[00:00:00] Speaker 02:
Our final case this morning is the 25-14-19 pact with Intel.

[00:00:08] Speaker 02:
Okay, Mr. Weisbursch.

[00:00:09] Speaker 01:
May it please the court, Sanford Weisbursch for plaintiff appellant pact. I'd like to begin with the 301 patent. The district court at appendix 22 to 23 stated that capability of executing a function doesn't mean the product infringes.

[00:00:26] Speaker 01:
That is contrary to this court's precedent, including Finjin, where this court stated... But you didn't raise the capability argument. We did raise the capability argument. I'd like to point specifically where in the record that we did so. I refer the court to Appendix 14685, which is our opposition memorandum in opposition to summary judgment, as well as Appendix 147... Wait a minute, let's take it one at a time. Yes, we'll take it one at a time. What's the first reference? 14685...

[00:00:55] Speaker 03:
That's the paragraph in your opposition to summary judgment where you don't really say, it doesn't matter whether it actually does it. You don't say that. That's the problem with that paragraph.

[00:01:07] Speaker 01:
I would say the problem with that paragraph is a different problem, which is that it's a grammatically incorrect... That would be another problem.

[00:01:13] Speaker 01:
Yes. And here's the answer to that problem. The sentence is a bit ambiguous. It says, it talks about whether...

[00:01:24] Speaker 01:
It says Dr. Conte, I'm paraphrasing, testified that Intel's products, and then it said sequentially processing data. And there's a missing word there.

[00:01:34] Speaker 01:
But we cited Dr. Conte's expert report, and I'd like to give you another reference, 24315.

[00:01:40] UNKNOWN:
24315.

[00:01:40] Speaker 01:
24315. That is Dr. Conte's report. Same volume. Yes. Okay, but that's not exactly...

[00:01:54] Speaker 02:
arguing capability.

[00:01:57] Speaker 01:
Well, it is because Dr. Conte said that Intel's products can be used to process data.

[00:02:05] Speaker 02:
Maybe he said that, but you didn't say it in the opposition is the problem.

[00:02:11] Speaker 01:
Granted, it's not a perfect situation, but I would refer the court to the Golden Bridge case. Similar situation of did a party preserve or not an argument. The party there had cited its expert report, and this court was willing to look at the portions of the expert report that were cited to determine whether the argument was preserved. It was not preserved in that case. It was here because 24315 we submit is clear on this point.

[00:02:38] Speaker 01:
Beyond that brief and the Conte report, however, I want to refer to Appendix 14720.

[00:02:45] Speaker 01:
Now, this is our response to Intel's statement of material facts as to which there is allegedly no dispute.

[00:02:52] Speaker 03:
And just... Look, I think I'm remembering this right. That does give you more. The problem is in the place in the legal contention that you needed to say, here's the reason the other side is wrong about summary judgment. You don't say, it doesn't matter whether Intel's processors actually do it. You just don't say that.

[00:03:14] Speaker 01:
We don't say it precisely enough, except if this court should, as we submit, consider content. But I want to go back. I think the statement is important, Judge Taranto, for the following reason. A principle on summary judgment is that the opponent of a summary judgment motion only has to respond to the assertions that are made in the motion. And I think it's really important to focus in on what assertion Intel made in its motion. And this is paragraph 3 of its statement, appendix 14720. Excuse me, paragraph 13 and 14 of its statement, appendix 14720.

[00:03:47] Speaker 01:
The assertion that they made was that Intel's products process instructions, period. And they may have cited some evidence as well. But it was not an assertion that Intel's products are not capable of processing.

[00:04:01] Speaker 03:
But how do you...

[00:04:04] Speaker 03:
How is it possible that you don't have to say that assertion is immaterial to our ability to survive summary judgment? It doesn't matter whether they do or don't, as long as they can.

[00:04:16] Speaker 01:
Respectfully, I submit that we did say that. Again, what they said is Intel's products... If we disagree with you, you lose, right? Well... On the 301, but let me just finish my answer to this and then I will move on to the 593.

