Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Google Technology

This Tiny Website Is Google's First Line of Defense in the Patent Wars (wired.com) 45

A trio of Google engineers recently came up with a futuristic way to help anyone who stumbles through presentations on video calls. They propose that when algorithms detect a speaker's pulse racing or "umms" lengthening, a generative AI bot that mimics their voice could simply take over. That cutting-edge idea wasn't revealed at a big company event or in an academic journal. Instead, it appeared in a 1,500-word post on a little-known, free website called TDCommons.org that Google has quietly owned and funded for nine years. WIRED: Until WIRED received a link to an idea on TDCommons last year and got curious, Google had never spoken with the media about its website. Scrolling through TDCommons, you can read Google's latest ideas for coordinating smart home gadgets for better sleep, preserving privacy in mobile search results, and using AI to summarize a person's activities from their photo archives. And the submissions aren't exclusive to Google; about 150 organizations, including HP, Cisco, and Visa, also have posted inventions to the website.

The website is a home for ideas that seem potentially valuable but not worth spending tens of thousands of dollars seeking a patent for. By publishing the technical details and establishing "prior art," Google and other companies can head off future disputes by blocking others from filing patents for similar concepts. Google gives employees a $1,000 bonus for each invention they post to TDCommons -- a tenth of what it awards its patent seekers -- but they also get an immediately shareable link to gloat about otherwise secretive work.

This discussion has been archived. No new comments can be posted.

This Tiny Website Is Google's First Line of Defense in the Patent Wars

Comments Filter:
  • They propose that when algorithms detect a speaker's pulse racing or "umms" lengthening, a generative AI bot that mimics their voice could simply take over.

    It's not April 1 yet, too early to post R&D spoofs.

    • Re:Ummm, no! (Score:4, Interesting)

      by HiThere ( 15173 ) <[ten.knilhtrae] [ta] [nsxihselrahc]> on Tuesday February 20, 2024 @01:50PM (#64255020)

      IIUC, the site isn't for good ideas, it's for ideas that aren't wroth developing, but which you don't someone else to ever be able to use against you.

      I agree the idea you quote is somewhere between stupid and dangerous, but that wouldn't keep someone else from patenting it and then arguing that something you do is derived from it.

      FWIW, I consider this particular idea to be "anti-evil". It's anti-evil in a way related to the BSD and GPL licenses. It makes ideas "open source", i.e. not patentable.

      • "it's for ideas that aren't wroth developing, but which you don't someone else to ever be able to use against you"

        It's for ideas that would be hard to prove infringement yet not so hard that a case could not be opened, thus the expected value would be negative from both a plaintiff's and a defendant's perspective.

  • by PPH ( 736903 ) on Tuesday February 20, 2024 @01:08PM (#64254898)

    ... to add to their website.

    The website is a home for ideas that seem potentially valuable but not worth spending tens of thousands of dollars seeking a patent for.

    How about we clean up the patent system so it doesn't cost "tens of thousands of dollars" to apply for one. And as a result, pretty much force individual (non wealthy) inventors to have to assign their inventions over to a corporation.

    • by suutar ( 1860506 ) on Tuesday February 20, 2024 @01:26PM (#64254942)

      I'd rather clean up the patent system so it actually checks for novelty, triviality, and instantiability, but that would make it more expensive, not less.

      • Why would an AI-checker make it more expensive?

        An AI could flag patents that were obvious, prior art, or poorly worded, saving time for the human examiners.

        That would make the patent system less expensive.

        • by suutar ( 1860506 )

          Sorry, I didn't see any mention of an AI checker previously. If it gets to the point where an AI checker can reliably give correct answers, that'd be great. I don't think we're close to that yet.

    • by kurkosdr ( 2378710 ) on Tuesday February 20, 2024 @01:49PM (#64255014)

      How about we clean up the patent system so it doesn't cost "tens of thousands of dollars" to apply for one. And as a result, pretty much force individual (non wealthy) inventors to have to assign their inventions over to a corporation.

      Applying for a patent will always cost a substantial amount of money because the people who determine the validity of the patent filing won't work for free, and the operational cost of running the whole bureaucracy related to patents isn't trivial either, and I don't want my tax money to be used to fund an office whose sole purpose is to give someone else a limited-time government-granted monopoly. Let them pay for it (if they want their limited-time government-granted monopoly so much) when they apply for a patent.

