The Case Against Patents (2013) • Hacker News Board

Re: The Case Against Patents (2013)

#11

Post by andrewla » Tue, Jul 10, 2018, 1:42 PM UTC

The authors summarize as:

> ... there is no empirical evidence that [patents] serve to increase innovation and productivity ...

This seems to be a fundamental misunderstanding as to what purpose patents serve. The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them. The cost to the public of patents are that such use must be deferred out of respect for the temporary monopoly granted to the patent holder.

That is, patents don't increase innovation or productivity -- that is by design. They are just a means of harnessing the innovation and productivity of individuals and encouraging them to make their work public instead of keeping it a secret. So it seems to me that the paper is arguing a strawman rather than making a useful point.

Re: The Case Against Patents (2013)

#12

Post by snarf21 » Tue, Jul 10, 2018, 1:43 PM UTC

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

I agree. I think that all patents aren't created equal. We'd have less issues if software patents were only valid for like 5 years. You still have protection but you get rid of a lot of the trolls and the IBMs of the world having 10K patents. For more physical things, the protection can be longer since it takes longer to recoup your investment.

Re: The Case Against Patents (2013)

#13

Post by rayiner » Tue, Jul 10, 2018, 1:46 PM UTC

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

Suggesting that they’re “impossible to execute” correctly is a bit odd, as all US technology arose in the context of the patent system (the first Patent Act predates the Constitution).

Re: The Case Against Patents (2013)

#14

Post by wolfgke » Tue, Jul 10, 2018, 1:47 PM UTC

snarf21 wrote: post #12

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

I agree. I think that all patents aren't created equal. We'd have less issues if software patents were only valid for like 5 years. You still have protection but you get rid of a lot of the trolls and the IBMs of the world having 10K patents. For more physical things, the protection can be longer since it takes longer to recoup your investment.

> I think that all patents aren't created equal. We'd have less issues if software patents were only valid for like 5 years.

And we'd have even less issues if software patents were only valid for 0 years (i.e. not exist).

Re: The Case Against Patents (2013)

#15

Post by beefield » Tue, Jul 10, 2018, 1:52 PM UTC

andrewla wrote: post #11

The authors summarize as: > ... there is no empirical evidence that [patents] serve to increase innovation and productivity ... This seems to be a fundamental misunderstanding as to what purpose patents serve. The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them. The cost to the public of pa…

> The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them

Can you come up with siginificant patented innovations that are not easy to reverse-engineer? I have thought that the raison d'etre of patents is to foster innovation.

Re: The Case Against Patents (2013)

#16

Post by api » Tue, Jul 10, 2018, 1:53 PM UTC

enriquto wrote: post #10

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

> I think the general idea of patents is good Also, there may be a fundamental difference between physical inventions, where patents seem to make sense, and mathematical theorems (and computer algorithms), where patents do not make sense at all.

The difference is between a specific invention vs. a law of mathematics or nature. The latter absolutely should not be patentable.

The negative opinion of patents around here comes from the fact that most people here are programmers and the vast majority of software patents are trash. I've seen very few patents on software that are not "obvious to anyone skilled in the art," patents on abstract mathematical concepts, or patents for which there is ample prior art in the literature or open source code repositories.

Telling the difference requires a great deal of conceptual thinking. Patent examining is "strong AI hard." I think it is possible to expand on these ideas to some useful extent. Obviousness for example could be explained as first-order connectedness is the evolutionary/learning state-space graph. Something is obvious if getting from current art to the new idea requires only one "hop" through state space or traversal of a densely connected region of state space. Non-obvious means some kind of great leap across state space or long journey through many indirections was required. It's basically a measure of creative labor.

The RSA patent is often cited as an example of a decent software patent: original, a specific well-bounded invention for a specific purpose, and non-obvious in the extreme. Another example would be if whomever Satoshi is had patented proof of work block chains since that would also meet those criteria. Of course patents on cryptographic systems tend to result in them not being used, but that's a separate issue.

