> If you fail to comply with our demands by the specified deadline, Synaptics will seek monetary
compensation from you in the amount of US$25,000 to recoup the costs and effort required to
create the attached documentation of your infringing acts.
Obviously this is intended as a scare tactic, but, surely that could never hold up in court? Surely the burden of proof is on the owner of the patents to prove that someone's work is infringing, and you can't then say "We are charging the defendant to prove that we have a valid case against them." ?
They can attempt to sue for whatever they want. The court may or may not accept the case. Seeking the recovery of costs for enforcement of ones patents if the defendant does not stop infringing is not under heard of, though it is typically attached to a case seeking the courts ruling that the defendant must stop infringing.
> I will also be sending to you our demand for substantial monetary compensation to recoup the cost and effort required to create such documentation of your infringing acts.
There is no universe in which this is a reasonable response.
I own two of their devices and some of the decomp fairies flittering through the gaming industry might just take a visit to their stack for some internal prodding…
Not really, all of law works that way. Everybody is allowed to demand pretty much anything from anybody for any reason, and include anything in any contract they're party of regardless of legality of any of the demands. It's the other party's responsibility to know which of these are legally valid and invalid.
DisplayLink is dying due to high DDR prices. Dock companies are cancelling their display link portfolios due to the unfavorable economics.
Good riddance if this is how OSS is treated by synaptics.
Most recent thunderbolt docks and usb4 docks have good display support, and fallback to DisplayPort alt mode if the host does not support tbt4. Which is much better supported over type-c. Intel and AMD upstream host support.
On my computer around 500 MB for two additional 4K monitors but I've heard up to 1 GB. With the RAM crunch, running DL + Docker + browser (esp Chrome), it gets pretty laggy sometimes.
What percentage does one gigabyte of RAM represent across the average amount of RAM in a laptop or computer? People who run multiple screens might spend more on RAM and storage, no?
I didn't think reverse engineering itself was a violation of a patent? The entire point of a patent is the public disclosure of the mechanism of an invention for the exclusive right to market that invention. Therefore arguing that reverse engineering is a violation of a patent is definitionally incorrect. They should be arguing it's a violation of trade secrets. Of course, they'd lose on that argument too because clean room reverse engineering of a trade secret is fine because you definitionally don't have a patent on it! That's the patent vs trade secret tradeoff!
Yes, but they're arguing in the letter that the work is _also_ disclosing information that would allow their competitors to know how DisplayLink works. You can't have it both ways.
Yeah, that statement clearly is making IP claims beyond patent infringement, they just didn't elaborate. They didn't say it's patent infringement because of that.
1. Reverse engineering for interoperability gives you a defense for copyright and trade secret claims in the US. It provides zero protection from patent claims.
2. I'm not saying their arguments are valid. It's a demand letter with multiple claims. There's many reasons it might be an invalid argument but listing more than one grievance and/or not explaining them in detail are not some of those reasons.
But lawyers are prone to using scary phrasing, even if that phrasing has no bite, and are free to write strongly-worded letters.
"I see that you've been publishing the privately-held, trade secret tomato processing technique I developed -- on the Internet! Now, tomato processors everywhere can see it! I'm really pissed! I demand that you stop doing this immediately! In fact, if you want a full enumeration of whar I really think about the terrible things you're doing, then I further demand that you pay me $25,000 for that effort!"
...it's got no bite, but it's a letter. It can be drafted and sent.
Is writting software "implementing," or is compiling it and releasing binaries (for free,) or just profiting from the release, from the perspective of patent law?
You can get around it by reverse engineering it as long as what's sold doesn't violate that patent.
Many companies have had to reverse engineer an existing patent to figure out what they can do instead.
> Design-arounds are considered to be one of the benefits of patent law. By providing monopoly rights to inventors in exchange for disclosing how to make and use their inventions, others are given both the information and incentive to invent competitive alternatives that design around the original patent. In the field of vaccines, for example, design-arounds are considered fairly easy. It is often possible to use the original patent as a guide for developing an alternative that does not infringe the original patent.
To be clear, a patent is an exclusive right to the specific functional invention, and if that invention happens to cover the one and only way to talk to the device, then it'll be hard to make a driver. In many cases you'll be able to make one for yourself due to some funny-business around patent exhaustion.
However, there is no universe where they can hold a valid patent that would encumber a driver for a regular USB display device. What they then might try to do is invoke DMCA, due to the device having DRM responsibilities (HDCP).
If they get trusted HDCP treatment from media players on some platforms, they might fight tooth and nail to not be broken open in public as that might end up with them losing trust from content providers...
Or maybe they're just assholes with no real rhyme or reason, who knows.
