Plagiarism as a Service.
We can all try really hard to pretend that's not the business model, but that's totally the business model.
We can all try really hard to pretend that's not the business model, but that's totally the business model.
1. The sheer amount of material on the internet that is "free to view but not free to use for any purpose" is the greatest resource of our time, and despite it being easy for individuals to take advantage of it (no one will take you to court for printing a newspaper comic and pinning it to your corkboard,) it's historically been difficult for corporations to exploit it (their best idea pre-AI is to encourage people to post it on social media walled-gardens where they can surround it with ads.)
2. The reason behind the impressive results of generative AI is because it exploits the above "free" resource, which is the greatest resource of our time. The reason behind the industry-wide push for AI and the insane amount of investment in it, is that they know it's their first real chance to exploit the greatest resource of our time. This is the gold rush.
3. Anthropomorphism is the wool that AI labs are pulling over legislators eyes so they can pull off this heist. If you see training and inference as a black box, a process that consumes a copyrighted work (among others) and produces something very similar to the original work that also competes directly with it, is clearly something that's against the spirit of copyright. But if you (afraid of being judged a luddite) see AI as a little man inside the computer who is "learning" and "creating," how could you deny him? Especially if it would deny your jurisdiction access to the above gold rush. A lot of scientific-sounding AI communication is propaganda for this way of thinking, like the Anthropic J-space stuff, which stops just short of claiming AI is conscious, despite leading the reader to that conclusion.
AI: scans all art, people can ask it to produce art they're searching for, no reference to the original or its author, people no longer pay artists/designers/...
I think there's a very obvious distinction there. The whole purpose of copyright is ensuring the financial viability of creating works. AI slop is a direct market substitute for the originals it ate up for training, so it goes directly against that. Meanwhile, search actually improves the reach of works and, well, snippets and summaries are somewhere in between...
In the US it is "To promote the Progress of Science and useful Arts".
Of course, it was also supposed to be a limited-time monopoly rather than perpetual...
To me that seems a exceedingly broad definition of fair use.
Perhaps we need to create a new form of intellectual property to protect innovations in style. But that’s not copyright, which protects existing works from unauthorized reproduction. It seems like it would be very difficult to adjudicate though; how do you assess whether a style is truly unique to an author or artist?
This is exactly the sort of anthropomorphism I'm talking about.
Yes, human artists reproduce styles all the time, but implying this equivalence between what a human does when they recreate a style, and what an AI model has done when its output is suspiciously similar to a single artist's "contributions" to the training set is reductive since the mechanism is obviously completely different. It doesn't say anything about whether one ought to be allowed, just because the other is, but the assumption that it should is doing the legwork for those who benefit from AI as a copyright washing machine.
My point is not that the process is the same. That’s not relevant to the law. What’s relevant is whether or not the material is a reproduction that substantially infringes upon the original.
I’m also not making an ought argument. If you want to change the law, write your congressman. What I’m arguing is that existing legal doctrine on copyright law very clearly does not protect mere ideas or styles.
> What I’m arguing is that existing legal doctrine on copyright law very clearly does not protect mere ideas or styles.
Like it or not, it is anthropomorphizing to claim that an algorithm can have ideas or use styles, rather than just saying that the apparent "ideas" and "styles" in its output are a mathematical derivation of the human ideas and styles in the copyrighted material in its training set.
1. https://storage.courtlistener.com/recap/gov.uscourts.nysd.64...
I continue to find that people strongly advocate for justice on exactly opposing sides, depending on who they have been told to think the “bad guy” is.
The same orgs that harassed and antagonized Aaron Schwartz are not going after AI for the same thing at a much much larger scale.
What's different? The size of their bank accounts.
One, time went on. The prosecution of Schwartz was seen as an overreach, as demonstrated by Ortiz’s failed political career thereafter.
Two, Schwartz’s charges were only ever charged. No jury or judge signed off on them.
Three, context changed. Schwartz wasn’t 5% of GDP. For better or for worse, that matters to voters.
But it's not. The only layer anyone has a handle on is consumers and companies paying AI ridiculous sums for AI. Some of those users are probably justifying it with labour replacement. But a lot may not be. So far, we haven't seen the employment effect outside recent college graduates at enterprise companies.
> as implemented in the United States it's a malignant tumor, a theft of labor by capital, and it must be (minimally) adddressed with confiscatory taxes applied to all those involved with it's creation and operation
It's also a godsend of economic growth. Growth other countries who are trying to balance their books would kill for. Without AI, we'd be in a failure state. Maybe we are, if this is all a bubble. But as it stands, there is paper wealth that can and has–limitedly–been taxed. That gives everyone options.
> If a single AI billionaire exists in the year 2030, then the US is a failed state
This is silly and projecting a narrow view of the world onto a larger voting population. Voters don't care so much that there are billionaires as that living standards haven't kept up with the rate at which they're being minted. Double tax brackets, add more on top, raise the minimum wage, raise Social Security taxes and benefits, expand Medicare, beef up antitrust, establish a progressive property tax on wealth that starts at 1,000x the median American's wage (about $65mm) and billionaires are fine.
