US Patent Office proposes rule to make it much harder to kill bad patents | Hacker News Reader

US Patent Office proposes rule to make it much harder to kill bad patents

techdirt.com

541 pointsberkeleyjunk100 comments

This is a bad proposed rule that should be killed, to be sure. But this article is missing the context about _why_ the Patent Office claims to think this is necessary in the first place. Patent trolls of a sort are setting up shop on the other side of where you would normally expect them in litigation, attempting to use the Inter Partes Review process in ways that look very much like an abuse of the system. For example, one company opened an IPR against a valuable patent and then offered to withdraw it for money. It was rightly called out as a shakedown.

https://www.jdsupra.com/legalnews/director-vidal-removes-ope...

https://news.bloomberglaw.com/ip-law/opensky-abuse-sanctions...

The US legal system generally requires parties to have standing before they can make use of courts or pseudo-courts like the PTAB. Situations like this make it clear why: when lots of money is on the line people will try to “hack” the system in ways that its designers didn’t expect. Uniquely open processes like IPR are vulnerable to exploitation in a way that is annoying to lawyers compared with what they’re used to, so they reflexively reach to reimpose standing requirements. That’s basically what is being proposed here.

Again, the above doesn’t mean I think this rule change should go forward: the IPR process and its openness are incredibly valuable and should be retained. but it’s misleading to portray this proposed rule as arbitrary or corrupt.

Yours is a great comment.

Since patent trolls already "hack" the system to essentially run a legalized protection racket, it only makes sense there should be some counter balance to their abuse. Fewer barriers to invalidate bad patents is such example.

And in this case, this new proposed rule is simply pushing the balance in patent abusers' favor, so it is corrupt because they are pushing it for their racketeering profits.

A better solution would be to prevent more bad patents in the first place (like forbid software patents for example) instead of making such kind of workarounds. But they aren't doing it either.

I'm in the patent biz. While they did not forbid software patents, it's currently the hardest it's ever been to get a new software patent. It's also easier than ever to invalidate a software patent. Even if something like this proposed rule came into effect, it would still be historically easy (quick and cheap) to challenge a software patent.

It's good, but not good enough. So any movement in the opposite (bad) direction should be seen with strong suspicion.

That's good context, thanks for sharing!

It makes sense, but I'm leery of trusting the USPTO because they have been saddled with (intentionally?) godawful incentives: they must pay for all their operations using patent fees. The more trolling they enable, the more money they get. Based on these incentives, I would expect the USPTO to oppose rule changes that effectively reduce trolling.

That’s a great point. I didn’t know the patent office was _entirely_ funded by fees but that does indeed appear to be the case.

https://www.uspto.gov/about-us/performance-and-planning/budg...

  > With full access to the fee collection estimate to offset total spending, the USPTO will use $32 million (net) from the combined operating reserves (ORs) in FY 2024, resulting in a net appropriation of $0.

> The more trolling they enable, the more money they get.

How do trolls pay anything to the USPTO? Usually the patents a troll tries to assert against others are purchased from others, often failed tech startups. (Note: People often forget that every “bad” patent was once someone’s startup dream. Edit: OK, maybe not every "bad" patent was once someone's startup dream.)

I think your Note at the end betrays your misunderstanding of the situation. There are people who make it essentially their full time job to fill out junk patents and see what they can get approved. They are paying USPTO fees to make this happen. They then use the portfolio so built to extort actual practitioners, or they sell/trade portfolio items to other trolls. There was never any startup dream involved

Often involves buying otherwise-worthless IP from defunct companies with the sole purpose of using it to sue others

If you look at the records of issued patents and which ones are owned by known "trolls," a very small minority of them come from this approach. Writing patents to anticipate future inventions is risky and expensive, and most of those patents will be worthless because a non-expert can't anticipate the technical direction of a field very well. Buying failed startups is cheaper and easier, and a lot less uncertain.

They encourage everyone else to patent everything that could be possibly be approved, regardless of value or intention to use in the business, as a defensive measure. This is the literal policy of every big tech company I've ever worked for.

Why not just simply raise the fees?

I don't really get it, anyone who is a neutral party in any kind of legal matter and takes money from an involved party for a specific action is guilty of criminal conspiracy, no? I mean can I citizen arrest people for other thefts in the US and offer to split them?

Offering to settle is usually fine for involved parties (as far as I understand, I’m not a lawyer). The “problem” (from the USPTO’s perspective) is that congress set up this process so any outside party can bring an IPR case, making it potentially profitable to pose as a public interest group when you might actually have intended all along to settle without invalidating the patent.

Yeah but people are missing the point that congress did that in order to make it trivially easy to attack small-time patent owners.

