Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
AI Patents The Courts

Website Attempts To Generate Every Possible Patentable Invention (allpriorart.com) 148

An anonymous reader writes: All Prior Art is a project attempting to algorithmically create and publicly publish all possible new prior art, thereby making the published concepts not patent-able. The concept is to democratize ideas and to preempt patent trolls. The work is released on-line and in files of 10,000 ideas under a creative commons license. The system works by pulling text from the entire database of US issued and published (un-approved) patents and creating prior art from the patent language. While most inventions generated will be nonsensical, the cost to computationally create and publish millions of ideas is nearly zero -- which allows for a higher probability of possible valid prior art.
This discussion has been archived. No new comments can be posted.

Website Attempts To Generate Every Possible Patentable Invention

Comments Filter:
  • by slazzy ( 864185 ) on Sunday April 10, 2016 @10:09AM (#51878983) Homepage Journal
    I already did that years ago and patented the idea.
    • I visited the web site. Looked for a way to submit an idea. Zip, nothing; not even contact page. Maybe there's a patent on contacting folks?
      • While most inventions generated will be nonsensical

        There's the first real problem with this idea: prior art has to be findable before it can be submitted as prior art. As it stands, patent searching is manageable only because keyword searches, sorting algorithms, and classification systems shrink down the number of documents to something that actual humans can actually scan and pick over. Those methods would utterly fail in this case because by the time the database has enough records to stand a chance of containing relevant prior art for a claim the search

        • by WarJolt ( 990309 ) on Sunday April 10, 2016 @12:07PM (#51879453)

          The cost of generating them? Nearly $0.

          The cost of using machine driven patents as prior art to fight a patent troll? Millions.

          A patent troll doesn't need a bullet proof patent to make your life miserable. It can be easier and cheaper to just pay them.

          • by ShanghaiBill ( 739463 ) on Sunday April 10, 2016 @12:37PM (#51879577)

            A patent troll doesn't need a bullet proof patent to make your life miserable. It can be easier and cheaper to just pay them.

            It can be even easier and cheaper to just ignore them. My company has been threatened many times by patent trolls, including Acacia Research [wikipedia.org] and Intellectual Ventures [wikipedia.org]. In every case, we chose to just ignore them. Sometimes they sent a follow up letter. We ignored those too.

            It costs more to pursue a patent lawsuit that to defend against one. Their business model depends on a quick payout. So they shotgun out lots of letters, waiting for some intimidated fool to bite. If you respond, you are basically saying "Hey, look at me, I'm a target!" Unless they have actually filed papers with the court, you have no obligation to respond.

            Of course, if you talk to a lawyer, they will be horrified at the idea of ignoring a legal threat, and will instead recommend that you spend a lot of money on your lawyer. Here's another free lesson in life: Your lawyer does not represent your best interests.

            • by Anonymous Coward

              If they mention the patents in question and you ignore the letter, you get to pay for "willful" infringement instead of just infringement, which is much worse.

              The US has recently modified its patent laws to become a "first to file" instead of "first to invent", which means the idea of just publishing ideas out there as "prior art" would need to be tested again in court. The defense might have to prove that the published ideas were well recognized in the industry--a high standard to meet.

              • Re: already patented (Score:5, Informative)

                by gavron ( 1300111 ) on Sunday April 10, 2016 @01:19PM (#51879717)

                Anonymous Coward that should NEVER EVER pretend to practice law said:
                "If they mention the patents in question and you ignore the letter, you get to pay for "willful" infringement instead of just infringement, which is much worse."

                No.

                Using imprecise words and just spreading misinformation and fearmongering. Better you would have just kept quiet. Didn't even have the decency to sign your name.

                A "mention" in a "letter" is nothing.

                If they serve you legal notice as per Rule 4.1 (US Federal Rules of Civil Procedure) then you can respond as appropriate.
                If they send you a letter and "mention" things (whatever the hell that means) you can safely use it as kindling for your weekend campfire.

                There are a lot of lawyers. Some actors play them on TV and in the movies. If you want to be one, just go do the whole law-school,
                bar-exam thing.

                Don't make up shit on slashdot. It's something people like to do... but it's not good to spread FUD and misinformation.

                Ehud
                Tucson AZ US

              • Re: (Score:2, Insightful)

                by Anonymous Coward

                If they mention the patents in question and you ignore the letter, you get to pay for "willful" infringement instead of just infringement, which is much worse.

