以下是与 David Ferrucci 的一段对话。他曾领导构建 Watson 的团队，IBM 的这个问答系统在 Jeopardy 比赛中击败了世界上最顶尖的人类。在与 David 共度了几个小时后，我看到了他真正的热情，不仅是对智能的抽象理解，更是为了在现实世界的最后期限和资源限制下，将其工程化以解决现实世界的问题。科学与工程相遇之处，便是卓越的、简单的创造力涌现之地。那些与之相邻工作的人，拥有通过两次失败和最终成功所获得的丰富智慧。David 也是 Elemental Cognition 的创始人、首席执行官兼首席科学家，该公司致力于构建能够像人类一样理解世界的 AI 系统。这是人工智能播客。如果您喜欢，请在 YouTube 上订阅，在 iTunes 上给予五星好评，在 Patreon 上支持我们，或者在 Twitter 上与我联系：Alex Friedman，拼写为 Fri D M a.m.。现在，这是我与 David Ferrucci 的对话。

您的本科专业是生物学，目标是进入医学院，之后您才继续攻读计算机科学博士学位。那么，让我问您一个简单的问题：生物系统和计算机系统之间有什么区别？当您坐下来仰望星空，进行哲学思考时，我常常会想，我常常会想，两者之间是否存在实质性的区别。我认为让我进入计算机科学和人工智能领域的原因正是这个前提：如果我们能让机器思考，或者我应该说，这个哲学问题，如果我们能让机器思考、理解、处理信息，就像我们一样，那么如果我们能描述一个过程，即使那个过程本身就是智能过程，那么区别又在哪里呢？所以，从哲学的角度来看，我并不是想说服您存在，我是说，您可以走向灵性，您可以走向灵魂，但就我们所能从智力和身体的角度体验到的而言，我不确定是否存在。显然，它们的实现方式不同，但如果您说，作为一个生物信息处理系统，其能力是否从根本上比我们可能用硅或其他基质构建的系统更强大？我不知道。

您认为生物实现方式距离我们有多远？所以，从根本上说，它们可能具有相同的能力，但它真的是一个巨大的谜团，需要大量的突破才能理解它，还是在很大程度上，在重要的方面，它与我们可能构建的系统具有相同的特征？是的，这很有趣。我的意思是，您的问题预设了目标是重现我们所认为的生物智能。我不太确定那是否是我的目标陈述方式。我的意思是，我认为研究……好的，我认为有几个目标。我认为理解人脑及其工作方式对我们诊断和治疗问题很重要，对我们理解自身的智力、心理和身体上的优势和劣势也很重要。所以，从这个角度来看，神经科学和理解大脑有一个清晰明确的目标。从我想模仿人类智能的角度来看，这更有趣。人类智能当然有很多我们羡慕的地方，但也有很多问题。所以，我认为我们可以退一步说，我们想要从中得到什么？我们想要从智能中得到什么？我们想如何与这种智能交流？我们想如何表现？我们想让它如何运作？当然，这是一个有点有趣的争论，因为我坐在这里，作为一个拥有生物大脑的人，我正在批评人类智能的趋势和弱点，并说我们有能力，仅仅是能力，退一步说，“天哪，什么是智能？我们到底想从中得到什么？”而这本身就表明，人类智能是相当可爱的，它可以……它可以反思，它可以那样反思。

您提到的那些缺陷，人类整体的……是的，但我认为，我认为人类整体所拥有的缺陷是极其偏颇和带有偏见的，它在得出许多推论时是这样的。您认为那些是……抱歉打断您。您认为那些是特征还是错误？您认为那些偏见、健忘、恐惧……列出所有其他缺陷……也许爱是一种缺陷？您认为这些都是阻碍智能的因素，还是智能的必要组成部分？嗯，再说一遍，如果您回去定义智能是能够准确、精确、严谨地推理，得出答案并以客观的方式证明这些答案，那么人类智能就有这些缺陷，它倾向于受到您所说的一些事情的影响，而且它在很大程度上是一个归纳过程，也就是说，它以过去的数据来预测未来。在某些情况下非常有利，但在其他情况下，它从根本上是带有偏见和偏见的，因为它会受到其先验知识的强烈影响，无论它们是否……从某种客观推理的角度来看，它们是对还是错，您都会倾向于它们，因为那是过去成功的决定，或者那些是过去成功的道路。我认为这种智能模式在您的主要目标是快速行动、生存和做出快速决定时是有意义的。而我认为，当您想更深入地思考并做出更客观、更合理的决定时，这些会产生问题。当然，人类有能力做到这两者，他们确实比做另一件事更自然地做其中一件事，但他们有能力做到这两者。

您说他们更自然地做那个快速响应的，对吗？因为那是你需要避免被世界上的捕食者吃掉的那种东西。但我的意思是，我们已经学会了通过逻辑进行推理，我们发展了科学，我们训练人们这样做。我认为这对个人来说更难。我认为这需要训练，以及……以及……以及教学。我认为我们人类……当然有能力做到，但我们觉得更困难。然后还有其他弱点，如果您愿意，就像您之前提到的，只是记忆容量，以及您可以在不迷失方向的情况下进行多少链式推理。所以，只是专注于……所以，您看待智能的方式，我们真的在进行一些哲学上的探讨，但我认为您是谈论这个的完美人选，因为我们将谈论 Jeopardy 以及更广泛的领域，那是在人工智能历史上最令人难以置信的成就之一。因此，哲学讨论。那么，让我问一下，您已经暗示了，但让我再次问一下：什么是智能？在我们将要进行的关于 Jeopardy 及更广泛领域的讨论中，您如何看待智能？它是足够复杂的问题吗？能够通过推理来解决这个问题，这就是您对智能的理解方式吗？

我认为智能主要有两种方式。一种是预测能力。也就是说，如果我有一个问题，会发生什么？我能预测接下来会发生什么吗？无论是预测一个问题的答案，还是说，“看，我正在观察所有的市场动态，我将告诉您接下来会发生什么。”或者您在一个房间里，有人走了进来，您将预测他们接下来会做什么，或者他们接下来会说什么。在充满不确定性的高度动态的环境中做到这一点，能够……锁定很多……变量越多，越复杂，可能性越多，越复杂。但我能否利用少量先验数据来学习模式，然后准确地预测接下来会发生什么？这当然是一种智能。顺便问一下，您需要什么？您需要了解世界运作的方式，才能将其展开到未来。好的，谢谢。需要预测一件事，这取决于您对理解的定义。我需要能够找到那个函数，这与深度学习、机器学习所做的非常相似。如果您给我足够多的先验数据，并告诉我重要的输出变量是什么，我将坐在这里，能够预测它。如果我能准确地预测它，这样我就可以……比不正确的时候更频繁地得到正确答案，我就是聪明的。如果我用更少的数据和更少的训练时间做到这一点，我就会更聪明。如果我能弄清楚什么值得预测，我就会更聪明，也就是说，我正在弄清楚哪条路径将引导我实现目标。

那么，关于选择目标呢？再次……嗯，这很有趣。选择目标本身就很有趣。我认为这就是您引入了什么？我们被预先编程做什么？我们谈论人类，人类被预先编程为生存。所以，他们的主要……您知道，驱动目标是什么？他们需要做什么才能做到这一点？而这可能非常复杂，对吧？所以，这不仅仅是弄清楚你需要逃离凶猛的老虎，而是我们以社会为例生存。所以，理解社会动态的微妙之处就成了生存、寻找配偶、繁殖的重要因素，对吧？所以，我们不断面临复杂的变量集、复杂的规则，如果您愿意，或者模式，我们学会了找到函数并预测事物。换句话说，高效地表示这些模式，并能够预测会发生什么。这是一种不需要任何特定东西的智能，除了找到那个函数并……预测那个正确答案的能力。这当然是一种智能。但是，当我们说，“我们互相理解吗？”也就是说，您是否认为我比那种预测能力更聪明？现在我可以预测，但我无法真正阐述我将如何……我的预测过程是什么？我的根本理论是什么？我也无法让您理解我在做什么，以便您能够跟随，您能够弄清楚如何自己做这件事，如果您没有……如果您没有，例如，我拥有的正确的模式匹配机制。现在我们可能会出现这种 breakdown，实际上，我是聪明的，但我有点像一个外星智能，相对于您。您是聪明的，但没有人知道。或者我可以看到……我可以看到输出……所以，您是说，让我们把这两件事分开。一是您解释为什么您能够预测未来，二是……二是让我……让我印象深刻，您很聪明，让我能够知道您成功地预测了未来。您认为那是……嗯，这不是一个令人印象深刻的事情。也就是说，您可能确信我以某种形式是聪明的。所以，我……我猜想……哇，您是对的。您……您比我更正确。您正在做一些有趣的事情。这是一种……这是一种智能。但是，如果我说，“您是怎么做到的？”而您无法与我沟通，也无法向我描述。现在，我将称您为天才。我的意思是，我可能会说，“您在做一些奇怪的事情，它……它对我来说并不那么有趣，因为您和我真的无法沟通。”所以，现在这很有趣，对吧？因为现在您处于一个奇怪的境地，那就是，要被您认为聪明，就像我一样聪明，那么您和我必须能够沟通。然后，我们开始互相理解，然后我的尊重、我的欣赏、我与您建立联系的能力就会改变。所以，您不再是外星智能了。您是……您是我们的人类智能，因为您和我能够沟通。所以，我认为当我们看待……当我们看待……当我们看待动物时，例如，动物可以做我们无法完全理解的事情，我们不知道它们是如何做到的，但它们无法真正与我们沟通。它们无法用我们的术语来表达它们所经历的事情。所以，我们认为它们是……它们是外星智能，它们并不真正……我们不认为它们有价值。因此，我们对待它们的方式不同。但很难，因为谁知道……谁知道发生了什么。

所以，只是对这一点进行一个简短的阐述。解释您很聪明，解释推理，一方面是预测，另一方面是……另一方面是我能够……让我印象深刻的是您很聪明。让我能够知道您成功地预测了未来。您认为那是……嗯，这不是一个令人印象深刻的事情。也就是说，您可能确信我以某种形式是聪明的。所以，我……我猜想……哇，您是对的。您……您比我更正确。您正在做一些有趣的事情。这是一种……这是一种智能。但是，如果我说，“您是怎么做到的？”而您无法与我沟通，也无法向我描述。现在，我将称您为天才。我的意思是，我可能会说，“您在做一些奇怪的事情，它……它对我来说并不那么有趣，因为您和我真的无法沟通。”所以，现在这很有趣，对吧？因为现在您处于一个奇怪的境地，那就是，要被您认为聪明，就像我一样聪明，那么您和我必须能够沟通。然后，我们开始互相理解，然后我的尊重、我的欣赏、我与您建立联系的能力就会改变。所以，您不再是外星智能了。您是……您是我们的人类智能，因为您和我能够沟通。所以，我认为当我们看待……当我们看待……当我们看待动物时，例如，动物可以做我们无法完全理解的事情，我们不知道它们是如何做到的，但它们无法真正与我们沟通。它们无法用我们的术语来表达它们所经历的事情。所以，我们认为它们是……它们是外星智能，它们并不真正……我们不认为它们有价值。因此，我们对待它们的方式不同。但很难，因为谁知道……谁知道发生了什么。

所以，只是对这一点进行一个简短的阐述。解释您很聪明，解释推理，一方面是预测，另一方面是……另一方面是我能够……让我印象深刻的是您很聪明。让我能够知道您成功地预测了未来。您认为那是……嗯，这不是一个令人印象深刻的事情。也就是说，您可能确信我以某种形式是聪明的。所以，我……我猜想……哇，您是对的。您……您比我更正确。您正在做一些有趣的事情。这是一种……这是一种智能。但是，如果我说，“您是怎么做到的？”而您无法与我沟通，也无法向我描述。现在，我将称您为天才。我的意思是，我可能会说，“您在做一些奇怪的事情，它……它对我来说并不那么有趣，因为您和我真的无法沟通。”所以，现在这很有趣，对吧？因为现在您处于一个奇怪的境地，那就是，要被您认为聪明，就像我一样聪明，那么您和我必须能够沟通。然后，我们开始互相理解，然后我的尊重、我的欣赏、我与您建立联系的能力就会改变。所以，您不再是外星智能了。您是……您是我们的人类智能，因为您和我能够沟通。所以，我认为当我们看待……当我们看待……当我们看待动物时，例如，动物可以做我们无法完全理解的事情，我们不知道它们是如何做到的，但它们无法真正与我沟通。它们无法用我们的术语来表达它们所经历的事情。所以，我们认为它们是……它们是外星智能，它们并不真正……我们不认为它们有价值。因此，我们对待它们的方式不同。但很难，因为谁知道……谁知道发生了什么。

所以，只是对这一点进行一个简短的阐述。解释您很聪明，解释推理，一方面是预测，另一方面是……另一方面是我能够……让我印象深刻的是您很聪明。让我能够知道您成功地预测了未来。您认为那是……嗯，这不是一个令人印象深刻的事情。也就是说，您可能确信我以某种形式是聪明的。所以，我……我猜想……哇，您是对的。您……您比我更正确。您正在做一些有趣的事情。这是一种……这是一种智能。但是，如果我说，“您是怎么做到的？”而您无法与我沟通，也无法向我描述。现在，我将称您为天才。我的意思是，我可能会说，“您在做一些奇怪的事情，它……它对我来说并不那么有趣，因为您和我真的无法沟通。”所以，现在这有趣，对吧？因为现在您处于一个奇怪的境地，那就是，要被您认为聪明，就像我一样聪明，那么您和我必须能够沟通。然后，我们开始互相理解，然后我的尊重、我的欣赏、我与您建立联系的能力就会改变。所以，您不再是外星智能了。您是……您是我们的人类智能，因为您和我能够沟通。所以，我认为当我们看待……当我们看待……当我们看待动物时，例如，动物可以做我们无法完全理解的事情，我们不知道它们是如何做到的，但它们无法真正与我们沟通。它们无法用我们的术语来表达它们所经历的事情。所以，我们认为它们是……它们是外星智能，它们并不真正……我们不认为它们有价值。因此，我们对待它们的方式不同。但很难，因为谁知道……谁知道发生了什么。

这个我从未听说过足球这款游戏，但如果它像篮球一样，目标就像篮筐，我必须把球投进篮筐，而且我有后卫，我有这个，我有那个，就像我们很奇怪一样，相似之处和不同之处在哪里？我现在有了一个基础来解读这些新信息，然后不同的群体，比如千禧一代，将拥有一个框架，然后，然后，嗯，你永远不知道，是的，嗯，克拉茨和共和党人，嗯，尼尔，没有人想要那个框架，嗯，我的意思是，我认为他理解它，对吧？我的意思是，你谈论的是政治和社会的方式来解读周围的世界，我认为这些框架在很大程度上仍然是相似的，我认为它们可能在一些基本假设和价值观上有所不同，现在从推理的角度来看，比如处理框架的能力，玛格娜可能没有那么不同，不同基本价值观或基本假设在那些框架中的含义可能得出非常不同的结论，所以从社会角度来看，结论可能非常不同，从智力角度来看，我，你知道，我只是跟着我的假设走，是的，产品本身的过程看起来会很相似，但这是一个有趣的观点，框架确实有助于塑造一个陈述如何被解读，我的意思是，拥有一个民主党和共和党的框架，并阅读完全相同的陈述，你得出的结论将完全不同，从广告的角度来看，这是迷人的，我们希望从人工智能中获得的是能够告诉你，这个视角，一个视角，一套假设将把你带到这里，在其他设置中，午餐将把你带到那里，事实上，你知道，帮助人们推理并说，哦，我看到了，我看到了我们的分歧所在，是的，你知道，我对那个有一个基本信念，我对那个有一个基本信念，是的，从我的角度来看，这是非常棒的，在自然语言处理方面，有一种观点认为理解一个陈述只有一种方式，但可能不是，可能有无数种方式，而且同样，嗯，有很多发现，有很多不同的解释，你知道，越广泛，你知道，内容越广泛，内容越丰富，所以，你知道，你和我对同一段文字可能有非常不同的体验，显然，如果我们致力于理解彼此，我们就开始，这是另一个重要的观点，比如如果我们致力于理解彼此，我们就开始分解和打破我们的解释，朝着越来越原始的组成部分，直到我们达到那个点，说，哦，我明白了为什么我们不同意，我们试图理解这种分歧到底有多根本，但这需要一种承诺，以逻辑的方式分解解释，否则，你知道，这就是为什么我喜欢，我认为人工智能确实是在补充和帮助人类智能克服它的一些偏见和它被说服的倾向，你知道，通过，但更肤浅的推理，我的意思是，我们克服了这种想法，嗯，我，你知道，你知道，我是对的，因为我是共和党人，或者我是对的，因为我是民主党人，有人贴上了民主党观点的标签，或者它有以下关键词，如果机器能帮助我们分解这个论点，并说，等等，你知道，你真正认为这是什么？所以基本上是让我们对进行更批判性的思考负责，坐下来思考这一点，因为我喜欢，我认为这是对公众话语的非常赋权的人工智能的使用，它目前完全在瓦解，我不知道，因为我们在社交媒体上学习如何做到这一点，所以人工智能历史上最伟大的成就之一是沃森在问答游戏中与人类竞争，你是那个关键部分的主要参与者，让我们从最基本的地方开始，问答游戏是什么？对我们人类来说，问答游戏是人类对人类的比赛，对吧？所以它是要提出一个问题并回答它，实际上不是，但它不是，它真的不是，它真的，它真的要提出一个问题并回答它，但我们称之为事实性问题，所以这种想法，比如它真的涉及到一些事实，很少有人会争论事实是否真实，事实上，大多数人在问答游戏中都依赖于这些陈述有事实答案的想法，想法是首先确定你是否知道答案，这是一个有趣的转折，所以首先要理解问题，你必须理解问题，它在问什么，这是一个很好的观点，因为问题不是直接提出的，对吧？它们都是，比如提问的方式是非线性的，有点诙谐，有点俏皮，有时有点棘手，是的，它们被提出，而且是以无数诙谐而棘手的方式提出的，正是它们在问什么并不明显，需要有经验的人花一段时间才能弄清楚它到底在问什么，你有一个有趣的认识，哦，问答游戏是一个问答节目，然后就像，我知道很多，然后你读它，你仍然在试图处理问题，而冠军们已经回答并继续前进，一次有三个问题在你弄清楚问题是什么意思之前，所以肯定有一种能力，可以解析出问题是什么，所以这当然具有挑战性，但从历史上看，如果你回顾更早的问答游戏，你知道，63年，我认为问题更直接，不像那样，它们变得越来越有趣，提问的方式也变得越来越有趣、微妙、细致、幽默和诙谐，这确实需要人类来建立正确的联系，弄清楚问题到底是什么意思，所以是的，你必须弄清楚问题到底是什么意思，然后你必须确定你认为你是否知道答案，因为你必须非常快地抢答，所以你必须尽快做出决定，否则你就会失去抢答的机会，在你真正知道你是否知道答案之前，你已经开始回答了，我认为，嗯，我认为很多人会假设他们会，他们会非常肤浅地看待他们的过程，换句话说，主题是什么？有哪些关键词？然后说，我是否了解这个领域？