Life is full of uncertainties. The more you talked about the issue, the more worries you raised. Perhaps, there is one more worry to take into account, which is how about the end of the world is arriving tomorrow. Jesus said "Who of you by worrying can add a single hour to your life? ". Who can tell when you will die, rich or poor, fit or ill?
老簫太過神化 AI，現在 AI frontiers 是大型言語模型，能夠明白了解用家，和現有知識 corpus，但 AI 本身並不創造知識。如果老簫團隊有時間又有能力，參考下 Patrick Boyle 引用 Harvard Business School 一篇研究報告： ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-FTs35x-xUg4.html 這是 empirical study，不是理論推敲 (aka 狗噏）。
@@alvinmark5558 我沒有說，LLM似乎并不能創出新方法。我只是說，LLM （應該是 ANN 的所有應用）不能創出現有知識沒有 imply 的新知識。不過，再討論下去很容易變從定義為起點的理論進路（e.g. 甚麼是知識，怎樣知道知識, etc.)，這是 top down approach, 但 ANN 偏偏相反，bottom up, 通過每個 training case 改變些微結構 so as to fit the case，十億百億個 training cases 之後變成 一個 best fit 的結構。Best fit 結構，是否「新」結構，這是定義問題。 記得好似沈郁輝有條片講過而家年輕人興用 bottom up approach 睇世界（條片懶得揾），老簫的 critical thinking 餐餐都是從定義出發，然後就是羅素的雞，否定所有 inductive thinking。人情朽於故見，老簫都是一個 bottom up 的 case.
No matter how clever AI is, how can it break through law of deminishing return in terms of achieving more and more circuit density, speed of electron movement. Is it really possible to limitless extension of DNA life and cell reproduction and aging
So AI will become God one day? Quality of life is much more important than longevity. Death will come as a blessing when you’re facing multiple health problems.
