Right now, if you go onto Espacenet and search "Quantum Computer", search by priority, and then see which nation has the most published applications for patents relating to quantum computer technology, you would see that roughly 95% of all recently published quantum computer patent applications are being filed in China. Does this mean that China is dwarfing all other nations such as the United States, India, Germany, and Russia in quantum computer development? Does this mean that scientists, engineers, and inventors in China really value patents more than those in other nations? Does this mean that researchers globally prefer to file Quantum Computing patent applications in China? The answer to all three questions is an equivocal no.
A lot of researchers in China stress patenting incremental, isolated improvements in the field of Quantum Computing, while researchers in other nations stress broad, holistic, sweeping patents in this field that describe lasting frameworks. For instance, consider this recently granted American patent, US10581616B1, "Managing nodes of a cryptographic hash tree in a hash-based digital signature scheme," which was granted to a Canadian company, Isara, on March 3, 2020. This patent describes an extremely novel way to use conventional electronic computers, i.e., NON-QUANTUM computers, to defend against cyber-attacks from quantum computers. This patent changes cryptography and cyber-security as we know it. Before any readers of this blog post say that this is not a quantum computer patent, know that this patent also profoundly changes quantum computing as well. Why? It reduces the short-term capabilities of quantum computers which reduces the utility of many quantum computing innovations. For instance, consider this patent application recently filed in China, CN110868297A, which roughly translates to "A Method of Improving RSA Decryption." The encryption techniques described in US10581616B1, if implemented, make the decryption methods found in CN110868297A largely obsolete. CN110868297A, in light of US10581616B1, would probably be found as lacking enough utility despite the fact that it appears to describe an incremental innovation in decryption. I think it's an example of different patent philosophies in different nations and different fields. What's so interesting to me is that other kinds of computer hardware patents in America have usually reflected the incremental philosophy that we're seeing in many quantum computing patents in China today.
0 Comments
I just wanted to quickly share this patent:
https://patents.google.com/patent/US4328470A/en In the wake of a global pandemic of a relatively unknown virus, COVID-19 a.ka. the coronavirus, one company is trying to have testkits pulled from the market. That company is called Labrador Diagnostics. Labrador bought several patents from a company called Softbank, who in turn bought them from a company called Theranos. Theranos is best known as the company that claimed to have invented a new and expedient kind of blood test that turned out to be a complete fraud. However, Labrador is insisting that coronavirus-testkit manufacturer and developer, BioFire, has infringed on its patents acquired from Theranos and is asking for an injunction of the sale of the testkits.
As a matter of public policy, this is dangerous. All other forms of property have temporary limits for emergency purposes, so it makes sense that patents should as well. Additionally, Labrador could simply ask for reasonable royalties, request negotiations for a sale of the patent, set up a patent pool for pandemic response, or attempt to market the testkits themselves in good faith. Instead of any of these sane responses to a global health crisis, Labrador has chosen to ask for an injunction on the sale of a product preventing millions of deaths. From a legal perspective, Labrador's request for an injunction certainly fails the eBay v. MercExchange standard needed for an injunction given the public need for testkits right now. However, there is a deeper problem--this Theranos patent, at least in certain aspects, is gibberish. One key patent in question is "Point-of-care fluidic systems and uses thereof"(US8283155B2). One key element of Claim 1, element b, reads "an external device configured to transmit a protocol to the communication assembly ". Unless this "external device" is literally building a new world-wide-web or alternative like Gopher, it can't "transmit a protocol." Protocols, in a Computer Science context, are how computers share data with each other. Data is transfered with protocols; the protocols are not transfered themselves. This reads like gibberish. The gibberish doesn't end there--Claim 30 of this patent seems to be gibberish too. There are many things that are strange about claim 30 such as nested 'comprising' terms without the phrase 'further comprising', but the biggest oddity of claim 30 is the phrase "wherein said protocol in turn effects a selection of at least one assay from said plurality of assays to be run with said assay assembly for generating said signal." 'Effect' is a noun 99% of the time in English, and this is a rare instance of it being used as a verb in a formal document. The use of 'effect' as a verb is important here because it is quite different from 'affect.' 'Affect' means to influence, but 'effect', as a verb, means to achieve. How could the mere selection of an assay, which is dictated by a processing algorithm according to the spec, be regarded as a specific achievement? Perhaps the author meant to say "affects a selection", which means that it causes the correct assay to take place, but it is really unclear. Last, but not least, I want to point out that Labrador Diagnostics' recent actions are really bad for Patent Law in so many ways. First of all, it hurts Patent Law's image with the general public. Patents serve an extremely important purpose--to incentivize the disclosure of innovations. Right now, a growing portion of the general public is turning against patents because of companies like Labrador Diagnostics who are egregiously abusing a system with almost psychopathic regard to society, and much of the public doesn't see the benefit that patents provide. Secondly, it weakens the patent system for everybody else. Some of the biggest news on the litigation side of Patent Law recently has been IBM's suit of Airbnb for patent infringement. IBM has become known for its litigation in recent years. In fact, a few years ago, some commentators were labelling the once-mighty tech giant a mere "patent troll" for its more recent litigation(https://www.markmanadvisors.com/blog/2018/8/2/is-ibm-a-patent-troll). However, once IBM joined LOT, it was thought that IBM would be far more hesitant to sue for patent infringement.