      Also, I am aware that the USPTO isn't known for doing a good job at determining the validity of patents considering the kind of junk that slips through (see Unified Patent's blog if you want examples), but my reasoning stands: The people who should determine the validity of the patent filings aren't going to work for free. If anything, it should cost more to file for a patent so the USPTO has more people to examine the patent filings more thoroughly instead of sacrificing scrutiny for speed.

      Also, if you have spent the R&D to make any modern invention worthy of a patent, your filing costs are trivial in comparison.

      • by PPH ( 736903 )

        because the people who determine the validity of the patent filing won't work for free

        AI will.

        • Considering the energy requirements of AI, I am doubtful if it will be even near-free, and by the time we reach a point where AI can determine the validity of a patent, we'd have probably reached post-labor utopia where everything is done by AI anyway. BTW this is a new spin on the talk about "singularity", in the sense that the initial hypothesis does all the heavy lifting but there is no evidence it will ever become true.
          • by the time we reach a point where AI can determine the validity of a patent, we'd have probably reached post-labor utopia where everything is done by AI anyway.

            A world where everything is done by AI and those that control them fight GovAI to ensure that others cannot make something without their approval. Sorry, that's dystopia.

            There's no morale or ethical reason to grant a government backed monopoly if your society is post-labor. There's no one to purchase it, as labor is worthless and done solely by machines. The best argument you can make for the practice is "keeping it from my enemies" but even then, it's made by an AI. AIs can be hacked just like any other

            • And you don't understand patents don't reward labor, they reward inventions. It's entirely possible the "singularity" future that people here are dreaming never pans out and AI does all the menial labor (including producing generic art as it does now) while being unable to invent things (or do creative work in general). In that future, humans will be left doing creative work and the reward for that work will be a limited-time government-granted monopoly in the form of copyrights and patents as it is now. Ke
              • BTW if we ever singularity (doubtful), patents would simply cease to have meaning, since machines would invent everything in seconds and then... stop. At this point, humanity can choose to dispense with the whole concept of patents, since there would be nothing to file for and no reason to reward inventions anymore. But again, singularity is the kind of thought experiment where the hypothesis does all the heavy lifting but is highly doubtful it will ever become a thing.
      • Re: (Score:2, Informative)

        by Anonymous Coward
        Applying for a patent will always cost a substantial amount of money because the people who determine the validity of the patent filing won't work for free

        Applying for a patent does not now nor ever has cost a substantial amount of money. You can, at this very moment, get on file with the USPTO for as little as $364, of which a mere $300 pays the Examiner for everything that person does. Who, by the way, doesn't "determine (its) validity" but rather presents a prima facie case (if any) as to why the app
      • the USPTO no longer checks for validity of patents. There are far too many patent submissions for them to handle them all directly. Congress, in its anti-tax wisdom, refuses to fund more of them or allow them to be paid through income from the patent filings. The intent of congress is for all executive functions to eventually fail, thus devolving back to individual states being supreme authorities.

        • As I've said above, I shouldn't have to pay for a governmental office whose sole purpose of existence is to give someone else a limited-time government-granted monopoly. Let the beneficiaries pay. Limited-time government-granted monopolies are a privilege, not a right. If thoroughly examining the average patent filing takes such and such amount of time and the people who examine them work for such and such rates, then that's how much filing a patent should cost (plus the bureaucracy costs per patent). Then
    • Hey I just got another patent, and enough of a pittance from the company that after taxes I can probably buy three pizzas this time!

  • This is just another effort to make everything we see or hear fake, generated by some algorithm whose only intent is to fool. We have politicians for that.

  • Old practice (Score:5, Informative)

    by dsgrntlxmply ( 610492 ) on Tuesday February 20, 2024 @01:22PM (#64254932)
    IBM long ago published the Technical Disclosure Bulletin to disclose inventions where patents were not pursued. Other companies had similar. These provided a prior art reference that could be cited.
    • by davidwr ( 791652 ) on Tuesday February 20, 2024 @01:33PM (#64254960) Homepage Journal

      They put them behind a paywall a decade or two ago. I haven't checked lately, maybe they are free again.

      From what I've been told, back in the old, old days (pre-1980s?), IBM "published" their technical support bulletins "in name only" - they were printed and stored in a small number of publicly-accessible locations, but unless you lived nearby you had to travel or get someone local to make copies for you. This "publishing in name only" was enough to fight off patents, while being obscure enough to make it hard for most small- and medium-sized companies to access them.

      • True, I read at one of the UCLA libraries early 80s. Recall that there were individuals and organizations who would photocopy articles from large libraries, and mail or fax these as a paid service.
    • IBM long ago published the Technical Disclosure Bulletin to disclose inventions where patents were not pursued. Other companies had similar. These provided a prior art reference that could be cited.