在他们真正知道答案之前，然后他们会抢答，然后我会抢答，然后思考一下，人类现在所做的很有趣，有些人，比如肯·詹宁斯，或者最近的那个伟大的问答选手，他知道那些，就假设他们知道，虽然问答游戏，我会把它抛出来，你知道，沃森有趣的是，它甚至没有接近知道所有问答游戏，沃森甚至在巅峰时期，甚至在，是的，所以例如，我的意思是，我们有这个叫做回忆的东西，就像所有问答问题中，有多少我们能找到正确的答案，任何地方，我们可以找到，我们可以想出，如果我们看，你知道，我们有一个庞大的知识库，大约几太字节，我的意思是，从网络规模来看，实际上非常小，但从书籍规模来看，谈论数百万本书，对吧？所以相当于数百万本书和百科全书，字典，书籍，这是一吨信息，你知道，对于，我认为是80，只有85%的答案可以在任何地方找到，嗯，所以你准备好了，你准备好了，只是为了开始，对吧？所以，重要的是要非常快速地了解你是否认为你知道这个问题的正确答案，所以我们必须尽快计算出这个信心，所以实际上是回答它，至少，你知道，花一些时间来回答它，然后判断我们，你知道，我们的答案是正确的，并决定我们是否足够自信地抢答，这取决于游戏中正在发生什么，因为这是一种风险，所以如果你真的处于我必须猜测的情况，我几乎没有什么可失去的，那么你就会以较低的信心抢答，所以这是为了计算不同竞争对手的财务状况，是的，游戏进行了多久，还剩多少时间，你在什么位置，等等，我们谈论的是多少百毫秒？你有没有概念，我们瞄准的是什么，因为我们，是的，我们瞄准的是，我的意思是，我们瞄准的是在三秒内回答，然后抢答，所以抢答的决定，然后实际回答，是这两者吗？是的，有两件事，事实上，我们有多个阶段，比如我们会说，让我们估计一下我们的信心，这是一种肤浅的回答过程，然后最终，然后最终决定抢答，然后我们可能会再花一秒钟左右，去那里做这件事，但总的来说，我们说，我们不能玩这个游戏，我们甚至不能竞争，如果我们不能平均在三秒或更短的时间内回答这些问题，所以你介入了，所以有，有这三个人在玩游戏，你介入了，想法是IBM沃森将取代其中一个人，并与两个人竞争，你能讲讲沃森接受这个游戏的故事吗？当然，这似乎异常困难，是的，所以故事是，嗯，它即将到来，我认为是大蓝，深蓝的十周年纪念日，IBM想做另一个，你知道，有趣的挑战，公众挑战，能够引起人们对IBM研究的关注，以及我们正在做的很酷的事情，我已经在IBM从事人工智能工作一段时间了，我有一个团队在做所谓的开放域事实性问题回答，这意味着，我们不会告诉你问题是什么，我们甚至不会告诉你它们是关于什么的，你能去那里找到这些问题的准确答案吗？这是一个我参与过的人工智能研究领域，所以这是一个很大的，这是我非常热衷的一个领域，语言理解，一直是我热衷的领域，关于是否能做任何事情，语言方面的一个狭窄切片是开放域的概念，也就是说，我可以问任何关于任何事情的问题，事实性问题，也就是说，它基本上有一个答案，你知道，并且能够准确而快速地做到这一点，所以这是我已经在研究的领域，然后完全独立地，几个，你知道，IBM，他们想，我们要做什么？下一个很酷的事情是什么？肯·詹宁斯正在连胜，这是，比如，2004年，我认为是在他连胜的时候，有人想，嘿，如果电脑能玩问答游戏，那将是多么酷啊，所以这是在2004年，他们到处兜售这个东西，每个人都在告诉研究主管，不可能，这太疯狂了，我们有一些相当，你知道，高级人物，他们会把它交给我，我当时想，但这正是我真正感兴趣的，但普遍存在一种感觉，我们不会拿IBM的声誉冒险，我们就是不做，这发生在2004年，发生在2005年，2006年底，它又来了，我刚结束，我正在做开放域问题回答的东西，但我刚结束了几个其他项目，我有更多的时间投入其中，我争辩说，这可以做到，我争辩说，不这样做是疯狂的，你能诚实地说吗？所以即使你为它辩护，你对自己有多大信心，这可以做到？我们完全讲了你如何讲故事来劝说他人，你有多自信？你当时对问题的估计是什么？所以我认为这是可能的，很多人认为这是不可能的，我认为这是可能的，我之所以认为这是可能的，是因为我做了一些简短的实验，我非常了解我们是如何处理开放域事实性问题回答的，我们已经做了很多年了，我看了问答的东西，我说，这对你之前提到的很多观点来说都会很难，很难理解问题，很难足够快地做到，很难计算出准确的信心，这些事情以前都没有做得足够好，但很多我们正在构建的技术，应该会起作用，但更重要的是，驱动我的是，我在IBM研究，我是IBM研究的高级领导，这是我们应该做的事情，我们基本上应该做登月项目，我的意思是，我们应该说，这是一个活跃的研究领域，我们有责任，如果我们有机会，就把它推向极限，如果它不起作用，就更深入地了解为什么我们做不到，所以我非常致力于这个想法，说，各位，这就是我们所做的，不这样做是疯狂的，这是一个活跃的研究领域，我们已经在这方面工作了很多年，为什么我们不接受这个宏大的挑战，并尽力而为？至少我们可以说，为什么这个问题如此之难，我们尝试了什么，以及我们是如何失败的，所以我从那个角度非常受科学家驱动，然后我又根据我们做的可行性研究进行了辩论，哦，为什么我认为它很难但可能，我举了例子，你知道，它在哪里成功，在哪里失败，为什么失败，以及为什么我们应该这样做的总体架构方法，但大多数时候，当时的执行官们只是在寻找一个足够疯狂的人来说是，因为在那几年里，每个人都说了不，我不想拿我的声誉和我的职业生涯冒险，你知道，关于这件事，显然你没有这样的恐惧，好吧，我没有说你死了，是的，而且据我所知，它一开始表现非常糟糕，那么最初的方法是什么？为什么它们会失败？有很多困难的方面，我的意思是，我们过去尝试过的先前方法失败的原因之一是因为问题很难，很难解释，你到底在问什么？很多时候，比如，如果问题非常直接，比如什么城市，你知道，或者什么，你知道，即使那样也可能很棘手，但是，你知道，什么城市或什么人，通常是它非常清楚地命名它，你就知道了，而且如果只有一小部分，换句话说，我们要问这五种类型，它将有一个答案，答案将是这个州的城市，或者这个国家的城市，答案将是某人，比如演员之类的，但事实证明，在问答游戏中，有数万种这样的东西，而且它是一个非常非常长的尾巴，这意味着，你知道，它一直在继续，所以即使你专注于试图编码顶部的类型，比如有五种是最多的，比如最常见的五种，你仍然只覆盖了很小一部分数据，所以你不能采取这种方法，说，我只是想收集关于这五种或十种或二十种或五十种类型的事实，所以这是第一件事，你该怎么办？所以我们想出了一个方法，这个方法看起来很有希望，我们继续改进我们处理这个问题的能力，整个项目，另一个问题是，从一开始我就说，我们不会，我承诺在三到五年内完成，所以我们用了四年，我运气好，但是，把这个，比如，设定一个目标，我知道语言的标准问题有多难，我说，我们不会真正理解语言来解决这个问题，我们不会解释问题和问题所指的知识领域，然后推理来回答这些问题，显然，我们不会同时这样做，简单的搜索也不足以自信地回答，你知道，一个正确的答案，首先，其他人，比如，这是多么棒的创新和实际工程的结合，所以你不是，你不是在试图解决一般的自然语言理解问题，你说，让我们以任何可能的方式解决它，哦，是的，我承诺说，看，我们要解决开放域问题回答问题，我们正在使用问答游戏作为驱动力，这是一个足够大的基准，正是如此，现在我们如何做到？我们就像，随便，随便想办法，因为我想回到学术科学界，说，我们尝试了什么，什么有效，什么无效，我不想进去说，哦，我只有一个技术锤子，我只会用这个，我将做任何事情，我将跳出框框思考，做任何事情，我也相信，我坚信，基本的自然语言处理技术和机器学习技术将足够，这是一个如何增强它们，如何集成它们，如何推进它们的问题，所以我有一个研究员来找我，他和我一起研究问题回答了很长时间，他说，我们需要麦克斯韦方程来回答问题，我说，如果我们，如果我们需要的公式能打破我们对语言的理解，我们就完蛋了，是的，我们不会从这里到达那里，我的意思是，我不是在指望，我的假设是，我不是在指望某个全新的发明，我指望的是能够利用以前所做的一切，找出如何很好地集成它的架构，然后看看它在哪里失败，并做出我们需要的必要改进，所以这个东西就能工作，是的，用力推，看看它在哪里失败，然后修补它，我的意思是，人们改变世界就是这样，这就是，你知道，大多数方法，火箭，SpaceX，这就是亨利·福特，等等，我碰巧，在这种情况下，我碰巧是对的，但是，就像我们不知道一样，对吧？但是你必须花一秒钟左右，你将如何运行项目，所以，是的，然后回溯到搜索，如果你要这样做，什么是暴力解决方案，你会搜索什么？所以你有一个问题，你如何搜索可能的答案空间？网络搜索已经走了很长一段路，即使在那之后，但当时，你知道，首先，我的意思是，问题有一些其他的限制，很有趣，所以你不能去网上，你不能搜索互联网，换句话说，人工智能实验是，我们想要一个独立的设备，如果设备像房间一样大，没关系，它像房间一样大，但我们想要一个独立的设备，你不能上网，你没有生命线，所以它必须，如果你愿意，装在一个鞋盒里，或者至少是几个冰箱的大小，无论它可能是什么，但是，你也无法摆脱，你不能离开网络，所以有这个限制，但然后我们做到了，但基本的事情是，去，去搜索网络，问题是，即使我们去搜索网络，我记不清确切的数字了，但有时，大约65%的时间，答案会在前10或20份文件中，所以首先，这甚至不足以玩问答游戏，你知道，即使你能提取，即使你能完美地从前20份文件中提取答案，前10份文件，无论是什么，我们不知道怎么做，但即使你能做到，你也会知道它是对的，你有足够的信心，对吧？所以你必须提取正确的答案，你必须有信心它是正确的答案，然后你必须足够快地做到这一点，然后去抢答，你仍然只有65%的正确率，九个甚至不能让你进入胜利者的圈子，胜利者的圈子，你必须超过70，而且你必须非常快地做到，而且你做得非常快，但现在的问题是，即使我在前10份文件中找到了答案，我如何确定答案在前10份文件中在哪里，以及我如何计算所有可能候选者的信心？所以这不像我进去就知道正确答案，我必须选择它，我不知道正确答案，我有一堆文件，正确的答案就在里面，我如何，作为机器，去找出哪个是正确的，然后我如何给它评分？所以，现在我如何处理我实际上不能上网的事实？首先，如果你停下来思考一下，如果你能上网，你认为这个问题可以解决吗？如果你只是停下来思考一下，即使是问答游戏之外，你认为阅读文本并找到答案的问题可以解决吗？我们解决了，在某种意义上解决了，考虑到问答游戏的挑战，你是怎么做到的？所以，你是如何获取工作成果和特定主题，并提取关键信息？所以，现在忘记网上巨大的信息量，对吧？所以现在我们必须弄清楚，我们做了很多来源研究，换句话说，哪些知识库足够小但又足够广泛，可以回答问答游戏？我们最终找到了那个知识库，我的意思是，它包括维基百科和其他一些东西，所以像百科全书类型的东西，我不知道你是否使用玛丽的，不同类型的语义资源，比如WordNet和其他类似的语义资源，以及一些网络爬行，换句话说，我们出去，获取内容，然后基于统计生产，你知道，统计生产，使用这些来搜索其他搜索，然后扩展，所以使用这些，比如扩展技术，我们出去，找到了足够的内容，我们说，好的，这很好，我们甚至完全，你知道，我们一直有资源威胁，一直在试图弄清楚哪些内容我们可以有效地包含，我的意思是，有很多流行的剪辑，比如，什么是教堂女士？我认为是其中一个，嘿，是的，我们准备好了，我认为那可能是一本百科全书，所以它是佩皮诺，但是，然后我们会，然后我们会把那些东西拿出来，然后我们会出去，我们会扩展，换句话说，我们会找到核心资源中没有的其他内容，并扩展，你知道，内容量增长了十倍，但仍然，所以再次，从网络规模来看，这是非常少的内容，它是非常有选择性的，然后我们把所有这些内容，我们，我们对它进行了大量的预分析，我的意思是，我们，我们解析了它，你知道，把它分解成所有这些单独的词，然后我们对它进行了语义统计和语义解析，你知道，让计算机算法对其进行注释，然后我们把它索引到一个非常丰富且非常快速的索引中，所以我们有相当大的数量，你知道，为了论证的缘故，两到五百万本书，我们已经分析了所有这些，尺寸甚至更大，因为现在有了所有这些元数据，然后我们丰富地索引了所有这些，并将其放入一个巨大的内存缓存中，所以沃森没有去磁盘，所以基础设施组件，如果你只是谈论它，有多难？我的意思是，我知道2000年，也许是2008年，你知道，那已经是很久以前了，对吧？使用多台机器有多难？基础设施部分，硬件组件，我们使用了IBM，所以我们使用了IBM硬件，我们有大约，我不知道确切数字，但2000到3000个核心，完全连接，所以有一个交换机，你知道，每个CPU都连接到每个其他CPU，它们以某种方式共享内存，劳伦，接近共享内存，对吧？所有这些数据都经过预先分析，并放入一个非常快速的索引结构中，所有这些都在内存中，然后我们把这个问题拿出来，我们分析问题，所以所有内容都经过预先分析，所以如果我，所以如果我试图找到一段内容，它会带回我们预先计算的所有元数据，你如何将那个问题塞进去？你如何连接，大的东西和元数据，大的知识库和元数据，并将其索引到简单、诙谐、令人困惑的问题？所以，沃森架构师就在这里，所以我们会把问题拿出来，我们会分析问题，所以这意味着我们会解析它，并以多种不同的方式来解释它，我们试图弄清楚它在问什么，所以我们会，我们会想出多种策略来，弄清楚它在问什么，这可能表示为一个简单的字符串，字符字符串，或者是一些我们可以连接回现有资源的各种语义类型，所以，总之，我们会做大量的分析，问题和问题分析必须完成，而且必须快速完成，所以我们做问题分析，因为然后从问题分析中，我们会产生搜索，所以我们会，我们已经构建了使用开源搜索引擎，我们修改了它们，我们有许多不同的搜索引擎，我们使用它们，它们具有不同的特性，我们进去，设计和修改了那些搜索引擎，最终，为了现在，把我们的问题分析，产生多个查询，基于对问题的不同解释，然后并行地发射大量的搜索，它们会产生，伴随着段落，所以这些是，被动的搜索算法，它们会返回段落，所以现在，假设你有大约一千个段落，现在对于每个段落，你再次并行化，所以你出去，你并行化，并行化搜索，每次搜索都会返回一大堆段落，也许你总共有大约一千或五千个不同的段落，现在对于每个段落，你不再弄清楚是否有候选者，它会称之为候选答案，所以你还有一大堆其他的，一大堆其他的算法，它们会找到候选答案，可能的答案，所以你有一大堆候选答案，候选答案生成器，一大堆，所以对于这些组件中的每一个，团队都在不断地研究，寻找更好的方法来从问题生成搜索查询，更好的方法来分析问题，更好的方法来生成候选者，以及速度，所以更好的准确性和速度，破解，所以，是的，速度和准确性，在很大程度上，它们是分开的，我们以不同的方式处理，我纯粹关注准确性，以及准确性，我们是否最终得到了更多的问题，并产生了更准确的信心，他们有一个完整的团队，一直在分析工作流程，寻找瓶颈，然后，如果你要并行化并驱动算法速度，但无论如何，所以，所以现在想象一下，你有一个巨大的扇出，对吧？因为你有，你有多个查询，现在你有，现在你有数千个候选答案，对于每个候选答案，你都要给它评分，所以你将使用所有构建的数据，你将使用问题分析，你将使用查询是如何生成的，你将使用段落本身，你将使用生成的候选答案，你将给它评分，所以现在我们有一群研究人员在给出分数，有数百种不同的分数，所以现在你又得到了一个扇出，从你拥有的候选答案数量到所有不同的评分者，所以如果你有200个不同的评分者，你有一千个候选者，现在你有两千个分数，所以，现在你必须弄清楚，你知道，我如何对这些答案进行排名，根据返回的分数，我想根据它们是问题正确答案的可能性来排名，所以每个分数都是一个独立的研究项目，你说的分数是什么意思？所以那是人类在说，这个答案，你认为，你认为，如果你想，你想想你在做什么，你知道，如果你想考虑一个人类会做什么，人类会看着一个可能的答案，他们会读，你知道，艾米莉·狄金森，他们会读发生这种情况的段落，他们会看着问题，他们会做出决定，鉴于这些证据和这段文字，艾米莉·狄金森是这个问题的正确答案的可能性有多大？明白了，所以，那是注释任务，斯坦·约翰逊的评分任务，但是评分意味着零到一的某种连续性，不是二元的，不，给它一个分数，给它一个零，是的，确切地说，这是人类所做的，给出不同的分数，所以你必须以某种方式标准化，处理所有这些，取决于你的策略，我们可以相对而言，我们可以，我们实际上也看了原始分数，标准化分数，因为人类不参与其中，人类不参与，抱歉，所以，我的意思是，我误解了，过程，这是段落，真实情况来自哪里？真实情况，有问题的答案，所以它是端到端的，它是端到端的，所以我们也，我一直在驱动，性能，一个非常有趣，一个非常有趣的工程方法，最终是科学和研究人员，个人，总是驱动端到端，那不是说我们不会提出假设，单个组件的性能以某种方式与端到端性能相关，当然我们会，因为人们必须构建单个组件，但最终，要让你的组件集成到系统中，你必须展示对端到端性能的影响，问题回答性能，因为有很多人在研究这个问题，他们基本上是在推销他们的想法，作为系统的一部分，那就是，他们会对他们的组件进行研究，他们会说，比如，我将改进这个作为候选生成器，我将改进这个作为问题评分者，或者作为段落评分者，我将改进这个，或者作为类型估计器，或者其他什么，然后我会说，我需要理解，改进那个计算机指标如何影响端到端性能，如果你不能估计它，也不能进行实验来证明它，它就不能进入，这是我听过的最好的AI项目，太棒了，那么突破是什么？比如，我确定有很多日常的突破，但是，有没有一个突破真正帮助提高了性能？比如，人们开始相信什么？或者这只是一个渐进的过程？嗯，我认为这是一个渐进的过程，但有一件事我认为让人们相信我们可以做到，是因为我们遵循了不同的想法，构建不同的组件，将它们插入架构，运行系统，看看我们做得如何，进行错误分析，开始新的研究项目来改进事物，以及，以及，以及非常重要的想法，即单个组件不必深入了解其他所有组件正在发生的一切，这就是我们以非常重要的方式利用机器学习的地方，所以虽然单个组件可以是统计驱动的机器学习组件，其中一些是你的，一些是机器学习组件，整个系统使用机器学习组合了所有分数，这是至关重要的，因为这样你就可以分而治之，所以你可以说，好吧，你负责你的候选生成器，或者你负责这个答案评分方法，你负责这个类型评分方法，你负责这个段落搜索方法，或者段落选择方法，等等，但是当我们，你只是把它插进去，我们有足够的训练数据说，现在我们可以，我们可以训练并弄清楚如何权衡所有分数，相对于彼此，基于预测结果，即在问答游戏中是正确还是错误，我们有足够的训练数据来做到这一点，所以这使得人们可以独立工作，并让机器学习来完成集成，美丽，所以，是的，机器学习正在进行融合，然后它是一个人类协调的合奏，没错，朋友的方法，仍然令人印象深刻，他们能够在几年内完成，如果我把自己置于那个心态，那对我来说并不明显，但是当你回顾问答挑战时，当你仰望星空时，你最自豪的是什么？