吴修铭：电脑弱化人脑

1912年的布拉格，当时的卡夫卡还是一名29岁的律师，9月22日晚上10点，他在打字机前坐下，开始写作。8小时之后，他完成了《审判》。

在当天的日记里，卡夫卡这样写道：“我几乎无法从桌子底下抽出双腿，久坐使它们变得僵硬。可怕的紧张与愉悦，故事在我面前呈现开来，如逆水而行。”后来他描述这种他所倾向的一坐到底、一气呵成的写作方法：“唯有这样的连贯一致，伴随肉身与灵魂的完全开放，才能完成写作。”

1951年4月，纽约雀儿喜区（Chelsea neighborhood）一栋褐砂石建筑的六楼，杰克•凯鲁亚克开始把描图纸用透明胶粘在一起，做成他称之为“卷轴”的120英尺（约合37米——译注）长的纸张。这样连续打字而无需停下换纸，三周后，他在“卷轴”上完成了《在路上》（On the Road）的初稿，既不分段，又无页边空白。

1975年，史蒂夫•乔布斯在雅达利（Atari）上晚班。他被要求在四天之内为《突出重围》（Breakout）这款电子游戏设计原型。他接下任务，向朋友史蒂夫•沃兹尼亚克寻求帮助。沃兹尼亚克这样描述当时的壮举：“四天？我以为我办不到。四天之内，我不眠不休。我和史蒂夫都患上了单核白血球增多症，一种睡眠方面的疾病，而我们交付了运行完好的《突出重围》。”

卡夫卡、凯鲁亚克和沃兹尼亚克的成就当然令人刮目相看，然而有才华的人在高度集中的状态下，有这样的成就并非完全闻所未闻。更有意思的问题是：要在当下达成他们的壮举，是更难了？还是更容易？

一方面，当今配备了编程与写作工具的电脑，比起二十世纪的任何工具都更为强大。然而在其他方面，他们每个人的工作都变得更加困难：在电脑上，每个人都要与注意力分散殊死搏斗。卡夫卡或许会像大部分律师一样，开始写作后又意识到最好去查查电邮，《审判》也就这样不了了之。凯鲁亚克或许会在推特上流连许久，或者是在博客上写旅行见闻。沃兹尼亚克或许会在工作中修改维基百科上有错误的条目，而最终使后来成为苹果的合作毁于一旦。

相对于我们来说，卡夫卡、克鲁亚克和沃兹尼亚克有一项优势：他们的书写工具不能同时处理几件事情，人们也因此不会轻易向纠结的欲望低头。虽然注意力的分散无处不在——看报纸，和朋友聊天——但是其间有本质上的区别。现在的电脑不单单分散注意力，而且是在促使这种分散。网络像叫卖者一样无时无刻不在召唤我们；电脑未能使我们专注于工作，反而使我们深陷泥潭，甚至是提供了使我们分心的诱惑。简而言之，我们创造了一代“分散注意力电脑”，使自己更加难以做到专心致志。

今时今日，我们应当创造的写作工具，需要帮助我们的大脑应对它所不擅长的，比如专心工作；帮助我们达成极度专注的状态，而不是分散我们的注意力。我们所需要的新一代科技，其功效应当是克鲁亚克的卷轴和卡夫卡的打字机。

要理解当下发生的一切，我们需要回溯到上个世纪六十年代。当时的电脑巨大而缓慢，同时为十几人甚至是上百人服务。这样的电脑需要一种办法来应对纷至沓来的资料处理需求。工程师开发了不同的技术来解决这一问题——从起初的时间共享到后来的多任务处理系统。本质上，多任务算法使用精确的技术，在多名使用者之间公平流畅地共享电脑资源。有了多任务处理，同一台电脑的共享者都会以为他们是在用只属于他们自己的电脑。

开发出时间共享和多任务处理的工程师们大概从来不曾想象过，他们的概念会被应用到个人电脑上——如果每位使用者都有了自己的电脑，又何必要多任务处理呢？当苹果II这样的个人电脑在七十年代末大规模上市的时候，它们的处理能力非常有限，一次仅能完成一项任务，或是编程，或是文字处理，两者无法同时进行。