Umm... what? Implementing the things described in a patent is literally a patent violation regardless of how you go about it. Reverse engineering only gets a pass for trade secrets. As you said, a patent literally describes how to do the thing.
DisplayLink is a "clever" software workaround to hardware limitations that shouldn't exist anymore and I hope it goes away. When I got stuck with a DisplayLink dock at work it caused no end of audio issues when connecting and disconnecting, the driver had to be restarted all the time, it was a pain
Even "modern" systems like the Apple M1 and M2 base chips had hardware limits of 1 external monitor, a massive regression from the previous Intel systems.
It has worked for me pretty great on linux for years.
The problem it solves is that if you have 3 screens (you can often get 3 gpu ouptus) then you are already OCD enough so you will need 4th screen to even them out.
Every DisplayLink hub or dock or adapter I've tried using has been a terrible nightmare. I've tried on Linux, macOS, and Windows. Now I actively avoid them.
I've had terrible experiences with most of the big computer name brands making Displaylink hubs. Frequent driver issues, unexplainable compatibility issues, hardware failures, and more.
My only "wow this thing just works everywhere" was actually the Plugable brand docs. Obviously can't do things like 3D gaming, but for regular desktop kind of applications dang have these things been largely bulletproof. Especially the older USB 3.0 ones.
I still remember running five displays (including the built-in) off a single cheap HP Split X2 back in the day thanks to a Plugable dock.
Is DisplayLink a company owned by Synaptics or just a file in a portfolio? How do they make money from DisplayLink except from selling the docks? Licensing?
Per the article, even if they weren't trying to play semantic games about subsidiaries (which OIN really shouldn't allow), OIN excludes "codec" patents and considers the DisplayLink patents to be "codec" patents.
Reverse engineering is mostly to address trade secret and copyright infringement.
Patents can be avoided if you can prove that a patent is "obvious to a practitioner in the field" but that's a highish bar to prove in court and requires expensive lawyers
Nevertheless Synaptics are behaving as arse here and as a purely practical matter should be welcoming people who will bring them more customers for free
A patent is a description of how something is solved, such that someone with expertise can recreate it. That patent then grants a monopoly on that method for a period of 20 years in the US, if you keep paying renewal fees.
The whole reverse engineering clean room thing, is to show no copyright infringement and / or trade secret (primarily trade secret). How could they intentionally copy the code, if they cannot see the code.
Patents are far stronger in that sense, but they do have an expiration date. People independently come up with techniques and infringe patents all the time. Look up the patent litigation history of CRISPR which are still going strong it …
It’s also disturbing they have the gall to demand tens of thousands of dollars for the task of specifying which of their 100+ maybe relevant patents he’s infringing.
I finally upgraded all my gear to USB3.1Gen2+ so I no longer need to use DisplayLink. It's so nice. No more messing around with weird broken drivers. I just plug things in and they work.
> The letter continued to claim that Lothian would be "single-handedly eviscerating DisplayLink's competitive advantage in this crowded market by educating our competitors and other third parties as to how we implement our proprietary technology"
If one guy can reverse-engineer everything with an LLM and some free time, I'm pretty sure their competitors can as well, and they won't be doing it out in the open like this guy was...
this is a good point. Isn't one of the pre-conditions for a patent (in the US at least, this case was UK) that it be "non-obvious" ? If LLMs can shit out working solutions that infringe patents, or "trivially" reverse engineer protocols, I wonder if we will see challenges in court that it was "obvious" to ask the LLM, and therefore the patents were trivial and therefore should be invalidated?
That seems to be a separate claim unrelated to their patent claims. Patents themselves are required to have all of the information that would allow someone to implement what is in the patent. Of course, there could be many ways to implement it. And a specific implementation could be further protected by trade secrets. The part that would be a trade secret is inherently not in the patent.
You have to wonder if this part was approved by a qualified lawyer at all. A trade secret would describe proprietary knowledge of how to implement the technology, but would have no protection under patent law.
They also repeatedly claimed "DisplayLink is not a member of OIN" when it actually is, and later when they acquiesced to the cease and desist, one of the actual engineers from the company tried to review the PR to ask for changes. It's not clear anyone there has any clue what's going on!
I suspect they know that, but it reads as if they were just lazily tacking that claim on for effect. This isn't a petition to a court, this is just a demand letter. Listing out a bunch of scary damages is how demand letters often work. They don't have to go into details as to the basis of their legal claims.
At some point there needs to be a test of when software is related to owning hardware.
Some questions that come to mind are:
Do you actually own the hardware?
Do you stop owning the hardware, meaning you're not allowed or able to use it once the company that's sold it to you stops providing or updating hardware?
Does every piece of hardware come with a guarantee that it will work on every piece of hardware that it's attached to forever?
