本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
这里的大图景是，2017年时，许多人认为我们生活在历史的终结，不会再有重大的地缘政治变动，联合国只需写几封措辞严厉的信就能阻止人们这么做。
The very big picture here is in 2017, a lot of people believed we lived at the end of history, that there was no more geopolitical movement of significance to happen, and that United Nations could write mean letters to dissuade people from Yeah.
扩张主义行动。
Expansionist actions.
愤世嫉俗者是对的。
The cynics were right.
那些说‘不’的愤世嫉俗者。
The cynics who said, no.
战争仍然是当下现实的一部分。
War is still a thing of the present.
你不能等到战争爆发后再开始研制炸弹，还指望它能起到威慑作用。
You can't start working on bombs after the war has started and expect to have any deterrent impact.
你只会沦为参与战争的一分子，而不是阻止战争。
You're just gonna be part of fighting wars instead of preventing them.
我一生中一直听人谈论的这些事，突然间都发生了。
All of this stuff that's been talked about my whole life is, like, suddenly happening.
AI显然正在发生。
AI is clearly happening.
我们相信加密货币将会兴起。
We believe crypto will happen.
你认为我们是否正进入一个可能前所未有的二三十年时期，
Do you think we're entering into this maybe unprecedented twenty, thirty year period of
没错。
its Absolutely.
让我给你谈谈我乐观的理由。
I'll give you my thesis for optimism.
在2017年，大多数人认为我们正生活在历史的终结。
In 2017, most people believed we are living at the end of history.
大国冲突已经结束，联合国只要发出措辞严厉的信件就足以制止一切。
Great power conflict was over, the United Nations could write strongly worded letters, and that would be enough.
然后俄罗斯入侵了乌克兰。
Then Russia invaded Ukraine.
帕尔默·卢基预见到了这一点。
Palmer Luckey saw it coming.
在入侵发生前，他曾试图向泽连斯基出售安多里尔的监控塔，泽连斯基曾在《连线》杂志上读到过这些设备，并希望将它们部署在乌克兰边境。
Before the invasion, he tried to sell Andoril's surveillance towers to Zelensky, who had read about them in Wired and wanted them on Ukraine's border.
美国国务院否决了这笔交易。
The State Department killed the deal.
他们说俄罗斯不会入侵。
They said Russia wasn't going to invade.
帕尔默于2017年与25人共同创立了安多里尔公司。
Palmer started Andoril in 2017 with 25 people.
彭博社称它是科技界最具争议的公司。
Bloomberg called it the most controversial company in tech.
《连线》杂志将他评为硅谷最糟糕的人。
Wired named him the worst person in Silicon Valley.
如今，这家公司拥有超过7000名员工和25款产品。
Today, the company has over 7,000 employees and 25 products.
他为SOCOM和海军陆战队制定了Drove Defense的正式项目。
He set the program of record for Drove Defense across SOCOM and the Marine Corps.
在安德鲁之前，帕尔默十几岁时创立了Oculus，以30亿美元将其出售给Facebook，并目睹Meta投入600亿美元实现他最初构想的愿景。
Before Andrew, Palmer founded Oculus as a teenager, sold it to Facebook for 3,000,000,000 and watched Meta pour 60,000,000,000 into the vision he started.
他也是比特币的早期持有者。
He was also an early Bitcoin holder.
他曾用8000个比特币购买了一部三星手机，差点成为Coinbase的首位商户客户。
He bought a Samsung phone for 8,000 Bitcoin and almost became Coinbase's first merchant customer.
这段对话是在一场16人活动中现场录制的。
This conversation was recorded live at a 16.
它探讨了如何大规模制造硬件、美国在国防上投入过多却收效甚微的原因，以及帕尔默为何仍对未来的三十年持乐观态度。
It covers what it takes to build hardware at scale, why The US is spending too much and getting too little on defense, and why Palmer is still optimistic about the next thirty years.
克里斯·迪克森采访了Andoril和Oculus VR的创始人帕尔默·拉基。
Chris Dixon speaks with Palmer Luckey, founder of Andriel and Oculus VR.
感谢你的到来。
Thanks for being here.
我刚看了你的《乔·罗根访谈》，每个人都应该看看。
I just watched your Joe Rogan interview, which everyone should watch.
里面谈到了会说话的鹦鹉、UFO，还有中国的UFO。
There's talking parrots, UFOs, China UFOs again.
中国UFO和中国，有很多有趣的内容。
China UFOs and China, just a lot of interesting stuff.
你打造的一款AR头盔。
A AR helmet that you built.
这段视频的结尾简直是极客的天堂，你们一定要去看看。
It's like nerd nirvana at the end this video, so you guys will have to watch it.
