China’s Deepseek is the Temu of AI
As the stock sell off in technology stocks due to China’s Deepseek AI technology has everyone worried that China has found a “magic pill” that has solved the power hungry issue of feeding AI models.
And yes China has found the “magic pill” and in fact found three of them.
One magic pill which is being presumed is causing the Earth’s day to grow longer: Three Gorges Dam. And the energy that Deepseek’s uses is destroying the Mekong River Valley that flows through Myanmar, Laos, Thailand, Cambodia, and on into Vietnam.
Three Gorges Dam in China is being speculated as the energy being used for Deepseek’s technology. NASA scientists calculated that the shift of water mass stored by the dams would increase the total length of the Earth’s day by 0.06 microseconds and make the Earth slightly more round in the middle and flat on the poles.
Second magic pill is being theorized as being the source to making it cheaper is enslaved labor that manufacture China’s homegrown AI chips using Uyghurs as political prisoners.
The Chinese government has used its justice system to sentence and imprison an estimated half-million people during the brutal crackdown in Xinjiang,. Xinjiang High People’s Procuratorate, which has continued publishing statistics that a total of 540,826 people have been prosecuted in the region since 2017. Given that China’s conviction rate is above 99.9 percent, almost all of these 540,826 people would have been convicted.
And the third magic pill is following a page from Temu, Deepseek is being run as a loss leader to push hype.
How perfectly timed the release of Deepseek to coincide with the tarrifs that Trump has proposed?
Don’t get me wrong, I am not a fan of Trump’s tariffs, nor did I vote for him, or do I even like him as a person. But I do find it remarkably well timed that as OpenAI, Oracle, Softbank, and MGX announce Stargate — and Davos is wrapping up — then suddenly a Chinese company releases the greatest AI model in history.
Isn’t a little bit ironic that the US and Taiwan’s darling NVIDIA just had $600 billion in marketcap erased by a magic AI startup that no one had heard of until a week ago?
Since Thursday of last week, I have been getting emails from friends in the AI/ML space all over the world saying the same thing, “Have you tried Deepseek?”
Well I indeed did try Deepseek but conveniently couldn’t gain access to it as they offer Google login but it doesn’t work. Instead, the website forces you to put your mobile, email address in to access the site and more details if you download the app. I refused to do this. Yes I know that if I had used my Google login and it actually worked — I would be sharing the same type of information but with Google as the middle person. But because it didn’t work, I refused to give more information nor use the model.
I believe Deepseek is a cybersecurity risk using social engineering.
In a typical social engineering attack, a cybercriminal will communicate with the intended victim by saying they are from a trusted organization. If the manipulation works (the victim believes the attacker is who they say they are), the attacker will encourage the victim to take further action.
We are only told that the costs around Deepseek are cheaper by China itself and as Meta has setup three war rooms to determine how Deepseek can indeed be cheaper and replicate how they did it — I think the answer will be: It’s not.
But by the time that its been determined that this was a social engineering exercise — all of our mobile numbers, emails, back doors via the app, training data uploaded to the model, prompt engineering, and code submitted — has now made China number one in the AI world not by innovation and creativity — but by stealing it all once again.
“I don’t care that they stole my idea. I care that they don’t have any of their own.” — Nikola Tesla
