Sacks claims 'substantial evidence' that DeepSeek distilled OpenAI's models

That's well possible. I asked DeepSeek how I should send them feedback, and the response was:

"If you’d like your feedback to be formally acknowledged or escalated (e.g., to highlight nuances about your work or correct details), you can submit it via OpenAI’s official feedback portal or share it through their support team."

Pretty remarkable.