OpenAI and Microsoft suspect that Chinese startup DeepSeek improperly trained their R1 model using OpenAI’s data, which is causing a stir in the AI world. Venture capitalist David Sacks claims that DeepSeek distilled knowledge from OpenAI’s models, sparking controversy. OpenAI is facing a lawsuit from the New York Times for training on their articles, defending their actions as fair use. DeepSeek managed to create a large language model more efficiently than OpenAI, using reinforcement learning. OpenAI argues that large language models require massive amounts of data for success. The debate continues on the ethics and legality of AI training methods.
https://www.404media.co/openai-furious-deepseek-might-have-stolen-all-the-data-openai-stole-from-us/