New York Times Tech Reporter, Cade Metz | The Human Podcast #14
发布时间 2022-07-12 19:39:48 来源
摘要
Cade Metz is a New York Times Technology Reporter, and the author of ‘Genius Makers: The Mavericks Who Brought A.I. to Google, Facebook, and the World’. Cade previously wrote for WIRED magazine.The Human Podcast is a new show that explores the lives and stories of a wide range of individuals. New episodes are released every week - subscribe to stay notified.WATCH - FILMED IN PERSON:https://www.youtube.com/channel/UC29JGmLUfv5eUeKzv3cRXGwSOCIAL:Twitter - https://twitter.com/heyhumanpodcastInstagram - https://www.instagram.com/heythehumanpodcastGUEST:Cade’s NYT Page: https://www.nytimes.com/by/cade-metzCade’s Twitter: https://twitter.com/cademetzCade’s Book ‘Genius Makers: The Mavericks Who Brought A.I. to Google, Facebook, and the World’: https://www.amazon.co.uk/Genius-Makers-Facebook-Artificial-Intelligence/dp/184794213XCade's WIRED Page: https://www.wired.com/author/cade-metz/ORDER OF CONVERSATION:0:00 - Intro0:28 - Interest In Technology/Engineers4:19 - IBM's Deep Blue v Gary Kasparov7:08 - Early Career Highlights8:44 - iPhone Launch11:17 - DeepMind's AlphaGo v Lee Sedol24:12 - AI Progress Since AlphaGo28:16 - Cade's Book, 'Genius Makers: The Mavericks Who Brought A.I. to Google, Facebook, and the World'34:30 - Interest In AI35:28 - AI Visionaries37:03 - Why Do You Enjoy Journalism?38:37 - Difficulties Of Journalism?40:50 - How Do You Feel About The Future?41:55 - Future Plans / Second BookMUSIC:Music from Uppbeat (free for Creators!): https://uppbeat.io/t/hartzmann/space-journeyLicense code: 4Y9SSRT4HAKSYWPFGUEST SUGGESTIONS / FEEDBACK: Know anyone who may like to speak about their life? Or have any feedback? Just message [email protected]
GPT-4正在为你翻译摘要中......
中英文字稿 
Hey everyone, welcome to today's episode with Cade Mats. Cade's been a technology reporter for over 30 years and he's currently at the New York Times. He released a fantastic book last year on the history and the story of AI and how I came to be. I hope you enjoy it, chat. If you do, please hit subscribe. It really helps the channel if you could do that. Thank you for watching.
大家好，欢迎收看今天和Cade Mats一起的节目。Cade是一位科技记者，已有30多年的从业经验，现在在纽约时报工作。他去年出版了一本很棒的书，讲述了AI的历史和故事，以及它如何出现。希望大家喜欢这个话题。如果您喜欢，请订阅我们，这对我们的频道非常有帮助。感谢您的观看。
You spent the first few years of your career as a playwright before switching over to tech journalism. What inspired that switch? Well, it wasn't necessarily a switch. I've always had this dual interest, a dual background. I was an English major in college. My senior thesis was a novella, but my father was an engineer. During college, I had an internship at IBM. He was a career IBMer. My father a programmer. Through him, I had a scholarship and part of the scholarship was a summer internship as a programmer.
你的职业生涯最初几年是当剧作家，后来转向了科技新闻。是什么激发了这个变化呢？嗯，不一定是转换。我一直拥有这种双重兴趣，双重背景。我在大学时是英语专业的。我的毕业论文是一篇小说，但是我父亲是一名工程师。在大学时，我在IBM有过实习。他的职业是IBM员工。我父亲是一名程序员。通过他，我获得了奖学金，其中一部分是作为程序员的暑期实习。
In college, I took programming courses as well and math and science courses. I always had this dual interest. I had a particular interest in writing about engineers. I felt like engineers and researchers were underrepresented, even in tech journalism. Tech stories in the mainstream press are typically about the entrepreneurs, the people building the companies as opposed to people really building the technology. I always felt like that was unexplored territory or underrepresented territory. I always had an interest in writing about not just the technology, but the people really building the technology.
在大学期间，我修了编程、数学和科学课程，一直对这两个领域感兴趣。我尤其对工程师的写作有浓厚兴趣。我觉得工程师和研究人员在科技新闻中未被充分报道。主流媒体中的科技报道通常是关于企业家和创业者，而不是那些真正建造技术的人。我一直认为这是未被发掘或者未被充分报道的领域。我一直热衷于写作，不仅关注技术本身，更关注那些真正构建技术的人。
The first thing that comes to mind is Steve Wozniak in the whole Apple story. You must have been thinking about him a bit then as opposed to the guy who has really talked about Steve Jobs. That's a good example. I always felt like engineers were as interesting as anybody else. The trope, the assumption is that engineers are somehow boring and uninteresting, but I think the opposite is true. My father, like I said, was an engineer in a career IBM or IBM. He had these amazing stories about the people that he worked with.
我想到的第一件事情是整个苹果故事中的史蒂夫·沃兹尼亚克。你当时一定有点想他，而不是那个真正谈论史蒂夫·乔布斯的人。这是个很好的例子。我总觉得工程师和其他人一样有趣。大家普遍认为工程师有点乏味、不有趣，但我认为相反。我父亲就是在IBM这家公司做工程师。他有很多关于工作伙伴的惊人故事。
He worked on, among other things, the UPC projects at IBM, so the universal product code, the bar code that is now on all our groceries. This is how we buy our groceries. He worked on that original project. He helped test the system. He had these incredible stories, not only about the people who built the system and first trained it up, but got him Joe Woodland, first envisioned this technology when he was on the beach back in the 50s. Also about how the technology affected people as it was pushed out into the world.
他曾在IBM的UPC项目中工作，其中包括现在出现在我们所有杂货上的通用产品代码和条形码。这就是我们购买杂货的方式。他参与了这个原始项目的开发，帮助测试系统。他有很多了不起的故事，不仅是有关构建系统的人和他们最初如何训练系统，还与乔伍德兰一起想象了这项技术，当时他们是在50年代的海滩上。还有关于技术如何在推向世界时影响人们的故事。
In the early 70s, as they began testing and deploying the system, there were literally protests against IBM and the system. People who saw it as the sign of the beast from revelations come to life. Why would you push this out into the world? Those types of stories, they were entertaining. They were amusing, but they also shown a light on the way technology can affect people and hit certain parts of their psyche in ways you might not expect. I've always been interested in exploring those types of things.
