DeepSeek, which recently unveiled DeepSeek-R1, an answer to OpenAI’s o1 “reasoning” model, is a curious organization. It’s ...