#490 – State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI

The discussion centers on the current state and future trajectory of AI, particularly large language models, considering both US and Chinese advancements. Key topics include the impact of open-source models, the balance between model intelligence and speed, and the role of data quality in pre-training. The participants debate whether scaling laws still hold, especially concerning pre-training, and explore the potential of reinforcement learning with verifiable rewards (RLVR). They examine the trade-offs between closed and open-source models, the challenges of tool use, and the potential for AI to transform industries. The conversation also touches on the ethical considerations of AI development, the importance of human agency, and the potential for AI to address societal challenges.