在70年代初期，当他们开始测试和部署系统时，实际上有人对IBM和这个系统进行了抗议。人们视其为启示录中标志着邪恶来临的迹象。你为什么要将其推向世界呢？这些故事非常有趣，很有趣，但它们也照亮了技术如何影响人们并以你未曾预料的方式打击他们心理的某些方面。我一直对探索这些类型的事情感兴趣。
Speaking of IBM, one thing I wanted to ask you was about, I saw online, which said that you reported on your reporting drawing in the 1997 match between IBM's Deep Blue and Gary Kasproff. Can you tell me a little bit about that time and the level of excitement and what it felt like to be in reporting on that? Right. I was in New York at the time and that event was put together in the city.
说到IBM，我想问你的一件事情是关于你在1997年IBM的深蓝和Gary Kasproff之间的比赛中进行报道的事情。我在网上看到了一些信息。你能向我介绍一下那个时候的情况以及当时的兴奋程度和进行报道的感受吗？是的，那个时候我在纽约，那个事件是在城市里组织的。
It was a remarkable event, in part because of the attention that was on it, in part because of the way it played out. How surprised Kasproff was by the machine that he was playing. That's what was most interesting is that it caught him, let alone everyone else, by surprise. It was interesting years later when I covered the Go match between AlphaGo, the system built by DeepMind, the AR Lab in London that had been purchased by Google, and Lee Cedal, who was one of the best Go players of the last decade. That played out in a similar way, but just on an even larger scale.
这是一个非凡的事件，部分原因是因为人们对它的关注，部分原因是因为它的发展方式。卡斯普罗夫对他正在玩的机器感到惊讶，最有趣的是，它让他，甚至其他所有人，都感到意外。几年后，当我报道阿尔法围棋（由DeepMind、被谷歌收购的伦敦AR实验室所打造的系统）和李世石的围棋对弈时，也是如此。尽管规模更大，但进展方式相似。
It was, it helped that I was able to contrast and compare that with what had happened years earlier in New York with Deep Blue and Gary Kasproff. I know the level of access you had at the AlphaGo match where you know you at the venue. Were you at the venue during the IBM match against Kasproff or were you kind of reporting on the news you heard? Well, why are you exactly? I was at the venue as well. It was in a hotel on the west side of Manhattan.
之所以帮助我，是因为我能够将这一点与几年前在纽约与 Deep Blue 和 Gary Kasproff 的比赛相比较。我知道你在 AlphaGo 比赛中拥有的接触水平，在比赛现场。你是否在 IBM 对战 Kasproff 的比赛现场，还是你在报道你听到的新闻？呃，你到底是怎么样的呢？我也在现场。比赛是在曼哈顿西区的一家酒店举行的。
It was, you know, I remember you had sort of stadium seating for a large TV or movie screen that would show the match and you had commentators, chess experts. We provided commentary in real time as if it was sporting the bit. But, though I was there for every match back then, and again, it was nice to compare and contrast that with what happened years later in Seoul with AlphaGo and Lee Seedol.
你知道的，我记得你那时有一个类似体育场的座位区，可以看到大电视或电影屏幕，播放比赛，还有象棋专家作为解说员。我们实时提供比赛评论，就像是体育比赛一样。尽管我当时参加了每一场比赛，但再次回想那时与后来在首尔所发生的AlphaGo和李世石之间的对比，感觉很不错。
Yes, so we'll come on to AlphaGo in a minute. I've got a few questions about that. But you're at the New York Times at the moment.
好的，那么我们马上就要谈谈AlphaGo了。我有几个关于它的问题。但是现在你在纽约时报工作。
After staying with, you know, PC Magazine for a number of years, what was some kind of highlights during their 15 or so years you were there? Did any specific inventions come out or were there any kind of moments that made you kind of stand out as being the most exciting during your career?
在与《PC Magazine》呆了几年之后，你觉得在那15年左右的时间里最引人注目的亮点是什么？是否出现了特定的发明，或者是在你职业生涯中有什么时刻让你表现得最令人兴奋？
Well, I think that the cast broth match was certainly one of those moments. You know, beyond that, nothing really, really compared to what would happen in later years with the type of technology we're going to talk about here, meaning AI technology, nothing happened for years and years.
嗯，我认为那个铸造清汤比赛肯定是其中一个重要时刻。你知道，在那之后，其实没有什么能和我们现在要谈论的技术 - 意思是AI技术 - 发生的事情相比。多年来，一直没有什么变化。
During that period, it was sort of the late 90s, early 2000s. It was what people often called an AI winter. There wasn't the interest or the funding going into the field that you would see in later years.
那个时期，大概是90年代末和2000年年初。人们通常称之为人工智能的冬季。当时没有像后来那样的兴趣或资金投入这个领域。
The field really, really changed in the late 2000s, early 2010s, right? There was an inflection point where the field really took off in some unexpected ways. And that's what is really been interesting.
这个领域在2000年代末、2010年代初真的很大变化，对吧？有一个转折点，这个领域真的以一些意想不到的方式开始崛起。这就是真正有趣的地方。
I'm not sure where you were exactly at the time, but I know off to PC Magazine you worked at the register for a few years. I'm just thinking of one event, like the iPhone launched in 2007. Was that an interesting moment? Where were you working then just to kind of think about another event outside of the AI space?
我不确定你当时确切在哪里，但我知道你在PC Magazine工作过几年，并且在收银处工作过。我只是想到了一个事件，就像2007年iPhone的推出一样。那是一个有趣的时刻吗？为了考虑AI领域之外的另一个事件，你当时在哪里工作？
Sure. I was at PC Magazine the time. I was at the event where Steve Jobs unveiled it. These are carefully designed and orchestrated press events. It's almost like a concert. It's at this place called the Rock Concert. It's called the Marconi Center in San Francisco.
没问题。当时我在PC Magazine。我参加了斯蒂夫·乔布斯揭幕的活动。这些是经过精心设计和策划的新闻发布会。它几乎就像一场音乐会。在一个叫做“摇滚音乐会馆”的地方举行。它在旧金山的马可尼中心举行。
They're carefully chosen songs playing as everyone is gathering and waiting for Steve Jobs to come on stage. His speech, while unforgettable, in hindsight, you realize just how carefully he's trying to pull the strings.
