从「制」造到「智」造，AI碰撞传统行业将催生东西方怎样的生产
字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:04 | We are very fortunate today to have six distinguished Yale and Stanford alumni with us. | 今天我们非常幸运 有六位杰出的耶鲁和斯坦福校友出席。 |
| 0:11 | They come from business, academia, entrepreneurship, so we will not run out of perspectives tonight. | 他们来自商业,学术界,企业家精神,所以我们今晚不会失去视野. |
| 0:19 | Our panelists are Lin Jie, CEO of Xinhe Group. | 我们的小组成员是辛亥集团CEO林杰. |
| 0:25 | Welcome. | 欢迎 |
| 0:30 | Yin Hang, Professor of Pharmaceutical Sciences at Tsinghua University. | 殷杭,清华大学药学教授. |
| 0:35 | Welcome. | 欢迎 |
| 0:38 | Our panelists online include Martin Ma, founder and CEO of Happy Universe. | 我们的在线小组成员包括快乐宇宙的创始人兼CEO马丁·马. |
| 0:48 | Marshall Van Austin, Allen and Kelly question professor in information systems at Boston University. | 马歇尔·范·奥斯汀(Marshall Van Austin),艾伦和凯利(Kelly)在波士顿大学问信息系统教授. |
| 0:57 | Welcome. | 欢迎 |
| 1:00 | Zhang Lu, founder and managing director of Fusion Fund and our panel moderator Jia Peilin, partner of ORH Partners. | 张鲁,Fusion Fund的创始人兼总经理,以及我们的小组主持人,ORH Partners的合伙人贾培林. |
展开字幕全文（1610 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | We are very fortunate today to have six distinguished Yale and Stanford alumni with us. | 今天我们非常幸运 有六位杰出的耶鲁和斯坦福校友出席。 |
| 2 | They come from business, academia, entrepreneurship, so we will not run out of perspectives tonight. | 他们来自商业,学术界,企业家精神,所以我们今晚不会失去视野. |
| 3 | Our panelists are Lin Jie, CEO of Xinhe Group. | 我们的小组成员是辛亥集团CEO林杰. |
| 4 | Welcome. | 欢迎 |
| 5 | Yin Hang, Professor of Pharmaceutical Sciences at Tsinghua University. | 殷杭,清华大学药学教授. |
| 6 | Welcome. | 欢迎 |
| 7 | Our panelists online include Martin Ma, founder and CEO of Happy Universe. | 我们的在线小组成员包括快乐宇宙的创始人兼CEO马丁·马. |
| 8 | Marshall Van Austin, Allen and Kelly question professor in information systems at Boston University. | 马歇尔·范·奥斯汀(Marshall Van Austin),艾伦和凯利(Kelly)在波士顿大学问信息系统教授. |
| 9 | Welcome. | 欢迎 |
| 10 | Zhang Lu, founder and managing director of Fusion Fund and our panel moderator Jia Peilin, partner of ORH Partners. | 张鲁,Fusion Fund的创始人兼总经理,以及我们的小组主持人,ORH Partners的合伙人贾培林. |
| 11 | We look forward to their insights and debates and likely disagreements, making this event a learning experience for everyone. | 我们期待着他们的见解和辩论以及可能的分歧,使这次活动成为每个人的学习经验。 |
| 12 | Finally, it is my great pleasure to turn the stage over to April Hu, the chair of the Yale International Alliance. | 最后,我非常高兴地将舞台翻到耶鲁国际联盟主席胡四月. |
| 13 | Hello, I'm April Sonderhu and I am the chair of Yale International Alliance, YIA. | 你好,我是艾普丽尔·桑德胡 我是耶鲁国际联盟主席,YIA |
| 14 | Our mission is to connect international Yale alumni and friends to understand each other better, to collectively learn, to make impact globally. | 我们的任务是把国际耶鲁大学的校友和朋友联系起来,以便更好地理解彼此,共同学习,在全球产生影响。 |
| 15 | As co-creators with Peilin Sun, we have curated a series of webinars called AI and the Global Citizen to showcase solutions on pressing world matters. | 作为与Peilin Sun共同创作者,我们主持了一系列名为AI和全球公民的网络研讨会,展示关于紧迫世界事务的解决办法。 |
| 16 | As a Chinese American Yale engineer, I am so excited to see so many of you on site in Beijing and also calling in by Zoom from all around the world. | 作为华裔美国人 耶鲁工程师,我很高兴看到你们这么多人 在北京现场,还有Zoom从世界各地打电话来 |
| 17 | YIA, Yale International Alliance, actually would invite other Yale organizations, including Yale clubs from different cities, different countries, different continents to collaborate together. | 耶鲁国际联盟(YIA),耶鲁国际联盟(Yale International Alliance)实际上会邀请其他耶鲁组织,包括来自不同城市,不同国家,不同大陆的耶鲁俱乐部一起合作. |
| 18 | Because this is what we're wanting to do and I thank you for joining this hybrid program bridging East and West that is actually quite timely. | 因为这是我们想要做的,我感谢你加入这个 连接东西方的混合方案, 这其实是非常及时的。 |
| 19 | So right now I encourage you all to feel curious and with awe to, with the feeling perhaps how you felt when you first walked into your first class at Yale or at your own education institution. | 所以,现在我鼓励大家感到好奇, 并怀着敬畏之心, 带着你第一次走进耶鲁大学或自己的教育机构时的感觉。 |
| 20 | And therefore right now I hand over to Jia Peilin who will be moderating our panel of experts from China and America. | 因此,我现在将交给贾佩林,他将主持我们来自中国和美国的专家小组。 |
| 21 | So and online here there's Marshall coming from Boston. | 网上有马歇尔从波士顿来 |
| 22 | Martin, I don't know who you're calling in from. | 马丁 我不知道你是谁打来的 |
| 23 | And there's also Lu Zhang and she's calling in from early, early morning from California. | 还有吕章,她从清晨从加州打来电话 |
| 24 | So thank you very much for joining us. | 非常感谢你加入我们。 |
| 25 | So over to you Jia Peilin. | 给贾培林 |
| 26 | Thank you. | 谢谢 |
| 27 | Okay, thank you, April. | 好的,谢谢,艾波莉 |
| 28 | Thank you very much for joining us, everyone. | 非常感谢你们加入我们 各位 |
| 29 | I'm honored to be moderating this event. | 我很荣幸能主持这次活动 |
| 30 | On site here with Professor Ying and Mr. Jie and online with Marshall, Professor Marshall Van Elsten and Martin Ma and Lu Zhang from Silicon Valley. | 在现场与英教授和杰先生,和马歇尔,马歇尔·范埃尔斯滕教授,马丁·马和来自硅谷的吕章一起上网. |
| 31 | So let's now start with our first question. | 现在让我们从第一个问题开始。 |
| 32 | Can our guests please share some of your experience with us? | 我们的客人能和我们分享一下你的经验吗? |
| 33 | How AI is transforming the industries you're covering and what new technology and new business models do you see in these industries? | AI正在如何改变你所覆盖的行业 以及你看到的这些行业中的新技术和新的商业模式? |
| 34 | Jie, maybe you can share the retail or manufacturing industry. | 小杰,也许你可以分享零售业或制造业. |
| 35 | Yeah, I'll talk a little bit about that. | 是的,我会说一点点。 |
| 36 | Well, I've been working in the retail industry for the past decades and covering both online and offline. | 我过去几十年都在零售业工作 覆盖在线和线下 |
| 37 | And we cover both the channel side and also the brand company side. | 我们既覆盖频道方面,也覆盖品牌公司方面。 |
| 38 | And I've been through what we call the AI transformation like twice already. | 我已经经历了我们所说的人工智能转变 就像已经两次。 |
| 39 | The first time is somewhere between 2015, just around that time. | 第一次是在2015年之间,就在那时左右。 |
| 40 | So computer vision is pretty hard. | 所以计算机视觉很困难。 |
| 41 | So that's where we have, first we have cameras and then we use cameras to identify objects and then human beings and also facial recognition. | 这就是我们拥有的地方, 首先我们有摄像机,然后我们使用摄像机来识别物体, 然后是人,还有面部识别。 |
| 42 | And then for some reason that the latest transformation comes from the AIGC, what we know about the large model and the AI agents. | 之后出于某种原因,最新的转变来自于AIGC,我们所了解的大型模型和AI代理. |
| 43 | So in terms of AI technology, really it's new. | 所以在AI技术方面,其实是新的. |
| 44 | There isn't one technology. | 没有一个技术。 |
| 45 | We are talking about a bunch of new cutting edge technology at different times. | 我们谈论的是不同时期的一系列新的尖端技术。 |
| 46 | Just for the time's sake, just share some interesting stories that we, as an active practitioner of technology, tell you some interesting stories. | 只是为了时间,只是分享一些有趣的故事, 我们作为一个积极的技术实践者,告诉你一些有趣的故事。 |
| 47 | And then I will start with my takeaway about how to adopt and apply AI technology for enterprise, mostly in retail business. | 然后,我将开始我的外卖 关于如何将AI技术 应用到企业, 大部分在零售业。 |
| 48 | I wasn't very familiar with biomed or healthcare or other industries. | 我不太熟悉生物医学 医疗或其他行业 |
| 49 | I'll start with my takeaway, which is there's a common mistake of using technology in general, let alone AI technology. | 我先从我的外卖开始, 这是一种常见的错误 使用一般的技术, 更不要说AI技术。 |
| 50 | That is, you mixed up ends with your means. | 也就是说,你把手段混为一谈 |
| 51 | And then oftentimes technology is not the end. | 技术往往不是终点。 |
| 52 | It's the means to get you to the end. | 令诸众生得解脱门. |
| 53 | But because of a common FOMO, fear of missing out, a lot of the enterprises here in China rush into adopting cutting edge technologies | 但是由于一个常见的FOMO,害怕被忽略,许多在中国的企业都急于采用尖端技术. |
| 54 | with the fear of if I don't do this, if my company doesn't do this, then my company will be wiped out. | 害怕如果我不这样做, 如果我的公司不这样做, 那么我的公司就会被消灭。 |
| 55 | I've seen that case many times with myself, with my clients, with companies that we work with. | 我和我自己 和我的客户 和我们合作的公司都见过这个案子 |
| 56 | And then as long as you hold that mentality, I've never seen a successful case where you make technology or apply AI technology per se as your end, as your goal. | 然后,只要你持有这种心态, 我从来没有见过一个成功的例子, 即你制造技术或应用AI技术本身 作为你的末日,作为你的目标。 |
| 57 | But in fact, it is only a tool. | 但事实上,它只是一个工具。 |
| 58 | It is only a means to get you to your end. | 这只是一个手段 让你的结局。 |
| 59 | You have to know what you want as a CEO of the company. | 你要知道作为公司的首席执行官你想要什么 |
| 60 | You really have to know what you want to use about this technology to get what you want. | 你一定要知道你想用什么技术来获得你想要的东西。 |
| 61 | If you really don't have the application, you don't have to use it. | 如果你真的没有应用程序,你不必使用. |
| 62 | So I will start out with the first funny, or the real case, currently happening in my own company, the latest company, which we are a FMCG company. | 所以,我先从第一个有趣的,或者真正的案例开始, 发生在我自己的公司,最新的公司,我们是一家FMCG公司。 |
| 63 | Fast-moving consumer goods. | 快速移动消费品. |
| 64 | We sell condiments, we produce vinegars, we produce soy sauce. | 我们销售调味料,生产醋,生产酱油. |
| 65 | So we have manufacturing factories, and then we have sales and marketing organizations to sell our products, to place them into supermarkets. | 所以我们有制造厂,然后我们有销售和营销组织 来销售我们的产品,把它们放到超市。 |
| 66 | So it's a very typical FMCG company. | 因此是一家非常典型的FMCG公司. |
| 67 | And then for some reason, this AIGC, or AIGBT, and then etc., excites. | 然后出于某种原因,这个AIGC,或者AIGBT,然后是兴奋. |
| 68 | And then the founder, all right, yeah, well, see, that's where technology doesn't work, right? | 然后创始人,好吧,好吧,你看, 这就是技术不起作用的地方,对不对? |
| 69 | So, you know, probably smart enough to know that I'm going to say something negative, so it cuts the connection. | 所以,你知道,也许足够聪明, 知道我会说一些负面的话, 所以它会切断连接。 |
| 70 | That's very smart. | 这很聪明 |
| 71 | So he wants to know what we can do. | 所以他想知道我们能做什么 |
| 72 | Well, honestly, I told him there's really not much we can in terms of using the large model and everything. | 老实说,我告诉他,在使用大型模型和一切方面,我们实在是无能为力。 |
| 73 | But then our technology team, the head of the technology, told me that, you know, you can't say no to Boss, right? | 但后来我们的科技团队,科技负责人,告诉我,你知道,你不能拒绝老板,对不对? |
| 74 | Boss wants to see, can you apply a large model to our company to make our life better? | 老板想看看 你能给我们的公司应用一个大模型 来改善我们的生活吗? |
| 75 | And then he ended up using a large model version of it locally, deployed in our server, and then used that as a tool to do web crawling. | 最后他在当地使用一个大型的模型版本, 部署在我们服务器上, 然后用它做网络爬行的工具。 |
| 76 | To do web, basically collecting data from the web, which you can honestly just use a web crawler to do it very efficiently, right? | 做网络,基本上从网络中收集数据, 你可以诚实地使用一个网络爬行器来非常高效地做,对吧? |
| 77 | Just write a Python, like 10 lines of code, and you could have done it, right? | 只要写一个Python, 像10行代码, 你可以做到这一点,对不对? |
| 78 | But then our tech team, yeah, think about, oh, let's do, let's make an AI agent. | 但是,然后我们的技术团队, 是的,思考, 哦,让我们做一个AI代理。 |
| 79 | And then we tell them, we talk to the AI agent and say, hey, I want some data from what year to what year can you get that for me from online? | 然后我们告诉他们,我们和人工智能特工谈 说,嘿,我要一些数据 从哪一年到哪一年 你能从网上给我吗? |
| 80 | And then this AI agent will say, okay, I'm going to analyze this, and I'm going to get this data for you. | 然后这个人工智能特工会说,好吧,我要分析一下这个, 我会得到这个数据给你。 |
| 81 | This honestly is funny, but to me it's absurd. | 老实说,这很有趣, 但对我来说,这是荒谬的。 |
| 82 | I mean, it is a typical example of you don't really know what you want to do, and then you just harshly push the application of the technology, which isn't the right technology. | 我的意思是,这是一个典型的例子, 你并不真正知道你想做什么, 然后你只是严厉地推动技术的应用, 这不是正确的技术。 |
| 83 | And another example was about 10 years ago. | 另一个例子是十年前。 |
| 84 | If you still remember, at that time in China, all the shopping malls went through a renovation. | 如果你还记得,当时在中国,所有的购物中心都经历了翻新. |
| 85 | They installed a lot of cameras in the shopping mall. | 他们在购物中心安装了许多摄像头. |
