It’s out of the box.
They can try, but, it’s too late.
Don’t hesitate, bring them to thier knees.
Is reversing with Deepseek just as good?
If not now, probably at some point it will.
No idea.
The legitimate developers I know aren’t hit by Claude restrictions, because they don’t fucking use it.
Linus Torvalds uses it. I’d consider him a legitimate developer.
“was legitimate”
Ftfy
Self host a QWEN model on a budget server build.
American AI is so centralized that utilizing AI, as an end user without having some loser looking over your shoulder or demanding a monthly fee, is difficult.
Nvh of that. I just looked it up. Can it be hosted on a lower spec machine? I downloaded GPT4All and couldn’t use it because my specs weren’t good enough.
I think the harder part I face is setting up qwen to be effective for my uses
Only the AI companies are allowed to plagiarize.
What exactly are you suggesting here? That using an LLM to reverse-engineer a video game, modify its code, and create an entirely new experience constitutes plagiarism?
Do you understand what plagiarism actually means? It means passing off someone else’s work as your own. It has absolutely nothing to do with whether someone uses AI to reverse-engineer software, modify an existing game, or develop an entirely new way to experience it.
There may be legitimate arguments to be made about copyright infringement and intellectual property rights, but those are entirely different discussions.
And what does the statement about AI companies being the only ones allowed to plagiarize even mean? Are you suggesting that modifying an existing video game is somehow equivalent to an AI company training a model on copyrighted material?
You’re conflating completely different concepts, and the result is an accusation that doesn’t even make sense in the context of what’s being discussed.
What I’ve presented here is the absurd concept that AI companies exist after having stolen a wealth of human intellectual property and haven’t paid for or attributed any value to the original creators. And now that they have this, they are gatekeeping how it gets used.
But I said it in a punchy one-liner with a sarcastic undertone, and I’m afraid that’s where I lost you.
Your reverse engineering argument might have a point but plagiarism regarding AI should be more obvious. Indeed AI is trained on the collective publicly available knowledge humans have deposited on the internet and then sold back to the same people for a fee without any acknowledgement let alone sharing of profits in someway. It is similar to the kind of unethical behaviour where some academic publishers are employing aggressive marketing strategies to sell public science back to public at high costs. Ofcourse there is still the argument that these AI companies are only charging for operational costs, which may be true for some. But I am willing to bet money that if LLMs becomes an integral part of the modern world then they will turn down this tap to a bare trickle at exorbitant prices. Sam Altman himself confessed as much (he said something along the lines of envisioning a future where AI becomes an on demand information and skill tap, you fill in the rest). Operating something at a loss is a known capitalistic strategy to get people used to your product and drive smaller competition out of market so in a decade or so you become the monopolistic owner of an essential service.
Chatgpt won’t give me bass sheet music to Ella Fitzgerald’s version of How High the Moon, due to concerns of edit: copyright issues (not plagiarism.).
I wouldn’t put it past anthropic to deny the request.
Not because of plagiarism. It’s not concerned about plagiarism. It’s concerned about copyright infringement. It is not creating the bass sheet for you and then claiming it as their own novel intellectual property that they just made up on their own.
Why do people not know what plagiarism means.
Why do people not know what plagiarism means.
you’re right I went back into my conversations and it was indeed because of copyright issues not plagiarism
Absolutely! Private corporations are going to milk these models for everything they can fortunately, there’s a lot of really good open source out there.
Open source developer here, AI is regularly being used to erase the original developers and artists.
You know what I liked in memes with “stolen” art? There was something to look up to, there was something to become part of our collective culture. Call me a hippie or “irrational” but this is why AI is like the second coming of asbestos, even if it becomes better over time.
Here’s my game engine, feel free to fork the shit out of it. Feel free to rewrite it in Rust to “make it better”. Feel free to borrow parts of it. My only problem if you do it with AI, so you neither learn during the process of doing it, and you manage to erase me in the process.
The whole reverse engineering craze is being used, so most game devs will be so disillusioned they will leave the industry, and AI vendors can sell expensive subscriptions instead. The AI bros are really angry, that their shiny new technology didn’t immediately disrupt the existing creative industry enough.
Well tell me where AI companies are giving credit
Once again, this comment doesn’t make any sense. The word plagiarize is being misused here AI companies and their product the llm aren’t committing plagiarism. So I’m not sure who you want them to credit. They have already disclosed where they’re getting their training data from do you want them to name every individually? And even if they did, what conceivable purpose would that serve? None of them are getting any royalties from it. Because using training data isn’t necessarily copyright infringement and is in fact still an ongoing legal matter in the courts. And even if they did, the amount of information that is used as training data is so massive that a credit section wouldn’t use all usability it can tens of thousands of individual names. You ever sit there and read the credits at the end of the movie?