Re: The Case Against Patents (2013)

#17

Post by jordigh » Tue, Jul 10, 2018, 1:59 PM UTC

snarf21 wrote: post #12

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

I agree. I think that all patents aren't created equal. We'd have less issues if software patents were only valid for like 5 years. You still have protection but you get rid of a lot of the trolls and the IBMs of the world having 10K patents. For more physical things, the protection can be longer since it takes longer to recoup your investment.

There often isn't even any "investment" in patents. Richard Feynmann would describe how his superiors would just walk around the labs and say, so! who has some ideas we could patent? And Feynmann's fellow scientists would say, well, you could grab this nuclear reactor and stick it into a submarine to power it, but that's pretty dumb and obvious. "Nonsense!" replied the superiors. "Patent that right away!"

I don't know how true it is that a lot of patents come just from employees sitting around navel-gazing in their free time and aggressive management telling them to write down those ideas so they can file more patents, but it seems believable that it would happen at least sometimes.

Re: The Case Against Patents (2013)

#18

Post by rbultje » Tue, Jul 10, 2018, 1:59 PM UTC

andrewla wrote: post #11

The authors summarize as: > ... there is no empirical evidence that [patents] serve to increase innovation and productivity ... This seems to be a fundamental misunderstanding as to what purpose patents serve. The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them. The cost to the public of pa…

> The public benefit of a patent is that the information is public [..] > The cost to the public of patents are that such use must be deferred

But who does this benefit? The author points out that it benefits big/old ("stagnant") corporations rather than innovators. And that still holds, no?

Here's a thought experiment: if you're a small innovator and you invent something new (let's be evil: a software algorithm), what benefit does disclosure bring you? It's better to keep it secret, because big/old ("stagnant") companies will find legal loopholes in your patent application to work around it. Or, they will run it in the cloud and not tell you at all that they're running your exact algorithm. The innovator would never know.

Big/old ("stagnant") companies, on the other hand, have enough legal manpower to sue you out of existence regardless of whether their patent is valid or not. It's not the patent, it's the legal process that will bankrupt the small innovator.

Can patents be molded to protect the small innovator? Because isn't that the intent? Regardless of whether that protection results in monopoly, disclosure or both.

Re: The Case Against Patents (2013)

#19

Post by jordigh » Tue, Jul 10, 2018, 2:01 PM UTC

beefield wrote: post #15

andrewla wrote: post #11

The authors summarize as: > ... there is no empirical evidence that [patents] serve to increase innovation and productivity ... This seems to be a fundamental misunderstanding as to what purpose patents serve. The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them. The cost to the public of pa…

> The public benefit of a patent is that the information is public -- inventions which might previously have been kept secret are instead put into the light where others may use them Can you come up with siginificant patented innovations that are not easy to reverse-engineer? I have thought that the raison d'etre of patents is to foster innovation.

I always mention the Michael Jackson smooth criminal patent. Is it easy to reverse-engineer? I don't know, it seems obvious once you see what it is, but it seems awfully clever if you don't know about it. A lot of people seem to believe that it would involve electromagnets, but the actual invention is more low-tech.

Re: The Case Against Patents (2013)

#20

Post by renox » Tue, Jul 10, 2018, 2:02 PM UTC

jordigh wrote: post #9

I never thought that the case for patents was to increase productivity. I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave. My favourite patent of all time is Michael Jackson's smooth criminal patent: https://mashable.com/2015/03/28/michael-jackson-shoe-patent/... This is the patent…

> I always thought that their purpose was to eliminate trade secrets: people are going to invent stuff whether you want to or not, but without patents, they'll take the secrets to the grave.

Except that if you look at the history of inventions you'll notice that quite frequently several people invented the same thing independently at almost the same time (the phone for example). So the risk of losing a 'trade secrets' isn't that high..