| 86 | The reason for that is they wanted to know how many people coming into my shopping mall, and then they do a statistic count. | 原因是他们想知道有多少人 进入我的购物中心, 然后他们做了一个统计统计。 |
| 87 | They wanted to know how many people go into this store, and how many people go into that store, and how long they have stayed in the store. | 他们想知道有多少人走进这家商店,有多少人走进那家商店,以及他们在商店呆了多久. |
| 88 | And then they asked me to help them to install that system, many of the shopping malls, almost everyone that you know about it. | 然后他们要求我帮助他们安装这个系统, 许多购物中心,几乎每个人都知道。 |
| 89 | And then I told them, I said, why do you want this, right? | 然后我告诉他们, 我说, 你为什么想要这个,对不对? |
| 90 | I said, oh, because I was from JD.com, I worked for Jingdong before, right? | 我说,哦,因为我来自JD.com, 我以前在京东工作过,对不对? |
| 91 | So he said, well, in the online world, you guys have all the data, all the traffic, all the UVs, right? | 他说,在网络世界, 你们拥有所有的数据, 所有的流量,所有的紫外线,对不对? |
| 92 | PVs and UVs and all the conversion rates, etc. | 光伏和紫外线以及所有换算率等. |
| 93 | Well, offline, we need to know the same thing. | 好了,离线,我们需要知道 同样的事情。 |
| 94 | Okay, that makes sense, but then what? | 好吧,这是有道理的,但然后呢? |
| 95 | And then what do you want to use about it? | 那你想用什么来着? |
| 96 | As the owner of my store, you charge me for the service, right? | 作为我店主,你付我服务费,对吧? |
| 97 | And you tell me how many people come into my store. | 你告诉我有多少人来我的店里 |
| 98 | Is it because of your technology that increases the visitors to my store? | 是因为你的技术 才增加了我店里的游客? |
| 99 | Or is it because my location is a premium location in the mall, therefore I have all these visits, right? | 还是因为我的位置是商场的贵重地点 所以我有这么多访问,对不对? |
| 100 | I mean, it's like means and ends. | 我的意思是,它就像手段和结束。 |
| 101 | It's not because you apply this technology so that I have more visitors into my store. | 不是因为你应用了这个技术 让我有更多人来我的店里 |
| 102 | It is because I rented, I paid the premium rent on the first floor next to the entrance of the shopping mall. | 是因为我租了,我付了购物中心入口旁边一楼的保费租金。 |
| 103 | And because my brand, for example, I'm a Starbucks at that time, or you are some pretty good brand. | 因为我的品牌,比如说,我是一个星巴克 在当时,或者你是一个相当不错的品牌。 |
| 104 | That's why that I have all this traffic coming in. | 所以我才有这么多人来 |
| 105 | But at that time, because of digitization, right? | 但当时,因为数字化,对不对? |
| 106 | Because Wang Jianlin basically said everyone that shopping mall needs to be digitized. | 因为王建林基本上都说购物中心需要数字化. |
| 107 | And then people are asking what does that even mean, right? | 然后人们问这是什么意思,对不对? |
| 108 | So that means that we burn a lot of money. | 这意味着我们烧了很多钱。 |
| 109 | Because chasing for the technology, China, my colleagues, my friends, people in this industry, | 因为追求技术,中国,我的同事,我的朋友, 在这个行业的人, |
| 110 | we literally relied on everything and we wasted a lot of money. | 我们完全依赖一切 我们浪费了很多钱 |
| 111 | Now, there are successful applications. | 现在,有成功的应用。 |
| 112 | I will start with my own company, yes, it's an FMCG company. | 我先从我自己的公司开始,是的,这是一家FMCG公司. |
| 113 | But there are a lot of automation technologies that have already been adopted, like robotics, right? | 但有很多自动化技术已经被采用,比如机器人技术,对吧? |
| 114 | So, yeah, robotics is a technology, is it AI? | 所以,是的,机器人是一种技术,是AI吗? |
| 115 | I mean, it's part of, right? | 我的意思是,这是一部分,对不对? |
| 116 | But it's long been used in the industry, the robotic arms. | 但它早已在工业中使用,机器人臂. |
| 117 | And then we also use computer, we also install cameras in our production line | 然后我们用电脑,在生产线安装摄像头 |
| 118 | to replace people to identify the defects of the products. | 替换人员以识别产品的缺陷。 |
| 119 | And then computer works far better and faster than human being eyes. | 然后计算机比人类的眼睛 工作得更好更快 |
| 120 | We used to have, for every production line, every 10 meters, | 我们以前每条生产线 每10米 |
| 121 | we have one worker standing there and looking at the bottle | 我们有一个工人站在那里,看着瓶子 |
| 122 | to see whether or not there's a defect on the product, on the packaging, basically. | 看产品,包装上是否有缺陷,基本上. |
| 123 | But then, yeah, that was replaced by computer, which is a lot faster. | 但后来,是的,它被电脑取代了, 它的速度要快得多。 |
| 124 | So there are very successful, and there are HEVs, right? | 因此,有非常成功的, 有HEVs,对不对? |
| 125 | When we do logistics, you see in our warehouse, | 当我们做后勤时,你看我们的仓库, |
| 126 | we do have these HEVs carrying cargoes and automatically stocking into the warehouses. | 我们确实有这些HEV 载货和自动储存进仓库。 |
| 127 | All of those technologies that are actually in place have been here for decades. | 所有那些实际上已经到位的技术都在这里已经几十年了。 |
| 128 | It honestly has nothing to do with the latest AI agent or large model, | 老实说,这和最新的AI代理或大模型无关, |
| 129 | but it has been in place. | 但它已经到位。 |
| 130 | And then that's useful, because that's because you have a goal. | 这很有用,因为你有个目标 |
| 131 | You wanted to do something. | 你想做点什么 |
| 132 | You want to realize something to improve your efficiency, | 你想意识到一些东西来提高你的效率, |
| 133 | and then you look for the right technology to apply and then get what you want. | 然后你寻找合适的技术 应用,然后得到你想要的。 |
| 134 | So that's what I meant by you have to meet up your ends with your means. | 所以这就是我的意思,你必须 满足你的末端 用你的手段。 |
| 135 | You can't not revert it around the other way. | 你不能掉头回去 |
| 136 | So I'll just pretty much stop here, and there are a lot of stories to tell, | 所以我就停在这里 有很多故事要说 |
| 137 | and we are more excited to hear more stories from other panelists. | 我们更兴奋地听到更多 来自其他小组成员的故事。 |
| 138 | Thank you very much. | 谢谢 |
| 139 | Professor Yin, would you share some of your perspectives? | 殷教授,你能谈谈你的观点吗? |
| 140 | Sure. | 当然 |
| 141 | 谢 you,裴林. | |
| 142 | So hello, everyone. | 大家好,你们好 |
| 143 | My name is Hang Hubert Yin. | 我叫杭休伯特・殷 |
| 144 | I go by Hubert. | 我从休伯特。 |
| 145 | Actually, Hubert is really my middle name. | 其实 休伯特是我的中名 |
| 146 | I adopted that when I was naturalized as a U.S. citizen. | 当我成为美国公民时,我接受了这个。 |
| 147 | So when I was in the U.S., people always laughed at me. | 所以当我在美国的时候 人们总是嘲笑我 |
| 148 | Hubert, that sounds like someone from the 60s or something. | 休伯特 这听起来像是60年代的某个人 |
| 149 | It's a long story anyway. | 说来话长 |
| 150 | But I want to thank Helen again, our excellent, fearless leader, | 但我想再次感谢海伦 我们出色的无畏领袖 |
| 151 | the president of Yale, Beijing. | 北京耶鲁大学校长. |
| 152 | Yale is always something that really has this whole sweet spot in my heart. | 耶鲁总是有 真正有 整个甜点在我心中。 |
| 153 | So just like Dr. Lin, I went to New Haven also back in 1999. | 和林医生一样 我1999年也去了纽黑文 |
| 154 | I think that's sometimes when many of you guys sitting in this room | 我想有时候你们这些家伙 会坐在这房间里 |
| 155 | were still not born yet. | 还没出生呢 |
| 156 | But anyway, so before that, I graduated from Beijing Peking University. | 但无论如何,所以在此之前,我毕业于北京大学. |
| 157 | We called it Peking University. | 我们称之为北京大学。 |
| 158 | And then after Yale, I went to Philadelphia for postdoc training | 耶鲁大学毕业后 我去费城接受博士后训练 |
| 159 | at UPenn Medical School. | 在乌本医学院 |
| 160 | And then after teaching 12 years at the University of Colorado Boulder, | 然后在科罗拉多波尔德大学教了12年 |
| 161 | I moved to Tsinghua. | 迁至清华. |
| 162 | That is a no-no for Peking alumnus. | 这是北京校友的禁言 |
| 163 | So my friend actually always asked me, how come you went to Tsinghua? | 我朋友总是问我 你怎么去清华? |
| 164 | I always told them, because Tsinghua is still just about a street away | 我总是告诉他们,因为清华还在街上 |
| 165 | from the world-class university. | 从世界一流的大学。 |
| 166 | So actually before this panel, I was like, | 所以,实际上在这个面板之前,我就像, |
| 167 | Penny and I would talk about, because she has this thing | 我和Penny会说 因为她有这个 |
| 168 | to move to Harvard after Yale. | 在耶鲁大学毕业后移居哈佛. |
| 169 | That is something we can't explain. | 这是我们无法解释的事情。 |
| 170 | But anyway, today's theme is about AI. | 但无论如何,今天的主题是AI. |
| 171 | I would just like to share with you some of this AI moment, | 我想和你们分享一下这个AI时刻 |
| 172 | because everyone sitting in this room, I think, | 因为大家坐在这里,我想, |
| 173 | would have some of this thing touched as Dr. Lin just mentioned. | 就像林医生刚才提到的 这件事情会触动 |
| 174 | I like to play Go. | 我喜欢玩Go |
| 175 | Like many of you guys like to play the game Go, right? | 你们很多人喜欢玩游戏,对吧? |
| 176 | So Lee Sido, as you guys know, used to be the best Go player. | 所以,李西多,你们知道, 曾经是最好的GO球员。 |
| 177 | The best Korean player in the world. | 世界最佳韩国选手. |
| 178 | So back in 2016, back then I still lived in Boulder, Colorado. | 那时我还住在科罗拉多州的博尔德 |
| 179 | 我听说有个新的AI叫Alpha Go | |
| 180 | They actually have this five-set game with Lee | 其实他们跟李有五打 |
| 181 | and the winner is going to take away with $1 million. | 赢家要拿100万美元 |
| 182 | So that's still like late night in the U.S. | 所以这还是像 深夜在美国。 |
| 183 | It's already our time to watch the game. | 我们该看比赛了 |
| 184 | And then they have the best player to, | 然后他们有最好的球员, |
| 185 | the illustrators to talk about. | 插画家们要谈论的。 |
| 186 | I still very much like, | 我仍然很喜欢, |
| 187 | they said it's a very vivid memory. | 他们说这是一个非常生动的记忆。 |
| 188 | At the beginning, they were like, | 一开始,他们喜欢, |
| 189 | let's see how Lee is going to basically just hammer this AI game. | 让我们看看李会怎样 基本上锤这个AI游戏。 |
| 190 | There's no way AI can challenge a human with the most sophisticated game | AI不可能用最复杂的游戏挑战人类 |
| 191 | human ever invented, which is Go, right? | 人类曾经发明过,谁是Go,对不对? |
| 192 | Even compared to chess, like for you guys who don't play Go, | 即使比起象棋, 就像你们不玩Go, |
| 193 | even compared to chess, Go is a game with like 19 by 19 | 和棋比起来,吴是十九比十九 |
| 194 | and only you have like two sets of like those things, | 只有你有两套类似的东西, |
| 195 | stones, I think what they call it, black and white. | 石头,我想他们怎么称呼它,黑白。 |
| 196 | So it's really like the astronomical possibility for those | 所以,这真的像天文的可能性 对于这些 |
| 197 | and people think computer, even Deep Blue, | 人们认为计算机,甚至深蓝, |
| 198 | beats the chess champion back in 97. | 97年击败围棋冠军. |
| 199 | The Go is like really like the last thing, | GO就像最后一样 |
| 200 | like the human just holds our honor. | 就像人类只是持有我们的荣誉。 |
| 201 | So in the beginning they have this thing, | 所以一开始他们有这个东西, |
| 202 | but actually like, | 但实际上喜欢, |
| 203 | so just have this optimism until like the middle of the first game | 所以,只要有这种乐观 直到像第一场比赛的中间 |
| 204 | they were like, the illustrator like, | 他们喜欢, 插画家喜欢, |
| 205 | let's just count the number, | 我们数数吧 |
| 206 | how much Lee is ahead of this AI. | 李在人工智能上领先多少 |
| 207 | And then to their surprise, | 然后让他们惊讶的是, |
| 208 | and almost to everybody's surprise, | 几乎令大家惊讶, |
| 209 | really Lee already like left or behind the AI. | 真正的李已经喜欢 左或后AI。 |
| 210 | So he lost the first game | 所以他第一次输了 |
| 211 | and then he lost the second game, | 然后他输了第二场比赛, |
| 212 | he lost the third game. | 他输了第三场比赛 |
| 213 | So I think actually at that moment, | 所以我想实际上在那一刻, |
| 214 | it really struck me really badly. | 我真的很惊讶 |
| 215 | Like it's the most, | 就像它是最伟大的, |
| 216 | like I think like the, you know, | 就像我觉得,你知道, |
| 217 | if you like the enemy, | 如果你喜欢敌人 |
| 218 | if you like really like fight with someone, | 如果你真的喜欢和别人打架 |
| 219 | the most terrible thing is that | 最可怕的事情是,我们... |
| 220 | you don't even understand his move | 你连他的动作都不懂 |
| 221 | because he's so much better. | 因为他好多了 |
| 222 | He's just acting like just like overwhelmed you. | 他只是表现得像 让你不堪重负。 |
| 223 | So at that moment I just feel really the AI arrow really came. | 所以那一刻,我感觉真的AI箭真的来了。 |
| 224 | Of course, Lee won the fourth game | 当然,李赢了第四场比赛 |
| 225 | and that became the only game | 这成了唯一的游戏 |
| 226 | that a human player ever won against AI. | 人类玩家曾经战胜AI |
| 227 | So of course, | 当然 |
| 228 | 开发AlphaGo的人 | |
| 229 | didn't mind. | 没关系的 |
| 230 | They actually took the Nobel Prize | 他们真的拿了诺贝尔奖 |
| 231 | in chemistry last year. | 去年在化学领域 |
| 232 | 也因为除了AlphaGo之外 | |
| 233 | 他们开发了AlphaFold | |
| 234 | predicting the protein folding problem, | 预测蛋白质折叠问题, |
| 235 | which is also a difficult one | 这也是一个困难的 |
| 236 | because, you know, we all use like proteins | 因为,你知道, 我们都使用像蛋白质 |
