WEBVTT 00:00:00.000 --> 00:00:02.640 Imagine for just a second that you're staring 00:00:02.640 --> 00:00:05.980 down a highly complex, incredibly high stakes 00:00:05.980 --> 00:00:09.119 puzzle. Yeah. And we are not talking about, you 00:00:09.119 --> 00:00:12.199 know, a casual weekend brain teaser here. We 00:00:12.199 --> 00:00:16.199 are talking about a scenario where deciphering 00:00:16.199 --> 00:00:20.359 the exact meaning of a single, seemingly ordinary 00:00:20.359 --> 00:00:23.859 word. Just one word. Right. Just one word could 00:00:23.859 --> 00:00:26.920 literally be worth. millions or maybe even billions 00:00:26.920 --> 00:00:29.219 of dollars. It's wild to think about. It really 00:00:29.219 --> 00:00:31.660 is. Welcome to the Deep Dive, everyone. We are 00:00:31.660 --> 00:00:34.280 thrilled you're joining us today because we are 00:00:34.280 --> 00:00:36.659 looking at a corner of the legal world that operates 00:00:36.659 --> 00:00:39.200 exactly like that high stakes puzzle. A massive 00:00:39.200 --> 00:00:41.200 puzzle. And it's one that plays out entirely 00:00:41.200 --> 00:00:43.500 behind the closed doors of United States district 00:00:43.500 --> 00:00:45.719 courts. Exactly. Specifically, we're going to 00:00:45.719 --> 00:00:48.939 be looking at the Markman hearing. Which, as 00:00:48.939 --> 00:00:51.460 anyone following patent litigation probably already 00:00:51.460 --> 00:00:54.600 knows, is a highly specialized judicial proceeding. 00:00:54.820 --> 00:00:57.219 It is designed exclusively for claims dealing 00:00:57.219 --> 00:00:59.759 with patent infringement. And during this hearing, 00:00:59.899 --> 00:01:02.619 the judge has one single solitary responsibility, 00:01:03.340 --> 00:01:05.540 claim construction. Or claim interpretation, 00:01:05.879 --> 00:01:09.340 yeah. Right. The judge has to sit down and meticulously 00:01:09.340 --> 00:01:12.739 define the exact scope and the exact meaning 00:01:12.739 --> 00:01:15.140 of the words and phrases written into a patent. 00:01:15.439 --> 00:01:19.000 And that definition. It usually ends up defining 00:01:19.000 --> 00:01:21.260 the entire scope of the patent. Absolutely. It 00:01:21.260 --> 00:01:23.359 swings the momentum totally in favor of the inventor 00:01:23.359 --> 00:01:27.500 or totally against them. So our mission for you 00:01:27.500 --> 00:01:30.439 today in this deep dive is to explore how this 00:01:30.439 --> 00:01:33.579 single legal choke point really shapes the fate 00:01:33.579 --> 00:01:36.120 of inventors and corporations everywhere. Because 00:01:36.120 --> 00:01:38.459 it really does. We're going to look at why judges 00:01:38.459 --> 00:01:41.060 and specifically not juries hold all the power 00:01:41.060 --> 00:01:44.180 in this situation. And we'll trace the really 00:01:44.180 --> 00:01:47.540 fascinating ripple effects. this procedural quirk 00:01:47.540 --> 00:01:51.040 has on the entire U .S. legal system. From draining 00:01:51.040 --> 00:01:53.340 judicial resources all the way to drastically 00:01:53.340 --> 00:01:55.640 extending the timelines of litigation. Yeah, 00:01:55.700 --> 00:01:57.920 it's a huge bottleneck. It is a procedure that 00:01:57.920 --> 00:02:00.219 fundamentally alters how innovation is protected 00:02:00.219 --> 00:02:03.459 and litigated in the United States. I mean, to 00:02:03.459 --> 00:02:05.359 understand the friction we see in modern patent 00:02:05.359 --> 00:02:07.000 courts today, you have to look at the massive 00:02:07.000 --> 00:02:10.360 shift that happened in the mid -90s. Okay, let's 00:02:10.360 --> 00:02:12.759 unpack this. The inception of all this drama. 00:02:13.350 --> 00:02:17.490 It starts with a 1996 Supreme Court case, Markman 00:02:17.490 --> 00:02:20.969 v. Westview Instruments, Inc. Which is obviously 00:02:20.969 --> 00:02:22.789 where we get the name. Right, which gives the 00:02:22.789 --> 00:02:25.590 Markman hearing its name. This was the absolute 00:02:25.590 --> 00:02:27.650 catalyst for the current system we have now. 00:02:27.889 --> 00:02:30.830 Yeah, and the tension at the heart of that 1996 00:02:30.830 --> 00:02:33.990 case is actually a constitutional one. Oh, right. 00:02:34.110 --> 00:02:36.629 The Seventh Amendment. Exactly. Ordinarily, under 00:02:36.629 --> 00:02:39.030 the Seventh Amendment, an individual is guaranteed 00:02:39.030 --> 00:02:41.530 the right to a jury trial in certain circumstances. 00:02:42.009 --> 00:02:45.689 So the prevailing assumption for a long, long 00:02:45.689 --> 00:02:48.610 time was that if you had a dispute over a patent. 