EN — LARRY ROMANOFF: China’s DeepSeek and the Criminal World of American AI — Part 2 – The Response

 

China’s DeepSeek and the Criminal World of American AI

Part 2 – The Response

By Larry Romanoff

 

CHINESE     ENGLISH

 

Everybody Praised DeepSeek . . .

Then Everybody Bashed it . . .

China Stole All OpenAI’s Data. Oh, no, wait . . .

Then the CIA Cyber-Attacks . ..

 

Everybody Praised DeepSeek . . .

 

 

 

Sam Altman, the CEO of OpenAI, said, “This will be the greatest technology humanity has yet developed”. Microsoft CEO Satya Nadella said the new DeepSeek model was both “super-impressive” and “super-efficient”, and the AI developments coming out of China needed to be taken “very, very seriously”. [1]French entrepreneur Arnaud Bertrand said, “There’s no overstating how profoundly this changes the whole game. And not only with regards to AI, it’s also a massive indictment of the US’s misguided attempt to stop China’s technological development, without which Deepseek may not have been possible…” [2]Political commentator Matt Bruenig posted on X that: “I have been extensively using Gemini, ChatGPT, and Claude for document summary for nearly a year. Deepseek is better than all of them at it. The chatbot version of it is free. Price to use [its] API is 99.5% below the price of OpenAI’s API. [3]

 

Alexander Wang, CEO of ScaleAI said, “What we found was that DeepSeek, the leading Chinese AI lab, their models are actually the top performing or roughly on par with the best American models.” [4]Aravind Srinivas, CEO of Perplexity AI, said it’s a myth that Chinese are good only at copying, so that if we stop writing research papers in America, if we stop describing the details of our infrastructure architecture, and don’t open-source, China will never be able to catch up. But the truth is that … DeepSeek is so good that the American firms will copy it all.[5] More than 670 derivative models have already been created from DeepSeek, all around the world, and that is the power of open-source models. [6]

 

Facebook (Meta) is integrating as much of DeepSeek as it can, into its own model. Mark Zuckerberg said, “Engineers are moving frantically to dissect DeepSeek and copy anything and everything we can from it. I’m not even exaggerating.” Amazon is offering it and also using and integrating it. Microsoft is hosting it on Azure, and even Nvidia is hosting DeepSeek R1 on its developer platform. One AI company said they had heard from many Fortune 500 companies who want to run DeepSeek and who are questioning whether they want to continue to pay OpenAI such high monthly fees when they could run DeepSeek themselves for free. [7]

 

Then Everybody Bashed it . . .

 

“What is the West doing about this? They are playing the Tik-Tok card. If you can’t defeat them, discredit them. If that doesn’t work, raise national security concerns, and if even that fails, then just go ahead and ban them.”[8]

 

Can’t let the Commies Have it

 

And that’s more dangerous because then they get to own the mindshare, the entire American AI ecosystem.

 

Sam Altman from OpenAI now claims “We worry a lot about authoritarian governments developing this… [having AI] in the wrong human hands” [9] Aravind Srinivas, the CEO of Perplexity AI, said “the most dangerous thing is that the Chinese have the best open-source model and all the Americans are now building on that. And that’s more dangerous because then they get to own the mindshare, the entire American AI ecosystem.“ [10]

 

National Security

 

The White House is looking into “national security” concerns with DeepSeek and, as the initial excitement subsided, US officials discovered “national security implications”, [11] which were so serious that the US Navy immediately banned the use of DeepSeek by any and all naval staff. [12] Italy said “the data of millions of Italians could be at risk”. Australia has told its citizens to “be very careful” when using the app; in other words, “Don’t use it.” Asked about national security concerns, Matt Sheehan, a fellow at the Carnegie Institute for Perpetual War, said “Yes, almost certainly”; whether their data is going to China, and could be accessed by the Chinese Communist Party. [13]

 

Graphic Processing Units (GPUs)

 