回顾那些日子，我最自豪的是我的承诺，以及我团队的承诺，忠于科学，不怕失败，这是美丽的，因为有如此大的压力，因为这是一个公开的活动，这是一个公开的节目，你致力于这个想法，没错，你认为它在全世界的眼中是成功的吗？它是一个成功，在你，我确信，非常高的标准下，有没有什么遗憾，你会做些不同的事情？它是一个成功，它是我们目标的成功，我们的目标是建立最先进的开放域问题回答系统，我们回到了我们过去试图解决的老问题，我们在所有问题上都做得更好，我们也赢了问答游戏，所以我们赢了问答游戏，所以它是，它是成功的，它不是自然语言理解的成功，但那不是目标，是的，那是，但我想说，我明白你的意思，从科学的角度来看，但我会说，它的灵感，对吧？它不是在解决自然语言理解方面的成功，它是一个激励未来挑战的成功，绝对推动未来的努力，人类在问答游戏中竞争的方式与沃森的方式有什么区别？这在智力方面很重要，是的，所以，这实际上在项目早期就出现了，事实上，我有一些人想加入这个项目，他们很早就，他们在我承诺去做之后就找到了我，他们想过人类是如何做的，他们来自认知角度，比如人类认知，以及它应该如何发挥作用，我不会让他们加入项目，因为另一个假设，或者我设定的另一个目标是，我真的不在乎，至少在这个项目的背景下，在自然语言理解的背景下，以及在构建理解如何与人类沟通的人工智能的背景下，我真的很在乎，是的，所以不是我不在乎，事实上，作为一个人工智能科学家，我非常关心这一点，但我也是一个务实的工程师，我承诺完成这项工作，我不会分心，我必须说，看，如果我要完成这项工作，并且它遵循这条路径，这条路径说，我们将工程化一个机器，它将完成这项工作，而且我们知道搜索和自然语言处理可以做什么，我们必须在此基础上构建，如果我进来并采取不同的方法，开始思考人类的思想可能或不可能这样做，我不会在那个时间里完成，你知道，在那个时间框架内，我认为这是一个领导团队的好方法，但现在它已经完成了，然后，当你回顾时，对吧？所以分析一下区别，对吧？所以，我有点惊讶，事实上，随着时间的推移，当这种情况偶尔出现，我反思它，以及与肯·詹宁斯聊了一点，听到肯·詹宁斯谈论他如何回答问题时，我有点惊讶，因为人类可能比我以前想象的更接近人类回答问题的方式，因为像肯·詹宁斯这样水平的人，在问答游戏中，可能也在作弊，对吧？现在，没有人是肤浅的，他们做得尽可能快，他们进行肤浅的分析，所以他们非常快速地分析问题，并想出一些，你知道，关键的，你知道，关键的向量或线索，如果你愿意，然后他们把这些线索，他们非常快速地浏览，比如他们的资料库，而不是深入思考正在发生的事情，然后，比如，很多不同的，我们称之为这些，这些分数，它们以非常肤浅的方式给分数，然后说，哦，砰，你知道，就是这样，所以，当我们反思这一点时很有趣，所以我们可能正在做一些与人类做的方式不太遥远的事情，但我们当然没有通过说，你知道，你怎么做这个，在元素认知中，比如我现在领导的项目，我们一直在问这些问题，因为最终，我们想做一些事情，让机器中的智能和人类中的智能非常兼容，嗯，兼容，也就是说，它们可以相互沟通，并且可以进行推理，具有共享的理解，所以它们如何思考事物，以及它们如何构建答案，如何构建解释，成为一个非常重要的问题，那么，开放域但冷构建的问题回答，或者问答游戏，与需要理解才能与人类和机器进行共享沟通的东西有什么区别？是的，嗯，这回到了我们之前谈论的解释，无论如何，问答游戏，系统正在试图解释问题，而不是解释内容，而是推理，并就任何特定框架而言，我的意思是，它是解析它，就像解析内容，并使用语法线索之类的东西，所以如果你认为语法在某种意义上是人类框架，以及它，但是当你进入更丰富的语义框架时，人们是如何思考的？是什么激励他们？正在发生的事件是什么？为什么会发生？什么导致了其他事情发生？以及，以及，以及事物在时间和空间中的位置？就像当你开始思考人类如何形成和构建他们头脑中获得的知识时，沃森并没有做任何这些事情，你认为自由交流的本质挑战是什么？自由交流的对话，比问题回答，即使有共享的解释对话？是的，你认为自由交流的对话比问题回答更困难，即使有共享的，所以对话在很多方面都很重要，我的意思是，这是一个挑战，所以首先，当我想到一个理解语言的机器，并最终能够以客观的方式进行推理，能够获取它通过语言或其他方式感知的信息，并将其与这些框架联系起来，进行推理并解释自己，该系统最终需要能够与人类交谈，或者我需要能够与人类互动，所以，在某种意义上，对话，这并不意味着，比如，有时人们谈论对话，他们认为，你知道，人类如何交谈？你如何与他们进行随意的交谈？然后你可以模仿随意的交谈，我们不是在模仿随意的交谈，我们实际上是在产生机器，目标是帮助你思考，帮助你推理你的答案，并解释原因，所以，而不是像和街上的朋友谈论抽烟，和街上的朋友进行闲聊，这更多的是，就像你会与星际迷航中的电脑沟通一样，你想思考什么？你想推理什么？我会告诉你我拥有的信息，我会总结它，我会问你问题，你会回答这些问题，我们会来回交流，我会弄清楚你的心智模型是什么，我会，我会把它和你拥有的信息联系起来，并以你能理解的方式呈现给你，我们可以问后续问题，所以你想构建的是这种对话，它更结构化，更有目标导向，但它需要流畅，换句话说，它不能，它必须引人入胜且流畅，它必须富有成效且不分散注意力，所以必须有一个词的模型，机器必须有一个模型，关于人类如何思考和讨论它们，所以基本上是一场富有成效的丰富对话，不像这个部分，是的，但我喜欢认为它更类似于这个部分，就像开玩笑一样，我也会问你关于幽默的问题，但是，最难的部分是什么？因为这似乎我们作为社区离那很远，所以，能够，所以一个是我们拥有共享的理解，就像我认为你说的关于框架的很多东西都非常棒，但是创造一个流畅的对话，感觉很笨拙，现在，嗯，你刚刚指定的整个问题的哪些方面，比如进行富有成效的对话是最难的，或者也许是任何方面，你可以评论一下，因为它如此神秘，所以，我认为，你必须有创造力，在以下方面，如果我要这样做，纯粹是机器学习方法，有人说，学会进行一次好的、流畅的、结构化的知识获取对话，我会出去说，好吧，我必须收集大量人们这样做的，人们推理的数据。

嗯，进行一次结构良好的对话，既能高效地获取知识，又能在此过程中产生答案和解释，而你却在挣扎，我不知道，选择一天来收集数据，因为我不知道有多少数据是这样的，我想，好吧，好吧，那么这个有一个人类，但即使它在那里，说它在那里，你喜欢，好吧，所以我想，我想，就像一个易于访问的，对吧，所以我想，任何像这样的问题，你没有足够的数据来代表你想要学习的现象，换句话说，如果你有足够的数据，你就可以潜在地学习模式，在这个例子中，很难做到这一点，这是你知道的，苏茜，一种人类该做的事情，你最近出来，IBM是辩论或项目，当然，对吧，因为现在你有，你确实有这些结构化的对话，这些辩论的东西，他们确实使用了机器学习技术来生成，你知道，生成这些辩论，在我看来，比生成一个结构化的论证要难一些，因为你有很多其他的结构化论证，你可以注释这些数据，你可以说这是一个好的回应，一个坏的回应，在一个特定的领域，我必须做出回应，我必须机会主义地对待人类在说什么，所以我的目标是解决问题，我想获取必要的知识，但我也必须机会主义地和回应性地对待人类在说什么，所以我想，我们不能仅仅通过训练大量数据来做到这一点，但我们可以引导它，换句话说，我们可以有创意，我们可以说，我们认为一个好的对话结构是什么，它能很好地做到这一点，我们可以开始创造它，如果我们能，如果我们能以编程方式至少以编程方式创造它，以启动这个过程，我就可以创建一个工具，现在有效地与人类互动，我可以开始生成数据，我可以开始人类学习过程，我可以更新我的机器，但我也可以开始自动学习过程，但我必须理解要学习哪些特征，所以我必须先引导这个过程，这是一个创造性的设计任务，然后我可以将其作为输入到更自动的学习任务中，这是创造力和引导，好吧，你认为对话的哪些元素你想看到，对我来说，一个基准是幽默，这似乎是最难的，对我来说，最大的对比是来自沃森，有史以来最伟大的喜剧素描之一，是SNL名人 Jeopardy，与亚历克斯·特雷贝克和肖恩·康纳利和伯特·雷诺兹等等，与肖恩·康纳利评论亚历克斯·特雷贝克的微笑，很多，我想他们都在负分，为什么，所以他们显然都在Jeopardy比赛中输了，但在喜剧方面他们都赢了，所以你对幽默在这个整个互动中，在生产性的对话中，或者即使是任何东西，人类对我来说代表的是你说的关于框架的同一个想法吗，因为幽默只存在于特定的人类框架中，所以你对幽默有什么看法，你对像幽默这样的事物有什么看法，它与你提到的那种创造力有关，我认为那里有几件事在发生，所以我，我，我有点觉得，我可能太乐观了，但我想，我们在这方面做了一些事情，关于双关语，在Jeopardy中，我们真的坐下来说，嗯，你知道双关语是如何运作的，你知道，就像文字游戏，你可以形式化这些东西，所以我认为幽默有很多方面你可以形式化，你也可以学习新的幽默，你也可以说人们笑什么，如果你有足够的数据来代表这种现象，你知道，你可能能够，你知道，权衡特征，找出，你知道，人类觉得什么有趣，什么不有趣，你可能，机器可能无法解释为什么，我的朋友，除非我们，除非我们坐下来更正式地思考，我认为再次，我认为你做两者的结合，我一直是一个坚定的支持者，我认为，你知道，健壮的架构和方法总是有点结合，我们反思和创造性地思考事物是如何构建的，以及如何形式化它们，然后利用大量数据进行学习，并找出如何结合这两种方法，我认为还有人与人之间的另一个方面，这涉及到我感觉我能与讲故事的人联系起来，讲故事的人，我认为这是人工智能主题中的一个有趣主题，那就是，当我得知它是一个机器人时，我感觉不同吗，当我得知我设想有一个机器人，它不像我一样有意识，当我设想机器人没有像我一样的经验时，我发现它，你知道，有趣吗，还是因为它不那么相关，我设想那个人没有像我一样把它联系起来，我认为这在艺术和娱乐中也能看到，比如，你知道，有时你会遇到一些在某件事上非常出色的人，无论是雕塑，音乐还是其他什么，但最受关注的人是那些能够唤起类似情感反应的人，能够让你产生共鸣的人，关于他们如何，换句话说，他们基本上能够将艺术品，音乐，绘画，雕塑与情感联系起来，让你与他们分享这种情感，然后，那时它就变得引人入胜了，他们正在以完全不同的水平进行交流，他们不仅仅是在交流艺术品，他们是在交流他们对艺术品的情感反应，然后你就会觉得，哇，我能理解那个人，我能与那个人联系起来，我能与那个人联系起来，所以我想幽默也有那个方面，所以你可以与那个人联系起来的想法，人是关键，但我们也能够很好地拟人化物体， pretty robots and AI systems pretty well, so we're almost looking to make them human, there may be from your experience with Watson, maybe you can comment on, did you consider that as part, well obviously the problem of Jeopardy doesn't require int the promotoras ation, but nevertheless, well there was some interest in doing that, and I've that's an that's another thing I didn't want to do, so I didn't want to distract from the from the actual scientific test nights, so you're absolutely right, I mean humans do anthropomorphize, and and without necessarily a lot of work, I mean just put some eyes in a couple of eyebrow movements, and you're getting humans to react emotionally, and I and I think you can do that, so I didn't mean to suggest that that that connection can't cannot be mimicked, I think that connection can be mimicked, and can get you to can produce that emotional response, I just wonder though, if you're told what's really going on, if you know that the machine is not conscious, not having the same richness of emotional reactions and understanding, that doesn't really share the understanding, but is essentially just moving inside brow or drooping its eyes or making them big or whatever it's doing that's getting the emotional response, will you still feel it interesting, I think you probably would for a while, and then when it becomes more important that there's a deeper under depreciate understanding, it may run flat, but I don't know, I'm pretty I'm pretty confident that it will, the majority of the world, even if you tell them how, no matter well, it will not matter, especially if the Machine herself says that she is cautious, that's very possible, so you the scientists that made the machine is saying that this is how the algorithm works, everybody will just assume you're lying, and that there's a conscious being there, so you're deep into the science fiction shop, you're on right now, but yeah, I think it's actually psychology, I think it's not science fiction, I think it's reality, I think it's a really powerful one that will have to be exploring in the next few decades, it's a very interesting element of intelligence, so what do you think, we've talked about social constructs of intelligences and and frameworks and the way humans kind of interpret information, what do you think is a good test of intelligence in your view, so there's the Alan Turing with the Turing test, Watson accomplished something very impressive with Jeopardy, what do you think is a test that would impress the heck out of you that you saw that a computer could do, they say this is crossing a kind of threshold that's that gives me pause in a good way, expectations for a are generally high, what does high look like by the way, so not the threshold test as a threshold, what do you think is the destination, what do you think is the ceiling, I think machines will in many measures will be better than us, will become more effective, in other words, better predictors about a lot of things, and then then then ultimately we can do, I think where they're gonna struggle is what we talked about before, which is relating to communicating with and understanding humans in deeper ways, and and so I think that's a