【双语音+文稿】（完整版）谷歌 DeepMind 的联合创始
共 1 集
剧集目录
| 集 | 标题 | 时长 |
|---|---|---|
| P1 | 【双语音+文稿】（完整版）谷歌 DeepMind 的联合创始 | 52:15 |
P1 · 【双语音+文稿】（完整版）谷歌 DeepMind 的联合创始人兼CEO德米斯·哈萨比斯做客坦福商学院 p01 英文版 (P1)
字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:22 | Demis, it's great to have you here at Stanford. | 戴米斯,很高兴你能来斯坦福 |
| 0:24 | It's fantastic to be here. | 来这里真是太棒了 |
| 0:25 | Thanks everyone for coming. | 谢谢你的光临 |
| 0:26 | Really appreciate you doing this. | 非常感谢你这样做。 |
| 0:29 | So we're gonna, I'm gonna ask you some questions. | ┮и璶拜ㄇ拜肈 |
| 0:31 | We'll have some questions from students. | 我们来问学生一些问题 |
| 0:35 | And we're looking forward to hearing your thoughts. | 我们期待着听到你的想法。 |
| 0:39 | Maybe we would, you've been chronicled a lot recently, a movie, a book. | 也许我们会,你最近被编了很多, 一部电影,一本书。 |
| 0:45 | So some, many people have heard about your trajectory. | 很多人都听说过你的轨迹 |
| 0:50 | It is quite remarkable. | 这是相当显著的。 |
展开字幕全文（838 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | Demis, it's great to have you here at Stanford. | 戴米斯,很高兴你能来斯坦福 |
| 2 | It's fantastic to be here. | 来这里真是太棒了 |
| 3 | Thanks everyone for coming. | 谢谢你的光临 |
| 4 | Really appreciate you doing this. | 非常感谢你这样做。 |
| 5 | So we're gonna, I'm gonna ask you some questions. | ┮и璶拜ㄇ拜肈 |
| 6 | We'll have some questions from students. | 我们来问学生一些问题 |
| 7 | And we're looking forward to hearing your thoughts. | 我们期待着听到你的想法。 |
| 8 | Maybe we would, you've been chronicled a lot recently, a movie, a book. | 也许我们会,你最近被编了很多, 一部电影,一本书。 |
| 9 | So some, many people have heard about your trajectory. | 很多人都听说过你的轨迹 |
| 10 | It is quite remarkable. | 这是相当显著的。 |
| 11 | Chess prodigy, video game developer, scientist, | 围棋天才,电子游戏开发者,科学家, |
| 12 | tech entrepreneur and leader, Nobel laureate. | 技术企业家和领袖,诺贝尔奖获得者. |
| 13 | That's just the first half of your career. | 这只是你职业生涯的前半部分 |
| 14 | So if you were gonna try to draw a through line | 所以如果你想画一条线 |
| 15 | through all those different things that you've done, what would it be? | 通过你所做的所有这些不同的事情, 它会是什么? |
| 16 | Well, I think there's several through lines actually with what seems maybe | 嗯,我认为有几条线 实际上与看起来可能 |
| 17 | somewhat unconnected subjects. | 有点无关紧要的话题 |
| 18 | First of all, I've always really enjoyed working at the intersection of creativity | 首先,我一直很喜欢在创造力的交汇点工作 |
| 19 | and technology, and very broadly construed. | 技术,以及广义的解释。 |
| 20 | So actually the games industry, the video games industry, | 所以实际上游戏行业,电子游戏行业, |
| 21 | which is my first very early part of my career in the 90s, | 这是我90年代职业生涯的第一个很早期的部分, |
| 22 | was one of the most creative spaces in any industry. | 是任何行业中最有创意的空间之一。 |
| 23 | That was using cutting edge technology with art and design | 这是利用尖端技术 艺术和设计 |
| 24 | to sort of create an entirely new entertainment medium. | 来创造全新的娱乐媒体 |
| 25 | So that was really an amazing time. | 这真是一个令人惊叹的时刻。 |
| 26 | In fact, some of the most fun times I've had in my career was early in the 90s. | 事实上,我职业生涯中一些最快乐的时光是在90年代早期. |
| 27 | The chess and the neuroscience I did, all of those things I've tried to... | 棋子和神经科学,我所做的, 所有这些我尝试... |
| 28 | I had this idea of working on AI and AGI being the most important thing | 我有一个想法是 AI和AGI是最重要的事情 |
| 29 | and most interesting one thing one could spend your career working on. | 有趣的是,有件事你可以花你的事业工作。 |
| 30 | Also from a very early age. | 也从很小的时候开始 |
| 31 | So as a teenager, probably I read too much science fiction, | 所以作为一个青少年, 也许我读太多科幻小说, |
| 32 | 阅读诸如Gödel Escher Bach, 这些类型的书籍, | |
| 33 | and biographies of some of my scientific heroes, Turing and Feynman and so on. | 以及我的一些科学英雄图灵和费曼等的传记. |
| 34 | So all of these were serving to inspire me to try and understand the world around us | 所以这些都是为了激励我 试着了解我们周围的世界 |
| 35 | in a really deep way. | 以非常深刻的方式。 |
| 36 | And then building AI was my expression of that mission | 然后,建筑AI是我对任务的表现 |
| 37 | to try and build the ultimate tool for science. | 尝试建立科学的最终工具。 |
| 38 | And I've tried to, because life's short, | 我尝试过,因为生命很短 |
| 39 | I've tried to reuse and repurpose every experience I've had | 我已经尝试了再利用和重新利用 我所有的体验 |
| 40 | in service of that bigger North Star mission that I've had for more than 30 years. | 服务于我30多年的北极星任务 |
| 41 | So my chess training is the way that I think about business | 所以,我的围棋训练是 我思考商业的方式 |
| 42 | and organizing things and planning and how I think I've been able to break down | 组织事情和计划 以及我如何想 我是如何打破 |
| 43 | very ambitious plans into smaller, more manageable steps. | 非常雄心勃勃的计划 进入较小,更可管理的步骤。 |
| 44 | That all comes from kind of chess thinking, I would say. | 这一切都来自一种象棋的想法,我想说。 |
| 45 | And then using games, first of all, building games, | 然后使用游戏,首先,建设游戏, |
| 46 | learning about engineering projects at scale, running companies, start-ups, | 了解规模工程项目、经营公司、开办企业、 |
| 47 | and then fusing this creativity with engineering. | 然后把创造力和工程学相融合 |
| 48 | Actually, it's what we do today with AI, it's an engineering science. | 实际上,这是我们今天做的AI, 这是一个工程科学。 |
| 49 | So you're fusing creative work, scientific work, | 所以你在灌输创造性的作品 科学的作品 |
| 50 | with very hardcore, cutting-edge engineering. | 与非常坚硬,尖端的工程。 |
| 51 | So that all served together. | 因此,所有服务在一起。 |
| 52 | And then finally on games, as everyone knows, | 最后,在游戏中,大家都知道, |
| 53 | we used games in the early days of DeepMind | 我们在DeepMind早期使用过游戏 |
| 54 | as the perfect proving ground for testing out algorithmic ideas, | 作为测试算法思想的绝佳证明, |
| 55 | 也许最著名的是AlphaGo | |
| 56 | which I think we've just had the 10-year anniversary of | 我认为我们刚刚有 十年的周年纪念 |
| 57 | and was really, looking back now, maybe the start of the modern AI era. | 现在回顾过去 也许是现代AI时代的开始 |
| 58 | When you went into AI professionally in 2010 or so, | 2010年左右你进入AI时 |
| 59 | you started DeepMind, you had this very ambitious vision. | 你开始了DeepMind, 你有一个非常雄心勃勃的愿景。 |
| 60 | You were going to solve intelligence and then solve everything else. | 你本打算解决情报问题 然后解决其他问题 |
| 61 | How's it going? | 怎么样? |
| 62 | Let me expand a little bit. | 让我再扩展一点。 |
| 63 | What has gone according to plan and sort of what has been off-plan? | 计划是什么,计划是什么? |
| 64 | Well, the broad arcs of it have gone unbelievably well, perhaps. | 嗯,它的宽弧 已经取得了令人难以置信的好,也许。 |
| 65 | When we started DeepMind in 2010, | 2010年开始深明时. |
| 66 | can you imagine we used to try to go to VCs in the UK, | 你能想象我们曾经尝试去英国的越共, |
| 67 | which there weren't very many, and with that as the business plan, | 没有太多, 和作为商业计划, |
| 68 | it was literally step one, solve intelligence, | 它其实是第一步, 解决智能, |
| 69 | and step two, use it to solve everything else, | 第二步,用它解决所有问题 |
| 70 | and people were quite confused. | 人们相当困惑。 |
| 71 | But we really meant it, and actually, | 但我们真的是认真的,实际上, |
| 72 | we may go back to exactly using that mission statement, | 我们也许可以重新使用任务说明 |
| 73 | because by solving intelligence, we meant build AGI, | 因为通过解决智能,我们的意思是建设AGI, |
| 74 | ideally also understand the nature of intelligence on the way to building AGI | 最好也理解智能的性质 在建设AGI的路上 |
| 75 | and perhaps using AGI to help us understand our own brains and minds better, | 也许利用AGI来帮助我们更好地理解我们的大脑和头脑, |
| 76 | things like the nature of consciousness, what is creativity, dreaming, | 比如意识的本质 创造力 梦想 |
| 77 | all of these deep mysteries of the mind. | 所有这些深奥的心灵。 |
| 78 | One of the reasons I studied neuroscience was try to learn | 我研究神经科学的一个原因 就是想学习 |
| 79 | from what we understood about the brain as inspiration for algorithmic ideas. | 从我们对大脑的理解来看, 是对算法思想的启发。 |
| 80 | Step one was to try and build AGI, | 第一步是尝试建造AGI, |
| 81 | and then we always had in mind what sort of happened, | 然后我们一直想着发生了什么 |
| 82 | which is that, of course, it's a general purpose technology, | 当然,这是通用技术 |
| 83 | maybe the general purpose technology, | 也许通用技术, |
| 84 | and if it was built in the right way, | 如果它是用正路建造的, |
| 85 | so it was a learning system that was very general, | 所以这是一个非常普遍的学习系统, |
| 86 | what would be the limit of what it could be applied to? | 其适用范围的限度是什么? |
| 87 | It could be applied to almost anything, was the dream, | 几乎可以应用到任何东西,是梦想, |
| 88 | and I think that's what's born out. | 我认为这就是诞生。 |
| 89 | I had specifically in mind for that step two, advancing science and medicine, | 我特别想到了第二步 推进科学和医学 |
| 90 | so that's what I meant by using it to solve everything else, | 所以这就是我用它来解决其他问题的意思 |
| 91 | I meant the big questions in science, all of them. | 我是指科学中的大问题,全部 |
| 92 | I was fascinated by all of them, the nature of time, the nature of reality, | 我迷上他们所有人 时间的性质 现实的性质 |
| 93 | maybe that's the most fundamental one, | 也许这是最根本的 |
| 94 | and I loved physics when I was at school, that was my favourite subject, | 我喜欢物理 当我在学校, 这是我最喜欢的主题, |
| 95 | and I think when you're interested in the big questions, | 我想当你对大问题感兴趣时 |
| 96 | you end up doing physics, probably, | 你最终做物理,也许, |
| 97 | but the reason I decided that there were too many interesting big questions, | 但我决定有很多有趣的大问题 |
| 98 | so how was one to try and tackle all of that in a lifetime, | 如何在一生中努力解决这一切, |
| 99 | and that meant, in my view, building new tools to aid us, | 这就意味着,我认为, 建造新的工具来帮助我们, |
| 100 | the best experts, to make much faster progress in the fields | 最佳专家,以在各领域取得更快的进展 |
| 101 | that they were tackling and the big questions | 他们正在处理和大问题 |
| 102 | and important questions that they were tackling. | 以及他们正在处理的重要问题。 |
| 103 | And then, of course, AI in itself is also a fascinating artefact in itself, | 当然,人工智能本身也是一种迷人的艺术 |