The only way to fix situations like this is to make it possible to extract heavy penalties for frivolous patent litigation. The cost benefit analysis will always fall too much in favor of patent trolls otherwise.

That is a fine idea in the abstract. But there are a lot of high-value patents that are, for example, very badly written, or they have other attributes that make them vulnerable. This makes them vulnerable to a shakedown. Discerning if a badly written patent deserves to stand would give me a bigger headache than the already soul crushing task of reading such patents.

I am not sure if I understand your point. Are you saying that poorly written patents should be litigated, and that such a law would make this much harder to do?

I can see how something like this could be an issue. Ultimately, I don't think there is a perfect solution, but I feel the current way the system works is very far from optimal in its costs and benefits.

I do not think there is a clear answer to this. I can understand some people taking the position that a poorly written, hard to understand patent does not deserve to survive, especially if if the people who wrote it are trying to be clever by having it mean what they say it means if it gets litigated.

OTOH it can also be argued that trolls posing as public interest groups are operating in bad faith and the fact a patent is hard to understand has been par for the course for so long, that making patent holders review their patents for clearly defensible language is unfair.

EFF is an industry group posing as a public interest group, operating in bad faith in order to dissolve copyright and patent law.

> Are you saying that poorly written patents should be litigated,

Why shouldn't they? Courts don't necessarily invalidate contracts just because one flaw might be present. The point of the system isn't to just invalidate patents because someone else wants to... it's to fairly adjudicate the boundaries of granted patents. You are starting from 2nd base!

It sounds like a class action in a normal court, where a lead plaintiff legitimately can get better treatment to a point. Viewed from that perspective, it's a wonder IPR has been so noncontroversial for so long, because the rules around class actions are extremely contentious.

> the IPR process and its openness are incredibly valuable and should be retained. but it’s misleading to portray this proposed rule as arbitrary or corrupt.

Both things can be true. I'm sure there is a need for changes to IPR, but that doesn't mean they didn't know what they were doing when this was proposed. It's hard to imagine a good faith effort to end abuse of the system that also just happens to protect bad actors. Is it really likely that they had no idea that this would drastically weaken the ability to challenge bad patents? I suspect they knew exactly what the proposed rules would mean, that there are other options that would help prevent abuses like the one you pointed out without making it needlessly difficult for legitimate challenges.

Sure. It certainly seems that this is being proposed for reasons more aligned with the interests of patent holders than patent challengers.

I’m just pointing out that without a discussion of the recent events which clearly motivate this rule change, the narrative is missing important context that makes the USPTO look like mustache-twirling villains for proposing this. They may in fact be villainous, but readers still deserve the full backstory to come to that conclusion themselves.

> Inter Partes Review process in ways that look very much like an abuse of the system

That's rich considering that IPRs were basically invented to give big companies the upper hand. For a while, they ran in parallel to actual district court litigations over the subject patents! Even now, they'll stay litigations! The same exact defendant, instituting a parallel challenge of the same exact patent, but in a pseudo court! IPRs are abuses of the system!

> But this article is missing the context about _why_ the Patent Office claims to think this is necessary in the first place.

Their claimed rationale for these changes is simply dishonest. Hundreds of patent trolls have been abusing pro-patent-holder rules for years and they never blinked an eye. USPTO only cares about "abuse" when they can use it as an excuse to gut a mechanism designed to (and working well to) combat patent troll abuse. They are an organization entirely captive to patent holders with zero interest to benefit innovation or society at large.

The legal system is a game that is played for profit. Being able to sue anyone for anything is advantageous for the big dogs. So it's imperative that as many laws as possible cover as much seemingly innocuous human conduct as possible with the highest stakes possible. Unrestricted expansion of intellectual property, the ability to lay claim to arbitrary regions of the ideosphere, makes perfect sense.

> Being able to sue anyone for anything is advantageous for the big dogs

Unless they can justify very high damages, they are predisposed to settling out of court. When you hire top lawyers, or have a massive legal department, you're paying a lot of money. Going to court is at best a gamble unless you have an obvious and solid case.

What this means is they use threats of going to court wrapped up in legalese in the hopes of getting their way out of court.

Unfortunately, the degree to which various districts earn reputations around being pro or anti patent means they're also advantaged in "shopping around", so to speak, to get any case they bring moved to a favorable court. This is the biggest thing that they can do that your average "small dogs" have a harder time with.

The worst has changed in the last year:

https://news.bloomberglaw.com/ip-law/patent-plaintiffs-scram...

but it'd be nice to see ways to make it more difficult to game the system by "judge shopping".

This comes up at least once a month on HN.

