The following Op-Ed was written and submitted by Paul Plante.
By way of review, the black swan theory or theory of black swan events is a metaphor that describes an event that comes as a surprise, has a major effect, and is often inappropriately rationalized after the fact with the benefit of hindsight, the term being based on a Latin expression which presumed that black swans did not exist, and indeed, people, yesterday, that being 27 January 2025, a black swan did come along in the form of a small Chinese start-up company with an AI program named DeepSeek, that was developed in only two months and took less than $6 million to build, using reduced-capability chips from Nvidia, called H800s, and that black swan landed square on the head of former Biden commerce secretary Gina Raimondo, and once landed, took a real good dump there, as she deserved, which takes us back in time to an Asia Financial article titled “US Starts Funding Awards to Ramp up Chip Production at Home” on December 12, 2023, where we were informed as follows on that subject of Joe Biden’s Soviet-era “command economy,” the economic system of Cuba, North Korea, Belarus, China, Iran, and Russia, which Trump has now inherited the fruits of, as follows:
US Commerce Secretary Gina Raimondo has begun announcing major funding awards that aim to fire up US production of computer chips.
Meanwhile, California chip giant Nvidia was being encouraged to sell semiconductors to China – but not their top-of-the-range AI products.
“We have national security goals and we need to make our investments to hit those goals and we’re going to do that,” she added.
Raimondo said US officials were talking with Nvidia Corp about permissible sales of artificial intelligence chips to China but emphasized that it cannot sell its most advanced semiconductors to Chinese firms.
The Commerce secretary said Nvidia “can, will and should sell AI chips to China because most AI chips will be for commercial applications.”
She added: “What we cannot allow them to ship is the most sophisticated, highest processing power AI chips, which would enable China to train their frontier models.”
end quotes
And that is why the black swan chose the head of Gina Raimondo to take a good crap on, because all of her efforts and all of the efforts of Joe Biden to literally cripple China’s AI capabilities just failed miserably and spectacularly, as we see in the Reuters story titled “DeepSeek sparks global AI selloff, Nvidia losses about $593 billion of value” by Sinéad Carew, Amanda Cooper and Ankur Banerjee on January 27, 2025, as follows:
NEW YORK/LONDON/SINGAPORE, Jan 27 (Reuters) – Investors sold technology stocks across the globe on Monday as they worried that the emergence of a low-cost Chinese artificial intelligence model would threaten the dominance of current AI leaders like Nvidia, shaving $592.7 billion off the chipmaker’s market value.
Last week, Chinese startup DeepSeek launched a free AI assistant that it says uses less data at a fraction of the cost of incumbent services.
By Monday, the assistant had overtaken U.S. rival ChatGPT in downloads from Apple’s app store.
This led the tech-heavy Nasdaq to fall 3.1% on Monday.
Nvidia was the Nasdaq’s biggest drag, with its shares tumbling just under 17% and marking a record one-day loss in market capitalization for a Wall Street stock, according to LSEG data.
The Nasdaq’s next-biggest drag was chipmaker Broadcom Inc, which finished down 17.4%, followed by ChatGPT backer Microsoft, which fell 2.1% and then Google parent Alphabet, which ended down 4.2%.
U.S. equity declines followed a selloff that started in Asia, with Japan’s SoftBank Group finishing down 8.3%, and moved through Europe where ASML fell 7%.
“If it’s true that DeepSeek is the proverbial ‘better mousetrap,’ that could disrupt the entire AI narrative that has helped drive the markets over the last two years,” said Brian Jacobsen, chief economist at Annex Wealth Management in Menomonee Falls, Wisconsin.
“It could mean less demand for chips, less need for a massive build-out of power production to fuel the models, and less need for large-scale data centers.”
The hype around AI has powered a huge inflow of capital into equities in the last 18 months, inflating valuations and lifting stock markets to new highs.
As recently as Wednesday, U.S. AI-related stocks had rallied sharply after President Donald Trump announced a private-sector plan for what he said would be a $500 billion investment in AI infrastructure through a joint venture known as Stargate.