| 104 | a scientific object, one could call, worthy of study itself, | 一个科学的物体,人们可以称为,值得研究本身, |
| 105 | it's almost a new field. | 这几乎是一个新的领域。 |
| 106 | To me, the most fascinating and most important thing to spend one's life on, | 对我来说,最迷人和最重要的事情 一个人的一生, |
| 107 | and I would have been doing it even if it hadn't worked out, | 就算没有成功 我也会这么做 |
| 108 | I would have found some way to be doing this in academia or wherever, | 我会想办法在学术界或任何地方这样做的 |
| 109 | this is what I always planned to spend my life working on, | 这就是我一直计划 花我的一生工作, |
| 110 | and all those things I did earlier were different expressions, | 我之前做的那些事都是不同的表达方式 |
| 111 | gathering the experience and, I suppose, the knowledge | 收集经验和,我想,知识 |
| 112 | to be able to attempt something like DeepMind in 2010 | 在2010年尝试像DeepMind那样的尝试 |
| 113 | when we felt we were ready to make fast progress. | 当我们觉得我们准备 取得迅速的进展。 |
| 114 | And, of course, the second part of that, use it to solve everything else, | 当然,第二部分,用它来解决其他问题 |
| 115 | is now much broader than just science and medicine, | 现在比科学和医学要广泛得多 |
| 116 | although that's where I've tried to personally do my work in, | 虽然这就是我努力去工作的地方 |
| 117 | as well as running the overall organisation. | 以及管理整个组织。 |
| 118 | But, obviously, it's going to be amazing for productivity | 但是,很明显,它将是惊人的生产力 |
| 119 | and many other things in the world outside of just science and medicine. | 以及世界上除了科学和医学以外的许多其他事情。 |
| 120 | As you've been building these different models at DeepMind, | 你在DeepMind建造这些不同的模型 |
| 121 | you started with games and then you went into science, | 你从游戏开始,然后进入科学, |
| 122 | were there particular moments where, | 有特殊的时刻, |
| 123 | I mean, you started with a lot of conviction, | 你从很多信念开始 |
| 124 | but I'm curious if there were particular moments where you saw | 但我好奇你有没有看到 |
| 125 | this is actually going to work, | 这其实是要去工作, |
| 126 | 与 AlphaGo 并播放 MoveThru, | |
| 127 | the Move Against the Necessity. | 《反对必要性的行动》。 |
| 128 | There were many moments where I thought it wasn't going to work, put it that way. | 有很多时候我以为它不会起作用 这么说吧 |
| 129 | Some of the ones I remember really well are, | 有些我记得很清楚的是, |
| 130 | we started with games because they're self-contained, | 我们从游戏开始 因为它们是自成一体的 |
| 131 | they obviously were designed by other humans to be challenging or fun | 它们显然被其他人类设计成具有挑战性或乐趣 |
| 132 | for other humans to play. | 给其他人玩 |
| 133 | I love games, they're often microcosms of a lot of real-world scenarios. | 我喜欢游戏,它们往往是许多现实世界情景的缩影. |
| 134 | If you think of Go or Poker or Chess, | 如果你想到Go或者扑克或者国际象棋 |
| 135 | I often thought one of the courses I'd have in an MBA or business school course | 我经常想在MBA或商学院的课程里 选一个 |
| 136 | would be a games module to study those types of games, diplomacy. | 将是一个游戏模块 研究这些类型的游戏,外交。 |
| 137 | They all have really interesting aspects of the best games of real life, | 他们都有真正有趣的方面 最好的游戏现实生活, |
| 138 | and you can obviously practice many times in a kind of safe scenario. | 你显然可以在安全的情况下 练习很多次 |
| 139 | That's what I think games are really useful for, | 我认为游戏真的有用 |
| 140 | and that applies to AI systems that are learning too. | 这也适用于正在学习的人工智能系统。 |
| 141 | They're neat environments, they're challenging, | 它们环境整洁,挑战性强 |
| 142 | and they have clear objective functions, | 他们有明显的功能, |
| 143 | which also was very important for our early days of reinforcement learning. | 这对我们早期的强化学习也很重要。 |
| 144 | Almost no one had used reinforcement learning for any kind of scaled-up problem. | 几乎没有人利用强化学习来解决任何扩大的问题。 |
| 145 | It was obviously an academic discipline, | 这显然是一个学术学科, |
| 146 | but it was used mostly for toy problems like little grid worlds. | 但它主要用于玩具问题 像小网格世界。 |
| 147 | It wasn't clear it could scale up to anything major, | 并不清楚它能扩大到什么大, |
| 148 | and so we started with, I would say, | 所以我们开始,我要说, |
| 149 | the most famous set of but most basic games that had become world-popular, | 最有名但最基本的游戏 已经成为世界流行, |
| 150 | which were Atari games from the 70s. | 那是70年代的阿塔里游戏 |
| 151 | We started with the simplest game of all, which was Pong, | 我们从最简单的游戏开始 就是庞 |
| 152 | the bat and the ball, just two bats and a ball. | 蝙蝠和球,只有两只蝙蝠和一个球。 |
| 153 | There's an inbuilt AI system that controls your opponent | 有一个内置的AI系统控制你的对手 |
| 154 | and uses all the information that the game has about where the ball is | 并使用游戏中关于球所在位置的所有信息 |
| 155 | {\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}等等 把球棒移开 | |
| 156 | What we wanted to do was could you play Pong just from the pixels on the screen, | 我们想做的是 你能从屏幕上的像素上弹出Pong吗? |
| 157 | so the raw data, the raw visual input, and no other information, | 所以原始数据,原始视觉输入, 没有其他信息, |
| 158 | no privileged information about no access to the insides of the program | 没有关于无法访问程序内部的特权信息 |
| 159 | about where the ball is or the speed of it and so on, | 关于球在哪里或者它的速度等等, |
| 160 | which obviously the program knows. | 很明显,程序知道。 |
| 161 | But we didn't give the DQN system, as it was called, our Atari system, | 但我们没有给DQN系统, 它被称作,我们的阿塔里系统, |
| 162 | any of that information. | 任何这些信息。 |
| 163 | It just got the 20,000 pixels on the screen. | 它刚刚把两万像素放在屏幕上 |
| 164 | And 20,000 pixels seems sort of trivial now, | 而两万像素现在似乎有点微不足道, |
| 165 | but back in 2010, that was an enormous amount of input data. | 但早在2010年, 输入的数据就非常多。 |
| 166 | No one had ever dealt with something that complex | 没有人处理过这么复杂的事情 |
| 167 | and then multiplied by all the frames that you were doing. | 然后乘以你做的所有框架。 |
| 168 | And for about, it felt like six months. | 大约六个月了 |
| 169 | Maybe it was only two months, | 也许只有两个月 |
| 170 | but we couldn't win a single point at Pong. | 但我们在邦一分都赢不了 |
| 171 | So it was jerking the bat around. | 所以它正在打球棒周围。 |
| 172 | It was like, oh, is it ever going to even be able to control the bat? | 这就像,哦,它会 甚至能够控制蝙蝠? |
| 173 | And of course, it had no notions of any of these things. | 当然,它对这些东西没有任何概念。 |
| 174 | It was just losing 21-0 to the inbuilt AI. | 它刚刚损失21-0到内置AI。 |
| 175 | And I did think, and we had a couple of different ways of trying to attack this, | 我想,我们有几种不同的方式 试图攻击这个, |
| 176 | and we had almost no money, the runway, | 我们几乎没有钱,跑道, |
| 177 | the couple of million dollars of funding that we had, | 几百万美元的资金,我们有, |
| 178 | which wouldn't even cover an intern these days, | 这几天连实习生都没有 |
| 179 | which is good for all of you students, | 这对你们所有的学生都是有益的, |
| 180 | was our entire funding. | 是我们全部的资金 |
| 181 | And we were taking no salary, | 我们没有工资, |
| 182 | and it was running out, the money. | 钱用完了 |
| 183 | And I was like, oh, well, maybe it turns out, | 我当时想,也许结果 |
| 184 | maybe we are still 10 years too early. | 也许我们还有10年还太早 |
| 185 | Maybe we're 20 years too early. | 也许我们早20年了 |
| 186 | And then magically, it got a point, and it was like, oh, maybe it was just luck. | 然后神奇的,它得到了一个点, 它就像,哦,也许只是运气。 |
| 187 | And then it started winning a lot of points, | 然后它开始赢得很多积分, |
| 188 | and then it started winning the games. | 然后它开始赢得比赛。 |
| 189 | And then it was like, OK, we have liftoff now. | 然后它就像,OK, 我们已经升空了。 |
| 190 | So now, and those of you working machine learning will know this, | 所以,现在,你们这些工作机器学习的人 将会知道这一点, |
| 191 | if you get a foothold, you can usually hill climb your way out of that. | 如果你有立足之地,你通常可以爬山 从那里爬出来。 |
| 192 | That's been the history of AI, I would say. | 我想这是AI的历史。 |
| 193 | Once you have something working, there's usually a way of optimizing it more. | 一旦你有了工作, 通常有一个方法 优化它。 |
| 194 | And that's what turned out with Atari. | 而阿塔里的情况就是这样. |
| 195 | So that was our first big result, and our first nature paper | 所以这是我们的第一个大结果 也是我们的第一个自然论文 |
| 196 | was really the first deep reinforcement learning model, certainly at scale. | 真正是第一个深入强化的学习模式, 肯定在规模上。 |
| 197 | Combining deep learning to learn the domain and deal with the perceptual inputs | 结合深入学习学习领域,处理知觉投入. |
| 198 | and the complexity of the inputs, find the patterns in it, | 以及投入的复杂性, 找到其中的模式, |
| 199 | and then reinforcement learning built on top of that | 然后再加强学习 在此基础上 |
| 200 | to kind of make the decisions and do the planning. | 做决定和计划 |
| 201 | And then, of course, that culminated in AlphaGo, which was always our aim. | 然后,当然,这 最终在AlphaGo, 这一直是我们的目标。 |
| 202 | Dave Silver and I was head of that project. | Dave Silver和我是这个项目的负责人 |
| 203 | We were friends as undergrads at Cambridge, | 我们是剑桥大学的同学 |
| 204 | and we were discussing it since our undergrad about, you know, | 我们从本科毕业后就开始讨论这个问题 |
| 205 | we were there in the mid-90s, the Deep Blue Kasparov match happened | 我们在90年代中期在那里, 深蓝卡斯帕罗夫比赛发生 |
| 206 | while we were at college. | 当我们在大学。 |
| 207 | Of course, I was fascinated by it both from the chess and the AI point of view, | 当然,从国际象棋和人工智能的角度,我对它很着迷, |
| 208 | but I was more impressed with Kasparov's brain than I was with Deep Blue, | 但我对卡斯帕罗夫的大脑印象比对深蓝更深刻 |
| 209 | because Kasparov, you know, with his incredible mind, | 因为卡斯帕罗夫,你知道, 他不可思议的头脑, |
| 210 | still one of the biggest chess geniuses that there's been of all time, | 仍然是有史以来最大的棋手之一 |
| 211 | he was able to basically compete on an equal footing | 他基本上可以平等竞争 |
| 212 | with this supercomputer brute force machine next to him. | 他旁边有一台超级计算机野蛮的机器 |
| 213 | But, of course, he could do all the other things with his mind, | 但是,当然,他可以做 所有其他事情,用他的思想, |
| 214 | speak five languages, do his politics, drive cars, | 说五种语言,做政治,开车 |
| 215 | all the rest of the things humans can do. | 人类能做的其余一切 |
| 216 | To me, that was, you know, incredible. | 对我来说,那是,你知道,不可思议的。 |
