Jack Dorsey, co-founder of Twitter (now X) and Square (now Block), sparked a weekend’s worth of debate around intellectual property, patents, and copyright, with a characteristically terse post declaring, “delete all IP law.”
X’s current owner Elon Musk quickly replied, “I agree.”
So delete all pharmaceutical IP to make drugs accessible to everyone and save taxpayers trillions?
“Noooo, not like that!”
This is why it’s a mixed bag for me. IP law is kinda important in a capitalist system, which, for better or worse, that’s what we have. If someone comes up with a wonder drug that outright cures addiction or something, you’d want that person to be able to recoup their costs before a bigger organization with more capital swoops in and undercuts them on production costs until they’re the sole supplier of the drug. The hepatitis C cure drug selling for $70,000 is a great example of this quandary; there’s millions of dollars worth of research and clinical trials that went into developing the drug, you’d want the company to be able to recuperate the costs of developing it or else there’s less incentive to do something similar for other diseases down the line. Also, though, $70,000 or go fucking die is an outrageous statement.
Of course, what we have for IP law in practice is a bastardized monster, where corporations exploit the fuck out of it to have monopoly control over important products like insulins and life-saving medications that cost cents to produce and allow them to sell for hundreds a dose. That’s not the intent of IP law, IMO, and that doesn’t really serve anyone.
In the US the tax payer subsidizes almost all drug research. Between 2010 and 2019 the NIH spent $184 Billion on all but 2 drugs approved by the FDA.
It worked out to about $1.5 Billion for each R&D product with a novel target and about $600 mill for each R&D product with multiple targets.
https://pmc.ncbi.nlm.nih.gov/articles/PMC10148199/
Or
https://jamanetwork.com/journals/jama-health-forum/fullarticle/2804378
The cost to develop each drug is between about $1 and $2.5 Billion
I’m not sure how much is subsidized outside of NIH but I’d imagine other countries are doing the same.
Why should companies own the whole IP or perhaps why should they have any ownership if most of the funding is from the public?
Do it., but also ensure that all work enters the public domain and is free for anyone to use, modify, commercialize, or basically whatever the GPL says.
… Delete… all… IP law?
So… just literally make all piracy legal, switch all gaming and tv show and movie production/consumption… to an optional donation model?
Fuck it, why not.
I am both an avid pirate and have a degree in econ, wrote papers as an undergrad on how to potentially reform the DMCA… and uh yeah, at this point yeah no one has any fucking idea how any thing works, everyone is an idiot, sure fuck it, blow it all up, why not.
Yeah except you know it isn’t going to be that
They’re going to go “yeah but not like that”
They’ll just remove consumer protections and make it so you own even less and if you try to fight it, you’ll have the full weight of the court system to make you poor
Is musk supports it, that’s exactly what he’s hoping will happen. The rich will be able to take advantage of it and the poor will either stay the same or get worse
This, he means abolish IP Law in terms of consumer protections.
also jam in there protections for AI training so they don’t have to deal with those pesky rent-seeking “authors”
IP law does 3 things that are incredibly important… but have been basically irrelevant between roughly 1995-2023.
- Accurate attribution. Knowing who actually made a thing is super important for the continued development of ideas, as well as just granting some dignity to the inventor/author/creator.
- Faithful reproduction. Historically, bootleg copies of things would often be abridged to save costs or modified to suit the politics of the bootlegger, but would still be sold under the original title. It’s important to know what the canonical original content is, if you’re going to judge it fairly and respond to it.
- Preventing bootleggers from outcompeting original creators through scale.
Digital technology made these irrelevant for a while, because search engines could easily answer #1, digital copies are usually exact copies so #2 was not an issue, and digital distribution made #3 (scale) much more balanced.
But then came AI. And suddenly all 3 of these concerns are valid again. And we’ve got a population who just spent the past 30 years living in a world where IP law had zero upsides and massive downsides.
There’s no question that IP law is due for an overhaul. The question is: will we remember that it ever did anything useful, or will we exchange one regime of fatcats fucking over culture for another one?
How does genai make those concerns valid again?
Make yourself the question: how does genai respect these 3 boundaries set by IP law? All providers of Generative AI services should be forced by law to explicitly estate this.
I’m still not getting it. What does generative AI have to do with attribution? Like, at all.
I can train a model on a billion pictures from open, free sources that were specifically donated for that purpose and it’ll be able to generate realistic pictures of those things with infinite variation. Every time it generates an image it’s just using logic and RNG to come up with options.
Do we attribute the images to the RNG god or something? It doesn’t make sense that attribution come into play here.
I would like to take a crack at this. There is this recent trend going around with ghiblifying one’s picture. Its basically converting a picture into ghibli image. If you had trained it on free sources, this is not possible.
Internally an LLM works by having networks which activate based on certain signals. When you ask it a certain question. It creates a network of similar looking words and then gives it back to you. When u convert an image, you are doing something similar. You cannot form these networks and the threshold at which they activate without seeing copyrighted images from studio ghibli. There is no way in hell or heaven for that to happen.
OpenAI trained their models on pirated things just like meta did. So when an AI produces an image in style of something, it should attribute the person from which it actually took it. Thats not whats happening. Instead it just makes more money for the thief.