然后又是UFO。
And then UFOs again.
是的。
Yeah.
提了三四次。
It's three or four times.
而且还有乔·罗根总要提到的外星海洋，关于人类进化或者某种提升理论。
And there's, like, the alien sea this Joe Rogan has to always talk about the alien sea human evolution or some or the uplift theory.
是这个吗？
Was that it?
他谈到了提升理论，即把非意识物种提升到或超越人类意识的水平。
Talked about uplift theory, taking non sentient species and bringing them up to or past the point of human consciousness.
这简直就是老派乔·罗根的风格，完全深陷每一个奇思妙想的洞穴里。
A much it's a much it's like old style Joe Rogan of just, like, deep in every rabbit hole.
我们当时就沉浸其中。
We were going for it.
这太惊人了。
It was incredible.
如果你有机会，我推荐去听乔·罗根的节目，但有个问题：在接下来至少48小时内，你所有的通讯工具都会完全无法使用，因为像收件箱、短信、邮件这些，都依赖于信息量有限这个前提。
I recommend going on Joe Rogan if you get the chance, but the one problem is that all of your messaging tools become completely unusable for a period of at least forty eight hours because a lot of things like inboxes, your SMS, email, it relies to some degree on there being a limited quantity.
是的。
Yeah.
是的。
Yeah.
所以如果你突然收到上千条消息，来自你认识的每一个人，都说‘兄弟，你在乔·罗根那期太酷了’，那中间的其他消息就全被淹没了。
And so if you have over a thousand messages come in from everyone you've ever known saying, bro, so cool on Joe Rogan, then everything else in between.
然后他们说：‘帕尔默，你今天必须完成安全培训，否则你会失去你的权限。’
And then they're like, Palmer, you must finish your security training today or you will lose your clearance.
然后他们说：‘帕尔默，你收到消息了吗？’
And they're like, Palmer, did you get the message?
没有。
No.
没有。
No.
等一下。
Hang on.
兄弟，兄弟，兄弟，兄弟，兄弟。
Bro, bro, bro, bro, bro.
在你的收件箱里。
On your inbox.
没错。
That's right.
乔·罗根，太棒了。
Joe Rogan, that's awesome.
不管怎样，我很幸运在你还在Oculus的时候见过你。
Anyway, so I was lucky enough to meet you back in your Oculus days.
没错。
That's right.
我记得那时马克、安德烈斯和我飞过去，看到了墙上的二维码演示。
Which was I think I remember Mark, Andres, and I flew down and saw I remember the demo with the QR codes on the wall.
我稍微转一下。
I'm gonna turn a little bit.
墙上的二维码。
The QR codes on the wall.
我可没那么会作弊。
I can't cheat that well.
我不是百老汇演员。
I'm not like a Broadway actor.
是的，克里斯。
Yes, Chris.
我记得。
I remember.
所以
So
所以那个墙上有二维码的演示。
So so the demo with the QR codes on the wall.
对。
Yep.
对吧？
Right?
是的。
Yeah.
显而易见。
Obviously.
对于那些不知道的人，当时第一个开发者套件虽然有追踪功能，但必须在外部放置盒子。
And like for those who don't know, like at the time, like the first developer kit, you'd had tracking, but you had to have the boxes outside.
其实，第一个开发套件根本没有空间追踪功能。
Well, what it was is the first development kit didn't have any spatial tracking.
你无法追踪在空间中的移动。
You couldn't track movement through space.
它只能追踪你的视线方向。
It was just where you were looking.
哦，所以它只有三个自由度
Oh, so it was three three three degrees
或者类似的情况。
off or whatever.
没错。
Exactly.
所以我们当时在开发一个六自由度系统。
And so then we were developing a six off system.
那就是后来的Oculus Rift DK2。
That was what became the Oculus Rift DK two.
没错。
That's right.
这实际上是第一个现代的
Like, that was the real first modern
那是外部追踪。
That was the outside tracking.
对吗？
Is that right?
我们也是。
We we too.
我们实际上正在试验不同的系统。
We're experimenting with different systems, actually.
我的意思是，你提到的是一个内向外追踪系统。
I mean, that was an inside out system that you're referring to.
我们最终还是为DK2甚至CV1使用了外向内追踪系统。
We ended up still using an outside in system for the DK two, and then even for the CV one.
我们花了好几年时间才让内向外追踪足够稳定，能够走出我们向投资者展示的房间，真正进入人们的客厅。
It took us years to get inside out tracking robust enough that we could take it out of a room that we would show investors and actually get into people's living rooms.
