5.3 KiB
翻译 · 片段1(《Reason is more than a tool》)
原题:Reason is more than a tool 副题:If intelligence is merely optimisation then machines will outrun us. Kant tells us why human reason is so much more. 译文:理性不只是一件工具 / 如果智能仅仅是优化,机器终将超越我们。康德告诉我们,人类理性为何远不止于此。 生词:点击英文原文中的生词可跳转到 01-生词表-片段1 对应词条。
原文与译文对照
If intelligence is 01-生词表-片段1#merely /ˈmɪəli/ 01-生词表-片段1#optimisation /ˌɒptɪmaɪˈzeɪʃən/ then machines will 01-生词表-片段1#outrun /ˌaʊtˈrʌn/ us. Kant tells us why human reason is so much more 如果智能仅仅是优化,那么机器终将超越我们。康德告诉我们,人类理性为何远不止于此。
A few months ago, I sat in a 01-生词表-片段1#packed /pækt/ university 01-生词表-片段1#auditorium /ˌɔːdɪˈtɔːriəm/ listening to a 01-生词表-片段1#celebrated /ˈselɪbreɪtɪd/ computer scientist explain the 01-生词表-片段1#brilliance /ˈbrɪliəns/ of large language models (LLMs). 几个月前,我坐在座无虚席的大学报告厅里,听一位知名的计算机科学家盛赞大语言模型(LLM)的精妙之处。
These systems 01-生词表-片段1#optimisation /ˌɒptɪmaɪˈzeɪʃən/, he said. And they improve at 01-生词表-片段1#optimisation /ˌɒptɪmaɪˈzeɪʃən/ on the basis of past experience. He paused. That, he said with a smile, is intelligence. The audience nodded along approvingly. Meanwhile, a philosophical alarm bell began sounding in my head. 他说,这些系统会进行优化;而且它们会基于过往经验,不断改进自身的优化能力。他停顿了一下,微笑着说道:这就是智能。听众纷纷点头称是。而与此同时,一个哲学意义上的警报在我脑中响起。
Not because the speaker's claim was 01-生词表-片段1#implausible /ɪmˈplɔːzəbl/. On the contrary, as a description of much 01-生词表-片段1#contemporary /kənˈtemprəri/ artificial intelligence, it seemed 01-生词表-片段1#apt /æpt/. AI systems are built around objectives, losses, rewards, predictions, preferences, 01-生词表-片段1#benchmark /ˈbentʃmɑːk/ and feedback loops. They are trained to perform better relative to some standard of success. In one form or another, they 01-生词表-片段1#optimisation /ˌɒptɪmaɪˈzeɪʃən/. That is a great part of their technical power. 不是因为这位演讲者的论断站不住脚。恰恰相反,作为对当代大量人工智能的描述,它似乎相当贴切。AI 系统围绕目标、损失、奖励、预测、偏好、基准和反馈回路而构建。它们被训练得能在某一成功标准面前表现得更出色。无论以何种形式,它们都在优化——这正是其技术力量的很大一部分。
What troubled me was the 01-生词表-片段1#ease /iːz/ with which a definition 01-生词表-片段1#appropriate /əˈprəʊpriət/ to machines sounded like a definition of intelligence as such. That 01-生词表-片段1#slide /slaɪd/ from engineering model to human 01-生词表-片段1#anthropology /ˌænθrəˈpɒlədʒi/ is easy to miss. We begin by saying: this is how machines work. Then, almost without noticing, we say: this is what intelligence is. And then, more dangerously still: perhaps this is what we are. 真正让我不安的,是一个适用于机器的定义,竟如此轻易地听起来像是对智能本身的定义。从工程模型滑向人类学的这一步,极容易被忽略。我们先是说:机器就是这样运作的。然后,几乎在不知不觉中,我们说:这就是智能。再然后,更危险的是:也许,我们就是这样。
An old argument about human beings had entered the room without announcing itself. It sounded, instead, like a definition. The AI revolution has brought one of the 01-生词表-片段1#Enlightenment /ɪnˈlaɪtnmənt/'s deepest 01-生词表-片段1#quarrel /ˈkwɒrəl/ back to life – and made one side of it seem like plain common sense. In my view, the wrong side. 01-生词表-片段1#hence /hens/ the alarm bell. 一场关于人类的古老争论,就这样不请自来地走进了房间。它听起来不像争论,反而像一个定义。AI 革命让启蒙运动最深刻的一场争论重新复活——并让其中一方看起来就像不证自明的常识。在我看来,那恰恰是错的一方。警报由此而响。
段落大意(供复习速览)
| 段落 | 大意 |
|---|---|
| 第1段(场景) | 作者听知名计算机科学家称"优化就是智能",听众信服,作者却警觉 |
| 第2段(让步) | 承认该说法作为对当代 AI 的描述确实贴切:AI 确实围绕优化构建 |
| 第3段(转折) | 不安之处:适用于机器的定义被悄悄当成了"智能本身"乃至"人的本质"的定义 |
| 第4段(点题) | 这是启蒙运动遗留的古老争论;AI 革命让错误的一方看似常识 |