Alexander Wang said, “The Chinese labs have more H100s (the most powerful GPUs) than people think. My understanding is that DeepSeek has about 50,000 H100s, which they can’t talk about, obviously, because it’s against the export controls the US has put in place.” [14]JP Morgan, Citibank and others have accused the Chinese firm of having as many as 50,000 GPUs instead of the 2,000 they claimed. [15]Elon Musk accused DeepSeek of lying about its breakthrough model. He says the Chinese company is lying about how many chips it has purchased. [16]Joseph Carlson, a self-styled AI investment expert, said, “I don’t believe you, DeepSeek. I think you spent a lot more money. There are rumors out there that they had 50,000 GPUs, which sounds a lot more reasonable.” [17]

 

Psy-op; send a message

 

“It’s symbolic that DeepSeek released their first model on Christmas Day when the rest of us are celebrating…”

 

A few people were quick to say that DeepSeek was just “A CCP psy-op”, conducting economic warfare against the US, and making American AI unprofitable. [18] Alexander Wang from ScaleAI said, “It’s symbolic that DeepSeek released their first model on Christmas Day when the rest of us are celebrating…” [19] “China” released DeepSeek R1 on the very same day Trump was inaugurated, to “send a stern message” to the US government. [20]

 

Cheap China, low quality

 

“Investors are thinking there is more hype than substance in DeepSeek. DeepSeek has shown only that you can train a small model off a larger one. Deepseek hasn’t shown they are capable of training the next iteration.” [21] What does this mean? It is like saying that you can build a beautiful and perfect 2-story house, but that is irrelevant and useless because you haven’t proven you can build a 3-story house. American reasoning at its finest. Singapore Live TV said that DeepSeek may be “a viable solution for the lower end of the market but not for the higher end”. [22] This is called “damning with faint praise”.

 

Summary? China Copies, Steals, and Lies About Everything

 

Kevin O’Leary, Chairman O’Leary Ventures: “There’s a lot of skepticism about this DeepSeek product. A $6 million dollar price from the desk of a hedge fund manager; I don’t believe it. No one does. We are in a propaganda war, an economic war, a technological war with China. Very risky to put this stuff on your phones. What they did was to copy technology that was already existing. It was innovated in the United States. [23] Matt Sheehan, Carnegie Institute for Perpetual War, “We always need to take information with a grain of salt when it’s coming from China”. [24]

 

Commentary

 

“China” released DeepSeek; “Beijing’s latest thrust”; “The Chinese Communist Party’s offensive”. What rubbish. It wasn’t “China” who created DeepSeek, and had nothing to do with either Beijing or the Chinese Communist Party. How dare the inscrutable Chinese release something during Christmas, when the Americans aren’t watching? American conspiracy theorists may be deeply disappointed to learn that nobody at DeepSeek gives a shit what is happening in Washington on any particular day.

 

All of this contrary noise was just the daily China-bashing that occurs in the Western media: “If anyone from China said it, it cannot be true”. Of course, there is never any evidence presented; every dirty accusation against China is treated as fact – no evidence required. Data privacy and national security noise are nonsense, since DeepSeek can be downloaded and run on computers in the US, with no connection or contact with China.

 

China made awesome mobile phones and the US banned them. China made fantastic and inexpensive EV cars and the US banned them. China makes the best small drones in the world, so the US banned them. China made one of the most popular social media platforms in the world, and the US banned it. Now a Chinese company produced an amazing and inexpensive AI model, and already the US is looking for excuses to ban it.

 

If the US makes a breakthrough it’s a giant step for mankind, but if a Chinese firm makes a breakthrough, it’s a threat to the world’s national security. A small Chinese firm creates one AI model, and suddenly China wants to dominate the world. So much irrationality and hysterical nonsense. If the US were a person, it would be a 14-year-old girl.

 

On the Other Hand . . .

