Ideas are global though, and the patent system isn't.

In order to turn a patent in to a literal recipe for a competing product all you have to do is step out of a registered jurisdiction.

So people who actually come up with something meaningful have the financial burden of registering in numerous major jurisdictions at a minimum.

However, each comes with its own quirks, costs, and processes, some of which can be slower (eg. Japan) or qualitatively different (actual novelty checks performed vs. only cursory inspection) and thus vary from verging on legally dubious (more like "option to expend lawyers on creating problems for competitors in the future" than actual patent) to total lockdown.

It's a mess. Mostly it only keeps patent lawyers in business. We need to get rid of it.

> In order to turn a patent in to a literal recipe for a competing product all you have to do is step out of a registered jurisdiction.

When you have a registered patent in the USA and the EU, there is no real alternative market for competition except India and China - because both the US and EU will reject any import of a patent-violating product.

Well those are some big generalizations. Objectively, USA and EU are neither the largest and certainly not the fastest growing markets for many products. In other words: even if that is true - who cares? Also, traditionally Japan is considered a major developed world market by patent law.

weird juxtaposition of arguments. you start by arguing that one system's rules aren't relevant when there are many jurisdictions.

then you talk about each jurisdiction having very unique flavors & characters to them.

the favt that the uspto governs such a vast share of products-sold carries a lot of weight. I tend to think increasing the cost of patents is a silly idea, because right now they cost nearly nothing if you file & do the reject/re-apply dance yourself (and know how to). unless we the people are going to start charging billions for good ideas being monopolized, I tend towards a fair reasonable & non-discriminatory access model for all ideas, with a depreciating over time scale. i'd rather open the closed system than try to find the right price for it. but either way, I think the uspto & Congress has a lot of power to fix this practically free hand out that ongoingly locks down this world & obstructs progress time & time again.

If we begin with the unspoken assumption that the system is supposed to be useful, then it's effectively dead in the water because ideas aren't under national jurisdictions. I think this "weird juxtaposition" is the innate problem with the status quo.

Nobody, and certainly not the USPTO, is able to determine the future value of any patent.

Nobody can predict the future value of pretty much anything, yet trading happens all the time.

To improve the predictive capability of the USPTO, they could host a bidding process in which other companies can participate.

I have pretty low faith in USPTO (or, to be fair, any other organization), making reasonable predictions on the technology value. And even if they did, I think inventors would feel low-balled: "those idiots did not understand the revolutionary benefit of my invention", etc.

Rather than trying to estimate the value of the idea we could drastically reduce the time for patent protection. Shortening protection period to 2-5 years should still encourage inventions, but would discourage early carpet-bombing filings (which scare many inventors away) as well as move successful inventions into public domain while they are still relevant. My 2c.

5 years is plenty of time for patenting some clever software trick that took a minute to think up and a week to implement (eg. minigames in loading screens). But some inventions take years or even decades of research (eg in materials science) and, even after being invented, may take a year or more to spool up the production line, let alone the procurement process.

Would it make sense to have the patent office choose the patent length between say 2 to 20 years?

Estimating the value seems pretty much impossible, but estimating a length seems more reasonable. Clever software trick, quick to invent, easy to capitalise -> 2 years. Ground breaking new drug/material with slow market deployment -> 20 years.

There's still some subjectivity involved (but that's already the case with the current patent system), but you could try to base it on some objective metrics.

Or perhaps it could be a fixed duration per field, somehow, so that you know before filing what you will get.

It would be easy to discriminate between 20 years for drugs vs. 5 for anything software, but I guess it's tricky to decide whether something like e-paper should count as fast-moving consumer-facing tech or hard-won physics/chemistry knowledge.

If something is in research phase it should not be patented yet. And once invented, a 5-year protection still gives the inventor a 5-year advantage. If it cannot leverage it effectively, sorry but too bad.

No method is perfect, but I think that in most cases a 5-year protection is enough to encourage invention.

Many products contain hundreds of distinct "inventions" in the patent sense. One can't necessarily hold off on registering a patent for an individual invention for 5 years while the rest of the product is developed, as you'd risk a competitor inventing the same before you have patent protections.

Patent does not protect a product, but a specific technology. If 1 technology out of 10 in a product is under patent protection the competitor (and the public) cannot copy the product as is, but is welcome to use any expired technologies any way it sees fit.

In general, it is a business decision on whether to patent a technology that it is not ready for the production: the risk of not making a product in 5 years vs the risk of competitor getting there first. It is not perfect, but I would prefer that inventions that the owner wants to own (and suppress competition for) but cannot commercialize in 5 years enter public domain. My 2c.

That's kind of what the private market already does, without the PTO's involvement. If people think your patent has value, they can license or or buy rights to it.

No, because if the patent applicant keeps the technology secret, there is no patent. Meaning the road is open for other companies to invent the same technology.

The point of patents is that the invention should be described, and patents are publicly accessible. This is what allows you to discuss infringements in a regular court.

Just to point out, there do seem to be certain (niche) fields where Patents can be kept "secret". eg National Security, Nuclear, (etc) type of things:

https://slate.com/technology/2018/05/the-thousands-of-secret...

it being impossible to get 100% right doesn't mean we must be afraid to try to assess merit.

perhaps a fair value assessment might include some % of revenue or profit, to self adjust?

It's not likely to be remotely correct. The only people who have a chance at it are those who are willing to risk their own money on it.

I want to nitpick. Monopolies aren't bad. Monopolies tend to be dangerous because they aren't self regulating and they can easily extort consumers. But monopolies can also be very useful to consumers if they are acting in their interests, particularly in goods that have network effects (where the number of participants directly increases the product's utility or makes it cheaper. As opposed to a S curve). Bell labs existed because the government gave Bell a monopoly over telephone networks so there wouldn't be hundreds of wires everywhere (literal network), but the lab was part of the deal.

Why would the USPTO not rig things in favor of US companies by underpricing the value of patents from foreign companies?

Compulsory licensing is already a thing in many countries. It's mostly used in drug manufacturing.

Copyright also provides a monopoly, that’s the whole point, if it didn’t then the GPL, etc would be useless.

Why should I invest in R&D then

That's right, the greedy approach would free a patent now to unlock a technology, but what do we do about not freezing creativity over the next decade?

I am no fan of the patent system, but what you propose is worse. Companies can get a license but by a price defined by the market. Giving this power to some burocrat seems odd.

If you merely double the cost though it removes the human element for error.

For every complex problem there is an answer that is clear, simple, and wrong.

H. L. Mencken