00:02:49.099 --> 00:02:51.340 Well, a jury of your peers would look at the 00:02:51.340 --> 00:02:53.280 fact. Listen to the experts, look at the diagram. 00:02:53.599 --> 00:02:55.400 Right. And then they decide who is infringing 00:02:55.400 --> 00:02:57.879 on what. But the Supreme Court in the Markman 00:02:57.879 --> 00:03:00.599 case made a massive pivot. They held that interpreting 00:03:00.599 --> 00:03:03.139 patent claims is actually a matter of law rather 00:03:03.139 --> 00:03:05.520 than a question of fact for a jury to decide. 00:03:05.740 --> 00:03:08.020 But wait, I mean, we trust juries with incredibly 00:03:08.020 --> 00:03:11.340 complex, highly technical matters all the time, 00:03:11.340 --> 00:03:14.039 don't we? We do. We asked them to decide on medical 00:03:14.039 --> 00:03:17.060 malpractice suits involving obscure surgical 00:03:17.060 --> 00:03:21.300 procedures. or antitrust cases with convoluted 00:03:21.300 --> 00:03:23.800 market dynamics. Totally. So why shouldn't we 00:03:23.800 --> 00:03:27.020 just trust juries to listen to the patent experts 00:03:27.020 --> 00:03:30.439 and decide what a technical word means? Well, 00:03:30.479 --> 00:03:33.199 the court's reasoning really comes down to treating 00:03:33.199 --> 00:03:35.939 patents not just as technical descriptions of 00:03:35.939 --> 00:03:39.800 technology, but as legal contracts. Ah, the contract 00:03:39.800 --> 00:03:42.300 analogy. Yeah. If you look at legal history, 00:03:42.539 --> 00:03:45.180 contracts have historically been reviewed and 00:03:45.180 --> 00:03:48.250 interpreted by the courts, not juries. Judges 00:03:48.250 --> 00:03:50.530 are the ones who parse the fine print of a corporate 00:03:50.530 --> 00:03:53.710 merger, right? Or a complex real estate agreement. 00:03:53.889 --> 00:03:55.870 Because it's about legal boundaries. Precisely. 00:03:56.219 --> 00:03:58.340 Because patent language establishes the legal 00:03:58.340 --> 00:04:00.520 boundaries of a property right, the Supreme Court 00:04:00.520 --> 00:04:02.960 determined that judges are just far better equipped 00:04:02.960 --> 00:04:05.319 to address those boundaries. They are trained 00:04:05.319 --> 00:04:07.659 to parse out the legal language in a way a jury 00:04:07.659 --> 00:04:10.139 wouldn't be. So the judge is essentially defining 00:04:10.139 --> 00:04:12.759 the property lines before the jury is even allowed 00:04:12.759 --> 00:04:15.400 to decide if someone trespassed. That is a perfect 00:04:15.400 --> 00:04:17.699 way to put it, yes. But that contract analogy, 00:04:17.959 --> 00:04:20.560 it shifts the entire power dynamic of a trial. 00:04:21.300 --> 00:04:23.959 The moment judges took the wheel on defining 00:04:23.959 --> 00:04:27.120 these terms, they ran into a practical nightmare. 00:04:27.139 --> 00:04:31.180 If a judge has the absolute power to determine 00:04:31.180 --> 00:04:34.199 what a patent's words mean, they need strict 00:04:34.199 --> 00:04:36.100 rules on what kind of evidence they are allowed 00:04:36.100 --> 00:04:38.329 to look at. Right. I mean, a federal judge might 00:04:38.329 --> 00:04:40.990 be a brilliant legal mind, but they aren't necessarily 00:04:40.990 --> 00:04:43.750 electrical engineers. Or biochemists. Yeah. Right. 00:04:43.829 --> 00:04:45.850 They can't just Google complex semiconductor 00:04:45.850 --> 00:04:48.629 terms or ask a friend in the industry. No, they 00:04:48.629 --> 00:04:51.550 cannot. So in analyzing this patent language, 00:04:51.910 --> 00:04:54.509 the courts needed a formalized framework for 00:04:54.509 --> 00:04:56.610 what sources of information they could turn to 00:04:56.610 --> 00:04:59.509 for guidance. And over time, the courts split 00:04:59.509 --> 00:05:02.350 the acceptable evidence into two very distinct 00:05:02.350 --> 00:05:05.889 buckets, intrinsic evidence and extrinsic evidence. 00:05:06.250 --> 00:05:08.250 OK, let's start with intrinsic. Intrinsic evidence 00:05:08.250 --> 00:05:10.610 is the core material. It's the first line of 00:05:10.610 --> 00:05:13.269 defense. And there are exactly three things that 00:05:13.269 --> 00:05:15.470 fall into this bucket. The first one is pretty 00:05:15.470 --> 00:05:18.689 obvious. It's the patent itself. The actual words 00:05:18.689 --> 00:05:21.110 written on the page are the single most important 00:05:21.110 --> 00:05:24.410 source of evidence. The court analyzes these 00:05:24.410 --> 00:05:26.769 words based on their plain, ordinary meaning. 