大家正在聚集等待史蒂夫·乔布斯上台，播放的歌曲都是精心挑选的。回顾起来，他的讲话令人难忘，但你意识到他是如何小心翼翼地操纵局面的。
He comes out and says, I'm going to unveil three devices today. One's a camera of a camera while unforgettable, right? You know, in hindsight, you realize just how carefully, you know, he's trying to pull the strings, right?
他站出来说，我今天要揭示三款设备。其中一个是一个不可错过的相机，对吧？回想起来，你会意识到他有多么小心翼翼地试图掌控一切。
He comes out and he says, I'm gonna unveil three devices today, right? One's a phone, you know, one's a camera and one's an internet connection device, right? And then the big reveal is, it's all one device, right? It's all, you know, and all the Apple Faithful are there.
他走出来说：“今天我要推出三款设备，是吧？一款是手机，一款是相机，还有一款是上网设备。”然后他露出惊天大秘密：“其实它们全部合在一个设备里面，没错。”所有的苹果追随者都在场。
You know, those are events are interesting as sort of, as a window into the way Steve Jobs would operate and kind of pull the strings on the general public and the Apple Faithful. But it's far less interesting to me than an event like the Go Match in Korea, because something is playing out there in real time, where no one quite knows what's going to happen.
你知道，那些事件非常有趣，就像一个窗口，可以让我们看到 Steve Jobs 是如何在大众和苹果粉丝中拉动其背后的线。但对我来说，它远不如在韩国进行的围棋比赛这样的事件有趣，因为在那里发生着一些实时的东西，没有人完全知道会发生什么。
And you see the technology affect people in unexpected ways. That's far more interesting to me.
而且你会看到技术以意想不到的方式影响人们。对我来说，这更有趣。
So after you're at the register, you move to Wired Magazine. Yeah, and good to talk about AlphaGo now, because I think you're at Wired when the AlphaGo challenge match happened.
所以当你到达收银台后，你就可以移步到《连线》杂志了。是的，现在谈论AlphaGo很好，因为我想你是在《连线》当AlphaGo挑战赛发生时所在的。
And this is how I first kind of heard of you. I saw you on the documentary, this brilliant documentary I'd recommend everyone watches, you know, produced by DeepMind on the AlphaGo Match. So yeah, just to say a little bit more about it and I've said a bit already, but can you just talk a bit more about what that experience was like your week or so there was like and how it felt, excitement.
这就是我第一次听说你的方式。我在纪录片上看到了你，这是一部非常出色的纪录片，我推荐每个人都去看一下。这纪录片由DeepMind制作，讲述的是AlphaGo比赛。所以，就多说一点吧，我已经说了一点，不过你能再多说一点你在那里度过的一周是什么感觉，兴奋吗？
What I often tell people is that even though I was just an observer to this match, right, I was not a participant, I was an observer.
我常常告诉人们，尽管我只是这场比赛的观察者，而不是参与者，但我是个观察者。
It was one of the most amazing weeks of my life. The part of it was that the entire country, meaning Korea, South Korea was focused on this match. You would walk out of the streets and people would be gathered outside the four seasons.
这是我人生中最惊奇的一周之一。其中最令我惊叹的是整个国家，也就是韩国，南韩，都集中关注这场比赛。当你走在大街上时，人们都聚集在四季酒店外面。
So to watch these sort of giant television screens, showing the match, it was on the front page of every paper, every day. And as the match would sway back and forth, you know, towards AlphaGo or back towards Lee Sido, you could kind of feel the whole country sway in the same way.
所以观看这些巨大的电视屏幕展示比赛，每天都在每份报纸的头版头条。随着比赛的来回摇摆，你可以感觉到整个国家也跟着摇摆。有时候是支持 AlphaGo，有时候是支持李世石。
When, certainly when Lee Sido lost the second match, you could feel his collective sadness across the country, right, he was a national hero. And it wasn't just about him losing, but I think we all sort of feel this pang when that sort of thing happens. You know, you relate to him because he's human, right?
当李思多输掉第二场比赛时，整个国家都能感受到他的悲伤，他是民族英雄。而且，这不仅仅是因为他输了，而是当这种事情发生时，我们所有人都会感到一种刺痛。你知道，你和他有联系，因为他也是人类，对吧？
And when he is beat by this machine, we all feel that sort of sadness. That was real and it was palpable. And then at the same time, when he came back and won game four, that sort of a relation we all could relate to.
当他被这台机器打败时，我们都感到那种悲伤。那是真实的，而且是可以感觉到的。而当他回来赢得第四场比赛时，我们所有人都能感同身受。
And that was an relation that really reverberated across the whole country. Being there wasn't an amazing, amazing thing. And it was something that maybe people back in the UK or back in the US didn't feel as much, right?
那确实是一种产生了深远影响的关系，对整个国家都有反响。在那里的经历是一件令人惊叹的事情。也许，在英国或美国的人们没有那么强烈的感受到。
Go as a national game in places like Korea and China and Japan in a way it isn't in the UK or in the US. So you really felt an importance being there in the country as that was playing out.
在像韩国、中国和日本这样的地方，围棋是国家游戏，而在英国或美国则不是。所以当你在那些国家玩游戏时，你真的感到自己很重要。
As you say, it was one of the most, you know, kind of interesting exciting weeks of your life. Did that take you by surprise a lot? Or did you go there thinking something, you know, very interesting is going to happen here?
就像你说的那样，这是你生命中最有趣、最令人兴奋的一周之一。那么这是否让你感到非常惊讶？还是在去那之前你就觉得一定会发生很有趣的事情呢？
Well, you know, I did make an effort to go, right? I knew that something was going to happen. It was going to make a good story. I remember pitching to my editor the story and him saying, well, most people assume that Lee Seedle is going to win, right? And I said, yes, but I think what you're going to see is that the machine is going to win this match.
嗯，你知道，我确实努力去了，对吧？我知道会发生什么事。这个故事肯定很有趣。我记得向我的编辑推销这个故事，他说，大多数人都认为李西德尔会赢，对吧？我说，是的，但我认为你会看到这场比赛是机器赢了。
I think what people, you know, go experts are not taking into account is that this system that DeepMond is building has continued to be improved over the month since it last played a match, right? It had beaten the European Go Champion behind closed doors previously. And from those matches, you can sort of, you know, ascertain the level that this system had achieved.
