This week, Chinese AI company DeepSeek introduced its reasoning model, R1.
![]()
DeepSeek Launches R1 Model: A Big Step in AI
The model has impressed many in the tech world, with some saying it performs as well or even better than OpenAI's models on certain tests. Many are calling it a major breakthrough in AI technology.
![]()
Affordable AI Despite U.S. Sanctions
DeepSeek claims to have trained R1 for just $5.6 million, much less than U.S. companies typically spend on similar models. This is especially notable since DeepSeek faced U.S. sanctions that block access to advanced chips but still managed to develop the model efficiently.
Concerns and Criticism Over DeepSeek’s Strategy
Some critics, like Curai CEO Neal Khosla, have questioned DeepSeek’s low-cost claims, suggesting they could hurt U.S. AI companies. However, others believe the lower cost of AI models could lead to faster adoption and growth in the industry.
Open-Source Models Gaining Strength
Meta’s Yann LeCun pointed out that DeepSeek’s success highlights the growing importance of open-source AI. By using open-source tools, DeepSeek was able to innovate and create R1, showing how sharing knowledge and resources can help advance AI development.
The post of this news is https://x.com/pmarca/status/1882719769851474108.