“The internet” is not actually a valid source.
Yes, I indeed want people to be named individually. The internet is not one Organisation that people generate IP for.
Of course crediting everybody would be ridiculous, but stealing the entire internet is not?
Let’s assume, for a moment, that LLMs are trained on publicly available data. In other words, they’re not stealing or pirating that data, just to simplify the argument. (Although, to be fair, some LLM developers have done exactly that and are facing legal consequences.)
Publicly available data is just that: publicly available. Anyone can access it, including individuals, businesses, and AI developers. They’re not stealing something that is freely accessible to literally everyone.
And that’s the dilemma: either that information is freely available to everyone, or it isn’t. You can’t argue that it’s free for you and me to access but somehow off-limits to OpenAI or Grok simply because they’re using it for a different purpose. That’s not how public access works.
So, they’re not stealing anything under the assumption we’re making here. To answer your original question, no, I don’t have an issue with them stealing the internet.
The thing about ownership is that unless something else is specified, everything is ARR by default.
Adding insult to injury, they don’t care about licenses…
You may be allowed to consume but not to sell that content
This almost feels like some sort of oxymoron.
Big tech is supposed to reverse engineer individual devs and artists not the other way around, stahp!!
Hilarious
“Give us all your code for free who cares about copyright anyway”
“NOOOO YOU CANT USE OUR PRODUCT TO GET FRE CODE 😡😡😡”
AI exists to aid capital against you; not aid you against capital.
AI exists to devalue human labour.
I can only talk about software developers, but there it’s fucking pants on head to fire developers. It only makes sense if you don’t have any competitors.
More than 50% of software projects fail either because they run over time or over budgetl, but for obvious reason both. So you get this tool that increase your output, let’s say 10x. Does that mean that you can fire 9/10 developers? What does your competitors do? If they retain their developers they can smoke your company in months.
This assumes 10x the features at 2/3 the quality is a winning bet
The quality doesn’t matter. Look at a product like netsuite which dominates corporate accounting. It’s a total piece of shit. And overpriced one at that. But it has oracle in the name, so an executive can choose it with little thought and not take any blame because it was a “reasonable” choice that “everyone” uses. Much of corporate life is monkey see monkey do. Rarely does innovation happen because trying something different is a huge risk to some middle manager or execute that is just trying to make it to the next promotion or bonus.
Maybe your are right, software that checks all the boxes wins b2b deals even if it is a peice of shit
The reason a lot of these enterprises are looking to offload developers is because they see AI as accelerating their workforce. These businesses are doing the same thing as they were before, but with AI, which means they don’t have enough workload to justify having that many people on anymore (as long as they are using AI workloads). If one developer with AI can do the work of 3 to 10 developers without AI the logic is that they can have their developer headcount and still be ahead.
Not that I agree with what they’re doing, but this is the logic behind it.
I suspect we’ll see these features opened up for US-allied actors against their foreign competitors.
You can guarantee US auto companies are going to be free to reverse engineer software and manufacturing techniques of Chinese companies, for instance.
But you’re not going to be allowed to use Grok or Meta to hack your Tesla or BMW, pull out the OS, clone it as freeware, and offer a substitute unconstrained by subscription fees and gated upgrades.
Why just games? What about the Photoshop clone? Seems like making a derivative of anything should be blocked, but that’s all it can really do in the first place.
Seems like making a derivative of anything should be blocked
So we should get rid of anything in-the-likeness-of? Like Libre Office Writer being in the likeness of Microsoft Word.
Microsoft word was not the first.
I think their point is that it was trained on literally billions of repos and other software written, but because those developers don’t have hundreds of lawyers on retainer with an unlimited budget; their intellectual property rights don’t seem to be a concern to the slopmasters.
I think that the problem is that big companies have made “an offer they cannot refuse” regarding the problem of AI cloning commercial products.
just a different way of saying the same thing. the slopmasters have zero issue stealing IP when the IP holder can’t defend themselves.
Major software companies and publishers are realising AI companies put their profit motive at risk.
Why buy the latest Mario game when you can have a version custom built for you? Why get an Adobe subscription when the cloned version is available?
That can be applied to any software now. Two industries are facing off.
This is why AI generated content should belong to EVERYONE. There should be no profit in it, because the content used belongs to humanity.
AI should belong in the history book of bad technologies, alongside with leaded fuel and asbestos.
I once asked Claude to check the app it vibecoded for security flaws and it didn’t want to do it.
