Computers have failed to improve education. That's because no one's gotten disruptive with technology.

    Why do U.S. public schools struggle so to improve? Everyone has a theory. Is it a lack of money? Maybe, but the U.S. spends more per student （$9,000） on K-12 public education than all but a few countries and still lags in results. Also blamed: student disaffection, parental neglect, intransigent teachers unions and flaws in the way we measure performance.

    Elements of all these play a part, but the underlying problem is deeper. It comes down to the fact that scholls aren't motivating the children, and they are unmotivating because they are far too monolithic and standardized. The system doesn't account for the fact that every student learns in a different way. Harvard psychologist Howard Gardner first posited the theory of "multiple intelligences" in the 1980s, and it has gained wide acceptance since. His classification system now numbers eight kinds of intelligence. You cannot compare the wiring of Michael Jordan's brain （a kinesthetic intelligence） to that of a Frank Lloyd Wright （spatial） or a Walt Whitman （linguistic）. Every student has a different blend of intelligence and within that a different learning style and pace.

    In the 1800s teachers in a one-room schoolhouse would have no problem customizing their approach to each student. But at the turn of the 19th century, as schools filled up with 30 or 40 kids in a room, standardization became the norm. Schools turned into factories and ever since have resisted all efforts to break from a monolithic batch-process approach. Students who succeed today do so because their intelligence happens to match the dominant paradigm in use in a particular classroom, or they've somehow found a way to adapt to it.

    If the goal is to educate all students so they have an all-American shot at realizing their dreams, we must find a way to disrupt the monolithic classroom and move toward a student-centric model. The way to get there is with computer-based learning. Technology offers students the ability to learn in ways that match their intelligence types in the places and at the pace they prefer. The hardware exists. The software is emerging. Now all that has to change is the system around it. Change will face mighty resistance, but we predict it will happen in the next ten years.

    Skeptics will say the U.S. has spent $60 billion putting computers into schools over the last two decades and has still failed to transform the classroom-save possibly to increase costs and draw resources away from other school priorities. This should come as no surprise. Schools have done what virtually every organization does when implementing an innovation: cram it into its existing operating model to sustain what it already does. Merrill Lynch developed a solid online trading system to be used by customers. It could have fended off the E-Trades of the world. But all Merrill did was give the new system to its army of brokers so they could handle existing clients better. It didn't transform anything. This is perfectly predictable, perfectly logical-and perfectly wrong.

    The way to implement an innovation so it will transform an organization is to implement it disruptively. That means not attaching it to the existing paradigm and serving existing customers but targeting those not being served or not buying what's served, people we call nonconsumers. That way, all the new approach has to do is be better than a nonexistent alternative.

    Disruptive innovations tend to be simpler and more affordable than existing products. This allows them to take root in undemanding applications within a new market or arena of competition. They start to handle more complicated problems, and then they take over and supplant the old way of doing things. Sony chose to sell its tinny little transistor radio to teenagers who had never had a radio because they couldn't afford a tabletop RCA model. Bit by bit, that radio improved until, at some point, it became a superior alternative. Japanese car companies did this to Detroit. Nucor's mini mills did it to U.S. Steel. Google started out with ads from bicycle repair shops.

    Computer-based learning is a radar blip now but is moving up the adoption curve we've seen in many industries （see chart below）. Enrollments in state-accredited online courses went from 45,000 in 2000 to roughly 1 million today. That accounts for 1% of all courses, but we estimate that, given a looming shortage of teachers and widespread state budget crises, online learning will continue to gain market share until, by 2019, it surpasses live instruction.

    There are many areas of nonconsumption within schools where this is already taking place. One is Advanced Placement classes, those college-level courses offered to high school students. Schools offer only a fraction of the 34 courses for which AP exams are available. One-third of high school students attend schools that make no advanced courses available at all, according to a 2007 Department of Education report.

    Other pockets of nonconsumption include rural or small schools that are unable to offer breadth; prekindergarten courses; remedial courses students must take to graduate; and homeschooling, the choice for an estimated 2 million students today.

    As online classes improve with better video and social networking tools, they can get more customized and engaging. Costs should fall. Already it costs less to educate a student online （$200 to $600 per course） than it does in a classroom （$600）.