These are just a few of the really troubling questions when people forget that software has to run on hardware. And the closer the software is to the core operation of a piece of hardware and equipment, the more topics like right to repair or right to maintain or right to optimize come up.
I never thought that a video card replacement driver would make me think of companies that are limiting Farmers from repairing and updating their own tractors.
> The letter is dated "October 5, 2026". WTF are these lawyers smoking that they think 5 days is any reasonable time frame.
They just try to keep the pressure up so that the other entity gives up before they have time to estimate whether any of the claims are valid. One person against big corporation, which is threatening with big penalties. A common tactic.
Seriously, use OCuLink. It costs $15 for the pcie card, $10 for a cable, and $100 for a GFX card dock, and a PSU.
OCuLink doesn't even use drivers. It's literally just a PCIE passthrough. All usual tools apply, whereas DisplayLink encrypts all data, so a big fuckyou for getting logs.
Quit giving trash companies pounds of money for abuse at scale.
LWN operates a kind of soft paywall as a form of try-before-you-buy
> The following subscription-only content has been made available to you by an LWN subscriber. Thousands of subscribers depend on LWN for the best news from the Linux and free software communities. If you enjoy this article, please consider subscribing to LWN. Thank you for visiting LWN.net!
> It's ok to post stories from sites with paywalls that have workarounds.
> In comments, it's ok to ask how to read an article and to help other users do so. But please don't post complaints about paywalls. Those are off topic. More here.
According to patent prosecution records, the independent claims for GB2484736 originally lacked any network transport limitations and were deemed entirely unpatentable against the prior art listed above.
To get the patent granted, DisplayLink’s attorneys amended the claims to specifically add delivery of the display data over a "general-purpose data network." The validity challenge relies on proving that adding a general-purpose transport layer (like USB or Ethernet) to existing image-compression methodologies was an obvious step and completely unoriginal in 2010, given the widespread existence of USB monitors and network-attached graphics at the time.
The patent generally covers methods and architectures for transmitting compressed screen updates over a general-purpose data network, such as USB or Ethernet, the patent specifically claims Haar-transform decoding.
So only the combination may be new. But this reminds me of a USPTO thing from my past.
The landmark ruling you are referring to is the U.S. Supreme Court case KSR International Co. v. Teleflex Inc. (2007), which led directly to the USPTO publishing its highly influential 2010 Examination Guidelines for Determining Obviousness.
Under this framework, if element A is known and element B is known, combining them is not patentable if the combination yields nothing more than predictable results.
So, general-purpose data network, such as USB or Ethernet was known (A), claims Haar-transform decoding was known (B) [DE69909880T2 / EP0955609]. Even if we don't find this combination in prior art, it is dubious that this is patent worthy based on USPTO guidelines (combine A with B).
But I see a lot of stuff that suggests prior art:
1. Noriaki Yamazaki and Tetsuo Tanaka (2007). “ZYWRLE: Enhanced VNC Codec for Motion Picture.” IEICE Transactions, J90-D(7), pp. 1790–1798.
Brief summary: Adds wavelet compression to VNC’s ZRLE codec for transmitting moving images, reducing network traffic to 20–50% of standard ZRLE.
https://search.ieice.org/bin/summary.php?abst=j&category=D&i...
2. Joshua G. Senecal, Peter Lindstrom, Mark A. Duchaineau and Kenneth I. Joy (2004). “An Improved N-Bit to N-Bit Reversible Haar-Like Transform.” Proceedings of Pacific Graphics 2004, pp. 371–380. DOI: 10.1109/PCCGA.2004.1348368.
Brief summary: Introduces PLHaar, a reversible Haar-like transform supporting lossy and lossless image compression without increasing coefficient bit width, suitable for bandwidth-limited video and graphics hardware.
https://www.osti.gov/servlets/purl/15014239
3. Joshua G. Senecal, Mark A. Duchaineau and Kenneth I. Joy (2004). “Reversible n-Bit to n-Bit Integer Haar-Like Transforms.” Proceedings of the Data Compression Conference 2004, p. 564.
Brief summary: Introduces lookup-table Haar-like transforms that preserve bit width and reversibility, with applications to fast image compression and streaming hardware such as DVI channels.
https://www.osti.gov/servlets/purl/15013668
Is this a potential use case for having a gen AI agent grind to put a prior art claim together? Feels like scanning and building prior art to challenge existing patents is a great use of AI agents, but I might be mistaken.
You assume there is some logic to software patents. It looks like an extortion racket to me. If you don’t pay enough fire insurance the mob lets your competition burn your business down.
If software patents are slop, you fight slop with slop.
