DeepSeek, China's AI model: News & Discussion

Elon Musk is not the *truth*.

Again you are missing the point completely.

Your point is efficiency, mine is access to the most advanced chips/compute.

Compute/chips will win out in the end and Deepseeks founder acknowledges that above. Without it you can’t achieve the leading edge models and ultimately ASI.
 
Your point is efficiency, mine is access to the most advanced chips/compute.

Compute/chips will win out in the end and Deepseeks founder acknowledges that above. Without it you can’t achieve the leading edge models and ultimately ASI.


20x less cost is not efficiency but a revolutionary model to train AI.

Anyway Huawei is now just 2 years away from leading Nvidia tech with Ascend 910C and their revolutionary model will allow much better Chatbots than anything Google or OpenAI can come up with
 
20x less cost is not efficiency but a revolutionary model to train AI.

Anyway Huawei is now just 2 years away from leading Nvidia tech with Ascend 910C and their revolutionary model will allow much better Chatbots than anything Google or OpenAI can come up with

You’re coping, NVIDIAs moat is significant, and SMIC can’t even mass produce A100 level chip performance.

China is only a fast follower at this point. The US remains at the frontier.
 
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.


Deepseek admits to using H100s…..illegally 🤣
 
You’re coping, NVIDIAs moat is significant, and SMIC can’t even mass produce A100 level chip performance.

China is only a fast follower at this point. The US remains at the frontier.


It does not matter if SMIC starts off with low yields for Ascend 910C(H100 performance) as these AI Chips are manufactured in the hundreds of thousands and not millions every year.

Huawei/SMIC are not building GPUs to satisfy millions of gamers who want to play CyberPunk 2077.

The yields will go up as they managed to get to 50% with Ascend 910B - A100 level performance.

Due to the much more efficient "training" process of Deepseek they would still need much less money to develop their chatbots than either Gemini or OpenAI.
 
In the end, whoever has the market has the game.
 

UK will not be able to resist China's tech dominance​

Jan. 28 2025

China's success in technology has not come out of thin air, even given the unlikely origins of the DeepSeek deep shock.

The obscure Hangzhou hedge fund that coded a ChatGPT competitor as a side project it claims cost just $6m emerges from a concerted effort to invest in future generations of technology.

This is not an accident. This is policy.

The raw materials of artificial intelligence (AI) are microchips, science PhDs and data. On the latter two, China might be ahead already.

There are on average more than 6,000 PhDs in STEM subjects (science, technology, engineering and mathematics) coming out of Chinese universities every month. In the US it is more like 2,000-3,000, in the UK it is 1,500.

In terms of patents generally, more are being filed in China than in the rest of the world put together. In 2023 China filed 1.7 million patents, against 600,000 in the US. Two decades earlier China had a third of the patents filed by the US, a quarter of Japan's and was well behind South Korea and Europe.

While there are some questions about the quality, on some measures China now exceeds the US on what is known as "citation-weighted" patents too, which adjusts for how often new scientific papers are referred to.

Chinese lithium-ion electric batteries now cost per kWh about a seventh of what they cost a decade ago. DeepSeek is doing in AI exactly what China has done elsewhere.

While the impact of this was most visible in electric vehicles (EVs), where China is now the world's biggest exporter, having cornered the supply chains and the science for battery technology, it stretches well beyond.

Even in auto the Chinese manufacturers are now pushing the concept of "electric intelligent vehicles", in which conventional carmakers cannot compete, especially on software development.

China's consumer electronics companies are shifting into car manufacturing, with "dark factories" operated 24/7 by armies of AI-powered robots, now also increasingly made in China.

This innovation is partly of necessity. China does not have indigenous fossil fuels, and is electrifying at an astonishing rate, and is referred to by some researchers as an "electro state". It now files three-quarters of all clean tech patents, versus a twentieth at the start of the century.

Last year the US National Science Board asserted China's objective of being the world's leading science and engineering nation was on the verge of being achieved. "We already see this in artificial intelligence, where China out publishes us, has more patents, and produces more students than the United States," they wrote.

Delegates who accompanied the UK Chancellor Rachel Reeves to China earlier this month marvelled at how the Beijing air had been cleaned up, and indigenous electric cars were everywhere. Another UK CEO told me of a visit to Huawei's Oxbridge-style campus complete with spires and bridges, and its own Tube line, purely for its scientists.