    Plenty of companies have sprung up around online learning. Apex Learning, started by Microsoft cofounder Paul Allen, started out by offering online AP courses that schools couldn't afford to offer. In the 2003–04 school year it enrolled 8,400 students. Last year that number was 30,200.

    A professor at Brigham Young University created the Virtual ChemLab in 2003, and it now serves 150,000 high school chemistry students across the country. The professor took 2,500 photos and 220 videos and got videogame designers to create a simulated laboratory to allow students to do online many of the same things they would do with real Bunsen burners. While virtual labs aren't as good as real ones, they are better than the alternative, which is nothing.

    A group of 50 Minnesota teachers laid off from rural schools because of declining enrollments banded together in 2004 to start an online charter school called Blue Sky. They each have the same number of students as they did before （150）, but the teachers describe their bond with each student as far stronger and more nuanced. The students call or e-mail at all hours because they are taking courses at all hours. Online, each student is an individual. Even children who were formerly behavior problems in school seem to have shaped up. It's hard to be a behavior problem in a class of one.

    電腦并沒有改善學習，那是因為沒有人被教會這種顛覆性技術。

    究竟為什么美國的公立學校要這么努力來變得更好？他們各自都有自己的理由。是缺乏資金嗎？也許吧，但是美國在從幼兒園到12年級的兒童教育上花在每個學生上的錢（9000美元）要比除了幾個國家的要多得多，結果仍然是滯后的。人們開始指責：學生的叛離、家長的忽視、頑固的教工聯(lián)，還有衡量績效方法的缺陷。

    所有這些因素都占了一部分的原因，但是根本的問題是更深層次的。它歸結為一個事實，即學校沒有去激勵孩子們，他們之所以沒有去激勵是因為他們都過于統(tǒng)一和標準化了。這個系統(tǒng)沒有證實每一個學生的學習方法都存在差異性這一事實。哈佛的心理學家哈沃德·加德納在二十世紀八十年代首次提出了"多元智能理論",并且至今已得到了廣泛的接受。現(xiàn)在他的分類系統(tǒng)將智力分為八個階段。你不能將邁克爾·喬丹（擅長動覺智力）的大腦和弗蘭克·勞埃德·懷特（擅長空間智力）的大腦或是沃爾特·惠特曼的（擅長語言的）來進行比較。每個學生有一個不同的混合的智力類型，其中是不同的學習和節(jié)奏型。

    19世紀老師們都在單室學校授課，這樣也不存在問題，老師可以將他們的方法個性化地傳授給各個學生。但是在19世紀交替之際，因為學校里每個教室有30或40個孩子，標準化就成了王道。學校變成了工廠，并從此一直抵制一切妄想打破這一統(tǒng)一批"處理"做法的努力。今天取得成功的學生們之所以能夠成功是因為他們的智力模型碰巧和在這個特定教室里使用的優(yōu)勢范例相匹配，或者是因為他們已經(jīng)在某種程度上適應了它。

    如果學校的目標是教育好所有的學生，那么他們應該嘗試去發(fā)現(xiàn)他們所有人的夢想，我們必須要找到一種打破統(tǒng)一的教學，把注意力放在以學生為中心的模式上的方法。實現(xiàn)這個方法要配合以計算機為基礎的學習。計算機技術提供給了學生學習的能力--通過在這里尋找符合他們智力類型的和適合他們學習節(jié)奏的方式。硬件有了。軟件正在開發(fā)。總之現(xiàn)在需要改變的是連接它們的系統(tǒng)。變化將面臨強大的阻力，但我們預測這將在未來的10年內(nèi)發(fā)生。

    對此持懷疑觀點的人會說，美國已經(jīng)花費了600億美元在過去20年里把電腦帶進了學校，并且仍然沒有改變這種課堂教學--盡可能節(jié)儉以防可能增加的成本、與其他學校的一度優(yōu)勢的方面也拉開了距離。這不應感到驚訝。學校已經(jīng)做了幾乎每一個組織在實施一項革新時都會做的：將它填補到其現(xiàn)有的教學模式，以保持之前的措施。美林公司開發(fā)了一項供客戶使用的可靠的網(wǎng)上交易系統(tǒng)。它能取代全球的電子貿(mào)易。但是美林所做的一切只是將新系統(tǒng)投放到掮客隊伍中，這樣他們可以更好地處理現(xiàn)有的客戶。它并沒有改變?nèi)魏螙|西。表面上如此完美地預期、完美地合乎常理--但它卻是一個徹徹底底地錯誤。

