‘Two sessions’ 2024: China’s top advisory body told AI gap with US is widening

Hamartia Antidote

Elite Member
Joined
Nov 17, 2013
Messages
46,559
Reaction score
26,674
Reputation
613.4
Country of Origin
Country of Residence

  • Head of major state-owned chip maker says his firm has ‘long way’ to go to catch up with American AI companies
  • Amid US tech curbs, CPPCC members told China will fall further behind if no decisive measures are taken
d3657c09-53c2-40b1-a13d-c19206c2dfc3_30656b7b.jpg

Zeng Yi, head of a state-run tech company, told the CPPCC that as new developments in AI emerge, China is at risk of seeing an even wider gap with US tech leaders if no decisive and groundbreaking measures are taken.

At a major state-owned chip company in the Chinese technology hub of Shenzhen, there is growing anxiety about a “widening gap” in artificial intelligence (AI).
During a sideline meeting on Tuesday in Beijing at the Chinese People’s Political Consultative Conference (CPPCC) – the country’s top political advisory body – Zeng Yi, a delegate who heads the Shenzhen-based China Electronics Corporation, said his firm has a “long way” to go to catch up with the United States.

“Objectively speaking, despite the great efforts we have been making, our difference [with the US] is still huge,” Zeng told the gathering of a few dozen representatives of the science and technology community.

At the annual “two sessions” meetings, Beijing has continued to highlight self-reliance in science and technology as the key to transforming the Chinese economy at a time when development is under pressure from increasing US technology curbs. But Zeng, who said the work his company is doing is “at the forefront of China-US science and technology competition”, said the country is still falling behind.

“In a certain sense, as new developments in AI emerge exponentially, if no decisive and groundbreaking measures are taken, we are at risk of seeing an even wider gap,” he said.


“Problems in the development of information technology cannot be solved by creating applicable scenarios, nor through breakthroughs in specific technologies … it is not even a problem about talent and basic research,” Zeng said.

“It is about many things from all aspects, we are all very anxious.”

Zeng leads a company that develops national cyber and information security technology, chip research and design, semiconductor equipment, advanced manufacturing processes, and operating systems.

China has relied heavily for its technology development on US chip imports and other crucial Western equipment. But US sanctions have constricted Chinese access to key tools such as advanced graphics processing units from Nvidia, the world’s leading AI chip designer, which had up to 90 per cent of market share in China’s AI chip market.

Now, both state-owned and private companies are under pressure to ramp up the country’s domestic technology development.

While China’s AI industry was generally seen as more competitive in the varied tech rivalry between the two countries, the latest launch of OpenAI’s Sora and ChatGPT has raised questions about China’s progress in catching up with the US.

To cope, Chinese tech companies, which have insisted that chip restrictions would not affect AI their development in the short term, have relied on their existing inventories, or have turned to domestic AI chip makers.

But after Sora was launched last month, the State Council’s State-owned Assets Supervision and Administration Commission, which manages companies like Zeng’s, urged firms under direct control of the central government to “embrace the profound changes brought about by AI”.
 
Both India and China have entirely designed their school system to jobs, not research or inventions type of innovation. By and large therefore they have hit their limits at process and design facades. The higher education environment of the US, Canada and western europe still have the wide moat.
 

  • Head of major state-owned chip maker says his firm has ‘long way’ to go to catch up with American AI companies
  • Amid US tech curbs, CPPCC members told China will fall further behind if no decisive measures are taken
d3657c09-53c2-40b1-a13d-c19206c2dfc3_30656b7b.jpg

Zeng Yi, head of a state-run tech company, told the CPPCC that as new developments in AI emerge, China is at risk of seeing an even wider gap with US tech leaders if no decisive and groundbreaking measures are taken.

At a major state-owned chip company in the Chinese technology hub of Shenzhen, there is growing anxiety about a “widening gap” in artificial intelligence (AI).
During a sideline meeting on Tuesday in Beijing at the Chinese People’s Political Consultative Conference (CPPCC) – the country’s top political advisory body – Zeng Yi, a delegate who heads the Shenzhen-based China Electronics Corporation, said his firm has a “long way” to go to catch up with the United States.

“Objectively speaking, despite the great efforts we have been making, our difference [with the US] is still huge,” Zeng told the gathering of a few dozen representatives of the science and technology community.

At the annual “two sessions” meetings, Beijing has continued to highlight self-reliance in science and technology as the key to transforming the Chinese economy at a time when development is under pressure from increasing US technology curbs. But Zeng, who said the work his company is doing is “at the forefront of China-US science and technology competition”, said the country is still falling behind.

