Will the Chinese DeepSeek AI upset the AI/ML race?
- Thread starter Arthur Hanson
- Start date
I agree with the idea that by making AI cheaper it will ultimately mean AI will become more widely deployed. The question is which businesses will benefit and which ones will be disrupted. When the Japanese started making more fuel efficient cars, gas demand may have rebounded... but that wasn't to the benefit of the US auto industry.
DeepSeek will be sensitive to topics China is sensitive to, ChatGPT is also sensitive to certain topics. Either way you are consuming propaganda by using any of these models.
I showed this to my supervisor. It seems OpenAI's RLHF incorporated elements of DEI.
Attachments
If your supervisor is in the US, that’s a wrong answer. Protected classes don’t stem from DEI, but rather the Equal Employment Opportunity Act of 1972.
Thank you for pointing that out. That is the reason I put a question mark there.
Is that law clause related to the following announcement?
Ending Illegal Discrimination And Restoring Merit-Based Opportunity
By the authority vested in me as President by the Constitution and the laws of the United States of America, it is hereby ordered:Section 1. Purpose.
www.whitehouse.gov
Not sure of your question. But Executive Orders don’t change legislation.Thank you for pointing that out. That is the reason I put a question mark there.
Is that law clause related to the following announcement?
Ending Illegal Discrimination And Restoring Merit-Based Opportunity
By the authority vested in me as President by the Constitution and the laws of the United States of America, it is hereby ordered:Section 1. Purpose.www.whitehouse.gov
That means the prompt explicitly disparages some group of people, based on being in some class or category.
My understanding is that all classes should be equal and merit-based. I didn’t get the reason for embedding preferential bias when reading that technical material.
Right, a protected class would be for example, gender, in itself (not female gender).
"I’ve said it before, and I’ll say it again—OpenAI may well become the WeWork of AI."
fortune.com
Gary Marcus is a well-known AI researcher and perhaps somewhat outspoken. My supervisor thinks he sometimes seems a bit negative toward current AI developments. Comparing OpenAI to WeWork is certainly interesting, but the fact that Masayoshi Son is once again involved in these dynamics is even more intriguing.
en.wikipedia.org
China just redefined the global AI race—with massive implications for OpenAI, Nvidia, and foreign policy
Silicon Valley’s initial advantage in LLMs evaporated quickly despite export controls, writes AI expert Gary Marcus.
Gary Marcus is a well-known AI researcher and perhaps somewhat outspoken. My supervisor thinks he sometimes seems a bit negative toward current AI developments. Comparing OpenAI to WeWork is certainly interesting, but the fact that Masayoshi Son is once again involved in these dynamics is even more intriguing.
Gary Marcus - Wikipedia
DeepSeek-V3 Technical Report
We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeek MoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks. The model checkpoints are available at https://github.com/deepseek-ai/DeepSeek-V3.
DeepSeek-V3 Technical Report
We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeek MoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks. The model checkpoints are available at https://github.com/deepseek-ai/DeepSeek-V3.
DeepSeek-V3 Technical Report
Wow! 
Nvidia’s $589 Billion DeepSeek Rout Is Largest in Market History
(Bloomberg) -- Nvidia Corp.’s plunge, fueled by investor concern about Chinese artificial-intelligence startup DeepSeek, erased a record amount of stock-market value from the world’s largest company.
Nvidia shares tumbled 17% Monday, the biggest drop since March 2020, erasing $589 billion from the company’s market capitalization. That eclipsed the previous record — a 9% drop in September that wiped out about $279 billion in value — and was the biggest in US stock-market history.
The drop rippled through the rest of the market due to how much weight Nvidia has in major indexes. Including Monday’s slump, Nvidia selloffs have caused eight of the top ten biggest one-day drops in the S&P 500 Index, based on market value, according to data compiled by Bloomberg. The S&P 500 fell 1.5% Monday and the Nasdaq 100 tumbled nearly 3%.