00:05:27.149 --> 00:05:29.029 But they also have to look at them specifically 00:05:29.029 --> 00:05:31.839 in light of the invention itself. Yes, they're 00:05:31.839 --> 00:05:33.860 looking for the intended scope of protection 00:05:33.860 --> 00:05:36.839 based on the text the inventor actually submitted 00:05:36.839 --> 00:05:38.860 to the government. So the text is paramount. 00:05:38.959 --> 00:05:41.500 The text is paramount. But as we know, words 00:05:41.500 --> 00:05:44.300 can be flexible. Which brings us to the second 00:05:44.300 --> 00:05:47.100 piece of intrinsic evidence, the patent specification. 00:05:47.699 --> 00:05:50.120 Okay. This is often referred to in the legal 00:05:50.120 --> 00:05:53.139 world as the dictionary of a patent. Yeah, I 00:05:53.139 --> 00:05:55.139 saw in the sources that there's actually a specific 00:05:55.139 --> 00:05:58.199 statute demanding the Patent Act almost like 00:05:58.199 --> 00:06:01.029 its own dictionary. Section 112. Right. 35 USC. 00:06:01.290 --> 00:06:04.250 Section 112 basically says you have to describe 00:06:04.250 --> 00:06:07.250 your invention so fully, so clearly and concisely 00:06:07.250 --> 00:06:09.670 that anyone skilled in that specific field could 00:06:09.670 --> 00:06:11.829 read it and immediately know how to make and 00:06:11.829 --> 00:06:14.449 use the invention. Exactly. And the underlying 00:06:14.449 --> 00:06:16.910 logic there is that inventors are allowed to 00:06:16.910 --> 00:06:19.029 act as their own lexicographers. They can make 00:06:19.029 --> 00:06:21.709 up their own definitions. Basically, yes. If 00:06:21.709 --> 00:06:24.089 an inventor decides to use a very common word 00:06:24.089 --> 00:06:27.029 but gives it a highly specific, unique meaning 00:06:27.029 --> 00:06:30.060 just for their invention. That dictionary of 00:06:30.060 --> 00:06:32.019 the patent is where the judge finds that out. 00:06:32.279 --> 00:06:35.339 The judge looks at the specification right alongside 00:06:35.339 --> 00:06:38.000 the patent claims to see if the inventor used 00:06:38.000 --> 00:06:40.540 any terms in a way that is inconsistent with 00:06:40.540 --> 00:06:43.139 ordinary everyday language. Which leads directly 00:06:43.139 --> 00:06:45.879 to the third piece of intrinsic evidence, which 00:06:45.879 --> 00:06:48.240 is the patent history. Yes, the paper trail. 00:06:48.399 --> 00:06:51.100 Right. This includes... Every single proceeding 00:06:51.100 --> 00:06:53.439 and communication that happened prior to the 00:06:53.439 --> 00:06:55.680 patent actually getting approved. We're talking 00:06:55.680 --> 00:06:58.060 about the entire back and forth, all the representations 00:06:58.060 --> 00:07:00.660 the applicant made to the patent and trademark 00:07:00.660 --> 00:07:03.259 office. That history is vital because it locks 00:07:03.259 --> 00:07:06.100 the inventor into their previous claims. It provides 00:07:06.100 --> 00:07:08.279 the court with a much more in -depth understanding 00:07:08.279 --> 00:07:10.660 of what the inventor originally intended to protect. 00:07:11.000 --> 00:07:13.399 You can't backtrack. Exactly. If the inventor 00:07:13.399 --> 00:07:15.259 told the patent examiner during the application 00:07:15.259 --> 00:07:18.860 process, my invention specifically does not perform. 00:07:19.120 --> 00:07:22.300 function X just to get the patent approved. They 00:07:22.300 --> 00:07:25.540 cannot turn around years later in a Markman hearing 00:07:25.540 --> 00:07:28.779 and argue that their patent actually covers function 00:07:28.779 --> 00:07:31.439 X. Right. They are bound by that history. Okay. 00:07:31.500 --> 00:07:34.779 So the judge exhausts the patent, the specification, 00:07:34.899 --> 00:07:36.899 and the history. Those are the three intrinsic 00:07:36.899 --> 00:07:39.759 sources. But what happens when the judge looks 00:07:39.759 --> 00:07:42.360 at all three of those and the technical language 00:07:42.360 --> 00:07:45.100 is still completely ambiguous? That is where 00:07:45.100 --> 00:07:47.939 we shift to the second bucket. Extrinsic evidence. 00:07:48.339 --> 00:07:51.100 The outside information. Right. This is outside 00:07:51.100 --> 00:07:53.379 information typically taking the form of expert 00:07:53.379 --> 00:07:56.079 testimony. This is where a litigator brings in 00:07:56.079 --> 00:07:58.579 a scientist or a specialized industry engineer 00:07:58.579 --> 00:08:01.319 to sit on the stand and explain what a specific 00:08:01.319 --> 00:08:03.959 term means in their highly specialized field. 00:08:04.259 --> 00:08:06.019 OK, that makes sense. But what's fascinating 00:08:06.019 --> 00:08:09.199 here is the absolute civil war this sparked within 00:08:09.199 --> 00:08:11.759 the federal circuit courts regarding when or 00:08:11.759 --> 00:08:14.300 even if you could use this extrinsic evidence. 