DeepSeek, a Chinese company, has released an open-source reasoning model, R1, which reportedly performs on par with leading U.S. models while costing significantly less to develop. This release has caused a stir within the AI community, prompting discussions about the perceived American dominance in AI technology. Analysts indicate that R1 poses a serious challenge to established products like those from OpenAI, suggesting a shift in the future trajectory of AI development away from expansive, expensive data centers to potentially more efficient alternatives.
Last week, something potentially enormous happened for artificial intelligence: The Chinese company DeepSeek released an open-source, free-to-use reasoning model that is by crude measures, at least, on par with the best American equivalents.
The A.I. analyst Zvi Mowshowitz quickly called it the first serious challenge to OpenAI's o1.
DeepSeek offered one cost estimate: Its new R1 model was built for one-thirtieth the cost of OpenAI's flagship product.
The first possibility is that the much-ballyhooed American advantage on A.I. may be much smaller than has been widely thought.
Collection
[
|
...
]