个人电脑多任务处理能力的兴起，与其他的研发是分不开的，起初是六十年代开始出现的桌面/windows界面，在八十年代通过苹果的Mac系统走进大众视野。有不同“窗口”的“桌面”这一概念本身，意味着使用者可以在不同工作项目之间切换。在七十年代的施乐公司（Xerox），艾伦•凯（Alan Kay）是最早的窗口系统研发者之一，在一次访问中他谈到：“一般来说我们都想同时编辑浏览几个场景——这可能就是简单的图像与文字的合并，或者是处理多个任务，或者是比较同一模式的不同面向。”

计算机的多任务处理功能反而使人们难以集中于自己手头的工作

多任务处理的目的，从支持同一台电脑的多个使用者，转向了支持同一个使用者同一时间的多项需求。前者解决了多人间的冲突；后者反而是带入了一种内在的冲突，仔细想想，试图同时达成多项需求恰恰就是与专注相违背。

另一项重大进展是过去三十年里电脑处理器速度的飞跃。也只有配备了这样的能力，个人电脑才能以人们可以接受的方式进行多任务处理。毋庸置疑，多任务处理较之于“单任务处理”电脑，体现了重要的科技进展。举例来说，旧版本的苹果处理系统使用手册宣称：“当Mac电脑还是新生事物的时候，处理系统照理应当是一名使用者处理一个程序。这显然已经过时。今时今日，我们需要电脑做得更多更快，而我们需要做的工作越来越少。”

当然，多任务处理电脑在技术层面上更为先进。然而我们已经可以看到有些事情走偏了。我们并不需要电脑做得更多——而是作为人类的我们要把事情做完。这一小点才是最重要的，也表明了让电脑重回最基本功用的需要。

六十年代，J.C.R•立克里德（J. C. R. Licklider最早预测类似互联网的系统会将全世界的人和计算机联系在一起的人之一——译注）和道格拉斯•恩格波特（Douglas Engelbart鼠标发明者——译注）提出，电脑的终极任务是人类强化的工具，他们改变了电脑发展的走向。他们认为，电脑不应当像星球大战中的R2-D2机器人一样具有独立的人工智能。相反，电脑应当服务于人类的大脑，使之变得更强大。立克里德将这一概念称为“人机共生（man-computer symbiosis）”。

从这个角度来说，电脑的多任务处理能力有时也是一种进步——也仅仅是有时而已。人们可以在浏览页面与待办事项之间切换，或者是在Skype通话的同时阅读文件。但很多时候，我们使用电脑是为了完成需要我们聚精会神的工作，也正是在一点上，电脑降低了人类的潜能。

尽管人类大脑有诸般专擅，它同时也有其弱点。在保持持续的注意力后，大脑不太能达到极度专注的状态。由于大脑很容易被有意无意的要求所影响，因此要对同一对象保持持久的注意（也就是佛教所说的专注冥想），需要大量的训练与努力。其次，大脑并不擅长自觉的多任务处理，也不能同时主动活跃地集中在多项事务上。也许电脑设计者们曾希望，电脑可以训练大脑更有效地处理多项任务，然而最近的研究表明，这一努力已告失败。

简而言之，我们很容易分心，不擅长同时做两件以上的事情。虽然电脑理所应当为我们服务，可它却不断使我们分心，要求我们同时处理一连串的不同信息。我们不得不问，这种时候，究竟谁是主导？

确实，人们已经作出努力去应对我所描述的问题。“自由程序”（Freedom program）的设计者切断因特网这一注意力的首要分散源，以此提升使用者的生产效率。有人使用咖啡因或是“聪明药”（adderall）来帮助集中注意力，有人借助期限将至或是可能被裁员这样的恐惧情绪来达到专注状态。

然而我们应当寻求的解决办法不应该依靠药物或是迫近的失业。我们需要的电脑，是从一开始就特意降低分心，帮助我们持久专注于棘手的工作上。我们需要的电脑和设备，应当回到人类强化的任务上来，认真对待大脑的极限，并帮助我们克服这些极限。

我不能完全确定这意味着什么。我所确定的是，我们应该这么做。或许我们需要的只是可以锁定不同模式的电脑：琐事模式，交流模式，工作模式。在工作模式里，电脑应该尽其所能让我们不偏离轨道。设计者要认识到大脑的弱点，尽力消除或降低不必要的分心，诸如提醒电邮的嘟嘟声，跳动的图标和毫无必要的弹出窗口。