Deutsche Bank analyst Adrian Cox wrote in a research note on Monday that DeepSeek was “sowing seeds of doubt to the ‘bigger is better’ approach that has fuelled the AI race up to now.”
After the release of the first Chinese ChatGPT equivalent, made by search engine giant Baidu, there was widespread disappointment in China over the gap in AI capabilities between U.S. and Chinese firms.
But the apparent quality and cost-efficiency of DeepSeek’s models changed this view, with Silicon Valley executives showering praise on DeepSeek-V3 and DeepSeek-R1.
Little is known about the Hangzhou startup behind DeepSeek, whose controlling shareholder is Liang Wenfeng, co-founder of quantitative hedge fund High-Flyer, based on records.
Its researchers wrote in a paper last month that DeepSeek-V3 model, launched on Jan. 10, used Nvidia’s lower-capability H800 chips for training, spending less than $6 million.
DeepSeek-R1, released last week, is 20 to 50 times cheaper to use than OpenAI’s o1 model, depending on the task, according to a post on DeepSeek’s official WeChat account.
Chinese startup DeepSeek’s launch of a free open-source AI model to rival OpenAI’s ChatGPT is battering the tech sector.
Among other stocks, Vertiv Holdings, which builds data center infrastructure, slumped 29.9%.
Investors also sold off shares of power utilities, which had recently rallied sharply on hopes for a massive demand surge from power-hungry data centers needed for AI.
Vistra shares fell 28.3% while Constellation Energy shares fell 20.8% and NRG Energy lost 13.2%.
end quotes
And they never saw that coming only because they believed the Chinese incapable of doing more with less, which takes us back to January 13, 2025 and a Reuters story titled “US tightens its grip on AI chip flows across the globe” by Karen Freifeld from the last days of the flailing and adrift Biden administration, where we had old Joe’s efforts to hamstring China on display, to wit:
NEW YORK, Jan 13 (Reuters) – The U.S. government said on Monday it would further restrict artificial intelligence chip and technology exports, divvying up the world to keep advanced computing power in the United States and among its allies while finding more ways to block China’s access.
end quotes
What bullying Bidenesque bull**** on the one hand, and on the other, the Chinese just proved they never needed that access, which gives the up-raised middle finger to the failed Biden administration America has turned the page on, which takes us back to Reuters for more, to wit:
The new regulations will cap the number of AI chips that can be exported to most countries and allow unlimited access to U.S. AI technology for America’s closest allies, while also maintaining a block on exports to China, Russia, Iran and North Korea.
Unveiled in the final days of outgoing President Joe Biden’s administration, the lengthy new rules go beyond China and are aimed at helping the United States maintain its dominant status in AI by controlling it around the world.
“The U.S. leads AI now – both AI development and AI chip design, and it’s critical that we keep it that way,” Commerce Secretary Gina Raimondo said.
end quotes
FAMOUS LAST WORDS, Gina, which is why today the top of her head is adorned with a huge mound of black swan poop, which takes us back to Reuters again, to wit:
The regulations cap a four-year Biden administration effort to hobble China’s access to advanced chips that can enhance its military capabilities and seek to maintain U.S. leadership in AI by closing loopholes and adding new guard rails to control the flow of chips and global development of AI.
The regulation divides the world into three tiers.
About 18 countries, including Japan, Britain, South Korea and the Netherlands, will essentially be exempt from the rules.
Some 120 other countries, including Singapore, Israel, Saudi Arabia and the United Arab Emirates, will face country caps.
And arms-embargoed countries like Russia, China and Iran will be barred from receiving the technology altogether.
In addition, U.S.-headquartered providers likely to receive global authorizations such as Amazon Web Services and Microsoft will be allowed to deploy only 50% of their total AI computing power outside the United States, no more than 25% outside of the Tier 1 countries and no more than 7% in a single non-Tier 1 country.
“How effective the rule ends up being in the next 10 to 15 years is now up to the incoming team,” said Meghan Harris, a national security official during the first Trump administration.
“They are well aware that ensuring a dominant domestic industry is a core element of competition with China.”
China’s Commerce Ministry said in response to the new rules that China will take necessary measures to safeguard its “legitimate rights and interests”.