It would take much human time to search for prior art across all of human knowledge for existing patents and new patents being proposed. But GPUs are tireless, and can continually grind towards success. They build their case for dismantling all existing active patents as long as they are kept fed.
If I were a frontier lab, I would be finding the most valuable patents, starting the research agents up, and then shorting companies that would see enterprise value decline if subject patents were invalidated (think Hindenburg or Culper Research activist investors). Just one way of monetizing the fancy search engine and token dispenser.
> I predict a high chance that they lose this patent. Mark my words.
After how many dollars spent? I mean, it is good that eventually bad patents get killed, but it still costs too much to do it, one at a time. We need a more streamlined process, maybe? Maybe an academic-review-like process at filing? Send the claims anonymously to a few experts in the field, randomly selected. If they all say it is obvious, kill it immediately and fine the company for wasting the patent office's time?
The most disturbing part is that a single developer, with some AI assistance, was able to improve hardware support that the manufacturer couldn't or wouldn't fix. Instead of embracing that work, Synaptics threatened to financially ruin him. As AI lowers the barrier to reverse engineering and interoperability, I suspect we'll see a lot more companies reaching for lawyers when their technical moats start disappearing.
> If you fail to comply with our demands by the specified deadline, Synaptics will seek monetary compensation from you in the amount of US$25,000 to recoup the costs and effort required to create the attached documentation of your infringing acts.
Obviously this is intended as a scare tactic, but, surely that could never hold up in court? Surely the burden of proof is on the owner of the patents to prove that someone's work is infringing, and you can't then say "We are charging the defendant to prove that we have a valid case against them." ?
They can attempt to sue for whatever they want. The court may or may not accept the case. Seeking the recovery of costs for enforcement of ones patents if the defendant does not stop infringing is not under heard of, though it is typically attached to a case seeking the courts ruling that the defendant must stop infringing.
"You've imposed costs on us, because of all the effort we had to go to to show that you've imposed costs on us" ??
> I will also be sending to you our demand for substantial monetary compensation to recoup the cost and effort required to create such documentation of your infringing acts.
There is no universe in which this is a reasonable response.
Synaptics: “Look what you made me do to you!”
Open source community "Look what you made us do":
DisplayLink on the driver side just compresses frames to send them over USB.
It would be interesting to analyse if Synaptics haven't stolen code from open source ffmpeg or VLC to do the compression.
I own two of their devices and some of the decomp fairies flittering through the gaming industry might just take a visit to their stack for some internal prodding…
I know, right!
“You are stealing our ideas! But we won’t tell you which ones unless you pay us”
In any other domain this would be called a protection racket.
Not really, all of law works that way. Everybody is allowed to demand pretty much anything from anybody for any reason, and include anything in any contract they're party of regardless of legality of any of the demands. It's the other party's responsibility to know which of these are legally valid and invalid.
DisplayLink is dying due to high DDR prices. Dock companies are cancelling their display link portfolios due to the unfavorable economics.
Good riddance if this is how OSS is treated by synaptics.
Most recent thunderbolt docks and usb4 docks have good display support, and fallback to DisplayPort alt mode if the host does not support tbt4. Which is much better supported over type-c. Intel and AMD upstream host support.
How much ram does DisplayLink use?
On my computer around 500 MB for two additional 4K monitors but I've heard up to 1 GB. With the RAM crunch, running DL + Docker + browser (esp Chrome), it gets pretty laggy sometimes.
What percentage does one gigabyte of RAM represent across the average amount of RAM in a laptop or computer? People who run multiple screens might spend more on RAM and storage, no?
500€ for 32 GB or ram is a stupid price no matter how many displays you own
I didn't think reverse engineering itself was a violation of a patent? The entire point of a patent is the public disclosure of the mechanism of an invention for the exclusive right to market that invention. Therefore arguing that reverse engineering is a violation of a patent is definitionally incorrect. They should be arguing it's a violation of trade secrets. Of course, they'd lose on that argument too because clean room reverse engineering of a trade secret is fine because you definitionally don't have a patent on it! That's the patent vs trade secret tradeoff!
no, a patent is exclusive right to implement, you can't get around it by reverse engineering
the tradeoff is that it becomes free eventually, unlike trade secrets
Yes, but they're arguing in the letter that the work is _also_ disclosing information that would allow their competitors to know how DisplayLink works. You can't have it both ways.
Yeah, that statement clearly is making IP claims beyond patent infringement, they just didn't elaborate. They didn't say it's patent infringement because of that.
Merely reverse engineering for interoperability is not a violation of any type of IP in the US.
1. Reverse engineering for interoperability gives you a defense for copyright and trade secret claims in the US. It provides zero protection from patent claims.
