and that judge is going to hear the matter, for all reason, for all purposes, I suppose you can change your lawyer. If your analysis shows that way, if your lawyer went to a small state school, you would have a material better chance of success with this particular judge, than the President’s graduate of Harvard who’s representing you, I suppose that’s something that you could do.

13:32

Sure. But actually, when it comes to even federal judges, we had a client that was in front of a judge in the Southern District of Florida, large, fairly decent sized case. And we went ahead and ran our analysis on her to determine whether or not they would rule in their favor, they were on the defense side, their analysis said, No, that’s not gonna happen, it’s gonna go to discovery, and then presumably, some, you know, multimillion dollar settlement aside from the discovery costs, but they had the opportunity to transfer venue, the Central District of California. Now, of course, before a case is assigned, you don’t know which judge in the central district that you’re going to get, and we’re very judged focused. So the way that we deal if you will, with the lack of determination. So the judges, we run our analysis on every single judge in the Central District of California, you provide an overall score, and then we also understand what each of those judges will do. And you could also if you will filter out for senior status judges, because they’re less likely to take the case. And when you do that, you see you have a much higher percentage of success of the motion be granted than the case to go away. But you’re essentially taking you know, millions of dollars, and you’re reducing it to zero, simply by using the AI to your advantage. Now, of course, that ultimately may be assigned to a different judge, but just think of that, like, on the one hand, you’ve know with almost 85 degree of 85% degree of certainty that this is going to go to discovery. And if you had even a chance That’s that that wouldn’t happen. And a good chance. Of course, you would avail yourself to that, I would think I would think so.

John Quinn  15:09

This is fascinating. Are there things that you’re working on? Being able to apply this technology to other types of decisions other than decisions and motions to dismiss? Yeah.

15:21

So first of all, we we began, when we launched, we had complete federal coverage of civil cases, motions to dismiss, I can’t tell you about the recent indictment and how that’s gonna go. But so so we have coverage of pretty much all civil cases at federal court, we acquired the assets of a company at the beginning of the year that had large amounts of state data. But we’re in the process of applying our approach and our and our models to state court data. So we’ll be able to provide those same predictions at the state court level for motions to dismiss. We’re also working on summary judgment, and trying to determine what we can do and what approach we might do apply there. Additionally, as you mentioned, as it relates to attorneys, we’re almost finished building a module for mainly for in house counsel, that would enable them as most large corporations have panel counsel, for us to assess, at least at the motion to dismiss stage whether or not picking a particular counsel may have an impact on the motion to dismiss. So exactly the scenario that you suggested, is what we’re rolling out in the near future.

John Quinn  16:30

I would think I mean, judges certainly have reputations for their willingness to grant summary judgment motions. So assuming that judges, there’s something to that that judges act consistently, and that they’re, you know, in your data assumes that judges will sort of act consistently, that seems to be an obvious potential application for this software.

16:56

Yes, it absolutely is. But But again, I want to distinguish between what the reputation is and then what the data actually bears out. Right. Right. Of course, those can be two very different things.

John Quinn  17:05

Sure. I mean, that’s, that’s anecdotal. lawyers think they have a read on that. But you you’ve got some actual data that will indicate whether that’s true or not. Yeah,

17:16

I mean, we had a fascinating discussion with a potential client that was a large plaintiffs firm, very large, plaintiffs firm. And in order to test our models, if you will, they sent us 25 of the most recent cases. Now, of course, 25, technically, is probably too small of a dataset, but you know, you have a potential client, do what they asked. And they had provided us with the names of the cases and the judges, and the numbers, but they also gave us a column with the outcome, which I was kind of surprised, but whatever, you know, I could figure it out anyways, on my own, and it was 25. And when I ran out analysis, we correctly anticipated 20 out of 25. When I sent it to him, he actually said, we are 20 is incorrect, because one of the cases he said was quote, unquote, miscoded. And really, it’s 21 out of 25. I can’t say if that was delivered or not. But be that as it may 11, out of 25, were dismissed. And this is from plaintiff side. Right? Right. So for those 11 out of 25, it was a total loss for them. Total Loss, and these were big cases, and nearly half talked about, like anecdotal evidence, right, so nearly half of the cases that they thought were going to be winners or losers, now, that to me, every one of the 11, by the way, we correctly predicted to be dismissed. Now with that information, knowing that it might be dismissed, knowing that we can do a jurisdictional analysis, even before cases filed, you could take those 11 losses based on presumably what these lawyers thought would be, you know, good jurisdictions, good judges, because they always appear before them, you know, and turn those into wins. And one of the the, the issues with law, generally, I would argue is, is that it’s fairly conservative meaning and that manifests itself in a variety of ways. One of which is on the plaintiff side, for example, they always file in the same district, once a determined that a district is favorable, then they just go there. And same thing on the defense side that this, this district is better for our corporate clients. Now, that may or may not be true, but it could, it might also be true that they’re 20, others that are equally good, if not better. But because you’re conservative, you just stick with what you know, and you don’t run an analysis because of course, it’s too difficult for any given lawyer to do that. It requires crunching, you know, dozens of years of data looking at, you know, potentially hundreds of judges looking at 1000s or 10s of 1000s or hundreds of 1000s of decisions. So they basically just go with what they know. Without knowing whether or not that is actually the best approach.

