It's no coincidence R1 was released on inauguration day. Export controls will start to tighten further regardless of release date, so they flexed with Challenge Accepted as hard as possible lol P.S. Elon's prob right and that 500B number is full of Sham Altman inspired hokum.
so that means DeepSeek has used 50.000 NVDA chips to help it to build the model / do the research since Biden, in his last days in office, had placed stiffer restriction on exporting NVDA chips to china, limiting the amount of future research/modeling by the PRC company
sanction works no better eg than the sanction against Putin, what was supposed to be, at most, a month-long take-over of Ukraine, has turned into 3-yr nightmare, with no end in sight, and global humuliation
DeepSeek-R1 is impressive, but ChatGPT’s product advantage is far from over https://venturebeat.com/ai/calm-dow...-chatgpts-product-advantage-is-far-from-over/ Just a week ago — on January 20, 2025 — Chinese AI startup DeepSeek unleashed a new, open-source AI model called R1 that might have initially been mistaken for one of the ever-growing masses of nearly interchangeable rivals that have sprung up since OpenAI debuted ChatGPT (powered by its own GPT-3.5 model, initially) more than two years ago. But that quickly proved unfounded, as DeepSeek’s mobile app has in that short time rocketed up the charts of the Apple App Store in the U.S. to dethrone ChatGPT for the number one spot and caused a massive market correction as investors dumped stock in formerly hot computer chip makers such as Nvidia, whose graphics processing units (GPUs) have been in high demand for use in massive superclusters to train new AI models and serve them up to customers on an ongoing basis (a modality known as “inference.”) Venture capitalist Marc Andreessen, echoing sentiments of other tech workers, wrote on the social network X U.S. President Donald Trump said on Monday that Chinese startup DeepSeek's technology should act as spur for American companies; at the same time, he provided incorrect/baseless claim, saying that "it was good that companies in China have come up with a cheaper, faster method of artificial intelligence." the report from China did say the the AI performed by DeepSeek was cheaper and faster, without any details. Trump did not believe China's report on the origin of COVID, why should he believe in China's report on AI from Deepseek, 深 度 求 索 !!
more from https://venturebeat.com/ai/calm-dow...-chatgpts-product-advantage-is-far-from-over/ DeepSeek-R1 was trained on synthetic data questions and answers and specifically, according to the paper released by its researchers, on the supervised fine-tuned “dataset of DeepSeek-V3,” It seems pretty clear-cut to say that without GPT-4o to provide this data, and without OpenAI’s own release of the first commercial reasoning model o1 back in September 2024, which created the category, DeepSeek-R1 would almost certainly not exist. Furthermore, OpenAI’s success required vast amounts of GPU resources, paving the way for breakthroughs that DeepSeek has undoubtedly benefited from. The current investor panic about U.S. chip and AI companies feels premature and overblown. While DeepSeek-R1 has been impressive, it lacks several features that make ChatGPT a more robust and versatile tool today. DeepSeek-R1’s accomplishments are impressive and signal a promising shift in the global AI landscape. However, it’s crucial to keep the excitement in check. For now, ChatGPT remains the better-rounded and more capable product, offering a suite of features that DeepSeek simply cannot match.
invalid comparison, as Sputnik was the first successful space launch. in the realm of AI, OpenAI has been the first to build automonous systems; some 2 years later, Deepseek just reported that it can replicate OpenAI's accomplishments more apt comparison would be Fuji films supplanting Kodak Japanese copiers supplanting Xerox machines Excel supplanting Lotus 123 mini computers supplanting DEC mainframe computers etc
They spent 1/100th of the price to make something just as good. OpenAI value probably dropped 90%.Plus they just released the weights and paper.
The funny thing with the damage control posted here (did a grumpy lil someone lose a grip of paper money today?) is that the industry already knows but are still put on notice in spite of it. It's still a big deal for many different reasons...
The United States and Western powers panicking because they realized their plan of turning Asia into a cheap labor dumping ground shot themselves in the face and realizing they underestimated the capability of the Chinese people because they thought China would perpetually be some peasant serfdom and now are trying to go all protectionist is.... Kinda funny? Like come on this has to make people kinds chuckle a bit right? The meme coin president has surrounded himself with the most advanced sophisticated grifters on the planet with the likes of Sam Altman and Musk. Let's see how they fight for the working class.