 

Gary Marcus, US university professor and AI expert, was asked if we can trust what DeepSeek has said. His response was that it’s very likely that what DeepSeek described, is correct. He said he read their papers, and it all makes sense, and their results have already been replicated by others. What they did is plausible; they worked harder on optimization than other people. DeepSeek did it in super-impressive fashion, and it is likely legitimate. [25]

 

It doesn’t matter if it’s a Chinese government psy-op or not. The technological innovation is real, and having a model train itself through reinforcement learning is impressive. DeepSeek’s technical paper says clearly that the $6 million does not include costs associated with prior research on architectural algorithms. [26]DeepSeek’s research papers were so detailed in describing how they did what they did, and many companies have already reproduced their results, so there is no room for doubt that their claims were factual.

 

China Stole All OpenAI’s Data. Oh, no, wait . . .

 

 

OpenAI went to the Financial Times with a complaint that DeepSeek may have used OpenAI’s data to train its model. They claimed they had seen “some evidence” of distillation which they “suspect” “may have come from” DeepSeek, but of course they declined to provide details of their evidence. But where did OpenAI obtain all the data to train its own model? It stole it from websites all over the world. Every website is copyrighted, and every country has copyright laws stating it is illegal to take all of the content from a website, then repackage and sell it commercially. Yet this is precisely what OpenAI did.

 

The truth is that OpenAI built their entire business on data they stole from everyone else.Sam Altman, the CEO of OpenAI, has admitted this many times, and Satya Nadella, the CEO of Microsoft, said that anything that can be accessed on the internet is in the public domain and can simply be taken by anyone who wants it. But it’s a felony crime if you do it to them. This is exactly the same as calling the police to report that someone stole a car from you, that you stole from someone else yesterday.

 

“OpenAI pulled a fast one. They began accusing DeepSeek of copying their text. They say the Chinese company used OpenAI’s data to train their model, and OpenAI says they have the evidence although they refuse to release it – which means they don’t have it. “It is hard to miss the irony here. OpenAI is having a meltdown because someone may have done what they have been doing for years. [OpenAI] is a company that has been training on copyrighted material with no shame; books, news articles, music, Youtube videos. OpenAI has literally been ripping off other people’s work, so what’s the fuss about now? And this is not even a secret. OpenAI’s leadership has talked about it. Sam Altman himself has acknowledged it, so why are they playing the victim now? They should know that the copy-paste button works both ways.” [27]

 

DeepSeek “stole” OpenAI’s Data? What makes you think it was OpenAI’s data in the first place? Why are we so obsessed with the possibility of Chinese companies stealing our data when American companies actually ARE stealing our data? Every website has a ‘sign’ that says you cannot take their content and use it for your commercial purposes without permission and payment. But it seems OpenAI didn’t see that sign, and stole all the data from millions of websites, without permission and without paying. Well, OpenAI, it seems that maybe DeepSeek missed seeing that sign on your website, too. [28]

 

But apparently not everyone sees the irony in this. David Sacks, White House AI and Crypto Czar: “There is substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models, and I don’t think OpenAI is very happy about this. [29] [We need to] slow down some of these copycat models.” [30]

 

Yet while OpenAI is crying “foul”, they are already embroiled in copyright lawsuits in many countries. A huge “landmark” lawsuit was filed in India by the Indian News Agency ANI back in 2024, who accused OpenAI of improperly using copyrighted content. [31] “They have scraped and used copyrighted material from all the news websites without permission.” Several other major news websites in India have already joined the lawsuit. OpenAI is embroiled in other lawsuits as well. They are being sued by the New York Times, other NY newspapers, various authors and journalists, and all indications are that this may snowball into something huge. [32][33][34] Microsoft is being sued as well.

 

OpenAI’s troubles are just beginning. They issued a press release saying they might have to totally withdraw ChatGPT from Europe, because of their privacy and copyright infringements. This is so serious for OpenAI that “persons” in the UK are floating new legislation to make this particular kind of copyright violation “legal”. If you or I copy something from a website without permission, we might be bankrupted or even go to prison, but if the large tech companies steal data from millions of websites, that somehow becomes praiseworthy and deserving of immunity. [35]

 

In fact, OpenAI is being sued by many groups in countries all around the world, now embroiled in multiple legal battles for everything from privacy issues, to copyright violations, to character defamation, and is even being investigated by US FTC and SEC. [36] OpenAI is being sued in Germany for unlicensed use of song lyrics. [37] The list goes on.