John Quinn  19:48

So I’m gonna say where you have options about the different jurisdictions and venues in which to file this gives you an ability to make an intelligent, evidence based decision about where to find And sometimes, you know, making motion to dismiss can be expensive. And a lot of times you wonder, is it worth it? Is it worth investing in this? There’s a cost, not only in attorneys fees, but you give up an element of surprise, you lay out your arguments, you educate the other side. So there’s a cost in that sense, too. So I can see that being able to decide. And when there’s a low probability of success, we’ll save the money and we want to educate the other side. And having some confidence in making that decision can be really valuable.

20:30

Yes, exactly. And that doesn’t have to do with the strength of your case, it simply has to do with the judge that you’re assigned to. Right.

John Quinn  20:38

So do you see a time where you can actually you’ll actually read the briefs? And instead of just get the docket number, but you know, be able to code for arguments, if you will? And help lawyers make decisions about what arguments to make to a particular judge? Is that down the road?

20:58

No, because we again, view the decision maker as the most important component of any given case, in terms of prediction. Lawyers have a pretty good sense of arguments and strengths and weaknesses. And this is not intended to displace that one of the things that you constantly hear with any new technology, how many peoples is going to displace or are we all going to lose our jobs. And I think, frankly, that’s that’s kind of a juvenile approach. And if you look at the history of technology, that certainly has not been the case, generally instead, but it enables us to do is to be more efficient, or to be smarter, right to use the technology to our advantage, to crack open a black box that we otherwise were unable to get to, in this instance, the judge in terms of the arguments in terms of the case law, you know, the variety of the commercially available databases entry, like which cases the judge cites more frequently, not that that necessarily tells you about weight or applicability, but I’m sure the AI will get better that might be able to identify more compelling arguments. But in terms of our capability, what we found getting to 85% Is that the briefs don’t have or don’t enable us to arrive at any better prediction. I mean, to the extent that the 15% is there, you know, based on our analysis, a lot of that just has to do with a lack of information at this point in time, meaning, like the judge, all things being equal would in fact, rule in favor of your motion, but for whatever reason, they got in a fight with their spouse on the way out the door, their partner, and therefore, it just throws off the way that they act. Look. 15% is something but it’s a very, very small amount when you think about it. So there are some variables, potentially, maybe we could close that shift small 15% gap with some analysis of the briefs. But when we were looking at when we were building our models, I mentioned the limitations of Natures of Suit. One of the analysis that we did do was through complaints and trying to determine which Natures of Suit better conformed or were more expressive of what was actually in the complaint. So we certainly did look at documents. But in terms of the the briefs, and the arguments that we sort of leave that to the attorneys to continue slaving at and billing by the hour.

John Quinn  23:14

Well, I mean, just you can imagine, I mean, some recurring issues, and I’m struggling to think of one, you know, say federal preemption, or, you know, some judges are, seem to be, have a greater proclivity to invalidate patents or trademarks and other judges, or I’m gonna You can imagine the legal issues where that are recurring all the time. And you collect data points on how judges rule on these recurring issues.

23:43

So so the way that we look at that is it’s not recurring in the same sense that you are using that word, it’s not recurring simply because it’s a patent. And they are more skeptical of patent or trademark claims. It’s recurring, recurring, because it could be the party types are the same, it could be that the lawyers are the same people that that are appearing for them over and over again, that’s what’s recurrent about if you want to look at the capability to find that pattern, that recurrent pattern that you could point your finger at and say, if these criteria are here, we know how the judge the judge will be necessarily skeptical, the judge is going to invalidate the patent. But it’s finding those recurrences, if you will. That is the key. And the recurrency is not simply the subject matter.

