It is rather humorous to see OpenAI complaining about Deepseek being trained using distillation, i.e. stealing ChatGPT's knowledge. OpenAI trained their models almost exclusively on data scraped from the Internet.

This is the pot calling the kettle black.