key point, like we can create the super parrot, what I mean by the super parrot is, given enough data, a machine can mimic your emotional response, can even generate language that will sound smart, and what someone else might say under similar circumstances, look how its paws on that, like that's a super parrot, right, so given similar circumstances, moves its face, its faces in similar ways, changes its tone of voice in similar ways, produce the strings of language that you know would similar that a human might say, not necessarily being able to produce a logical interpretation or understanding that would ultimately satisfy a critical interrogation or a critical understanding, I think you guys describe me in a nutshell, I think I think philosophically speaking, you could argue that that's all we're doing as human beings to war, so I was gonna say, it's very possible, you know, humans do behave that way too, and so upon deeper probing and deeper interrogation, you may find out that there isn't a shared understanding, because I think humans do both, like humans are statistical language model machines, and and and they are capable reasoner's, you know, they're they're both, and you don't know which is going on, right, so and I think it's, I think it's an interesting problem, we talked earlier about like, where we are in our social and political landscape, can you distinguish some who can string words together and sound like they know what they're talking about from someone who actually does, can you do that without dialogue, without integrity of a programming dialogue, so it's interesting, because humans are really good at in their own mind justifying or explaining what they hear, because they project their understanding on onto yours, so you could say, you could put together a string of words, and and someone will sit there and interpret in a way that's extremely biased, this is the way they want to interpret it, they want to assuming you're an idiot, and they'll true put it one way, they've all seen you're a genius, and interpreted another way that suits their needs, so this is tricky business, so I think the answer your question, as AI gets better and better at better and better mimic you, we create the super parrots, we're challenged, just as we are with we're challenged with humans, do you really know what you're talking about, do you have a meaningful interpretation, a powerful framework that you could reason over and justify your answers, justify your predictions and your beliefs, why you think they make sense, can you convince me what the implications are, you know, can you, so can you reason intelligently and make me believe that those um, the implications of your prediction and so forth, so what happens is it becomes reflective, my standard for judging your intelligence depends a lot on mine, but you're saying that there should be a large group of people with a certain standard of intelligence that would be convinced by this particular AI system, then there should be by, I think one of the depending on the content, one of the problems we have there is that if that large community of people are not judgment judging it with regard to a rigorous standard of objective logic and reason, you still have a problem, like masses of people can be persuaded, the Millennials, yeah, to turn them turn their brains off, right, okay, sorry, I have nothing against the warning, I just, so you, you're a part of one of the great benchmarks challenges of AI history, what do you think about alpha zero, open AI five, alpha star accomplishments on video games recently, which are also, I think, at least in the case of go, without fagala, now for zero playing go, was a monumental accomplishment as well, what are your thoughts about that challenge, I think it was a giant lamare, I I think it was phenomenal, I mean, as one of those other things, nobody thought like solving go was gonna be easy, particularly because it's again, it's hard for particularly hard for humans, our team is to learn how for humans to excel at, and so it was up another measure, a measure of intelligence, it's very cool, I mean, it's very interesting, you know, what they did, I mean, and I loved how they solved like the data problem, which again, they bootstrapped it, and got the machine to play itself to generate enough data to learn from, I think that was brilliant, I think that was great, and and and of course, the result speaks for itself, I think it makes us think about again, it is okay, what's intelligence, what aspects of intelligence are important, can the can the go machine help me make me a better go player, is it an alien intelligence, it was, is am I even capable of like, again, if we if we put in very simple terms, it found the function, we found the go function, can I even comprehend the go function, can I talk about the go function, can i conceptualize the go function, like whatever it might be, so one of the interesting ideas of that system is it plays against itself, right, yeah, but there's no human in the loop there, so like you're saying, it could have by itself created an alien intelligence, how torta torta gorrik, imagine you're sentencing, you're judging, you're sentencing people, or you're setting policy, or you're, you know, you're making medical decisions, and you can't explain, you can't get anybody to understand what you're doing or why, so it's, it's, it's an interesting dilemma for the applications of AI, do we hold AI to this accountability that says, you know, humans have to be, humans have to be able to take responsibility, you know, for for the decision, in other words, can you explain why you would do the thing, well, you will use get up and speak to other humans and convince them that this was a smart decision, is the AI enabling you to do that, can you get behind the logic that was made there, do you think, sorry to link on this point, because it's a, there's a fascinating one, that's a great goal for AI, do you think it's achievable in many cases, or do you, okay, there's two possible worlds that we have in the future, one is where AI systems do like medical diagnosis, or things like that, would drive a car without ever explaining to you why it fails when it does, that's one possible world, then we're okay with it, or the other, where we are not okay with it, and we really hold back the technology from getting to good before it gets able to explain, which of those worlds are more likely, do you think, and which are concerning to you, or not, I think the reality is, it's gonna be a mix, you know, I'm not trying a problem with that, I mean, I think there are tasks that perfectly fine with machines show a certain level of performance, and that level of performance is already better, it is already better than humans, so for example, I don't know that I get tape driverless cars, if driverless cars learn how to be more effective drivers than humans, but can't explain what they're doing, but bottom line, statistically speaking, there, you know, ten times safer than humans, I I don't know that I care, I think when we, we have these edge cases, when something bad happens, and we want to decide who's liable for that thing, and who made that mistake, and what we do about that, and I think in those, those educators are interesting cases, and now do we go to designers of the AI, and the I says, I don't know, if that's what it learned to do, and it says, well, you didn't train it properly, you know, you, you were you were negligent in the training data that you gave that machine, like how do we drive down and realize, oh, so I think those are, I think those are interesting questions, so the optimization