总会有人说，解决这些问题的不二法门就是训练与意志力——即使生活在2013年，卡夫卡还是会像1912年一样专注。对此我不太确定。训练当然有用，但环境和工具也的确有用。奇怪的是，我们现在拥有了改变环境的技术能力——这对于前人来说是无法想象的，不过我们并没有认清大脑的缺点就在使用这些技术了。

或许，一条简单的规则就足够表达：电脑不应该让我们变得更蠢。

（本文载于《纽约客》网站2013年9月9日，原标题How Today’s Computers Weaken Our Brain ；任致均/译）

翻页请看英文原文

How Today’s Computers Weaken Our Brain

By Tim Wu

September 9, 2013

At 10 P.M. on September 22, 1912, Franz Kafka, then a twenty-nine-year-old lawyer, sat down at his typewriter in Prague and began to write. He wrote and wrote, and eight hours later he had finished “Das Urteil” (“The Judgment”).

Kafka wrote in his diary, “I was hardly able to pull my legs out from under the desk, they had got so stiff from sitting. The fearful strain and joy, how the story developed before me, as if I were advancing over water.” He later described the one-sitting method as his preferred means of writing. “Only in this way can writing be done, only with such coherence, with such a complete opening out of the body and soul.”

In April, 1951, on the sixth floor of a brownstone in New York’s Chelsea neighborhood, Jack Kerouac began taping together pieces of tracing paper to create a hundred-and-twenty-foot-long roll of paper, which he called “the scroll.” Three weeks later, typing without needing to pause and change sheets, he’d filled his scroll with the first draft of “On the Road,” without paragraph breaks or margins.

In 1975, Steve Jobs, working the night shift at Atari, was asked if he could design a prototype of a new video game, Breakout, in four days. He took the assignment and contacted his friend Steve Wozniak for help. Wozniak described the feat this way: “Four days? I didn’t think I could do it. I went four days with no sleep. Steve and I both got mononucleosis, the sleeping sickness, and we delivered a working Breakout game.”

The accomplishments of Kafka, Kerouac, and Wozniak are impressive, but not completely atypical of what can be achieved by talented people in states of supreme concentration. The more interesting question is this: Would their feats be harder today, or easier?

On the one hand, today’s computers feature programming and writing tools more powerful than anything available in the twentieth century. But, in a different way, each of these tasks would be much harder: on a modern machine, each man would face a more challenging battle with distraction. Kafka might start writing his book and then, like most lawyers, realize he’d better check e-mail; so much for “Das Urteil.” Kerouac might get caught in his Twitter feed, or start blogging about his road trip. Wozniak might have corrected an erroneous Wikipedia entry in the midst of working on Breakout, and wrecked the collaboration that later became Apple.

Kafka, Kerouac, and Wozniak had one advantage over us: they worked on machines that did not readily do more than one thing at a time, easily yielding to our conflicting desires. And, while distraction was surely available—say, by reading the newspaper, or chatting with friends—there was a crucial difference. Today’s machines don’t just allow distraction; they promote it. The Web calls us constantly, like a carnival barker, and the machines, instead of keeping us on task, make it easy to get drawn in—and even add their own distractions to the mix. In short: we have built a generation of “distraction machines” that make great feats of concentrated effort harder instead of easier.

It’s time to create more tools that help us with what our brains are bad at, such as staying on task. They should help us achieve states of extreme concentration and focus, not aid in distraction. We need a new generation of technologies that function more like Kerouac’s scroll or Kafka’s typewriter.

To understand what has happened, we need to return to the nineteen-sixties, when computers were giant, slow machines that served dozens and sometimes hundreds of people at once. Such computers needed a way to deal with competing requests for processing resources. Engineers devised various techniques for handling this problem—known first as time-sharing, and later as multitasking, operating systems. In essence, multitasking algorithms used clever techniques to share the computing power available among multiple users as fairly and smoothly as possible. With multitasking, it was possible with a single computer for many people to have the illusion of having their own machine.

The engineers who designed time-sharing and multitasking probably never imagined that their ideas would be used for personal computers—if each user already had a computer, why would he or she need multitasking? And when the first mass-market personal computers, like the Apple II, arrived in the late seventies, their highly limited processing power was used to perform a single task at a time. It was programming or word processing, but not both at once.