LOT stands for License on Transfer, and it's basically an organization founded upon companies promising not to sell their patents to a "patent assertion entity" WITHOUT immediately licensing the said patents to every member in LOT. So if IBM has patent X, and IBM decides to sell patent X to a patent troll, every member of LOT immediately gains a license to patent X. Airbnb joined LOT on January 22, 2020, literally two weeks after IBM joined. However, this lawsuit against Airbnb is not precluded by LOT membership because it is over tech that IBM claims it developed on its own and still owns the rights to. LOT would not protect Airbnb against such a lawsuit. This lawsuit suggests that LOT is changing patent litigation, but not necessarily curtailing it in the long term. IBM is asking for an injunction on Airbnb. In the post eBay v. Merc-Exchange world, this is a very, very tall order. For starters, there is no presumption of an injunction as the appropriate remedy to infringement. Public good is also considered. Stay tuned for more. Google and Amazon are giants in today's corporate culture. Much like Standard Oil of 1900, GM and IBM of 1970, and Walmart of 1995, Google(Alphabet technically) and Amazon stand alone when discussing companies today. Google's largest competitor on search is literally Amazon's search bar. Historically speaking, there's nothing unusual about a capitalist climate being dominated by two or three major players, even with fair competition. However, there is one thing that sets this BigTech era of Google and Amazon apart from past eras of corporate dominance: the asset they hold. Big Tech doesn't control land, natural resources, or even most machines(there are a few interesting exceptions, but only in niche areas). No, Big Tech controls intelligence. The biggest asset of Google and Amazon are the intelligence of their employees.
Naturally, they would love to increase their control of intelligence by making more of it through "Artificial Intelligence" and boy have they been doing so. They have a good strategy: find really smart people to develop AI, woo them with a really nice corporate culture, and have tough trade-secret rules to discourage their departure. This tactic would literally lead to Google and Amazon controlling the best AI and starting a singularity for ultimately decisive corporate domination if not for one thing, software patents on AI. The patent process is becoming increasingly beneficial to smaller companies that want to develop AI without seeing that AI swallowed whole by Google, Amazon, and Facebook. Google's DeepMind division is extremely secretive, but appears to have the technology to literally just comb and reverse engineer the Internet for novel software and algorithms developed by other companies and incorporate them into its own AI--only existing intellectual properties laws are stopping Google's AI divisions like DeepMind from doing this. This is why smaller AI companies are seeking patents on their technology, to protect their intelligence. There are two competing patent strategies for small AI startups though: one strategy is to patent rarely but powerfully with a couple of really broad patents, and the other strategy is to patent smaller innovations more aggressively. Consider this patent from DarwinAI: System and method for automatic building of learning machines using learning machines https://patents.google.com/patent/US20190138929A1 This is only one patent, but it is a huge patent that claims a lot. It is literally singularity-style tech that deals with artificial intelligence designing other forms of artificial intelligence. Whoever controls this kind of innovation will inevitably control the future as the inexorable advance of AI continues. DarwinAI seems to like the broader patent approach. I wonder if DarwinAI faced a restriction requirement from the USPTO when they initially filed for this patent. Kyndi is the opposite. Kyndi likes to patent frequently for it's innovations that allow for specific problems to be solved. Consider this patent, which will be discussed in more detail in another Quantum Computing entry: Quanton representation for emulating quantum-like computation on classical processors https://patents.google.com/patent/US20160328253A1 This is a very specific innovation that is basically dealing with ways to simulate quantum computing on convention computers. It's claim is specific. That's right, I wrote 'claim' not 'claims' as there is only one claim in the application. I think the difference in patent-strategy between Kyndi and DarwinAI reflects differing roles of AI. For Kyndi, AI is really just a tool for inventing other methods and machines, whereas for DarwinAI, AI is the end goal of its innovation. According to several sites on the Internet, on March 3, 1941, Forrest Mars, the candy tycoon, received a patent for a method of making what would later be called M&M's. Of the sites that source this claim such as the Wikipedia article on Forrest Mars, all appear to link to this MIT page:
https://lemelson.mit.edu/resources/forrest-e-mars-sr None of the sites provide the patent number. I've tried Espacenet and Google Patents and neither database shows Forrest Mars being awarded a US patent in March 1941. I can't use the USPTO database because the USPTO database only allows searches by patent number or family for patents prior to 1976. This is frustrating, and the moral of the story is: ALWAYS CITE THE PATENT NUMBER IF YOU'RE GOING TO REFERENCE SOMEONE RECEIVING A PATENT. There is a legal expression, "res ipsa loquitur", which means "the thing speaks for itself." I think that fits for the patent we're talking about here. In this patent, the inventors basically came up with a slightly improved and specialized version of Skype with a conventional chatbot, but one would think they invented some sort of revolutionary, cybernetic, paradigm-shifting, network of robotic doctors from reading the patent.
