OpenAI CEO, Sam Altman: "you don't need to write prompts anymore" overstates it, but the real argument underneath it is worth taking seriously. engineer sean goedecke makes the actual case. for chat use specifically, the skill isn't crafting the perfect prompt, it's understanding what llms can and can't do well, and asking hard questions across multiple models at once instead of iterating fifteen minutes in a single chat. his practical tip: ask a hard question to four or five frontier models simultaneously rather than refining one prompt repeatedly. one of them often lands on a good answer immediately, and comparing responses tells you more than any single well-crafted prompt would. the caveat he includes himself, worth keeping: if you're writing code that calls an llm, prompting still matters a lot. this shift applies specifically to how you use chat, not to production systems.(CyrilXBT)