The semiconductor maker led a broader selloff in technology stocks after DeepSeek’s low-cost approach reignited concerns that big US companies have poured too much money into developing artificial intelligence. The Chinese firm appears to provide a comparable performance at a fraction of the price.
All About DeepSeek and Its Lower-Cost AI Model: QuickTake
The latest AI model of DeepSeek, released last week, is widely seen as competitive with those of OpenAI and Meta Platforms Inc. The open-sourced product was founded by quant-fund chief Liang Wenfeng and is now at the top of Apple Inc.’s App Store rankings.
“Concerns have immediately emerged that it could be a disruptor to the current AI business model, which relies on high end chips and extensive computing power and hence energy,” Jefferies analysts said in a note to clients.
Nvidia has been the biggest beneficiary of the influx in spending on AI because they design semiconductors used in the technology. While that heavy spending looks poised to continue, investors may grow wary of rewarding companies that aren’t showing a sufficient return on the investment.
Meta announced plans on Friday to boost capital expenditures on AI projects this year by about half to as much as $65 billion, sending its shares to a record high. That came on the heels of OpenAI, SoftBank Group Corp. and Oracle Corp. announcing a $100 billion joint venture called Stargate to build out data centers and AI infrastructure projects around the US.
In a bid to stall China’s progress in AI, the US has banned the export of advanced semiconductor technologies to the country and is limiting sales of advanced Nvidia AI chips to others. But DeepSeek’s progress suggests Chinese AI engineers have found a way to work around the export bans, focusing on greater efficiency with limited resources.
Nvidia said in a statement Monday that DeepSeek’s model is an “excellent AI advancement” and indicated that the Chinese company didn’t violate US restrictions that limit access to advanced US chips in creating its technology. It also added that inference, or the work of running AI models, requires “requires significant numbers of Nvidia GPUs and high-performance networking.”
Nvidia’s $589 Billion DeepSeek Rout Is Largest in Market History
(Bloomberg) -- Nvidia Corp.’s plunge, fueled by investor concern about Chinese artificial-intelligence startup DeepSeek, erased a record amount of stock-market value from the world’s largest company.
Nvidia shares tumbled 17% Monday, the biggest drop since March 2020, erasing $589 billion from the company’s market capitalization. That eclipsed the previous record — a 9% drop in September that wiped out about $279 billion in value — and was the biggest in US stock-market history.
The drop rippled through the rest of the market due to how much weight Nvidia has in major indexes. Including Monday’s slump, Nvidia selloffs have caused eight of the top ten biggest one-day drops in the S&P 500 Index, based on market value, according to data compiled by Bloomberg. The S&P 500 fell 1.5% Monday and the Nasdaq 100 tumbled nearly 3%.
The semiconductor maker led a broader selloff in technology stocks after DeepSeek’s low-cost approach reignited concerns that big US companies have poured too much money into developing artificial intelligence. The Chinese firm appears to provide a comparable performance at a fraction of the price.
All About DeepSeek and Its Lower-Cost AI Model: QuickTake
The latest AI model of DeepSeek, released last week, is widely seen as competitive with those of OpenAI and Meta Platforms Inc. The open-sourced product was founded by quant-fund chief Liang Wenfeng and is now at the top of Apple Inc.’s App Store rankings.
“Concerns have immediately emerged that it could be a disruptor to the current AI business model, which relies on high end chips and extensive computing power and hence energy,” Jefferies analysts said in a note to clients.
Nvidia has been the biggest beneficiary of the influx in spending on AI because they design semiconductors used in the technology. While that heavy spending looks poised to continue, investors may grow wary of rewarding companies that aren’t showing a sufficient return on the investment.
Meta announced plans on Friday to boost capital expenditures on AI projects this year by about half to as much as $65 billion, sending its shares to a record high. That came on the heels of OpenAI, SoftBank Group Corp. and Oracle Corp. announcing a $100 billion joint venture called Stargate to build out data centers and AI infrastructure projects around the US.