| 237 | and for proteins, | 对于蛋白质, |
| 238 | they also take, you know, | 他们也采取,你知道, |
| 239 | more than the time of say the universe | 超过说宇宙的时间 |
| 240 | like would actually like form | 喜欢实际上喜欢形式 |
| 241 | to calculate all of the possibility. | 来计算所有的可能性。 |
| 242 | But protein can fold to like one structure | 但蛋白质可以折叠成一个结构 |
| 243 | to carry out all of the biological function | 履行所有生物功能; |
| 244 | and AlphaFold seems to solve the problem. | 而AlphaFold似乎解决了这个问题. |
| 245 | So now as Petey mentioned, | 所以现在,正如皮特提到, |
| 246 | I'm a professor of pharmaceutical science at Tsinghua. | 清华药学教授. |
| 247 | So I work in the field of | 所以,我的工作领域 |
| 248 | biotechnology and pharmaceutical industry. | 生物技术和制药业。 |
| 249 | From 2020, I started serving as an advisor | 从2020年开始,我开始担任顾问 |
| 250 | for SoftBank. | 为软银行。 |
| 251 | And that is the VC | 这是越共 |
| 252 | actually has quite much impact | 其实影响很大 |
| 253 | in like many different areas. | 在很多不同的领域。 |
| 254 | Actually, Katherine here, | 其实,凯瑟琳在这里, |
| 255 | she's a postdoc in my lab | 她是我实验室的博士 |
| 256 | and she actually was my co-worker | 她其实是我的同事 |
| 257 | actually we work with the SoftBank team | 其实我们和软银行团队合作 |
| 258 | with Eric Chen, | 和陈埃里克 |
| 259 | maybe you guys know him | 也许你们认识他 |
| 260 | 作为SoftBank的领导者. | |
| 261 | So as you guys know, | 所以,你们知道, |
| 262 | SoftBank invested Alibaba back in | 软银行投资阿里巴巴 |
| 263 | I don't know, 20 years ago | 我不知道,20年前的时候 |
| 264 | or something like that. | 或者类似的东西。 |
| 265 | They actually like picked the AI for drug. | 他们其实喜欢选择AI的药物。 |
| 266 | Like now people talk about AI Plus, | 就像现在人们谈论AI Plus, |
| 267 | AI for everything. | 万物的人工智能。 |
| 268 | So back 2020, | 因此,回到2020年, |
| 269 | like they actually get into the AI healthcare, | 就像他们真正进入AI保健, |
| 270 | AI drug development area. | AI药物开发区. |
| 271 | So the first company we invested | 我们投资的第一家公司 |
| 272 | is CrystalPie. | 这是水晶派。 |
| 273 | 有多少人认识CrystalPie? | |
| 274 | No, that is a drug company. | 不,那是个毒品公司 |
| 275 | That's actually a company | 那其实是一家公司 |
| 276 | went public last year. | 去年公开了 |
| 277 | The first one in Hong Kong | 香港第一个 |
| 278 | actually according to the 18C rule | 实际上根据18C规则 |
| 279 | as the rule said. | 如规则所说。 |
| 280 | So but this actually pretty crazy. | 但其实这很疯狂 |
| 281 | So back in 2020, SoftBank invested | 在2020年,软银行投资 |
| 282 | let's see that one, | 让我们来看看, |
| 283 | 370 million US dollars. | 3.7亿美元。 |
| 284 | That's actually a lot of money. | 这其实是一大笔钱。 |
| 285 | But the company was found | 但公司被发现了 |
| 286 | by three physics postdocs | 由三个物理后文档 |
| 287 | left like leaving | 离开就像离开 |
| 288 | I think they left MIT back in 2015. | 我认为他们在2015年离开了麻省理工学院。 |
| 289 | And then they actually like the first | 然后他们其实喜欢第一个 |
| 290 | like the first round | 就像第一轮一样 |
| 291 | annual round | 年度回合 |
| 292 | 实际上是从Tencent也。 | |
| 293 | So 2020 they actually like | 所以2020年 他们其实喜欢 |
| 294 | released 370 million | 释放3.7亿 |
| 295 | from SoftBank than others. | 从软银行比其他人。 |
| 296 | But what is amazing | 但是,这是惊人的 |
| 297 | because actually I wrote the technical report | 因为实际上我写了技术报告 |
| 298 | and then | 接着 |
| 299 | actually like how do you say that | 其实你怎么说 |
| 300 | Price Waters Cooper, | 普莱斯·沃特斯·库珀 |
| 301 | PCW wrote this financial report | PCW撰写了这份财务报告 |
| 302 | over five years | 五年以上 |
| 303 | not actually the annual income | 实际不是年收入 |
| 304 | the accumulated income over five years | 五年累计收入 |
| 305 | is 50 million RMB | 5千万人民币 |
| 306 | 50 million RMB | 5000万人民币 |
| 307 | supporting over 3 billion value. | 支持价值超过30亿美元。 |
| 308 | So at that moment I was like | 所以,那一刻我就像 |
| 309 | so this is a different story | 所以这个故事不一样 |
| 310 | we came to a different time | 我们来到了不同的时间 |
| 311 | because like in the drug industry | 因为像在毒品行业 |
| 312 | we call like | 我们叫它 |
| 313 | we got like those | 我们得到了像那些 |
| 314 | pre-clinical phase one, two, three | 临床前阶段一、二、三 |
| 315 | clinical trials | 临床试验 |
| 316 | after clinical phase two | 临床第二阶段之后 |
| 317 | we call that late stage | 我们称之为晚期 |
| 318 | it has started having some real world value | 它开始具有真实的世界价值 |
| 319 | but I feel that is the moment | 但我觉得现在 |
| 320 | we don't really use the old rules | 我们不用旧规矩 |
| 321 | to evaluate AI drug company anymore. | 来评价AI毒品公司 |
| 322 | So crystal pilots have so much money | 所以水晶飞行员有这么多钱 |
| 323 | they act like at some point | 他们好像在某个时候 |
| 324 | they have over one million US dollar cash | 他们有超过100万美元的现金 |
| 325 | in hand | 手执 |
| 326 | they became even VC themselves | 他们甚至成了越共本身 |
| 327 | to buy other companies | 购买其他公司 |
| 328 | so as you remember | 正如你所记得的 |
| 329 | I'm not sure how many of you know | 我不知道你们有多少人知道 |
| 330 | a VC 称为 DST | |
| 331 | that is also a famous one | 这也是一个著名的 |
| 332 | 他们很早就投资了Facebook | |
| 333 | so I remember | 所以我记得 |
| 334 | Katherine and I went to | 凯瑟琳和我去了 |
| 335 | partners | 合作伙伴 |
| 336 | Michael Chen | 陈迈克尔 |
| 337 | in Beijing | 北京会议 |
| 338 | so they have a Michelin three star restaurant | 他们有一个米其林三星级餐厅 |
| 339 | if you don't know that place | 如果你不知道那个地方 |
| 340 | because actually I wrote a technical report | 因为我写了一份技术报告 |
| 341 | for crystal pilot | 晶体飞行员 |
| 342 | working for soft banks | 为软银行工作 |
| 343 | we look over all of these AI drug companies | 我们看看所有这些AI毒品公司 |
| 344 | in China and also like overseas | 在中国和国外 |
| 345 | so actually like a lot of people came to talk to me | 所以实际上就像很多人来跟我说话 |
| 346 | so we talked with them like overnight | 所以我们和他们谈了一夜 |
| 347 | not overnight | 不是一夜之间 |
| 348 | like over three years | 像三年多一样 |
| 349 | at the end of this talk conversation | 在谈话结束时 |
| 350 | Michael was like | 迈克尔就像 |
| 351 | hey Hubert | 嘿,休伯特,你好吗? |
| 352 | do you think with AI | 你觉得AI |
| 353 | the drug companies | 制药公司 |
| 354 | like we came to a paradigm shifting moment | 就像我们来到一个范式转变的时刻 |
| 355 | so when we were talking about that | 所以当我们谈论 |
| 356 | that sort of makes sense | 这样才有意义 |
| 357 | really with AI | 和人工智能真的一样 |
| 358 | drug development | 毒品开发 |
| 359 | is different | 不一样 |
| 360 | from the past | 从过去 |
| 361 | years, decades | 年份、几十年 |
| 362 | even centuries | 甚至几个世纪 |
| 363 | so we really don't use the same | 所以,我们真的不用同样的 |
| 364 | evaluation rules | 评价规则 |
| 365 | as before | 和以前一样 |
| 366 | but of course I want to mention | 但我当然想提一下 |
| 367 | people talk about AI all the time | 人们总是谈论AI |
| 368 | but really like into the VC field | 但真的喜欢进入越共领域 |
| 369 | after 2022 | 2022年后 |
| 370 | and other VC stop investing | 其他维也纳公约停止投资 |
| 371 | AI | 大赦国际 |
| 372 | drug companies anymore | 毒品公司现在 |
| 373 | because I think that the key here | 因为我认为这里的钥匙 |
| 374 | is we don't figure a way to make money | 我们没有办法赚钱 |
| 375 | I think that's also a challenge | 我觉得这也是个挑战 |
| 376 | at least in my area | 至少在我的地区 |
| 377 | in healthcare and biomedical | 卫生保健和生物医学 |
| 378 | and biotechnology | 生物技术 |
| 379 | still some great challenges we are facing | 我们仍然面临一些巨大的挑战 |
| 380 | I'll stop here | 我就停在这儿 |
| 381 | thank you Professor Yun | 谢谢你,云教授 |
| 382 | our online guest | 我们的在线嘉宾 |
| 383 | Martin | 马丁 |
| 384 | and Ms. Lu Jiang | 卢江女士 |
| 385 | and Professor Van Elsteen | 范埃尔斯廷教授 |
| 386 | do you have any perspectives | 你有观点吗? |
| 387 | that you can share with us | 你可以和我们分享 |
| 388 | maybe I can get started quickly | 也许我可以快点开始 |
| 389 | hello everyone | 你们好 |
| 390 | so I'm glad to see so many friends here | 很高兴见到这么多朋友 |
| 391 | including some other board members | 包括其他一些董事会成员 |
| 392 | from Yale Graduate School Alumni Association | 耶鲁大学研究生院校友会 |
| 393 | welcome | 欢迎 |
| 394 | we are very happy to work with Helen | 我们很高兴和海伦一起工作 |
| 395 | and haven't seen you in a long time, Pei Ling | 很久没见到你了,裴玲 |
| 396 | so to have this AI series | 因此,有这个AI系列 |
| 397 | I think it's right on time | 我觉得很准时 |
| 398 | because we | 因为我们 |
| 399 | as a practitioner | 执业者 |
| 400 | in this industry | 在这个行业 |
| 401 | I really see many new applications | 我真的看到很多新的应用 |
| 402 | many new requirements | 许多新要求 |
| 403 | applying AI in their business | 在他们的企业中适用大赦国际 |
| 404 | so | 这样 |
| 405 | just a brief introduction about myself | 只是一个关于我的简介 |
| 406 | I went to Yale in 2001 | 我在2001年上过耶鲁大学 |
| 407 | before that I got my bachelor degree in computer science | 在那之前,我获得了计算机科学的学士学位 |
| 408 | from Tsinghua University | 清华大学 |
| 409 | so I spent six years getting my PhD | 所以我花了六年时间拿到博士学位 |
| 410 | at Yale | 在耶鲁大学 |
| 411 | so I'm now a board member | 所以我现在是董事会成员 |
| 412 | of Yale Graduate School Alumni Association | 耶鲁大学研究生院校友会 |
| 413 | which we help to coordinate | 我们帮助协调 |
| 414 | many events like this | 很多这样的事 |
| 415 | to connect all the alumni | 连接所有校友 |
| 416 | so | 这样 |
| 417 | many of the speakers like Linjie and Professor Yun | 林杰和云教授等许多发言者 |
| 418 | we met a lot | 我们经常见面 |
| 419 | offline | 离线 |
| 420 | so we know each other quite well | 所以我们很了解对方 |
| 421 | for myself | 为了我自己 |
| 422 | I first became a consultant | 我第一次当顾问 |
| 423 | after my PhD | 在我博士毕业后 |
| 424 | with the Boston Consulting Group | 与波士顿咨询集团 |
| 425 | helping many people | 帮助很多人 |
| 426 | expanding their business globally | 在全球扩大业务范围 |
| 427 | and after that | 在那之后,我们... |
| 428 | I was a VC for a few years | 我当了几年越共 |
| 429 | in Beijing | 北京会议 |
| 430 | so after that I started | 之后我开始 |
| 431 | my own company | 我的亲公司 |
| 432 | which is called Happy Universe | 叫做快乐的宇宙 |
| 433 | we started by helping | 我们从帮助开始 |
| 434 | many gaming companies | 许多游戏公司 |
| 435 | game developers and app developers | 游戏开发者和应用程序开发者 |
| 436 | from China to grow their business | 从中国发展他们的生意 |
| 437 | globally | 全球 |
| 438 | that was involved | 这涉及到 |
| 439 | with a lot of efforts | 付出了很大努力, |
| 440 | like online marketing and product | 像在线营销和产品 |
| 441 | modification | 修改 |
| 442 | user study localization | 用户学习本地化 |
| 443 | after that we see | 之后,我们看到 |
| 444 | many skills we developed | 我们培养了许多技能 |
| 445 | can help | 能够帮助 |
| 446 | larger companies | 大型公司 |
| 447 | especially when they want to | 特别是当他们想 |
| 448 | expand globally | 在全球扩展 |
| 449 | or locally in China | 在中国或当地 |
| 450 | but they don't have enough insight | 但他们没有足够的洞察力 |
| 451 | on their user behavior | 关于他们的用户行为 |
| 452 | and on the application design part | 和应用程序设计部分 |
| 453 | so we started helping | 所以我们开始帮忙 |
| 454 | many bigger companies | 许多大公司 |
| 455 | which is what we call | 我们称之为 |
| 456 | digitalization we call it | 我们称之为数字化 |
| 457 | which means they want to digitize | 这意味着他们想要数字化 |
| 458 | their process in their business | 企业的经营过程 |
| 459 | either in retail | 无论是零售 |
| 460 | sometimes in pharmaceutical | 有时在药店里 |
| 461 | so even in internet business | 即使是在互联网业务 |
| 462 | so they want to make their business | 所以他们想做生意 |
| 463 | more digitized | 数字化 |
| 464 | so that's about | 所以这是关于 |
| 465 | that's something we have been doing | 那是我们一直在做的事情 |
| 466 | for five years | 为期五年 |
| 467 | and we see many new demand | 我们看到许多新的需求 |
| 468 | recently on AI | 最近在大赦国际 |
| 469 | because | 因为 |
| 470 | a few big | 几个大 |
| 471 | biggest application scenario | 最大应用情景 |
| 472 | 和AIGC有关 | |
| 473 | stuff so that can enhance their | 东西,这样可以增强他们的 |
| 474 | productivity a lot | 生产率很高 |
| 475 | for example in the gaming industry | 例如,在游戏业 |
| 476 | we used to have about 30 people | 我们以前大概有30个人 |
| 477 | in our game art team | 在我们的游戏艺术团队中 |
| 478 | to design all the | 设计所有 |
| 479 | original graphics | 原始图形 |
| 480 | 3D model | 三维模型 |
| 481 | animation special effects | 动画特效 |
| 482 | you see many cool stuff in games | 你在游戏中看到很多很酷的东西 |
| 483 | maybe you guys can guess | 也许你们可以猜到 |
| 484 | how many people we have | 我们有多少人 |
| 485 | in that team | 在团队中 |
| 486 | so it's a big change | 变化很大 |
| 487 | so right now we only have three | 所以现在我们只有三个 |
| 488 | so we were able to | 所以我们可以 |
| 489 | to leverage | 以杠杆作用 |
| 490 | the AI tools to save about | 用于保存关于 |
