While irony is to be treasured and to hear whining from people who foist the most ridiculous terms of service on humanity is ironic, the more Germaine question is, Where was the cyber security? You have an AI company worth tens of billions of dollars with Microsoft as a major shareholder who has world class cyber security and these people just drove up to the back of the store and stole the inventory?? That is nuts 🌰
OpenAI can just start generating its responses by calling deepseek in the background. Save a few bucks along the way, and the content is all the same I guess (i.e. somebody else’s)
The issues of copyright and privacy infringement remain unresolved.
Beyond the ongoing reckless acquisition of training data—including personal information, web scraping, and harvesting content from social networks—there is also a persistent lack of attribution to original authors. AI models do not disclose their data sources, let alone credit the creators or specify how the data is being used.
A recent example is AI agents scraping website data without consent, as if it were freely available for the taking. This not only adds unnecessary traffic to websites, increasing server load and reducing service capacity for regular users, but also collects data for undisclosed purposes, including personal user information. As a result, some websites are actively blocking AI agents to protect their content and users.
This practice is akin to a fleet of heavy-duty lorries clogging up a local shopping centre, disrupting parking and access for legitimate customers—all without warning or permission.
Given this, it is both hypocritical and embarrassing to see OpenAI criticising DeepSeek for failing to play by the rules when OpenAI itself has repeatedly broken them.
The rules are simple: Act respectfully, or face the consequences. For AI agents, this could mean being blocked from all websites.
In general, always consider the impact of your actions. For AI agents and web scraping, it is best to contact the site owner first—otherwise, expect to be blocked.
Your continued behaviour, however, speaks for itself. This is why OpenAI is rated very low when it comes to shady practices.
I'm going to assume you've never done network administration or network security work. Blocking is hard, or at least blocking bad actors without also blocking legitimate traffic you want. Say you identify an origin that's an AI spider, somehow, because it's not going to advertise itself as such, and block the IP address. The spiders of a multi-billion dollar company is going to have a range of IPs, and can make use of VPN services just like the rest of us.
The 'no spiders' setting in your HTTP header only works if the owner of the spider agrees to it. If OpenAI is disregarding that it's an awful move, but nothing anyone can stop.
And every single owner of every single website would have to spend dozens of hours of a mid to high level admins time every year blocking the things.
The only way to stop it would be by law and international treaty. Specifically BRIC, EU and US would all have to agree to snub billionaires in order to protect the IP of bloggers, podcasters, and legacy media. Good luck with that.
Maybe the case law could be established via a suit from someone wealthy like music/film companies. You just have to rich enough to win the case, not richer than the tech companies.
I agree, and I think that's what will ultimately destroy OpenAI and all the rest of the LLM BS. The executive boards and c-suites of WB, Disney and the big book publishers are all in bed with the VCs and likely turning a blind eye to the mass IP theft for the sake of their own stock portfolios.
But they just got destroyed in the markets, and these are not nice people willing to overlook a loss of tens or hundreds of millions of dollars. They don't need that money, but it's how they keep score.
Now we might see those organizations get a lot more interested in protecting their IP if Chinese chatbot apps are pretending to be generic cartoon mouse or princess using text from Disney IP.
There is a big difference. OpenAI plundered intellectual property and then sells it to you with a subscription. Deepseek (allegedly) plundered this property from OpenAI and is giving it away.
No. That's not correct. You can take the DeepSeek model and deploy it on your own hardware. This is the primary contrast with OpenAI. OpenAI runs the model for you, and collects your prompts and your subscription money. DeepSeek just uploaded a model that you can download and never have anything to do with them again ...
I don't know, that's a tough one. You sure "open"ai and facebook and ilks are not taking your data thru some backdoors? who's to say who is more benevolent?
Well I'm pretty sure that the Chinese Communist Party is less benevolent towards the interests of the average American than Sam and Zuck. But I do expect the CCP to win in the end. Hope you're cool with that outcome.
And as far as what I am cool with? I am cool with the whole world working together to solve big problems facing all of mankind, not differentiating between us and them, not declaring either you are with us or you are against us. How about that for a change?
