#198 - DeepSeek R1 & Janus, Qwen2.5, OpenAI Agents
Last Week in AI
Reinforcement Learning and AI Optimization
This chapter explores the complexities of reinforcement learning, emphasizing the importance of reward models in AI training. It discusses the innovative DeepSeq R1-0 model, which embeds reasoning within defined thinking tags, and the challenges of language switching and human interpretability. The conversation also highlights advancements in AI technologies, including the DeepSeek R1 model and Janus Pro text-to-image generator, alongside the implications for compute resource availability and future developments.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