      Absolutely, this idea isn't at all new. Defensive disclosure has long been used to ensure that someone can't come back later and claim to have invented an idea you used. The only novelty here is the idea of doing defensive disclosure Wikipedia-style; a resource open to anyone who wants to submit disclosures, on the web where it's cheap to operate and universally accessible.

      • by ip_vjl ( 410654 )

        ip.com has been doing defensive publishing for over 20 years in their "Prior Art Database" product. (Which includes old IBM TDB docs).

        It's not free, but part of the cost is that they produce printed materials with hashes of submitted documents whose time of origin can be legally defended because you have the bibliographic information and the document hash in a printed volume at a particular date.

        https://kb.ip.com/pad/knowledg... [ip.com]

        • Interesting. I guess Google's attorneys decided the printed materials weren't necessary. TDCommons datestamps all submissions electronically.
    • Yes, absolutely. For anyone who has worked for a big tech company with Invention Disclosure procedures, you may know that the process is:

      1. File an invention disclosure
      2. The Patent Review (or IP review) committee will review it and see if it is worth processing further - at this point, you usually receive your disclosure bonus of $100 or whatever.
      3. If it is worth proceeding with, then you start the process of filing a patent application and work with the attorney on crafting it, review prior art, etc., et

  • This doesn't really help anyone get over their fear of public speaking. And, it doesn't help anyone overcome their shyness with the other sex, either. No wonder they didn't patent it.
  • Posting, "My invention is having an electrode in someone's head they can control robots with" is useless.

    Posting a technical description someone could use as a road map to create my mind controlled evil robot army would be useful.

    Until shortly after I build it and take over the world.

    • Posting, "My invention is having an electrode in someone's head they can control robots with" is useless.

      But if you are working on actual engineering of how to control robots with a brain implant, if some amateur patents "having an electrode in someone's head they can control robots with", you would very much like to have that publication to point to, useless or not, so you can say "that's not patentable because it was already published."

      Because it's much easier to invalidate a patent with "that idea was alreay published" than it is to invalidate one with "the patent shouldn't have been granted because it wasn

      • by iAmWaySmarterThanYou ( 10095012 ) on Tuesday February 20, 2024 @02:38PM (#64255186)

        I've worked as a legal assistant/researcher for a patent law firm. The one liner would never get a patent. The USPTO overall does a pretty good job of filtering out trash. The problem is there's a *huge* number of patents coming in every year and sometimes properly written patent applications are granted a patent but shouldn't have been for any number of reasons. The patent examiners are just people doing the best they can, sometimes they fuck up, and those outlying cases are the only ones we hear about.

        "having an electrode in someone's head they can control robots with" nor anything similar will even get examined as it does not conform to proper patent application format. Whoever opens that envelope will immediately bounce it. But a properly written patent for a brain connected mind controlled robot system that was cartoonish bullshit might get through, yes, unfortunately.

        However, when a patent is compared to earlier works of art the thing they look at is the claims. The claims are a series of statements which define the patent in technical terms that someone else familiar with the state of the art could use to duplicate your invention. If your robot control patent says, "use the bio-mech transthombulator to establish link to the robot's positronic pathways", that's actually ok because no one else's work will ever duplicate those silly claims so if anyone ever does come up with a real way to control an evil robot army from a brain tap, the silly cartoon patent will not be prior art blocking the real invention.

        The opposite is also true. A real patent will not block the cartoon patent if filed later as the claims will be different.

        • They may do a good job of filtering out trash, but sometimes trash does get through. And some of it really does have no usable details. Probably not quite as bad as the joke example you gave, but pretty bad.

          ...The patent examiners are just people doing the best they can, sometimes they fuck up, and those outlying cases are the only ones we hear about.

          Unfortunately, it doesn't help that much when you have to litigate a patent troll to say "oh, that was just an outlier case."

          • I agree. The results of those errors are very bad, but as a percentage they're doing better than say .. Boeing. :-)

            • I agree. The results of those errors are very bad, but as a percentage they're doing better than say .. Boeing. :-)

              Really? There were 2.3 million Boeing flights in the US last year. One had a problem.

  • What a splendid idea. Everybody will think you are the most eloquent and elegant idiot! Something to aspire to!

    (Yes, I know this is a side-aspect of the story.)

  • Of which half of them are still valid, we have enough patents for corporations to make money off, let's do something more productive with our intelligence instead.

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Working...