problem there, sorry, is to create a system that's able to explain the lawyers away, there you go, um, I think that, uh, uh, I think it's gonna be interesting, I mean, I think this is where technology and social discourse are gonna get like deeply intertwined, and how we start thinking about problems, decisions, and problems like that, I think in other cases, it becomes more obvious, where, you know, it's, I got like, why did you decide to give that person, you know, a longer sentence, or or to deny them parole, again, policy decisions, or why did you pick that treatment, like that treatment up killing that guy, like why was that a reasonable choice to make, so so and people are gonna demand explanations, now, there's a reality though here, and the reality is that it's not, I'm not sure humans are making reasonable choices when they do these things, they are using statistical hunches, biases, or even systematically using statistical averages to make Osmonds, is what happened my dad, if you saw that target gave about that, but you know, I mean, they decided that my father was brain dead, he had went into cardiac arrest, and it took a long time for the ambulance to get there, and wasn't not resuscitated right away, and so forth, and they came, they told me he was brain dead, and why was he brain dead, because essentially they gave me a purely statistical argument, under these conditions, with these four features, 98% chance he's brain dead, innocent, but can you just tell me, not inductively, but deductively, go there and tell me his brain stopped functioning, is the way for you to do that, and they, and their, the protocol in response was, no, this is how we make this decision, I said, this is adequate for me, I understand the statistics, and I don't have, you know, there's a two percent chance he's, so like, I just don't know the specifics, I need the specifics of this case, and I want the deductive logical argument about why you actually know he's brained it, so I wouldn't sign that do not resuscitate, and I don't know, it was like, they went through lots of procedures, as a big long story, but the bottom was a fascinating story, by the way, but how I reasoned, and how the doctors reasoned through this whole process, but I don't know, somewhere around 24 hours later, or something, he was sitting up, that would zero bushido brain damage, any, what lessons do you draw from that story, that experience, that the data that they're you, that the data that's being used to make sophistical inferences, doesn't adequately reflect the phenomenon, so in other words, you're getting shit Ramsar, you're getting stuff wrong, because you're your model is not robust enough, and you might be better off not using statistical inference and statistical averages in certain cases, when you know the models insufficient, and that you should be reasoning at about the specific case more logically and more deductively, and hold yourself responsible, to hold yourself accountable to doing that, and perhaps AI has a role to say, the exact thing we just said, which is perhaps this is a case, you should think for yourself, you should reason deductively, so it's hard, it's, it's so, it's hard because it's hard to know that, you know, you'd have to go back, and you'd have to have enough data to essentially say, and this goes back to how do we, this goes back to the case of how do we decide whether the AI is good enough to do a particular task, and regardless of whether or not it produces an explanation, so, um, and and what standards do we hold, right, for that, so, um, you know, if you look at, you, you look more broadly, for example, as my father, as a metal kick, medical case, the medical system ultimately helped him a lot throughout his life, without it, he probably would have died much sooner, so overall, sort of, you know, work for him, and sort of a net, in that kind of way, actually, I don't know that's fair, um, but it maybe not in that particular case, but overall, like, oh, the medical system overall, that's more, given a system overall, you know, was doing more more good than bad, now, is another argument that suggests that wasn't the case, but for the, for the sake of argument, let's say, like, that's, let's say a net positive, and I think you have to sit there, and there, and take, take that into consideration, now, you look at a particular use case, like, for example, making this, this decision, have you done enough studies to know how good that prediction really is, right, and how you have you done enough studies to compare it to say, well, what if we, what if we dug in, and in a more direct, you know, let's get the evidence, let's let's do the deductive thing, and not use the statistics here, how often would that have done better, right, you just, so you have to do this studies to know how good the AI actually is, and it's complicated, because depends how fast you have to make decision, so if you have to make the decision superfast, do you have no choice, right, if you have more time, right, but if you're ready to pull the plug, and this is a lot of the argument that I had was a doctor, I said, what's he gonna do, if you do it, what's gonna happen to him in that room, if you do it my way, you know, if you do, well, he's gonna die anyway, so let's do it my way, though, I mean, it raises questions for our society to struggle with, as was the case with your father, but also when things like race and gender start coming into play, when when judgments are made made based on things that are complicated in our society, at least in this course, and it starts, you know, I think I think I'm safe to say that most of the violent crimes committed by males, so if you discriminate based, you know, as a male versus female, saying that if it's a male, more likely to commit the crime, so this is one of my, my very positive and optimistic view views of why the study of artificial intelligence, the process of thinking and reasoning logically and statistically, and how to combine them, is so important for the discourse today, because it's causing a regardless of what what state AI device devices are or not, it's causing this dialogue to happen, this is one of the most important dialogues that, in my view, the human species can have right now, which is how to think well, yeah, how to reason well, how to understand our own cognitive biases, and what to do about them, that has got to be one of the most important things we as as as a species can be doing, honestly, we are reached, we've created an incredibly complex society, we've created amazing abilities to amplify noise faster than we can play amplifies signal, we are challenged, we are deeply deeply challenged, we have, you know, big segments of the population getting hit with enormous amounts of information, do they know how to do critical thinking, do they know how to objectively objectively reason, do they understand what they are doing, nevermind with their AI is doing, this is such an important dialogue, you know, to be having, and and and, you know, we are fundamentally are thinking can be, and easily becomes fundamentally bias, and there are statistics, and we shouldn't blind our, so we shouldn't discard statistical inference, but we should understand the nature of such this conference, as us as a society, as you know, we decided to reject statistical inference to favor individual understanding, and and deciding on the individual, yes, we we consciously make that choice, so even if the statistics said, even if the Cystic said males are more likely to have, you know, to be violent criminals, we still take each person as an individual, and we treat them