Drug and medical patents have at least some moral claim to virtue (don't come at me on that one). Whenever you see an op-ed about how patents are so valuable, those are always the poster children.

It's never software patents, because absolutely no one but lawyers thinks those are a good thing. But they're subject to the same laws as the "good" patents (whatever you think those are).

So the solution is clear: Congress must legislate that software is not patentable subject matter. We can do this.

I don't know any lawyers who think dumb software patents are good, including my own patent lawyers.

> dumb software patents

their definition of "dumb" is probably quite different from yours (or most software engineers).

To clarify, I think both of us agree that things like weird data structures, compression/encryption algorithms, and a few other things that are actually non-obvious hard tech probably should be patentable (although it's usually a shitty business strategy to patent these things), but that most other software (including things like Amazon's one click ordering patent) shouldn't be. I will admit that I don't know where his moral line on this is exactly, and I know he is very happy to draft patents beyond it for strategic reasons.

I think many smart patent lawyers are ultimately legal realists here - they have their own philosophical views, but their job involves working in the real system.

>To clarify, …

Not the parent, but: in that case you and I agree to disagree: any weird encryption algorithm is the usage of mathematics which should not be patentable in my opinion.

The simplest solution in my mind is simply abolishing patents. Just as e.g. the COVID vaccines were developed using public funds, there is enough ‘protection’ in things like trade secrets (‘recipe of coca cola’) or copyright/-left (‘GPL’ or ‘iOS’) to keep us innovative enough.

I honestly don't think the patent office can draw the right line on what mathematical formulas are discovered vs invented, so that seems reasonable to me.

I think you are looking at the abolition of patents from a very "software" perspective. Copyleft license terms have basically never been tested in court, and could just be unenforceable, but the tradition of software engineering seems to keep the peace there - GPL'ed and other open source hardware (which frequently has copyleft licensing) frequently gets ripped off without credit and nobody does anything about it. These would have bad facts in a court case, since the copyright doesn't cover the thing, just the schematics.

Trade secrets also generally don't protect you when your product is easy to reverse-engineer and you have customers or competitors with a lot of money to do it. A drug is a prime example of this.

Step by step: first remove software from the domain of patent law. That IS a heavy lift, but abolishing all patents is an impossible one.

Then the question of "OK, what IP protections should apply to software, if any?" can be addressed. I don't claim to have the answer all worked out

That seems like a reasonable approach to me, and may come with a natural experiment built in.

However, it's hard for me to suggest that an algorithm like PageRank, for example, is less inventive or risky to develop than a typical drug, which would suggest that some software ought to be available to patent.

As Ripley would say, nuking it from orbit is the only way to be sure.

Whatever rules you attempt to come up with, the lawyers will nitpick them to death and find some way to keep junk patents alive. For years, everyone in the legal profession looked forward to the CLS Bank v. Alice case at the Supreme Court as the Final Word on software patents. That was 2014.

My article:

https://albertcory50.substack.com/p/lets-vote-on-it

lists some of the patents that have been upheld since CLS Bank.

Very interesting blog. I might be n = 1 as a software engineer who occasionally reads patents to learn how they work, but I also have some experience with the patent system and can read "patentese" better than most. They often leak more details than their papers.

I think a key point here is that Alice is clear as dirt on its intentions. SCOTUS could have come up with a much cleaner test but didn't, and has had the option to clarify their thinking several times since then but has never taken it. I guess patents aren't "sexy" enough.

I'm not sure how you think copyright would protect that which protects patents, but if your position is that patents are bad, so that in the alternative, corporations should keep everything secret forever (that's what you have to do to have a trade secret...) seems a bit absurd to me, to be perfectly honest.

I love how software engineers think they have a clear understanding of what should or shouldn't be patentable, yet cannot come up with words to describe the delineation more accurately than "good" or "bad" or "weird", or other ways of just facially repeating patent requirements like "non-obviousness", yet don't seem to grasp this is the exact challenge that the law has to deal with with regards to patents. It's not an easy challenge. It turns out, it's actually incredibly hard to come up with a law that describes what should or shouldn't be patentable, with regards to anything that has yet to be invented.

You seem to think that, having mastered what the law IS, you're also an authority on what it should be.

Wrong. The people who live under the law are equally well qualified to judge that.

That's not what I think at all - I'm actually begging you to come up with a better delineation.. that's the challenge the law faces in regards to what it should be.

We can't write a law that says "good patents are good, bad patents are bad" there needs to be a tangible, comprehensible delineation between good and bad patents that can be applied to things not yet invented... that's the challenge.