VCs are paying ridiculous sums for AI. Consumers are not - we get tokens subsidised by VCs.
> if this is all a bubble
Interesting to see that revenue growth for both Anthropic and OpenAI has levelled off recently. Once this fact percolates through to the VCs, it's going to cause problems. All those valuations are based on projections of vastly greater revenue than they're getting now (and, ofc, achieving AGI and "winning" everything immediately that happens). This is looking less and less likely - the current batch of AIs are very, very, useful tools, but as we learn how to use them commercially they're not generating those limitless revenues that were anticipated.
This tech, like all the rest, will go through the Gartner Hype Cycle, and that includes the Trough of Despair where it all looks shit and the bubble pops. I think we're approaching that rapidly.
Right now it's replacing "recent college graduates", which makes sense, because they're the least differentiated white collar workers. However it's advancing quickly, and it will (quite obviously) eat more and more white collar jobs.
> It's also a godsend of economic growth.
Fuck no. There is unprecedented Capex that is propping up the economy, as companies are rushing to create capacity that they intend to use to destroy jobs. So yes, those datacenter buildouts and chip purchases and power infrastructure buildouts are creating economic activity... all with the hope that someday companies will be able to fire a huge percentage of workers.
You wrote it as though the gains were from productivity, which they really aren't... but even if they were (and that is likely to happen at some point), it would only be beneficial to actual people if that also results in greater distributions to labor. But a technology like this disfavors labor, so we can anticipate aggregation toward capital, and a slower velocity of money overall.
> Voters don't care so much that there are billionaires as that living standards haven't kept up with the rate at which they're being minted.
I suspect that voters will, at some point, care quite deeply that a lot of awful people got INSANELY rich by destroying tens of millions of lives, and making the K shaped economy have a much smaller segment of winners
The policies you suggest could help... but we have an administration (all three branches) that oppose anything of the sort. They're much more likely to simply try to deploy AI to further oppress the people whose careers they destroyed, than to try to create soft-landings or fair outcomes.
The party most supportive of mass AI deployment is a party that despises the poor, and would happily simply lock them up. This... is not a good mix.
I hope that your optimism turns out to be warranted, but I think you're a fucking idiot, defending reckless bullshit that is executed as part of the biggest heist in modern history.
Where was I optimistic?
The technology is here. If Anthropic and OpenAI or DeepSeek develop it, it’s still going to do its labor replacement to the degree it will. The difference is whether it occurs under our tax base or not.
I’m currently seeing America squander this boon. But it’s objectively better to be in a world being transformed by AI when the AI is in your jurisdiction than sitting on the sidelines while suffering all of the same ill effects.
You can’t create a soft landing if you haven’t the funds for feathers. We do. We’re just blowing it.
Sorry, what? You're claiming that which countries exactly are failed states because of a lack of AI? And that the US would have failed, what, in the past four years if ChatGPT hadn't released or something? That's an absolutely insane thing to claim, but I have no clue how to else to interpret what you're saying.
No. I'm saying the difference between the United States and other profligate countries is that our economy is, even if just on paper, growing at the pace of a low-tier developing economy.
That gives us living-standard and tax-base margin that others don't have. It means we have policy wiggle room. That doesn't guarantee we'll use it wisely. (We are not, currently.)
> that the US would have failed, what, in the past four years if ChatGPT hadn't released or something? That's an absolutely insane thing to claim
I should have said state of failure. Not failure state was unnecessarily ambiguous. Sorry.
If our economy were growing at 1% or nil, and we were busy prosecuting a war in Iran and tariffing everyone, I think we'd likely tip into a recession and a political crisis more intense than the one we have. AI's largesse is buying us roadway. That roadway, in turn, may take us through to the midterms.
Ignoring the obvious “citation needed” and taking this as accurate, wiping out entry level jobs at massive employers is a serious problem with longterm effects we likely only understand a part of at best.
We gleefully moved manufacturing overseas for decades and finally realized the extent of the cons after it was too late. We clearly needed a more balanced approach. Something tells me we’re setting ourselves up for the same mistake.
I don’t believe for a second Schwartz wouldn’t have been charged if his parents were rich. He would have been better equipped to fight it. But that’s it.
Yes, our political economy is more corrupt today. But it’s silly to project the Schwartz example onto AI companies given the former is seen as a mistake and the latter are orders of magnitude more potentially valuable. And yes, if you can swing a state’s tax coffers meaningfully that’s going to influence voters and thus prosecutors.
Schwartz wouldn’t be prosecuted today. In part because Schwartz’s prosecution tanked Ortiz’s career. There isn’t a hypocrisy between these two examples.
(That doesn’t mean it isn’t a good bit of political sloganeering. It will rally some folks and I’d suggest someone in a D primary use it in the right circumstances. But it isn’t technically true.)