based on the logic and the knowledge of that situation, we purposefully and intentionally reject the statistical once we do that, at a respect for the individual, for the individual, yeah, and then that requires reasoning and cracking, looking forward, what Grand Challenges would you like to see in the future, because the the Jeopardy challenge, you know, captivated the world, alpha go, alpha zero, cap day of the world, deep blue, certainly beating Kasparov, Gary's bitterness aside, and captivated the world, what do you think, do you have ideas for next grand challenges, for future challenges of that, oh, you know, I look, I mean, I think there are lots of really great ideas for Grand Challenges, I'm particularly focused on one right now, which is Kent, you know, can you demonstrate that they understand, that they could read and understand, that they can, they can acquire these frameworks, and communicate, you know, reason and communicate with humans, so it is kind of like the Turing test, but it's a little bit more demanding than the Turing test, it's not enough, it's not enough to convince me that you might be human, because you could, you know, you can parrot a conversation, I think, you know, the the this standard is a little bit higher, is for example, can you, you know, the santa is higher, and I think one of the challenges of devising this grand challenge is that we're not sure what intelligence is, we're not sure how to determine whether or not two people actually understand each other, and then what depth they understand it, they, you know, and what to what depth they understand each other, so the challenge becomes something along the lines of, can you satisfy me that we have a shared understanding, so if I were to probe and probe, and you probe me, can, can, can machines really act like thought partners, where they could satisfy me that they, that we have a share our understanding, is shared enough that we can collaborate and produce the answers together, and that, you know, they, they can help me explain and justify those answers, so maybe here's an idea, so we'll have a Isis run for president, and convinced, that's too easy, from sorry, oh, no, you have to convince the voters that they should vote for it, so they s, what I would again, again, I, that's why I think this is such a challenge, because we go back to the emotional persuasion, we go back to, you know, now we're checking off an aspect of human cognition, that is in many ways weak or flawed, right, we're so easily manipulated, our minds are drawn for often the wrong reasons, right, not the reasons that ultimately matter to us, but the reasons that can easily persuade us, I think we can be persuaded to believe one thing or another, for reasons that ultimately don't serve us well in the long term, and a good benchmark should not play with those elements of emotional manipulation, I don't think so, I think that's where we have to set the set the higher standard for ourselves, of what, you know, what does it mean, this goes back to rationality, and it goes back to objective thinking, and can you produce, can you acquire information, and produce reasoned arguments, and to those reasons arguments pass a certain amount of muster, and is it, and can you acquire new knowledge, you know, can you, can you under, can you reason, oh, I have acquired new knowledge, can you identify where it's consistent or contradictory with other things you've learned, and can you explain that to me, and get me to understand that, so I think another way to think about it, perhaps, is kind of machine teach, you can the hell really less than that's where to put it, can you understand something that you didn't really understand before, where's where is, you know, it's taking it, so you're not, you know, again, it's almost like, can it, can it teach you, can it help you learn, and and in an arbitrary space, so it can open those domain space, so can you tell the Machine, and again, this borrows from some science fiction's abut, can you go off and learn about this topic that I'd like to understand better, and then work with me to help me understand it, that's quite brilliant, what the machine that passes that kind of test, do you think it would need to have self-awareness, or even consciousness, what do you think about consciousness, and the importance of it, maybe in relation to having a body, having a presence, an entity, do you think that's important, you know, people used to ask if Watson was conscious, and I used to think, and he said, he's the conscious of what exactly, I mean, I think, you know, main cell, it depends what it is that you're conscious, I mean, like, so, you know, did it, if you, you know, it's certainly easy for it to answer questions about, it would be trivial to program it, so the answer questions about whether or not it was playing Jeopardy, I mean, it could certainly answer questions that will imply that it was aware of things, exactly what does it mean to be aware, and what does it mean to conscious, and it's sort of interesting, I mean, I think that we differ from one another, based on what we're conscious of, but wait, wait, for sure, there's degrees of consciousness in there, so it, well, in those areas, like, it's not just agrees, what do you, what do you, what are you aware of, like, what are you not aware, but nevertheless, there's a very subjective element to our experience, let me even not talk about consciousness, let me talk about another, to me, really interesting topic, immortality, fear, or mortality, Watson, as far as I could tell, did not have a fear of death, certainly not, most most humans do, wasn't conscious of death, wasn't that, so there's an element of finiteness to our existence, that I think, like, we, like I mentioned survival, that adds to the whole thing, that I mean, consciousness is tied up with that, that we are us thing, it's a subjective thing, that ends, and that seems to add a color and flavor to our motivations, in a way that seems to be fundamentally important for intelligence, or at least the kind of human intelligence, well, I take for generating goals, again, I think you could have, you could have an intelligence capability, and a capability to learn, I capability to predict, but I think without, I mean, again, you get a fear, but essentially without the goal to survive, so you think you can just encode that, without having to million code, I mean, can you create a robot now, and you could say, you know, and plug it in, and say, protect your power source, you know, and give it some capabilities, and we'll sit there and operate to try to protect this power source, and survive, I mean, I, so I don't know that that's awfully a hard thing to demonstrate, it sounds like a fairly easy thing to demonstrate, that you can give it that goal, we'll come up with that goal by itself, as you have to program that goal in, but there's something, because I think, as we touched on, intelligence is kind of like a social construct, the the fact that a robot will be protecting its power source, would would add depth and grounding to its intelligence, in terms of us being able to respect, I mean, ultimately, it boils down to us acknowledging that it's intelligent, and the fact that it can die, I think, is an important part of that, the interesting thing to reflect on is how trivial that would be, and and I don't think if you knew how trivial that was, you would associate that with being intelligence, I mean, I literally put in a statement of code, that says, you know, you have the following actions, you can take, you give it a bunch of actions, like you mount a laser gun on her, or you may do you