I'm not sure how you came to your response at all, to be perfectly honest. I'm certainly not saying that lawyers are the only people who should be involved. I'm saying, if software engineers want to help... they need to actually help instead of just pretending like its trivial to delineate the good from the bad. It isn't. Really, the point I'm making is that, as experts in the field, if you cannot come up with a better delineation than simply stating "good" or "bad" - how is that you expect lawyers to do a better job?

To further my point, that you responded the way you did is evidence that software engineers aren't actually engaging in the problem and are just sitting on the sidelines having tantrums because things aren't going the way they want.

OK, so do I understand you correctly that you'd like to write a law that separates the good patents from the bad ones?

Tell us: is this a "good patent":

https://en.wikipedia.org/wiki/Enfish,_LLC_v._Microsoft_Corp.

How about this?

https://patents.google.com/patent/US7818399B1/en?oq=us7%2c81...

Both of these are considered valid, post-Alice.

I'm not sure you are getting my point, so I will repeat myself. The challenge in patent law is to come up with an actual law that delineates good patents from bad patents. This is an extremely difficult task, if not impossible. So, simply stating, from the sidelines "bad patents shouldn't exist" doesn't actually help anything, because it provides no way for the law to filter out bad patents.

My point being, I don't think you or others here appreciate that challenge, and instead just do things like you are doing here, where you ignore the questions I pose to you, and instead ask me to defend these patents. I never said either of those were good patents or should be patents... so why am I being asked to defend them?

The crucial question is... can you provide a better delineation between good and bad patents besides just facially calling them good or bad? It doesn't appear to be that you can do that. Given that you know more about the technology than lawyers do, how is it that you suppose attorneys make this delineation?

As a matter of convenience, I'm not going to respond if you don't answer my question, because it will be clear to me that you are not reading or engaging with my posts because this will be the third time I've posed this issue to you, when you've completely ignored it the first two times in order to step up onto your soapbox while repeating the categorical fallacy I accused you of.

So, as I understand the deal here: if I come up with an actual law, then you will tell us whether the patents I named should survive or not? Is that the bargain?

Or are you just going to critique what I say?

There's no deal. You keep avoiding my questions. It's very simple.

If technologists can't come up with a delineating line between valid and invalid software patents, how do you propose a judge to do so?

You sit here and say good patents are okay, but bad patents must go. So I ask you again and again, define a bad patent... no response. You ignore and you flip the conversation to ask me about already existing patents. They aren't the issue, the issue is coming up with a way to filter them out...

> If technologists can't come up with a delineating line between valid and invalid software patents, how do you propose a judge to do so

I didn't say a judge should do it. I said the Congress should do it. And what Congress should do is a political question, not a legal one, which is why our opinion is as valid as yours.

As for 'just answer my question': I think you mainly want engineers to say something concrete so you can critique it. So how about if you tell us: what are some examples of good software patents?

>I didn't say a judge should do it. I said the Congress should do it. And what Congress should do is a political question, not a legal one, which is why our opinion is as valid as yours.

Is there something wrong with you? I'm clearly valuing your opinion... this will be like the 7th time I've asked you for it, yet you refuse to give it and now insist that I'm being some sort of gatekeeper for patent opinions... this is bizarre. Hoping its because english isn't your first language.

>As for 'just answer my question': I think you mainly want engineers to say something concrete so you can critique it. So how about if you tell us: what are some examples of good software patents?

Lol dude why are you being this obtuse? It's not a trick. It's just that hard to come up with a workable line here. It really has nothing to do with my opinion on enfish. Engineers are here saying it should be easy to filter out the bad patents... okay, if so, tell us how...

I'm here waiting.

Since you are now reduced to hurling personal insults, we're done. Dude.

I said up at the top, "Congress must legislate that software is not patentable subject matter." So apparently you're the one who can't understand English.

Buh bye, dude.

I think the definition, for a lot of people, is like the famous definition of pornography - I can't tell you what it is, but I know it when I see it. This is a prime area for courts to provide some judgment, but only if they had a clear set of criteria (not necessarily a bright line, but a set of criteria as to what makes a software patent sufficiently "inventive") on which to judge software patents, which they definitely don't have today.

>I think the definition, for a lot of people, is like the famous definition of pornography - I can't tell you what it is, but I know it when I see it. This is a prime area for courts to provide some judgment, but only if they had a clear set of criteria (not necessarily a bright line, but a set of criteria as to what makes a software patent sufficiently "inventive") on which to judge software patents, which they definitely don't have today.

There's a lot more patent litigation than obscenity litigation, so I don't think that's a workable stadnard. Also with much more money at stake. And given the responses here... strong feelings as well. So I don't think that really works given the complaints here.