“The U.S. has to be prepared for rapid increases in AI’s capability in the coming years, which could have transformative impact on the economy and on our national security,” U.S. National Security Adviser Jake Sullivan said.
end quotes
And yesterday, it just did!
Stay tuned, more is yet to come.
So, people, how was it that we came to have Gina Raimondo as Joe Biden’s Commerce Secretary so that she would be in a position to have a black swan come and take a good crap on her head?
According to history as written, in early February of 2020, Gina Raimondo appeared alongside former Republican New York City Mayor and Democratic presidential hopeful Michael Bloomberg at the Wexford Innovation Center in Providence to endorse his candidacy, a move she described as “an easy call,” and Gina was named a national co-chair for the Bloomberg campaign.
According to history as written, her press secretary, Jennifer Bogdan Jones of the Governor’s Office, told The Providence Journal that Raimondo was “prepared to do whatever it takes to support Mike and defeat President Trump.”
As Mike Bloomberg’s campaign co-chair, Gina would have “provided advice and attended events,” which must have been beyond her level of competence because less than a month later, Bloomberg dropped out of the race and endorsed former Vice President Joe Biden, who we have just turned the page on, along with Kamala Harris, Timmy Walz, Gwen Walz and Gina Raimondo, and big surprise, on the same day Mike Bloomberg dropped out of the presidential race, Gina, ever the opportunist, also endorsed Biden.
According to Gina, who was Mike Bloomberg’s national co-chair for the Bloomberg campaign, Bloomberg, her chosen candidate, “obviously” performed poorly on the debate stage but for her, supporting his candidacy “was an easy decision” for her at the beginning, but supporting Biden “was an easy decision, too,” with Raimondo concluding that it was now time “to unify behind Joe Biden,” and out of that somehow, we ended up with Gina Raimondo as Joe Biden’s Commerce Secretary.
As to her record of incompetence prior to Joe Biden picking her from out of the crowd to be his Commerce Secretary, on November 2, 2010, Raimondo was elected as general treasurer of Rhode Island by a margin of 62% to 38%, and during her first year as general treasurer, she prioritized reforming Rhode Island’s public employee pension system, advocating for benefit cuts as the solution to Rhode Island’s pension problems, and under Raimondo’s tenure, the pension fund was criticized for underperforming when compared with its peers, with Raimondo’s critics attributing the underperformance to a sharp increase in fees paid to hedge fund managers, and hey, they are people, too, and like everybody else, they need to eat, so of course, it is only natural to pay them more, even if it adversely impacts the common folks the pension fund is allegedly there for.
But that alone was not enough stacked-up incompetence for Joe Biden to notice her.
So how about the subject of RI DCYF fatalities and near-fatalities?
Under Gina as governor of Rhode island, the Rhode Island Department of Children, Youth & Families came under fire for the rate of deaths and near-deaths of children in its care.
In a period between January 2016 and December 2017, there were 31 fatalities or near fatalities of children in its care, with eight being confirmed fatal.
Raimondo appointed Trista Piccola as her new DCYF director in January 2017.
Piccola’s term was marked by the death and near-deaths of children, high staff turnover, votes of no confidence, and high budget deficits, and there, people, we are getting into the kind of incompetence that Joe Biden regarded highly and rewarded with cabinet positions in his administration!
Getting back to that part of the story, Rep. Patricia Serpa and Rep. Charlene Lima called for Piccola’s resignation, which finally occurred in July 2019.
In October 2018, the United States Department of Health and Human Services’ Administration for Children and Families ordered the Raimondo Administration DCYF to improve in 33 of 36 areas assessed.
The federal report noted that DCYF services were “inadequate, not developed when needed, or lacked consistent monitoring.”
Harvard Kennedy School professor and former Obama Administration official Jeffrey Liebman agreed with the recommendations and analysis of the report from the U.S. Department of Health and Human Services and claimed that the DCYF was “the most messed-up agency ever.”
And as a result, Joe Biden, Hussein Obama’s successor in office to continue the Obama LEGACY, hired Gina Raimondo to be his Commerce Secretary, and the rest is now history – OUR history, and people, we are stuck with it!