“In a certain sense, as new developments in AI emerge exponentially, if no decisive and groundbreaking measures are taken, we are at risk of seeing an even wider gap,” he said.


“Problems in the development of information technology cannot be solved by creating applicable scenarios, nor through breakthroughs in specific technologies … it is not even a problem about talent and basic research,” Zeng said.

“It is about many things from all aspects, we are all very anxious.”

Zeng leads a company that develops national cyber and information security technology, chip research and design, semiconductor equipment, advanced manufacturing processes, and operating systems.

China has relied heavily for its technology development on US chip imports and other crucial Western equipment. But US sanctions have constricted Chinese access to key tools such as advanced graphics processing units from Nvidia, the world’s leading AI chip designer, which had up to 90 per cent of market share in China’s AI chip market.

Now, both state-owned and private companies are under pressure to ramp up the country’s domestic technology development.

While China’s AI industry was generally seen as more competitive in the varied tech rivalry between the two countries, the latest launch of OpenAI’s Sora and ChatGPT has raised questions about China’s progress in catching up with the US.

To cope, Chinese tech companies, which have insisted that chip restrictions would not affect AI their development in the short term, have relied on their existing inventories, or have turned to domestic AI chip makers.

But after Sora was launched last month, the State Council’s State-owned Assets Supervision and Administration Commission, which manages companies like Zeng’s, urged firms under direct control of the central government to “embrace the profound changes brought about by AI”.
This man said a lot of nonsenses. Nvidia chips and OpenAI again and again. Every boday knows that. I don't know why the government wasted money on the useless forum.

GPU chips is not a big problem. Huawei has sold 300k GPU chips already. Not as good as Nvidia's. But it's just cost problem. The government can provide subsidy to Huawei chips. Like it did to EV.

As for OpenAI, as I said, both ChatGPT and Sora are not so useful for promoting a country hard power. China's "AI serves industry" strategy is the right pass for AI development. Time will prove my view is correct.
 
Last edited:
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Zeng Yi
 
Last edited:

China offers AI computing ‘vouchers’ to its underpowered start-ups​


Chinese tech giants are hogging scarce cloud resources to develop generative AI models as US restrictions bite

ftcms%3A294201ed-d363-4399-83e0-25982c09f8d5

China’s AI start-ups are getting subsidised access to cloud services to train their products, but big tech companies are giving themselves priority when it comes to using Nvidia’s latest AI processors

China is taking steps to level the playing field for its artificial intelligence start-ups, as the country’s tech giants hog AI-training computing resources already squeezed by US chip restrictions.

At least 17 city governments, including the largest, Shanghai, have pledged to provide “computing vouchers” to subsidise AI start-ups facing rising data centre costs as supplies of crucial chips become more scarce.

The vouchers will be typically worth the equivalent of between $140,000 and $280,000, according to official announcements. They can be used for time in AI data centres to train and run the companies’ large language models (LLMs) that understand and generate natural language, and other content, to perform a wide array of tasks.

Industry insiders said the move to help AI start-ups pay for computing costs comes after internet companies with cloud computing services cancelled contracts, as stricter US controls prompted them to “hog the GPUs [graphics processing units] for themselves”, according to one AI founder.

Internet giants Alibaba, Tencent and ByteDance have taken measures to limit the rental of Nvidia’s GPUs and have reserved the majority of these stockpiled AI processors for internal use and important clients, according to multiple sources familiar with the situation.

Over the past two years, the Biden administration has tightened China’s access to crucial AI chips, leading companies to stockpile ahead of bans, repurpose Nvidia gaming chips or resort to the black market.

Alibaba Cloud has distributed the majority of its advanced GPUs to business units within the Alibaba Group, said two employees of the cloud business. Acquiring high-performance chips has become an “arduous task”, said one manager.

ByteDance declined to comment. Tencent and Alibaba Cloud did not respond to requests for comment.Subsidies will only tackle part of the problem, according to analysts. “The voucher is helpful to address the cost barrier, but it won’t help with the scarcity of the resources,” said Charlie Dai, an analyst at US research company Forrester.

A person close to regulatory authorities said Beijing would soon roll out a subsidy programme for AI groups that were using domestic chips, as it tries to ramp up efforts to displace foreign components.China has also been creating an alternative to Big Tech’s data centres and cloud services. A network of state-run data centres and online platforms has been built over the past year where AI companies can rent computing power.