> This is a prime area for courts to provide some judgment, but only if they had a clear set of criteria (not necessarily a bright line, but a set of criteria as to what makes a software patent sufficiently "inventive") on which to judge software patents, which they definitely don't have today.

That's not true. Patents today have to be just as inventive as they were before.

To show how you are proving my point... I asked... "hey software engineers, help us attorneys out and tell us where the line should be" your response was "well, there is no clear line and I can't help you find one but judges should really provide one..."

I just want to reiterate that if technologists cannot delineate a clear line between the good and bad software patents... please literally explain how a judge will be able to...

As I said above: I will propose a law (or rather amendments to existing law).

And in return, you will tell us whether some of the patents that have been upheld by CAFC post-CLS Bank v. Alice deserve to survive or not?

Are we agreed?

I don't see what one has to do with the other. The issue isn't that you or I agree or disagree. The issue is that no one has offered a workable solution to the problem, yet people here insist that its some sort of trivial thing. You refuse to acknowledge that and just keep asking me about enfish... why would I agree to play your game? You can't even honestly discourse with me

It IS simple; it's just not easy.

I've been around negotiators, and you always want to make the other party commit themselves without offering anything yourself. So: would you keep any software patents if it were up to you, or would you not?

If you don't want to talk about Enfish, then pick some others.

I think I finally understand what's going on. You don't have a way of keeping out bad software patents while allowing good ones in. You just don't want any software patents at all, and that's your solution. I think this conversation is probably over.

As I said at the beginning, "Congress must legislate that software is not patentable subject matter." I'm glad you finally understand that.

> There's a lot more patent litigation than obscenity litigation, so I don't think that's a workable stadnard. Also with much more money at stake. And given the responses here... strong feelings as well. So I don't think that really works given the complaints here.

I will just point out that the amount of litigation that has occurred and will occur in the future is actually a better justification for the "clear criteria, unclear boundary" approach (like how we define pornography) rather than a one-size-fits-all boundary, which would be pretty much impossible to draw anyway given the breadth of human knowledge. I believe that you are asking for the latter, correct?

> That's not true. Patents today have to be just as inventive as they were before.

There is no question that they have to be inventive. The question is how "inventiveness" is defined with respect to software under the (unquestionably true) idea that laws of nature cannot be invented. Since algorithms are arguably like laws of nature, there are very unclear rules right now about what kinds of software innovations reflect "invention" vs "discovery," particularly post-Alice.

>I will just point out that the amount of litigation that has occurred and will occur in the future is actually a better justification for the "clear criteria, unclear boundary" approach (like how we define pornography) rather than a one-size-fits-all boundary, which would be pretty much impossible to draw anyway given the breadth of human knowledge. I believe that you are asking for the latter, correct?

No it's not. It makes litigation much more expensive and leads to the exact problem people here are complaining about "bad software patents shouldn't exist!"

>There is no question that they have to be inventive. The question is how "inventiveness" is defined with respect to software under the (unquestionably true) idea that laws of nature cannot be invented. Since algorithms are arguably like laws of nature, there are very unclear rules right now about what kinds of software innovations reflect "invention" vs "discovery," particularly post-Alice.

I'm well aware. I'm pointing out that saying good and bad hasn't gotten us the solution so until someone proposes something better, software engineers need to realize that we are stuck here. You are once again, not providing a delineation between what is validly inventive and invalidly inventive. If the technologists can't do it, how can a judge.

Dumb or not, no software patents are good, they shouldn't exist. Patent troll outfits are often run by lawyers, so they clearly see it as "good" for their protection racket. MPEG-LA for example.

True, but he meant there are lawyers who'll cheerfully admit that a lot of patents are crap, but they'll defend plenty of other ones.

I'm agreeing with you: Congress can end the legal hairsplitting, and just establish that software's not patentable.

Before we stopped funding things properly, medical research was undertaken in publicly funded bodies. Thats where Insulin, and Penicillin emerged from. Prior to this of course drugs like sulpha and their precursors lay in the fully commercial dyestuff industry of "Mauve" fame (the Bayer-Hoechst thing)

A shitload of science is done on the public purse.

I don't want to objectify you, and you aren't wrong because stuff is now being done speculatively for profit day 1, but the "moral" aspect of this is really dubious to me: If we want to fund public health drug research, we can. And if we did, there would be no reason to vest patents over the discovery, we sunk the public purse into the drug design for public good outcomes.

I back India every time on this: Fuck the patents, they just make the drugs.

> I don't want to objectify you

What??? Read the post again: it was studiously neutral on the topic.

I'm describing a political reality: software should be carved out of their space. They have a ton of clout in DC, and if you insist on killing drug patents, too, then nothing will ever be done. Guaranteed.