But stay tuned, for more is yet to come!
My understanding is (a) the Chinese team really did some novel and innovative science to get the result, (b) the “$6 million” figure is probably an exaggeration that excludes a lot of major costs involved in getting to the final model, and (c) the jury is out as to whether they actually had the top-of-the-line NVIDIA chips – they may have been getting them through shell companies in Singapore. So, novel and innovative science for sure, but the extent of cost saving over more transparent American companies is possibly/probably exaggerated. Plus, they’re still copying American models a lot.
Interesting thoughts, Virginia Gentleman, but speaking as an engineer here, who learned engineering mechanics from Stepan Prokopovich Timoshenko, a Russian who wrote seminal works in the areas of engineering mechanics, elasticity and strength of materials, I have a lot of trouble with the concept of “American models,” as if NOBODY in the world could possibly come up with an idea about anything but an American.
Ideas know no borders and ideas don’t care about national identity.
And like phpBB, that AI stuff, which has been hyped and hyped and hyped, as if it truly were some kind of revolutionary idea that only an American could come up with, and is going to radically change the world, is open source, which term refers to any program whose source code is made available for use or modification as users or other developers see fit, and unlike proprietary software, open source software is computer software that is developed as a public, open collaboration and made freely available to the public.
And anyway, this story is not really about DeepSeek.
It is about Gina Raimondo, who thank the heavens is now gone, and her arrogant empty head, her main attraction for Joe Biden to make her his commerce secretary, who thought she could issue a bunch of dictatorial orders that were going to send China back to the stone ages, technologically speaking, to supposedly protect our national security, which was never so weak and flimsy as it was under Gina Raimondo, scared of her own shadow, and worried that some Chinese person might sneak up behind her and give her a wedgie, and the doddering, senile Joe Biden.
When all is said and done here, Gina may well have weakened this country economically with her stupid orders, and that does not do our national security any good whatsoever.
Consider the Reuters article “US finalizes up to $6.75 billion in chips awards for Samsung, Texas Instruments, Amkor” by David Shepardson on December 20, 2024, where we were informed as follows:
WASHINGTON, Dec 20 (Reuters) – The U.S. Commerce Department said on Friday it was finalizing an award of up to $4.745 billion to South Korea’s Samsung Electronics and up to $1.61 billion for Texas Instruments to expand chip production.
In total, Commerce has finalized over $33 billion of the over $36 billion in proposed incentives funding.
“With this investment in Samsung, the U.S. is now officially the only country on the planet that is home to all five leading-edge semiconductor manufacturers,” said Commerce Secretary Gina Raimondo.
end quotes
Said another way, on the one hand, the Commerce department under Gina “THE WITLESS” Raimondo has pumped BILLIONS of our tax dollars into five huge corporations, who are each other’s competitors, and with the other hand, the same Commerce department has taken away their markets for their products, which takes us to the Reuters story titled “Samsung warns of slow AI chip sales in Q1, hurt by US restrictions on China” by Joyce Lee and Hyunjoo Jin on January 31, 2025, where we have this to consider:
SEOUL, Jan 31 (Reuters) – Samsung Electronics on Friday warned of sluggish sales of its artificial intelligence chips in the current quarter due to U.S. export restrictions to China, and as it worked towards launching an improved version of its high-end chips.
end quote
And in the meantime, out pops DeepSeek!
For dimes compared to the BILLIONS BLOATED AMERICA has to pay for the same thing.
It used to be said that engineers do for a dime, what it costs someone else to do.
Now, it has been modified to read a Chinese engineer and can for a dime what it costs an American engineer BILLIONS to do.
Way to go, Gina Raimondo!
Sigh, DeepSqueak has been revealed to be yet another Chinese lie. Runs on the NVidia 100, not the other one and the total cost was in the billions, but they lied and only claimed the last load test as the ‘cost’.