the ability to scream a screech, or whatever, and you know, and you, you say, you know, if you see your power source, then you could program that in, and you know, you're gonna print it, you're gonna take these actions to protect it, you know, you teach it checking it on a bunch of things, so, and and now you're gonna look at that, and you say, well, you know, that's intelligence, because it's protecting power source, maybe, but that's again, at this human bias, that says, the thing I had, then I identify my intelligence, and my conscious, so fundamentally, with the desire, or at least the behaviors associated with the desire to survive, that if I see another thing doing that, I'm going to assume it's intelligent, what timeline year will society have a something that would that you would be comfortable calling an artificial general intelligence system, well, what's your intuition, nobody can predict the future, certainly not next few months, or twenty years away, but what's your intuition, how far away are we, I the ideas hearts make these predictions, and I would be, you know, I would be guessing, and there's so many different variables, including just how much we want to invest in it, and how important it, you know, and how important we think it is, what kind of investment are willing to make in it, what kind of talent we end up bringing to the table, all, you know, the incentive structure, all these things, so I think it is possible to do this sort of thing, I think it's, I think trying to sort of ignore many of the variables, and things like that, is it a ten-year thing, as a 23, it's probably closer to a 20-year thing, I guess, but not as little, no, I don't think it's several hundred years, I don't think it's several hundred years, but again, so much depends on how committed we are to investing, and incentivizing this type of work, this type of work, and it's sort of interesting, like, I don't think it's obvious how incentivize we are, I think from a task perspective, you know, if we see business opportunities to take this technique, is a technique to solve that problem, I think that's the main driver for many from any of these things, from a from a general Tosta, seems kind of an interesting question, are we really motivated to do that, and like, we just struggled ourselves right now, to even define what it is, so it's hard to incentivize, when we don't even know what it is, we're incentivized to create, and if you said mimic a human intelligence, I just think there are so many challenges with the the significance and meaning of that, there's not a clear directive, there's no clear directive to do precisely that thing, so assistance in a larger and larger number of tasks, so being able to a system that's particularly able to operate my microwave, and making a grilled cheese sandwich, I don't even know how to make one of those, and then the same system would be doing the vacuum cleaning, and then the same system would be teaching my kids, that I don't have math, I think that when when you get into a general intelligence for learning physical tasks, and again, yeah, I want to go back to your body questions, it's on your body question was interesting, but you want to go back to, you know, learning abilities, do physical tasks, you might have, we might get Majan in that timeframe, we will get better and better at learning these kinds of tasks, whether it's mowing your lawn, or driving a car, or whatever it is, I think we will get better and better at that, where it's learning how to make predictions over large bodies of data, as if we're going to continue to get better and better at that, and machines will out, you know, outpace humans, and and a variety of those things, the underlying mechanisms for doing that, may be the same, meaning that, you know, maybe these are deep Nats, there's infrastructure to train them, reusable components to get them to different classes of tasks, and we get better and better at building these kinds of machines, you could see argue that the general learning infrastructure in there, is a form of a general type of intelligence, I think what starts getting harder, is this notion of, you know, can we, can we effectively communicate and understand, and build that shared understanding, because of the layers of interpretation that are required to do that, and the need for the machine to be engaged with humans at that level, at a continuous basis, so how do you get in, how do you get the machine in the game, how do you get the machine in the intellectual game, yeah, and to solve AGI, you probably have to solve that problem, you have to get the machine, so it's a little bit of a bootstrapping, can we get the machine engaged, and, you know, in the intellectual calling a game, but in the intellectual dialogue with the humans, are the humans sufficiently an intellectual dialogue with each other, to generate enough, to generate enough data in this context, and how do you bootstrap that, because every one of those conversations, every one of those conversations, those intelligent interactions require so much prior knowledge, that is a challenge to bootstrap it, so that's, so as, so the question is, and how committed, so I think that's possible, but when I go back to, are we incentivized to do that, I know we're incentivized to do the former, are we incentivize to do the latter, significantly enough, to people understand what the latter really is, well, part of the elemental cognition mission is to try to articulate that better and better, you know, through demonstrations, and to trying to craft these grand challenges, and get people to say, look, this is a class of intelligence, this is a class of AI, do we, do we want this, what, what is the potential of this, what are the business, what's the business potential, what's the societal potential, to that, and so, you know, and to build up that incentive system around that, yeah, I think if people don't understand yet, I think they will, and is a huge business potential here, so it's exciting that you're working on it, you've kind of skipped over, but I'm a huge fan of physical presence of things, do you think, you know, Watson had of body, do you think having a body as to the interactive element between the AI system and a human, or just in general to intelligence, so I think, I think going back to that shared understanding bit, humans are very connected to their bodies, I mean, is one of the reasons, one of the challenges in getting an AI to kind of be a compatible human intelligence, is that our physical bodies are generating a lot of features that make up the input, so in other words, where our bodies are, are the the tool we use to affect output, but they're also, but they also generate a lot of input for our brains, so we generate emotion, we generate all these feelings, we generate all these signals that machines don't have, so missions that have this as the input data, and they don't have the the feedback that says, okay, I've gotten this, I've gotten this emotion, or I've gotten this idea, I now want to process that, and then it affects me as a physical being, and then I, and I, and I can play that out, in other words, I could realize the implications of tax implications, again, on my bond, mind body complex, I then process that, and the implications, again, are internal features, are generated, I learn from them, they have an effect on my mind body complex, so it's interesting, when we think, do we want a human intelligence, well, if we want a human compatible intelligence, probably the best thing to do is to embed it, embedded in a human body, just to clarify, and both concepts beautiful, is humanoid