Yep. This is true. Until we get a good old fashioned FDR style socialist government back, its stupid to think about US centrally funded public works in any dimension.

It was studiously neutral. But, the world is bigger than the USA and US patents leak into the world, which is a problem: other economies do it differently.

Drug patents are bad. The pragmatic good in the short term is that they got us the covid vaccine, and will get us malaria, and better flu shots and probably personalised cancer treatments. All of that should be nationalised, and made available worldwide for public good.

Fuck the shareholders.

(not you. you're just calling it as you see it. I'll say it again: you aren't wrong)

Software patents are both bad, and stupid.

> Fuck the patents, they just make the drugs.

> Drug patents are bad. The pragmatic good in the short term is that they got us the covid vaccine, and will get us malaria, and better flu shots and probably personalised cancer treatments. All of that should be nationalised, and made available worldwide for public good.

> Fuck the shareholders.

I wonder if you have any experience at all in a non-trivial business where heavy investment is a necessary precursor to getting something to market. For example, a hypothetical where you have to invest a million dollars cash and a year of work by a team of ten well-paid professionals (which is expensive) to get a product to market.

I say this because anyone who has this experience would not say the things you are saying. Not because you altruistic sentiment is wrong. It isn't. At all. Your perspective is rooted in public good for the greatest number of people. Again, nothing wrong with that at all.

The issue is one related to business reality. If the prevailing mode of operation is for society to just take what you develop through heavy investment of cash and time and give it to everyone, the problem quickly becomes evident:

Nobody is going to take the risk or make the investment.

"Fuck the patents, they just make the drugs" is a commendable position that can only be held by those who do not have to make the investment necessary to create and support those drugs or products. From that perspective, it is easy to steal anything. Yes, it is stealing when you take someone's work-product and do not compensate them at all.

To use an imperfect example to illustrate the idea, it's like a situation where you spend five years saving money and slowly using it to build a house, only to have a bunch of people move-in and take it once built...and they don't pay you anything for it. In parts of the world this kind of thing actually happens, I know someone who had this problem in South America.

It's fine to want better things for everyone. However, these ideas have to be rooted in reality. If everyone steals nobody is going to have any interest or motivation to take risks and build better things.

Thanks. Somehow it's not even possible to talk about "software patents" without someone insisting on dragging in "drug patents." Why would you deliberately make yourself a target for one of the biggest lobbies in DC?

They are separate discussions.

What? I am literally replying to someone who said:

> Drug patents are bad.

Not sure what twist you might be referring to.

Look, I agree. I, mostly don’t like patents. I think most everything being patented are implementations, not inventions.

Implementation, in this context, means that if I ask a group of reasonably-trained engineers to create an LED light with multiple color channels that can be controlled individually, they will instantly reach for any a range of well known techniques —like PWM— and just implement it.

Yet, years ago the USPTO granted lots of patents on this very topic to a scammer under the company name “Color Kinetics”. Complete horseshit, and yet the USPTO thought they actually invented something any engineer could implement in a day or two.

Where patents make sense is in any work that represents real invention and requires non-trivial investment and work to bring to market.

Any patent. Type should not matter so long as the USPTO competently filters for real invention.

It should be hard, very hard, to get past that filter. And, if it was that difficult it deserves recognition and protection for a period sufficiently long to both allow the inventor to justify the risk, effort and investment and motivate them to take the leap.

Without this engineers and companies will choose to create ways for people to click ads and pass on the hard stuff anyone can steal.

It isn’t as simple as “I hate patents”, one has to understand how and why things are created and what it costs.

OK, I've seen your posts before and getting triggered by a bad word is not something I'd expect from you.

> Until we get a good old fashioned FDR style socialist government

Until then, we can excise our field from that cesspit. That's something could be done without a revolution.

> What??? Read the post again: it was studiously neutral on the topic.

Bullshit. Your post starts with the assumptions that: 1) good patents exist, and 2) medical patents are good.

Neither of those assumptions are "neutral".

> Bullshit. Your post starts with the assumptions that: 1) good patents exist, and 2) medical patents are good.

You're so articulate.

"Drug and medical patents have at least some moral claim to virtue " is not an assumption that "medical patents are good." No it doesn't. Unless subtlety is lost on you.

There was a big discussion of the EFF post this one riffs on last week https://news.ycombinator.com/item?id=36198329

Also see https://www.linuxfoundation.org/blog/stand-up-for-open-sourc... which didn't get much discussion https://news.ycombinator.com/item?id=36154988

Our entire patent system in the US is designed to benefit lawyers (keep them employed) and large corporations. Fostering innovation and supporting inventors is now completely secondary (or not a goal at all)... probably been that way for a long time but it is only getting worse.