我认为人们，你知道的，专家们没有考虑到这个DeepMind正在建立的系统自上一次比赛以来已经不断改进了一个月，对吧? 它在之前的一场私下比赛中击败了欧洲围棋冠军。从这些比赛中，你可以推断出这个系统达到的水平。
What people were not taking into account is that the, you know, part of the system literally learns from, from repeated play. And that sort of learning aspect of the system was what, you know, a lot of the experts were not, were not acknowledging as they tried to predict what would happen in Korea. I was confident that the machine was going to win.
人们没有考虑到的是，你知道的，这个系统的一部分实际上是通过重复的游戏来学习的。这种学习方面的系统是专家们试图预测韩国将发生什么而未能认识到的。我有信心机器会赢得比赛。
And, you know, I didn't know it was going to win. But that's what I was expecting. And, you know, it played out in ways, certainly that didn't expect. Like I didn't expect the machine to be that dominant. And, you know, I didn't expect that twist at the end when, when Lee Cito, you know, took game four.
而且，你知道的，我并不知道它会赢。但那正是我期望的。而且，你知道的，它的表现方式绝对是我没有预料到的。比如，我没想到机器会那么优势。而且，你知道的，我也没想到李西托在第四场比赛中会出现那样的转折。
I remember, you know, after game three, when effectively, you know, the machine had won the match, right, in one three out of three out of, you know, five games. And my wife said, where are you coming home now? And I said, no, I'm going to stay and see how this plays out. You know, and luckily I did, I did stay because that was one of the most interesting moments when, when Lee Cito came back and won the fourth game.
我记得，在三场比赛结束之后，机器已经赢了整场比赛，赢了五局中的三局。我太太问我回家了吗？我说不，我要留下来看看结果。幸运地是我留下来了，因为那是李·西托反击赢得第四局的最有趣的时刻之一。你知道的。
Yeah, there is a lovely moment in the documentary where you, I think you say you're nearly about to tear up, recalling the moment, you know, where Lee comes back and wins. As you said, there was a shared interest at one point, you know, from, I guess you wanted them to at least get one match, even if you know, because that made all the difference, you know, for making them feel a bit better and stuff. I think he said, we're just winning one match was enough or something, they kind of consoled him a bit.
嗯，在纪录片中有一个美好的时刻，你提到你几乎要哭了，回忆起那个时刻，你知道，那时李打回来赢了。正如你所说，曾经有一个共同的兴趣点，你希望他们至少能打一场比赛，即使你知道，这对于让他们感觉好些来说是至关重要的。我想他说过，“赢一场比赛已经足够了”，他们有点安慰了他。
Oh, absolutely, right? This isn't, you know, this is about more than just, you know, the statistics of that match, right? It's, you know, it's about something more human than that, more important than that. And it's about this larger arc of AI machines and its relationship to us humans.
噢，完全正确吧？你知道的，这不仅仅是那场比赛的数据统计，对吧？它是关于比那更人性化、更重要的东西的。它是关于人工智能机器与我们人类之间的更大趋势及其关系的。
Are there any kind of interest in behind the scenes stories or insights, you know, conversations you had with I've a go experts or people at DeepMind? Yeah, a lot of this is in the piece I eventually published at Wired and in a book, you know, I later wrote about the history of neural networks, which is a key technology that's used in AlphaGo.
你有没有对幕后故事或深度洞察感兴趣，比如说你和某些AlphaGo专家或DeepMind的人士的交流？嗯，我的Wired文章和我后来写的神经网络历史书里提到了很多这方面的内容，神经网络是AlphaGo所使用的关键技术之一。
But, you know, the most amazing moment, and the most amazing character, you know, as far as I was concerned, you know, who was involved in that match was a kind of fun way. And he was the European Go champion who had lost the match to AlphaGo behind closed doors. And, you know, he's not a native English speaker.
但是，你知道的，我觉得那场比赛中最令人惊奇的时刻和最令人惊奇的角色就是一个有趣的人。他是欧洲围棋冠军，在闭门比赛中输给了AlphaGo。他不是母语为英语的人。
But, you know, I ran into him or talked to him in the wake of AlphaGo winning game too, right? Which was sort of the real, like sort of devastating moment for a lot of people. And, you know, he had this wonderfully poetic way of describing the system and the way it played, and the beauty of this system, which had, you know, made this sort of this transcendent move to effectively win that match, a move that David Silver, one of the deep-mind engineers later told me, was a move that a human most likely would never have made, right, according to this machine's calculations, which are based on real games involving human players.
但是，你知道的，我也在AlphaGo赢得比赛后遇到了他或与他交谈，对吧？这是许多人实际上感到沮丧的真正时刻。他有这种奇妙的诗意方式来描述系统和它的游戏方式，以及这个系统的美丽，它做出了一些超越性的举动，有了这些举动，它有效地赢得了比赛。根据这台机器基于涉及人类玩家的真实游戏计算的结果，Deepmind工程师之一David Silver后来告诉我，这是一个人类很可能永远不会做出的举动。
Its calculation was that a human, the chances of human making that move, an expert player were one in 10,000. But, based on the machines, the machine had, based on the games, the machine had effectively played with itself, it decided to make that move anyway. And, the way that Fawnway described that move in the moment was remarkable. He's a neat guy who shows up in that documentary and you talked about as well, and one of the important characters in that piece. It's a great movie.
机器认为人类的走法成功率只有一万分之一，而作为一名专业的玩家，人类也只能在一万次中出现一次成功。但是，根据机器在游戏中的表现，它已经与自己打得不亦乐乎，决定尝试那个走法。当时Fawnway对这个走法的描述非常出色。他是那部纪录片中出现的一个整洁的家伙，你也谈到了他，是那个作品中的重要角色。这是一部很棒的电影。
I was surprised at how effective that documentary was. Yeah, really is a beautiful piece. What was it like for watching the documentary back, seeing yourself in it, seeing the images for when you were there?
我对那部纪录片的效果感到很惊讶。是啊，真的是一件美丽的作品。当你重新看这部纪录片时，看到自己的形象，看到你曾经在那里所经历的情景，感受如何呢？
Well, I thought it did a great job of capturing what it was like to be there. From the small moments with people like Fawnway, to the larger scope of things, and how this whole country reacted to the event, it really captured what it was like to be there.