Compute, however, clearly is not.
Here are a couple articles about the charges and the case against Swartz [1][2]. Compare the elements of the various charges to what the AI companies are doing and there isn't really a good match.
[1] https://volokh.com/2013/01/14/aaron-swartz-charges/
[2] https://volokh.com/2013/01/16/the-criminal-charges-against-a...
While another difference is that AI companies violate copyright in a less obvious way, it's different for them, they do want to make money from the copyrighted works they train on.
That they usually don't strive for exact reproductions of works is another difference.
Their wealth has exceeded an escape velocity beyond which they won't be put in prison (or if they are they would quickly be pay-for-play pardoned) unless they are seen as a threat to even wealthier people.
See, for example: Devon Archer, Jason Galanis, Benjamin Delo, Arthur Hayes, Samuel Reed, Trevor Milton, Carlos Watson, Paul Walczak, Todd and Julie Chrisley, Lawrence Duran, Marian Morgan, Imaad Zuberi, Changpeng Zhao (CZ), Joseph Schwartz, et al.
Some of these people are broke bitches compared to the group of people you're talking about now, and yet still hit the threshold of being above the law as long as they play the corruption game.
I mean... sure... we're all going to die eventually.
But in the meantime it would be nice if the rule of law existed regardless of wealth, but that isn't the world we live in.
If anyone can just prompt all their basic “information needs” however how sloppy, then what remains of the economy? Health care, child care, handyman?
Most people won’t even pay for ad free YouTube. I don’t think any software business can survive AI as a substitute good even if it’s inferior (and it might not be).
Have people ever really broadly cared about the IT professionals behind their devices?
You’re making a lot of things objectively better, but none of them are the essentials that people need.
I’m not saying it’s bad to make AI bots or video games or network apps. I’m saying that the blanket statement “we make things better” is oblivious to a lot of realities.
somehow I still keep having work to do!
That kinda describes a lot of “not the USA” developed countries since quite some time.
It's also forgery as a service.
The problem is that copyrighted material is intermingled with non-copyrighted material in a way where it's not obvious how to solve it. But I think in this case, AI is so powerful that we should (gasp) cut it some slack. This would be the perfect example of throwing the baby out with the bathwater if OpenAI were to be sued into oblivion.
Layering and placement of copyrighted materials to evade copyright.
AI is here. We've had a good long look at what it does and what it's used for, and it's not going to be something else. This is what it's for. It's for copyright washing other people's work (shitily). It's for astroturfing social media with product placement comments. It's for presidents to make videos of themselves dropping poop on protestors from airplanes. It's for souless "creators" to earn updoots from other bots for their "street photography" generated images of neon lights on puddles in Tokyo. It's rube goldberg automations that almost always accomplish nothing. It's fanatics claiming that it has multiplied their productivity by some incredible factor, but never showing the receipts, or when they do it's always something trivial like a calorie counter app.
This is it. This is the AI we've heard so much about. I'd say I can't wait for the hype to end, but after living through several cycles, I'm almost certain that whatever hype cycle emerges from the IT sector next will be even worse.
I am curious, your type seems to become a rare species - I suppose you have never worked with a modern coding agent recently?
Otherwise you have noticed also here many, if not the majority expressing they rarely if ever write code anymore?
If it is a hype, then one that produces lots of working code. Way faster than I can type it. And I am fast.
I covered that:
> It's fanatics claiming that it has multiplied their productivity by some incredible factor, but never showing the receipts, or when they do it's always something trivial like a calorie counter app.
There is no shortage of "working code" out there. GitHub is reportedly falling over from all the working code. Where's your business? Who's using it? Why aren't products better yet? It's a mirage. Your "working code" is the same thing as gen-AI "street photography" images. No one wants to see them, yet it's pumped out in vast enough amounts that it's choking the spaces that are ostensible for photography. The low quality, low investment nature of it excites the lazy wanna-bes who leave a trail of half-baked, soon-forgotten detritus behind them.
Show us the receipts.
And I mean no wonder github is drowning in garbage, now that anyone can "code"
"The low quality, low investment nature of it excites the lazy wanna-bes who leave a trail of half-baked, soon-forgotten detritus behind them."
But would you also call Linus Torvalds a lazy wannabe?
If there are any face-melting characteristics to this whole thing it's how much money has been poured into it.
Ok, so as an example, I have a old 2D game of mine. Custom written voxel engine for a shooter game with completely destroyable world. Works great, but 2D.
I always dreamed of making it 3D, but never found the time. It would have been a huuuge effort doing it on my own.
Some days ago I pointed fable at the engine and basically said, make it 3D.
And it did. Took some hours. Not by copying other voxel engines, but by making my engine 3D. (I know because I know there is no game on the market like this)
So yes, I knew the domain and thought long and hard about it and the core is handwritten and battle tested and that probably helped steering it in the right direction. But I did not wrote a single line of this new code. And it works awesome.
This is not what I call marginally improvement. But you do you.