00:08:14.660 --> 00:08:17.600 A civil war over experts. Yeah, you had different 00:08:17.600 --> 00:08:19.800 appellate panels applying totally contradictory 00:08:19.800 --> 00:08:23.160 rules, which created a massive headache for litigators 00:08:23.160 --> 00:08:26.180 trying to build a case. For example, in a 1996 00:08:26.180 --> 00:08:29.259 case called the Tronics Court v. Conceptronic 00:08:29.259 --> 00:08:33.419 Inc., the court drew a very hard line. How hard. 00:08:33.559 --> 00:08:35.860 They ruled that extrinsic evidence cannot be 00:08:35.860 --> 00:08:38.120 used at all if the intrinsic evidence is sufficient 00:08:38.120 --> 00:08:41.559 to resolve the ambiguous claims. Meaning, if 00:08:41.559 --> 00:08:43.399 the patent and its history solve the puzzle, 00:08:43.639 --> 00:08:46.179 outside experts are strictly barred from influencing 00:08:46.179 --> 00:08:48.580 the judge's interpretation. Honestly, that seems 00:08:48.580 --> 00:08:50.879 like a solid rule to keep trials efficient and 00:08:50.879 --> 00:08:53.100 focused on the actual documents. It does. But 00:08:53.100 --> 00:08:56.259 then, just three years later, in a 1999 case 00:08:56.259 --> 00:08:58.700 called Pitney Bowes, Inc. v. Hewlett Packard 00:08:58.700 --> 00:09:01.720 Company, the same court seemingly fractured its 00:09:01.720 --> 00:09:04.470 own precedent. They determined that judges actually 00:09:04.470 --> 00:09:06.929 could turn to extrinsic evidence for guidance 00:09:06.929 --> 00:09:09.649 and background context, even in circumstances 00:09:09.649 --> 00:09:11.970 where the intrinsic evidence was completely adequate 00:09:11.970 --> 00:09:14.429 on its own. You can imagine the chaos this caused 00:09:14.429 --> 00:09:17.230 for legal strategy. Total chaos. Litigators preparing 00:09:17.230 --> 00:09:20.070 for a Markman hearing had no idea which standard 00:09:20.070 --> 00:09:22.950 the judge was going to apply. I mean, do we spend 00:09:22.950 --> 00:09:25.289 hundreds of thousands of dollars retaining industry 00:09:25.289 --> 00:09:28.370 experts just in case the judge follows the Pitney 00:09:28.370 --> 00:09:32.029 Bowes flexible approach? Or do we rely entirely 00:09:32.029 --> 00:09:34.649 on the paper trail, hoping the judge follows 00:09:34.649 --> 00:09:37.289 the Vitronic strict approach? It was a huge gamble 00:09:37.289 --> 00:09:39.809 and it took years for this procedural nightmare 00:09:39.809 --> 00:09:43.210 to get sorted out. It wasn't until 2015 that 00:09:43.210 --> 00:09:45.909 the Supreme Court finally stepped in to settle 00:09:45.909 --> 00:09:48.009 the debate. And what was the final rule? They 00:09:48.009 --> 00:09:51.399 clarified it simply. In claim construction, judges 00:09:51.399 --> 00:09:54.740 must use only intrinsic evidence first. They 00:09:54.740 --> 00:09:56.980 have to exhaust those three intrinsic elements. 00:09:57.500 --> 00:10:00.200 Only if the claim terms remain genuinely ambiguous 00:10:00.200 --> 00:10:02.720 after that internal review can they open the 00:10:02.720 --> 00:10:05.139 door to extrinsic evidence like expert testimony. 00:10:05.480 --> 00:10:07.460 Here's where it gets really interesting, though, 00:10:07.500 --> 00:10:10.000 because figuring out the evidence hierarchy was 00:10:10.000 --> 00:10:12.899 only half the battle for these courts. Yep. Once 00:10:12.899 --> 00:10:14.620 the district court judge finally makes their 00:10:14.620 --> 00:10:16.860 decision on what the patent means, the losing 00:10:16.860 --> 00:10:18.779 side is almost certainly going to appeal it. 00:10:19.000 --> 00:10:21.519 The stakes are simply too high not to. And the 00:10:21.519 --> 00:10:23.779 appeals process for Markman hearings turned into 00:10:23.779 --> 00:10:27.559 an absolute statistical disaster. We need to 00:10:27.559 --> 00:10:29.779 talk about the reversal problem. The appellate 00:10:29.779 --> 00:10:32.539 drama is arguably a most impactful part of this 00:10:32.539 --> 00:10:35.480 entire system. After the Supreme Court created 00:10:35.480 --> 00:10:38.360 the Markman hearing in 1996, the Federal Circuit 00:10:38.360 --> 00:10:40.559 had to figure out what standard of review they 00:10:40.559 --> 00:10:42.779 were going to use when these patent interpretations 00:10:42.779 --> 00:10:45.879 were inevitably appealed. Right. So in 1998, 00:10:46.200 --> 00:10:50.500 a case called CyberCourt FAS -Tex, Inc. answered 00:10:50.500 --> 00:10:53.159 that question. The Federal Circuit affirmed that 00:10:53.159 --> 00:10:55.639 the standard of review would be de novo. Okay, 00:10:55.700 --> 00:10:57.740 since our listeners follow the legal space, we 00:10:57.740 --> 00:11:00.080 know that a de novo standard means the appellate 00:11:00.080 --> 00:11:02.700 court is giving zero deference to the lower court. 