[00:04:29] Speaker 02:
Yes, if you're wrong. If we conclude that you didn't raise it, you lose.

[00:04:33] Speaker 01:
We would lose in that case. But just one final point on why we did raise it. In the response, so Intel's statement again is Intel's products process instructions, period. we responded that the fact that something processes instructions does not mean that it cannot process data as well. And cannot process data is, we respectfully submit, a reference to this capability idea. So we do think you should judge what we said in the context of what they said. And in the context of what they said, we both responded that it's not mutually exclusive to process instructions and to process data.

[00:05:07] Speaker 01:
You can do both. And we said, doesn't mean that we cannot process data. So we did actually make the capability argument. I'd like to move to the 593 argument.

[00:05:17] Speaker 01:
This is a case where I want to first start out just by orienting ourselves. And I think it may be helpful to have a couple of pages of the blue brief in mind as I make this argument. Blue brief, page 11, is an annotation that we prepared of an Intel document.

[00:05:34] Speaker 01:
And it shows an Intel product. And it's got a line that's going across.

[00:05:38] Speaker 03:
And just, that figure is secret.

[00:05:43] Speaker 01:
The, yes, and I'm going to... Mark in blue, so blue is your version of... Yes, I think I'm going to describe the document consistent with the words in the district court opinion that are public. I'm going to try to do that. But what you have in that document, it's our position, is a portion represented by a bump in the middle that is an exclusive segment that is only for the processor and the memory unit that are paired. And then you've got another line in the document that is shared. And we analogize this, and this is Blue Brief 3, to a highway with an overpass.

[00:06:17] Speaker 01:
The overpass is the exclusive portion. The highway is the shared portion. The local roads are shared as well. And the question before this court and before the district court is, Is it enough that there's an exclusive segment somewhere on that path in order to bring it within the claim physically dedicated connection?

[00:06:35] Speaker 03:
That's the key word of the problem. The problem, subject to your last little argument about the two endpoints, is that even on your highway overpass thing, there's... pieces at the processor end and the memory end that are shared. You have this little argument about how they're not really part of the path, and that's why you say your main argument is that as a matter of claim construction or something, as long as it's not enough for there to be a segment in the middle that is not shared.

[00:07:11] Speaker 03:
Right.

[00:07:16] Speaker 03:
Why in this context would that make any sense? Isn't the point of the dedicated connection that I can get from A to B without any traffic jam? It doesn't matter where that traffic jam would be. It might just be on 10 feet in the middle. It doesn't have to be on the whole thing, but the point of the dedicated connection that's required by the claim is that nobody's interfering anywhere from A to B. Why is that not right?

[00:07:50] Speaker 01:
Yes. The ultimate purpose, and this is column two of the patent, is indeed to reduce waiting time, to reduce latency by creating this special connection between the two units.

[00:08:00] Speaker 01:
To answer that question, let me try to refer to figure seven of the patent, appendix 143. That is a depiction of a memory unit in an embodiment of the patent. Of course, embodiments are presumed to practice the patent. And if you look on the left side, upper left and the bottom left, you'll see a single data input going into the memory unit and a single data output. Those necessarily have to be shared by everything in the schematic. So the patentee understood that there would in fact, and we of course, as Your Honor Judge Taranto noted, we call that an interface, that's a separate argument that we make, but it's necessary that that last piece is shared.

[00:08:40] Speaker 01:
and yet the patentee viewed it as satisfying the definition of physically dedicated connection. Now, the district court itself also agreed with us the first time around, and this is at Appendix 9. District court denied summary judgment by saying that, hey, there are separate paths. Reasonable jurors could find that there are separate paths. Intel disagrees. We've got to put this to the jury.

[00:09:01] Speaker 01:
The only reason, as we understand it, that the district court changed its view by the second order was based on prosecution history disclaimer. I'd like to address that before I sit down.

[00:09:11] Speaker 01:
We think the judge was wrong. It's a very demanding standard. It requires clear and unmistakable evidence of disclaimer, as the court knows.

[00:09:17] Speaker 02:
No, that's not the only thing that you use the prosecution history for. You use it to interpret the past.