| 217 | That's much more impressive. | 这更令人印象深刻。 |
| 218 | So there was something missing from the Deep Blue system. | 所以深蓝系统还少了一些东西 |
| 219 | And, obviously, those techniques, those expert system techniques | 很明显,这些技术,这些专家系统技术 |
| 220 | where you hand-curate the heuristics and then you use brute force search on top, | 在那里,你用手校准 高压,然后你使用野蛮的搜索 在上面, |
| 221 | which is still how a lot of traditional chess programs work today, | 这也是很多传统围棋项目今天如何运作, |
| 222 | that works for chess, but it's never worked for Go, | 做象棋,但它从来没有为吴, |
| 223 | because Go's too esoteric a game. | 因为Go太喜欢玩游戏了 |
| 224 | It hasn't got material. | 它没有材料。 |
| 225 | Every piece is worth the same. | 每块都值钱 |
| 226 | It's all about patterns and intuition. | 这都是规律和直觉 |
| 227 | Even the top Go players, that's how they play it. | 即使是顶尖的Go球员, 他们就是这样玩的. |
| 228 | So we realized, we sort of thought, | 所以,我们意识到, 我们有点想, |
| 229 | okay, if someone could actually get to world champion level at Go, | 好吧,如果有人能真正 到世界冠军级别 在吴, |
| 230 | it's not just about, really, that was an aside getting to that level. | 这不仅仅是关于,真的, 这是一个边 进入这个水平。 |
| 231 | It was more about the approach he would have taken | 更多是他会采取的方法 |
| 232 | would probably be a really interesting algorithmic approach, | 可能是一个非常有趣的算法方法, |
| 233 | and maybe and hopefully would generalize to other domains. | 并且也许而且希望能够概括到其他领域。 |
| 234 | AlphaGo就是这样的 | |
| 235 | And then it went beyond our wildest dreams, really, | 然后它超越了我们最疯狂的梦想,真的, |
| 236 | because not only did it win the match against Lisa Dole in 2016, | 因为它不仅在2016年对阵丽莎·多尔的比赛中获胜, |
| 237 | it also created, famously, new strategies that had never been seen before, | 也创造了前所未有的新策略 |
| 238 | even though we've played Go. | 尽管我们玩过Go |
| 239 | Go is the oldest game humanity has invented, 2,000 years old, | Go是人类发明的最古老的游戏 已有2000年历史了 |
| 240 | 2,000 plus years old, and been played professionally for hundreds of years. | 两千多岁,被戏弄数百年. |
| 241 | And we hadn't discovered those strategies. | 我们没有发现这些策略。 |
| 242 | So that was sort of double whammy for me. | 对我来说,这有点夸张。 |
| 243 | I was waiting for that moment that AI was able to come up | 我等的那一刻,AI能够出现 |
| 244 | with something novel. | 有小说 |
| 245 | There's more levels of creativity than that, beyond that, | 创造力比这还要高, 不仅如此, |
| 246 | but at least it was a novel idea. | 但至少这是一个新想法。 |
| 247 | And then that, for me, was what I was waiting for, | 然后,对我来说,这就是我等待的, |
| 248 | to then start using AI for science. | 然后开始将AI用于科学。 |
| 249 | So the moment we got back from Sol, we started the AlphaFold project. | 我们从索尔回来的那一刻 我们开始了AlphaFold项目 |
| 250 | So let's talk about the science a bit, | 我们来谈谈科学吧 |
| 251 | because then you went into the protein folding problem, | 因为你进入蛋白质折叠的问题, |
| 252 | and again, you picked a problem where there was data | 再说一次,你挑了一个问题 在有数据 |
| 253 | and where there was a clear objective function | 在有明确客观职能的地方 |
| 254 | in terms of thinking about protein folding, and it worked. | 在蛋白质折叠的思考上,它奏效了. |
| 255 | I mean, you actually managed to solve this long-standing problem | 我的意思是,你实际上 解决了这个长期的问题 |
| 256 | of predicting protein structure. | 预测蛋白质结构 |
| 257 | You did something very interesting when you came up with AlphaFold, | 你提出AlphaFold时做了一件很有趣的事 |
| 258 | which was obviously a huge, noble, worthy scientific breakthrough, | 这显然是 一个巨大的,崇高的,有价值的科学突破, |
| 259 | probably also of commercial value. | 可能也具有商业价值。 |
| 260 | And you just gave it away for free. | 你只是免费送走它 |
| 261 | I'm curious, how did you come to that decision? | 我很好奇 你是怎么做出这个决定的? |
| 262 | Did you think about other ways of going about it? | 你考虑过其他方法吗? |
| 263 | Why give it away? | 为什么送人? |
| 264 | Yeah, so we picked the protein folding problem. | 是啊,所以我们选择了蛋白质折叠问题。 |
| 265 | I had my kind of eye on that since also my undergrad days at Cambridge. | 自从我在剑桥大学毕业后 我就一直关注这个 |
| 266 | Just when I first came across it, I had a few biologist friends | 刚发现的时候 我有几个生物学家的朋友 |
| 267 | who were obsessed with the protein folding problem, | 他们迷恋蛋白质折叠问题, |
| 268 | and actually they ended up becoming structural biologists, of course, | 他们最终成为了结构生物学家,当然, |
| 269 | in their career. | 在他们的职业生涯。 |
| 270 | And one specifically, I remember, he was, you know, | 一个特别的,我记得,他是,你知道, |
| 271 | every time we were in the pub playing table football or something, | 每次我们在酒吧玩桌球什么的, |
| 272 | he would be talking about, obsessively, | 他会在谈论,执着地, |
| 273 | how this was the most important problem in biology. | 这是生物学中最重要的问题 |
| 274 | And more importantly, I think of it as a root-note problem, | 更重要的是,我认为这是一个根注问题, |
| 275 | like if you could unlock that and find the structures of proteins, | 就像你可以解锁 找到蛋白质的结构, |
| 276 | that would unlock whole, you know, new avenues of research, | 会打开整个,你知道, 新的研究途径, |
| 277 | things like drug discovery, obviously we're trying to push that, | 比如毒品的发现, 显然我们正试图推动, |
| 278 | but also fundamental biology and disease understanding. | 还有基本的生物学和疾病认识 |
| 279 | So this was a problem worth really spending a lot of attention and time on | 因此,这是一个问题 值得真正花费 很多关注和时间 |
| 280 | because of the downstream effects that it would have. | 因为它会产生下游影响 |
| 281 | It also felt to me, it was a fascinating problem, | 我也觉得,这是一个令人着迷的问题, |
| 282 | it felt to me like the ultimate puzzle, you know, 3D puzzle of | 对我来说就像一个三维谜题 |
| 283 | how does this sort of amino acid sequence, | 这种氨基酸序列是怎么形成的? |
| 284 | think of it as genetic sequence, fold up into this 3D structure. | 把它看作是基因序列, 折叠成这个三维结构。 |
| 285 | It's an amazingly interesting, intricate thing. | 这是一个惊人的有趣,复杂的事情。 |
| 286 | And the more I looked into proteins, | 我越看蛋白质, |
| 287 | the more incredible my respect and wonder is for biology, | 我对生物学的尊敬和好奇 |
| 288 | like these unbelievable little bio-nano machines, | 像这些难以置信的小 生物纳诺机器, |
| 289 | you know, everything on life obviously depends on proteins | 你知道,一切关于生命 显然取决于蛋白质 |
| 290 | and as you start looking at their structure, | 当你开始观察它们的结构时, |
| 291 | you start understanding their function. | 你开始理解他们的功能。 |
| 292 | So this is fascinating to me as a science question. | 因此,这对我来说是一个有趣的科学问题。 |
| 293 | And then, yes, there was the clear objective, | 然后,是的,有明确的目标, |
| 294 | it's sort of like minimizing the free energy in the system, | 它有点像 尽量减少系统中的自由能量, |
| 295 | presumably this is how physics is doing, | 假设物理学就是这样 |
| 296 | it's why the body, you know, these proteins fold in milliseconds in your body, | 这就是为什么身体,你知道, 这些蛋白质折叠 在你的身体的毫秒, |
| 297 | you know, billions of times a second. | 你知道,一秒钟几十亿次。 |
| 298 | So somehow physics has solved this. | 所以物理学已经解决了这个问题。 |
| 299 | And it can't be, there must be some topology, let's say, | 这不可能,一定有一些地形学,比如, |
| 300 | that you could learn, maybe with a deep learning system, | 你可以学习,也许 通过深层的学习系统, |
| 301 | 就像我们和AlphaGo一样 | |
| 302 | to find a great move in Go, a great strategy out of the, you know, | 找到一个伟大的动作在吴, 一个伟大的策略出,你知道, |
| 303 | more possibilities than there are atoms in the universe in Go, | 比宇宙中的原子还可能 |
| 304 | and protein folds are even larger search space than that. | 蛋白质折叠的搜索空间甚至更大 |
| 305 | But there's some way to narrow that down in a sensible way. | 但有办法以明智的方式缩小范围。 |
| 306 | You learn a kind of heuristic using the deep learning models | 你用深层的学习模式来学习一种热力 |
| 307 | to then guide your search for that to become tractable. | ,然后引导您搜索它成为可引导。 |
| 308 | And it felt like a really analogous problem in science to what we'd solved in Go, | 这感觉就像一个非常相似的 科学问题 我们解决在吴, |
| 309 | sort of applying some of those same approaches, | 某种程度上是采用其中一些相同的做法, |
| 310 | those same theories to this domain. | 同样的理论 这个领域。 |
| 311 | And then the other thing was there was obviously 50 years' worth of | 还有一件事情是 很明显有50年的时间 |
| 312 | painstaking crystallography, structural biology work | 辛苦的晶体学、结构生物学 |
| 313 | by many great labs and people. | 许多伟大的实验室和人。 |
| 314 | And after all of that effort, there was about 150,000 structures | 经过所有的努力, 有大约15万个结构 |
| 315 | in the PDB, the main database, which isn't a lot, actually. | 在PDB中,主数据库,实际上并不多. |
| 316 | Obviously, it's a huge amount of effort that's gone into that, | 显然,这是一个巨大的努力 已经进入, |
| 317 | but there's 200 million proteins, | 但有2亿个蛋白质 |
| 318 | and 150,000 also for machine learning systems is a very small amount of data. | 15万也用于机器学习系统 数据数量很少 |
| 319 | So most people thought it was at least 10, 20 years away | 所以大多数人都认为 至少有10,20年了 |
| 320 | before we would have enough data and the right types of algorithms to tackle that. | 之前,我们将有足够的数据 和正确的算法类型 来解决这一点。 |
| 321 | But we felt that using every technique we knew in the end | 但是我们觉得利用我们最后掌握的所有技术 |
| 322 | that we could make progress with that. | 我们能在这方面取得进展。 |
| 323 | And it turned out to be the case. | 结果却成了这样. |
| 324 | And then when we decided how would we make the maximum impact with this, | 然后当我们决定 如何发挥最大的影响 |
| 325 | it was obvious to me that we should fold all the proteins | 很明显,我们应该把蛋白质都叠起来 |
| 326 | because not only was AlphaFold accurate, it was extremely fast. | 因为不仅AlphaFold准确,而且速度极快. |
| 327 | It could fold a protein in a matter of seconds | 它可以在几秒钟内折叠蛋白质 |
| 328 | and then collaborate with, in the end, European Bioinformatics Institute in Cambridge | 最后与剑桥欧洲生物信息研究所合作 |
| 329 | which hosts many of the biggest databases, biology databases scientists use, | 拥有许多最大的数据库 生物学数据库的科学家们 |
| 330 | and just hosts the entire 200 million protein structures on their database | 只要把整个2亿个蛋白质结构放在数据库里 |
