EMJ Capital President and founder Eric Jackson analyzes the Star Gate under the Trump administration and the tough AI competition of cryptocurrency cyber security.
The latest artificial intelligence (AI) model, launched by China's Startup DeepSeek, has promoted the turmoil in the technology sector following the appearance of a potential rivalry to the United States.
Last month, DeepSeek wrote in Paper that the Deepseek-V3 model was 2,000 NVIDIA H800 chips with less than $ 6 million, and that it achieved the same performance as Openai and Meta's most advanced models. 。
These chips are not more advanced than the most advanced chips in the market for export control, but Deepseek claims that it has overcome its disadvantages with innovative AI training technology. DeepSeek's AI Assistant is equipped with a Deepseek-V3 model, exceeding Openai's Chatgpt as a free application in the US Apple App Store.
The emergence of companies based in China has questioned questions about investing in US high -tech companies, which have invested billions of dollars in advanced chips and large -scale data centers used in the training of AI models. It was raised. Following the awareness that the United States has a superior to geopolitical rivals in emerging fields, it also functions as an AI competition between the United States and China.
Chinese App Deepseek Hammers U.S. stock with low stock AI model
Deepseek's AI Assistant has exceeded Openai's Chatgpt on the Apple App Store. (Photo alliance by Christoph Dernbach / Getty Images / Getty Images)
The quality of the DeepSeek model and the cost efficiency have changed the story that Chinese AI companies are tracking US counter parts.
The Deepseek-R1 model was released last week, according to the company's official WeChat account, 20 to 50 times cheaper than O1 models of OPENAI according to the task.
The R1 model is also an open source and users can use it for free, but Openai's Chatgpt Pro plan is $ 200 per month.
Silicon Valley praise China's AI Startup DeepSeek: “Deep Gifts for the World”
According to Deepseek, the model was executed at the latest Openai and human models at only a few minutes of cost. (Getty Image /Getty Image)
Deepseek was created in late 2023 after dominating Liang Wenfeng, a co-founder of the quantitative hedge fund HIGH-FLYER. [artificial general intelligence]。 “
The artificial general information (AGI) is a goal pursued by Openai, defining AGI as an autonomous system that exceeds humans with most economically valuable tasks.
Click here to get a FOX business on the go
“DeepSeek has done something really impressive with employees and funds that are much less than Openai. Malek pointed out that DeepSeek is” not competitive with Openai “, and some of the more famous AI apps. I explained.
“Openai (and Google's Gemini) is a wide -range general -purpose tool based on a vast information coat. It can also be used to create applications with a smaller focus. Openai has a huge number of features in natural language processing. Malek said, but Deepseek is created. “Now, it doesn't mean that Deepseek is not good. It's important to understand that it's not the same as Openai.
Reuters contributed to this report.