                Willful infringement requires that you actually saw their claim and recognized it as such. "Mistakenly" confuse it for a coupon for something you didn't want and trowing it in the bin isn't illegal or willful.
                Whenever confronted, pretend that is the first time you heard about them. Now they need to prove that they informed you of the "infringement".
                Their business model is based on it being more expensive to prove them wrong than to pay them.
                Make sure that they don't get a response without doing actual work.

              • by allo ( 1728082 )

                The question is, do you want to make money or do you want to prevent patents? The second one is easy with this machinery.

            • Here's another free lesson in life: Your lawyer does not represent your best interests.

              This is so true.

            • Very informative post.

              I'm on the other end of such letters, but am not a patent troll. Several people have copied a set of patents I invented and own. Recently, some other patents have appeared that copy mine without any improvement at all. Mine is a foundational one, meaning that no one can license and actually 'use' their patent without infringing on mine – improvement or not.

              Here is how far they will go:
              * Copy-past entire paragraphs from my Detailed Description, with a single ch

        • I suspect that this initiative will receive harsher scrutiny in terms of 'findable'; purely because of its overt bad attitude(from the perspective of people who treat patents as a serious and respectable business); but it is relatively hard to claim with a straight face that we currently have a situation where attempting to either discover prior art or existent patents you might need to avoid infringing on is adequately practical.

          Even major enterprises with substantial legal resources get torpedoed regul
          • by Anonymous Coward
            Forgot to mention: some fine fellow has already put together a PRNG-based Library of Babel [libraryofbabel.info] both aesthetically based on Borges' description and about as close to his description of its function as one can reasonably expect in a universe where a benevolent, and/or pitiless, and/or inscrutable god hasn't provided us with unlimited RAM.

            I don't expect the imminent end of copyright because of this; but if your creative work can be expressed in the symbol set used, it's already in the library. They have an ana
          • Attempting to construct a 'Library of Babel' and put a stake through either patent or copyright on that ground (while almost certainly technically feasible, PRNGs are handy like that) are more or less certain to be tossed out the moment they hit a court that doesn't appreciate some snide CS punk getting clever;

            I think if someone managed to create an actual useful prior art document by this method and it was presented to challenge a patent it could be successful - once. That's how long it would take for the federal government to ram through a change in patent law to disallow prior art that has served no purpose other than to serve as prior art - sort of like tax shelters that serve no true business purpose outside of avoiding tax are deemed abusive.

            They could just disallow combinatorically generated documents

    • by Anonymous Coward

      Simpsons did it.

    • by Dunbal ( 464142 ) *
      I believe I said that 15 years ago, those very words. My copyright attorney will be in touch.
    • by PolygamousRanchKid ( 1290638 ) on Sunday April 10, 2016 @11:26AM (#51879291)

      I believe my patent for "A Method and Process for Doing Stuff with Things" has just about everything covered.

      • Nonsense: My 'Apparatus and method for doing stuff to a plurality of things' is totally different, though your pitiful little 'patent' may actually be nothing more than an infringing special case of the matters covered by my patent. See you in court!
      • by NormalVisual ( 565491 ) on Sunday April 10, 2016 @05:03PM (#51880691)
        I believe my patent for "A Method and Process for Doing Stuff with Things" has just about everything covered.

        Unfortunately, I hold the patent for "A Method and Process for Doing Stuff With Things Using a Computer", so watch your step!
    • Charles Stross wrote about this very idea in the science fiction novel Accelerando, pubished in 2005. So unless you beat him, he gets the credit.

      He may have borrowed the concept in turn, I don't know.
  • it is an invention

    • by alexhs ( 877055 )

      I hope they patented this, it is an invention

      But there's prior art [wikipedia.org].
      Of course, this time, it is "on a computer", which as we all know means that it is completely novel.

  • by Layzej ( 1976930 ) on Sunday April 10, 2016 @10:11AM (#51878993)
    It was the best of times it was the BLURST of times? You stupid Monkey! - https://www.youtube.com/watch?... [youtube.com]
  • by Anonymous Coward

    Having looked at the site, I'm not too impressed. All they're doing is generating abstracts, not specific embodiments.