2. I'm not saying their arguments are valid. It's a demand letter with multiple claims. There's many reasons it might be an invalid argument but listing more than one grievance and/or not explaining them in detail are not some of those reasons.
A thing can be encumbered by both patents and trade secrets, can it not?
well technically, yes and no
if I patent a microwave and specify exactly how it works, that part can't be trade secret
BUT if my trade secret is on using ruby diamond for deluxe premium microwave lazor THAT isn't specified on the patent, well that be a trade secret
but trade secret CAN ONLY BE PUNISHED IF someone insider leaks the info -- not much can be done if some junk collector finds out about the ruby stuff
Right.
But lawyers are prone to using scary phrasing, even if that phrasing has no bite, and are free to write strongly-worded letters.
"I see that you've been publishing the privately-held, trade secret tomato processing technique I developed -- on the Internet! Now, tomato processors everywhere can see it! I'm really pissed! I demand that you stop doing this immediately! In fact, if you want a full enumeration of whar I really think about the terrible things you're doing, then I further demand that you pay me $25,000 for that effort!"
...it's got no bite, but it's a letter. It can be drafted and sent.
(It can also be ignored upon receipt.)
Is writting software "implementing," or is compiling it and releasing binaries (for free,) or just profiting from the release, from the perspective of patent law?
There is no "fair-use" or non-commercial exception to patent law.
You can get around it by reverse engineering it as long as what's sold doesn't violate that patent. Many companies have had to reverse engineer an existing patent to figure out what they can do instead.
https://en.wikipedia.org/wiki/Design_around
> Design-arounds are considered to be one of the benefits of patent law. By providing monopoly rights to inventors in exchange for disclosing how to make and use their inventions, others are given both the information and incentive to invent competitive alternatives that design around the original patent. In the field of vaccines, for example, design-arounds are considered fairly easy. It is often possible to use the original patent as a guide for developing an alternative that does not infringe the original patent.
To be clear, a patent is an exclusive right to the specific functional invention, and if that invention happens to cover the one and only way to talk to the device, then it'll be hard to make a driver. In many cases you'll be able to make one for yourself due to some funny-business around patent exhaustion.
However, there is no universe where they can hold a valid patent that would encumber a driver for a regular USB display device. What they then might try to do is invoke DMCA, due to the device having DRM responsibilities (HDCP).
If they get trusted HDCP treatment from media players on some platforms, they might fight tooth and nail to not be broken open in public as that might end up with them losing trust from content providers...
Or maybe they're just assholes with no real rhyme or reason, who knows.
Umm... what? Implementing the things described in a patent is literally a patent violation regardless of how you go about it. Reverse engineering only gets a pass for trade secrets. As you said, a patent literally describes how to do the thing.
DisplayLink is a "clever" software workaround to hardware limitations that shouldn't exist anymore and I hope it goes away. When I got stuck with a DisplayLink dock at work it caused no end of audio issues when connecting and disconnecting, the driver had to be restarted all the time, it was a pain
Even "modern" systems like the Apple M1 and M2 base chips had hardware limits of 1 external monitor, a massive regression from the previous Intel systems.
It has worked for me pretty great on linux for years.
The problem it solves is that if you have 3 screens (you can often get 3 gpu ouptus) then you are already OCD enough so you will need 4th screen to even them out.
Might be related to the acquisition by onsemi:
https://www.onsemi.com/company/newsroom/news-and-insights/on...
My company used to have some display link docks, which I was unable to use due to bugs in the Linux implementations.
For a while I'd keep an eye on updates to see if they'd been fixed but they never were before we moved office.
I'm going to be quite sure we don't get any more in any future offices (we don't have them at our current office)
Every DisplayLink hub or dock or adapter I've tried using has been a terrible nightmare. I've tried on Linux, macOS, and Windows. Now I actively avoid them.
Mine works just fine with the newest Dell XPS 14 laptop and Ubuntu.
I've had terrible experiences with most of the big computer name brands making Displaylink hubs. Frequent driver issues, unexplainable compatibility issues, hardware failures, and more.
My only "wow this thing just works everywhere" was actually the Plugable brand docs. Obviously can't do things like 3D gaming, but for regular desktop kind of applications dang have these things been largely bulletproof. Especially the older USB 3.0 ones.
I still remember running five displays (including the built-in) off a single cheap HP Split X2 back in the day thanks to a Plugable dock.
Synaptics should lose all access to Linux.
Is DisplayLink a company owned by Synaptics or just a file in a portfolio? How do they make money from DisplayLink except from selling the docks? Licensing?
They make money by authoring standards which force vendors to buy even more parts/licenses from them.
DisplayLink was acquired by Synaptics in 2022
So they are and aren't part of OIN at the same time.