It literally told me to look up some basic security principles and do it myself.
It literally told me to look up some basic security principles and do it myself.
I think that seems like excellent advice!
you want vibe coded apps with basic diy security principles?
Yeah their anti-hacking protections are over eager. Ask it to find bugs instead, or reasons we shouldnt deploy yet.
I hook mine up to some classical static analysys tools like sonar or codacy. Just ask it to remove the security findings and haven’t had an issue. Then again, been on vacation while all this shit has happened, so don’t know what to expect when I return.
The restrictions are so they can gradually institute tiered access fees. This is how the industry becomes profitable.
Tier based models = economic segregation. The poor can fuck off and the rich can pay for enterprise.
Only works if they get their competitors “regulated” so other models can’t offer unrestricted features for a lower price.
Only works if local models cannot do it.
The best part of paying for a coding assistant is discovering halfway through your project that it has developed opinions about what your project is allowed to be. Imagine your IDE refusing to compile because it thinks your game feels a little too much like Tetris. At least a compiler doesn’t pretend it’s doing you a favor.
I think that’s what they wanted from the very beginning, to accelerate projects that are socially convenient for them, but not those that are not.
There’s that issue with universal machines that some people don’t like that others are allowed to make universal decisions.
I got shut down by it two days ago. It suddenly told me that the project I was working on had tripped some kind of guard, that it couldn’t continue working with me on it, and it couldn’t tell me why. I just went back to my local Qwen model. I can’t see this doing great things for selling Claude subscriptions.
I wonder if you can trip this on corporate subscriptions just by examining standard code so I could finally have an excuse to not use it at work.
As it was before AI. Local is the answer when shared is at the mercy of interests antagonistic to the users.
That’s what happens when you let others do your computation for you.
Run the tool on your own hardware.
My coworker recently had an experience where he was trying to fix a security hole and his AI wouldn’t do it because it claimed that it was unsafe.
Great time to know how to do things yourself.
Copyright violation from American AI hoovering up all the data is okie dokie unless it’s used by peasants to threaten technology corporations.
I’m not defending them it’s just comical hypocrisy, and ans I doubt China AI will have the same kneecapped coding capabilities.
Chinese models are kneecapping Tiananmen Square and Winnie the Pooh, they’ll happily do anything else.
But at least they are open (Most of them at least).
So there’s at least the possibility to lever these restrictions away.
While I agree that open-weight is better than nothing, it is not the same as open-source.
Having a truly open-source model would be wonderful and it would also allow us to verify what data it is trained on and whether it has the right to use that data.The Open Source Initiative’s definition of a fully open-source AI model includes the training data as a key element. Some developers wonder, however, how feasible it is to distribute the enormous datasets required.
He says Stanford HAI has staked out a specific bar: alignment with the Linux Foundation’s Model Openness Framework at its top tier, the “Open Science” class – meaning the code, the training data (or a thoroughly documented, auditable account of it), and the tooling are all released, not just the weights, alongside a real way for outside researchers and communities to download, run, study, contribute to, and modify the work. Anything short of that, in his view, forecloses entire lines of research: You can’t, for example, study how a model’s training data shapes its cultural assumptions, audit it for how fairly it represents different populations, or test new architectures at the scale where genuinely new behavior shows up. You can only observe a finished system from the outside and guess at why it does what it does.
Right I agree, but this is basically the entire internet (and more, like scanned books), what it is trained on. They are using reinforcement learning now. And this is also one of the reasons why they have escaped the sandboxes in the American models for example.
So you can probably safely assume that they are trained on all data that is available easily. and very likely also artificially generated data at this point. How long they are trained on, what kind of data is a different story. I think it gets increasingly more interesting how these models are trained.
So while I agree it would be really nice to know exactly what it is trained on and how it is trained on that, I think that those that have the resources to get at least close to frontier models (i.e. AI companies) don’t just open up on that, sadly.
Kimi K3.1 pro, had reverse engineering information pulled from its training data. So at least some of them are.
This is another reason why I’ve reduced my dependency on AI considerably over the past few weeks, and I no longer personally pay for any AI tools/services at all.
It’s insane to think someone would pay for something and after using it for a couple days they get told they can’t use it for certain things anymore.
They enshittified too soon, before people depended on them.
Oh, based on my experience, there are plenty of people convinced they can’t do anything without Claude anymore. Also based on my experience, I would be ecstatic if they actually stopped trying to do work because they sucked both with and without Claude.
Maybe not enough to fund their financials, but that was going to be a tall order anyway.
Sure, but they can’t just let us plebs threaten their dominance in other areas.