我觉得它很好地展现了当时的氛围。从与Fawnway这样的人一起度过的小时刻，到整个国家对事件的反应，它真正地展现了当时的情景。
Did much change in your perspective of AI technology and the future from being that event? I mean, you said that you predicted that the machine would most likely win. So after it had finished, did you think much differently about the future AI, or had you kind of expected that to happen? And it was obviously an amazing experience, but nothing had changed much in your minds.
你对AI技术和未来的看法有了很大的变化吗？我的意思是，你曾经预测机器最有可能获胜。所以在比赛结束后，你对未来的AI技术有了很大的不同看法，还是你已经预料到会发生这种情况？显然这是一个令人惊叹的经历，但并没有在你的想法中产生太大的变化。
Well, I mean, I think it was an important moment because a game like that is something we can all understand, we grow up playing games. And when you have a moment like that, when a machine can beat one of the world's top players at a game like that, it's a way that everyone can understand how the technology is progressing.
嗯，我想这是一个重要的时刻，因为像那样的游戏是我们都可以理解的东西，我们从小就玩游戏。当你有像那样的时刻，当一台机器可以在那样的游戏中击败世界顶尖的玩家时，这就是每个人都可以理解科技进步的方式。
And because of that phenomenon, that aspect of the system that learns from repeated play, you can see this technology continuing to advance. Like I said, it's based on a technology called a neural network, which is a mathematical system that literally learns tasks by analyzing data. So that might be analyzing go moves from expert players, or it might be images or sounds, right?
因此，由于那个从不断的游戏中学习的系统方面，你可以看到这项技术不断发展。就像我说的，它基于一种被称为神经网络的技术，它是一种实际上通过分析数据学习任务的数学系统。因此，它可以分析来自专家玩家的围棋棋谱，也可以是图像或声音等等，对吗？
A neural network is what allows Siri to understand the words that we say. It analyzes thousands of hours of spoken words, that learns to recognize the words that we speak. You can feed it, feed a neural network thousands of cat photos and it can learn to identify the patterns that allow it to identify a cat.
神经网络是让Siri能够理解我们所说的话的东西。它分析了成千上万小时的口语，学会了识别我们说的话。你可以给它喂上成千上万张猫的照片，它能够学会识别允许它识别猫的模式。
And you could see a path where this technology could continue to improve certain computer skills, right? So image recognition, speech recognition, translation, that was the next area where this technology really, really improved things. And we continue to see this basic technology improve what scientists call natural language understanding, the ability to understand the languages that you and I and others on our speak, and responding to that.
你能看到这项技术可以继续提高某些电脑技能的方向，对吧？所以图像识别、语音识别、翻译，这些是接下来这项技术真正取得进步的领域。我们继续看到这项基础技术改进科学家所谓的自然语言理解，也就是理解你、我和其他人说话的语言，以及对此做出回应。
So you're seeing systems now that use this same fundamental technology to apply that phenomenon to all sorts of other natural language skills, whether it's question and answer or dialogue, right? You know, we're seeing increasingly, you know, systems that can, you know, move towards carrying on a conversation. We're not there yet, but you can, at the time, you could see the progress in these areas continuing.
那么现在你看到有些系统使用这种基本技术，将该现象应用于其他所有的自然语言技能，无论是问答还是对话，对吧？你知道的，我们越来越多地看到，这些系统可以进行对话。虽然我们还没有完全达到这个程度，但你可以看到这些领域的进展。
And you know, it's something that I moved to the times that I pitched to my editors, right? This idea would continue to show progress in the years to come.
你知道的，这是我向编辑们推荐的时代的一个观点。这个想法将在未来的几年中继续显示进展。
Since that, since the AlphaGo challenge match, we've had some interesting things happen with all the technologies you've just discussed, you know, we had deep-mind with AlphaFold and OpenAI, you know, with a few things, GPT-3 and Dali-2. Can you envision and picture what you think may be maybe the next moment, this is exciting as, you know, for example, the AlphaGo match, is there anything that you think may kind of stand out as a next big milestone that you're excited to see?
自从Alphago挑战比赛以来，我们就经历了一些有趣的技术进展，你知道的，比如DeepMind的AlphaFold和OpenAI的一些东西，比如GPT-3和Dali-2。你能想象并描述一下你认为可能会出现的下一个时刻嘛？这是令人兴奋的，就像Alphago比赛一样。你是否认为有什么可能会脱颖而出成为下一个令人激动的里程碑呢？
Well, I think, you know, at this point, we need to look beyond the games and look beyond the interesting demos towards where this might really change things in real ways. AlphaFold is a good example, right? That technology, again, based on a neural network, fundamentally, you know, change biological research. And, you know, it may help scientists, for instance, when it comes to drug discovery, right? Developing new vaccines and medicines.
嗯，我认为你知道，此时此刻，我们需要超越游戏和有趣的演示，向着这种技术在真实应用中可能真正改变事物的方向看齐。AlphaFold就是一个很好的例子，对吧？这种基于神经网络的技术，从根本上改变了生物研究。你知道的，它可能帮助科学家，比如在药物发现方面，开发新的疫苗和药品。
And, you know, when you think about what OpenAI has done with these, what they call large language models. So, you know, neural networks that can understand language in the way that I described earlier. We need to, you know, really think about where that can be of use. And we're starting to see it be of use in certain ways.
你知道，当你想到OpenAI用所谓的大型语言模型所做的事情时，那些能够以我之前描述的方式理解语言的神经网络。我们需要认真考虑它可以用在哪些地方。我们正在开始看到它在某些方面发挥作用。
Like that sort of technology has now been deployed with software developer, right? It can, it can, in a way, generate pieces of software code that developers can then make use of, right? It's not perfect, right? It can't generate software code to the point where it can replace all coders, but it can help software developers do their job. Generate snippets of code that they can then use and shape and insert into their projects. And that's where it's starting to be useful.
