Chinese company caught brazenly copying
Surprised_Pikachu.jpg
“They can’t just help themselves to all the data they can get their hands on because they feel entitled to it for training! Wait…”
we must stop these computers from copying the numbers in their memory banks!
They have pretty much lost already. The US will probably try to fight in some way, but they have a very little moat. Even if they actually “stole”, it’s not as if that Anthropic had the moral high ground here.
Some chinese models like GLM 5.2, Kimi K2.7 Mimo V2.5, Deepseek V4 Pro and Minimax 3 cost close to nothing and have wild usage limits if you subscribe.
You can also run these Chinese models on European infrastructure through Cortecs.ai or similar. And you have actually have privacy! Privacy + Cheap vs Expensive and slightly better.
Opus 4.6 is almost 5€ in and 24€ out. Sonnet 4.6 is 3€ in and 15€ out. Minimax 3 is 0,3€ in and 1,7€ out.
All frontier models are so good that the differenciator going forward is going to be price. The models keep improving, but there is a clear trend of diminishing returns.
Wow, this one hits close to home.
Pot. Kettle. Black.
Depending on where you live it’s a good thing or a very bad thing.
It’s just two parasites slinging mud at each other.
“They’ve stolen our rightfully stolen data!”, said the spokesman.
Aww, did the serial copyright violator get copyright violated?
No, machine generated output is not eligible for copyright protection. https://en.wikipedia.org/wiki/Threshold_of_originality
Yeah, it wouldn’t be copyright. It might be trade secrets, though. And trade secrets can be made out of public data, but arranged in a way that gives competitive advantage (for example, customer lists themselves might be trade secrets, even if each entry is a publicly available set of name/contact information/job title/company).
If a company voluntarily discloses a trade secret to a member of the public, it ceases to be a trade secret, so I doubt that would apply here either.
Sharing trade secrets under the terms of a contract that dictates how one can use the information still retains trade secret protections.
Without a contract: intentional disclosure to the person who receives it generally destroys the trade secret status of the information, because the “owner” of the information didn’t do a good job trying to protect it.
With a contract: intentional disclosure to a person under the terms of the contract makes the contract’s own protections of the information relevant, and misuse of the information by the recipient can get them sued under the contract. Plus, the information itself probably retains trade secret protection so that even if that person gives the information to a third party who can’t be sued under a contract they never agreed to, there are still rights to protect that trade secret as property.
I’d be shocked if any paid API use isn’t under a robust, enforceable contract. The only question is whether the contract language itself effectively prohibits distillation.
Depends on the agreement. Contracts (like EULAs) can cover a fair bit
EULAs aren’t legally binding in sane countries.
Can you name a country where signing up for a paid account to an online service, and using the service and paying the invoice that comes in, doesn’t form a legally binding contract between the customer and the vendor?
Most countries in the EU don’t allow for consumer rights to be overridden by an EULA.
Similarly, I can’t have a contract with you to murder me. It’s illegal and me being a willing participant in it does not make it legal, even if I sign a contract.
An EULA is legally binding, but only the parts that aren’t in conflict with consumer rights, meaning most of any EULA is going to be invalid.
I wonder if US courts would agree.
Plagiarism machine plagiarizes plagiarism machine. Film at 11.
Anthropic, I mean this with the upmost sincerity:
No one gives a fuck.
If your competitor can put out a model that functions really similarly to yours for $2 less per month, and your entire userbase can just leave and move to them… explain to me why investors would want to pump hundreds of billions into your business to be ‘first to market’? That’s a really dumb thing to admit for Anthropic.
Who is ‘first to 100 million users’ is utterly irrelevant under a business model where your sole value is Intellectual Property (IP) and that IP can be “illicitly extracted” by a clever competitor without ever hacking into your nextwork or doing anything explicitly illegal.
I’ve had to explain this to a lot of people who seem to think Anthropic/OpenAI are incredibly valuable companies because “they’ll make money long-term so long as they keep being pumped full of it investment cash to be the first to earn a big userbase”, but that just doesn’t make sense. OpenAI owns no datacenters…zero. Theyre 100% IP. Anthropic “is building” some datacenters, but they exist on paper only so far, so they’re also presently 100% IP.
Can this obvious scam just collapse already so I can upgrade my PC without a personal loan?
I think your take is completely reasonable but I think the ‘first to 100 million users’ is actually noteworthy because if they can become entrenched and people become unwilling to learn anything else, they’ve won and can charge nearly whatever the fuck they want (at least in the medium term). See Microsoft and Adobe. They charge whatever they want for their subscription programs because what else are you going to do, use GIMP? Even in situations where the FLOSS alternative is legitimately good, a lot of people will still refuse to switch. I don’t think Anthropic can survive long enough for them to become the only thing Susan from HR knows or is willing to use, but I think there’s a path to profit somewhere here.
I’d argue that agentic AI by nature makes transition to a different model easy.
Yeah this is a key realization that I suspect most investors aren’t privy to. With proven viable local, accessible, scalable, and energy-efficient 2TB infiniband clusters and routed multi-agentic stacks of open source models constantly nipping at their heals, achieving longterm market dominance for any of these AI developers is simply a tenuous prospect.
The only legitimate option is to maintain a meaningful lead at the cutting edge of performance and/or offer a superior efficiency/value proposition via SLA guarantees. Beyond that, the brute force options are limited to things like short-term market manipulation (such as outbidding everyone else for existing talent pool, chip manufacturing capacity, etc) or suppression of competition via regulatory capture.
In every case, above or below board, there is no permanent longterm global breakaway strategy, only treading water as long as investors are willing to inject enough funds to temporarily outrun market efficiency.
Once that reality sinks in… pop.
There’s nothing to “learn”. Using one of these is in no way different than using the other.
Unless you start using fancy little features that let you do things the others don’t do quite as well.
No, because any cool feature will be immediatelly replicated everywhere.
This isn’t a real product. Just bullshit generator.
That’s not exactly true. The implementation details around context management matter to the user a use case. It’s totally feasible for providers to go into different directions, especially if they’re hoping to target different subsections of the same market.
See Microsoft and Adobe.
Except Microsoft and Adobe never bankrupted a company by getting adopted. It was a tax that companies could afford since they were still rounding errors compared to labor.
If the adoption of a tech can be measured as being roughly equal to higher than the labor expense of a company, that decision isn’t going to be dictated by what Susan in HR knows.
Lol. Stupid thieving fucks whine that their stolen data gets copied?
Alibaba picking up Anthropic’s fair use strategy?
Edit: is there an argument for letting the US ruin its economy and environment to train all these models and then just swooping in before it turns into a mild madmaxian hellscape to distill and/or extract the knowledge? Beats having to do this on your own, doesn’t it?
Yeah. Any Co2/other climate change regressions that the US makes affect everyone globally, and while water use is local, its also as-needed, so post-collapse you have to use up all your water anyway.
AI could use solely graywater/non-water cooling and renewable energies, and that’s the answer, just takes slowing down, building specific and rigorous facilities, . Letting the US speed along just hurts everyone due to climate change.
That and every major company economically depends on each other, and disconnecting from the US in a way that doesn’t cause backlash also takes time.
Fuck america but don’t let them drill holes in the boat we’re all riding.
Laws for thee
I’m still on the side of treating AI development with more caution than less. So depending where you live this could be a very good thing or a very bad thing in the long haul.
You know what they say, no honor among thieves.
