DeepSeek is sending Silicon Valley into a panic. The China-developed LLM performs as well as OpenAI’s GPT-1o and cost a fraction of the cost to train, under $6M for DeepSeek compared to $80 - $100M for GPT-1o. The US contrained China’s access to chips, and DeepSeek innovated a win. This seems like a pivotal moment for driving down costs, which may mean trouble for Nvidia.
https://www.axios.com/2025/01/27/deepseek-ai-model-china-openai-rival
[#]AI #openAI #DeepSeek #Nvidia
=> More informations about this toot | More toots from bretcarmichael@mastodon.social
text/geminiThis content has been proxied by September (3851b).