In a recent podcast episode, Nathan Lambert and Florian Brand survey the state of open models, focusing on developments around Kimi K3 and Qwen 3.8. The conversation also touches on broader industry signals, including Xi Jinping's speech at WAIC, which adds a policy dimension to the technical news.
A recurring theme is distillation, the process that allows open models to learn from larger or more capable systems. According to the episode's framing, distillation helps explain why the capability gap between open and closed models may be shrinking. The title frames this as an ongoing shift rather than a settled conclusion.
Since the article draws on a single source, there are no contrasting views to note. Still, the recap positions the open ecosystem at a moment of rapid change, with model releases, policy statements, and technical innovations all pointing toward a more competitive open-model landscape.