| 491 | 80 to 90% of the work | 80%至90%的工作 |
| 492 | and the reason for that | 和原因 原因 |
| 493 | is many people become freelancer | 很多人成为自由职业者 |
| 494 | as well so | 也一样 |
| 495 | we actually need to do | 我们实际上需要做 |
| 496 | the coordination work in-house | 内部协调工作 |
| 497 | we actually outsource many works | 我们实际上外包了许多作品 |
| 498 | to to the freelancer | 致自由职业者 |
| 499 | and they actually can | 他们实际上可以 |
| 500 | do things much | 做点什么 |
| 501 | quicker with AI tools | 更快地使用 AI 工具 |
| 502 | and at much lower price | 价格低得多 |
| 503 | so our total cost | 所以我们的总成本 |
| 504 | decreased about | 约减少 |
| 505 | I would say 70% | 我会说70% |
| 506 | so like we don't keep so many employees | 所以像我们没有这么多员工 |
| 507 | and also we outsource the work | 我们把工作外包出去 |
| 508 | so the total cost saved more than half | 所以总成本节省了一半以上 |
| 509 | so that's a change we have seen | 所以,这是一个变化 我们已经看到 |
| 510 | in the gaming industry | 游戏业 |
| 511 | and in some other industries | 其他行业 |
| 512 | I see different | 我看不一样 |
| 513 | similar requirement as well | 同样的要求 |
| 514 | on improving the productivity | 提高生产率 |
| 515 | so I think we can chat more | 我想我们可以多聊聊 |
| 516 | about that later | 晚点再说吧 |
| 517 | but that's what happening in my industry | 但我的行业就是这样 |
| 518 | so thank you | 谢谢 |
| 519 | oh interesting Martin | 哦,有趣的马丁 |
| 520 | just to echo your point on the | 只是为了响应你的观点 |
| 521 | video and graphics production | 录像和图形制作 |
| 522 | related work | 相关工作 |
| 523 | I recently met an advertising | 我最近遇到一个广告 |
| 524 | company | 企业 |
| 525 | they were able to use AI | 他们能够使用人工智能 |
| 526 | to generate | 生成 |
| 527 | advertising videos | 广告录像 |
| 528 | and it's amazing animation | 这是惊人的动画 |
| 529 | and they are reducing | 他们正在减少, |
| 530 | the cost by 90% | 费用增加90% |
| 531 | yes | 对 |
| 532 | they told me | 他们告诉我 |
| 533 | so it's quite a tremendous cost saving | 因此,这是一个相当巨大的成本节省 |
| 534 | there | 那边 |
| 535 | okay | 还好 |
| 536 | Ms. Lu Zhang are you online | 吕章小姐,你上线了 |
| 537 | can you share some of the insights | 你可以分享一些见解 |
| 538 | from Silicon Valley | 从硅谷 |
| 539 | I know it's very early | 我知道现在还很早 |
| 540 | in your time zone | 在您的时区 |
| 541 | I'm a founder, manager, partner | 我是创始人,经理,合伙人 |
| 542 | of Fusion Fund | 合并基金 |
| 543 | actually I started investing in AI | 其实我开始投资AI |
| 544 | almost 10 years ago | 将近10年前 |
| 545 | since 2015 | 自2015年以来 |
| 546 | since we actually | 自从我们真的 |
| 547 | launched the firm | 发起公司 |
| 548 | even before that I was | 之前我还是 |
| 549 | a mature scientist | 一个成熟的科学家 |
| 550 | later became an entrepreneur | 后来成为企业家 |
| 551 | and for the company I built myself | 也为了我自己建立的公司 |
| 552 | were using AI for health care | 在医疗方面使用AI |
| 553 | so as some other panelists mentioned | 正如其他一些小组成员所提到的那样, |
| 554 | AI being in the industry for a long time | AI长期在业界工作 |
| 555 | but now finally | 现在终于 |
| 556 | we are using this option | 我们正在使用这个选项 |
| 557 | across the industry | 全行业 |
| 558 | with AI | 与大赦国际 |
| 559 | to focus on both business automation | 以两个业务自动化为重点 |
| 560 | which is the industrial automation | 这是工业自动化 |
| 561 | we've been talking about | 我们一直在谈论 |
| 562 | and also business optimization | 以及业务优化 |
| 563 | which is really treating AI | 这是真正治疗AI。 |
| 564 | not only as a tool | 不仅是工具 |
| 565 | but also the future of the digital labor | 还有数字劳工的未来 |
| 566 | AI labor | 大赦国际劳工 |
| 567 | be able to expand outside the tech industry | 能够扩展到技术产业之外 |
| 568 | but into the service industry | 但进入了服务业 |
| 569 | into like health care insurance | 像医疗保险一样 |
| 570 | financial logistic | 财务后勤 |
| 571 | to benefit from the | 从《公约》 |
| 572 | AI powered business optimization opportunity | AI 给企业提供优化机会 |
| 573 | so in the past 10 years | 过去10年里 |
| 574 | we invested so many companies | 我们投资了这么多公司 |
| 575 | and the opportunity | 机会和机会 |
| 576 | is really rising very very fast | 涨得很快 |
| 577 | I would say | 我会说 |
| 578 | when we talk about AI | 当我们谈论AI时 |
| 579 | the three things really matters | 这三件事真的很重要 |
| 580 | one is the compute | 一个是计算 |
| 581 | second is the data | 第二是数据 |
| 582 | and also the last thing is | 还有最后一件事 |
| 583 | really just the application use cases | 真的只是应用程序使用大小写 |
| 584 | another good thing is | 还有一件好事 |
| 585 | the amount of the data | 数据数量 |
| 586 | we have available for AI application | 我们有AI应用程序 |
| 587 | is quite fascinating | 相当迷人 |
| 588 | partially also due to the COVID | 部分原因还包括COVID |
| 589 | that lots of industry traditional sector | 很多行业的传统部门 |
| 590 | they have to adopt a digital tool | 他们必须采用一个数字工具 |
| 591 | which collect in terms of industry data | 收集工业数据 |
| 592 | used for AI training | 用于人工智能培训 |
| 593 | and across the industry | 和整个行业 |
| 594 | I'm personally really passionate about | 我个人对 |
| 595 | AI in health care | 保健方面的大赦国际 |
| 596 | lots of people probably didn't know | 很多人也许不知道 |
| 597 | that in human society | 在人类社会中 |
| 598 | over 30% of the data we have | 超过30%的数据 |
| 599 | is related | 相关 |
| 600 | but we only leverage less than 5% | 但是我们只有不到5%的杠杆 |
| 601 | of the data to help us | 帮助我们的数据 |
| 602 | with different type of innovation | 不同类型创新 |
| 603 | now with AI enabled opportunity | 现在有了AI允许的机会 |
| 604 | we are seeing | 我们看见了 |
| 605 | digital diagnostic | 数字诊断 |
| 606 | digital therapeutics | 数字治疗 |
| 607 | digital life science solution | 数字生命科学解决方案 |
| 608 | for cancer heart disease | 癌症心脏病 |
| 609 | and mental disease | 精神病 |
| 610 | so it's a prime time for AI | 所以这是AI的黄金时间 |
| 611 | in health care right now | 现在在医疗领域 |
| 612 | and also besides that | 除此之外 |
| 613 | there's other industry people | 还有别的行业人士 |
| 614 | huge amount of high quality data | 大量高质量的数据 |
| 615 | although there's a regulation compliance issue | 虽然有遵守规章的问题 |
| 616 | but adoption of the federal learning | 但通过联邦学习 |
| 617 | and the data encryption | 和数据加密 |
| 618 | AI governance technology | AI 治理技术 |
| 619 | make it possible to utilize AI | 使利用大赦国际成为可能 |
| 620 | and also one industry | 以及一个行业 |
| 621 | I really want to highlight | 我真的很想强调一下 |
| 622 | is the space industry | 是空间产业 |
| 623 | space tech | 空间技术 |
| 624 | and now since we have much lower launching costs | 现在我们的发射成本要低得多 |
| 625 | and every satellite you can see | 你可以看到的每一个卫星 |
| 626 | that the edge devices | 边缘设备 |
| 627 | will be able to collect | 能够收集 |
| 628 | lots of high quality data | 大量高质量的数据 |
| 629 | which is the safe data | 这是安全数据 |
| 630 | for AI training | 人工智能培训 |
| 631 | so that's also the opportunity | 这也是机会 |
| 632 | to build up application there | 以建立应用程序 |
| 633 | and since we talk about | 自从我们谈论 |
| 634 | a lot about data | 大量关于数据 |
| 635 | another thing is really about | 还有一件事 |
| 636 | the discussion between the model size | 模型大小之间的讨论 |
| 637 | the large model, vertical small model | 大模型,垂直小模型 |
| 638 | of course you guys heard | 当然你们听到了 |
| 639 | a lot of amazing launches | 很多惊人的发射 |
| 640 | from the major players | 从主要角色 |
| 641 | like Google Germany | 像谷歌德国 |
| 642 | like cloud | 像云一样 |
| 643 | and also open AI | 并开放AI |
| 644 | but besides the large model | 但除了那个大模型 |
| 645 | we also see a rising | 我们还看到上升 |
| 646 | of the small vertical model | 小型纵向模型 |
| 647 | for industry application | 工业应用 |
| 648 | and which doesn't necessarily | 也不一定会 |
| 649 | require lots of compute | 需要大量计算 |
| 650 | or energy | 或能量 |
| 651 | the cost could be really cheap | 费用可能很便宜 |
| 652 | and also model size could be as small | 模型大小可能也很小 |
| 653 | as less than a billion tokens | 不到十亿令牌 |
| 654 | so we can deploy on the local device | 这样我们就可以在本地设备上部署 |
| 655 | and in the local network | 在地方网络中 |
| 656 | which can really provide | 能够真正提供 |
| 657 | better security | 提高安全性 |
| 658 | and also compliance | 以及遵守情况 |
| 659 | requirement for AI adoption | 收养要求 |
| 660 | and we're also seeing | 我们还看到 |
| 661 | lots of open source community | 许多开源社区 |
| 662 | doing AI innovation | 创新 |
| 663 | that's kind of changing | 这有点改变 |
| 664 | the traditional company driven | 传统公司的驱动力 |
| 665 | AI 创新 | |
| 666 | computer driven AI innovation | 计算机驱动的AI创新 |
| 667 | so yeah there's a lot of things | 所以是有很多东西 |
| 668 | happening right now | 现在发生 |
| 669 | and we're also seeing more active | 我们也看到更活跃 |
| 670 | integration of the AI tech | 集成人工智能技术 |
| 671 | ecosystem | 生态系统 |
| 672 | this year alone we had five | 光是今年 我们就有五个 |
| 673 | exits already | 已经退出 |
| 674 | 2家公司被NVIDIA收购 | |
| 675 | one was acquired by Apple | 一个被苹果收购 |
| 676 | there's also the other two acquired | 还有另外两个 |
| 677 | by the public as a company | 公众作为公司 |
| 678 | that's also an indication of | 这也说明 |
| 679 | how fast the industry | 工业速度有多快 |
| 680 | is adopting AI | 正在通过大赦国际。 |
| 681 | right now | 现在 |
| 682 | and we're seeing | 我们看到 |
| 683 | even faster collaboration | 更快捷的合作 |
| 684 | between startups and large corporations | 初创企业与大公司之间 |
| 685 | one of our companies | 我们的一家公司 |
| 686 | 其实是JPMorgan大通的 | |
| 687 | within a couple of weeks | 两周内 |
| 688 | you know typically you only expect | 你知道你通常只期待 |
| 689 | the bank industry | 银行业 |
| 690 | to be able to | 能够 |
| 691 | finalize such a sensible check | 完成这种明智的检查 |
| 692 | with a startup within | 在内部启动 |
| 693 | even within three months now | 甚至三个月内 |
| 694 | we're talking about weeks | 我们在谈论几个星期 |
| 695 | so it's a great time | 所以,这是一个伟大的时间 |
| 696 | it's definitely an amazing opportunity | 这绝对是一个惊人的机会 |
| 697 | ahead of us in terms of | 就我们而言, |
| 698 | adopting AI and also as I said | 如我所说,通过大赦国际 |
| 699 | eventually it's not only just a tool | 最后,它不只是一个工具 |
| 700 | we need to use | 我们需要使用 |
| 701 | in the future of the digital labor | 在数字劳工的未来 |
| 702 | and us as a future | 我们是一个未来 |
| 703 | as a leader also potentially | 作为领导者也有可能 |
| 704 | need to manage both | 需要管理两者 |
| 705 | human labor and the digital AI labor | 人类劳动和数字人工智能劳动 |
| 706 | that's also some interesting | 这也很有趣 |
| 707 | opportunity and also challenge | 机会和挑战 |
| 708 | for us to consider how to evolve | 让我们考虑如何进化 |
| 709 | from there | 从那里 |
| 710 | I'll pause here and happy to share more | 我会在这里停下来,并乐于分享更多 |
| 711 | Thank you Lu | 谢谢你,鲁 |
| 712 | just to add some point | 只为了增加一点 |
| 713 | on interesting | 有关有趣的 |
| 714 | AI applications | AI 应用程序 |
| 715 | last year we invested in | 去年我们投资 |
| 716 | an interesting startup | 一个有趣的启动 |
| 717 | they produced an app | 他们制作了一个应用程序 |
| 718 | powered by AI | 由AI供电 |
| 719 | that enables | 能够 |
| 720 | pet users | 宠物用户 |
| 721 | pet owners to emphasize better | 宠物主人更强调 |
| 722 | with their pets | 和他们的宠物 |
| 723 | so it's typically a vertical | 所以典型的纵向 |
| 724 | model on dog behavior | 狗行为模型 |
| 725 | and dog language | 和狗语 |
| 726 | and initially we thought it's | 一开始我们以为 |
| 727 | interesting because you know | 很有趣,因为你知道 |
| 728 | people are increasingly adopting pets | 人们越来越多地收养宠物 |
| 729 | as their family member | 作为家庭成员 |
| 730 | so we think it will be | 所以,我们认为这将是 |
| 731 | a tool for pet users | 宠物用户的工具 |
| 732 | and perhaps we can build | 也许我们可以建造 |
| 733 | e-commerce on top of that | 电子商务除此之外 |
| 734 | but now we find that | 但现在我们发现 |
| 735 | a lot of hardware companies | 许多硬件公司 |
| 736 | are very interested in | 有兴趣 |
| 737 | technology | 技术 |
| 738 | because a lot of hardware | 因为很多硬件 |
| 739 | like cars | 像汽车一样 |
| 740 | they want to have | 他们想要 |
| 741 | more pet friendly features | 更亲切的宠物特征 |
| 742 | building their car systems | 建造汽车系统 |
| 743 | so large car manufacturers | 这么大的汽车制造商 |
| 744 | are actually talking with | 其实是说话 |
| 745 | a company about | 公司关于 |
| 746 | how to build | 如何建设 |
| 747 | into their in car system | 进入他们的汽车系统 |
| 748 | and more recently | 最近发生的 |
| 749 | even robot companies | 甚至连机器人公司 |
| 750 | are approaching the company | 已经接近公司了 |
| 751 | to talk about collaborating | 谈谈合作 |
| 752 | you know pet | 你知道宠物 |
| 753 | care features | 护理特征 |