The following is a link to a genuine patent, which means an examiner actually approved it: https://patents.google.com/patent/US20160119582 Title: Neurosynaptic network connectivity and collaborative knowledge exchange with visual neural networking and packetized augmented cognition The first claim is about 800 words. It features two non-functional elements: "A network systems apparatus"and "a tele-visual imagery informatics management system". All the other elements of the claim are functional and the claim is loaded with mean-plus-function language. The functional language is where this media software patent just becomes meaningless. Consider one portion of the same claim, Claim 1: enabling at least one or more collaborants to concurrently view, communicate, collaborate, and consult among collaborants using at least one or more sources of streaming imagery data acquired and transmitted by tele-visual imagery informatics management system clini-docks, including live streaming imagery data, archived streaming imagery data, appended streaming imagery metadata, collaborant annotations, and archived collaborated imagery files during a synchronous or asynchronous collaboration session; This entire functional element is basically VOIP and an interface. There are so many of these functional elements that it is essentially impossible to infringe on this patent. Now, what about the specifications? Well, the spec contains language such as: "Medical dicom vismemes allow for networked cognitive enrichment through recursive annotation, tagging, encapsulation and saving, together with value chain knowledge exchange. Value chain knowledge exchange includes knowledge acquisition and mapping, creation and destruction, integration and sharing, replication and protection, as well as performance evaluation and innovation. The invention also provides for networked collaborative innovation and value chain knowledge exchange with multisensory cognitive communications and interactive data visualization. Multisensory cognitive communications includes multisensory [sight-sound-touch] digital data exchange with vision, audition and sensation [semiotics, semantics, somesthetics.]" This isn't gibberish, but it's pretty darn close. Basically, the patent drafter is saying that the software makes people smarter by using AI to exchange various forms of medical data during conversations. All the rest of this is basically just a bunch of big words thrown around, presumably to bother the examiner into granting the patent without much fuss. The invention is like an automated version of a sidekick that chimes in at just the right moment with theoretically useful information during a conversation, except this is for medical conversations through the Internet. Much of the patent just consists of lots of potential examples of simple and obvious features that are just thrown in there to overwhelm the examiner. Now, James Paul Smurro has filed another application heavily borrowing from the 2016 patent shown above: https://patents.google.com/patent/US20190355483 There is also this, very similar patent: https://patents.google.com/patent/US10332639B2/en?inventor=James+Paul+Smurro The current assignee of this patent is a company, Foresight Imaging, and it is a legitimate company in Chelmsford, which is not too far from UNH Law in Concord, NH. I'm guessing that the inventor, James Paul Smurro, filed this by himself. I will try to get in touch with Mr. Smurro and Foresight Imaging because I want to learn more about these patents. A key section of The Atomic Energy Act of 1954, 42 U.S.C. §2181(a), reads:
No patent shall hereafter be granted for any invention or discovery which is useful solely in the utilization of special nuclear material or atomic energy in an atomic weapon. Any patent granted for any such invention or discovery is revoked, and just compensation shall be made therefor. This is a relatively obscure aspect of patent law that essentially amends 35 U.S.C. §101. This leads to an interesting question: 'just what is an atomic weapon?'. The question is far tougher to answer than it may seem at first glance. Are atomic weapons any kind of device that is used induce nuclear explosions? If so, then tons of patents related to theoretical nuclear-fueled spacecraft are either unpatentable outright, or only have limited protection. In 1958, shortly after the passage of The Atomic Energy Act of 1954, the government, in conjunction with General Atomics, began researching the possibility of using nuclear weapons to propel spacecraft in "Project Orion". Although the Space Race, and the bureaucratic paradigm shift engendered by the emergence of NASA, quickly lead to scientists and engineers to abandon the project for more imminently feasible spacecraft, interest persists to this day. Carl Sagan, in his famous Cosmos program, suggested using nuclear weapons as propulsion(https://www.youtube.com/watch?v=JQnka2wNa_M). Carl Schroeder, in his Lockstep novels, describes a futuristic space-faring society where nuclear propulsion reaching 3% lightspeed is the only known means of long-distance space travel. So this field has been stimulating the imaginations of visionaries for a long time. Due to the lack of immediate benefit, the difficulties procuring the means to invent in this