| 331 | and just allow it to be as simple as a kind of Google search | 让它和谷歌搜索一样简单 |
| 332 | to just find your protein structure, | 去找你的蛋白质结构 |
| 333 | along with the confidence intervals the machine learning system had | 以及机器学习系统的信心间隔 |
| 334 | about which parts of the protein structure it was confident on, | 关于蛋白质结构的哪个部分,它相信, |
| 335 | which is very important for our biologists to know. | 这对我们的生物学家很重要 |
| 336 | So we put that all together, | ┮ир硂ㄇ癬 |
| 337 | and it was, of course, could have been very valuable. | 当然,它可能非常宝贵。 |
| 338 | I don't know how many billions of dollars or whatever. | 不知百千亿那由他等. |
| 339 | I mean, it depends how you calculate it. | 我的意思是,这取决于你怎么计算它。 |
| 340 | To do that experimentally would be incalculable cost, | 实验性地这样做是无法估量的成本, |
| 341 | but it would have been hugely valuable to keep proprietary. | 但保持专有性会很有价值 |
| 342 | But for us, it felt like we would only be able to scratch the surface of the downstream impact | 但对我们来说 感觉我们只能刮下游撞击的表面 |
| 343 | that putting all those structures out in the world could have on our own | 让世界上所有的结构 都靠我们自己 |
| 344 | because there's three million researchers around the world that use AlphaFold pretty much every day, | 因为全世界有三百万研究人员 每天使用AlphaFold, |
| 345 | almost every biologist medical researcher in the world. | 几乎是世界上每个生物学家的医学研究员. |
| 346 | There's no way one organization could have done that. | 一个组织不可能这么做的 |
| 347 | So it was obviously the right thing to do. | 所以这显然是正确的事情。 |
| 348 | We also had depended on public data to train the first versions of AlphaFold, | 我们还依靠公共数据来训练AlphaFold的第一个版本, |
| 349 | so it only felt right to give back to that community, | 所以,它只是觉得正确的 回到社区, |
| 350 | the structural biology community, | 结构生物学界, |
| 351 | this amazing resource that was amplifying the resource they had basically built. | 这惊人的资源 正在扩大 他们基本建设的资源。 |
| 352 | And so it wasn't even a question for me. | 所以这对我来说甚至不是一个问题。 |
| 353 | And it was great that also the executives at Google also love science and totally got that. | 谷歌的高管们也喜欢科学 并且完全明白这一点,这真是太好了。 |
| 354 | I don't think all companies would have made that decision. | 我不认为所有公司都会做出这一决定. |
| 355 | So I give them a lot of kudos on that, too. | 所以,我给他们很多 kudos,太。 |
| 356 | That was an easy discussion. | 讨论很轻松 |
| 357 | And then we've tried to ourselves push that downstream with isomorphic labs, | 然后我们试图自己推 下游与异形实验室, |
| 358 | an alphabet spin out that is building, you can think, | 一个字母旋转 正在建设,你可以认为, |
| 359 | 还有几类AlphaFold级突破, | |
| 360 | putting them together into a way that will accelerate hopefully drug discovery, | 把它们拼凑在一起 希望能加速毒品的发现 |
| 361 | take it down from years to months, maybe even one day weeks, | 数年到数月, 甚至一天的周, |
| 362 | just like we did with protein structures, | 就像我们做的蛋白质结构, |
| 363 | which used to take years for a single one, | 以一年为一年, |
| 364 | and then we could do it in seconds. | 然后我们可以在几秒钟内完成。 |
| 365 | That's one of the really exciting areas of the future with AI. | 这是未来AI最令人兴奋的领域之一. |
| 366 | I want to turn for a minute to something you said earlier this week. | 我想谈谈你本周早些时候说的话。 |
| 367 | You were in the news this week because at a big Google event, | 你本周上新闻 因为谷歌的大型活动 |
| 368 | you said that we're in the foothills of the singularity. | 你说我们在奇特的山上 |
| 369 | Yes, it got quite a lot of pickup, that line. | 是的,它得到了很多的皮卡, 该线。 |
| 370 | It got a lot of pickup. | 有很多货 |
| 371 | And I understand that maybe the Google Press team might not have been so thrilled about it. | 我理解,也许Google新闻团队 可能没有这么激动。 |
| 372 | But since you're out there saying that, what did you mean by that? | 但是既然你在外面说,那是什么意思? |
| 373 | The full thing I said to close the conference with was, | 我说结束会议的全部事情是, |
| 374 | when we look back at this time, | 当我们回头看这一刻, |
| 375 | I think that maybe I'm thinking 10 years from now, | 我想,也许我在想 10年后, |
| 376 | I think we'll realize that we were standing in the foothills of the singularity now. | 我想我们会意识到 我们当时站在奇特的山坡上 |
| 377 | What I mean by that, and the reason I chose that word, | 我的意思是,我选择这个词的原因, |
| 378 | is that there's the technology, which is AGI. | 技术就是AGI |
| 379 | We've been calling AGI this next version of really general artificial intelligence. | 我们一直称AGI是下一个真正的普通人工智能版本。 |
| 380 | I believe that we're only a few years away from that. | 我相信我们离那只有几年时间了 |
| 381 | Maybe like 20, 30, plus or minus a year, which is astounding to think, really. | 也许像20,30,加或减一年, 这是惊人的思考,真的。 |
| 382 | And then the era, I think that will be such an enormous transformative technology, | 之后的时代,我认为这将是 一个巨大的变革性技术, |
| 383 | it's going to effectively be a new human era. | 这将是一个新的人类时代。 |
| 384 | And that's what I'm meaning by the singularity is that, | 这就是我所谓的奇特之处 |
| 385 | and what many science fiction writers have written about that, | 以及许多科幻小说作家所写的 |
| 386 | is that it's describing the era that we will be in, | 是它描述的时代 我们将进入, |
| 387 | in and around when the advent of AGI happens. | 当AGI出现时,在周围 |
| 388 | So, and I think we can feel this year, I would say, | 所以,我想我们可以感觉到 今年,我会说, |
| 389 | even though I've been working towards this for 30 years, | 尽管我为此努力了30年 |
| 390 | I think this year, with the way the agents are working and tool use, | 我认为今年,随着 特工的工作方式和工具的使用, |
| 391 | it started to become really useful for, you know, still early days of it, | 它开始变得非常有用, 你知道,它仍然早期, |
| 392 | but genuinely useful in people's workflows. | 但真正有用 在人们的工作流程。 |
| 393 | And we can sort of see what extra things need to be done, | 我们可以看看还有什么需要做的, |
| 394 | and all of us, the leading labs, are working on that. | 我们所有人,主要的实验室, 都在努力。 |
| 395 | I think this is the beginnings of that. | 我想这是开始。 |
| 396 | But the foothills, I still think there's a lot more work, | 但是在山上,我仍然认为还有许多工作要做 |
| 397 | and it's just the beginnings, and it's not any one thing, | 这仅仅是开始, 而且它不是任何一件事, |
| 398 | it's several different technologies, several use cases that I see, | 这是几种不同的技术, 几种使用的案例,我看到, |
| 399 | several things that I thought were maybe a bit further out, | 几件事,我认为 也许更远, |
| 400 | turned out to be now, that are coming together, | 原来是现在, 那些正在走到一起, |
| 401 | that make me feel that in aggregate, | 我总觉得 |
| 402 | and to the extent that I wanted to say that. | 并且就我想说的而言 |
| 403 | Because I think society needs to hear that, | 因为我认为社会需要听到这一点, |
| 404 | because we don't have long to prepare for what that means. | 因为我们没有很长的时间来准备那是什么意思。 |
| 405 | And it's going to be enormously profound, I think. | 我认为,这将是非常深刻的。 |
| 406 | And the future, in my view, is still to be written. | 而未来,在我看来,仍然有待写作. |
| 407 | But these next few years are going to be very critical | 但接下来的几年将是非常关键的 |
| 408 | as to which way that will go, | 至於那条路, |
| 409 | and how we collectively want that to look like. | 以及我们集体希望它看起来像什么。 |
| 410 | If you look at surveys of how people perceive AI, | 如果你看看调查 人们如何看待AI, |
| 411 | in this country in particular, it's very negative right now. | 特别是这个国家,它是非常负面的 现在。 |
| 412 | And maybe more negative here than even in other countries. | 也许比其他国家更消极。 |
| 413 | And there's probably lots of things driving that, | 可能有很多东西在驱动 |
| 414 | concerns about privacy, or state control, | 对隐私或国家控制的关切, |
| 415 | or the size of the tech companies, or jobs. | 或技术公司的规模,或工作。 |
| 416 | You're running one of the leading labs, | 你在经营一个主要的实验室 |
| 417 | and how do you think about that public concern about the technology? | 你觉得公众对技术的关注如何? |
| 418 | I think the public's right to be concerned. | 我认为公众有权利关心 |
| 419 | I think that there are things that, | 我认为有些事情, |
| 420 | and I'm concerned about several aspects of what the technology is. | 我担心技术的几个方面。 |
| 421 | It's a dual-purpose technology. | 双功能技术. |
| 422 | It's something this profound. | 实在是太深刻了 |
| 423 | I sometimes describe it, quantify it, | 我有时会描述它,量化它, |
| 424 | as ten times the impact the Industrial Revolution was, | 工业革命的十倍影响 |
| 425 | ten times faster. | 10倍快 10倍快 10倍快 |
| 426 | And so it's taking place over a decade instead of a century. | 因此,它发生 十多年而不是一个世纪。 |
| 427 | So that's like a hundred X of the Industrial Revolution. | 这就像工业革命的一百X。 |
| 428 | And it's probably an underestimate, to be honest, | 老实说,这可能是一个低估, |
| 429 | but that's probably enough for us to try and comprehend and deal with. | 但是这也许足以让我们 尝试理解和处理。 |
| 430 | And so, of course, I think it's super exciting. | 所以,当然,我觉得这太刺激了。 |
| 431 | There's going to be amazing things that are going to happen. | 将会发生惊人的事情。 |
| 432 | We're trying to do that with solving all disease. | 我们试图用解决所有疾病的方法做到这一点。 |
| 433 | I think a lot of the other challenges facing society today, | 我认为今天社会面临的许多其他挑战, |
| 434 | from climate to energy to disease, will be helped by AI. | 从气候到能源到疾病,将由大赦国际帮助。 |
| 435 | I'm sure of it. | 我敢肯定这一点。 |
| 436 | And in fact, I'd be much more worried about those challenges | 事实上 我更担心那些挑战 |
| 437 | if I didn't think something like AI was coming down the line. | 如果我不认为 像AI的东西 正在下线。 |
| 438 | But it's going to cause a lot of change and disruptions, | 但它会引起很多变化和干扰 |
| 439 | and actually both on the technical side, economic and philosophical. | 实际上,在技术方面,经济方面和哲学方面都是如此。 |
| 440 | And I think we've got to think through very thoughtfully | 我想我们必须仔细思考 |
| 441 | and bring together all parts of society to discuss this, | 并召集社会各界讨论此事, |
| 442 | not just the technologists. | 不只是技术人员 |
| 443 | The technology and the safety of the technology is just one piece of this. | 技术和技术的安全只是其中的一部分. |
| 444 | It needs economists, social scientists, human and humanity experts | 它需要经济学家、社会科学家、人类和人类专家 |