“Chinese governments are dedicated to improving resource allocation efficiency,” said Dai, referring to Beijing’s additional plan to build 10 data centre clusters. The so-called “East Data West Computing” project is “essential for lowering energy resource consumption and powering AI workloads in an orchestrated approach,” he added. One government official who works on data centre construction said the vouchers would reduce AI companies’ computing costs by around 40 to 50 per cent if they opted for government-run data centres.They added that while this will “alleviate financial pressure on start-ups” in the short term, there were still stringent requirements for applicants, including a minimum revenue threshold or being part of a government-sponsored research project.

Beijing has approved at least 40 LLMs for public use, as regulators accelerate adoption of AI technology while maintaining strict oversight over its application.

This includes centralising the distribution of computing power using state-run trading platforms that broker access to cloud computing resources. Last year, Suzhou, a city west of Shanghai, set up a trading platform to “co-ordinate the city’s computing resources”, according to its website.

Traces of China’s underground chip trade can be found on such brokering sites. The Suzhou one, which partners with AliCloud and Tencent Cloud, is advertising computing clusters that include advanced Nvidia H100 chips, which the US has always banned from sale to China.Additional reporting by Wenjie Ding in Beijing
 
A good thing, facing shortcomings calmly, science and technology should already have various aspects of information available to the public.
 

Funds Pour Into Chinese AI Startups Challenging OpenAI’s Sora in Video Generation​


In February, the American tech giant OpenAI, known for developing ChatGPT, unveiled a text-to-video model named Sora

In February, the American tech giant OpenAI, known for developing ChatGPT, unveiled a text-to-video model named Sora

Venture capital investors are pouring hundreds of millions of yuan into China’s burgeoning video generation startups, betting on their potential in the artificial intelligence race to rival OpenAI’s text-to-video model Sora.


This week, two leading Chinese generative AI startups announced the successful closure of their latest funding rounds, with each securing more than 100 million yuan ($14 million). Both aim to catch up with Sora in the coming months.
 
The different between USA and China on AI.

USA has a plan for AI, they have goal that has been imagined and shaped since decades ago.

While China is just following what USA does, in hope to surpass USA.


When USA is digging a hole.

China will do the same, but more holes, deeper and wider holes.

China seems to be a winner in this race, but it turns out USA is digging a hole for planting a tree.

USA has a tree, and China has a lot of useless holes on the ground.


China should have their own plan on AI, separated from USA.

Racing for our own goal is the formula for winning and staying ahead.


China lacks of software industry, China doesn't have its own OS, ecosystem, including the software.

Developing an AI like ChatGPT, MidJourney, and Sora is something useless for China.

China doesn't have a photo editor software, so why should China develop an image generator then?

The same with video editor software, so why should China develop a video generator capability?
 
While China is just following what USA does, in hope to surpass USA.

This probably has more to do with the Chinese paranoia of "loss of face" if somebody tries to do something new on their own...and fails...versus traveling down a route previously found to be successful and matching it.

Elon Musk once said something along the lines of "If you aren't failing..you aren't innovating. Failure IS an option".

Meanwhile the largest spacecraft put into orbit was achieved..after 2 failures...leading to design adjustments.
 

OpenAI’s Sora pours ‘cold water’ on China’s AI dreams, as text-to-video advancements prompt more soul-searching​

  • The introduction of Sora has forced China’s AI industry to again grapple with how it will catch up to the latest tech in the face of escalating US sanctions
  • China had envisioned itself as dominating the AI race through its vast troves of data, but quality training data has now proven critical
 
Actually, we can do a lot of things with AI.

But Chinese people lack of imagination and vision.

What they do is doing what others do, in the hope they can do better.

AI is not just about video.

But you can also do many things, like dishwashing, cleaning house, building house, tending garden, traffic management, etc.

And why aren't Chinese people doing it? Because Americans are not doing it???

Chinese cities are getting older, all the new shining buildings, all the infrastructures are getting rusty and dusty, yet, no one ever think to clean it using bot powered by AI.

This is one example.


1.4 billion people lack imagination and creativity, everyone is doing the same thing.

If someone is making a toy, everyone is also making a toy, flooding the market with the same products and then price war, killing all of them.
 
Both India and China have entirely designed their school system to jobs, not research or inventions type of innovation. By and large therefore they have hit their limits at process and design facades. The higher education environment of the US, Canada and western europe still have the wide moat.
I wont include China in this but I agree with India although I am finally seeing seeing changes in last decade in schools. I cannot say about collages now but during my time was all ratta.

Personally I think India indeed need to learn a lot from China both in Good and Bad policies.
 

Users who are viewing this thread

Pakistan Defence Latest

Back
Top