field, and the longtime role of patent pools in the field of aircraft and spacecraft that patenting would be commonplace for these ideas, but it isn't. Looking at the applications in the area may explain why. Consider US20060056570, "Fission fragment propulsion for space applications," and read his independent claim(claim 10): 10. A spacecraft propulsion engine that directly uses the kinetic energy of nuclear fission fragments to produce spacecraft thrust, comprising: a. a heat sink with one or more of its surfaces not located within a containment structure or any other form of outer shell. b. a heat exchanger for removal of nuclear fission waste heat. c. a sub critical-mass fission zone external to the spacecraft, not located within a containment structure or any other form of outer shell. He can't actually claim anything unique about the fuel here because of the Atomic Energy Act of 1954. His claims 1-9 were canceled, likely because of this law as well. Many of the novel aspects of such a craft would likely be identical to atomic weapons, just used in a different way. If you check the inventor, Donald G. Sutherland, you'll see that he applied for a related patent but abandoned it as well(US20070127617A1 "Nuclear fission fragment kinetic energy rocket engine"). Honestly, I think it is time to amend the Atomic Energy Act of 1954 for more clarification on the patentability of nuclear-fueled spacecraft. In this post, we continue our discussion of the rapidly emerging field of quantum computing patents. Consider US 10275721, which is a "Quantum computing machine learning module". Particularly, we discuss the significance of Zlatko Minev's groundbreaking research into quantum jump prediction and its implications for the enablement requirements under 35 USC 112(a) and the utility requirements under 35 USC 101. Quantum computing works by utilizing really weird properties of electrons that appear to defy conventional Physics. Specifically, conventional Quantum Physics has taught for decades that electrons seem to often be in multiple places at once and spin in particular directions for no predictable reason. Applying the principle of "quantum superposition", i.e., the notion that an electron may be in multiple states at the same time, one may store a lot more data than they can with only conventional digital logic gates.
Under this massive assumption of traditional Quantum Physics regarding superposition, an extremely powerful method of encryption has emerged. To play around with a free Quantum Computer, see D-Wave's website for a trial(https://www.dwavesys.com/) or see this great simulator(http://www.quantumplayground.net/#/home). From these trials or simulators, you can see that the encryption rests on the idea that one cannot predict an electron's state and transition from one state to another. The enablement requirements in U.S. Patent No. 10,275,721 clearly rest on this idea that quantum jumps are unpredictable because a PHOSITA would necessarily interpret the claims to mean that quantum superposition is the driving principle behind the module. Specifically, the drafters write "It is understood that quantum memories are devices that can store quantum data for a long time with high fidelity and efficiency, e.g., light-matter interfaces where light is used for transmission and matter for storing and preserving the quantum features of quantum data such as superposition or quantum coherence. " If Dr. Minev's experiment holds true, then this above statement is false. This essentially means that a PHOSITA would be unable to use quantum superposition to store the relevant machine learning data, as the methods described to do so are in fact false. In our next entry, coming soon, we will discuss this issue in greater detail and also discuss what quantum jump prediction would imply in light of the Alice decision. Pepsi, which owns Lays, has decided to sue four Indian farmers for infringing on US20050081269A1, which is a species of potato. (https://patents.google.com/patent/US20050081269A1/en). Recall the 2013 case, Bowman v. Monsanto for another recent example of a corporate giant going after a farmer for infringing on a patent. In case anyone is wondering, neither of these patents were plant patents--they were both utility patents, and you can always spot a plant patent by the letters "PP" in the name.
Legally, this suit was definitely probably inspired by Bowman v. Monsanto and SCOTUS' narrow view of the "patent exhaustion doctrine". However, there are enormous practical differences here. In the Bowman case, the farmer was successful and making a lot of money(about $75K per year) in the global scheme of things. In this case, it isn't clear just how successful the four farmers are. Statistically speaking, they likely earning about $4,000 per year(If you want to know how I estimate that, comment, but otherwise, see this https://www.hindustantimes.com/india-news/rs-6-000-is-6-of-a-small-farmer-s-annual-income-according-to-nsso-data/story-rddMw0hk6cSbxjo7E1GyKK.html). From Pepsi's perspective, the case could be about making sure the potato patent is being enforced to deter other infringers. If some kind of action is not brought against these farmers, other companies could infringe on a much larger scale. |
Archives
November 2021
Categories |