robots, so robots that look like humans, is one, or did you mean actually sort of what Hamas was working with, neural link, really embedding intelligence systems that the ride-alongs human bodies, no, I was riding along is different, I meant like, if you want to create an intelligence that is human compatible, meaning that it can learn and develop a shared understanding of the world around it, you have to give it a lot of the same substrate, part of that substrate, you know, is the idea that it generates these kinds of internal features, like sort of emotional stuff, it has similar senses, it has to do a lot of the same things with those same sentences, um, right, so I think if you want that, again, I don't know that you want that, like man, like that's not my specific goal, I think that's a fascinating scientific goal, I think it has all kinds of other implications, that's sort of not to go, like I want it, I want to create, I think of it as, I create intellectual thought martyrs for humans, so that kind of that kind of intelligence, I know other companies that are creating physical thought partners, the fiscal partners to figure out for you, but that's kind of not where we're, you know, I'm at, but but but the the important point is that a big part of how of what we process, is that physical experience of the world around us, on the point of thought partners, what role does an emotional connection, or forgive me, love, have to play in that thought partnership, is that something you're interested in, put another way, sort of having a deep connection beyond intellectual with the AI, yeah, with the a, between human and ass, is that something that gets in the way of the the rational discourse, is there something that's useful, I worry about biases, you know, obviously, so in other words, if you develop an emotional relationship with the machines, do all of a sudden, you start, are more likely to believe what it's saying, even if it doesn't make any sense, so I, you know, I worry about that, but at the same time, I think the opportunity to use machines to provide human companionship, is actually not crazy, and it's again, the intellectual and social companionship is not crazy, the idea, do you have concerns, as a few people do, you know, Musk, Sam Harris, about long-term existential threats of AI, and perhaps short-term threats of AI, we talked about bias, we talked about different misuses, but do you have concerns about thought partners, systems that are able to help us make decisions together, humans somehow having a significant negative impact on society in the long term, I think there aren't things to worry about, I think the giving machines too much leverage is a problem, and what I mean by leverage is, is too much control for things that can hurt us, whether it's socially, psychologically, intellectually, or physically, and if you give them machines too much control, I think that's a concern, you forget about the AI, just when you give them too much control, human bad actors can hack them, and produce havoc, so, um, you know, that's a problem, and you imagine hackers taking over the driverless car network, and you know, creating all kinds of havoc, but you could also imagine, given given the ease at which humans could be persuaded one way or the other, and now we have algorithms that can easily take control over over that, over that, and amplify noise, and move people one direction or another, I mean, humans do that to other humans all the time, and we have marketing campaigns, we have political campaigns, that take it to image of our our emotions, or our fears, and this is done all the time, when but with machines, machines are like giant mecha phones, right, we can amplify this, and orders of magnitude, and can fine-tune its control, so we can tailor the message, we can now very rapidly and efficiently tailor the message to the audience, taking taking advantage of, you know, of their biases, and amplifying them, and using them to pursue a them in one direction or another, in ways that are not fair, not logical, not objective, not meaningful, and humans, the machines and power that, so, so that's what I mean by leverage, like, it's not new, but wow, it's powerful, because machines can do it more effectively, more, more, you know, more quickly, and we see that already going on, and and and social media, not the plays, and other places, that's scary, and and that's why, like, I'm, that's why I go back to saying, one of the most important public dialogues we could be having, is about the nature of intelligence, and the nature of inference, and logic, and reason, and rationality, and us understanding our own biases, us understanding our own cognitive biases, and how they work, and then how machines work, and how do we use them to complement, and sit, basically, so that in the end, we have a stronger overall system, that's just incredibly important, I don't, most people understand that, so, so like, telling telling your kids, or telling your students, this goes back to the cognition, here's how your brain works, here's how easy it is to trick your brain, right, there are fundamental cognitive, but you should appreciate the different, the different types of thinking, and how they work, and what you're prone to, and, you know, and what, and what do you prefer, and under what conditions does this make sense, versus that makes sense, and then say, here's what AI can do, here's how it can make this worse, and here's how it can make this better, and then that's where the as a role is to reveal that, then the that trade-off, so if you imagine a system that is able to, beyond any definition of the Turing test of the benchmark, really an AGI system, as a thought partner, that you one day will create, what question, what topic of discussion, if you get to pick one, would you have with that system, what would you ask, and you get to find out the truth together, so you threw me a little bit with finding the truth at the end, but this is a whole nother topic, but the, I think the beauty of it, I think what excites me, is the beauty of it, is if I really have that system, I don't have to pick, so in other words, I can, you know, I can go to, and say, this is where I care about today, and and and that's what we mean by, like, this general capability, go out, read this stuff, in the next three milliseconds, and I want to talk to you about it, I want to draw analogies, I want to understand how this affects this decision, or that decision, what if this were true, what if that were true, what what knowledge should I be aware of, that could impact my decision, here's what I'm thinking, is the main implication, can you find, can you prove that out, can you give me the evidence that supports that, can you give me evidence supports this, oh, there's a boy, that would that be incredible, you would that be just incredible, just a long discourse, just to be part of, whether it's a medical diagnosis, or whether it's, you know, the various treatment options, or whether it's a legal case, or whether it's a social problem that people are discussing, like, be part of the dialogue, one that holds itself and us accountable to reasons, an objective dialogue, you know, I just, I get goosebumps talking about it, right, so when when you create it, please come back on the podcast, well, the discussion together, and make it even longer, this is a record for the longest conversation, now, there's an honor, it was a pleasure, David, thank you so much for, thanks so much, a lot of fun, you

Get Our Mobile App

David Ferrucci: IBM Watson, Jeopardy & Deep Conversations with AI | Lex Fridman Podcast #44

Transcription