| 754 | so perhaps in future | 也许将来 |
| 755 | we can have robots | 我们可以有机器人 |
| 756 | feed your pets | 养宠物 |
| 757 | at home and perhaps walk your dogs | 在家里,也许走你的狗 |
| 758 | so these are all very interesting | 所以这些都很有趣 |
| 759 | applications that we never thought | 我们从未想过的应用程序 |
| 760 | about initially | 初约. |
| 761 | AI and robotics | 人工智能和机器人 |
| 762 | a different kind of | 一种不同的类型 |
| 763 | new technologies | 新技术 |
| 764 | so that's quite interesting | 这很有趣 |
| 765 | Professor Van Esteen | 范埃斯滕教授 |
| 766 | can you also share | 也可以共享 |
| 767 | some of your experience | 你的一些经验 |
| 768 | from Boston? | 波士顿来的? |
| 769 | Sure I'm delighted to | 当然,我很高兴 |
| 770 | I gotta say here I am | 我得说,我在这里 |
| 771 | from the belly of the beast | 从野兽的肚子里 |
| 772 | adjacent to Harvard | 毗邻哈佛大学 |
| 773 | so yes I almost entered | 所以,是的,我几乎进入 |
| 774 | a little bit of enemy territory | 一点点的敌人领土 |
| 775 | but we started in computer | 但我们从电脑开始 |
| 776 | artificial intelligence | 人工智能 |
| 777 | technology using very different | 使用非常不同的技术 |
| 778 | methods at the time | 方法在当时 |
| 779 | logical inferencing | 逻辑推断 |
| 780 | it's quite fascinating | 相当迷人 |
| 781 | that the AI techniques | 人工智能技术 |
| 782 | now for the neural network | 现在进入神经网络 |
| 783 | technologies are quite different | 技术差异很大 |
| 784 | than those that were used a while | 比那些被暂时使用过的人 |
| 785 | ago and these are extraordinarily | 和这些特别 |
| 786 | different many of you may know | 你们中很多人可能知道 |
| 787 | the similar foundational models | 类似的基础模型 |
| 788 | are built on technology | 以技术为基础 |
| 789 | that's quite recent | 这是最近 |
| 790 | based on some insights from 2017 | 基于2017年的一些见解 |
| 791 | when AI is transforming | 当AI正在改变 |
| 792 | dramatically I think everyone | 令人惊讶的是,我想每个人 |
| 793 | anticipates the content creation | 预期内容创建 |
| 794 | and there's always some concern | 总是有些担心 |
| 795 | for the plagiarism detection | 用于侦测盗版 |
| 796 | or producing someone else | 或产生他人 |
| 797 | you know work that's not your own | 你知道这不是你自己的工作 |
| 798 | things of that sort but I think | 这样的事情,但我想 |
| 799 | some of the bigger transformations | 一些更大的转变 |
| 800 | are actually in content creation | 实际在内容创建中 |
| 801 | to tailor content | 以裁剪内容 |
| 802 | to generate whole new content | 生成全新内容 |
| 803 | courses to tailor to folks | 适合人们的课程 |
| 804 | interest it's interesting | 有兴趣,很有趣 |
| 805 | that there's academic evidence | 有学术证据 |
| 806 | AI很难 | |
| 807 | but if you put them in competition | 但是如果你把它们放在竞争中 |
| 808 | then in fact you might actually get | 然后,事实上你可能会得到 |
| 809 | them to work harder | 他们更努力地工作 |
| 810 | so maybe you can actually design | 所以也许你可以设计 |
| 811 | some incentive systems to get | 一些奖励制度 |
| 812 | better results out of it | 更好的结果出来 |
| 813 | another area that's quite interesting | 另一个有趣的领域 |
| 814 | colleagues at Siemens are actually | 西门子的同事其实是 |
| 815 | looking at industrial uses of AI | 研究AI的工业用途 |
| 816 | many of the ones that we talk about | 我们谈论的很多 |
| 817 | here have been consumer applications | 这里有消费者应用 |
| 818 | or you know solving the game of go | 或者你知道如何解决游戏 |
| 819 | or things of that sort | 或者那种东西 |
| 820 | most of the training sets are then | 大部分的训练是 |
| 821 | for video | 视频 |
| 822 | and they're huge opportunities | 他们是巨大的机会 |
| 823 | additionally in the industrial | 工业 |
| 824 | era for engineering for | 工程学的时代 |
| 825 | grasping for robotics | 抓住机器人 |
| 826 | for design under constraints | 有限情况下的设计 |
| 827 | for building construction | 建筑工程 |
| 828 | and I think some of those are | 我觉得其中一些是 |
| 829 | interesting new opportunities that | 有趣的新机会 |
| 830 | have not been as well explored | 还没有很好地探索 |
| 831 | so again as an academic we kind of | 所以,作为一个学术 我们算是 |
| 832 | study different other areas when we | 当我们学习不同的其他领域时 |
| 833 | give you facts or data | 提供事实或数据 |
| 834 | from other | 其他人员 |
| 835 | research on impacts of AI | 关于大赦国际影响的研究 |
| 836 | give you a few examples | 举几个例子 |
| 837 | hall center workers that have adopted | 收养的工人 |
| 838 | it become 14 more percent | 这又增加了14% |
| 839 | more productive in terms of | 提高生产率 |
| 840 | projects solved or in terms | 已解决或已解决的项目 |
| 841 | of satisfaction and turnover | 满意程度和更替率 |
| 842 | reduces but it's fascinating | 减少但很迷人 |
| 843 | {\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}新手会增加 | |
| 844 | productivity at 34 percent | 生产率为34% |
| 845 | and experts almost not at all | 和专家几乎完全没有 |
| 846 | so it's quite interesting that | 所以这很有趣, |
| 847 | there's a differential impact in the | 会有不同的影响 |
| 848 | productivity and the techniques | 生产力和技术 |
| 849 | of the best folks are being ported | 最好的人被移植 |
| 850 | over to the least | 倒计时 |
| 851 | experts the | 专家名单 |
| 852 | learning curve falls from about 10 | 学习曲线从大约 10 落下 |
| 853 | months to about three months | 月份至约3个月 |
| 854 | in making the | 用于制作 |
| 855 | productivity gains | 生产率的提高 |
| 856 | one fascinating element is that | 一个迷人的因素是 |
| 857 | you'll need to redesign your reward | 你需要重新设计你的奖励 |
| 858 | systems to make that happen | 实现这一点的系统 |
| 859 | because the experts don't want to be | 因为专家不想成为 |
| 860 | displaced or have their expertise | 流离失所者或具有专门知识者 |
| 861 | taken away in order that | 被带走是为了 |
| 862 | others can simply produce at the | 其他人可以简单生产 |
| 863 | same level the way most self | 和最自我一样 |
| 864 | 由Google驾驶 | |
| 865 | experiences 79 percent fewer | 经验减少79% |
| 866 | retention so there's some big | 留着吧,有大块头 |
| 867 | impact in that domain | 该领域的影响 |
| 868 | another fascinating experiment | 又一个迷人的实验 |
| 869 | showed that | 显示 |
| 870 | large language model voice recruiter | 大型语言模式语音招募者 |
| 871 | we're using it to actually hire | 我们用它来雇人 |
| 872 | people | 人员 |
| 873 | generated 12 percent more | 增加12% |
| 874 | offers 18 percent | 18%的报价 |
| 875 | more starts and 17 percent | 更多开始和17% |
| 876 | higher retention | 留存额较高 |
| 877 | a month after the job had been | 工作一个月后 |
| 878 | hired so there's actually some big | 雇了这么多人 |
| 879 | impact there | 撞击那里 |
| 880 | another one | 又一个 |
| 881 | someone looked at the | 有人看着 |
| 882 | searches for companies | 搜索公司 |
| 883 | using the term artificial | 使用人工术语 |
| 884 | intelligence in their employment | 就业时掌握的情报 |
| 885 | and correlated that with state | 并与之相关联 |
| 886 | hiring and state level gross | 雇用和州一级毛额 |
| 887 | domestic product | 国内生产总值 |
| 888 | fascinatingly in the | 令人着迷 |
| 889 | month following bursts in | 连续一个月 |
| 890 | activity and searches for AI and | 对大赦国际的活动和搜索 |
| 891 | hiring the GDP | 雇用国内生产总值 |
| 892 | in those states fell by point | 这些州逐点下降 |
| 893 | of one percent for a month | 一个月1% |
| 894 | and then a year later | 一年过去了 |
| 895 | grew by point three six | 三点六分 |
| 896 | percent so the impact seemed to | 所以影响似乎 |
| 897 | have a lag over time | 时间有滞后 |
| 898 | so there's actually a big impact | 因此,实际上有一个巨大的影响 |
| 899 | but it happens much later | 但后来发生很多事 |
| 900 | and then I attended a conference | 然后我参加了一个会议 |
| 901 | there in Beijing only this past | 在北京只有过去 |
| 902 | summer and there's some wonderful | 夏天和一些美好的东西 |
| 903 | Kevin Dye CEO提供的数据 | |
| 904 | Tongdao Li Pin group there in | 东岛李平组 |
| 905 | China so let me just share a little | 中国,让我分享一下 |
| 906 | one or two data points from that | 一个或两个数据点 |
| 907 | that I thought were fascinating | 我觉得很迷人 |
| 908 | he's suggesting that and they run | 他这么说,他们就跑了 |
| 909 | a job recruiting site | 招聘地点 |
| 910 | the 220 million jobs in China | 中国2.2亿个就业岗位 |
| 911 | are already used by 78 percent of | 78%的人已经使用过 |
| 912 | firms already using it | 已经使用它的公司 |
| 913 | there's a sharp decline in | 急剧下降 |
| 914 | entry level work and yet a huge | 入门工作,但规模很大 |
| 915 | surge in AI and senior | AI和高年级学生激增 |
| 916 | level work and also | 各级工作,以及 |
| 917 | new job requirements | 新的工作要求 |
| 918 | are kind of different than some of | 有点不同 |
| 919 | the older job requirements | 较旧的工作要求 |
| 920 | they're seeing | 他们看到 |
| 921 | increased quest for an aesthetic | 增加追求美学 |
| 922 | sense being self-driven | 自我驱动的感觉 |
| 923 | being a deep thinker | 作为深思熟虑者 |
| 924 | fast open-minded learner | 快速开明的学习者 |
| 925 | and of course artificial | 当然是人为的 |
| 926 | intelligence and AI | 情报和大赦国际 |
| 927 | so all of those are kind of | 所以所有这些都是一种 |
| 928 | interesting new skills | 有趣的新技能 |
| 929 | one other thing that you'd asked | 还有一件事你问 |
| 930 | about was the business models | 关于商业模式 |
| 931 | and I want to hear a little bit | 我想听一点 |
| 932 | more about that because most of my | 更多关于这一点,因为我的大多数 |
| 933 | own expertise is really on platform | 自己的专长真的在平台上 |
| 934 | business models matter of fact we | 事实上,我们 |
| 935 | have a book on platform business | 有一本关于平台业务的书 |
| 936 | models and how those operate so | 模型及其运作方式 |
| 937 | just to share a few thoughts on the | 只是为了分享一些关于 |
| 938 | artificial intelligence business | 人工智能业务 |
| 939 | models it's interesting that some | 模特儿 有趣的是有些 |
| 940 | of the breakthroughs have initially | 最初的突破 |
| 941 | come from these big platform | 来自这些大平台 |
| 942 | companies in the United States of | 美国境内的公司 |
| 943 | course it's Google Microsoft | 当然是谷歌微软 |
| 944 | Apple and others they're the ones | 苹果和其他人就是他们 |
| 945 | with the massive amounts of data | 大量的数据 |
| 946 | that are possibly used for | 可能用于 |
| 947 | training sets | 培训集 |
| 948 | it's also the case that they're | 这也是他们 |
| 949 | usually typically | 通常情况下 |
| 950 | funding those models either you | 你也可以资助这些模式 |
| 951 | know through venture capital | 通过风险资本知道 |
| 952 | investments or through | 投资或通过 |
| 953 | subscriptions some other things I | 订阅一些其他东西 I |
| 954 | think are going to be coming down | 以为会下来 |
| 955 | {\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}那匹小马和那匹马 | |
| 956 | a lot of data that's being | 许多数据是 |
| 957 | used creating data which generates | 用于创建生成数据 |
| 958 | usage creating | 创建用法 |
| 959 | massive industry concentration | 大规模工业集中 |
| 960 | if you doubt that | 如果你怀疑的话 |
| 961 | you should think of the salaries | 你应该想想薪水 |
| 962 | that are being offered | 正在提供 |
| 963 | for the top AI scientists in | 高级人工智能科学家 |
| 964 | moment the competition is so fierce | 当比赛激烈的时候 |
| 965 | in the United States imagine it's | 在美国 想象一下 |
| 966 | similar in China | 中国的类似情况 |
| 967 | but the competition for AI top | 但是AI顶级的竞争 |
| 968 | programmers is so fierce | 程序员这么凶猛 |
| 969 | you know packages of tens of | 你知道几万个包 |
| 970 | millions of dollars for signing | 百万的签名 |
| 971 | up for AI there's a there's an | 人工智能上有一条 |
| 972 | anticipation of huge | 即将到来 |
| 973 | gains which kind of forecasts | 何种预测收益 |
| 974 | industry concentration | 工业集中 |
| 975 | and what let me give you one last | 让我给你最后一个 |
| 976 | thought just based on some | 认为只是基于一些 |
| 977 | research that we did in the content | 我们所做的研究的内容 |
| 978 | projected going forward so I don't | 预测前进,所以我不 |
| 979 | know if I'll be able to share | 如果我能分享的话 |
| 980 | one screen on your behalf | 一个屏幕代表您 |
| 981 | let's see if this works I don't even | 看看这是否可行 我甚至不 |
| 982 | know if sharing was enabled here | 知道是否在此启用共享 |
| 983 | yep that seems as though it was so | 是的,似乎它是如此 |
| 984 | I'm going to try to share one | 我要试着分享一个 |
| 985 | one slide | 一个幻灯片 |
| 986 | let's look at APIs or application | 让我们看看API 或应用程序 |
| 987 | programming interfaces | 编程接口 |
| 988 | one of the things that we found | 我们发现的东西之一 |
| 989 | fascinatingly is that firms | 令人着迷的是公司 |
| 990 | 打开 API 的 | |
| 991 | grew in market cap | 市场上限增长 |
| 992 | 4% within two years | 两年内4% |
| 993 | and 36% within | 和36% |
| 994 | 16 years well what happened | 16年好,发生了什么事 |
| 995 | they became the center | 他们成了中心 |
| 996 | of a connected social network | 社会网络 |
| 997 | of information flows | 信息流动 |
| 998 | it's almost like social networks | 就像社交网络 |
| 999 | where you become a power broker | 在那里你成为一名权力经纪人 |
