A massive amount of OpenAI’s compute is dedicated to next-generation research, whereas we are stretched thin — just meeting ...