4 ms·

The broader issue is that the USPTO will let you patent anything. They have a long review process, but so many stupid things get approved that the agency is a

The broader issue is that the USPTO will let you patent anything. They have a long review process, but so many stupid things get approved that the agency is a sad joke. The latest example: https://www.eff.org/deeplinks/2019/07/someone-suing-companies-using-sms-messages-2019 https://www.eff.org/deeplinks/2019/07/someone-suing-companie...

ilaksh 7y ago
Nice one. How about this Google patent for "Generating output sequences from input sequences using neural networks" http://www.freepatentsonline.com/10402719.html http://www.freepatentsonline.com/10402719.html
- pnw_hazor 7y ago
  This is what is protected by this patent: "A method comprising: obtaining an input sequence having a first number of inputs arranged according to an input order; processing each input in the input sequence using an encoder recurrent neural network to generate a respective encoder hidden state for each input in the input sequence; and generating an output sequence having a second number of outputs arranged according to an output order, each output in the output sequence being selected from the inputs in the input sequence, comprising, for each position in the output order and beginning at an initial position in the output order: generating, using the encoder hidden states, an attention vector for the position in the output order; generating, using the attention vector, a softmax output for the position in the output order, wherein the softmax output scores each position in the input order; determining, using the softmax output, a pointer to a particular position in the input order; and selecting, as the output for the position in the output order, an input from the input sequence that is located at the particular position in the input order identified by the pointer."
DannyB2 7y ago
The patent examination process is not well understood by most people. Once a patent is received, the patent examiner carefully places the application into a room full of other patent applications. Then kittens are released into the room with PATENT GRANTED stamps affixed to their feet. The kittens are then returned to their holding area to await the next round of patent examination. The patent examiners collect the applications from the floor and look to see which patent applications were granted.
- gumby 7y ago
  The use of kittens was specifically specified by lobbyists as they knew that would make it impossible to change the procedure. Who’s going to vote against kittens? Squee!
- api 7y ago
  This is easily the best HN comment of 2019. It's also 100% true and completely accurate.
bcaa7f3a8bbc 7y ago
Why does the USPTO do this? Is it a technical problem or a political one? I remembered reading a story from the early industrial era, about an inventor who challenged another inventor's validity of patent and pressured the patent office to hold a demonstration for verification before believing the invention is true. Software patent is probably a culprit that makes some of the previously impossible patents possible by doing it on a computer. But I don't think it can explain everything. Why has the USPTO became something beyond stupid nowadays? I don't think the patent office verifies anything these days, they only log everything and left all the actual business to courts, I'm not sure it's the best system. Is it because the model of a centralized patent office simply doesn't scale well for the tremendous amounts of patent? Or because of lobbyists? Or both (e.g. patent office doesn't scale well and the lobbyists and patent lawyers exploit it to apply stupid patents)?
- dredmorbius 7y ago
  It's an economic one feeding a political one. Defining de novo property interests without physical bound inherently creates a vested rentier class interested in the persistence of those properties. New entrants to a field generally oppose patents. Incumbants generally support them. I've watched numerous firms (Oracle, Microsoft, Google) transition from one side of this dynamic to the other. It's a tragic inevitability, in the classical Greek sense. The (further) tragedy of the tech world is that it thinks this is a problem only with, take your pick, software or information-technology subjects. The problems with (and debates over) patents go far further back. Arguably, patents delayed the start of the Industrial Revolution by 25 years, as England awarded extensions to James Watt for his steam engine. Which may, of course, mean we merely postponed the inevitable global warming by a couple of decades. The histories of light bulbs, phonograph, telephone, radio, television, and more, are other arguments against patents. https://www.washingtonpost.com/news/wonk/wp/2012/10/10/the-case-against-patents/ https://www.washingtonpost.com/news/wonk/wp/2012/10/10/the-c...
  - rayiner 7y ago
    On the flip side, Watts developed his engine using financing from a coal industrialist, John Roebuck. Would he have secured that financing in the absence of a potential monopoly profit on his engine? The examples you mention illustrate the importance of patent rights to securing financing that bankrolls R&D. Bell had investors who pressed him to patent the telephone. Edison’s research was bankrolled by a series of contractual arrangements with companies like Western Union—where patent rights protected technology that he was demonstrating and trying to sell to Western Union. He also sold patent rights to Jay Gould. His Menlo Park R&D lab tested 3,000 lightbulb designs before filing the first patent. The same month as that patent filing, Edison founded what would become General Electric using financing from JP Morgan and others. Although you didn’t mention flight—the famous Wright-Curtis lawsuit actually involved two patent holders. Curtis was bankrolled by, among other people, Bell, and had its own aircraft patents. See here’s for more of the history: https://poseidon01.ssrn.com/delivery.php?ID=498002099093106085081003006071087070028083066008012016111002113081101122119072071096121054122119001002048075099076094031066002031087047003043075077098113104088083031081001022012064100021000077117072113101016066110082096090097084064113026086015018104070&EXT=pdf https://poseidon01.ssrn.com/delivery.php?ID=4980020990931060... Patents almost certainly delay the widespread adoption of a new technology. That’s the whole point. But the counter-factual scenario is not, as you suggest, what would have happened had the invention not been patented, but what would have happened if patents had not been available. Would the alternative means that exist to finance R&D (government spending, mostly) have produced the invention in the same time frame? Also, there is the problematic fact that monopolies are, for a time anyway, good for innovation. Someone else might have invented the telephone. But Bell’s monopoly ended up inventing the transistor, C, and UNIX. Someone else might have invented the copier. But Xerox’s monopoly ended up investing the GUI, email, and networking. GE, IBM, AT&T, Microsoft, Xerox, HP, Google-these are all companies that added way more to the state of the art during their quasi monopoly periods than at inception.
ozzmotik 7y ago
wow i violated this patent years ago when I made a microsocial network that used an email gateway to interact with users via sms. guess I should have patented it first ¯\_(ツ)_/¯