| 1000 | the more contacts and the bigger | 接触越多 接触越多 |
| 1001 | is that you can anticipate this | 你可以预料到这一点 |
| 1002 | exact same phenomenon | 完全一样的现象 |
| 1003 | will happen among AI firms | 将发生在AI公司之间 |
| 1004 | there's a new model context | 有一个新的模式背景 |
| 1005 | protocol or 80 agent to | 协议或80名代理人 |
| 1006 | agent protocol | 代理协议 |
| 1007 | and to the extent that others are | 并且只要其他人是 |
| 1008 | going to be using your individual | 将使用您的个人 |
| 1009 | resources and you can program them | 资源并可以编程 |
| 1010 | around them you too may be | 在他们周围,你也可能 |
| 1011 | a nexus of hiring | 雇用关系 |
| 1012 | or a nexus of interactions | 或相互作用的关系 |
| 1013 | for data and | 数据及 |
| 1014 | for | (单位:千美元) |
| 1015 | power | 电源 |
| 1016 | you can invest in those same kinds | 你可以投资同样的种类 |
| 1017 | {\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}这简直就是 | |
| 1018 | looking backwards at APIs | 向后查看 API |
| 1019 | and projecting forwards to this | 并投影到此 |
| 1020 | next generation of technology | 下一代技术 |
| 1021 | as good places to invest | 作为投资的好地方 |
| 1022 | so you'd ask for just some | 所以你会要求一些 |
| 1023 | background and perhaps some | 和也许一些 |
| 1024 | business models so | 商业模式 |
| 1025 | I'll stop there and see what else | 我停下来看看还有什么 |
| 1026 | folks would like to hear about | 人们想听听 |
| 1027 | great great thanks for sharing | 非常感谢分享 |
| 1028 | that's very very interesting | 这很有趣 |
| 1029 | statistics and we have | 统计与我们 |
| 1030 | talked about you know successful | 说到成功 |
| 1031 | adoption | 收养 |
| 1032 | and are there also bad cases | 还有坏事 |
| 1033 | of using AI | 使用人工智能 |
| 1034 | I don't know if | 我不知道,如果 |
| 1035 | any of you have any insights on | 你们中谁对 |
| 1036 | that maybe Jay can start | 也许杰伊可以开始 |
| 1037 | with some of your experience | 有你的一些经验 |
| 1038 | failed cases of | 失败的个案 |
| 1039 | AI adoption | 大赦国际的收养 |
| 1040 | in in | 输入 |
| 1041 | traditional industries | 传统工业 |
| 1042 | yeah well yeah well there are | 是啊,是啊,还有 |
| 1043 | more failed failures than success | 失败多于成功 |
| 1044 | obviously | 很明显 |
| 1045 | without | 无 |
| 1046 | mentioning the name but I can talk | 提到这个名字,但我可以说 |
| 1047 | about one robotics | 大约一个机器人 |
| 1048 | installation failure | 安装失败 |
| 1049 | one of the | 一个 |
| 1050 | logistic | 后勤 |
| 1051 | warehouses before | 之前的仓库 |
| 1052 | well yeah well | 不错,不错 |
| 1053 | there are not many | 没有多少 |
| 1054 | warehouses using | 使用 |
| 1055 | robotics | 机器人 |
| 1056 | so you probably guessed who but | 所以你可能猜到是谁了 |
| 1057 | anyway that's back about | 无论如何,这是回来 |
| 1058 | 20 | 20国 |
| 1059 | that's actually 2015 | 其实是2015年 |
| 1060 | AGV 开始 | |
| 1061 | with Amazon first of all | 首先与亚马逊 |
| 1062 | Amazon use a lot of AGVs | 亚马逊使用了很多 AGV |
| 1063 | to move around good so | 转过身来 |
| 1064 | in China there are companies | 在中国有公司 |
| 1065 | producing AGVs at a time | 一次生产 AGV |
| 1066 | and then provided to | 然后提供给 |
| 1067 | to this to this client | 到此客户端 |
| 1068 | to use in their | 用于其 |
| 1069 | automation | 自动化 |
| 1070 | picking lines in the | 选择线条 |
| 1071 | in the warehouse | 在仓库里 |
| 1072 | now the failure was | 现在的失败是 |
| 1073 | the design of the automatic | 自动设计 |
| 1074 | picking line | 选择线条 |
| 1075 | is designed to have a maximum | 被设计为最大 |
| 1076 | capacity of | 能力 |
| 1077 | I can't remember exactly number | 我不记得确切的数字了 |
| 1078 | but something around | 但周围的东西 |
| 1079 | a hundred thousand | 十万块 |
| 1080 | packages per hour | 每小时包数 |
| 1081 | that's the | 这是 |
| 1082 | that's the picking line | 这就是选择线 |
| 1083 | that's another maximum | 这是另一个最大 |
| 1084 | capacity was designed | 设计能力 |
| 1085 | and then before the robotics | 然后在机器人之前 |
| 1086 | was installed | 已安装 |
| 1087 | we use human beings right we have | 我们利用人类的权利 |
| 1088 | people at different | 不同人间 |
| 1089 | locations around the | 地点 |
| 1090 | the picking line so when the | 选择线,这样当 |
| 1091 | package was sorted over to | 软件包已排序到 |
| 1092 | to your station and then we | 到你的站 然后我们 |
| 1093 | have people we make | 有我们制造的人 |
| 1094 | predictions we say okay this | 预言我们说好的 |
| 1095 | station probably needs two | 车站可能需要两个 |
| 1096 | people that station probably | 可能就是那个站的人 |
| 1097 | needs five people depending on | 需要5个人 |
| 1098 | the destination right so it | 目的地是这样的 |
| 1099 | works fine it works | 不错,不错 |
| 1100 | when the automation line | 当自动化线 |
| 1101 | was installed and then the | 已安装,然后 |
| 1102 | speeds pick up and then all | 快速捡起然后全部 |
| 1103 | you need to do is adding more | 您需要做的是添加更多 |
| 1104 | in line | 线条 |
| 1105 | 但当AGV是 | |
| 1106 | installed | 已安装 |
| 1107 | AGV 速度为 | |
| 1108 | usually like | 通常喜欢 |
| 1109 | one meter per second | 每秒1米 |
| 1110 | that's pretty fast already | 已经快了 |
| 1111 | otherwise you can go too fast | 否则你走得太快了 |
| 1112 | indoor right for the AGVs | AGVs的室内右侧 |
| 1113 | running around in the | 到处乱跑 |
| 1114 | in your warehouse | 在你的仓库 |
| 1115 | then the disaster happened | 然后灾难就发生了 |
| 1116 | that is you know | 这就是你知道 |
| 1117 | the maximum capacity | 最大容量 |
| 1118 | of this automatic | 此自动 |
| 1119 | sorting line | 排序行 |
| 1120 | reduces to something like | 减为类似的东西 |
| 1121 | five thousand packages an hour | 每小时五千包 |
| 1122 | because the robotics | 因为机器人 |
| 1123 | the AGVs are just | AGV是正义的 |
| 1124 | not fast enough to | 不够快 |
| 1125 | pick up the goods coming from | 拿起货来 |
| 1126 | the sorting line and ship it | 分拣线和船运 |
| 1127 | over to the warehouse | 到仓库去 |
| 1128 | so that was a huge | 所以,这是一个巨大的 |
| 1129 | waste actually | 浪费实际上 |
| 1130 | then today there's a smarter | 那么今天有个更聪明的人 |
| 1131 | way to do it today because | 如何做到这一点,因为 |
| 1132 | today people invented | 今天有人发明了 |
| 1133 | 首先是AGV | |
| 1134 | is more automatic | 更自动 |
| 1135 | and carry more weights | 并携带更多的重量 |
| 1136 | so instead of carrying | 而不是携带 |
| 1137 | one package the AGV nowadays | 如今的 AGV 软件包 |
| 1138 | they carry the whole stack | 他们把整个堆 |
| 1139 | right so this whole stack | 对,所以整个堆 |
| 1140 | of like ten layers | 大约十层 |
| 1141 | and then you probably have | 然后你也许有 |
| 1142 | you know more than a thousand | 你认识一千多人 |
| 1143 | packages over in one stack | 包在一个堆栈中结束 |
| 1144 | and then you shovel the whole | 然后,你铲整个 |
| 1145 | stack into the warehouse | 堆入仓库 |
| 1146 | spaces instead of | 空格替换为 |
| 1147 | showing one package over | 显示一个软件包 |
| 1148 | so but then back | 所以,然后回来 |
| 1149 | and well it's the first | 而且这是第一个 |
| 1150 | doctor right so | 对,医生 |
| 1151 | but that was a huge | 但它是一个巨大的 |
| 1152 | failure and then | 失败,然后 |
| 1153 | which and then I know | 后来我知道 |
| 1154 | another concrete example | 另一个具体例子 |
| 1155 | which I did myself | 我亲手做的 |
| 1156 | was I was doing | 我当时在做 |
| 1157 | another startup | 另一个启动 |
| 1158 | about oh yes | 大约是的 |
| 1159 | ten years ago | 10年前,我们... |
| 1160 | I do this unattended | 我做这个无人管 |
| 1161 | convenience store | 便利商店 |
| 1162 | followed by we follow | 然后我们跟着 |
| 1163 | Amazon Go right | 亚马逊 右转 |
| 1164 | and then so I have no people | 所以我没有人民 |
| 1165 | actually on duty | 实际值班 |
| 1166 | it's all camera picks up who | 所有摄像头接谁 |
| 1167 | enters the store pickups what | 进入商店皮卡 |
| 1168 | and then and then check out | 然后检查出来 |
| 1169 | automatically it was all | 都自动了 |
| 1170 | fun but then | 好玩,不过 |
| 1171 | the first store I opened | 我开的第一家店 |
| 1172 | was near Hongqiao | 刚才在香港附近 |
| 1173 | Airport back then that | 当时的机场 |
| 1174 | area is like desert | 区域就像沙漠 |
| 1175 | I mean today you go there you | 我是说今天你去那里 |
| 1176 | have all these buildings | 拥有这些建筑 |
| 1177 | whatsoever but ten years ago | 10年前,除了别的 |
| 1178 | it's almost desert | 它几乎是沙漠 |
| 1179 | so you have this | 所以你有这个 |
| 1180 | there are more workers there | 那里还有更多的工人 |
| 1181 | construction workers there | 那里的建筑工人 |
| 1182 | than the office white collar | 比办公室的白领 |
| 1183 | right so one day | 对啊,有一天 |
| 1184 | then all of a sudden one day | 突然有一天 突然间 |
| 1185 | I found the | 我发现 |
| 1186 | abnormal data | 异常数据 |
| 1187 | sales data coming out of the | 销售数据 |
| 1188 | store at one time for 15 | 一次存储 15 |
| 1189 | minutes during lunchtime | 午餐时间 |
| 1190 | that is the shelf | 这是架子 |
| 1191 | this while I have a sensor | 趁我有感应器 |
| 1192 | which under the shelf | 在货架下 |
| 1193 | so the shelf told me | 所以货架告诉我 |
| 1194 | there are a lot of | 还有很多 |
| 1195 | products are taken away | 产品被拿走 |
| 1196 | from the shelf because the | 因为... |
| 1197 | weights right the weights | 权重 权重 |
| 1198 | change but then there's no | 变了,但是没有 |
| 1199 | check out so and then | 检查出来,然后 |
| 1200 | and then what I realized later | 后来我意识到 |
| 1201 | was the computer at that | 是电脑在那 |
| 1202 | time when we do computer | 我们做电脑的时间 |
| 1203 | vision people face recognition | 愿景:人们面临承认 |
| 1204 | and then gesture recognition | 然后手势识别 |
| 1205 | these construction workers | 这些建筑工人 |
| 1206 | they came in a dozen of | 他们进来了十几个 |
| 1207 | them after a shift | 他们下班后, |
| 1208 | they all wear the same uniform | 他们都穿同样的制服 |
| 1209 | right and they all wear the | 他们都戴着 |
| 1210 | helmet safety hazmat | 头盔安全罩 |
| 1211 | so when the first | 所以,当第一个 |
| 1212 | first person scan his face | 第一人称扫描他的脸 |
| 1213 | coming in | 进来 |
| 1214 | the second person coming in | 第二个人进来 |
| 1215 | when these dozen people come | 这十几个人来的时候 |
| 1216 | into my store | 进入我的店里 |
| 1217 | the algorithm just doesn't | 算法不会 |
| 1218 | work I mean it all looks the | 工作,我的意思是,一切看起来 |
| 1219 | same so these people | 这些人也一样 |
| 1220 | grab a lot of food | 多拿点吃的 |
| 1221 | from off my shelf | 从我的架子上 |
| 1222 | and then the computer | 然后是电脑 |
| 1223 | couldn't know which one to | 不知道是哪个 |
| 1224 | charge to so that's what | 所以,这就是 |
| 1225 | really happened but then | 真的发生了,但后来 |
| 1226 | I call it a failure the | 我称之为失败 |
| 1227 | reason is because we | 原因是因为我们 |
| 1228 | thought at that time | 那个时候的想法 |
| 1229 | as the technology | 作为技术 |
| 1230 | is good enough to | 足够好 |
| 1231 | or even to perfect | 甚至为了完美 |
| 1232 | not perfect but good | 不完美但不错 |
| 1233 | enough to do business | 足够做生意 |
| 1234 | but the mistake we made | 但我们所犯的错误 |
| 1235 | was it might be okay in | 也许可以进去 |
| 1236 | other cases when you have | 其他情况 |
| 1237 | in retail I mean | 在零售业,我的意思是 |
| 1238 | the margin of a typical rate | 典型比率的比值 |
| 1239 | to convenience store is like | 到便利店就像 |
| 1240 | net margin is like five percent | 净比值是5% |
| 1241 | right when you have this | 当你有这个的时候 |
| 1242 | twenty percent error | 20%的误差 |
| 1243 | and sometimes you charge people | 有时你收人钱 |
| 1244 | more right sometimes you | 有时候你更正确 |
| 1245 | charge more and sometimes you | 收钱多,有时你 |
| 1246 | charge less but with that | 少收钱,但用那个 |
| 1247 | variation so that technology | 技术的改变 |
| 1248 | at that time even today | 那时候,即使是今天 |
| 1249 | though by the way it's just | 顺便说一句 |
| 1250 | not suitable for this | 不适合这个 |
| 1251 | particular application where | 在下列情况下特别适用: |
| 1252 | we think for example | 我们想,比如 |
| 1253 | when you have when we | 当你有,当我们 |
| 1254 | have all these cameras | 拥有所有这些摄像头 |
| 1255 | in the airport monitoring | 机场监测 |
| 1256 | abnormal behaviors and | 异常行为 |
| 1257 | counting traffic | 计算流量 |
| 1258 | etc. | 页:1 |
| 1259 | that you don't need | 你不需要钱 |
| 1260 | a hundred percent accuracy I | 100%的准确性I |
| 1261 | mean you only need to | 就是说你只需要 |
| 1262 | know a ballpark of how many | 知道有多少个球板 |
| 1263 | people are walking through | 人们走过 |
| 1264 | this camera at this time | 此相机此时 |
| 1265 | blah blah that is okay | 没关系的 |
| 1266 | but unfortunately | 但很遗憾 |
| 1267 | eight years ago we were | 八年前我们 |
| 1268 | bold bold enough | 大胆一点 |
| 1269 | and naive enough to | 和天真的足够 |
| 1270 | trust that that works | 相信这有效 |
| 1271 | and then I'll tell you what's | 然后我会告诉你什么是 |
| 1272 | the solution now solution is | 现在的解决办法是 |
