Tag
Sarah Guo's framework on open models, model labs vs agent labs, and the concept of 'untrainable' is discussed, emphasizing that applications win through unglamorous integration work and that intent is a scarce input.