OpenAI Chief Scientist: DeepSeek Independently Discovered Some Core Concepts of o1
DeepSeek challenges OpenAI with low-cost AI models, sparking debates on AI hardware demand, cost efficiency, and open-source principles. How far will DeepSeek go?
"AI Disruption" publication New Year 30% discount link.
Lower Costs, Higher Demand?
In the past few days, DeepSeek has dominated discussions in the AI community, while NVIDIA’s stock fluctuations have shocked global investors.
Simply put, DeepSeek recently released two models—DeepSeek-V3 and DeepSeek-R1, achieving performance comparable to OpenAI’s counterparts at significantly lower costs.
This has sparked concerns in the market about AI hardware demand. Investors worry that future demand for NVIDIA’s high-end chips might decline.
At the same time, discussions about DeepSeek’s technological innovations have surged.
Many believe that DeepSeek, constrained by limited computing resources, was forced to take a different path from OpenAI and others, who rely on massive computational power. Instead, DeepSeek used a series of technical innovations to reduce compute dependency while improving performance.
These achievements have earned recognition from AI leaders, including Sam Altman.
Following this, OpenAI’s Chief Scientist, Mark Chen, also weighed in. He acknowledged that DeepSeek independently discovered some of the same core ideas OpenAI's o1 model employs. However, he quickly added that this doesn’t mean OpenAI’s heavy investment in computing is unjustified.