最近看到很多人在談 AI 的未來。哪些工作會消失、哪些能力會變重要，甚至幾年後我們會怎麼生活。看得越多，我反而越想問：我們是不是太快把一種可能，當成了未來的答案？
There are plenty of predictions about AI. The more I read, the more I wonder: are we treating possibilities as answers too soon?
今天的 AI，不一定能讓我們看懂明天
Today's AI may not tell us what comes next.
Ilya Sutskever 在訪談中提到，我們很難想像還不存在的 AI 系統。看著今天的工具，連做 AI 的人，也不容易真正理解未來更強大的系統會帶來什麼改變。他自己也提出預測，但這個提醒值得留下：我們熟悉的使用經驗，可能限制了我們的想像。訪談原文
Ilya Sutskever points out how hard it is to imagine AI systems that do not yet exist. Even researchers can struggle with this. He makes predictions too, but his point reminds me that today's experience can limit our imagination.
這讓我想到，現在我們常用「寫得更快、做得更多」來理解 AI。但未來的變化，也許不只是同一件事情加速，而是工作怎麼分配、怎麼合作，甚至什麼事情值得做，都跟著改變。這是我的延伸思考，不是已經確定會發生的結果。
We often describe AI as a faster way to do things. But it could also change how we work together and what we choose to work on. That is a possibility I see, not a promise about the future.
承認不知道，不代表什麼都不做
Not knowing does not mean doing nothing.
Yoshua Bengio 在談 AI 風險的文章中，坦承未來充滿不確定性，也提醒研究者避免過度自信，願意在新證據出現時改變看法。他並不是說「既然不知道，就不用擔心」，而是認為我們仍然需要在不確定中採取預防措施。文章原文
Yoshua Bengio acknowledges uncertainty and asks researchers to stay open to new evidence. His message is not to ignore the risks. It is to take precautions even when we do not have all the answers.
把這兩個觀點放在一起，我的理解是：我們可能想不到未來的全貌，但還是能決定現在要怎麼學、怎麼試，以及哪些風險不能隨便放過。
Together, these ideas remind me that we can prepare without knowing the whole future. We can keep learning, test carefully, and take risks seriously.
我想保留的，是修改想法的空間
I want to leave room to change my mind.
對我來說，與其急著相信某一套未來劇本，不如先把眼前的問題弄清楚。挑一個能驗證的小地方試試，看看 AI 到底幫上什麼忙，又在哪裡需要人來判斷。如果結果跟原本想的不一樣，就調整做法。
For me, a useful starting point is a real problem and a small test. See where AI helps and where human judgment still matters. If the results challenge my assumptions, change the approach.
不是所有預測都沒有價值，而是預測應該拿來幫助準備，不是代替思考。知道 AI 會帶來改變，跟知道改變後的世界長什麼樣，是兩件不同的事。
Predictions can help us prepare. They should not replace our own thinking. Expecting change is different from knowing what the changed world will look like.
我不一定能猜對未來，但我希望自己保持好奇，也保留說「我之前想錯了」的能力。
I may not get the future right. But I want to stay curious and be able to say, “I was wrong.”