The rise of multitasking capabilities in personal computers cannot be separated from other developments, beginning with the introduction of the familiar desktop/window interface that began in the sixties and reached the public in the eighties, via the original Apple Macintosh. The very idea of a “desktop” with different “windows” implies a user who can switch between tasks. As Alan Kay, one of the inventors of the first functioning window-style system, at Xerox in the seventies, explained in an interview, “We generally want to view and edit more than one kind of scene at the same time—this could be as simple as combining pictures and text in the same glimpse, or deal with more than one kind of task, or compare different perspectives of the same model.”

The purpose of multitasking had gone from supporting multiple users on one computer to supporting multiple desires within one person at the same time. The former usage resolves conflicts among the many, while the latter can introduce internal conflict; when you think about it, trying to fulfill multiple desires at once is the opposite of concentration.

A second crucial advance was the huge increase in the speed of computer processors over the past three decades. Only with this kind of power could personal computers multitask in an acceptable way. It was immediately assumed that, once achieved, multitasking represented an important technical advance over “single-tasking” machines. For example, an old guide to Apple operating systems declared, “Way back when Macs were new, operating systems were meant to be operated by one user working with one program. Obviously, this is no longer the case. Today, we want our computers to do more, faster, with less work on our part.”

Of course, in a technical sense a multitasking machine is more advanced. But we can already see where things might be going astray. We don’t really want our computers to accomplish more—it’s us, the humans, who need to get things done. This subtle point is all-important, and shows a need to return to the basics of what computers are for.

When, in the sixties, J. C. R. Licklider and Douglas Engelbart proposed that computers should ultimately serve as a tool of human augmentation, they changed what computers would come to be. The computer, they argued, shouldn’t try to be independently intelligent, like R2-D2. Rather, it should be a tool that works with the human brain to make it more powerful, a concept that Licklider called “man-computer symbiosis.”

From this perspective, the multitasking capabilities of today’s computers are sometimes a form of augmentation—but only sometimes. It can be helpful to toggle between browser pages and a to-do list, or to talk on Skype while looking at a document. But other times we need to use computers for tasks that require sustained concentration, and it is here that machines sometimes degrade human potential.

While the brain is good at many things, it is rather bad at others. It’s not very good at achieving extreme states of concentration through sustained attention. It takes great training and effort to maintain attention on one object—in what Buddhists call concentration meditation—because the brain is highly susceptible to both voluntary and involuntary demands on its attention. Second, the brain is not good at conscious multitasking, or trying to pay active attention to more than one thing at once. Perhaps computer designers once hoped that our machines could train the brain to multitask more effectively, but recent research suggests that this effort has failed.

In short, we are easy to distract, and very bad at doing two or more things at the same time. Yet our computers, supposedly our servants, constantly distract us and ask us to process multiple streams of information at the same time. It can make you wonder, Just who is in charge here?

To be sure, efforts are being made to deal with the problems I’ve described. The designers of the Freedom program give users a way to boost productivity by switching off the Internet, the chief source of distraction in our times. Some people turn to caffeine or Adderall as an aid to concentration, or achieve similar effects through the use of emotions like the fear created by deadlines or the possibility of being fired.

But we should be searching for solutions that don’t rely on drugs or imminent job loss. What we need are machines that are built from the ground up purposely to minimize distraction and help us sustain attention for hard tasks. We need computers and devices that return to the project of human augmentation by taking the brain’s limits seriously, and helping us overcome them.

What this looks like, I’m not exactly sure, although I am sure we should be trying to find out. Perhaps all we need are computers that lock into different modes: chore mode, communication mode and concentrated work mode. In the work modes, the machine would do what it could to keep you on track, in ways both subtle and less so. We also need designers cognizant of the brain’s weaknesses, who strive to eliminate or minimize unnecessary distractions, such as beeps for e-mails, bouncing icons and unnecessary pop-up windows.

There will always be some who say that all anyone needs to deal with these problems is better discipline or will power—that Kafka, being Kafka, would stay on task in 2013 just as well as in 1912. I’m not so sure. Discipline is useful, but so is an environment and tools that actually help, rather than hinder. The strange part is that we now have technological powers to shape our environment that were unimaginable to earlier generations, yet we don’t use them with a realistic view of the brain’s weaknesses.

Perhaps a single rule is enough: our computers should never make us stupider.