Part 2/10:
Deep Seek, a newly developed model, seems to signal a paradigm shift in artificial intelligence. While previous models, like ChatGPT from OpenAI, focused heavily on training data, Deep Seek appears to leverage a process that utilizes synthetic data, engaging in massive prompting of existing models to extract insights about the world. Essentially, Deep Seek has adopted a “query and extract” method, which has proven to be more efficient technologically and financially. By using comparatively lesser hardware and significantly reduced costs for training, Deep Seek has challenged established models, inciting a surge in interest and speculation.