Especially recently we've seen the widespread acceptance of copyright and license laundering through large machine learning systems such as copilot and chatgpt, backed by enough microsoft laywers to prevent meaningful push back from the copyright owners. They'll output verbatim copies of what was previously copyrighted code.
If it is true that software copyright can be cleaned away in such a fashion, maybe patenting the algorithm itself actually is necessary to avoid big tech companies like microsoft from forcibly strip mining the ideas of independent creators.
> forcibly strip mining the ideas of independent creators.
this is the same argument that patent proponents had, because patents protects the little guy.
And yet, most copyright enforcement occurs via large corporations who have an interest in doing so to further some other goal as part of a strategy to gain monopoly. Patents as well as copyright is but a tool, and any tool can be more powerfully wielded if the wielder is more powerful!
The little guy is better off without copyright imho (but that horse has already bolted).
Patents protect the little guy from the manufacturers.
A patent is supposed to secure the effort of R&D without necessarily having the budget for massive manufacturing. It prevents the manufacturer from taking off with your idea and giving you nothing for it. It seperates the manufacturing cost and r&d cost.
In software, there is 0 manufacturing, it is entirely R&D. As soon as you have developed your product, its ready. Hence why patents feel so useless in software.
Also why software startups are so attractive to investors.
> Patents protect the little guy from the manufacturers.
As long as the little guy's patent can improve the process without stepping on any of the countless pre-existing patents manufacturers have accumulated. Now how often is that the case?
How well has copyright really worked for the little guy? Even when there are wins in the GPL violation suits, the payout barely covers the cost of litigation (i'm imputing the cost even if it is pro bono).
Not to mention tivoization which trivially bypasses GPLv2.
By the way, it is much easier to enforce a patent as a "little guy" than a copyright. Because patents entitle you to huge damages and are harder to evade, many layers will take complex patent cases on contingency (as long as you have a solid read on the claims) and there is a lot of financing available otherwise. On copyright enforcement, you're pretty much stuck.
any mechanism will be used by someone to protect and entrench themselves. Small or large.
The more difficult and expensive it is to use, the more it will be the domain of those with more resources, and the more the small will be defenseless.
The answer is not to eliminate copyright or patents, but rather to make them simple and easy to apply for, defend, and find. The current bar is so high it favors the big guys only, and gives them leverage against smaller folks who can’t afford to even show up to defend themselves usually.
Or eliminate it entirely, but you’ll not be able to do that because Disney, Microsoft, Oracle, and the like would murder/compromise/destroy anyone who might be able to do that.
The thing is that it probably isn't. The thing you call "laundering" is probably actually fine. The thing that people misunderstand is that copyright is written such that it never protects function. If your code is open source, then I can write code that does the same thing, and if I can't because your code perfectly encapsulates the function, it's inseparable from that function and thus isn't eligible for copyright protection at all.
> maybe patenting the algorithm itself actually is necessary to avoid big tech companies like microsoft from forcibly strip mining the ideas of independent creators
a. Independent discovery of same algorithm is very likely before 17 years.
b. Big companies abuse such systems by patenting trivial ideas.
The nice thing is that this works in both directions. Just as Microsoft and Co. can train on people's data, nothing stops the people from training on Microsoft's chatbot output.
As long as we don't get any lawsuits derailing this practice, we actually might end up fine, as there is nothing that'll allow anybody to get ahead in AI without automatically provide a mountain of training data for the competition.
If that means the end of copyright right as we know it, I am all for it.
Unsurprisingly, the "creators getting ripped off is a small price to pay for progress" stance is less popular here than when HN discusses AI-generated art.
Yes. Copyright already protects what a patent granted for source code would protect.
The problem you describe exists in both scenarios, and is resolved with enforcement. You can sneakily break the law, and you run the risk of getting caught for fraud.
> Copyright already protects what a patent granted for source code would protect.
This is not correct. Patent protects the abstract functional design, copyright protects the reduction to practice. Two implementations of the same algorithm have independent copyrights.
You see the same thing in physical engineering too. A novel chemical process is protected by patent and each reduction to practice is protected by copyright. They are separately licensable.
To play the devil's advocate: Is it though?
Especially recently we've seen the widespread acceptance of copyright and license laundering through large machine learning systems such as copilot and chatgpt, backed by enough microsoft laywers to prevent meaningful push back from the copyright owners. They'll output verbatim copies of what was previously copyrighted code.
If it is true that software copyright can be cleaned away in such a fashion, maybe patenting the algorithm itself actually is necessary to avoid big tech companies like microsoft from forcibly strip mining the ideas of independent creators.