    It's normal for any patent to come up against prior art (there's about 100 million patents on Espacenet, similar in the US archives). To get a patent granted you have to pick those apart and figure out what makes your invention novel. It can often be a minor difference that allows you to get a valid patent.

    Since the abstracts generated by this site are not very specific, it'd be a trivial ex

  • Years ago I registered the site publicpatent.org. I had planned for people to just submitted random stuff (i.e. idea A + idea B), so that stupid, obvious stuff would be in the public domain and no one could try to patent it. Using NLP to automate the stupid idea generation is great!
  • Library of Babel (Score:5, Insightful)

    by sbaker ( 47485 ) on Sunday April 10, 2016 @10:52AM (#51879137) Homepage

    This is essentially no more than the concept of a library containing books with every possible letter combination:

                    https://en.wikipedia.org/wiki/... [wikipedia.org]

    Such a library necessarily contains every work that has ever been (or will ever be) written.

    The problem with such a library (and the problem with All Prior Art) is that of search. Finding prior art that disrupts a patent that you need to make "go away") is just as impossible as finding the cure for cancer in the Library of Babel,

    So the important question is whether you can go to a court of law and say "My opponent's patent is provably invalidated because it's already explained in the Library of Babel"? If that's a valid legal argument - then perhaps this is of use. But I strongly suspect it's a complete waste of time.

    Of course one might argue that a physical embodiment of the Library of Babel (or *ALL* prior art) is impossible - but I might also argue that I've merely done text compression by writing;

            while ( 1 )
                    for ( int i = 0 ; i MAX_PATENT_LENGTH ; i++ )
                            putchar ( "abcdefghijklmnopqrstuvwxyz0123456789., " [ rand() % 39 ] ) ;

    • Also, your post, and the entire /. database are contained in pi and e.

    • Such a library necessarily contains every work that has ever been (or will ever be) written.

      With that set of characters. But what if there are ideas which can't be expressed with that set of characters?

    • Your argument is fundamentally flawed because the Library of Babel is infinite while AllPriorArt is a finite subset of the Library of Babe that is domain specific and encodes semantic patterns in the form of valid n-gram chains. There is a high probability that AllPriorArt could produce the exact wording for part of a yet to be approved patent and that part will describe a key concept in that patent thereby invalidating it.
    • Re:Library of Babel (Score:4, Informative)

      by Theaetetus ( 590071 ) <theaetetus@slashdot.gmail@com> on Sunday April 10, 2016 @05:51PM (#51880951) Homepage Journal

      This is essentially no more than the concept of a library containing books with every possible letter combination:

      https://en.wikipedia.org/wiki/... [wikipedia.org]

      Such a library necessarily contains every work that has ever been (or will ever be) written.

      The problem with such a library (and the problem with All Prior Art) is that of search. Finding prior art that disrupts a patent that you need to make "go away") is just as impossible as finding the cure for cancer in the Library of Babel,

      So the important question is whether you can go to a court of law and say "My opponent's patent is provably invalidated because it's already explained in the Library of Babel"? If that's a valid legal argument - then perhaps this is of use. But I strongly suspect it's a complete waste of time.

      I believe you're right. In patent law, there's a doctrine that says that prior art is only useable for what it teaches. For example, H.G. Wells' "Time Machine" is valid prior art for the concept of time machines, such that you couldn't get a patent that claimed "1. A system comprising: a time machine," but if you went one step farther and included a claim directed to your temporal disruption modulator, Wells' story wouldn't be applicable against that.

      Here's an example of the output of this database:

      A faucet includes a spout, a handle, and a capacitive sensor operably coupled to at least one of the spout and the handle. The systems, methods, and programs may extract a language pattern from the stored text sample and may create an authorized profile based on the language pattern. The second part is put on standby and storing parameters of the action are stored in the chip card.

      That's a nice string of phrases, but they teach nothing. You couldn't even use that to invalidate a patent on a faucet that extracts language patterns from a stored text sample and stores it in a chip card, because it has no teaching of how to build such a thing. Or even what that really means.

      At best, you could use those sentences individually to invalidate a first patent that claimed "a faucet including a spout, handle, and capacitive sensor operably coupled to at least at least one of the spout and the handle" and a second patent that claimed "extract[ing] a language pattern from the stored text sample and may create an authorized profile based on the language pattern and a third that claimed "putting [something] on standby and storing parameters of the action in the chip card."