John Quinn  24:26

You mentioned that at one point you worked in a large international law firm How did your career evolved to the point that you know you were an associate and I don’t know maybe a partner to in a big international firm and develop this AI capability to predict what judges are going to do.

24:43

So I am based in DC and I like to think of a fairly traditional ad DC career directory up to a point started at a large firm I went to school in the DC area started a large firm, went to the Department of Justice to do litigation went in house, defense government contractor think all that is go boring chatter at a cocktail party, certainly in the DC area, then ultimately, I ended up at a small data analytics company that had access to a number of analytic platforms, or mainly being used in the intelligence space for a variety of intelligence agencies. And it was there where I started to understand the way that we look at things in if you will, the real world is very different than in the way that you want to run analysis. In that instance, you know, where the next terrorist attack may, may come from, or who’s a threat or who’s not a threat. And a lot of that has to do with the approach to data understanding that we are, if you will, one person, but we have a lot of component parts, right. So if you think about the way that we break down the judge, you know, and their component parts and, and perhaps when, when, when you’re thinking about those, you know, actors, and again, in the intelligence space, you want to find commonality between them, but it’s not simply going to be their name. It’s not simply going to be where they live. It’s these other characteristics that you have to sort of understand and use a massive data set in order to find those patterns, and trends. So that’s where I had the idea potentially, that you might be able to harness big data, and then specifically, focus more on the individual that is the judge, rather than the legal aspects of the case. What’s

John Quinn  26:22

fascinating, what, how do you see this? I mean, you’ve talked a little bit about potential applications for summary judgment. Do you see any other applications for this approach to using data analytics to predict decisions made by judges? What’s in the future? What’s in the future?

26:43

So I think that you know that this approach of looking at the decision maker is, is very important. And that can be applied in any number of of these legal contexts. It might be you’re looking at, for example, you know, anti trust or something else, whether or not the government will approve or deny or decide to intervene. If you understand who the actors are in that decision. And you have enough information that’s a potential use, of course, any motion that’s involved in litigation. But then there there there are other ways of looking at predictive technology and sort of expanding its use. So if we’re capable of using predictive technology, of course, we’re able to look beyond the law and Beyond Counseling once clients. So if you want to understand, let’s say, publicly traded companies, impact of litigation events, of course, knowing the outcomes is very important and can provide an advantage in terms of litigation funders, really being able to do an objective assessment of the case beyond the merits. You know, you’re you’re anyways building financial and economic models. So that certainly is a distinct application. But then in terms of just how you deploy it, it’s really taking the the the central premise that the actors matter more, and what matters to the actors is not necessarily the substance. It’s these, if you will, seemingly external elements in any decision making

John Quinn  28:09

was really fascinating stuff. Dan, how can people find you if they’re interested in following up your name, your business is pre dicta PR e slash dicta DIC TA,

28:21

B, era slash or dash di CTA have a website more than happy to reach out to us. I would just mentioned just because you brought up like, like how it can be used just sort of to emphasize the power. One of the questions I get asked this, what do you do with newly appointed judges? Right? I’ve talked about judges and how much we know about judges. What do you do with newly appointed judges, and this is where the biographical characteristics really come into play. Because with the newly appointed judge, but we don’t have their decisional history, we do have their biographical information. And we can then match those two judges that have been on the bench for years or decades. We use those judges as proxies for what this newly appointed judge will do. And admittedly, we aren’t quite as accurate from 85, we dropped 81. But the reason that that’s such a minimal drop, is because the biographical matching those not obvious patterns is so important in determining in determining the decisions. Were Put another way, a newly appointed judge, we can tell you today, what they’re going to do for the next 10 or 15 years of their career, simply by applying tenants of big data, AI and so on. So I just think you know that back to your question about how this can be used, I really think that it’s incredibly powerful. And this, I would say, almost nearly unlimited approach. It really just depends on how creative the person doing. But certainly to come back to your question how they can find me the internet is wonderful. Firm, a firm believer in Google, and it’s approach. You may get started getting targeted ads from us. I can’t say one way or another.

John Quinn  29:55

Thank you, Dan. We’ve been talking to Dan Rabinowitz have predicted Okay, As a fascinating predictive software capability to predict to 85% accuracy, how federal judges are going to rule on motion to dismiss and a lot of other potential future applications. This is John Quinn and this has been law disrupted.