It has been almost two months since the release of DeepSeek’s R1 model, an AI reasoning model that rivals the performance of the models of US giants like OpenAI but, according to its creators, at a fraction of the cost and using less advanced chips