[00:09:24] Speaker 01:
However one uses it. although we disagree with Intel on that point.

[00:09:28] Speaker 02:
We do think that there is this high standard for a reason, but let me point to what I think is... It's explicit that you use the specification of prosecution history to interpret the patent. It doesn't say it has to be a disclaimer.

[00:09:41] Speaker 01:
Well, here's the key point, regardless of how the court views it. At Appendix 3, so the district court believed, apparently, it was a little unclear from the decision, that...

[00:09:51] Speaker 01:
a portion of it being shared would be enough to take something outside the scope of the path and not satisfy physically dedicated. The district court did not consider all the statements in the prosecution history. Refer the court to Appendix 33958, which is one page after the Foo diagram. Foo is the prior art. And the point that Pack was making there is that all segments in the Foo are shared. So Pack didn't have a reason to get into this distinction of at least one versus many segments The priority was distinguishing all were shared.

[00:10:22] Speaker 01:
I'd like to read this sentence from Appendix 33958.

[00:10:26] Speaker 01:
A connection of any of the processors to any of the memories requires use of a data path of the FCU that is also equally used by other processors for connection to the memories so that the connection cannot be considered dedicated. The district court did not address that sentence. There's a similar sentence at 33955. And the point here is that PACS was focused not on this at least one versus more segments question. It was focused on the examiner's misapprehension of what the word dedicated means.

[00:10:56] Speaker 01:
The examiner thought dedicated means direct. It doesn't have to mean uniquely for a pair. Just a direct line between two elements. And that's what we were addressing in that prosecution history. We didn't have occasion to get into this distinction that was later used against us. And in fact, we made statements, the one I just read.

[00:11:14] Speaker 03:
Just to be clear, in that sentence, you're giving... a particular meaning to the phrase use of a data path? Yes, so I think... I think you're suggesting it means the very same data path from beginning to end?

[00:11:31] Speaker 01:
From beginning to end.

[00:11:32] Speaker 03:
We think that that sentence... If there are two different processors and say one memory, it's not possible that the entire data path from location one... to the memory is the same as location two to the memory because location one is different from location. It can't possibly mean the entire path, can it?

[00:11:52] Speaker 01:
Well, if you had two adjacent processors, it would be, and this is the diagram at three, three, nine, five, seven. It's basically a grid.

[00:12:02] Speaker 01:
Whether any two units share the same path, you can always find some other, either BBU it's called, or another processor that shares some of that path for the entire thing so that every segment in who we submit is shared. And that's why we didn't have occasion to distinguish that. Again, our focus was on the word dedicated and whether that means direct versus uniquely devoted. If the court will permit me, I'll save the rest of my time.

[00:12:45] Speaker 00:
Thank you, Judge Dyke. It may please the court, John O'Quinn, on behalf of Intel. The district court properly granted summary judgment as to both the 301 and the 593 patents. As the district court observed, PACT created a moving target on many issues, first in the PTO, then in the district court, and even now on appeals. But however viewed, there's no material evidence that Intel's microprocessors infringe and this court should affirm. I'll just very briefly address the 301 patent unless the court has questions about that.

[00:13:16] Speaker 00:
But the only argument that PAC made in its opening brief is that the relevant claim limitation is directed to mere capability, asserting that that's the same thing that it argued to the district court in opposing summary judgment. And in trying to do so, PAC does, in the district court's words, it tries to rewrite history. The district court recognized that that was not the argument that PAC made. The district court was clear about that in addressing this in the second rehearing motion. And this is not, just to be clear, this is not about a typo.

[00:13:47] Speaker 00:
This is not about a typographical error, because at the exact same... I think there is a bit of a typo in that. I agree with you, there is a bit of a typo, Judge Carranza.

[00:13:53] Speaker 03:
How do you correct it? How do you correct it? Get rid of the I-N-G or put words in front of the word processing?

[00:14:01] Speaker 00:
Yeah, so three responses to that, Judge Toronto. The first is they actually corrected it themselves in the second motion for reconsideration. If you look at Appendix 27,796, they interpreted it the way that we do, which is products sequentially process data. Is it 27 what?