Per the article, even if they weren't trying to play semantic games about subsidiaries (which OIN really shouldn't allow), OIN excludes "codec" patents and considers the DisplayLink patents to be "codec" patents.
By some definition, anything is a codec :clownface:
Schrödinger’s membership. Exists when it helps them, they’ve never heard of it when it doesn’t
>> An independently written or clean-room implementation does not necessarily avoid patent infringement if it implements patented inventions.
Can someone explain this to me? Isn't this the whole premise of reverse engineering?
Reverse engineering is mostly to address trade secret and copyright infringement.
Patents can be avoided if you can prove that a patent is "obvious to a practitioner in the field" but that's a highish bar to prove in court and requires expensive lawyers
Nevertheless Synaptics are behaving as arse here and as a purely practical matter should be welcoming people who will bring them more customers for free
A patent is a description of how something is solved, such that someone with expertise can recreate it. That patent then grants a monopoly on that method for a period of 20 years in the US, if you keep paying renewal fees.
The whole reverse engineering clean room thing, is to show no copyright infringement and / or trade secret (primarily trade secret). How could they intentionally copy the code, if they cannot see the code.
Patents are far stronger in that sense, but they do have an expiration date. People independently come up with techniques and infringe patents all the time. Look up the patent litigation history of CRISPR which are still going strong it …
https://www.reuters.com/legal/litigation/nobel-prize-winners...
There’s a million examples.
For context, here’s my company IP Copilot - https://ipcopilot.ai/
It’s also disturbing they have the gall to demand tens of thousands of dollars for the task of specifying which of their 100+ maybe relevant patents he’s infringing.
Does UK patent law somehow work differently?
Microsoft Is Making An Astonishing $2 Billion Per Year From Android Patent Royalties[0].
[0]: https://www.businessinsider.com/microsoft-earns-2-billion-pe...
> This money, says Sherlund, helps Microsoft hide the fact that its mobile and Xbox groups are burning serious cash.
Little did they know how much bigger that pyre would get lmao
I finally upgraded all my gear to USB3.1Gen2+ so I no longer need to use DisplayLink. It's so nice. No more messing around with weird broken drivers. I just plug things in and they work.
> The letter continued to claim that Lothian would be "single-handedly eviscerating DisplayLink's competitive advantage in this crowded market by educating our competitors and other third parties as to how we implement our proprietary technology"
If one guy can reverse-engineer everything with an LLM and some free time, I'm pretty sure their competitors can as well, and they won't be doing it out in the open like this guy was...
this is a good point. Isn't one of the pre-conditions for a patent (in the US at least, this case was UK) that it be "non-obvious" ? If LLMs can shit out working solutions that infringe patents, or "trivially" reverse engineer protocols, I wonder if we will see challenges in court that it was "obvious" to ask the LLM, and therefore the patents were trivial and therefore should be invalidated?
That seems to be a separate claim unrelated to their patent claims. Patents themselves are required to have all of the information that would allow someone to implement what is in the patent. Of course, there could be many ways to implement it. And a specific implementation could be further protected by trade secrets. The part that would be a trade secret is inherently not in the patent.
You have to wonder if this part was approved by a qualified lawyer at all. A trade secret would describe proprietary knowledge of how to implement the technology, but would have no protection under patent law.
They also repeatedly claimed "DisplayLink is not a member of OIN" when it actually is, and later when they acquiesced to the cease and desist, one of the actual engineers from the company tried to review the PR to ask for changes. It's not clear anyone there has any clue what's going on!
You're saying I can't use a shell company acquisition as a means to shield certain technologies from my other agreements? Booo
I suspect they know that, but it reads as if they were just lazily tacking that claim on for effect. This isn't a petition to a court, this is just a demand letter. Listing out a bunch of scary damages is how demand letters often work. They don't have to go into details as to the basis of their legal claims.
I wish Synaptics a very "find this work fully completed by someone on https://gitflic.ru/" day.
I support Federico Faggins company
At some point there needs to be a test of when software is related to owning hardware.
Some questions that come to mind are:
Do you actually own the hardware?
Do you stop owning the hardware, meaning you're not allowed or able to use it once the company that's sold it to you stops providing or updating hardware?
Does every piece of hardware come with a guarantee that it will work on every piece of hardware that it's attached to forever?
These are just a few of the really troubling questions when people forget that software has to run on hardware. And the closer the software is to the core operation of a piece of hardware and equipment, the more topics like right to repair or right to maintain or right to optimize come up.
I never thought that a video card replacement driver would make me think of companies that are limiting Farmers from repairing and updating their own tractors.
If you want to know whether any relatively modern computer hardware is encumbered by IP restrictions, it's pretty simple, the answer is yes.