00:11:03.049 --> 00:11:05.549 zero deference. They aren't looking to see if 00:11:05.549 --> 00:11:07.690 the trial judge made a reasonable call. They're 00:11:07.690 --> 00:11:10.129 wiping the slate completely clean and reviewing 00:11:10.129 --> 00:11:12.610 the claim interpretation completely from scratch. 00:11:13.039 --> 00:11:14.860 The appellate judges essentially act as if the 00:11:14.860 --> 00:11:16.919 lower court's Markman hearing never happened. 00:11:17.139 --> 00:11:19.960 They don't consider the weeks or months the district 00:11:19.960 --> 00:11:22.240 court judge spent agonizing over the technical 00:11:22.240 --> 00:11:24.580 language or listening to the experts. They just 00:11:24.580 --> 00:11:26.879 start over with the raw materials. So when the 00:11:26.879 --> 00:11:29.139 appellate court started wiping the slate clean 00:11:29.139 --> 00:11:32.419 with this de novo review, what did that actually 00:11:32.419 --> 00:11:35.019 do to the reversal rates? It must have caused 00:11:35.019 --> 00:11:38.200 a massive spike in overturned cases. The statistical 00:11:38.200 --> 00:11:40.990 fallout was staggering. Before the Markman hearing 00:11:40.990 --> 00:11:43.330 was established, studies showed that the reversal 00:11:43.330 --> 00:11:45.509 rate for patent claim construction on appeal 00:11:45.509 --> 00:11:49.690 was roughly 20 .8%. Okay, but one in five. Right. 00:11:49.789 --> 00:11:52.490 But after the cyber decision entrenched that 00:11:52.490 --> 00:11:55.590 de novo standard, the reversal rate skyrocketed 00:11:55.590 --> 00:11:59.889 to 32%. Nearly a third of all patent claim constructions 00:11:59.889 --> 00:12:01.789 were being thrown out by the higher court. A 00:12:01.789 --> 00:12:05.210 32 % reversal rate is mind -blowing. Think about 00:12:05.210 --> 00:12:07.610 the massive inefficiency that introduces into 00:12:07.610 --> 00:12:10.070 the corporate world. It's huge. You have companies 00:12:10.070 --> 00:12:12.210 spending tens of millions of dollars on legal 00:12:12.210 --> 00:12:14.629 fees and district court judges spending vast 00:12:14.629 --> 00:12:16.750 amounts of their docket time defining these terms, 00:12:16.889 --> 00:12:19.470 only to have the appellate court bulldoze the 00:12:19.470 --> 00:12:22.289 entire foundation one out of every three times. 00:12:22.509 --> 00:12:24.610 And it makes settlements nearly impossible because 00:12:24.610 --> 00:12:26.769 the losing party always assumes they have a one 00:12:26.769 --> 00:12:28.809 in three chance of getting a total do -over on 00:12:28.809 --> 00:12:31.250 appeal. Why settle when the odds of a reversal 00:12:31.250 --> 00:12:33.789 are that high? Exactly. It just incentivizes 00:12:33.789 --> 00:12:36.490 endless appeals. And the legal system recognized 00:12:36.490 --> 00:12:39.690 this lack of uniformity was unsustainable. It 00:12:39.690 --> 00:12:42.049 undermined the entire purpose of having a district 00:12:42.049 --> 00:12:44.330 judge interpret the claims in the first place. 00:12:44.730 --> 00:12:47.830 So the Supreme Court had to step in again to 00:12:47.830 --> 00:12:50.710 fix the mess created by cyber. This was in 2015, 00:12:50.970 --> 00:12:53.649 right? Yes. In 2015, they took up a case called 00:12:53.649 --> 00:12:58.909 Teva Farms, USA, Ink v. Sandoz, Inc. The Supreme 00:12:58.909 --> 00:13:01.029 Court officially reversed the cyber precedent. 00:13:01.519 --> 00:13:03.559 They did away with the pure de novo standard 00:13:03.559 --> 00:13:05.899 across the board and created a much more nuanced 00:13:05.899 --> 00:13:08.320 hybrid standard. So they forced the appellate 00:13:08.320 --> 00:13:10.620 courts to finally respect the groundwork laid 00:13:10.620 --> 00:13:13.379 by the trial judges. How does that hybrid standard 00:13:13.379 --> 00:13:15.559 actually function when a case goes up for appeal 00:13:15.559 --> 00:13:17.779 now? It ties directly back into the evidence 00:13:17.779 --> 00:13:20.220 rules we discussed earlier. Under the Teva standard, 00:13:20.480 --> 00:13:22.179 the appellate court still reviews the ultimate 00:13:22.179 --> 00:13:24.669 interpretation of the patent. But if the district 00:13:24.669 --> 00:13:27.169 court had to rely on extrinsic evidence, say, 00:13:27.250 --> 00:13:29.269 they brought in a software engineer to explain 00:13:29.269 --> 00:13:31.889 a complex algorithm, and the trial judge made 00:13:31.889 --> 00:13:34.210 secondary factual findings about the credibility 00:13:34.210 --> 00:13:36.429 of that expert, the appellate court cannot just 00:13:36.429 --> 00:13:38.710 ignore that anymore. They have to respect the 00:13:38.710 --> 00:13:41.330 trial judge's assessment of the expert. Right. 