| 445 | to kind of chart out what is going to happen next. | 来找出接下来会发生什么 |
| 446 | And I think one of the reasons it's negative here is that, specifically, | 我认为这里负面的原因之一是... |
| 447 | because it's different in other countries. | 因为它在其他国家是不同的。 |
| 448 | For example, when we came back from the summit in India, | 例如,当我们从印度的峰会回来时, |
| 449 | it's hugely popular with the youth of India | 受到印度年轻人的欢迎 |
| 450 | because they see the opportunities that it's going to democratize for them, | 因为他们看到机会 他们将民主化, |
| 451 | having access to basically the same tools that you would have needed | 拥有基本相同的工具 |
| 452 | to go to Silicon Valley for. | 去硅谷为。 |
| 453 | We're in an amazing moment in that everyone can access | 我们正处在一个令人惊叹的时刻 每个人都可以进入 |
| 454 | pretty much what's going on in the frontier labs, | 几乎在前沿实验室发生了什么 |
| 455 | but only with a delay of just a few months. | 但只拖延了几个月 |
| 456 | That's unheard of, really, if you think about that. | 这是没有听说过的,真的, 如果你想想。 |
| 457 | So there's incredible things. | 故有不可思议之事. |
| 458 | But I think it's also partly the way some of my peers are articulating. | 但我认为,这也是我的一些同行 正在阐述的部分方式。 |
| 459 | I think they're being very careful with their communication and understanding. | 我认为他们对于沟通和理解非常小心. |
| 460 | They're being way too certain, I would say, with some of their pronouncements, | 我敢说,他们太确定 他们的一些声明, |
| 461 | where I think actually there's just huge uncertainty, | 在那里我认为,实际上, 只是巨大的不确定性, |
| 462 | and that is worrying in itself, | 这本身就是令人担忧的, |
| 463 | but it also means that nothing is decided, in my opinion. | 但它也意味着任何决定,在我看来。 |
| 464 | I think it's kind of unknown, and I think anyone who says, | 我认为这是未知的, 我认为任何人谁说, |
| 465 | I think directionally I could tell you some things, | 我想我可以告诉你一些事情 |
| 466 | but I think a lot of it really depends on the actions taken in the next few years. | 但我认为,其中很多确实取决于今后几年采取的行动。 |
| 467 | And also what the youth of today, the many students in the room today, | 还有今天的青春,今天的很多学生, |
| 468 | you're going to be forming, you're the first generation to kind of grow up | 你将要形成, 你是第一代 种成长 |
| 469 | in AI native, should we say, like I did, computer native. | 在AI本地,如果我们说, 像我一样,计算机本地。 |
| 470 | And just like every generation, you'll master these technologies, | 就像每代人一样 你会掌握这些技术 |
| 471 | become super productive with them, | 与他们一起成为超级生产力者, |
| 472 | and I actually think over the next 10 years, at least, hard to predict beyond that, | 我实际上认为在未来10年里,至少, 很难预测, |
| 473 | you'll almost be superpowered with those if you use them in the right way. | 如果你用正确的方式使用它们,你几乎会拥有超能力。 |
| 474 | The amount of creativity and projects you can do | 创造力和项目量 |
| 475 | and the amount an individual will be able to do. | 和一个人能做的金额。 |
| 476 | But maybe that will change the nature of jobs. | 但也许这会改变工作的性质。 |
| 477 | There'll be more entrepreneurial, small entrepreneurial things | 会有更多的企业,小企业的东西 |
| 478 | rather than big companies. | 而不是大公司。 |
| 479 | I don't know. | 师曰. |
| 480 | It's going to change a lot. | 将大变化. |
| 481 | And I think part of that is for society to come together | 我认为其中的一部分是 社会聚集在一起 |
| 482 | and really take this exponential seriously, | 认真对待这个指数, |
| 483 | not just the technologists, economists and others, | 不只是技术人员,经济学家和其他人, |
| 484 | we were discussing this last night, | 我们昨晚在讨论这个 |
| 485 | need to take this seriously right now, | 现在需要认真对待这一点, |
| 486 | and start charting out what does that look like | 并开始绘制 是什么样子 |
| 487 | if we're in a post-scarcity world, for example, | 举例来说,如果我们处于饥饿后的世界, |
| 488 | how does everyone benefit from that? | 每个人如何从中获益? |
| 489 | A lot of it's about how does, it's obviously not correct | 很多事情都是关于怎么做的 很明显是不对的 |
| 490 | for just a few people or even a few companies | 只有少数人甚至几个公司 |
| 491 | or even a few nations to be benefiting from this technology. | 或甚至几个国家将受益于这种技术。 |
| 492 | It needs to be broad, it's going to affect all of humanity. | 需要宽广,影响全人类. |
| 493 | It has to broadly accrue the benefits to everyone. | 它必须广泛地为每个人带来好处。 |
| 494 | But how is that going to be done? | 但是,如何做到这一点? |
| 495 | And we really need, a lot of us have been talking about this for a while, | 而我们真的需要, 我们很多人已经谈论了一段时间, |
| 496 | but we really need answers now and concrete things and actions to be taken. | 但我们现在确实需要答案,需要具体的东西和行动。 |
| 497 | And I plan to do my bit on that. | 我打算尽我的责任 |
| 498 | I've been thinking a lot about this over the years | 多年来我一直想着这个 |
| 499 | and planning and building sort of influence on this. | 并对此进行规划和建设。 |
| 500 | And I will do what I can. | 我会尽我所能的 |
| 501 | Obviously, we're an important actor, but we're only one actor in this space. | 显然,我们是一个重要的演员,但我们只是这个空间中的一个演员. |
| 502 | But I hope, I know that the good news is, | 但我希望,我知道好消息是, |
| 503 | I know that all of the sort of leading labs and the leaders of those labs, | 我知道所有的主要实验室 和这些实验室的领导人, |
| 504 | although they disagree on a lot of things, | 虽然他们有很多分歧 |
| 505 | they do worry about these sorts of issues coming down the line. | 他们确实担心这些问题会落到一线。 |
| 506 | But we need more forums to allow them to, | 但我们需要更多的论坛,以便使它们能够, |
| 507 | for us to come together to discuss these things more candidly. | 来更坦率地讨论这些事情 |
| 508 | And I think that's what probably the public is detecting, | 我认为这也许是 公众发现, |
| 509 | is just sort of slightly skewed discussions about what's going to happen | 只是有点曲解 讨论会发生什么 |
| 510 | and maybe there's some ulterior motives behind some of that messaging, | 也许有些信息背后有别有用心的动机 |
| 511 | raising money, other things. | 集资,其他东西。 |
| 512 | But I think we need to get to use the scientific method, | 但我认为我们需要使用科学方法 |
| 513 | be really rigorous and thoughtful about this critical moment in history. | 对历史的这一关键时刻,要非常严格和深思熟虑。 |
| 514 | And then maybe the final thing I would say is, | 然后也许我最后要说的是, |
| 515 | I would love to see, I think it's incumbent on the industry and the field | 我很想看,我觉得这是行业和领域的责任 |
| 516 | to show more unequivocally what the benefits are. | 更明确地显示好处是什么。 |
| 517 | And not just talk about them, but demonstrate them. | 不只是谈论他们,而是展示他们。 |
| 518 | So in health, in medicine, in science, | 因此,在健康, 在医学,在科学, |
| 519 | these things are all, in my view, are sort of unequivocal goods, right? | 在我看来,这些东西都是 一种明确的商品,对不对? |
| 520 | Like alpha fold. | 像阿尔法折叠。 |
| 521 | But there aren't enough examples. | 但没有足够的例子。 |
| 522 | There should be 20 alpha folds, right? | 应该有20个字母折叠 对吧? |
| 523 | And there should be, we've got to stop talking sort of in the hypothetical | 应该有,我们必须停止 说话那种假设 |
| 524 | about curing cancer and actually cure cancer, right? | 治疗癌症,实际上治疗癌症,对不对? |
| 525 | And so these are the things that I think are going to be needed | 所以这些是我认为需要的东西 |
| 526 | to demonstrate to the public why are those of us who are excited about it, | 向公众展示为什么我们中那些对此感到兴奋的人, |
| 527 | and many of us are in this room, why are we excited about this? | 我们很多人都在这间屋子里 为什么我们对此感到兴奋? |
| 528 | Why have we spent our whole life building towards this? | 为什么我们一生都是为了这个? |
| 529 | And also how are we going to concretely mitigate the risks | 我们如何具体减轻风险? |
| 530 | while enabling all of the amazing things that we would like to see | 同时也让所有我们想看到的奇妙事物 |
| 531 | and I think society needs? | 我觉得社会需要吗? |
| 532 | I think a lot of great points there that if there were some tangible benefits | 我认为很多伟大的观点 如果有一些实际的好处 |
| 533 | that were realized because of AI breakthroughs, | 因为AI的突破, |
| 534 | say to human health or drug discovery, | 说到人类健康 或药物发现, |
| 535 | that that might change people's perception in some ways | 可能在某些方面改变人们的看法 |
| 536 | and I love the suggestion of sort of trying to think farther out | 我喜欢那种建议 试图更深入地思考 |
| 537 | about a world that might look very different in terms of productivity | 关于一个在生产力上可能很不一样的世界 |
| 538 | and so forth. | 诸如来. |
| 539 | It's hard actually to do that. | 实际上很难做到这一点。 |
| 540 | Rarely in social science people can get out of the current frame they're in | 在社会科学领域,人们很少能摆脱他们目前所处的环境 |
| 541 | and actually project way forward. | 并实际预测前进的方向。 |
| 542 | I think of Keynes's great article during the Depression | 我想凯恩斯在萧条时期的精彩文章 |
| 543 | when he looks at the economic lives of our grandchildren. | 他看着我们孙子们的经济生活 |
| 544 | It's a rare case that you were saying last night. | 你昨晚说的这个案子很罕见 |
| 545 | You thought we needed another Keynes right now. | 你认为我们现在需要另一个凯恩斯。 |
| 546 | Maybe there's someone in the audience who will do that. | 也许观众中有人会这么做 |
| 547 | One of the things you've talked about for a lot of years | 一个你已经谈论了很多年的事情 |
| 548 | is the need for the frontier labs to, in a sense, regulate themselves. | 边界实验室在某种意义上需要自我调节。 |
| 549 | That is to sometimes sort of not release certain kinds of technologies | 这就是说,有时 那种不释放 某些类型的技术 |
| 550 | that might be threats to safety and so forth. | 可能威胁到安全等等。 |
| 551 | Right now it's pretty clear that the labs are just at a breakneck competition. | 现在很清楚 实验室正在一场破解比赛中 |
| 552 | They're investing everything. | 他们在投资一切 |
| 553 | They're going all out. | 他们都出去了。 |
| 554 | Do you still feel the labs ought to be self-regulating? | 你还觉得实验室应该自我管制吗? |
| 555 | Do you think the government ought to step in and regulate AI in some way? | 你觉得政府应该介入 以某种方式管理AI吗? |
| 556 | How do you see the current dynamic relative to the way you talked about it in the past? | 你如何看待现在的动态 相对于你过去谈论它的方式? |
| 557 | Well, look, first of all, just to give some historical context to this, | 好吧,听着,首先, 只是为了给一些历史背景, |
| 558 | this is not what... | 这不是什么... |