It seems that bad patents would hurt less if they didn’t persist for so long. Why can’t we come up with some sort of earlier phase-out, where patents must be re-litigated or justified to remain in effect?

Because the foxes are in charge of the hen house. The lobbyists, the congressmen, the patent trolls, the patent office - none of them would benefit from weakening patents.

I tend to agree, but the introduction of the Inter Partes Review process at issue in this rule is itself an important counter example.

https://en.m.wikipedia.org/wiki/Leahy–Smith_America_Invents_...

It was introduced not that long ago as part of bipartisan legislation that meaningfully reformed the existing patent system, including by introducing systems like IPR which clearly weaken the rights of patent holders (who have been fighting ever since to overturn it). If the system were utterly irredeemable, reforms like this should never survive.

The patent system is flawed enough that I think a rewriting from scratch is required.

It was created or a different time, and its governing rules reflect it.

How to rewrite it is no simple task to figure out and is well beyond what I am able to do.

I do think the bar for applying ought to be higher and the categories that can be patented significantly lowered.

I am not sure, but it seems that the time it takes to grant A patent requires changes as well.

Nearly everything new has "pending patents"

A good thing would be changing the ethics of how people think about patents and applying for them.

A better, quicker and easier way to prove prior art should be developed. (and if it is not already the case, the entity that applied for a patent with prior art should be made to cover all legal and paperwork etc. for doing so.

The existence of the patent system makes me never want to invent anything ever. Having to deal with that thing? That would be like working in Academia. Wouldn't wish it on anyone.

The US is still making itself less competitive while China is closing the tech gap at a breaking speed. I really don’t understand what they think this will accomplish.

I don't even know why they bother at this point. Why not just cut to the chase and say if you give us enough money we'll give you the power to stop whoever you want doing whatever they are doing right now.

There's not even a pretence this is helping innovation any more.

Because if they did that, then you would get corporate white knights pretending that any of this makes sense in a healthy society.

The pretense of a rules based order is what holds the sharade together.

Can't we let the free market solve the problems with patents?

Like, if the government (or anybody, really) wants an invention to be open, they can offer the inventor/company some sum of money, and if they accept then they will have to explain how it works. Otherwise, it will remain a trade secret until someone else figures it out.

Patents already explain how things work. The problem is limiting other people's ability to do something that someone claimed as their own with basically no effort.

That's effectively what the patent system already does. The "sum of money" you refer to is the market value of the monopoly on the idea embodied in the patent. This solves the problem of what the price should be.

It's not the same because a patent is a time-limited monopoly on a technology. This means that patents can be used to block the progress of other companies.

E.g. if A patents X and some large company B wants to use X and A refuses, then B cannot use X even if they offer millions.

You are also right. I was responding to this. (Italics mine)

> Like, if the government (or anybody, really) wants an invention to be open,

The government literally does want inventions to be open. It's in Article 1, Section 8, Clause 8 of the constitution. [0]

  [The Congress shall have Power . . . ] To promote the Progress of Science and 
  useful Arts, by securing for limited Times to Authors and Inventors the 
  exclusive Right to their respective Writings and Discoveries.

[0] https://constitution.congress.gov/browse/essay/artI-S8-C8-1/...

> E.g. if A patents X and some large company B wants to use X and A refuses, then B cannot use X even if they offer millions.

How does this situation improve in any way if X is a trade secret? Under what circumstance would company A be willing to accept payment to make X open, but not willing to license X as a protected patent?

It sounds like you'd be in exactly the same situation except that there would be a massive collective action problem trying to figure out how much the world should pay A to make the invention open and who should pay it, and also A and B would have multi-million-dollar incentives to engage in corporate espionage and counter-espionage that achieves no progress.

Patents exist specifically to avoid a fully free market. The trade-off is that in return for a time-limited legal monopoly on certain ideas and obvious derivatives, you have to publicly disclose what it is that you own. This also serves the practical purpose of allowing you to determine if you're potentially infringing anyone else because you can look at the patents. The free market approach would be no patents at all.

Patents exist to compel inventors to divulge their secrets, to allow future humanity to be able to replicate the processes when those with the secrets move on, by offering a time-limited monopoly in return as payment.

It is being suggested that if a secret is worth exposing, the people can simply make a direct offer for the things they want to know and let things they don’t care about, like how to buy an item online with one click, remain ‘secrets’. Patents wouldn't go away, but would by priced by their recognized value.

