And what we are learning about the role of trees
The AI landscape now appears to be entering a new phase, following the explosive debut of a Chinese chatbot called DeepSeek.
The privately held Chinese company was founded in July 2023 by Liang Wenfeng, a graduate of Zhejiang University, one of China’s top institutions. According to the MIT Technology Review, Liang funded the start-up through his hedge fund.
Meanwhile, Silicon Valley is reportedly on edge, with concerns growing that tech companies may be overspending on advancements in the AI sector, according to Angelo Zino, senior equity analyst at CFRA Research.
Its arrival sparked a massive sell-off in US tech stocks, erasing a staggering trillion dollars from the market.
Market experts echo this concern. "The fact that this technology is more energy-efficient and cost-effective than US-based models has US technology investors very concerned," said Jay Woods, chief global strategist at Freedom Capital Markets.
DeepSeek is an open-source large language model that uses a method called "inference-time computing."
ChatGPT's meteoric rise from a nascent concept to an influential AI powerhouse vividly illustrates the breakneck speed of artificial intelligence development.
This groundbreaking model not only advanced AI research, but also sparked transformative changes across a wide range of industries.
When the start-up OpenAI launched an early demo of ChatGPT on November 30, 2022, the chatbot went viral almost instantly, with users flooding social media to share examples of its capabilities.
The impact of ChatGPT on the stock market has been equally significant. Although OpenAI remains private and isn’t traded on stock exchanges, the surge in AI's public awareness sparked by ChatGPT has propelled companies in the field to new heights.
Amazed users shared how the AI language model effortlessly crafted detailed travel itineraries, penned witty jokes and captivating short stories, and even generated functional code for complex computer programs.
The chatbot experienced explosive growth, surpassing one million users in just five days.
After ChatGPT, many big players introduced their own versions of AI chatbots, such as Google’s Gemini, Microsoft’s Copilot, and Anthropic’s Claude.
As explained by Giuseppe Sette, president of the AI market research firm Reflexivity, this means the model activates only the most relevant parts of its system for each query, which helps save both money and computational power.
Nvidia, the leader in high-performance AI chips, has risen to the top of global corporate rankings, becoming the world’s most valuable company by market cap in 2024.
The Chinese company claims to have achieved AI breakthroughs using a fraction of the chips deployed by US rivals, raising eyebrows and casting a shadow over the billions of dollars US tech giants have poured into AI hardware.
One of DeepSeek’s research papers revealed that the company had used around 2,000 Nvidia H800 chips, designed to meet US export controls imposed in 2022. Experts told Reuters that these measures would have little impact on slowing China’s progress in AI.
Initially, the US banned Nvidia's most advanced AI chips, the H100s, from export to China in September 2022. Recognizing the need for the Chinese market, Nvidia introduced the less powerful H800s, only to have those subsequently banned in October.
This is where DeepSeek truly differentiates itself: its ability to seemingly achieve comparable performance to other chatbots while consuming fewer resources.
Imagine AI chips as the brains of AI systems–they're specially designed computer microchips that are optimized for the complex calculations that power artificial intelligence.
To function effectively, AI models require a tremendous amount of computing power. This requires specialized hardware like AI chips.
By making these chips smaller and more numerous, AI models can perform more calculations with less energy. It's like having a larger workforce that's more efficient, leading to faster and cost-saving results.
A 2020 study by the Center for Security and Emerging Technology revealed that training a top-tier AI algorithm can consume a month of continuous processing and cost a staggering US$100 million. This highlights the immense resource demands of the AI revolution.
According to reports from MIT, Liang of DeepSeek began purchasing Nvidia chips back in 2021—chips that are now banned from being exported to China. These chips became the foundation for DeepSeek's technology.
Experts argue that the US government’s measures are specifically aimed at slowing China’s advancements in critical fields like supercomputing and AI, with the ultimate goal of preventing the development of advanced weaponry, including nuclear arms.
On January 27, Nvidia’s shares plummeted 17%, closing at US$118.58, while its competitor, Advanced Micro Devices (AMD), also saw a significant drop, with shares falling over 6% to $115.01.
Wall Street is still evaluating the long-term impact of a low-cost AI tool from China that competes with ChatGPT and other generative AI applications.
While some experts have lauded DeepSeek’s performance, with renowned tech investor Marc Andreessen praising the tool on X on January 24 as "one of the most amazing and impressive breakthroughs" he's ever seen—and "a profound gift to the world," some Wall Street analysts believe the recent stock sell-off was an overreaction.
They argue that the immense demand for AI will continue to drive growth for key players in the sector. Still, the true impact of DeepSeek's new AI model remains to be seen, with both government and corporate measures playing a crucial role in shaping its future.
Sources: (CBS News) (The Guardian) (Reuters) (MoneyWeek)
See also: Trump announces US$500 billion investment to boost AI infrastructure
However, users of the chatbot quickly encountered China’s strict censorship and information control. While the AI model provides answers similar to US-based models like OpenAI’s GPT-4, Meta’s Llama, and Google’s Gemini, it shifts when asked about topics restricted in China, reflecting the country’s tight information control.
Concerns are growing that China’s rapid advancements in AI could pose a serious threat to major tech giants in the US and Europe. But what does this mean for the future of AI and the global tech race? Click on to learn how DeepSeek is shaking up the industry and why it has investors on edge.
DeepSeek offers glimpse into China's censorship and information control
Users report shifts on topics restricted in China
LIFESTYLE Ai
Chinese start-up DeepSeek has recently surged into the spotlight with its new AI chatbot. It's risen to the top of Apple's App Store, and has caused a tech sell-off that erased billions from Silicon Valley giants and challenged the notion of America’s tech dominance.
However, users of the chatbot quickly encountered China’s strict censorship and information control. While the AI model provides answers similar to US-based models like OpenAI’s GPT-4, Meta’s Llama, and Google’s Gemini, it shifts when asked about topics restricted in China, reflecting the country’s tight information control.
Concerns are growing that China’s rapid advancements in AI could pose a serious threat to major tech giants in the US and Europe. But what does this mean for the future of AI and the global tech race? Click on to learn how DeepSeek is shaking up the industry and why it has investors on edge.