"Chinese Communist Party is less benevolent towards the interests of the average American than Sam and Zuck" You have evidence for that?
Took me 10 seconds to find the below comment on this YouTube clip https://www.youtube.com/watch?v=0PxbXnIlRno . There are comments like this all over the Youtube space and elsewhere. Sometimes one really just have to look into the mirror and see where the problem is.
The courts have yet to rule on the training copyright issue, but quite independent of how they rule, it is the case that one has no more legal right to give away stolen property than one has to sell it.
If this were not the case, thieves could simply gift stolen property to whomever they please and get away Scott free.
Furthermore, those who accept/use stolen property are not blameless according to the law either. And the latter simply encourage the former.
With all the problems of "hallucinations" and unreliability, what is the benefit of an LLM such as ChatGPT? If you think deeply, you will find that ChatGPT is Altman's invention of a systemic way to steal without being caught.
Gary, I've got say as I read the clipping I was getting increasingly excited to get to the end and see your response, and I'm not happy. You're completely on point as usual, but bullet points 2 and then 3 made me suddenly laugh out so loud that I scared the sleeping cat off my lap in a flurry of panic and claws. She's sulking out with me, won't come back, and now my legs are scratched to all buggery.
OMG 😱. The Chinese would violate terms of service and steal data from someone who stole data?????
that IS what this world has come to
While irony is to be treasured and to hear whining from people who foist the most ridiculous terms of service on humanity is ironic, the more Germaine question is, Where was the cyber security? You have an AI company worth tens of billions of dollars with Microsoft as a major shareholder who has world class cyber security and these people just drove up to the back of the store and stole the inventory?? That is nuts 🌰
That is exactly the same question I have
Actually laughable his outsize hubris he can't recognise the irony of claiming copyright infringement
Technically it’s not copyright infringement (at least not of OpenAI) cuz OpenAI doesn’t own the rights to what their bot produces.
Well written, to the point. What a irony ... DeepSeek threatens OpenAI by being open AI.
The OpenAIrony goes deeper than DeepSeek
Beautiful summary
OpenAI can just start generating its responses by calling deepseek in the background. Save a few bucks along the way, and the content is all the same I guess (i.e. somebody else’s)
LOL!
The issues of copyright and privacy infringement remain unresolved.
Beyond the ongoing reckless acquisition of training data—including personal information, web scraping, and harvesting content from social networks—there is also a persistent lack of attribution to original authors. AI models do not disclose their data sources, let alone credit the creators or specify how the data is being used.
A recent example is AI agents scraping website data without consent, as if it were freely available for the taking. This not only adds unnecessary traffic to websites, increasing server load and reducing service capacity for regular users, but also collects data for undisclosed purposes, including personal user information. As a result, some websites are actively blocking AI agents to protect their content and users.
This practice is akin to a fleet of heavy-duty lorries clogging up a local shopping centre, disrupting parking and access for legitimate customers—all without warning or permission.
Given this, it is both hypocritical and embarrassing to see OpenAI criticising DeepSeek for failing to play by the rules when OpenAI itself has repeatedly broken them.
Rules?
What rules?
Rules are for losers, not for rulers.
Rulers have their own rulers.
And their own losers.
The rules are simple: Act respectfully, or face the consequences. For AI agents, this could mean being blocked from all websites.
In general, always consider the impact of your actions. For AI agents and web scraping, it is best to contact the site owner first—otherwise, expect to be blocked.
Your continued behaviour, however, speaks for itself. This is why OpenAI is rated very low when it comes to shady practices.
I'm going to assume you've never done network administration or network security work. Blocking is hard, or at least blocking bad actors without also blocking legitimate traffic you want. Say you identify an origin that's an AI spider, somehow, because it's not going to advertise itself as such, and block the IP address. The spiders of a multi-billion dollar company is going to have a range of IPs, and can make use of VPN services just like the rest of us.
The 'no spiders' setting in your HTTP header only works if the owner of the spider agrees to it. If OpenAI is disregarding that it's an awful move, but nothing anyone can stop.