00:13:41.409 --> 00:13:43.870 They must defer to the lower court's factual 00:13:43.870 --> 00:13:46.590 findings regarding that extrinsic evidence unless 00:13:46.590 --> 00:13:49.370 there is a clear error. So the appellate judges 00:13:49.370 --> 00:13:51.929 can't just substitute their own opinions on whether 00:13:51.929 --> 00:13:54.879 an expert was believable. If the trial judge 00:13:54.879 --> 00:13:57.000 looked the expert in the eye and deemed their 00:13:57.000 --> 00:14:00.100 testimony credible and foundational to the patent's 00:14:00.100 --> 00:14:02.740 meaning, the higher court has to accept that 00:14:02.740 --> 00:14:05.620 unless the trial judge made an obvious, undeniable 00:14:05.620 --> 00:14:09.259 mistake. Exactly. That requirement to defer to 00:14:09.259 --> 00:14:11.980 secondary factual findings injects a crucial 00:14:11.980 --> 00:14:15.039 layer of stability. In theory, it should bring 00:14:15.039 --> 00:14:18.240 that 32 % reversal rate down and give litigants 00:14:18.240 --> 00:14:20.480 much more confidence in the district court's 00:14:20.480 --> 00:14:23.080 initial Markman ruling. Though I imagine the 00:14:23.080 --> 00:14:25.500 long -term statistical impact of TAVA is still 00:14:25.500 --> 00:14:27.639 being analyzed by legal scholars. Oh, absolutely. 00:14:27.860 --> 00:14:30.799 It takes years for these trends to fully crystallize 00:14:30.799 --> 00:14:32.840 in the data. Okay, so we have covered who is 00:14:32.840 --> 00:14:34.899 making the decision, the strict rules on what 00:14:34.899 --> 00:14:36.639 evidence they can use, and how the appellate 00:14:36.639 --> 00:14:39.720 courts handle the fallout. But there is one final 00:14:39.720 --> 00:14:42.759 massive piece to this procedural puzzle. Timing. 00:14:42.980 --> 00:14:45.940 Timing is everything. If judges have all this 00:14:45.940 --> 00:14:49.090 power to define the boundaries of a patent, When 00:14:49.090 --> 00:14:51.330 do they actually drop the gavel on these definitions 00:14:51.330 --> 00:14:54.350 during the lifespan of a lawsuit? This is where 00:14:54.350 --> 00:14:56.730 the procedural ambiguity of the federal circuit 00:14:56.730 --> 00:15:00.769 really shines. The short answer is whenever the 00:15:00.769 --> 00:15:03.210 individual trial judge feels like it. Really? 00:15:03.309 --> 00:15:06.629 No set schedule? None. The federal circuit has 00:15:06.629 --> 00:15:09.610 pointedly refused to provide strict rules or 00:15:09.610 --> 00:15:13.129 deadlines to the lower courts on when a Markman 00:15:13.129 --> 00:15:16.080 hearing must occur. There are implied preferences, 00:15:16.259 --> 00:15:18.899 but ultimately it is entirely at the discretion 00:15:18.899 --> 00:15:21.740 of the district judge managing the docket. Generally, 00:15:21.759 --> 00:15:24.360 we see judges choosing one of three main phases 00:15:24.360 --> 00:15:26.720 in litigation to hold the hearing. OK, so if 00:15:26.720 --> 00:15:29.039 a judge has the flexibility to do this whenever 00:15:29.039 --> 00:15:30.840 they want, wouldn't they just want to get it 00:15:30.840 --> 00:15:32.840 out of the way immediately, like right at the 00:15:32.840 --> 00:15:34.740 start of the lawsuit during pre -discovery, before 00:15:34.740 --> 00:15:37.419 either side has spent millions digging through 00:15:37.419 --> 00:15:39.740 each other's hard drives? It seems logical. It 00:15:39.740 --> 00:15:42.100 seems like doing it early would force quick settlements. 00:15:42.559 --> 00:15:45.200 decrease corporate costs and centralize all the 00:15:45.200 --> 00:15:47.159 future discoveries strictly around the court's 00:15:47.159 --> 00:15:49.460 official interpretation of the patent. The early 00:15:49.460 --> 00:15:51.720 intervention approach sounds incredibly efficient 00:15:51.720 --> 00:15:54.000 on paper, but litigators will tell you it is 00:15:54.000 --> 00:15:57.159 often a disaster in practice. The fatal flaw 00:15:57.159 --> 00:15:59.519 of holding a Markman hearing during pre -discovery 00:15:59.519 --> 00:16:02.419 is the lack of context. Because they haven't 00:16:02.419 --> 00:16:04.879 seen the evidence yet. Right. The discovery phase 00:16:04.879 --> 00:16:07.480 has barely started, meaning the judge is being 00:16:07.480 --> 00:16:10.899 asked to interpret highly complex, abstract claim 00:16:10.899 --> 00:16:14.179 language in a complete vacuum. They don't have 00:16:14.179 --> 00:16:16.659 the full picture of the accused product or the 00:16:16.659 --> 00:16:19.080 broader market context yet. The flying blank. 