 

Then the CIA Cyber-Attacks . . .

 

The US government, perhaps abetted by its friends in OpenAI, immediately launched massive and protracted DDOS and other cyber-attacks on DeepSeek…

 

On January 28, 2025, DeepSeek announced it was limiting registrations and restricting access because its website was being subjected to massive, “large-scale and malicious” cyber-attacks from the US. [38] The reason was that American users flocked to DeepSeek’s website to download the app, making it the #1 download on both Apple’s app store and on Google.As soon as this happened, the US government, perhaps abetted by its friends in OpenAI, immediately launched massive and protracted DDOS and other cyber-attacks on DeepSeek.

 

Another article said, “DeepSeek is restricting registrations and limiting who can log into its website, allowing only China-based phone numbers, because of large-scale and malicious cyber-attacks. These attacks on DeepSeek add to the intrigue.” [39] It was interesting to see that DeepSeek must have been expecting this response, and had the technical skill to combat it. Their website was never forced off-line; they simply found it necessary to reduce new registrations. Well-done, China.

 

“With great fame comes great risk. The company was hit by a cyber-attack that was so bad they had to limit registrations. The timing was suspicious and the world wasted no time in determining who was the culprit.” [40] Sky News in Australia said that DeepSeek was so popular that downloads of the app were causing their website to crash, completely ignoring the statements on their own website of the malicious cyber-attacks. [41]

 

 

*

Mr. Romanoff’s writing has been translated into 34 languages and his articles posted on more than 150 foreign-language news and politics websites in more than 30 countries, as well as more than 100 English language platforms. Larry Romanoff is a retired management consultant and businessman. He has held senior executive positions in international consulting firms, and owned an international import-export business. He has been a visiting professor at Shanghai’s Fudan University, presenting case studies in international affairs to senior EMBA classes. Mr. Romanoff lives in Shanghai and is currently writing a series of ten books generally related to China and the West. He is one of the contributing authors to Cynthia McKinney’s new anthology ‘When China Sneezes’. (Chap. 2 — Dealing with Demons).

His full archive can be seen at

https://www.bluemoonofshanghai.com/  + https://www.moonofshanghai.com/

He can be contacted at: 2186604556@qq.com

 

*

NOTES

 

[1] How China’s new AI model DeepSeek is threatening U.S. dominance

https://www.cnbc.com/2025/01/24/how-chinas-new-ai-model-deepseek-is-threatening-us-dominance.html

[2] China’s DeepSeek AI Moves the Capital of Tech from Palo Alto to Hangzhou

https://www.unz.com/mwhitney/chinas-deepseek-ai-moves-the-capital-of-tech-from-palo-alto-to-hangzhou/