| 559 | In terms of what we talked earlier about how the technology's gone, | 关于我们刚才谈到的技术如何消失, |
| 560 | I think the technology has gone amazingly | 我觉得科技发展得惊人 |
| 561 | and maybe even on the better side of what I imagined 20 years ago. | 甚至更符合我20年前的想象 |
| 562 | But the environment it's kind of been birthing in is not the ideal. | 但它所诞生的环境并不理想。 |
| 563 | Far from it, right? | 远远没有,对不对? |
| 564 | I was very worried about 15 years ago, 10 years ago, | 我非常担心15年前,10年前, |
| 565 | about this race dynamic happening as more and more people, | 关于这个种族的动态 以越来越多的人, |
| 566 | more and more companies, more and more ambitious tech leaders | 越来越多的公司 越来越多的科技领袖 |
| 567 | realized what I had known for 20 plus years | 发现我认识20多年 |
| 568 | of how important this technology was going to be. | 这个技术有多重要 |
| 569 | And we talked about some of this in the room | 我们谈了一些 在房间里 |
| 570 | about the dangers of this kind of race dynamic. | 关于这种种族运动的危险。 |
| 571 | And unfortunately we've ended up, because of the way the technology's gone, | 不幸的是,我们结束了, 因为技术已经走了, |
| 572 | so if I could have waved a magic wand, | 如果我能挥动魔杖 |
| 573 | what I would have done was build AGI, the general technology, | 我要做的是建造AGI, 通用技术, |
| 574 | more in a research facility, perhaps like a CERN, | 更多在研究设施, 也许像CERN, |
| 575 | maybe all the best minds helping critique each other's ideas | 也许所有的最好的头脑 帮助批评对方的想法 |
| 576 | and making sure we were rigorous with the scientific method | 并且确保我们严格使用科学方法 |
| 577 | and the testing of it and understanding each step that we took. | 试验它,了解我们所采取的每一步。 |
| 578 | But then we wouldn't have to wait for that. | 但那样我们就不会等了 |
| 579 | Of course, that means AGI would arrive later, maybe 10 years later. | 当然,这意味着AGI会在十年后到达。 |
| 580 | But we wouldn't need to wait for that to get the societal benefits of it, | 但我们不需要等待 得到它的社会好处, |
| 581 | because at the same time we would break off bits of that | 因为与此同时,我们会 断掉其中的一点 |
| 582 | and use it for specialized systems like more alpha-folds, curing diseases. | 用来治疗疾病 |
| 583 | That can be done, because alpha-fold is a specialized hybrid system, | 这可以做到,因为alpha-fold是一个专门的混合系统, |
| 584 | uses a lot of the ideas the general purpose systems use, | 使用了很多通用系统使用的想法, |
| 585 | but it's specialized to protein folding. | 但它是专用于蛋白质折叠. |
| 586 | And you can see that was my vision for it, because that's what we were doing. | 你可以看到,这是我的愿景, 因为这就是我们在做什么。 |
| 587 | But then chatbots changed that because effectively, | 但后来聊天员改变了这一点 因为有效的, |
| 588 | and that was probably the only surprise to me over the last 15 years on the science side, | 这可能是我过去15年在科学方面 唯一一个惊喜 |
| 589 | is how effective Transformers ended up being for language. | 是变形器对语言的效用。 |
| 590 | And the fact that you could separate language and learn it just from the internet | 而事实上你可以分离语言 从互联网上学习 |
| 591 | without having to act in the world, either robotics or simulations. | 不需要在世界上行动, 无论是机器人还是模拟。 |
| 592 | It's kind of very interesting, and that would be a whole other topic why that was. | 这是非常有趣的, 这将是另一个话题为什么。 |
| 593 | And I have some theories on that. | 我有一些理论。 |
| 594 | Language is more grounded than linguists probably thought. | 语言比语言学家可能想到的更扎根. |
| 595 | There's some grounding coming from the reinforcement learning feedback | 强化的学习反馈中有些基础 |
| 596 | that the human testers are doing, because obviously we're grounded in the real world. | 人类测试者正在做的, 因为很明显,我们被困在现实世界。 |
| 597 | So when we say yes, no to certain things, our grounding is then ending up | 所以,当我们说是,不 对于某些事情, 我们的地面是最后 |
| 598 | in a very low bandwidth way, but still ending up modifying what the foundation model understands. | 以非常低的带宽方式,但最终仍然修改了基础模型所理解的. |
| 599 | So there was these sort of unexpected things, I would say, that happened. | 因此,发生了一些出乎意料的事情,我要说, |
| 600 | And then that made it a commercial, very important commercial technology | 然后,它成为一个商业,非常重要的商业技术 |
| 601 | that could be scaled with engineering and money and so on, which is what you see today. | 可以用工程学和金钱等等来放大, 这就是你今天看到的。 |
| 602 | And that changed the dynamic and then has created what we see today, | 这改变了动力,然后创造了我们今天所看到的, |
| 603 | which is probably the most ferocious competitive environment I would say there's ever been. | 这可能是最凶猛的竞争环境 我想说,曾经有过。 |
| 604 | I mean, certainly in the tech industry, tech era, maybe ever. | 我的意思是,肯定在技术产业,技术时代,也许永远。 |
| 605 | I may be other historians here from the business school will tell me otherwise. | 我可能是其他的历史学家 从商学院会告诉我其他。 |
| 606 | But it feels unbelievably intense being in the middle of it. | 但是在中间感觉很紧张 |
| 607 | And it feels like that for all of the participants. | 对所有参与者来说都是这样 |
| 608 | And then on top of that, you layer the geopolitical complexities. | 最重要的是,你把地缘政治的复杂性分层。 |
| 609 | So there's also a double race going on. | 故亦有双赛. |
| 610 | There's the race between the companies and it's pretty life or death for them. | 公司之间的竞争 对他们来说是漂亮的生死 |
| 611 | And then there's the U.S.-China dynamic and others, the geopolitical dynamic, and there's a race there. | 之后还有美中动态等,地缘政治动态,那里也有比赛. |
| 612 | So it's a double layered one, very tricky. | 这是一个双层的,非常棘手。 |
| 613 | Now, I still have hope that there can be some cooperation and coordination between, | 现在,我仍然希望,在以下各方面可以进行一些合作与协调: |
| 614 | we certainly discussed this as lab leaders, on the safety elements and the security elements. | 我们当然是作为实验室负责人 讨论安全因素和保安因素 |
| 615 | Everybody wants that. | 人人都想这样 |
| 616 | Nobody wants something catastrophic to go wrong. | 没人想让灾难发生 |
| 617 | The problem is we're in a kind of prisoner's dilemma where anyone who, by definition, | 问题是,我们处于一种囚犯的两难境地, 根据定义, |
| 618 | if you take more time to release something or make something safer, | 如果你花更多时间释放一些东西 或者让一些东西更安全, |
| 619 | that's harder than just putting it out there and letting it see what happens. | 这比把它放在那里 让它看看会发生什么更难 |
| 620 | So a defector has some advantage. | 所以,一个叛逃者有一些优势。 |
| 621 | And this is the classic problem with the race to the bottom dynamic. | 这是比赛到底部的经典问题。 |
| 622 | And we've got to change that somehow and I think urgently. | 我们必须改变这一点,我觉得很紧急。 |
| 623 | And I think part of that is some form of government involvement. | 我认为,其中一部分是某种形式的政府参与。 |
| 624 | The hard part there, of course, is that anything to do with regulation, it's too slow. | 当然,那里最难的是 任何与监管有关的东西 都太慢了 |
| 625 | Every week there's something new. | 每周都有新的东西 |
| 626 | If we were to regulate something two years ago, it's like ancient history now. | 如果我们在两年前管理一些东西, 现在就像古老的历史。 |
| 627 | So it would almost be, you know, almost certainly the wrong thing. | 所以这几乎是,你知道, 几乎肯定是错误的事情。 |
| 628 | So whatever is designed, and I have some ideas on this and I'll probably be talking about this later this year, | 所以不管什么设计, 我有一些想法 关于这一点,我可能会谈论 今年晚些时候, |
| 629 | is like it needs to be dynamic, which usually that word doesn't go with regulation, right? | 就像它需要动态, 通常这个词不与监管,对不对? |
| 630 | So it's got to be light, fleet-footed and able to inform by the latest developments | 因此,它必须是轻, 舰队脚和能够 了解最新的事态发展 |
| 631 | so that it can adapt to where the actual risk is rather than some kind of perceived risk | 以便适应实际风险而不是某种可察觉的风险 |
| 632 | that turns out not to be the case or not the critical thing many years before. | 事实证明,许多年前,情况并非如此。 |
| 633 | It's just not going to work for AI. | 它只是不会为AI工作。 |
| 634 | And even today, the leading scientists wouldn't necessarily agree on a short list. | 而即使在今天,领先的科学家们 也未必同意一个简短的名单。 |
| 635 | In fact, I know they definitely wouldn't agree on a short list of what checks and balances are needed. | 事实上,我知道他们肯定不会同意 列出需要什么样的制衡 |
| 636 | And that's because the science isn't settled. | 这是因为科学还没有解决。 |
| 637 | And that's partly the speed, but also the pace of the progress is running ahead of the understanding of it. | 这也是部分的速度, 同时也是进步的速度 在理解它之前。 |
| 638 | That's just how it is. | 就是这样 |
| 639 | It's part of the race dynamic. | 这是种族动态的一部分。 |
| 640 | But we need to kind of somehow rebalance that. | 但我们需要重新平衡 |
| 641 | And I think some form of really almost smart regulation is required that is dynamic | 我认为需要某种形式的 几乎是聪明的监管,这是动态的 |
| 642 | and can adapt with the times very quickly and probably informed by the leading labs | 并且可以适应时间 很快 并且可能从主要实验室得到信息 |
| 643 | because they're seeing what's actually at the coalface. | 因为他们看到什么 实际上是在煤面。 |
| 644 | All right. | 好吧,我们走。 |
| 645 | I think there's so much more to discuss there in terms of the prospects | 我觉得还有许多事情要讨论 关于前景 |
| 646 | of how you set up a regulatory system for AI and do it in a way that didn't prevent some of the breakthroughs, | 如何建立AI的监管系统 并且这样做的方式 并不妨碍一些突破, |
| 647 | positive breaks you're talking about to the geopolitics. | 你所说的对地缘政治的正面突破 |
| 648 | We want all that innovation, right? | 我们希望所有的创新,对不对? |
| 649 | We want to solve the disease. | 我们想要解决疾病。 |
| 650 | So exactly, how do you enable the good use cases and mitigate the bad? | 所以,确切地说,你是如何使 良好的使用案例和减轻坏? |
| 651 | I'm looking forward to when you bring out your plan for that this year. | 我期待你提出今年的计划 |
| 652 | I think that'll be fantastic. | 我觉得那会很棒 |
| 653 | And that'll give us a lot to talk about here on campus and everywhere. | 这会给我们很多 在校园和任何地方谈论。 |
| 654 | We've got a couple of student questions. | 我们有几个学生问题 |
| 655 | I want to give a chance for some student questions. | 我想给一些学生提问的机会. |
| 656 | Hi, Demis. | 嗨,黛米丝。 |
| 657 | I'm a second year at the business school. | 我在商学院第二年 |
| 658 | My question is, how do you balance pushing the frontier of AI | 我的问题是,如何平衡推进AI的前沿 |