The closest you can get is something like the Raptor Talos II, although parts of the hardware are still under patent.
[dupe] https://news.ycombinator.com/item?id=50363968
"Your actions are potentially putting tens of millions of dollars of yearly corporate revenue at risk."
https://static.lwn.net/images/pdf/Cease_and_Desist_Notice_Mi...
Fuck you, Synaptics.
Felony contempt of business model?
https://www.eff.org/deeplinks/2019/06/felony-contempt-busine...
> Fuck you, Synaptics.
Which is made funnier by the lawyer's name being Bateman. Sorry, could not resist that one.
What I find more ridiculous is the timing:
> You are to provide Synaptics, no later than 11:59 p.m. Pacific Time on October 11, 2026
The letter is dated "October 5, 2026". WTF are these lawyers smoking that they think 5 days is any reasonable time frame.
> The letter is dated "October 5, 2026". WTF are these lawyers smoking that they think 5 days is any reasonable time frame.
They just try to keep the pressure up so that the other entity gives up before they have time to estimate whether any of the claims are valid. One person against big corporation, which is threatening with big penalties. A common tactic.
>Yes, you should be concerned. Very concerned, actually
Unbelievable mafia tactics from those clowns.
Seriously, use OCuLink. It costs $15 for the pcie card, $10 for a cable, and $100 for a GFX card dock, and a PSU.
OCuLink doesn't even use drivers. It's literally just a PCIE passthrough. All usual tools apply, whereas DisplayLink encrypts all data, so a big fuckyou for getting logs.
Quit giving trash companies pounds of money for abuse at scale.
Available only to subscribers, so, most of HN can't see this page I guess.
EDIT: Fixed now.
AFAIK, links to LWN pages are intentionally always visible to non-subscribers.
In any case, I am not a subscriber and I see it.
Subscribers can create a free link to the content. And all articles become publicly available after one week of paywall.
It seems they made it available.
LWN operates a kind of soft paywall as a form of try-before-you-buy
> The following subscription-only content has been made available to you by an LWN subscriber. Thousands of subscribers depend on LWN for the best news from the Linux and free software communities. If you enjoy this article, please consider subscribing to LWN. Thank you for visiting LWN.net!
Yes, I'm able to see it now.
https://news.ycombinator.com/newsfaq.html
> Are paywalls ok?
> It's ok to post stories from sites with paywalls that have workarounds.
> In comments, it's ok to ask how to read an article and to help other users do so. But please don't post complaints about paywalls. Those are off topic. More here.
I predict a high chance that they lose this patent. Mark my words.
https://lkml.rescloud.iu.edu/hypermail/linux/kernel/2610.1/1...
According to patent prosecution records, the independent claims for GB2484736 originally lacked any network transport limitations and were deemed entirely unpatentable against the prior art listed above.
To get the patent granted, DisplayLink’s attorneys amended the claims to specifically add delivery of the display data over a "general-purpose data network." The validity challenge relies on proving that adding a general-purpose transport layer (like USB or Ethernet) to existing image-compression methodologies was an obvious step and completely unoriginal in 2010, given the widespread existence of USB monitors and network-attached graphics at the time.
> mapping claim 11 of DisplayLink GB Patent No. 2484736 and claim 11 of DisplayLink GB Patent No. 2485613 to your Vino driver source code
As far as I can tell, these patents only go up to claim 10 in the first place:
- https://patents.google.com/patent/GB2484736A/en
- https://patents.google.com/patent/GB2485613A/en
Strange, I noticed that too. I couldn't find any other patent reference for GB Patent No. 2485613 with more than 10 claims.
Secret FISA Patent Claim
U not cleared to see it
The patent generally covers methods and architectures for transmitting compressed screen updates over a general-purpose data network, such as USB or Ethernet, the patent specifically claims Haar-transform decoding.
So only the combination may be new. But this reminds me of a USPTO thing from my past.
The landmark ruling you are referring to is the U.S. Supreme Court case KSR International Co. v. Teleflex Inc. (2007), which led directly to the USPTO publishing its highly influential 2010 Examination Guidelines for Determining Obviousness. Under this framework, if element A is known and element B is known, combining them is not patentable if the combination yields nothing more than predictable results.
https://www.uspto.gov/web/offices/pac/mpep/s2143.html
So, general-purpose data network, such as USB or Ethernet was known (A), claims Haar-transform decoding was known (B) [DE69909880T2 / EP0955609]. Even if we don't find this combination in prior art, it is dubious that this is patent worthy based on USPTO guidelines (combine A with B).