[3] Why everyone in AI is freaking out about DeepSeek

[4] AI video

https://www.douyin.com/video/7464439238818385186

[5] Interview

https://www.douyin.com/video/7464665030865243451

[6] Hundreds of new models already

https://www.douyin.com/video/7465392756186254632

[7] Fortune 500 companies are maybe going to Deepseek

https://www.douyin.com/video/7466154959046774031

[8] Indian News Anchor

https://www.douyin.com/video/7465599297011174710

[9] Biggest one day stock drop. Nvidia biggest in history

https://www.douyin.com/video/7464850962012753152

[10] Interview

https://www.douyin.com/video/7464665030865243451

[11] Security risk and theft

https://www.douyin.com/video/7466356079782399267

[12] Commentary

https://www.douyin.com/video/7466124155751107859

[13] Biased view

https://www.douyin.com/video/7464781890348797184

[14] Wang Interview

https://www.douyin.com/video/7464439238818385186

[15] DeepSeek is Lying

https://www.douyin.com/video/7465193513030798632

[16] Elon Musk says Deepseek lying

https://www.douyin.com/video/7464889659298974986

[17] Reasoning

https://www.douyin.com/video/7464856751351844106

[18] Deep seek Psy-op

https://www.douyin.com/video/7464679744240422203

[19] Wang Interview

https://www.douyin.com/video/7464439238818385186

[20] More opinion

https://www.douyin.com/video/7464561079587310885

[21] Investors

https://www.douyin.com/video/7466117741838109964

[22] Viable solution for lower end

https://www.douyin.com/video/7466054808114416911

[23] Lies, accusations

https://www.douyin.com/video/7465312313210359080

[24] Biased view

https://www.douyin.com/video/7464781890348797184

[25] DeepSeek is Telling the Truth

https://www.douyin.com/video/7465040429461900580

[26] Nvidia selloff

https://www.douyin.com/video/7464767993176689971

[27] OpenAI pulled a fast one

https://www.douyin.com/video/7465933245880831267

[28] Security risk and theft

https://www.douyin.com/video/7466356079782399267

[29] Copycat model

https://www.douyin.com/video/7465318338038730018

[30] DeepSeek used OpenAI’s model to train its competitor using ‘distillation,’ White House AI czar says

https://fortune.com/2025/01/29/deepseek-openais-what-is-distillation-david-sacks/

[31] OpenAI faces landmark copyright infringement

https://legal.economictimes.indiatimes.com/news/litigation/openai-faces-landmark-copyright-infringement-case-in-india-legal-experts-weighs-in/117782660

[32] ‘The New York Times’ takes OpenAI to court. ChatGPT’s future could be on the line

https://www.npr.org/2025/01/14/nx-s1-5258952/new-york-times-openai-microsoft

[33] The New York Times sues OpenAI and Microsoft for using its stories to train chatbots

https://apnews.com/article/nyt-new-york-times-openai-microsoft-6ea53a8ad3efa06ee4643b697df0ba57

[34] Sarah Silverman sues OpenAI and Meta alleging copyright infringement

https://edition.cnn.com/2023/07/10/tech/sarah-silverman-openai-meta-lawsuit/index.html

[35] The irony of stealing

https://www.douyin.com/video/7466774155384360227

[36] ChatGPT maker OpenAI, valued at $157bn, sued by GEMA in Germany over unlicensed use of song lyrics

https://www.musicbusinessworldwide.com/openai-sued-by-gema-in-germany-for-unlicensed-use-of-song-lyrics/

[37] Legal Challenges Surround OpenAI: A Closer Look at the Lawsuits

https://analyticsindiamag.com/ai-trends/all-the-lawsuits-filed-against-openai/

[38] Videos Deepseek

https://www.douyin.com/video/7464779585578650895

[39] Cyber-Attack

https://www.douyin.com/video/7464779585578650895

[40] Indian News Anchor

https://www.douyin.com/video/7465336499320392999

[41] SkyNewsLies

https://www.douyin.com/video/7464855769477893434

 

*

This article may contain copyrighted material, the use of which has not been specifically authorised by the copyright owner. This content is being made available under the Fair Use doctrine, and is for educational and information purposes only. There is no commercial use of this content.

 

 Other works by this Author

 

BIOLOGICAL WARFARE IN ACTION

民主，最危险的宗教  

Democracy – The Most Dangerous Religion

建立在谎言上的国家–第一卷–美国如何变得富有
NATIONS BUILT ON LIES — VOLUME 1 — How the US Became Rich

美国随笔

Essays on America

美国警察国家》第一卷免费电子书

Police State America Volume One

宣传与媒体  

PROPAGANDA and THE MEDIA

     BOOKS IN ENGLISH

 

THE WORLD OF BIOLOGICAL WARFARE

NATIONS BUILT ON LIES — VOLUME 2 — Life in a Failed State

NATIONS BUILT ON LIES — VOLUME 3 — The Branding of America

False Flags and Conspiracy Theories

FILLING THE VOID 

Police State America Volume Two

BERNAYS AND PROPAGANDA

Kamila Valieva

LARRY ROMANOFF FREE E-BOOKS & PDF ARTICLES

 

Copyright © Larry Romanoff, Blue Moon of Shanghai, Moon of Shanghai,  2025