    用這樣的方法實施革新，實施起來會是組織遭到顛覆。那意味著不重視現(xiàn)有的模式和服務于現(xiàn)有的顧客，反而把那些沒有被給予服務的或沒有購買服務的人--我們稱之為"非消費群"作為目標。這樣一來，所有的新方法要做的就是實現(xiàn)比沒有選擇余地時做得更好。

    顛覆性創(chuàng)新往往是輕而易舉而又更比現(xiàn)有的產(chǎn)品花費更低。這使他們能夠扎根于要求不高的應用的一個新的市場或競爭性的領域。他們開始處理更復雜的問題，然后他們接管并取代舊的處事方式。索尼選擇了針對從沒有過收音機的青少年全體出售其迷你型的晶體管收音機，因為他們買不起臺式的RCA.一點一點的，收音機也不斷改進，最終，不知什么時候，它就成了不二的選擇。日本的汽車對底特律也是用同樣的手法；紐克的小廠成了現(xiàn)在的紐克鋼鐵公司；谷歌從以前的自行車修理廠變成了現(xiàn)在的廣告巨頭。

    以計算機為基礎的學習現(xiàn)在是雷達上的一點，但是我們能夠看見在多個行業(yè)中，它正沿著曲線不斷向上（見下表）.國家認可的在線課程的學員從2000年的4萬5千人猛增到今天的大約100萬。現(xiàn)在高中生里占有1%,但我們估計，由于教師短缺情況的出現(xiàn)和廣泛的國家預算危機，網(wǎng)上學習將繼續(xù)獲得市場份額，到2019年，將有超過50%的課程在網(wǎng)上教授。

    包括學校在內(nèi)的許多非消費領域，這已經(jīng)開始了。其中之一是跨級班，這些大學水平的課程提供給高中學生。學校只提供一小部分足以應對美聯(lián)社考試的34個課程。根據(jù)2007年教育部門的報告，有1/3的中學生參加網(wǎng)上課堂，這也使得沒有什么高級課程可供選用。

    其他零星的非消費團體包括農(nóng)村的或規(guī)模小的學校，他們無法提供廣泛的、學齡前的課程；學生畢業(yè)需要的補習課程；今天有大約200萬的學生會選擇家庭學校。

    隨著在線課程的改善--有了更好的視頻和社交網(wǎng)絡工具，他們可以得到更符合自身要求和吸引人的服務。花銷也降下來了。學生在線接受教育的花費（每堂課200--600美元）要低于在教室上課的花費（600美元）.

    大量的公司也迅速發(fā)展了在線學習。由微軟聯(lián)合創(chuàng)始人保羅艾倫始創(chuàng)的遠程教學，開始時是通過提供學校不能提供的在線課程。在2003--2004學年，招收了有8400名學生，去年是30200人。

    2003年，楊百翰大學的一位教授創(chuàng)建了虛擬化學實驗室，全國范圍內(nèi)現(xiàn)有15萬高中化學學生。該教授提供了2500張圖片和220個視頻，并請游戲設計師建立了一個模擬實驗，可讓學生用真的本生燈操做相同的實驗。雖然虛擬實驗室不如真實的，但是有總比沒有好。

    因為農(nóng)村學校生源的持續(xù)流失，明尼蘇達州的一群老師共50人下崗了，他們在2004年聯(lián)合起來在創(chuàng)建了一所網(wǎng)上學校--藍天。他們每人有和以前相同的學生數(shù)（150人）,但是他們坦言自己對每個學生付出的遠遠要更多更細致。學生們隨時都會打電話或是發(fā)郵件因為他們接受的是隨時授課。在網(wǎng)上，每一個學生都是單獨的個體。即使孩子們以前在學校里是有行為問題似乎已經(jīng)定型了。一個人的班級很難會有什么行為問題了。