就像现在的技术已经被软件开发人员所采用，是吗？它可以以某种方式生成程序代码片段，供开发人员使用，对吧？不完美，对吧？它不能生成代码到足以替代所有开发人员，但它可以帮助软件开发人员完成工作。生成代码片段使他们可以使用、塑造和插入到他们的项目中。这就是它开始有用的地方。
The next area on the horizon are what, what's, researchers call multi-modal systems. So, it's basically the same sort of model. These sort of large language models also applied to images. So, you have this system produced by OpenAI Presses called Dali. And what you can do is you can ask it to give you a photograph. An image, you can describe the image you want. You can say, I want an image of two cats playing chess and it will generate that image with photorealistic quality.
接下来的研究方向是什么研究人员称为多模式系统。所以，它基本上是相同类型的模型。这些大型语言模型也适用于图像。所以你有这个由OpenAI Presses生产的系统叫Dali。你可以要求它提供一张照片。你可以描述你想要的图像。你可以说我要两只猫下国际象棋的图像，它将生成具有光感传递质量的图像。
It's a remarkable system in a lot of ways. Surprising, entertaining. So, then the question becomes, how is this gonna impact our world in real ways? And that is sort of an open question. But, you know, much as these large language models can help developers build software code, something like Dali can help graphic designers as they're building images, right? You can have the system generate an image that you can then tweak and modify and insert into the work you're doing.
这是一个非常出色的系统，在很多方面令人惊喜且娱乐性十足。于是问题就来了，这会以什么真实的方式影响我们的世界呢？这是一个开放的问题。但是，正如这些大型语言模型可以帮助开发人员构建软件代码一样，像达利这样的系统可以帮助图形设计师构建图像，对吧？你可以让系统生成一张图像，然后对其进行微调和修改，并将其插入到你正在进行的工作中。
That's not the sort of, you know, super intelligent system that a lot of AI researchers have long dreamed of and long claim was on the horizon. These are systems that are best used in tandem with humans. But that's the sort of thing that that time been looking at and thinking about.
这不是那种你知道的，大量人工智能研究人员长期梦想并长期声称即将到来的超级智能系统。这些系统最好与人类一起使用。但那就是我们一直在关注和思考的东西。
That's interesting. So, we get to move on now to talk about the books that you released in the last year, which I've read and there's really brilliant, really interesting, titled Genius Makers, the Mavericks who brought AI to Google, Facebook and the world.
很有趣。所以，我们现在可以继续谈论你去年发布的那些书了。我已经阅读了其中的一些，它们真的很棒，非常有趣，其中一个书名为《天才制造者：将人工智能带到谷歌、Facebook和全世界的闯荡者们》。
So, in first of all, yeah, why did you decide to write this book on this topic?
首先啊，你为什么决定写这本书呢？
Well, I decided to write it after coming back from Korea and that go match. That event where you could see how the technology was affecting people in real time. And you could see the interesting characters behind the technology, Dimas Asabas, the CEO and founder of DeepMind, being one prime example. That's when I resolved to write the book.
嗯，我决定在韩国回来参加那场围棋比赛之后开始写这本书。在那个事件中，你可以看到技术如何实时影响人们。你也可以看到技术背后有趣的人物，比如DeepMind的CEO和创始人Dimas Asabas。那时我就决定写这本书了。
But as I started to pull together a pitch, to make, to publishers, I became even more interested with a guy named Jeff Hinton, who had worked on that idea, the idea of a neural network, since the early 70s. He embraced this idea at a moment when most people thought it would never work. And as I talked to him more and got to know him, the book, and it's focused shifted a lot. He became the central character, like the one human thread, you know, through the history of this idea, the idea of a neural network. And that really became the heart of the book, right? Any good book, any good story, is about people. And he became the central person in this book.
但是当我开始准备一份介绍材料去给出版商时，我对一个名为杰夫·辛顿的人产生了更浓厚的兴趣。他从70年代初开始就一直在从事神经网络的研究，这个想法让人们一直认为它行不通。而当我跟他聊得越多，越是了解他，这本书的定位也随之改变了很多。他成了这个想法发展历史中唯一的人类纽带。这真的成为了书的灵魂。任何好的书、好的故事都是关于人的。他成了这本书的中心人物。
I mean, just a bit intrigued about the process I've read in that book. So, how much of the stories in the book were, did you experience first-hand, did you speak much to these main characters in the book and about these stories in Catcher From Them, or was it stuff you'd found from research? How did you go about finding out all the content in the book?
我是说，关于我在那本书中读到的那个过程，我稍微有些感兴趣。那本书中的故事有多少是你亲身经历过的？你和书中的主角有多少交流？这些故事都是你通过研究找到的吗？你是如何获取那本书中所有的内容的呢？
It's based on dozens and dozens and dozens of interviews. Almost everyone who is mentioned in the book, I spoke to, you know, I mean, there are a few exceptions, but I spoke to almost everyone. And, you know, a lot of those stories were revealed for the first time in the book, or through my own reporting.
这本书是基于几十个访谈的。几乎书中提到的每个人，我都跟他们交谈过，就是有少数例外。这许多故事在书中首次被揭示，或通过我的调查报告中呈现。
You know, a couple of the big moments, you know, I had written about before, whether it's the Go-Match or a couple of other events. But most of that is original research. You know, I do go back in time, right? This is an old idea. The idea of a neural network was first proposed in the 50s. And so there's some historical research there.
你知道，有几个重要时刻，我之前已经写过了，不管是 Go-Match 还是其他几个事件。但是大部分是原创研究。你知道，我会回溯过去，这是一个古老的想法。神经网络的概念最早是在 50 年代提出的，所以这里有一些历史研究。
Another interesting character is a guy named Frank Rosenblatt, who was a psychologist in the 1950s, and someone who really championed this idea then. He died in the early 70s. So there's some historical research involved. But a lot of that also is firsthand, right? Through people like Jeff Hedden, who have worked in this area since the 70s, people like Jan LeCoon, who is now the head of research at Metta, formerly Facebook, and other people who have long worked on this idea.
还有一个有趣的人物是名叫弗兰克·罗森布拉特的人，他是20世纪50年代的心理学家，也是当时真正支持并推崇这个想法的人。他在70年代初去世了，所以这里涉及到一些历史研究。但很多情况下也是第一手资料，对吧？比如像杰夫·赫登这样的人，从70年代以来就一直在这个领域工作，像珍·勒昆这样的人，现在是前Facebook研究负责人的Metta的研究主管，还有其他一些长期从事这个想法的人。
You must have learned a huge amount of AI whilst writing the book. Were there any or many big questions or ideas that you've noticed you've changed your mind on that you wouldn't have thought previously before you read the book and I know what you did?