But I see a lot of stuff that suggests prior art:
1. Noriaki Yamazaki and Tetsuo Tanaka (2007). “ZYWRLE: Enhanced VNC Codec for Motion Picture.” IEICE Transactions, J90-D(7), pp. 1790–1798. Brief summary: Adds wavelet compression to VNC’s ZRLE codec for transmitting moving images, reducing network traffic to 20–50% of standard ZRLE. https://search.ieice.org/bin/summary.php?abst=j&category=D&i...
2. Joshua G. Senecal, Peter Lindstrom, Mark A. Duchaineau and Kenneth I. Joy (2004). “An Improved N-Bit to N-Bit Reversible Haar-Like Transform.” Proceedings of Pacific Graphics 2004, pp. 371–380. DOI: 10.1109/PCCGA.2004.1348368. Brief summary: Introduces PLHaar, a reversible Haar-like transform supporting lossy and lossless image compression without increasing coefficient bit width, suitable for bandwidth-limited video and graphics hardware. https://www.osti.gov/servlets/purl/15014239
3. Joshua G. Senecal, Mark A. Duchaineau and Kenneth I. Joy (2004). “Reversible n-Bit to n-Bit Integer Haar-Like Transforms.” Proceedings of the Data Compression Conference 2004, p. 564. Brief summary: Introduces lookup-table Haar-like transforms that preserve bit width and reversibility, with applications to fast image compression and streaming hardware such as DVI channels. https://www.osti.gov/servlets/purl/15013668
Good luck synaptic!
There's also a 2003 datasheet the describes basically the same compressed framebuffer transport as GB2485613: https://www.amd.com/content/dam/amd/en/documents/archived-te..., section 5.5.2
Also possibly related:
Method and apparatus for increasing memory resource utilization in an information stream decoder, https://patents.google.com/patent/WO1999057908A1/en
And a paper by a PHD student in 2005, Real-time Rendering with Wavelet-Compressed Multi-Dimensional Datasets on the GPU, https://cs.ucsb.edu/sites/default/files/documents/2005-05.pd...
Displaylink's own abandoned patent application talking about sending patches of compressed/uncompressed graphical data to the screen over USB: https://patents.google.com/patent/US20060119602A1/en
Using reference frames with Haar wavelet transforms in video compression, https://patents.google.com/patent/US6272180B1/en
Is this a potential use case for having a gen AI agent grind to put a prior art claim together? Feels like scanning and building prior art to challenge existing patents is a great use of AI agents, but I might be mistaken.
https://www.youtube.com/watch?v=v5d8nwgiSYI
Brilliant, thanks for the share.
You assume there is some logic to software patents. It looks like an extortion racket to me. If you don’t pay enough fire insurance the mob lets your competition burn your business down.
If software patents are slop, you fight slop with slop.
It would take much human time to search for prior art across all of human knowledge for existing patents and new patents being proposed. But GPUs are tireless, and can continually grind towards success. They build their case for dismantling all existing active patents as long as they are kept fed.
If I were a frontier lab, I would be finding the most valuable patents, starting the research agents up, and then shorting companies that would see enterprise value decline if subject patents were invalidated (think Hindenburg or Culper Research activist investors). Just one way of monetizing the fancy search engine and token dispenser.
https://www.cloudflare.com/jengo/sable-prior-art-search/
https://www.eff.org/issues/patent-busting-project
X11? RDP?
> I predict a high chance that they lose this patent. Mark my words.
After how many dollars spent? I mean, it is good that eventually bad patents get killed, but it still costs too much to do it, one at a time. We need a more streamlined process, maybe? Maybe an academic-review-like process at filing? Send the claims anonymously to a few experts in the field, randomly selected. If they all say it is obvious, kill it immediately and fine the company for wasting the patent office's time?
Ex Parte Reexamination (A streamlined request to have a USPTO examiner re-evaluate the patent) $6,775 (Streamlined, ≤ 40 pages)
I think we could get the 7 grand together here.
This is a UK patent, the USPTO has no jurisdiction over it.
In the UK, there is an administrative "reexamination" process that is significantly cheaper... only 250 UKP: https://www.gov.uk/government/publications/request-for-opini...
If someone wants to do the work to put together and file the claim, I'll provide the ₤250. Worth it to see attempted patent enforcement destroyed.
The most disturbing part is that a single developer, with some AI assistance, was able to improve hardware support that the manufacturer couldn't or wouldn't fix. Instead of embracing that work, Synaptics threatened to financially ruin him. As AI lowers the barrier to reverse engineering and interoperability, I suspect we'll see a lot more companies reaching for lawyers when their technical moats start disappearing.
Disturbing that this pretty reasonable take was flagged.
The patent system itself has been a moat and we will see more and more reasonable submissions that will force a rethink.