      But there's a problem with that, too. From the database's description:

      The system works by pulling text from the entire database of US issued and published (un-approved) patents and creating prior art from the patent language.

      In other words, each of those sentences already appear in issued or published patent applications, and are already prior art. In fact, if you wanted to really invalidate any of those, you'd pull up the corresponding patent they used to create the prior art database, and it would include that sentence plus a whole bunch of additional details.

      So, not only does this not actually create any new prior art, it's less useful than the original data they're pulling from.

      Disclaimer: I am a patent attorney. Feel free to accuse me of bias and say that my background invalidates my opinion, but that doesn't mean it's legally or logically incorrect or that federal circuit judges - many of whom are patent attorneys - will not come to the same conclusion.

    • How many unique English tweets are possible? How long would it take for the population of the world to read them all out loud? [xkcd.com]

      Even with only 140 characters, there are *a lot* of possible English phrases.

      And what if you tried to store them all for future reference? You'd end up boiling the world's oceans [oracle.com].

      • by KGIII ( 973947 )

        I'd never read that before. It was greatly amusing. The thing is... I'd kind of want to turn it on. We'll have to find another way to cool it. If we ran it on a planet that was insanely cool AND were able to capture the energy from the two being in proximity then we might be on to something. I kid when I say it but it'd be kind of awesome if turning it on caused the nearby stars to dim for a brief second. Yes, yes the picture in my head is most awesome. Thanks for the link. (Can you imagine the sound of it

    • Such a library necessarily contains every work that has ever been (or will ever be) written.

      The problem with such a library (and the problem with All Prior Art) is that of search. Finding prior art that disrupts a patent that you need to make "go away") is just as impossible as finding the cure for cancer in the Library of Babel,

      So the important question is whether you can go to a court of law and say "My opponent's patent is provably invalidated because it's already explained in the Library of Babel"? If that's a valid legal argument - then perhaps this is of use. But I strongly suspect it's a complete waste of time.

      Even assuming they could generate stuff that was recognizable as prior art (which I doubt) I'm dubious a court would treat it as such.

      An invention is an idea, the patent is part of your reward for sharing the idea with the world. But if the invention was generated by a machine and never read by a human being then the idea doesn't exist.

      Or to put it another way, you can't have discovered something if you don't even know it exists.

  • by Anonymous Coward

    what a lot of rubbish: patent lawyers will argue:
    prior art has to be public and it must be possible for the skilled person to find such prior art. Therefore, prior art needs to be embedded in a context where it could or would be found (text books, patent documents, publications on web whose intent is to publish non-jokes). That's not the case here: a database with largely shitty or random content would not be searched by the skilled person, and/or is not 'made public' in the sense of the relevant patent law

  • This is a sad commentary on how abuse of US patent and copyright law has flourished.
  • Won't help (Score:4, Interesting)

    by GameboyRMH ( 1153867 ) <`gameboyrmh' `at' `gmail.com'> on Sunday April 10, 2016 @11:14AM (#51879221) Journal

    Prior art has never been a hindrance before...

  • Programs are not creative. They do what they're told to do. I'd like to see it develop a warp drive engine that works.
    • by Anonymous Coward

      You don't need something that:
      1) works
      2) even exists
      3) is even logical
      to get a patent on it in the US.

      A large number of patents are so verbose and basic that they probably all violate each other given a good lawyer and enough bad blood between 2 companies.
      This is the reason many companies cross-licence their stuff between each other rather than suing the hell out of each other because it would be a MAD situation.

      • The problem with this approach to patents is that when I create a patent that covers similar technology as the "non-working" version, but actually works properly, I still get my patent granted.

        You *can* patent nonsense, but that doesn't prevent actual patents of existing inventions from parallel (non-infringing by addition of crucial invention) patents.

  • They have re-invented monkey by the typewriter, and they are using computer power to stitch not separate letters, but words and phrases.

    They have, however forgot several things.

    It is the math. Some numbers, representing a possible number of combination of letters and words and ideas, are so high and so large, that there is not enough quarks in the universe to represent the number. If you have an infinitely large number and you use automation and software to reduce, you will still have a very large still unfathomable number of possible combination of ideas left.

    Or put it the other way, you can employ not one but a trillion billion of monkeys and give each a super fast computer and the outcome will be exactly the same as having one monkey with the typewriter.