00:16:19.480 --> 00:16:22.179 Exactly. What inevitably happens is that the 00:16:22.179 --> 00:16:25.080 court makes an early interpretation. But as millions 00:16:25.080 --> 00:16:42.919 of paid... That makes perfect sense. You can't 00:16:42.919 --> 00:16:45.120 draw the property lines if you haven't even surveyed 00:16:45.120 --> 00:16:47.500 the land yet. Great way to phrase it. That pushes 00:16:47.500 --> 00:16:50.399 us to the second timing option. Summary judgment. 00:16:50.960 --> 00:16:54.500 This happens prior to the actual trial, but crucially... 00:16:54.809 --> 00:16:57.169 after the massive discovery process has been 00:16:57.169 --> 00:16:59.850 completed. Yes. Conducting the hearing at the 00:16:59.850 --> 00:17:02.509 summary judgment stage is the most widely adopted 00:17:02.509 --> 00:17:05.329 route by federal judges. It strikes the best 00:17:05.329 --> 00:17:07.970 balance. It's the Goldilocks zone. Pretty much. 00:17:08.150 --> 00:17:10.750 It allows the judge to utilize all the technical 00:17:10.750 --> 00:17:13.109 context and information gathered during the long 00:17:13.109 --> 00:17:15.589 discovery phase to accurately interpret the claim. 00:17:15.869 --> 00:17:18.750 It also helps identify exactly which claims are 00:17:18.750 --> 00:17:21.789 truly central to the dispute, allowing the judge 00:17:21.789 --> 00:17:24.490 to potentially dismiss the case right there before 00:17:24.490 --> 00:17:26.730 wasting weeks of the court's time with a full 00:17:26.730 --> 00:17:29.130 jury trial. But doing it at summary judgment 00:17:29.130 --> 00:17:31.490 still carries that historical shadow we talked 00:17:31.490 --> 00:17:34.440 about. For years, if a judge conducted a Markman 00:17:34.440 --> 00:17:36.359 hearing at summary judgment and dismissed the 00:17:36.359 --> 00:17:38.720 case based on their interpretation, it carried 00:17:38.720 --> 00:17:41.680 that massive 32 % risk of a de novo reversal 00:17:41.680 --> 00:17:44.380 on appeal. It was a very real fear. The new clear 00:17:44.380 --> 00:17:46.779 error standard from the Teva case adds some protection, 00:17:46.920 --> 00:17:49.460 but the institutional memory of those high reversal 00:17:49.460 --> 00:17:51.779 rates still makes summary judgment a high anxiety 00:17:51.779 --> 00:17:54.259 moment for corporate legal teams. The anxiety 00:17:54.259 --> 00:17:56.900 is definitely still there, which leads some judges 00:17:56.900 --> 00:18:00.299 to choose the third and final option. holding 00:18:00.299 --> 00:18:03.160 the Markman hearing at or after the trial itself, 00:18:03.579 --> 00:18:06.200 right before the judge gives the jury their final 00:18:06.200 --> 00:18:08.920 instructions. Waiting until the trial's almost 00:18:08.920 --> 00:18:11.039 over to tell everyone what the patent actually 00:18:11.039 --> 00:18:14.900 means feels like the ultimate legal procrastination. 00:18:15.220 --> 00:18:18.660 It sounds incredibly messy. It does seem counterintuitive, 00:18:18.720 --> 00:18:21.119 but from a judge's perspective, there are strong 00:18:21.119 --> 00:18:24.029 advantages. If a judge waits until the trial 00:18:24.029 --> 00:18:26.569 phase, they have absorbed all the necessary background 00:18:26.569 --> 00:18:29.089 information. They have a deep understanding of 00:18:29.089 --> 00:18:31.509 the specific technology at issue, and they have 00:18:31.509 --> 00:18:33.890 personally heard all the competing evidence presented 00:18:33.890 --> 00:18:36.809 by both sides over weeks of testimony. So they 00:18:36.809 --> 00:18:38.910 are fully educated on the topic. Right. They 00:18:38.910 --> 00:18:41.049 are in the best possible position to make a fully 00:18:41.049 --> 00:18:43.549 informed, bulletproof decision on what the claims 00:18:43.549 --> 00:18:46.569 mean. But the logistical cons of waiting that 00:18:46.569 --> 00:18:48.950 long are severe. You have a jury sitting in a 00:18:48.950 --> 00:18:51.609 box for weeks listening to highly technical evidence 00:18:51.609 --> 00:18:54.250 and expert testimony without actually knowing 00:18:54.250 --> 00:18:56.869 the court's official legal construction of the 00:18:56.869 --> 00:18:59.089 patent they are evaluating. It's a huge burden 00:18:59.089 --> 00:19:02.109 on the jury. Furthermore, it sparks endless unnecessary 00:19:02.109 --> 00:19:04.789 litigation arguments during the trial itself 00:19:04.789 --> 00:19:07.589 because both legal teams are still arguing about 00:19:07.589 --> 00:19:09.950 competing claim interpretations in front of the 00:19:09.950 --> 00:19:13.089 jury. It just turns the courtroom into a chaotic 00:19:13.089 --> 00:19:15.660 theater. It creates significant friction during 00:19:15.660 --> 00:19:18.400 the trial. That is exactly why summary judgment 00:19:18.400 --> 00:19:21.000 remains the preferred middle ground despite the 00:19:21.000 --> 00:19:23.720 reversal risks. But looking at all three of these 00:19:23.720 --> 00:19:26.200 timing options, it really highlights how the 00:19:26.200 --> 00:19:28.519 insertion of this single judge -only hearing 00:19:28.519 --> 00:19:31.000 completely disrupts the traditional flow of a 00:19:31.000 --> 00:19:33.460 lawsuit. So what does this all mean when we zoom 00:19:33.460 --> 00:19:36.279 out and look at the whole picture, the 1996 Supreme 00:19:36.279 --> 00:19:39.369 Court pivot? The battles over intrinsic versus 00:19:39.369 --> 00:19:42.029 extrinsic evidence, the wild reversal rates and 00:19:42.029 --> 00:19:44.829 the timing chaos. And what is the ultimate real 00:19:44.829 --> 00:19:47.210 world impact of the Markman hearing for inventors 00:19:47.210 --> 00:19:50.630 and the legal system? The most profound consequence 00:19:50.630 --> 00:19:53.390 is that the Markman hearing has essentially bifurcated 00:19:53.390 --> 00:19:56.210 patent litigation. We no longer have a single 00:19:56.210 --> 00:19:59.170 unified trial where a jury hears everything and 00:19:59.170 --> 00:20:02.019 decides the outcome. It's split in two. Exactly. 00:20:02.160 --> 00:20:04.539 We have a Markman hearing to define the words 00:20:04.539 --> 00:20:07.180 and then, much later, a trial to apply those 00:20:07.180 --> 00:20:10.599 definitions. This bifurcated system duplicates 00:20:10.599 --> 00:20:13.250 efforts across the board. Technical experts have 00:20:13.250 --> 00:20:16.089 to be prepped, paid and produced on two completely 00:20:16.089 --> 00:20:19.269 separate occasions. It monopolizes massive amounts 00:20:19.269 --> 00:20:22.069 of judicial time, bloats the litigation budgets 00:20:22.069 --> 00:20:24.490 for both the plaintiffs and the defendants and 00:20:24.490 --> 00:20:27.329 drastically drags out the timeline from the initial 00:20:27.329 --> 00:20:30.150 filing to a final verdict. It really grounds 00:20:30.150 --> 00:20:33.039 the entire topic for you as a listener. This 00:20:33.039 --> 00:20:35.380 procedural choke point proves that in the high 00:20:35.380 --> 00:20:37.960 -stakes world of patents and innovation, having 00:20:37.960 --> 00:20:40.339 a brilliant world -changing idea is literally 00:20:40.339 --> 00:20:43.339 only half the battle. Only half. The other half 00:20:43.339 --> 00:20:45.700 is surviving a gauntlet of federal judges and 00:20:45.700 --> 00:20:48.440 appellate courts to successfully defend the exact 00:20:48.440 --> 00:20:50.440 definition of the words you used to describe 00:20:50.440 --> 00:20:53.059 that idea. If you can't survive the Markman hearing, 00:20:53.279 --> 00:20:55.680 your patent might as well be blank paper. This 00:20:55.680 --> 00:20:57.559 raises an important question, something for you 00:20:57.559 --> 00:21:00.619 to ponder long after this deep dive ends. The 00:21:00.619 --> 00:21:02.720 Markman framework was largely built around physical 00:21:02.720 --> 00:21:06.019 inventions, mechanical gears, chemical compounds, 00:21:06.220 --> 00:21:08.980 tangible things you can touch. But as we accelerate 00:21:08.980 --> 00:21:11.880 into an era dominated by abstract AI algorithms, 00:21:12.119 --> 00:21:13.859 neural networks and machine learning models, 00:21:14.160 --> 00:21:17.220 how will federal judges trained in 18th century 00:21:17.220 --> 00:21:19.920 constitutional law, not 21st century computer 00:21:19.920 --> 00:21:23.079 science, adapt to interpreting the plain meaning 00:21:23.079 --> 00:21:25.259 of software code that even software engineers 00:21:25.259 --> 00:21:28.000 constantly debate? If the English language is 00:21:28.000 --> 00:21:30.359 so inherently ambiguous that even expert judges 00:21:30.359 --> 00:21:32.519 get their patent word definitions overturned 00:21:32.519 --> 00:21:34.319 by higher courts nearly a third of the time, 00:21:34.569 --> 00:21:36.829 Are patents actually protecting the bold ideas 00:21:36.829 --> 00:21:39.490 of inventors, or are they just fueling an expensive 00:21:39.490 --> 00:21:41.269 high -stakes game of semantics for lawyers?