The current window for open-source AI viability is narrowing. Nathan Lambert argues that open models face a critical six-month test to prove they can keep pace with closed systems. If they fail to bridge the capability gap now, developer adoption will stall. This puts immediate pressure on the Llama ecosystem.