What would be no doubt lost are the things people don’t know that they want to know — ideas that aren’t commercially viable at the time of conception, for example. On the other hand, with the patent record now impossibly large, does anyone truly scour patents to seek out inventions they didn’t know they needed and derive a real benefit from them upon discovery?

This is how trade secrets work, not patents. The main point of a patent is to encourage inventors to divulge their secrets, by promising them temporary exclusive rights to control their tech in exchange for a public, detailed description of how it works. Eventually, the patent will expire and the technology will enter the public domain for the good of society. Without something like patents, technology would be much more secretive.

My point is that there is more than one way to make someone give up their secrets. Giving someone a monopoly seems wrong, because they will use it to block progress.

Governments already provide funding for scientific research. Why can't the same principle be used to "fund" the disclosure of technological secrets?

The point of the patent system is the exact opposite of what you claim.

To get a patent you must provide the full details of the invention, then your competitors can look at that patent and develop new systems that might be superior. This even applies to "software patents" (the different compression systems to beat LZ, or new versions of arithmetic coding - the most commonly derided "software patents" often seem to process/business patents rather than actual software).

This is the idea at least. In the absence of patent protection you get other problems: say you spent a year of your life coming up with a thing to do X, if you ship it some company can turn around and copy it without having to have invested any effort or money in the development, so can then also undercut you. "trade secrets" (an actual legal concept) doesn't protect you unless they steal the details from your company, but simply buying a product and reverse engineering (including decompiling) it is fine.

The problem is that the patent office allows patents that don't describe how to make the thing. That's where you get the BS style "X but on a computer", "X but with the internet", "X but with AI", ... or "compress data using some mechanism".

The patent office should just be requiring exact and explicit details in a patent application, so that you can read the patent and implement it. If you can't do that, the patent shouldn't be granted. (People periodically suggest "require an example implementation", but that for many fields there are huge costs involved in actually constructing a device so suddenly only people with a tonne of money can get patent protection)

Opinion: a truly free market is one where patents do not exist, in which case yes, the problem would be solved.

Free, and dead

No doubt the system is abused, but the solution isn't to eliminate all protections...Apple or Google have the money to instantly clone anything

Exactly, but also for most products so does any company that does any manufacturing.

There are very few companies that manufacture goods themselves due to the massive capital costs. But that means anyone can buy one the devices you invented, tear it down and reverse engineer it, and then send plans of to a (maybe the same!) manufacturing facility. They can even undercut you on price because they don't have to spend any money actually designing anything.

The "remove patents" argument fails for the same reason "remove copyright" fails. Developing anything takes time and money, and if there is no ability to profit from that (because anyone else can copy it - in the case of software this is literally - they can sell it cheaper than you), so there are legal protections for the "intellectual property" rather than just the physical goods being sold. If you remove them all sorts of things break: OSS/copyleft requires copyright - if you remove copyright then any company can take any "opensource" software and use it wholesale, and you can't require them to contribute anything back, or provide sources, or even make the binaries available.

> Basically, if a patent holder is designated as an “individual inventor, startup” or “under-resourced innovator” then their patents are protected from the IPR process.

Just curious...is there a good faith argument why this change would be good?

The good faith argument is that the people who are predating on the poor, beleaguered genius inventor would be unable to avail themselves of a low-cost-for-the-challenger-high-cost-for-the-inventor mechanism for challenging patents.

What makes it bad faith is that the reason the mechanism exists is because the predators in the patent system are the inventors, not the challengers.

They did leave something extremely important out here that helps to explain the good faith argument. From the source:

> Petitions challenging under-resourced patent owner patents where the patentee has or is attempting to bring products to market;

So I’d say the good faith argument is that it’s designed to protect larger orgs from bullying actual inventors/startups by attempting to threaten the legitimacy of their patents. The bit about bringing products to market is supposed to exclude trolls. More from the source:

> Such limited resources may impact the perceived fairness of post-grant reviews. For example, some stakeholders in response to the RFC expressed concern that under-resourced inventors are unable to afford the costs involved in defending patents in post-grant review. Some stakeholders advocating for small businesses and individual inventors urged the Office to take into account the financial resources of a patent owner, and to limit reviews of patents owned by under-resourced entities who lack funding to defend challenges to their patents but who have sought to bring their inventions to market either themselves or through a licensee.

So yeah basically the argument is that defending themselves in this proceeds is too onerous for small entities.

So I think that’s the good faith argument. I’m skeptical, though I will happily admit that I don’t know much about the uses and abuses of the patent system.

On the face of it, it is good. The problem is trolls masquerading as individual/small-business inventors.

Looks like yet another attempt at regulatory capture.

Such an obviously corrupt idea.

Reply on news.ycombinator.com