In a bid to stall China’s progress in AI, the US has banned the export of advanced semiconductor technologies to the country and is limiting sales of advanced Nvidia AI chips to others. But DeepSeek’s progress suggests Chinese AI engineers have found a way to work around the export bans, focusing on greater efficiency with limited resources.
Nvidia said in a statement Monday that DeepSeek’s model is an “excellent AI advancement” and indicated that the Chinese company didn’t violate US restrictions that limit access to advanced US chips in creating its technology. It also added that inference, or the work of running AI models, requires “requires significant numbers of Nvidia GPUs and high-performance networking.”
Last edited:
Intel's former CEO says the market is getting DeepSeek wrong after AI chip stock rout
Pat Gelsinger
@PGelsinger
Wisdom is learning the lessons we thought we already knew. DeepSeek reminds us of three important learnings from computing history:1) Computing obeys the gas law. Making it dramatically cheaper will expand the market for it. The markets are getting it wrong, this will make AI much more broadly deployed.2) Engineering is about constraints. The Chinese engineers had limited resources, and they had to find creative solutions.3) Open Wins. DeepSeek will help reset the increasingly closed world of foundational AI model work. Thank you DeepSeek team.
Wow. the former CEO of THE x86 company endorsing open source. Hm...
siliconbruh999
Well-known member
Wow!
(Bloomberg) -- Nvidia Corp.’s plunge, fueled by investor concern about Chinese artificial-intelligence startup DeepSeek, erased a record amount of stock-market value from the world’s largest company.
Nvidia shares tumbled 17% Monday, the biggest drop since March 2020, erasing $589 billion from the company’s market capitalization. That eclipsed the previous record — a 9% drop in September that wiped out about $279 billion in value — and was the biggest in US stock-market history.
The drop rippled through the rest of the market due to how much weight Nvidia has in major indexes. Including Monday’s slump, Nvidia selloffs have caused eight of the top ten biggest one-day drops in the S&P 500 Index, based on market value, according to data compiled by Bloomberg. The S&P 500 fell 1.5% Monday and the Nasdaq 100 tumbled nearly 3%.
The semiconductor maker led a broader selloff in technology stocks after DeepSeek’s low-cost approach reignited concerns that big US companies have poured too much money into developing artificial intelligence. The Chinese firm appears to provide a comparable performance at a fraction of the price.
All About DeepSeek and Its Lower-Cost AI Model: QuickTake
The latest AI model of DeepSeek, released last week, is widely seen as competitive with those of OpenAI and Meta Platforms Inc. The open-sourced product was founded by quant-fund chief Liang Wenfeng and is now at the top of Apple Inc.’s App Store rankings.
“Concerns have immediately emerged that it could be a disruptor to the current AI business model, which relies on high end chips and extensive computing power and hence energy,” Jefferies analysts said in a note to clients.
Nvidia has been the biggest beneficiary of the influx in spending on AI because they design semiconductors used in the technology. While that heavy spending looks poised to continue, investors may grow wary of rewarding companies that aren’t showing a sufficient return on the investment.
Meta announced plans on Friday to boost capital expenditures on AI projects this year by about half to as much as $65 billion, sending its shares to a record high. That came on the heels of OpenAI, SoftBank Group Corp. and Oracle Corp. announcing a $100 billion joint venture called Stargate to build out data centers and AI infrastructure projects around the US.
In a bid to stall China’s progress in AI, the US has banned the export of advanced semiconductor technologies to the country and is limiting sales of advanced Nvidia AI chips to others. But DeepSeek’s progress suggests Chinese AI engineers have found a way to work around the export bans, focusing on greater efficiency with limited resources.
Nvidia said in a statement Monday that DeepSeek’s model is an “excellent AI advancement” and indicated that the Chinese company didn’t violate US restrictions that limit access to advanced US chips in creating its technology. It also added that inference, or the work of running AI models, requires “requires significant numbers of Nvidia GPUs and high-performance networking.”
Still 10%+ higher than they were in September a long long 4 months ago.
If they drop to their historic levels maybe something for the speculators to worry about