[00:14:23] UNKNOWN:
27,796.

[00:14:23] Speaker 00:
It's, I believe, in the second volume of the appendix. So this was their second motion for reconsideration, and in it they block-quote it. and they correct it to read exactly the way that the district court read it and exactly the way that we read it. That's point number one. Point number two is I'd actually point you to the same motion itself.

[00:14:46] Speaker 00:
If you look at the sentence before, so we're now back at appendix 14,685. If you look at what they said in the sentence before, they said Intel's motion does not explain why the accused products do not also process sequences of data. And in fact, Dr. Conte has offered evidence that the accused product sequentially, the logical implication is that it's processed data because that's what they just said in the sentence before.

[00:15:13] Speaker 00:
And then if you look earlier in the motion at Appendix 14679, where they were talking about the 047 patent, there they say the products, quote, do sequentially process data. End quote. And they cite the exact same evidence that they cite here. But actually that's not the most powerful point on this that they weren't making a capability argument. The most powerful point you can find at appendix 11682 to 11683.

[00:15:43] Speaker 00:
That's volume one of the joint appendix.

[00:15:51] Speaker 00:
That is where, in responding to our summary judgment motion on marking, they specifically and clearly made the argument that the language adapted for and adapted to was not directed to capability. They were arguing exactly the opposite, and they said that it was not enough that a processor device, quote, can be adapted for practicing the limitation. That's Appendix 11,682 to 683. The opposite of what they're now arguing.

[00:16:20] Speaker 03:
I'm not sure it's the opposite. The problem is that adapted to itself... sometimes means capable of denying that it's capable of being adapted to is not the same thing as denying that it's capable of performing the functions.

[00:16:37] Speaker 00:
That's fair, Judge Toronto, and you're right. The term adapted to can take on different meanings. This court addressed that specifically in Aspects Eyewear versus Marchand. To the extent that you get past the issue of forfeiture and you're asking the question of what are these claims directed to, Aspects Eyewear teaches that you look at, well, how are other words used in neighboring claims? And in this one, I would point you to two things. Number one, the very claim itself also later uses the term capable of operating.

[00:17:07] Speaker 00:
So that, under the logic of aspects I wear, the fact that you've used adapted to, or excuse me, adapted for in one context means it means something different than capable because just a couple of lines later in the same claim, it refers to being capable of operating. It's also different language from the 047 patent. The 047 patent refers to things that are adaptable as opposed to being adapted to. But in all events, I think between their affirmative argument at 11,682 and the clear context from the other statements that they made that I recited and the district courts was in the best position to assess what had been argued to the court and made it very clear and pretty emphatic.

[00:17:52] Speaker 00:
At appendix 27 to 28, this wasn't the argument they made. I'm happy to answer any other questions about this patent, but I'm also happy to turn to the 593 patent at this point.

[00:18:01] Speaker 00:
Why don't you move on? So turning to the 593 patent, Judge Dyke, the district court properly construed the dedicated connection term based on its plain meaning. And that was confirmed by PAC's re-examination statements to require, quote, direct interconnection of a particular processor to a particular memory via a link inaccessible to other devices and memories. That's the district court's interpretation. And PACT reiterated to the patent office that, quote, in the context of electrical components and hardware, the dedicated means, quote, designed to interconnect exclusively.

[00:18:41] Speaker 00:
That was their statement at Appendix 33,956. So whether you think it's disclaimer or not is very, very clear, and indeed their argument to the Patent Office was this was the plain and ordinary meaning of the term. And so the district court applied the plain and ordinary meaning, construed it entirely consistent with their statements, and then their statements about the few reference itself are confirmatory. What they do with the few reference is they show the examiner that at least some sharing was taking place, and because there was some sharing, of these lines that Foo did not read on the requirement of having a dedicated connection.

[00:19:29] Speaker 00:
Now, they get up and say, well, what we were really telling the patent examiner was that all of the segments within Foo are connected.