Don’t trust China, China is assshoe.
https://interestingengineering.com/culture/deepseeks-ai-training-cost-billion
That Interesting Engineering article titled “Research exposes DeepSeek’s AI training cost is not $6M, it’s a staggering $1.3Billion” by Kapil Kajal on January 31, 2025 you supplied us with, Civil Servant, was indeed quite interesting and informative, and nowhere in the article does it say the Chinese lied to anybody, and given all the lies we have been told these last four years by BIDEN INCORPORATED and its main-stream and legacy media supporters and enablers, it’s
interesting that you would shift the blame here and have it be the Chinese who are the liars.
According to your article, which is actually quite favorable to DeepSeek, a rising player in the AI landscape, it was spotlighted in its latest report by a crowd called SemiAnalysis, an independent research company.
As to the $6 million estimate, that primarily considers GPU pre-training expenses, neglecting the significant investments in research and development, infrastructure, and other essential costs accruing to the company, which costs accrue to every company engaged in that same field of endeavor.
According to the SemiAnalysis report, DeepSeek’s total server capital expenditure (CapEx) amounted to $1.3 billion, with much of this financial commitment directed toward operating and maintaining its extensive GPU clusters, the backbone of its computational power.
B y way of comparison, throughout 2024, Nvidia invested $1 billion in tens of AI companies as it solidified itself as a key backer of new start-ups looking to capitalize on the AI boom, and Meta plans to increase its spending to $65 Billion this year in its A.I. push, with much of the capital investment, a big jump from 2024, going to fund expansion of Meta’s data centers, which provide the computing power needed by A.I. products and algorithms, so $1.6 billion for DeepSeek really isn’t all that much.
According to the SemiAnalysis report, DeepSeek reportedly has access to approximately 50,000 Hopper GPUs, leading to some misconceptions in the industry, and SemiAnalysis clarifies that this does not equate to having 50,000 H100s, as some previously inferred.
Instead, the GPU inventory comprises a mix of models, including H800s, H100s, and the country-specific H20s produced by NVIDIA in response to U.S. export restrictions.
This nuanced understanding of their hardware inventory underscores the strategic decisions in sourcing and operational efficiency at DeepSeek.
And here your report, Civil Servant, goes on to praise DeepSeek’s organizational structure, informing us that unlike some of the larger AI laboratories, DeepSeek operates its data centers and employs a streamlined model that aids in its agility and efficiency, and as the AI landscape grows increasingly competitive, that ability to adapt quickly becomes a vital asset.
As to performance, the analysis you presented us with indicates that DeepSeek’s R1 model demonstrates comparable reasoning capabilities to OpenAI’s o1, which positions DeepSeek at a crossroads where balancing performance and cost is key to its future success.
As to the engineering, according to the SemiAnalysis report, the Multi-Head Latent Attention (MLA) technology employed by DeepSeek is a groundbreaking innovation highlighted in the report, which cutting-edge approach significantly slashes inference costs by an impressive 93.3% through reduced usage of key-value (KV) caching, representing a major leap toward cost-effective AI solutions, and experts suggest that innovations that emerged from DeepSeek are likely to be swiftly adopted by Western AI labs eager to remain competitive.
In conclusion, your SemiAnalysis report paints a complex picture of DeepSeek’s current standing within the AI realm and the revelations regarding its cost structure, GPU utilization, and innovative capabilities position DeepSeek as a formidable player, as the firm continues to evolve, the industry watches closely — eager to see how it will respond to emerging challenges and opportunities in an ever-changing landscape.
So thanks for sharing that with us, Civil Servant.
Your attention to our continuing education in here is well appreciated.
Not to mention, Civil Servant, that Alphabet said it plans to spend an astonishing $75 billion on capital expenditures as it builds out its AI offerings and races against megacap rivals to build out data centers and new infrastructure, with that figure being much higher than the $58.84 billion expected by Wall Street analysts, according to FactSet.
Makes the $1.3 billion DeepSeek total server capital expenditure (CapEx) amount seem like mere chicken feed, doesn’t it?
The machines are taking control and:
“Most human beings
have an almost
infinite capacity for
taking things for granted.”
A. Huxley
The machines can only take control because people willingly surrender their own control and cede it to the machines.
In the end, only a few preferred liberty!
As for the rest, all they wished for was a kind master.
– Sallust