Reliability Beats Brilliance
Insight
Yao argues AI research does not really need raw brainpower; the trait that matters most is being 靠谱 (dependable), working carefully, and owning the outcome of what you build. A clever algorithm that beats others on a benchmark is less valuable than knowing it stays stable and holds up at true production scale, which takes systems understanding and responsibility rather than cleverness.
Source Context
"AI这个事，本来也不太需要脑子，这个行业最重要的特质就是靠谱，就是做事细，然后对自己做的事负责任。" "你可能可以做一个算法，出来效果比别人好，但是可能更关键的问题是……在真正大的尺度下，我是不是能够保持稳定性、做得一样好。" — Whisper transcript [00:02:08] ("不太需要脑子……最重要的特质，就是靠谱，做事儿细，对自己做的事负责任"), host preview at [00:00:50]; scale-stability point in 组织搭建 chapter [~03:18:18]. Quotes verified against 虎嗅 digest #2, #6. Gloss: AI barely needs brains; dependability, care, and responsibility (does it hold at scale?) matter most.