ilaksh 7y ago

Nice one. How about this Google patent for "Generating output sequences from input sequences using neural networks" http://www.freepatentsonline.com/10402719.html http://www.freepatentsonline.com/10402719.html

pnw_hazor 7y ago
This is what is protected by this patent: "A method comprising: obtaining an input sequence having a first number of inputs arranged according to an input order; processing each input in the input sequence using an encoder recurrent neural network to generate a respective encoder hidden state for each input in the input sequence; and generating an output sequence having a second number of outputs arranged according to an output order, each output in the output sequence being selected from the inputs in the input sequence, comprising, for each position in the output order and beginning at an initial position in the output order: generating, using the encoder hidden states, an attention vector for the position in the output order; generating, using the attention vector, a softmax output for the position in the output order, wherein the softmax output scores each position in the input order; determining, using the softmax output, a pointer to a particular position in the input order; and selecting, as the output for the position in the output order, an input from the input sequence that is located at the particular position in the input order identified by the pointer."

DannyB2 7y ago

The patent examination process is not well understood by most people. Once a patent is received, the patent examiner carefully places the application into a room full of other patent applications. Then kittens are released into the room with PATENT GRANTED stamps affixed to their feet. The kittens are then returned to their holding area to await the next round of patent examination. The patent examiners collect the applications from the floor and look to see which patent applications were granted.

gumby 7y ago
The use of kittens was specifically specified by lobbyists as they knew that would make it impossible to change the procedure. Who’s going to vote against kittens? Squee!
api 7y ago
This is easily the best HN comment of 2019. It's also 100% true and completely accurate.

bcaa7f3a8bbc 7y ago

Why does the USPTO do this? Is it a technical problem or a political one? I remembered reading a story from the early industrial era, about an inventor who challenged another inventor's validity of patent and pressured the patent office to hold a demonstration for verification before believing the invention is true. Software patent is probably a culprit that makes some of the previously impossible patents possible by doing it on a computer. But I don't think it can explain everything. Why has the USPTO became something beyond stupid nowadays? I don't think the patent office verifies anything these days, they only log everything and left all the actual business to courts, I'm not sure it's the best system. Is it because the model of a centralized patent office simply doesn't scale well for the tremendous amounts of patent? Or because of lobbyists? Or both (e.g. patent office doesn't scale well and the lobbyists and patent lawyers exploit it to apply stupid patents)?

dredmorbius 7y ago
It's an economic one feeding a political one. Defining de novo property interests without physical bound inherently creates a vested rentier class interested in the persistence of those properties. New entrants to a field generally oppose patents. Incumbants generally support them. I've watched numerous firms (Oracle, Microsoft, Google) transition from one side of this dynamic to the other. It's a tragic inevitability, in the classical Greek sense. The (further) tragedy of the tech world is that it thinks this is a problem only with, take your pick, software or information-technology subjects. The problems with (and debates over) patents go far further back. Arguably, patents delayed the start of the Industrial Revolution by 25 years, as England awarded extensions to James Watt for his steam engine. Which may, of course, mean we merely postponed the inevitable global warming by a couple of decades. The histories of light bulbs, phonograph, telephone, radio, television, and more, are other arguments against patents. https://www.washingtonpost.com/news/wonk/wp/2012/10/10/the-case-against-patents/ https://www.washingtonpost.com/news/wonk/wp/2012/10/10/the-c...
- rayiner 7y ago
  On the flip side, Watts developed his engine using financing from a coal industrialist, John Roebuck. Would he have secured that financing in the absence of a potential monopoly profit on his engine? The examples you mention illustrate the importance of patent rights to securing financing that bankrolls R&D. Bell had investors who pressed him to patent the telephone. Edison’s research was bankrolled by a series of contractual arrangements with companies like Western Union—where patent rights protected technology that he was demonstrating and trying to sell to Western Union. He also sold patent rights to Jay Gould. His Menlo Park R&D lab tested 3,000 lightbulb designs before filing the first patent. The same month as that patent filing, Edison founded what would become General Electric using financing from JP Morgan and others. Although you didn’t mention flight—the famous Wright-Curtis lawsuit actually involved two patent holders. Curtis was bankrolled by, among other people, Bell, and had its own aircraft patents. See here’s for more of the history: https://poseidon01.ssrn.com/delivery.php?ID=498002099093106085081003006071087070028083066008012016111002113081101122119072071096121054122119001002048075099076094031066002031087047003043075077098113104088083031081001022012064100021000077117072113101016066110082096090097084064113026086015018104070&EXT=pdf https://poseidon01.ssrn.com/delivery.php?ID=4980020990931060... Patents almost certainly delay the widespread adoption of a new technology. That’s the whole point. But the counter-factual scenario is not, as you suggest, what would have happened had the invention not been patented, but what would have happened if patents had not been available. Would the alternative means that exist to finance R&D (government spending, mostly) have produced the invention in the same time frame? Also, there is the problematic fact that monopolies are, for a time anyway, good for innovation. Someone else might have invented the telephone. But Bell’s monopoly ended up inventing the transistor, C, and UNIX. Someone else might have invented the copier. But Xerox’s monopoly ended up investing the GUI, email, and networking. GE, IBM, AT&T, Microsoft, Xerox, HP, Google-these are all companies that added way more to the state of the art during their quasi monopoly periods than at inception.

ozzmotik 7y ago

wow i violated this patent years ago when I made a microsocial network that used an email gateway to interact with users via sms. guess I should have patented it first ¯\_(ツ)_/¯