Description
1.7 Societal concerns about AI code assistants like Copilot
1.7 大众对于 AI 编程助手的担忧
There’s societal uncertainty right now about AI code assistants like Copilot. We thought we’d end the chapter with a few questions and our current answers. Perhaps you’ve been wondering about some of these questions yourself! Our answers may turn out to be hilariously incorrect, but they do capture our current thoughts as two professors and researchers who have dedicated their careers to teaching programming.
如今,社会各界对于 Copilot 等 AI 编程助手的态度还有些摇摆不定。我们打算在本章的结尾处提出一些常见问题,并附上我们当前的看法。这些问题可能也正是你心中所疑惑的。虽然我们的回答可能随着时间的推移逐渐显得荒谬,但这些回答确实反映了眼下我们两位作为长期投身于编程教育领域的教授和研究者的真实观点。
Q: Are there going to be fewer tech and programming jobs now that we have Copilot?
问:现在有了 Copilot,技术和编程岗位会减少吗?
A: Probably not. What we do expect to change is the nature of these jobs. For example, we see Copilot as being able to help with many tasks typically associated with entry-level programming jobs. This doesn’t mean that entry-level programming jobs go away, only that they change as programmers are able to get more done given increasingly sophisticated tools.
答: 应该不会。不过我们预计这些岗位的性质将会发生变化。比如,我们知道 Copilot 能辅助处理许多与初级编程岗位相关的任务。这并不意味着初级编程岗位将直接消失,只不过随着程序员能够借助越来越先进的工具完成更多任务,这些岗位的性质将发生改变。
Q: Will Copilot stifle human creativity? Will it just keep swirling around and recycling the same code that humans have already written, limiting introduction of new ideas?
问:Copilot 会扼杀人类的创造性吗?它会不会只是在不断地回收利用人类已经编写的代码,从而限制新观点的引入?
A: We suspect not. Copilot helps us work at a higher level, further removed from the underlying machine code, assembly code, or Python code. Computer scientists use the term abstraction to refer to the extent that we can disconnect ourselves from low-level details of computers. Abstraction has been happening since the dawn of computer science, and we don’t seem to have suffered for it. On the contrary, it enables us to ignore problems that have already been solved and focus on solving broader and broader problems. Indeed, it’s been the advent of better programming languages that have facilitated better software—software that powers Google search, Amazon shopping carts, and macOS weren’t written (and likely could not have been written) when we only had assembly!
答: 我们认为不会。Copilot 使我们能够在更高层面上进行工作,远离了底层机器码、汇编语言或 Python 代码。计算机科学家用 “抽象” 这一术语来描述我们与计算机底层细节脱离的程度。抽象自计算机科学诞生之初就在进行,我们并没有因此遭受损失。相反,它让我们能够忽略那些已经解决的问题,专注于解决越来越广泛的问题。事实上,正是更高级编程语言的出现,推动了更高质量软件的开发——那些驱动 Google 搜索、亚马逊购物车和 macOS 的软件,并非在我们仅有汇编语言时编写的(可能靠汇编也根本写不出来)!
Q: I keep hearing about ChatGPT. What is it? Is it the same as Copilot?
问:我一直听人在说 ChatGPT,它是什么?它和 Copilot 是一回事吗?
A: It’s not the same as Copilot, but it’s built on the same technology. Rather than focus on code, though, ChatGPT focuses on knowledge in general. And as a result, it has insinuated itself into a wider variety of tasks than Copilot. For example, it can answer questions, write essays, and even do well on a Wharton MBA exam [7]. Education will need to change as a result: we cannot have people ChatGPT’ing their ways to MBAs! The worthwhile ways in which we spend our time may change. Will humans keep writing books and, if so, in what ways? Will people want to read books knowing they were partially or fully written by AI? There will be effects across industries, including finance, health care, and publishing [8]. At the same time, there is unfettered hype right now, so it can be difficult to separate truth from fiction. This problem is compounded by the simple truth that no one knows what’s going to happen here in the long term. In fact, there’s an old adage coined by Roy Amara (known as Amara’s Law) that says, “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” As such, we need to do our best to be tuned into the discussion so that we can adapt accordingly.
答: ChatGPT 和 Copilot 并不相同,但它们是基于同一种技术构建的。与专注于编程的 Copilot 不同,ChatGPT 适用于更广泛的知识领域。这使得它能够胜任更多样的任务,比如回答问题、撰写文章,甚至在沃顿商学院的 MBA 考试中取得优异成绩 [7]。这意味着教育也需要随之变革:我们总不能让人们靠 ChatGPT 就获得 MBA 吧!同样,我们花费时间的方式也需要转变。人类还会继续写书吗?以什么方式来写?当人们知道书籍可能部分或完全由 AI 编写时,他们还会愿意读书吗?这将对金融、医疗保健、出版等行业产生深远影响 [8]。与此同时,这项技术也被过度炒作,这也让我们一时难以辨别传言的真假。如果拉长时间跨度,这些问题会愈发难以回答,没有人能预测未来到底会发生什么。实际上,罗伊·阿玛拉有一句老话(阿马拉定律)就曾指出:“我们倾向于高估一项技术的短期影响,而低估其长期影响。” 因此,我们需要密切关注这一领域的讨论,以便及时适应变化。
In the next chapter, we’ll get you started using Copilot on your computer so you can get up and running writing software.
在下一章中,我们将引导你在自己的电脑上启动并使用 Copilot,让你能够迅速开始编写软件。