[00:19:38] Speaker 00:
All of the segments within Foo are shared. That's actually not true, but let's just set that aside for a second because clearly the examiner understood that it was enough that some segments within Foo were shared. And the examiner applying their definition, this is Appendix 33,970, says, "...the plain and ordinary meaning is designed to directly interconnect exclusively..." A portion of the connection between the core and the memory is a shared connection.

[00:20:08] Speaker 00:
Hence, the connection is not a dedicated connection as defined above. That's 33,971. That is the examiner. And, of course, as this court has recognized in cases like LK v. EBCO and Sinclair v. Cannon and Biogen v. GSK, that's very informative as to understanding what it was they were arguing. And that should be the end of this case.

[00:20:33] Speaker 03:
What about the very last little argument that even under your construction, there really isn't any shared piece because the shared pieces are these just near the end points. And they talk about how they're, I don't know, Sometimes, maybe always, they say they're interfaces. They don't make much of that.

[00:21:03] Speaker 03:
My understanding, and just tell me if this is wrong, that the claim says you have to get from the processor to the memory. Their mapping for their infringement were very specific things for the processor and the memory. And these, what they're now calling interfaces, are actually not part of those two endpoints.

[00:21:28] Speaker 00:
I agree with that.

[00:21:29] Speaker 03:
So they have to be in the lake.

[00:21:31] Speaker 00:
I agree with that, Judge Taranto. The so-called interfaces are not the processor, they're not the memory. And indeed, they actually never presented any evidence to suggest that they were.

[00:21:47] Speaker 02:
So if you read their... They did have expert testimony that was part of the interface. What's missing, I think, is expert testimony that... that the interface, as it's called, is not part of the path.

[00:22:00] Speaker 00:
I think what's missing, I think what's missing, Judge Zach, I mean, I agree with that. I think what's missing is evidence that the interface is part of the memory. If you look at Appendix 33,980-983, yes, this is their expert, Conti, and the point that he was making was that the so-called secondary buss That that ends at the interface. That's the only point that he was making. There's no suggestion by him that it's not part of the path that right that it was not part of the secondary bus path that one part of the secondary bus path.

[00:22:40] Speaker 00:
Because the way the claim is written is that you have data processing cores include a physically dedicated connection to one of the memory units, and that that memory unit is also accessible by another of the data processing cores via secondary bus paths of the bus system. And his position is that the ring is the only thing that matters for purposes of the secondary bus path. But that doesn't answer the question of is there actually a physically dedicated connection that runs all the way from the processor to the memory?

[00:23:15] Speaker 00:
Because if it's, and it's indisputable that there is sharing. in order to get to the memory. Conti essentially concedes this at paragraph 180 of his opinion. They don't have an answer to what is figure two of the intel documents where we've got the circle showing the part that is shared. They don't have an answer to that other than to say it doesn't count. And their argument on appeal that it doesn't count is to say, well, maybe a fact finder could find the interface is part and parcel of the memory, but Conti doesn't say that.

[00:23:54] Speaker 00:
They have no expert to back them up on that. And indeed, under that approach, if you end up with a world where... everything is the interface, well then we win because there is no physically dedicated connection at all because there would be literally nothing outside the interface. The entirety of the so-called physically dedicated connection is inside the so-called interface and so they can't have it both ways.

[00:24:28] Speaker 00:
So to kind of summarize this point, their interface argument fails as a matter of law really for two independent reasons. One is it fails under the district court's construction of dedicated connection because the dedicated connection means there can be no sharing at all. and it has to run from the processor to the memory, and they have no evidence to support a different line drawing. The second reason it fails is because there is no genuine issue of material fact. That is, if you look at the confidential Intel figures, there is indisputably sharing before you get to the so-called interface.

[00:25:05] Speaker 00:
I mean, that is plain in the figure two that they rely on. If you look at it, the thing... You can just record.

[00:25:12] Speaker 03:
I won't make that point.

[00:25:14] Speaker 00:
The district court's point at Appendix 44 was really to say that the entire ring infrastructure ends up being shared. That is, that while the ring itself is not.