I think your understanding of generative AI is incorrect. It’s not just “logic and RNG” It is using training data (read as both copyrighted and uncopyrighted material) to come up with a model of “correctness” or “expectedness”. If you then give it a pattern, (read as question or prompt) it checks its “expectedness” model for whatever should come next. If you ask it “how many cups in a pint” it will check the most common thing it has seen after that exact string of words it in its training data: 2. If you ask for a picture of something “in the style of van gogh”, it will spit out something with thick paint and swirls, as those are the characteristics of the pictures in its training data that have been tagged with “Van Gogh”. These responses are not brand new, they are merely a representation of the training data that would most work as a response to your request. In this case, if any of the training data is copyrighted, then attribution must be given, or at the very least permission to use this data must be given by the current copyright holder.
I think your understanding of generative AI is incorrect. It’s not just “logic and RNG”…
If it runs on a computer, it’s literally “just logic and RNG”. It’s all transistors, memory, and an RNG.
The data used to train an AI model is copyrighted. It’s impossible for something to exist without copyright (in the past 100 years). Even public domain works had copyright at some point.
if any of the training data is copyrighted, then attribution must be given, or at the very least permission to use this data must be given by the current copyright holder.
This is not correct. Every artist ever has been trained with copyrighted works, yet they don’t have to recite every single picture they’ve seen or book they’ve ever read whenever they produce something.
Talking about “IP” as if it were a single thing confuses any debate. Copyright is not a patent, which is not a trademark - they do different things.
Software patents actually should be deleted. It is impractical to avoid accidentally infringing as there are multiple ways to describe the same system using totally different technical descriptions. Copyright for software was enough.
Capital finally taking the market out behind the barn.
IP law is used to reinforce capital tho??
Yes. It’s attacking its own institutions. Strange, isn’t it?
They don’t want to delete all IP law, they just want to delete the IP law which is preventing them from postponing the collapse of the AI hype a little bit more.
I’m fully in favour of abolishing IP law for everyone, ideally globally.
Public domain everything.
I suspect that isn’t the picture these two have in mind. It’s going to be the same as Musk’s demand for free speech, which just turns out to mean “let me be an asshole and you’re not allowed to complain.” This one is going to be “I get to profit off your ideas, but you’re not allowed to use mine.”
This would be disastrous for actual manufacturing because a patent is the only thing that makes it worthwhile to spend a bunch of money upfront to develop a new technology. Unlike with software where you don’t have nearly as much up front capital investment to develop something, it costs millions of dollars to get a manufacturing process up and running and in a good enough state to where it can actually work out financially. Without patents, your competitor can just take all of that work and investment and just copy it with the benefit of doing it right the first time, so they’re able to undercut you on cost. The alternative is that everyone is super secretive about what they’re doing and no knowledge is shared, which is even worse. Patents are an awesome solution to this problem because they are public documents that explain how technologies work, but the law allows a monopoly on that technology for a limited amount of time. I also feel that in the current landscape, copyright is probably also good (although I would prefer it to be more limited) because I don’t want people who are actually coming up with new ideas having to compete with thousands of AI slop copycats ruining the market.
TL;DR- patents are good if you’re actually building things, tech bros are morons who think everything is software.
Patent documents are rarely useful because they’re kept as general and opaque as possible to cover as many innovations as possible. I agree that it’s important to protect manufacturing, but patents are not the right way to go about it for at least two reasons: (1) they block innovation by design (e-ink screens are great examples) and (2) they create a huge barrier to entry for new ideas (think about how many lawyers are making a living on this) I disagree with the senders on so many things. But patents were invented in a world of monarchies and craftsmen. Time to go!
Patents would be fine if the bar for “innovation” would be much higher, software patents weren’t a thing, there was way more research done into prior art, and there would be different (shorter) lengths for patents depending on what industry they target.
Like, if it’s manufacturing or something like drugs where it takes years before you can start making profit, sure, make them 10-20 years. If it’ something you make money off of immediately, it should be shorter.
I actually agree that the patent system could be improved a lot. Not all things are bad about it.
What do you mean with “innovation”? How would that be defined?
Protecting innovative stuff is literally the point of patents and why the system exists. Anything “new” is by definition innovation, except the bar is really low currently, with very little research being done into prior art.
Patented stuff should be non-obvious, and not a simple derivative of existing stuff (i.e. when there are square buttons and circle buttons you shouldn’t be able to patent a button that has 2 corners square and 2 circle just because it’s “novel” because it’s just a very simple and logical step).
So basically, make the bar for a patent much higher, and require some proof into the research of prior art and explaining why/how your patent is different.
Also, patents should expire early/not be renewable if you don’t actually use them (so move a certain number of units / generate some amount of revenue using your patents). So you couldn’t patent random BS in the hopes someone else will break your patent by accident.
Or even better, just outright punish patent trolls.
Patent documents are rarely useful because they’re kept as general and opaque as possible to cover as many innovations as possible.
I think this is a problem that can be fixed inside of patent system. Make it so by the end of patent life there is “how to build production line of this” manual.
