AI researcher Nathan Lambert explains that open-source models trail closed frontier labs primarily due to compute constraints rather than talent gaps. He argues that distillation has limits, and exponential scaling costs make a sudden intelligence explosion unlikely.
FireBrief
FireBrief