https://www.wsj.com/opinion/deepsee...ald-trump-big-tech-9bb91f78?mod=hp_opin_pos_0 The DeepSeek AI Freakout The Chinese startup’s model stuns Big Tech—and Wall Street—with its capability and cost. By The Editorial Board Jan. 27, 2025 at 6:10 pm ET Who saw that coming? Not Wall Street, which sold off tech stocks on Monday after the weekend news that a highly sophisticated Chinese AI model, DeepSeek, rivals Big Tech-built systems but cost a fraction to develop. The implications are likely to be far-reaching, and not merely in equities. The tech-heavy Nasdaq fell 3.1%, driven by a 16.9% dive in Nvidia shares. Nvidia dominates the market in advanced AI chips. Its stock had surged more than 10-fold since early 2023—achieving a more than $3.3 trillion market valuation until Monday—as tech giants announced hefty outlays on AI. Enter DeepSeek, which last week released a new R1 model that claims to be as advanced as OpenAI’s on math, code and reasoning tasks. Tech gurus who inspected the model agreed. One economist asked R1 how much Donald Trump’s proposed 25% tariffs will affect Canada’s GDP, and it spit back an answer close to that of a major bank’s estimate in 12 seconds. Along with the detailed steps R1 used to get to the answer. More startling, DeepSeek required far fewer chips to train than other advanced AI models and thus cost only an estimated $5.6 million to develop. Other advanced models cost in the neighborhood of $1 billion. Venture capitalist Marc Andreessen called it “AI’s Sputnik moment,” and he may be right. DeepSeek is challenging assumptions about the computing power and spending needed for AI advances. OpenAI, Oracle and SoftBank last week made headlines when they announced a joint venture, Stargate, to invest up to $500 billion in building out AI infrastructure. Microsoft plans to spend $80 billion on AI data centers this year. CEO Mark Zuckerberg on Friday said Meta would spend about $65 billion on AI projects this year and build a data center “so large that it would cover a significant part of Manhattan.” Meta expects to have 1.3 million advanced chips by the end of this year. DeepSeek’s model reportedly required as few as 10,000 to develop. DeepSeek’s breakthrough means these tech giants may not have to spend as much to train their AI models. But it also means these firms, notably Google’s DeepMind, might lose their first-mover, technological edge. Google shares fell 4% on Monday. DeepSeek’s model is open-source, meaning that other developers can inspect and fiddle with its code and build their own applications with it. This could help give more small businesses access to AI tools at a fraction of the cost of closed-source models like OpenAI and Anthropic, which Amazon has backed. There are advantages to such closed-source systems, especially for privacy and national security. But open-source can foster more collaboration and experimentation. It’s notable that DeepSeek is a startup founded by Liang Wenfeng, a Chinese hedge fund trader. Americans think of China’s economy as run top-down, and much of it is. But its growth over the last few decades, especially in tech, has been spurred by entrepreneurs. Alibaba, Tencent and ByteDance were all once startups that now rival U.S. tech giants. This is another reason for the U.S. to avoid the trap of thinking it must imitate Chinese industrial policy to succeed in the AI race. A bipartisan Senate AI report last spring called for Congress to pass $32 billion a year in “emergency” spending for non-defense AI, supposedly to better compete with China. What a waste of money that would be. *** DeepSeek is vindicating President Trump’s decision to rescind a Biden executive order that gave government far too much control over AI. Companies developing AI models that pose a “serious risk” to national security, economic security, or public health and safety would have had to notify regulators when training their models and share the results of “red-team safety tests.” Mr. Biden said such tests are needed to eliminate biases, limitations and errors. But open-source models allow the public to review and test systems. Some have pointed out that DeepSeek doesn’t answer questions on subjects that are politically sensitive to Beijing. DeepSeek should also cause Republicans in Washington to rethink their antitrust obsessions with big tech. Bureaucrats aren’t capable of overseeing thousands of AI models, and more regulation would slow innovation and make it harder for U.S. companies to compete with China. As DeepSeek shows, it’s possible for a David to compete with the Goliaths. Let a thousand American AI flowers bloom. Appeared in the January 28, 2025, print edition as 'The DeepSeek AI Freakout'.
chief market strategist at B. Riley Wealth Management, urged investors to take DeepSeek’s claims “with a grain of salt” because there is no concrete proof to back them up.
Take the claims about how much they spent or what gpus they used with a grain of salt. It's half showmanship and half geopolitics. But everything behind their results is in their whitepaper and you can't fake open source. If you like AI or "progress", I think it's a good development. Then again, I don't know if there really any "good guys" here. Just rich ass folks wanting to make even more money and take even more control.
Real collaboration and competition always win. Isolation, on the other hand, loses in the long run. With that said... if it's open source and accessible, what’s stopping META, OpenAI, and others from studying their "formula" or "secret OPEN source" and leveraging it with 100x more compute power? (Nvidia stock took a big hit yesterday but bounced back over 10% today - chips are still essential for training, lol. And if you can have 1,000 DeepSeeks instead of 5 big players, you're going to need more chips than ever). How Chinese AI Startup DeepSeek Made a Model that Rivals OpenAI | WIRED
The training models are not open source, only the codes. How they tweaked the severs, load balancing, CPU management etc are not open to public. That is why TSMC can make 2nm chips while intel can only make 10 nm chips with the same equipment.
After that, what stops those geeks from China from doing it again? It's not the chips' computing power alone. It's the efficiency of their code as they were forced to improve in that aspect because of the chip restrictions. Surely, the earlier AI makers can match or maybe even surpass the "brain power" part. Welcome to the AI arms race.