Clearly, however, there are concerns about censorship, democracy and security. One of the drivers of the Chinese AI industry has been access to extraordinary amounts of data, which is more difficult to get hold of in the West.

If the US Congress was sufficiently concerned about TikTok to ban it, then surely a table-topping AI program could be highly problematic. President Trump's argument this morning was that DeepSeek's innovation was "positive" and "a wake-up call". China has not been prominent as the first target of Trump tariffs.

There is still an obvious balancing act for the UK government here. But this sort of innovation and its impact on the world was exactly why the chancellor visited Beijing a fortnight ago.

She said at the time she wanted a long-term relationship with China that is "squarely in our national interest" with the visit part of a "commitment to explore deeper economic co-operation" between Prime Minister Sir Keir Starmer and President Xi.

Other European nations such as Spain have encouraged China not just to set up factories but to transfer its advanced battery technology, for example, into Europe.

The West wants China to make its T-shirts, its tables, its TVs and EVs. But could that really now stretch into DeepSeek data-hungry AI models too? It is a deep tremor, not just for tech, but for economics and geopolitics as well.
 
It does not matter if SMIC starts off with low yields for Ascend 910C(H100 performance) as these AI Chips are manufactured in the hundreds of thousands and not millions every year.

Huawei/SMIC are not building GPUs to satisfy millions of gamers who want to play CyberPunk 2077.

The yields will go up as they managed to get to 50% with Ascend 910B - A100 level performance.

Due to the much more efficient "training" process of Deepseek they would still need much less money to develop their chatbots than either Gemini or OpenAI.

China is only fast following at this point. They attempt to match US capability at a cheaper price, but still no leading edge capability in the hardware or LLM space.

Efficiency won’t win in the end. Funding is not an issue for US AI frontier companies, and Trump has made that clear.

Chips/compute will ultimately decide who makes it to ASI first.
 
If only China can translate its AI success into the biotechnology / Medical field, offering / inventing / discovering new treatments and medications for untreatable diseases, that would be revolutionary.
 
China is only fast following at this point. They attempt to match US capability at a cheaper price, but still no leading edge capability in the hardware or LLM space.

Efficiency won’t win in the end. Funding is not an issue for US AI frontier companies, and Trump has made that clear.

Chips/compute will ultimately decide who makes it to ASI first.

They have destroyed America’s AI companies.

Chips is next.
 
China is only fast following at this point. They attempt to match US capability at a cheaper price, but still no leading edge capability in the hardware or LLM space.

Efficiency won’t win in the end. Funding is not an issue for US AI frontier companies, and Trump has made that clear.

Chips/compute will ultimately decide who makes it to ASI first.


Hold on.

So now you are saying that US has unlimited money and manufacturing capabilty to exceed the 20x superior efficiency of China as proven by Deepseek?

Once those Ascend 910Cs are being pumped out in the 10s of thousands per month by SMIC later on this year, expect another "surprise" soon afterwards from Deepseek and other Chinese AI companies that shows up Google and OpenAI for the mediocrity they really are.
 
China is only fast following at this point. They attempt to match US capability at a cheaper price, but still no leading edge capability in the hardware or LLM space.

Efficiency won’t win in the end. Funding is not an issue for US AI frontier companies, and Trump has made that clear.

Chips/compute will ultimately decide who makes it to ASI first.
DeepSeek spent 6 millions dollar to prove as a lie the USA AI power consumption and that moved billions in stocks markets when investments firms saw the deception.

What can do USA tech companies with 6 millions dollar?

How can a 6 millions USD company beat a billions dollar company? Because the billions company is a lie and has no real rivals due to regulations and a fake captive market.
 
Hold on.

So now you are saying that US has unlimited money and manufacturing capabilty to exceed the 20x superior efficiency of China as proven by Deepseek?

Once those Ascend 910Cs are being pumped out in the 10s of thousands per month by SMIC later on this year, expect another "surprise" soon afterwards from Deepseek and other Chinese AI companies that shows up Google and OpenAI for the mediocrity they really are.
from what I see the chinese use a different approach to problems than the americans.
let’s say you have a math equation you can do brute force or you do the smart way.
brute force requires lots of resources, lots of money. smart way uses instincts and learned experiences.
given a, what is b, what is c?
a, b, c belong to N.

1738090418113.png
 

Users who are viewing this thread

Pakistan Defence Latest

Latest Posts

Back
Top