| 1273 | not an artificial intelligence | 不是人工智能 |
| 1274 | it's a human intelligence | 这是人类的智慧 |
| 1275 | so they have still cameras | 所以他们还有摄像头 |
| 1276 | in the store but they have | 在商店里,但他们有 |
| 1277 | this human representative | 这个人类代表 |
| 1278 | sitting somewhere in India | 坐在印度的某个地方 |
| 1279 | maybe I don't know and then | 也许我不知道,然后 |
| 1280 | she or he he watches the | 她或他看着 |
| 1281 | camera right so big daddy | 摄影机好大 爸爸 |
| 1282 | is watching and then | 正在观看,然后 |
| 1283 | he sees this person picks up | 他看见这个人接走了 |
| 1284 | and then charges them much | 并收取他们很多钱 |
| 1285 | but don't laugh but that's a | 但别笑,不过那是 |
| 1286 | actually that's a innovative | 其实,这是一个创新 |
| 1287 | business model right because | 商业模式正确,因为 |
| 1288 | when you have a physical store | 当你有一个物理商店 |
| 1289 | once people can only stay in | 一旦人们只能留在 |
| 1290 | this store and account you | 这家店和你帐号 |
| 1291 | serve for this store right | 为这家店服务 |
| 1292 | but this model a person sits | 但这个模型一个人坐 |
| 1293 | behind the desk looking at | 在桌子后面看着 |
| 1294 | he can look at 20 cameras | 他可以看到20个摄像头 |
| 1295 | simultaneously so he can | 同时又能让他 |
| 1296 | actually service 20 different | 实际服务20种不同 |
| 1297 | stores at one time | 一次存储 |
| 1298 | so that's actually pretty | 所以这其实很漂亮 |
| 1299 | smart so and now | 这么聪明,现在 |
| 1300 | it is there are still | 还有呢 |
| 1301 | applications and I would say | 申请,我会说 |
| 1302 | pretty successful successful | 相当成功 |
| 1303 | applications of unattended | 未处理的申请 |
| 1304 | convenience store models | 便利店型号 |
| 1305 | in China especially in | 中国,特别是 |
| 1306 | the southern part in | 南部地区 |
| 1307 | Guangdong and then they're | 广东然后是 |
| 1308 | actually making money and | 其实赚钱和 |
| 1309 | then they're actually using | 然后他们实际上在使用 |
| 1310 | technologies is with with | 技术与 |
| 1311 | the help of AI AI helps | 大赦国际的帮助 |
| 1312 | to give you one result | 给你一个结果 |
| 1313 | but then the human being is | 但人则 |
| 1314 | the final judge to say okay | 终审法官说好的 |
| 1315 | whether or not I use | 无论我是否使用 |
| 1316 | the recommended solution or | 建议的解决办法或 |
| 1317 | not and then when at this | 不,然后,当在这里 |
| 1318 | time where you have 80 percent | 有80%的时间 |
| 1319 | accuracy that's good enough | 准确性足够好 |
| 1320 | it filters 80 percent of the | 它过滤了80%的 |
| 1321 | orders right and then the | 命令右转,然后 |
| 1322 | person sitting behind the desk | 坐在办公桌后面的人 |
| 1323 | looking at cameras only need to | 看着摄像机只需要 |
| 1324 | go through like 20 percent of | 经历的20% |
| 1325 | the orders and that's pretty | 命令,这是漂亮的 |
| 1326 | remarkable adoption so I'll | 了不起的收养,所以我会 |
| 1327 | stop stop here okay that's an | 停下来,这是个 |
| 1328 | amazing example so you can use | 惊人的例子,这样你就可以使用 |
| 1329 | AI but you cannot completely | 人工智能,但你无法完全 |
| 1330 | use AI but you still need | 使用人工智能,但你还需要 |
| 1331 | some human intervention | 一些人类干预 |
| 1332 | our online guest | 我们的在线嘉宾 |
| 1333 | Miss Lu Zhang | 吕章小姐 |
| 1334 | maybe can you give us some | 也许你可以给我们一些 |
| 1335 | closing remarks as you have to | 结束语 |
| 1336 | leave early I think Zhang Lu | 早点走,我想张鲁 |
| 1337 | she's left already she has she | 她已经离开了 她有她 |
| 1338 | has a flight to catch but I | 有飞机要赶,但我... |
| 1339 | maybe I'll just add two | 或许我再加两个 |
| 1340 | comments about her role | 关于她作用的评论 |
| 1341 | in joining the panel | 加入小组 |
| 1342 | she is a female she actually | 其实她是个女人 |
| 1343 | one of the very first | 是第一个 |
| 1344 | women investor in Silicon | 硅的女性投资者 |
| 1345 | Valley and she is also | 谷子和她是一样 |
| 1346 | a and I think with that | 和我想着 |
| 1347 | viewpoint I just wanted to | 观点 我只是想 |
| 1348 | particularly highlight the need | 特别强调 |
| 1349 | for more women to be involved in | * 让更多妇女参与 |
| 1350 | what is right now more | 现在的更多 |
| 1351 | male-dominated industry of | 以男性为主的工业 |
| 1352 | not only an AI but also in | 不仅是大赦国际,而且 |
| 1353 | the VC field also Zhang Lu | 也是张鲁 |
| 1354 | has also continuously traveled | 也不断旅行 |
| 1355 | in bridging bringing | 搭桥 |
| 1356 | technologies interesting | 有趣的技术 |
| 1357 | projects from Silicon Valley | 硅谷项目 |
| 1358 | also to Asia and | 向亚洲和 |
| 1359 | in fact I've seen her | 其实我见过她 |
| 1360 | I'm based in Singapore | 我总部在新加坡 |
| 1361 | and I also wanted to highlight | 我还想强调 |
| 1362 | what I see is greater bridging | 我看到的是更大的桥梁 |
| 1363 | of the East and the | 东部和东部 |
| 1364 | West and hopefully beyond | 西部和希望之外 |
| 1365 | sensitive anything sensitive | 敏感 任何敏感 |
| 1366 | sensitive industries I do think | 我认为,敏感的行业 |
| 1367 | AI has has accelerated | 大赦国际加快了 |
| 1368 | some of the discussions | 一些讨论 |
| 1369 | and collaboration so | 合作与协作 |
| 1370 | overall I'm optimistic | 总的来说我很乐观 |
| 1371 | and I know | 和我知道,我爱 |
| 1372 | about you she | 关于你,她 |
| 1373 | agrees with this about | 同意这一点: |
| 1374 | the opportunities and | 机会和机会 |
| 1375 | excitement for an AI | 对AI的兴奋 |
| 1376 | and also to encourage people | 和鼓励人们 |
| 1377 | to be more open to | 更开放 |
| 1378 | to share and explore | 共享和探索 |
| 1379 | new markets a new | 新市场 新市场 |
| 1380 | application beyond | 应用程序超出 |
| 1381 | national borders so | 国家边界 |
| 1382 | anyway just want to give | 无论如何,只是想给 |
| 1383 | that comment on behalf of | 代表 |
| 1384 | Martin and Marshall | 马丁和马歇尔 |
| 1385 | would you like to add some | 您想要添加一些 |
| 1386 | more closing remarks | 更多结束语 |
| 1387 | I was thinking more in the | 我当时在想 |
| 1388 | failure modes from your last | 上一个失败模式 |
| 1389 | question I had two or three | 我问了两三个问题 |
| 1390 | examples that might be | 实例 |
| 1391 | interesting first I would | 很有趣,首先我会 |
| 1392 | simply comment it's been a | 只是评论一下而已 |
| 1393 | long-standing information | 长期信息 |
| 1394 | technology practice that | 技术实践 |
| 1395 | indeed you really do need | 您确实需要 |
| 1396 | I've emphasized the point | 我强调过这一点 |
| 1397 | that we've made earlier | 我们之前做的 |
| 1398 | have a strategy don't just | 有个策略 不只是 |
| 1399 | adopt the technology without | 采用不使用 |
| 1400 | a goal in mind have a | 目标有一个 |
| 1401 | strategy and apply it to | 战略及其应用 |
| 1402 | If you don't think about the longer term, | 如果你不考虑长远, |
| 1403 | you might actually get some very different results, | 你可能会得到一些非常不同的结果, |
| 1404 | whether this is the human labor level or at the firm level. | 无论是人类劳动还是企业劳动。 |
| 1405 | Often folks think, well, | 人们常常会想 |
| 1406 | what about job safety with respect to artificial intelligence? | 人工智能的就业安全怎么办? |
| 1407 | We had that interesting example where it increases | 我们有一个有趣的例子 它增加 |
| 1408 | the productivity of call center workers, | 呼叫中心的工人的生产力, |
| 1409 | 14 percent or 34 percent. | 14%或34%。 |
| 1410 | The rubric had sometimes been, | 标题有时是, |
| 1411 | you may not be replaced by AI, | 你可能不会被AI取代, |
| 1412 | but you may be replaced by someone using AI. | 但你可能会被使用AI的人取代 |
| 1413 | Well, that may or may not be true. | 说不定是真的 |
| 1414 | You have to look at the long-term trends in that. | 你必须看看这方面的长期趋势。 |
| 1415 | If you go back to some of the data from Kevin Dai, | 如果你回到Kevin Dai的一些数据 |
| 1416 | the call center jobs are disappearing at an extraordinary rate. | 呼叫中心的工作正在以惊人的速度消失。 |
| 1417 | So simply the use of AI alone will not protect those jobs. | 因此,仅仅使用人工智能并不能保护这些工作。 |
| 1418 | I have a couple of recommendations maybe in some of | 我有一些建议 也许在一些 |
| 1419 | the other remarks on what you might want to do going forward. | 关于你今后可能要做什么的其他评论。 |
| 1420 | If you wanted to ask later about | 如果你想问以后 |
| 1421 | looking for ways to identify future opportunities or what things are happening, | 寻找查明未来机会或情况的方法, |
| 1422 | there may be some things to follow up. | 可能有些事情要跟进 |
| 1423 | There's another one also on decision rights. | 还有另一个关于决定权的 |
| 1424 | Let me give you a fascinating example. | 让我给你一个有趣的例子。 |
| 1425 | Consider three different cases from radiology. | 考虑三个不同的病例 与放射学。 |
| 1426 | In the identification and accurate diagnosis of disease based on chest x-rays, | 在根据胸部X光鉴定和准确诊断疾病时, |
| 1427 | consider three cases. | 考虑三起案件。 |
| 1428 | Expert doctors diagnosing by themselves, | 专家医生自己诊断, |
| 1429 | artificial intelligence diagnosing by itself, | 人工智能诊断本身, |
| 1430 | and doctors plus AI diagnosing together. | 医生和人工智能一起诊断 |
| 1431 | Which of those three cases do you think was most accurate? | 你觉得这三件案子中哪件最准确? |
| 1432 | By a show of hands, I don't know. | 通过举手,我不知道。 |
| 1433 | How many of you think the doctors alone were the most accurate? | 你们中有多少人认为医生是最准确的? |
| 1434 | I'm not seeing any hands on that one. | 我可没看到有人碰那个 |
| 1435 | All right. | 好吧,我们走。 |
| 1436 | One in there. | 一个在那里。 |
| 1437 | How many of you think that the AI alone was the most accurate? | 你们之中有多少人认为AI是最准确的? |
| 1438 | I know one there. | 我知道一个在那里。 |
| 1439 | All right. | 好吧,我们走。 |
| 1440 | How many of you think the doctors plus the AI were the most accurate? | 你们中有多少人认为医生加上人工智能是最准确的? |
| 1441 | A majority of hands. | 多数手. |
| 1442 | The statistics were fascinating. | 统计数据令人着迷。 |
| 1443 | In fact, the doctors were 74 percent accurate. | 事实上,医生的准确度是74%。 |
| 1444 | The doctors plus AI were 76 percent accurate, | 医生和人工智能的准确度是76% |
| 1445 | and the AI by itself, 92 percent accurate. | 和AI本身,92%准确。 |
| 1446 | It was more accurate on its own than it was with the doctors | 自己比医生更准确 |
| 1447 | because they were not willing to give up their prior biases. | 因为他们不愿意放弃他们之前的偏见。 |
| 1448 | They weren't willing to accept that perhaps this technology, | 他们不愿意接受这个技术 |
| 1449 | by looking across millions of cases that they themselves wouldn't have had time to do | 看着成百上千万个案子 他们自己没有时间去做 |
| 1450 | and learn from each of them, could actually be more accurate than they themselves. | 从他们中学习, 可能比他们自己更准确。 |
| 1451 | This raises an organizational design question. | 这就提出了一个组织设计问题。 |
| 1452 | When will you cede decision rights and under what terms | 何时放弃决定权? |
| 1453 | will you cede decision rights to an AI and pattern recognition for important cases? | 你会把决定权让给AI 和模式化的承认重要案件? |
| 1454 | It also then raises the question, if you do cede decision rights, who's liable? | 这也引起了一个问题,如果你让出决定权,谁要承担责任? |
| 1455 | Is it the programmer? | 是程序员吗? |
| 1456 | Is it the party using the AI? | 是利用人工智能的派对吗? |
| 1457 | Or is the party acting on that decision? | 还是该党正在根据这项决定采取行动? |
| 1458 | Those are going to be some longer term issues that may be identifying failure modes | 这些将是一些长期问题,可能正在找出失败模式 |
| 1459 | in the long run that I think absolutely need to be thought through more carefully. | 从长远来看,我认为绝对需要经过更仔细的思考。 |
| 1460 | I'll stop there. | 我会停在那里。 |
| 1461 | If folks want some questions on ideas of what to do going forward, | 如果人们想问一些关于下一步行动的想法, |
| 1462 | I'm happy to go into some of those as well. | 我也很高兴能加入其中的一些 |
| 1463 | But let's take some questions from the audience. | 但是,让我们从观众那里来一些问题. |
| 1464 | OK, Martin, would you add some more remarks on this issue or other issue? | 好吧,马丁,你会补充一些 关于这个问题或其他问题的评论吗? |
| 1465 | Yeah, I think people, I just keep it very short because I think many audience want to ask questions. | 是的,我想人们,我只是 保持非常简短 因为我认为很多观众 想提问。 |
| 1466 | So a common mistake I see many business has today is that when they buy this, | 所以,我今天看到很多生意的一个共同错误是 当他们买下这个, |
| 1467 | they made huge investment on AI, especially after DeepSeek. | 他们为AI做了大量投资,特别是在DeepSeek之后. |
| 1468 | They think, OK, we need to own this. | 他们认为,好吧,我们需要拥有这个。 |
| 1469 | So let's build this with hardware and software and open source together | 我们用硬件、软件和开源软件来构建这个 |
| 1470 | so that we have AI in our company. | 所以我们有AI在我们的公司。 |
| 1471 | I think inspired by Professor, while asking the graph, I saw a better approach | 我觉得受到教授的启发, 一边问图表, 一边我看到了更好的方法 |
| 1472 | is to use the API from service providers instead of building everything in-house | 使用服务提供商提供的API,而不是在内部建造所有设备 |
| 1473 | because the whole industry is moving really rapidly. | 因为整个产业正在迅速发展。 |