你写作这本书的时候肯定学了大量的人工智能。你有没有发现有很多大问题或思想是你在读这本书之前没有想到过的，现在你改变了你的想法？我知道你做了什么。
Well, I think one thing that's that has been interesting and I continue to think about is, and I don't think this is widely understood among the general public. AI is an aspirational field, right? In the 1950s, when that term was coined, artificial intelligence, it was coined by this kind of John McCarthy, he was a Dartmouth College professor. He had, he pulled together this summer conference at Dartmouth in 1956, and brought together light-minded people who were interested in this new field. He decided to call it artificial intelligence.
我觉得有一件有趣的事情让我一直在思考，但我不认为这在普通大众中被广泛理解。人工智能是一个富有抱负的领域，对吧？在上世纪50年代，当这个术语被创造出来时，即“人工智能”，它是由这位约翰·麦卡锡创造的，他是达特茅斯学院的教授。他在1956年组织了这个夏季会议，聚集了志同道合、对这个新领域感兴趣的人。他决定称其为人工智能。
And these academics were sure that it wouldn't take long for them to create machines that could do what the human brain could do. Some predicted that within 10 years you would have a system that could beat the world's best players at chess or that could develop its own mathematical theorem. Well, you know, that didn't happen. You know, the chess piece didn't happen until 1997, as we discussed.
这些学者们确信，他们不需要很长时间就能制造出能够像人类大脑一样工作的机器。有些人预测10年内，会有一个系统能够击败世界顶级的国际象棋选手，或者自己发展出数学定理。但是，你知道吗？这并没有发生。就像我们讨论的那样，国际象棋不是在1997年才实现。
You know, artificial intelligence from the beginning was a misnomer, right? There were building technology that was nowhere close to intelligent. But they were sure that they could do this. And that's a theme throughout the history of AI and that you continue to see, right? It's very much an aspirational field. And I think that that's something that people outside of the field don't quite understand.
你知道吗，人工智能从一开始就是一个错误的命名，对吧？它们建造的技术远远不能算是智能。但他们确信自己能够做到这一点。这是人工智能历史中一直存在的主题，而且至今仍然存在。这是非常有抱负的领域。我认为这是领域外的人不太理解的事情。
Is that the kind of main reason you're so interested in AI, that it's just this really aspirational field that you're trying to solve such difficult issues? Is that the reason you're particularly interested in that as opposed to, you know, all the other areas of technology?
你这么感兴趣的主要原因是不是因为AI是一个充满野心的领域，你正在尝试解决如此困难的问题呢？与其他技术领域相比，这是你特别感兴趣的原因吗？
I think that that's part of it. But part of what I want to do with the book or with my report at the time is is give people a realistic view of this, right? Because it's so aspirational, people often, people in the field, often talk about the technology in ways that exaggerate its powers. And I think part of my role is to give people a real view of what's going on here and show them exactly what the technology can do and what it can't. And explain this aspirational nature and help them understand that what some people are saying isn't necessarily true.
我认为那是其中一部分。但是我想在我的书或报告中做的一部分是给人们提供现实的观点，对吗？因为这个领域太有抱负了，人们经常夸大技术的力量。我认为我角色的一部分就是给人们真正的观点，展示技术能做什么，不能做什么，并解释这个抱负的本质，帮助他们理解有些人说的话并不一定是真实的。
One thing I've seen you say is that a lot of the visionaries in the field they continued going going with this when everyone else didn't believe in them. They believed in AI, everyone thought they were crazy but they continued. You've spoke to some of these people, maybe all of these people that I'm thinking of. What do you think it is about them that you think enabled them to have this belief and to pursue it in the face of doubt?
我听你说过的一件事是，若干先见之明的人在这个领域的时候，当所有人都不信任他们时，仍然坚持不懈地追求AI，每个人都认为他们是疯子，但他们仍在继续。你曾经和他们中的一些人，或者是我想到的所有这些人交谈过。你认为是什么让他们有这个信仰，并在面对怀疑时继续追求？
I admire that quality in anyone, right? Having a firm belief in what you're doing and sticking with it, even in the face of doubts from others, right? That's an abdiment quality and it's a quality that's at the heart of so many great stories. And it's a quality you see in Jeff Hinton, in particular. He believes very strongly what he believes and he's willing to pursue that. At the same time, he, like maybe some others, is grounded in a way and he sees the limits of the current technology and he's willing to talk about the limits. But he is also very firm in his beliefs and he's willing to talk about that. It's an abdiment quality and it's absolutely something that's tracked in me.
我觉得任何人都应该欣赏这种品质，对吧？就是坚信自己所做的事，并且即使面对他人的怀疑也会坚持下去。这是一种关键的品质，也是许多伟大故事的核心。而且，这种品质在杰夫·辛顿身上尤为明显。他非常坚信自己的理念，并且愿意为此不断努力。同时，他也深思熟虑，在某种程度上比其他人更加踏实，他看到了当前技术的局限性，愿意谈论这些限制。但他也非常坚定地坚持自己的信仰，并且愿意谈论这一点。这是一种关键的品质，而且它在我身上也可以看出来。
That's always pretty much booked together. So you're just moving on to some final questions. I kind of want to return to your career a bit more broadly, a bit more general. I mean, you beat a tech journalist for a number of years. What's your enjoy about the work that you do?
那个总是相当预订在一起。所以你现在只是会进入一些最后的问题。我有点想回到你的职业上，稍微更广泛、更普遍一些。我的意思是，你当了多年科技记者。你对自己所做的工作有什么喜爱之处呢？
Really it's about talking to people, right? That's the most interesting thing. That's really how the job works. You talk to one person and you get through the end of a good conversation and you say, who else should I talk to? They recommend a few people you talk to them and the cycle continues.
其实，工作就是和人交谈，是吗？这是最有趣的事情。这就是这份工作的实际操作。你和一个人交谈，你们举行了愉快的谈话，然后你会问：我还应该和谁交谈？他们会推荐给你一些人，你再和他们交谈，周而复始。
That's really what's most interesting. It's talking to and getting to know and getting to understand new people. And then helping others do the same, right? Taking what you have learned, taking these conversations you've had and turning them into a story or a book that others can appreciate.