| 659 | with ensuring that the health and scientific dividends | 确保健康和科学红利 |
| 660 | is evenly distributed in places like Africa and the global south where the need is the greatest, | 分布在非洲和全球南部等最需要的地方, |
| 661 | but the infrastructure for deployment and research is most limited? | 但部署和研究的基础设施却非常有限? |
| 662 | Yeah. | 对 |
| 663 | We think about that a lot, actually. | 实际上,我们想了很多。 |
| 664 | And that goes back to one of some examples of that I can give is back to the AlphaFold question | 这又回到我可以给出的一些例子之一 就是回到AlphaFold的问题 |
| 665 | where we folded all the proteins. | 我们把所有蛋白质都叠在一起 |
| 666 | We put that out on databases you could access from anywhere around the world. | 我们把这个放进数据库 你可以从世界各地访问。 |
| 667 | So these three million researchers come from 190 countries, just to be clear. | 所以这300万研究人员来自190个国家, 原因很简单。 |
| 668 | It's pretty much every country, every researcher. | 几乎每一个国家,每一个研究员。 |
| 669 | And it means that they... | 这意味着他们... |
| 670 | And what was great, what we did actually in the early days of seeding some collaborations, | 以及什么是伟大的, 我们实际上做了 在早期的种子 一些合作, |
| 671 | 你和AlphaFold有什么关系 我们和DNDI合作过 | |
| 672 | part of the WHO in Switzerland, | 卫生组织在瑞士的一部分, |
| 673 | which work on diseases in the poor places of the world that don't have good healthcare systems. | 治疗世界上没有良好医疗体系的贫困地区疾病。 |
| 674 | And some of those diseases are neglected, as you know, | 有些疾病被忽略了,你知道的, |
| 675 | because the big pharma can't make money in those markets. | 因为大药店不能在这些市场赚钱 |
| 676 | So then the diseases that affect primarily those areas of the regions of the world | 因此,主要影响世界各地区的疾病 |
| 677 | don't get as much research resources behind them. | 他们没有那么多的研究资源 |
| 678 | So what we were able to do in collaboration with this institute | 所以,我们能做的 与这个研究所合作 |
| 679 | and many actual universities on the ground is jump them straight to | 地面上的许多大学 直接跳到 |
| 680 | not needing to try to figure out malaria virus or the structures of Zika virus or something like that. | 不需要试图找出疟疾病毒 或者Zika病毒的结构之类的 |
| 681 | They would have had to do all the painstaking structural biology. | 他们不得不做所有艰难的结构生物学。 |
| 682 | They can just start that as a given and work straight away on the drugs. | 他们可以从这个开始 直接吸毒 |
| 683 | So that allows to speed up massively the whole process. | 这样就可以大大加快整个过程。 |
| 684 | They can sort of take the structure and of interest and move forwards from there. | 他们可以采取 结构和感兴趣的 从那里前进。 |
| 685 | Same with crop resilience affected by climate change. | 受气候变化影响的作物抗御能力也是如此。 |
| 686 | We work with Jennifer Doudna's Institute and many others on these things | 我们和Jennifer Doudna的研究所 和许多其他人合作 |
| 687 | because lots of plant proteins, we didn't know what the structures were of them. | 因为很多植物蛋白质, 我们不知道它们的结构是什么。 |
| 688 | Because obviously most of the structure work has gone into human proteins. | 因为很明显大部分的结构工作都已经进入了人类的蛋白质中. |
| 689 | So if it's animals or plants, there's a lot less data out there. | 因此,如果是动物或植物,那么那里的数据就少了很多. |
| 690 | So they were able to, it's even more differentially impactful in those types of areas. | 因此,他们可以, 更差别化的影响 在这些类型领域。 |
| 691 | And then the final thing I would say is like if we can, | 最后我要说的是 如果我们能的话 |
| 692 | and I think this is where the capitalist engine can actually work for good here is, | 我认为这里是 资本主义引擎可以真正工作 在这里是, |
| 693 | if we can make the drug discovery platform that we're working on isomorphic | 如果我们能做一个药物发现平台 我们的工作异形 |
| 694 | as efficient as I'm talking about, you know, down from years to months. | 和我说的效率一样,你知道的, 从几年到几个月。 |
| 695 | So instead of it costing billions of dollars, it costs tens of millions of dollars, | 因此,与其花费数十亿美元,还不如花费数千万美元, |
| 696 | maybe single millions of dollars. | 也许一百万美元。 |
| 697 | Then suddenly what I'm hoping we'll be able to do with isomorphic is, | 突然间我希望我们能处理异形症 |
| 698 | you know, we cure all these terrible diseases that maybe affect the richer parts of the world. | 你知道,我们治愈所有这些可怕的疾病 可能影响到世界富裕地区。 |
| 699 | That makes money and that fuels it, fuels the engine. | 这能赚钱,能给引擎加油 |
| 700 | But then we can do sort of philanthropically, | 但我们可以做一些慈善, |
| 701 | the company could find cures to diseases where we don't need to make any return | 公司可以找到治疗疾病的方法 我们不需要任何回报 |
| 702 | because it's fast enough and it's cheap enough that it can just be done and in a short amount of time. | 因为它的速度足够快,而且价格也足够低廉,可以仅仅在短时间内完成. |
| 703 | So I think that's sort of my dream for how I can make isomorphic help the whole world. | 所以我觉得这就是我的梦想, 我如何使异形帮助整个世界。 |
| 704 | Hi, Demis. | 嗨,黛米丝。 |
| 705 | Thank you so much for taking the time to talk to us. | 非常感谢你花时间和我们谈谈 |
| 706 | My name is Miki. | 我叫美纪 |
| 707 | I'm a senior in the Dorr School of Sustainability. | 我是多尔可持续性学院的高级学生 |
| 708 | And you've described extensively how AGI could be humanity's most transformative technologies. | 你广泛描述了AGI如何成为人类最具有变革性的技术。 |
| 709 | And I'm just curious the responsibility or how you think about the societal impacts alongside this intellectual pioneering | 我只是好奇你的责任 或你如何思考社会影响 与这个知识先锋 |
| 710 | and productivity that AGI presents, particularly when thinking about, you know, | 以及AGI带来的生产力,特别是在思考,你知道, |
| 711 | how is this going to redefine and reshape people, the challenges that we're trying to solve today, | 如何重新定义和重塑人们, 今天我们试图解决的挑战, |
| 712 | but kind of the downstream effects that that could bring forward. | 但下游可能带来的后果 |
| 713 | Thank you. | 谢谢 |
| 714 | Yeah, thanks for your question. | 是啊,谢谢你的问题。 |
| 715 | I think about this all the time and have done from the beginning because we were planning for success. | 我经常想着这一点,并且从一开始就这样做,因为我们计划取得成功。 |
| 716 | Right. | 对 |
| 717 | So even though it seemed very improbable back 15, 20 years ago. | 所以,尽管它看起来非常不可能 回到15,20年前。 |
| 718 | And I think this is why I like doing talks like this and meeting folks in in these kinds of places is I is a bit of a call to arms now. | 我想这就是我为什么喜欢做这样的演讲, 和人们在这样的地方见面, 现在我是一个有点 武装。 |
| 719 | I feel like it's very urgent that we really think about the second order consequences. | 我觉得我们真的考虑第二顺序的后果是十分紧迫的. |
| 720 | And I think many of you in the room and many of you in the humanity subjects. | 我认为你们中许多人在房间里,你们中许多人在人类主体中。 |
| 721 | It's now is your time, in my opinion, because, OK, we've got to get the technology right. | 现在轮到你了,我认为, 因为,好吧,我们必须把技术弄正确。 |
| 722 | But then if we do that, then there's the economics question. | 但是,如果我们这样做, 那么这就是经济学问题。 |
| 723 | And if we get in and if we get that right, there's the philosophical questions about the human condition. | 如果我们进去了,如果我们说对了, 就会有关于人类状况的哲学问题。 |
| 724 | And and I'm very excited about and I'm a big believe I'm very optimistic. | 我很兴奋 我非常乐观 |
| 725 | I should just very obviously I'm a cautious optimist is the way I would say it is. | 我显然应该是一个谨慎的乐观主义者。 |
| 726 | I'm very optimistic that we're going to get this right. | 我非常乐观,我们会得到这个正确的。 |
| 727 | And I'm a big believer in human ingenuity, especially when the pressure is on. | 我非常相信人类的智慧 特别是当压力大的时候 |
| 728 | I think humanity has always figured it out when when the chips are down and they are now. | 我认为当芯片掉下来的时候 人类总是会明白的 |
| 729 | But we do really need to start taking that. | 但我们确实需要开始接受它。 |
| 730 | I think the technologists are taking it seriously, but the other parts of society need to as well. | 我认为技术专家正在认真对待它,但是社会的其他部分也需要. |
| 731 | Economists, you know, I'm always a little bit astounded when I talk to economists about what's happening and it's sort of they're pretty skeptical. | 经济学家们,你知道, 我总是有点惊讶 当我跟经济学家谈论 正在发生的事情, 他们有点怀疑。 |
| 732 | Where is it? | 在甚么处. |
| 733 | Where is it coming in the GDP? | 其产值从何而来? |
| 734 | And it's like, look, it's ten times the Industrial Revolution. | 这就像,看,这是工业革命的十倍。 |
| 735 | Can we start planning for that now? | 我们能开始计划了吗? |
| 736 | Like, you know, and we'll be in a world where and we talked about this last night. | 就像,你知道,我们会在一个世界 在那里,我们谈论昨晚。 |
| 737 | I think we do need some giants of these fields like Keynes was for now. | 我认为我们确实需要一些 巨型的这些领域 像凯恩斯是现在。 |
| 738 | Like, why would that hold in a post scarcity world? | 为何在后稀缺世界中存在? |
| 739 | We're going to be in a world for the first time if we get the technology right, where we're non zero sum world for the first time in humanity's existence. | 如果我们能正确掌握技术,我们将第一次进入一个世界, 人类生存中我们第一次成为非零和世界。 |
| 740 | How can that not need a new type of economic system? | 这怎么可能不需要一种新的经济体制呢? |
| 741 | Right. | 对 |
| 742 | It has to. | 它必须。 |
| 743 | And I don't think it's any of the ones we've tried because they were all done under the guise of zero sum more, you know, and a limited a scarce world. | 我不认为我们尝试过的东西 因为它们都是在零和的幌子下完成的, 你知道,一个有限的稀缺世界。 |
| 744 | And I'm talking about, you know, traveling to the stars and utilizing all the all the resources that are out in the solar system, not just the limited ones on earth. | 我说的是,你知道, 前往恒星,利用太阳系中所有的资源, 不仅仅是地球上有限的资源。 |
| 745 | And I think that really is going to happen if we get the technology right in the next 10, 20, 30 years. | 我认为,如果我们在未来10、20、30年里 掌握正确的技术,那真的会发生。 |
| 746 | And then after all of that, there's the even harder question of how do we want to evolve our society and what is virtuous, what is meaning, what is purpose? | 毕竟,还有一个更难的问题,就是我们想怎样进化我们的社会,什么是良性,什么是目的? |
| 747 | And I think that's going to need lots of great philosophers. | 我认为这需要很多伟大的哲学家。 |
| 748 | So that would be my appeal to to people in those fields is now, in my view, could not be more of an exciting time if you're working on those types of projects. | 因此,这是我对这些领域的人们发出的呼吁, 我认为,如果你在研究这些类型的项目, 这将是更令人兴奋的时刻。 |