[00:25:30] Speaker 00:
If it's going from the processor to the memory, it's got to go through the SIBO and it's got to go through each of the three parts of the SIBO. The district court is absolutely correct about that and the intel figures on which they rely for their analysis. affirmative case demonstrate that. The only last thing I'll say is their so-called inconsistency argument about saying, well, the district court's inconsistent in what it did. First of all, that argument only applies under their claim construction. It doesn't apply under the district court's claim construction, number one.

[00:26:02] Speaker 00:
And number two, just to be clear, the district court didn't treat the question of whether or not something was an interface as being a relevant disputed fact. his question was whether or not there were separate paths within the interface because that's the claim because the claim requires that you have a dedicated connection and the whole idea is if it's dedicated well then it can't be shared I'm happy to address additional questions that the court may have on this I'm also happy to yield back the balance of my time thank you Judge Dyke Mr. Weiss first

[00:26:43] Speaker 01:
My friend on the other side said that Conte never made the ultimate point that the interface is not part of the connection. Instead, it's part of either the memory unit or the processor. Disagree with that. I refer the court to Appendix 33984, which is part of Conte's report. And what Conte says is that the fact that there are an overlapping portion at the very end of the road, if you will, does not negate the fact that the co-located CPU core, that's the processor, has a physically dedicated connection with the co-located LLC slice, which is the memory.

[00:27:17] Speaker 01:
Then I think it's helpful, he uses an analogy, which I found helpful, and he says, by analogy, a person of skill in the art would understand that different paths can be used to reach Rome by different travelers, even if those travelers are forced to walk up the same final steps to reach the gates of the city.

[00:27:34] Speaker 02:
Does he say that the interface is part of the memory?

[00:27:37] Speaker 01:
He's saying that the fact that there's this shared insignificant portion at the end does not negate... I'm looking for a language. Sure. Does not negate... It's in the fourth line of paragraph 179.

[00:27:50] Speaker 01:
does not negate the fact that the co-located CPU core has a physically dedicated connection. What he is saying there is this shared portion does not undermine that there is a physically dedicated connection. That's exactly the point. You have to determine under the clean language, is it part of the connection?

[00:28:07] Speaker 02:
He doesn't say it's part of the memory.

[00:28:09] Speaker 01:
Well, he's saying it's not part of the connection.

[00:28:13] Speaker 01:
You either have to be part of the connection or you have to be part of the memory or part of the processor core.

[00:28:19] Speaker 01:
In terms of what the examiner said beyond what we said, under Salazar and other cases, that does not count. We had no duty under the regulation that this court cited in Salazar to respond to the examiner's statements that went beyond ours.

[00:28:33] Speaker 01:
Now, I want to go to the district court's point about this inconsistency.

[00:28:38] Speaker 01:
I respectfully disagree because the district court When it came, whatever its claim construction was, when it came time for the district court to apply that to the evidence, the district court said, in Intel's products, there are no shared, there are no, excuse me, There are only shared segments. There's not even a single uniquely dedicated segment, and that is inconsistent, we submit, with his finding at Appendix 9. So the key passages, again, are Appendix 44 to 45 is the second order. Appendix 9 is the first. The district court seemed to be applying the claim construction that we think is appropriate there, but he applied it incorrectly to the facts.

[00:29:17] Speaker 01:
The segment that we pointed to, again, is this thing in the middle. There is something left of a connection, even if the other parts are part of the memory unit or core. It's the bump or the overpass that goes over the common line. The court has no further questions.

[00:29:32] Speaker 02:
Okay, thank you. Thank you. Just a caution. The confidentiality markings in the brief are very carelessly done. They mark as confidential things that are in the district court's unredacted opinion.

[00:29:48] Speaker 02:
that's not what you're supposed to do.

[00:29:50] Speaker 01:
Understood, Your Honor. Those were mostly at the initiative of Intel, whose confidentiality it was.

[00:29:57] Speaker 01:
Not packed, and we respected those, but perhaps inappropriately. I agree that these terms that I've tried to use today are in the district court's order publicly. Thank you, Your Honor.

[00:30:05] Speaker 02:
All right. Thank both counsel. The case is submitted. That concludes our session for this morning.