这真的是最有趣的地方。就是跟新朋友聊天，互相了解，更加理解他们。然后去帮助别人做同样的事情，对吧？把你所学到的东西，把你所进行的这些对话转化成别人可以欣赏的故事或书籍。
It feels like with all the technology work and reporting that you do, there are always technology stories that are deeply intertwined with people as opposed to just technology on its own. I guess that's the thing that you find most interesting.
用母语中文说话的感觉就像是在做所有的技术工作和报告时，总有与人紧密相连的技术故事，而不仅仅是技术本身。我想这就是你最感兴趣的事情。
Well, absolutely. I think that technology itself is certainly interesting. But it's even more interesting when you think about how this affects us all. And so, again, I think that not all technology writing does this, but it should. It should look at the intersection of technology and people in society. That is where the technology becomes most important.
嗯，当然。我认为技术本身确实很有趣。但当你想到这影响我们所有人时，就更有趣了。所以，我再次认为不是所有的技术写作都这样，但它应该是这样的。它应该关注技术和人类社会的交叉点。那才是技术最重要的地方。
Are there any aspects of being a journalist that stand out particularly difficult or the most challenging part of the job?
作为一名记者，有哪些方面是特别困难或最具挑战性的部分？
Well, I mean, that piece, I mean, these are complicated issues, complicated in a lot of ways, right? Technically complicated, the ramifications of the technology are complicated. And you have to take in all that complication and then find a way to distill it into something that anyone and everyone can understand. That's always hard.
嗯，我的意思是，那个问题，我是说，这些是非常复杂的问题，从很多方面来看都是复杂的，对吧？在技术上很复杂，技术的影响很复杂。你必须把所有这些复杂的因素考虑进去，然后找到一种方法来将其精简成任何人都能理解的东西。这总是很难的。
And you have to continue to struggle to do that. Because, again, ultimately, it's about imparting this knowledge to anyone. So it's a hard thing to do to distill that down into its essence into something that anyone can grasp and enjoy. But that's the job. I think you suddenly did that in the book, so that's for sure.
你必须继续努力才能做到这一点。因为最终，这是要将知识传授给任何人的事。所以将其精华蒸馏成任何人都可以理解和享受的东西是一件困难的事情。但那就是我们的工作。我认为你在书中成功做到了这一点，那是肯定的。
And I've seen you make the point a few times as well that one thing you really try and do is remain objective in your reporting and not to take sides and give your opinion too much. Is that difficult or is this the kind of way you enjoy working? Do you ever wish you could give a bit more of your opinion and your take on things or?
我也看到你很多次强调，你在报道中非常努力保持客观性，不偏袒任何一方，也不会过多表达你自己的看法。这个难吗？还是说你很享受这种方式工作？有没有想过更多地表达你自己的观点和想法？
It is the way that my mind works. And it's the way I really want to approach things. It's one of the reasons I'm at the times. Because the way I work aligns with the organization. And I think that that's really what's most effective. And I think that's one of the reasons the book is effective. Because it doesn't take anyone's side. It looks at the entire landscape. And it's the way that everyone, all the information they need to understand what is truly going on and they can make their own decisions.
这就是我的思维方式，也是我处理事情的方式。这也是我来到时代杂志的其中一个原因，因为我和这个机构的工作方式是相符的。我认为这种做法是最有效的。这也是这本书有效的原因之一，因为它不站在任何人的立场上，而是全面地看待整个局面。这样每个人都可以得到他们所需要的所有信息，从而理解实际情况并做出自己的决策。
After spending so many years reporting and thinking and writing about technology, how do you feel about the future and where we're headed?
在花费了这么多年的时间来报道、思考和写作科技方面，您对未来和我们的发展方向有何感受？
Well, there are some dark parts to our future. I'll tell you that. And a lot of them relate to AI. We haven't talked much about this. These large language models and these multi-modal systems we talked about. They're essentially generating fake content. They're generating their own images. They're generating texts. They're generating tweets, blog posts.
嗯，我们的未来会有一些黑暗的部分，我得告诉你。很多都涉及到人工智能。我们还没有谈论太多这一点。我们提到的那些大型语言模型和多模式系统，实质上是在生成虚假内容。它们可以生成自己的图像、文本、推文和博客文章。
We're going to reach a point where it's going to be hard to tell whether texts or images or sounds were created by a human or they were created by a machine. We are going to have to have a skepticism about everything that we see and hear online. And I wonder if we're as humans capable of that sort of skepticism. I think it's going to be a real shift.
我们将会到达一个难以区分文本、图像或声音是由人类还是机器创造的点。我们将必须对我们在网上所看到和听到的一切持怀疑态度。我想知道，作为人类，我们是否有能力拥有这种怀疑精神。我认为这将是一个真正的转变。
Do you have any plans for the future with your work and anything exciting coming up?
你们有没有未来的工作计划，有没有什么令人兴奋的事情即将发生？
Well, I'm going to continue to cover this. This area at the time, and all sorts of areas. I'm what they call the emerging technologies reporter. So any technology that has been beginning to come to the fore is a potential subject for my stories. Maybe another book at some point. Nothing eminent yet, but I'd like to do another.
好的，我打算继续关注这个领域，包括现在和各种其他领域。我被称为新兴技术记者。因此，任何开始出现的技术都有可能成为我报道的主题。也许会在某个时候写另一本书。现在还没有什么明确的计划，但我很想再写一本。
What hate to stop it from?
什么可以阻止它的仇恨呢？
Any thoughts on what the second book may be about?
你有没想过第二本书可能是关于什么的呢？
That's true, yeah. But it might be nice to get outside the AI field and do something entirely different.
那是真的，没错。但可能尝试离开人工智能领域，去做些完全不同的事情会很好。
That's great. Well, look forward to it.
太好了。好的，期待着。
Okay, thanks so much for speaking. It was great to hear a bit about your story, but also talk about some AI and take stuff. Appreciate your time.
好的，非常感谢你说话。听到你的故事和一些关于人工智能和现有技术的讨论真的很棒。谢谢你花时间和我交流。
Enjoy it. Thank you.
享受它吧，谢谢。
I hope you enjoyed this conversation. If you enjoy this human podcast, please don't forget to subscribe in.
我希望你喜欢这次对话。如果你喜欢这个人类播客，请别忘了订阅。
I hope to see you soon.
我希望能尽快见到你。