Huang defends Nvidia’s moat: ecosystem, supply chain and CUDA. And he argues that cutting off China would weaken the United States.

Sanfilippo argues that in coding agents, the brand of the model matters less than the ability to switch it often. His thesis: try everything, pay for a short time, don’t fall in love with anyone.

Sanfilippo explains why DeepSeek v4 Pro could be the first open-weights model really close to closed systems, and why price remains the decisive problem.

OpenAI misses targets, big tech raises capex, and AI’s real limit becomes energy, not enthusiasm.

Anthropic、Microsoft 和 Google：The Prime Time 的叙述者认为，这些信号显示 AI 经济正变得更艰难。

Naval Ravikant spiega come l’AI sta già cambiando il lavoro, perché teme concentrazione e rischi biologici, e perché difende l’ottimismo.

Anthropic, OpenAI, and the Pentagon clash over military AI. The video explains why this fight is about power, contracts, and accountability.