And every single owner of every single website would have to spend dozens of hours of a mid to high level admins time every year blocking the things.
The only way to stop it would be by law and international treaty. Specifically BRIC, EU and US would all have to agree to snub billionaires in order to protect the IP of bloggers, podcasters, and legacy media. Good luck with that.
Maybe the case law could be established via a suit from someone wealthy like music/film companies. You just have to rich enough to win the case, not richer than the tech companies.
I agree, and I think that's what will ultimately destroy OpenAI and all the rest of the LLM BS. The executive boards and c-suites of WB, Disney and the big book publishers are all in bed with the VCs and likely turning a blind eye to the mass IP theft for the sake of their own stock portfolios.
But they just got destroyed in the markets, and these are not nice people willing to overlook a loss of tens or hundreds of millions of dollars. They don't need that money, but it's how they keep score.
Now we might see those organizations get a lot more interested in protecting their IP if Chinese chatbot apps are pretending to be generic cartoon mouse or princess using text from Disney IP.
There is a big difference. OpenAI plundered intellectual property and then sells it to you with a subscription. Deepseek (allegedly) plundered this property from OpenAI and is giving it away.
Yeah, digital Robin Hood. I seriously hope it is true!
They are not "giving it away". They are taking YOUR data from YOU.
No. That's not correct. You can take the DeepSeek model and deploy it on your own hardware. This is the primary contrast with OpenAI. OpenAI runs the model for you, and collects your prompts and your subscription money. DeepSeek just uploaded a model that you can download and never have anything to do with them again ...
You can, but a huge number of users are just downloading the app, which does take your data.
And are you 100% sure there's no back door in that model when your computer is connected to the internet?
I don't know, that's a tough one. You sure "open"ai and facebook and ilks are not taking your data thru some backdoors? who's to say who is more benevolent?
Well I'm pretty sure that the Chinese Communist Party is less benevolent towards the interests of the average American than Sam and Zuck. But I do expect the CCP to win in the end. Hope you're cool with that outcome.
And as far as what I am cool with? I am cool with the whole world working together to solve big problems facing all of mankind, not differentiating between us and them, not declaring either you are with us or you are against us. How about that for a change?
"Chinese Communist Party is less benevolent towards the interests of the average American than Sam and Zuck" You have evidence for that?
Took me 10 seconds to find the below comment on this YouTube clip https://www.youtube.com/watch?v=0PxbXnIlRno . There are comments like this all over the Youtube space and elsewhere. Sometimes one really just have to look into the mirror and see where the problem is.
<quote from=https://www.youtube.com/watch?v=0PxbXnIlRno>
Personally, if I have a choice between my data being captured and analysed by one of:
- the US government
- a huge (US) mega corporate
- the Chinese government
..I'm choosing the Chinese government every single time
</quote>
The courts have yet to rule on the training copyright issue, but quite independent of how they rule, it is the case that one has no more legal right to give away stolen property than one has to sell it.
If this were not the case, thieves could simply gift stolen property to whomever they please and get away Scott free.
Furthermore, those who accept/use stolen property are not blameless according to the law either. And the latter simply encourage the former.
Isn't 'proprietary' and 'open source' an oxymoron?
With all the problems of "hallucinations" and unreliability, what is the benefit of an LLM such as ChatGPT? If you think deeply, you will find that ChatGPT is Altman's invention of a systemic way to steal without being caught.
Gary, I've got say as I read the clipping I was getting increasingly excited to get to the end and see your response, and I'm not happy. You're completely on point as usual, but bullet points 2 and then 3 made me suddenly laugh out so loud that I scared the sleeping cat off my lap in a flurry of panic and claws. She's sulking out with me, won't come back, and now my legs are scratched to all buggery.
I love it. Open AI is getting a taste of its own medicine. Tough luck.
Open AI’s terms of service? LOLOLOLOLOLOLOLOLOLOLOL
LOL! One of the world’s greatest plagiarizers is bent about having their intellectual property used without credit or compensation. Boo hoo!
OpenAIrony
All’s fair…
Karma is a bitch! indeed! thanks for giving me a smile in my morning