| 749 | As long as you understand what's and really viscerally sort of understand and lean into what's actually happening here. | 只要你能理解什么是真实的, 并真正地理解, 并倾斜到这里实际发生的事情。 |
| 750 | It's a good good charge to university. | 大学收费不错. |
| 751 | One more student question. | 又一弟子问. |
| 752 | Hi, Dennis. | 嗨,丹尼斯,你好吗? |
| 753 | I'm Janai. | 我是贾奈. |
| 754 | I'm a second year MBA student. | 我是二年级MBA学生. |
| 755 | My question to you is, what do you not want A.I. | 我问你,你想要什么? |
| 756 | to touch in this lifetime? | 来接触这一辈子? |
| 757 | And what do you hold secret from your perspective? | 从你的角度来说,你有什么秘密? |
| 758 | Thanks. | 谢谢 |
| 759 | Yeah, that's a great question. | 是啊,这是一个伟大的问题。 |
| 760 | 听着,A. I. | |
| 761 | is going to be in terms of the scientific world of things. | 将是科学世界的东西。 |
| 762 | It's a fully general technology, right? | 这是一个完全通用的技术,对不对? |
| 763 | You can think of it as a Turing machine. | 你可以把它当成图灵机 |
| 764 | That's the way I think about it is my favorite course at college. | 这是我在大学里最喜欢的课 |
| 765 | And I think our minds are actually fully general. | 我认为我们的思想是完全一般的。 |
| 766 | So we're kind of approximate Turing machine. | 所以我们大概是图灵机 |
| 767 | So as Turing showed, we anything that's computable, a Turing machine can compute. | 所以,图灵显示,我们任何可以计算的东西, 一个图灵机器可以计算。 |
| 768 | And most things we know about in the universe, non-quantum things are computable. | 在宇宙中我们所知道的大多数事情,非量子的事情都是可以计算出来的。 |
| 769 | So that's a pretty large set of general things that we can turn our minds to. | 因此,这是一个相当大 一组一般性的事情,我们可以改变我们的想法。 |
| 770 | Hence, we built modern civilization, which is miraculous if we stop to think about it. | 因此,我们建立了现代文明,如果我们停止思考,这是奇迹。 |
| 771 | And I don't think we wonder enough. | 我觉得我们不够奇怪 |
| 772 | We keep our sense, don't keep our sense of wonder for long enough about that. | 我们保持我们的感知, 不要让我们的好奇感 足够长的时间。 |
| 773 | But it also means these systems that we're building, they're also going to be sort of Turing powerful as well. | 但也意味着我们正在建造的这些系统, 它们也将是图灵强大的。 |
| 774 | One thing I would say is that there are very big questions to come that I think it would be better if we took more time over. | 我想说的是,有些大问题要问,我认为,如果我们花更多的时间来讨论,会更好。 |
| 775 | So one example that's pretty topical right now is consciousness. | 因此,现在相当热门的一个例子是意识。 |
| 776 | And, you know, it's not a very well posed problem from philosophy and neuroscience still. | 而且,你知道,这不是一个很好的问题 从哲学和神经科学仍然。 |
| 777 | Although I think we all have intuitions as to what are the important aspects of that. | 尽管我认为我们大家都有直觉,知道其中的重要方面是什么。 |
| 778 | My feeling is the current systems don't exhibit any, are not. | 我的感觉是,目前的系统 没有显示任何,不是。 |
| 779 | But others disagree. | 但其他人不同意这一点。 |
| 780 | What I would recommend, though, in terms of what areas should AI not touch is that we build our first systems as tools, intelligent tools. | 然而,我想建议,在AI不应触及的领域,我们应建立我们的第一个系统,作为工具、智慧工具。 |
| 781 | That's enough of a challenge already, in my opinion, because that's already AGI. | 这已经足够挑战了,我认为, 因为这已经是AGI。 |
| 782 | And then using those tools, I think we should study neuroscience and other things like that. | 然后利用这些工具, 我认为我们应该研究神经科学和其它类似的东西。 |
| 783 | And philosophy and actually come up with a more rigorous definition of things like consciousness. | 而哲学和实际上提出了更严格的意识等事物的定义. |
| 784 | I think that is possible and then test things against that. | 我认为,这是可能的,然后检验事情与此相对。 |
| 785 | And then maybe society decide if we want to cross the second Rubicon of trying to make entities that at least seem like conscious to us. | 也许社会会决定我们是否想跨越第二个Rubicon, 试图使那些至少在我们看来似乎有意识的实体。 |
| 786 | So we may not want to make that decision. | 因此,我们可能不想作出这一决定。 |
| 787 | I could write. | 我可以写作。 |
| 788 | I think that intelligence and consciousness are dissociable. | 我认为智力和意识是不可分离的. |
| 789 | So I don't think you have to do that to have an intelligent system. | 所以我不认为你必须这么做 才能有智能系统。 |
| 790 | I think it's a choice. | 我认为这是一个选择。 |
| 791 | Some of our. | 一些我们。 |
| 792 | And so you can probably feel that in the when you use some of the leading chatbots, there's differences in opinion that come through. | 所以,你也许可以感觉到,当你使用一些主要的聊天机器人时, 会有不同的看法。 |
| 793 | And my view is it'd be better to take that as two steps. | 我的看法是,最好把这作为两个步骤。 |
| 794 | They're both enormous for humanity rather than conflate the two. | 它们对人类来说都是巨大的,而不是将两者混为一谈. |
| 795 | Demis, we've got an auditorium with many students in it. | 戴米斯,我们有一个礼堂 里面有很多学生 |
| 796 | If you were back in school, how would you be thinking about what would you be studying? | 如果你回到学校,你会怎么想 你会学习什么? |
| 797 | How would you think about what? | 你会怎么想? |
| 798 | What should they what would be your advice on how they should be thinking about to study in their careers? | 他们应该有什么建议,你应该如何考虑在他们的职业生涯中学习? |
| 799 | Well, look, I would I would be really excited if I was back at college now. | 如果我现在回到大学 我会很兴奋的 |
| 800 | My recommendation would be those of you doing science and STEM subjects and mathematics and computer science still do those things. | 我的建议是你们做科学和STEM课程 数学和计算机科学仍然做这些事。 |
| 801 | I think you'll be able to take better advantage of these tools if you understand how they are put together and what they're capable of. | 如果你能理解这些工具是如何组合起来的,以及它们的能力,你就能更好地利用这些工具。 |
| 802 | I think that's going to be true for the next the next period, the next 10 years at least. | 我认为这在下一个时期,至少在未来10年里是真实的. |
| 803 | And I would also lean in, though, to not wish it away. | 我也希望不要放弃 |
| 804 | The genie is not going back in the bottle like like lean into what these tools can do. | 精灵不会回到瓶子里 像倾斜到这些工具能做的一样 |
| 805 | I can tell you that the leading labs are so busy making the tools that we have not. | 我可以告诉你,主要的实验室 忙于制造我们没有的工具。 |
| 806 | We probably scratch the surface of not even probably of what they can actually do. | 我们可能刮掉表面 甚至可能没有 他们能做什么。 |
| 807 | Even today's tools have this. | 即使是今天的工具都有这个。 |
| 808 | Sometimes people call it capability overhang. | 有时人们称之为能力悬浮。 |
| 809 | There's so much potential these things can do if you figure out how to pair them with other things or compare it with another domain you're expert in, build it into your workflow in an interesting way. | 如果你想出方法 把它们与其他东西配对起来 或者把它与你所擅长的另一个领域比较 用有趣的方式把它建进你的工作流程中 |
| 810 | You have those tools. | 你有这些工具。 |
| 811 | They're the most powerful tools anyone's got. | 他们是最强大的工具 任何人都有。 |
| 812 | You have them in the palm of your hand. | 你手里拿着呢 |
| 813 | There's so much more you can do as an individual. | 作为个人你还能做很多事 |
| 814 | I think it should unleash creativity like those of you studying humanities or or product or business. | 我认为它应该释放出创造力 就像你们那些研究人文或产品或商业。 |
| 815 | Maybe you didn't have coding skills before, but you don't need you don't you can produce a lot of what's in your mind now using these tools. | 也许你以前没有编码技能,但是你不需要你,你现在用这些工具不能产生很多你心中的东西. |
| 816 | Right. | 对 |
| 817 | And I but I think also the coders, the people who are expert that could do a hundred X more like the terms of the size of project you can do if you're expert coding. | 但我也认为,那些编码者, 那些专家,他们可以做一百X 更多的条件 的项目大小 如果你是专家编码。 |
| 818 | So I think it enables both the democratization and the people that are specialized in those areas. | 因此,我认为它既使民主化,也使专门从事这些领域工作的人民得以实现。 |
| 819 | So I think it's an amazing time, but it's also I get it's also worrying because everything's going to change. | 所以我觉得这是一个令人惊奇的时刻,但这也是我得到它也担心,因为一切都会改变. |
| 820 | So that's the only thing I can tell you for sure. | 所以,这就是我唯一可以肯定的事情。 |
| 821 | Everything is going to change in the next 10 years, probably more than people assume. | 未来十年一切都会改变, 可能比人们想象的要多。 |
| 822 | But that also in any time there's enormous change like that. | 但是,在任何时间都有如此巨大的变化。 |
| 823 | There's enormous opportunities. | 有巨大的机会。 |
| 824 | There has to be. | 必须有。 |
| 825 | Right. | 对 |
| 826 | And the world sort of your oyster really. | 这世界真是你的牡蛎 |
| 827 | And I kind of envy some of you now because you're the first generation that will be a native just like my generation was computer and Internet native. | 我有点羡慕你们中的一些人 因为你们是第一代人 就像我这一代人一样是本地人 电脑和网络本地人 |
| 828 | And it's going to be in your hands in the end. | 而终将归汝手. |
| 829 | The students in the room like how that future world gets built. | 房间里的学生喜欢 未来世界是如何建造的。 |
| 830 | And I think it's a very exciting time if you think about it in the right way from the right angle and with a lot of imagination and creativity. | 我认为这是一个非常激动人心的时代,如果你从正确的角度,用很多的想象力和创造力来思考这个问题。 |
| 831 | But I think that's always been true. | 但我认为这一直是事实。 |
| 832 | And maybe it's more so now in changes in periods of enormous change like this that accentuates it. | 也许现在在如此巨大的变化时期 变化会更强烈 |
| 833 | We were saying last night that in a period of a lot of change where you don't quite know what the future holds, but you have to be able to be adaptable and have broad domain of knowledge. | 我们昨晚说,在很多变化中, 你并不完全知道未来会是什么, 但你必须能够适应,拥有广泛的知识领域。 |
| 834 | It's going to be a golden era for liberal education. | 这将是自由主义教育的黄金时代. |
| 835 | So I think I mean, the main thing is to just make sure you double down your own agency, the future still to be written. | 所以,我想我的意思是,主要的事情是 只是确保你翻番 你自己的机构, 未来仍然有待写。 |
| 836 | I would say that's so don't listen to anyone who says it's not. | 我会说,所以不要听 任何人谁说它不是。 |
| 837 | Thomas, thank you for joining us. | 托马斯 谢谢你加入我们 |
| 838 | It's amazing. | 不可思议. |
该视频共有字幕 838 条。解锁更多字幕为会员功能，请移动到 价格