    • by PPH ( 736903 )

      They have re-invented monkey by the typewriter, and they are using computer power

      So, completely in keeping with the existing patent process then.

      Submit 'Something, something, blah blah blah' (all prior art) and append using a computer or using the Internet and you have a brand new patent.

  • by Anonymous Coward

    They changed it from first to invent to first to file if people remember in an effort to save time and money while speeding up the patent process. I also doubt it would help fight against the obvious clause as the person who does the patenting could say that a person didn't think of it and it was randomly generated and buried so the idea was in fact not obvious since a human did not think of it until now.

    • by Anonymous Coward

      The US switched from First-to-invent to First-to-File starting March 16, 2013.

      This website is about 5-10 years too late.

      • The US switched from First-to-invent to First-to-File starting March 16, 2013.

        This website is about 5-10 years too late.

        Just replying to try to mod this up a notch (my postings seem to come in at 2, and I don't have moderator points today).

        This needs to be emphasized: Prior art doesn't mean what it used to mean as of 2013 as said above. This further tilts in the direction of large companies who can better afford to "carpet bomb" the patent office with filings.

        Anyone who thinks the patent system has any resemblence to "fair" should try filing just one patent on their own, without legal representation. If you have a career

      • That does not stop a potential copyright claim against a patent, if your patent just happens to have a block of text that is exactly the same as an older one published in an AllPriorArt like system, but one that asserts it's copyright over the patterns of generated text. It is the fact that machine intelligence can do this at all that breaks the system, imagine what IBM could do if they had Watson generate patents and use it's legal skills to file them as well, then it did this at a strategic level to fence
  • the patent office isn't going to check this database and after a patent it approved, it (sadly) requires more than just prior art to have the patent invalidated. the patent system is borked but this isn't the solution.

  • by Anonymous Coward

    There is prior art for prior art generators:
    http://thesurrealist.co.uk/priorart.cgi

    The older one is certainly more amusing.

  • April 1st was over a week ago, but nice try.

  • This is like pointing to the sci-fi of E.E. "Doc" Smith and claiming you have prior art for a faster-than-light drive.

    Prior art has to be anchored in the real world of invention and application or it means nothing.

  • A computer-generated e-book with every possible melody using up to 10 notes of a chromatic scale (12 different frequencies in an octave, as shifting up or down an octave doesn't change the melody musically) would have a little over 10 billion entries. The first note can always be C because transposing a melody to a different key doesn't change the melody (that's the whole point of a chromatic scale). And rests can be covered by a null note (raising the total possible notes to 13). 13^9 = 10.6 billion.
  • I tried feeding it the opening line of one of my patents [google.com]. Sure enough, it found it verbatim [allpriorart.com], along with random lines from other patents. Maybe if they threw some "deep learning" at it, they'd get somewhere.
  • The algorithmically generated prior art is probably not prior art at all. To be prior art, the description has to be published at a specific date. I don't see any dates on this prior art (here for example http://allpriorart.com/1459996... [allpriorart.com]). Further, even if it had a date, it's unclear if this was ever "published". A single URL sitting on the internet is probably insufficient to meet the publication requirement if the link isn't accessible from some indexing or search system. Even the 10,000 prior art invent
  • This is so much more efficient than training 10,000 monkeys as patent attorneys and letting them loose to patent ideas.

  • Surely prior art has to be an actual, physical thingamagig,
    not just an 'idea' scribbled on the back of a napkin.

    • If you've seen what qualifies as actual patents these days, then you'd know that most of the time they haven't even bothered with scribbling on a napkin before it gets shoveled through and approved.

  • Peanut-butter powered horse launcher? Check!

  • It's only art if someone produces it. Unless you can show there's some sort of AI or other intelligence, I would think they could argue it's not art. It's machine output.

    There's also a heck of a lot of work to do for patents. You have to have a notebook, that has numbered pages. You need to show how you came to this invention. You also have to mention prior art. And so on.

    Being a patent troll isn't easy work. A whole lot easier than coming up with a real invention, however. I'd argue that if a company is

  • ... how most Internet of Things devices were invented.

    > A device for interconnecting the buttering of tcp-aligned toast for a network-aware breakfirst.

    > Wearable bluetooth-capable scarf designed to fulfill the modern needs of internet connected fitness tracking on the go

The Tao is like a glob pattern: used but never used up. It is like the extern void: filled with infinite possibilities.

Working...