| 1474 | So in order to enjoy the benefit of this whole improvement from the latest model. | 所以为了从最新的模型中享受到这个整体改进的好处. |
| 1475 | So you don't want to build everything in-house, | 所以你不想把一切都建在内部 |
| 1476 | unless you have a very security concerns for your data, for anything. | 除非你对数据有非常安全的顾虑 任何事情 |
| 1477 | But you still want to you don't want to make a huge investment just to buy hardware | 但你还是想不想大投资 只为了买硬件 |
| 1478 | because many businesses I see today, they did this in March this year. | 因为我今天看到很多生意 他们今年3月就这么做了 |
| 1479 | And in June, they realized, OK, they can their machines are already out of date. | 在6月份 他们意识到 他们的机器已经过时了 |
| 1480 | And plus, they don't have enough people to optimize it. | 此外,他们没有足够的人来优化它。 |
| 1481 | They don't have a very good application scenario. | 他们没有很好的应用方案。 |
| 1482 | So basically, it's a wasted investment. | 因此,基本上,这是一个浪费的投资。 |
| 1483 | So I think that's a common mistake I see today in business, applying AI. | 因此,我认为这是一个常见的错误,我今天在商业中看到,应用AI. |
| 1484 | So just keep in mind, API is still a better approach to leverage. | 所以记住,API仍然是更好的杠杆方法。 |
| 1485 | So, well, I really like what Marshall just mentioned about this | 我很喜欢马歇尔刚才提到的 |
| 1486 | diagnosis, accuracy, things definitely like really pretty like mind blowing. | 诊断,准确性, 事情绝对像真的漂亮 像心灵吹。 |
| 1487 | So it's actually interesting. | 所以这其实很有趣。 |
| 1488 | So I I went to some like MD defense in a medical school, | 所以我去了医学院的MD防御 |
| 1489 | like local medical school recently. | 就像最近当地的医学院 |
| 1490 | To my surprise, you know, like when this school like medical students, | 令我惊讶的是,你知道, 就像当这所学校 像医学生, |
| 1491 | by the way, Katherine is also MD, PhD also. | 凯瑟琳也是博士博士 |
| 1492 | So when this like, you know, potential physician chose department | 所以当这样,你知道, 潜在的医生选择部门 |
| 1493 | these days because they do rotations, they do rotations like with different departments. | 这些天,他们做轮换, 他们做轮换,就像与不同的部门。 |
| 1494 | And then when they chose the department to like, you know, | 然后当他们选择部门喜欢,你知道, |
| 1495 | like like to work really like one of the most like influential | 喜欢工作 真正像一个 最像有影响力的 |
| 1496 | factors these days is how soon his or her job is going to be replaced by AI. | 这些天的因素是,他或她的工作将很快被大赦国际取代。 |
| 1497 | I think that's like really like what, you know, Marshall mentioned, | 我觉得这就像什么,你知道,马歇尔提到, |
| 1498 | if like I work on diagnosis, like the AI can do much better. | 如果像我做诊断, 像AI可以做得更好。 |
| 1499 | If like with my input is only make the decision even worse. | 如果我的投入只会让决定更糟 |
| 1500 | Really, you know, what do you feel? | 你感觉如何? |
| 1501 | So I guess coming back to Patti's question, I would really call it failure. | 所以我想回到帕蒂的问题, 我真的会称之为失败。 |
| 1502 | Failure is maybe like a little harsh to the ear, but really like the question, | 失败对耳朵来说也许有点苛刻 但其实就像问题一样 |
| 1503 | another different perspective is what really like AI is good at | 另一种不同的观点是,真正像AI的东西是好的 |
| 1504 | or not that good at or maybe bad at at doing something. | 或没有那么好,或 可能不好做的东西。 |
| 1505 | So coming back to like, you know, my story is like I talk about back in 2020, 2021 | 所以回到喜欢,你知道,我的故事 就像我在2020年,2021年谈论 |
| 1506 | in the VC market for biotechnology, at least AI for drug. | 在生物技术的《维也纳公约》市场,至少是药品的AI。 |
| 1507 | Development is really hot. | 发展真的很热。 |
| 1508 | But even soft banks, they withdrew from that area, that direction | 但即使是软银行,他们也退出了那个区域,那个方向 |
| 1509 | around two, three years ago. | 两三年前 |
| 1510 | Really, like, again, I wouldn't call them like failure. | 真的,像,再次, 我不会把他们称为失败。 |
| 1511 | But many of these companies we look at, | 但我们所看到的很多公司 |
| 1512 | market companies we look at back in 2022, they already disappeared already. | 我们回顾2022年的市场公司 他们已经消失了 |
| 1513 | So I feel like one of these things like when we talk like today, | 所以我觉得像今天这样说话 |
| 1514 | we talk about the like AI enabled industrial revolution, whatever. | 我们谈论的像AI 使工业革命,不管。 |
| 1515 | So really like who is going to behind the wheel? | 所以,真的像谁要去 背后的车轮? |
| 1516 | Who is going to be in the driver's seat for this? | 谁会坐驾驶座? |
| 1517 | I feel like really like the data like really matters | 我觉得数据很重要 |
| 1518 | because like those many companies, Katherine and I, we look at like AI | 因为像很多公司 凯瑟琳和我 我们看着像AI |
| 1519 | drug companies, the difference. | 药物公司,区别。 |
| 1520 | I mean, you all use like like Marshall mentioned, the neural network | 我的意思是,你们都使用 像马歇尔提到,神经网络 |
| 1521 | transformer framework, the technology like fundamentals are pretty much the same. | 变压器框架,基础技术等技术基本相同. |
| 1522 | But really, who owns the most data for training really matters a lot | 但是,真的,谁拥有 最大的数据培训 真的很重要 |
| 1523 | because a lot of these companies, for example, to predict the drug, | 因为很多这些公司,例如, 预测药物, |
| 1524 | like how this drug would work from clinical trials. | 像是临床试验的结果 |
| 1525 | But what data we use to train this model, we really don't have like | 但是我们用来训练这个模型的数据 我们真的没有 |
| 1526 | because even for this investment cycle, for clinical phase, | 因为即使是在这个投资周期, 对于临床阶段, |
| 1527 | the average for new drug is like 10 years, | 新药的平均值是10年 |
| 1528 | 10 billion US dollars input. | 100亿美元投入. |
| 1529 | And numerous candidates, you came with one new drug. | 和众多的候选人, 你带来了一种新的药物。 |
| 1530 | So we even don't really like complete one circle. | 所以,我们甚至不 真正喜欢完整的一圈。 |
| 1531 | 因此,这与AlphaGoat不同。 | |
| 1532 | Remember, we talk about the Go game, right? | 记住,我们讨论"走"游戏,对吧? |
| 1533 | Why AI is so good at playing Go? | 为什么AI这么擅长玩Go? |
| 1534 | Because, you know, by this after this move, | 因为,你知道,在这个动作之后, |
| 1535 | eventually you're going to win or lose the game. | 最终你会赢或输 |
| 1536 | That is the definite deterministic almost outcome. | 这无疑是决定性的几乎结果。 |
| 1537 | With which you can train the model perfectly. | 你可以完美地训练模型。 |
| 1538 | You do not have such thing for clinical application, at least for my business | 你没有这种临床应用,至少我的生意 |
| 1539 | or my disciplines, that's still the case. | 或者我的学科, 仍然是这样。 |
| 1540 | I think really to us to think about who owns the critical data, | 我觉得对我们来说 真正值得考虑的是 关键数据是谁拥有的 |
| 1541 | it's going to matter so much. | 这会很重要。 |
| 1542 | That's like one thing I after I heard from Marshall's comments. | 这就像我听到马歇尔的评论后听到的一件事。 |
| 1543 | I'm just going to jump in because, you know, it's many of the people from overseas | 我只是要跳进去 因为,你知道, 这是很多人 来自海外 |
| 1544 | are using into the morning time. | 正在使用到上午的时间。 |
| 1545 | And I know it's also maybe late for those in Beijing or Asia. | 我知道对于北京或亚洲的人来说 也太迟了 |
| 1546 | But I just because I had, you know, Helen and I had opened the webinar | 但我只是因为我有,你知道, 海伦和我已经打开了网络。 |
| 1547 | about bridging the East and West. | 关于沟通东西方 |
| 1548 | So I really just want to ask each panelist to just say in one sentence, | 所以我想请每位小组成员 一句话就说 |
| 1549 | how do you think in AI with all this opportunities and challenges? | 在AI里你如何看待这些机会和挑战? |
| 1550 | And how would you think that this could be a bridge between | 你怎么会认为这能成为 |
| 1551 | particularly China and America, but East and West? | 特别是中国和美国,但是东西方? |
| 1552 | And to give us an optimistic and constructive comment | 并给我们一个乐观和建设性的评论 |
| 1553 | about bridging, as we all are here together. | 关于桥,因为我们都在这里。 |
| 1554 | There's actually 90 of us online and in person, which is an amazing | 其实我们有90个人在线和亲身经历 这真是太棒了 |
| 1555 | gathering from from around the world. | 来自世界各地 |
| 1556 | So I just would like to take advantage of this time for each panelist to say one line. | 因此,我只想利用这一次,让每位专题小组成员说出一行话。 |
| 1557 | Again, how do you think the East and the West can cooperate for AI? | 再说一遍,你觉得东西方如何为AI合作? |
| 1558 | Thank you. | 谢谢 |
| 1559 | OK, Professor Yin, I'll get started. | 好的,殷教授,我开始吧 |
| 1560 | We have more in common than being different with AI. | 我们有更多的共同点 而不是与AI不同。 |
| 1561 | I feel we can work together in a much better way. | 我觉得我们可以以更好的方式合作。 |
| 1562 | That's great. | 赞曰. |
| 1563 | And I think at least for the next 10 years, | 我觉得至少在未来十年里 |
| 1564 | I can't I can't say for too long, but for the next 10 years, | 我不能说太久, 但接下来的十年, |
| 1565 | China has the application and data and the US has the technology. | 中国拥有应用和数据,美国拥有技术. |
| 1566 | So I think that that'll be great. | 所以,我认为这将是伟大的。 |
| 1567 | I think there is a lot China and the US can work together on using | 我认为有很多中国和美国可以合作 |
| 1568 | and advancing AI and then with the manufacturing capability | 和推进AI,然后具有制造能力 |
| 1569 | and supply chain in China and the AGI | 中国的供应链和AGI |
| 1570 | and large language model development in the US. | 美国的大型语言模型开发。 |
| 1571 | Perhaps we can bridge the two to make AI | 也许我们可以把两者连接起来 做人工智能 |
| 1572 | more effective and more applicable in real world. | 更有效和更适用于现实世界。 |
| 1573 | Yeah, just very brief from one side. | 是啊,只是很简短 从一边。 |
| 1574 | So artificial intelligence doesn't differentiate countries. | 因此人工智能并不区分国家. |
| 1575 | It's shared by all human being. | 由一切众生共知. |
| 1576 | So we have to collaborate. | 所以我们必须合作。 |
| 1577 | So so I like that. | 所以我喜欢这样 |
| 1578 | So I would pose that as a question. | 所以,我会提出一个问题。 |
| 1579 | The first one is, you know, in the United States, can we replace our leadership with AI? | 第一个是,你知道,在美国,我们能用AI取代我们的领导吗? |
| 1580 | I'm very frustrated with our current leadership. | 我对我们目前的领导感到非常沮丧. |
| 1581 | Perhaps that would be an interesting, better solution. | 也许这将是一个有趣、更好的解决办法。 |
| 1582 | But the real answer, I think, is one I think is fairly interesting | 但真正的答案,我认为, 是一个我认为相当有趣的 |
| 1583 | and relates to one of the other questions asked. | 并涉及所提出的其他问题之一。 |
| 1584 | I think it was Karen Chow. | 我想是周凯伦 |
| 1585 | I actually think AI ethics is one of the most important things | 其实我觉得AI道德是最重要的事情之一 |
| 1586 | where we have a shared interest. | 我们有着共同的利益 |
| 1587 | And I think Eastern and Western views on AI ethics are going to be essential | 我认为东方和西方对AI伦理学的看法将是至关重要的 |
| 1588 | and need to be derived from first principles and added to these AI systems. | 并且需要从第一原则出发,并加入这些AI系统。 |
| 1589 | The reason I think this is so important is actually pretty simple. | 我认为这很重要的原因其实很简单。 |
| 1590 | We are on the verge of AI systems that are able to design themselves. | 我们濒临AI系统 能够自己设计。 |
| 1591 | When that happens, any constraint that we seek to put into an AI | 一旦发生这种情况,我们试图施加于AI的任何限制 |
| 1592 | can be designed out by an AI in its next generation. | 可以由下一代AI设计. |
| 1593 | So hard coded constraints aren't going to work. | 这么硬的密码约束是行不通的。 |
| 1594 | So how do we get them to exercise self-control? | 那么,我们如何让他们 行使自我控制? |
| 1595 | It's going to be ethics, but that's going to have to be derived by first principles. | 这会是道德的,但必须先从原则中推导出来. |
| 1596 | So I would argue that one of the best things that we could do | 所以我认为我们最好做的一件事 |
| 1597 | is to collaborate around ethics. | 是围绕道德进行协作。 |
| 1598 | And I think we would all win if we can do that. | 如果我们能做到这一点,我们都会赢。 |
| 1599 | That's a great point, especially on the eve of the realization of AGI. | 这是一个伟大的点,特别是在实现AGI的前夕. |
| 1600 | And I think currently people are predicting faster and faster, | 我认为目前人们预测的速度越来越快 |
| 1601 | you know, | 你知道吗? |
| 1602 | realization and that's kind of so. | 认识,就是这样。 |
| 1603 | If no other questions, then we'll conclude this event. | 如果没有其他问题,那么我们就结束这次活动. |
| 1604 | Thank you, everyone, for attending and hope we can see you again at our next event. | 感谢各位出席,希望我们能在下次活动中再次见到你。 |
| 1605 | And just to say, so as I say, we're doing AI for the global citizen | 顺便说一句,我说,我们正在做AI 全球公民。 |
| 1606 | and the future topics may be about sustainability and Shantai Wenming, | 未来的主题可能是可持续性和山泰文明 |
| 1607 | also in the arts and education. | 艺术和教育领域也是如此。 |
| 1608 | And so we welcome all members around the world to join on Pressing Matters | 因此,我们欢迎全世界所有成员加入《新闻事务》 |
| 1609 | using AI to improve, improve for a better world. | 利用人工智能改善、改善世界。 |
| 1610 | Thank you. | 谢谢 |
该视频共有字幕 1610 条。解锁更多字幕为会员功能，请移动到 价格