DeepSeek has done an amazing job disrupting everything we know about AI models, and what resources are needed to run them efficiently Up until now investors have been pouring money into a funnel waiting for the return, The DeepSeek model developed in China has been built with a fraction of the cost. Now we’re wondering how the model was trained and we have at least some answers.
OpenAI and Microsoft are investigating whether its data was being used to train the DeepSeek model. In the beginning, OpenAI faced some scrutiny from companies and website owners for using its crawlers to gather data without permission. Microsoft observed some large data transfers from OpenAI developer accounts late in 2024.
Of course, what makes this interesting is that OpenAI technically used its crawlers to feed its models across the Internet without consent from companies or websites.
Via Bloomberg