因为我们之前稍微谈过这一点，但边缘情况才是最致命的。
Because we were talking a little bit about this earlier, but the edge cases are what kills you.
在实验室里让它运行起来很容易。
It's easy to make it work in a lab.
非常、非常、非常容易。
Very, very, very easy.
也许不算真的容易，但相对而言还是挺容易的。
Not maybe not really easy, but it's relatively easy.
真正困难的是当你把它放到真实的房间里时。
What's hard is when you put it into a real room.
真实的房间里面有玻璃。
Real rooms have glass in them.
它们里面还有反射。
They have reflections in them.
还有其他人会走来走去。
They have other people walk through.
现在你场景的很大一部分是移动的。
Now a large chunk of your scene is moving.
你必须过滤掉这些移动的部分。
You have to reject that.
对于大部分情况来说，这听起来很容易，但那些边缘情况呢？
And that sounds easy for the bulk of it, but what about the things on the fringes?
像窗帘这样轻微晃动的东西怎么办？
What about things like curtains that move a little bit?
或者，你扫描了你的场景，然后回来，下次发现它只是微微移动了一下。
Or, like, you you scan your scene, and then you come back, and next time it's moved just a little bit.
还有风扇，吊扇，这是内向外追踪的噩梦。
Also, fans, ceiling fans, the death of inside out tracking.
拒绝所有这些因素成了一项艰巨的挑战，但最终我们还是做到了。
Just rejecting all of those things becomes Herculean challenge, and we eventually got it working.
那就是Oculus Quest。
That was Oculus Quest.
那你们的光学系统呢？比如镜头？
And did you the optics, like, the lenses?
没错。
That's right.
屏幕也是一个大问题。
The screens were a big issue.
你可以谈谈
You can talk
关于这个，比如对那些不了解的人，我记得只有三星生产过那种东西？
about, like, for those who don't know, like, I believe only Samsung made what was it?
那个系统
The system
哦，情况甚至更糟。
Oh, it was even worse.
三星是唯一一家能生产它们的公司。
Samsung was the only company that was sort of able to make them.
所以我们用的三星显示屏，比它们智能手机里的还要先进。
So we were using Samsung displays that were more advanced than was in their smartphones.
我们试图把那些原本只是工程样品和原型的东西，投入大规模生产。
We were trying to take things that really were engineering samples and prototypes, and we were trying to put them into full scale production.
是的。
Yeah.
因此，那些生产线的良品率非常低。
And so we had very low yield on those lines.
我们必须对每块显示屏进行测试，并剔除有坏点的。
We had to test every display and reject ones that had dead pixels.
在VR中，你根本无法容忍任何一个坏点。
And in VR, you really can't get away with one dead pixel.
别提亮像素了。
Forget bright pixels.
甚至连坏点都存在。
There's even dead pixels.
这是因为它们太明显了。
It's because it's so obviously apparent.
它就在你眼前。
It's right in your face.
它就在你眼前。
It's right in your face.
它被大大地拉伸了。
It's stretched out in a big way.
而且，还有你双眼之间的不匹配问题。
And then also it's the mismatch between your eyes.
我们曾经试验过一种方案：如果你有一颗坏点，就同时关闭另一只眼睛中对应的像素。
One of the schemes we experimented with actually was if you had a dead pixel, killing an equivalent pixel in your other eye.
至少这样看起来是对称的。
So at least it's symmetric.
这实际上让情况好很多了。
It actually made it much better.
但问题在于不对称。
But it was the asymmetry.
你有没有见过那种交叉眼谜图，当你交叉视线时，就能看出不同之处？
You've ever seen those cross eye puzzles where you cross your eyes, you could see what's different?
有坏点的感觉就是这样。
That's what it's like having a dead pixel.
你一直在看到这种不对称的失配图像，所以那些点显得格外突出。
You're constantly seeing this asymmetric mismatch image, and so those dots stand out so much.
总之，我们最终没有这么做，因为客户仍然不喜欢死点，但那确实是个艰难的时期。
Anyway, we ended up not doing that because it turns out customers still disliked dead pixels, but it was a tricky time.
对我来说，这让我明白我没有在硬件上投入足够，也让我认识到硬件为什么这么难做。
To me, was a lesson of I had not invested in hardware, and to me, it was a lesson of why hardware is so hard.
当时实际上只有一家公司，三星，除非你下个几亿美元的订单，否则他们根本不会考虑，这不就是你为什么和三星做这笔交易的原因吗？
There was literally one company, Samsung, and it, like, unless it was, like, you tell me, a couple $100,000,000 order, they weren't even gonna in fact, isn't that why you did the Samsung deal?
三星Gear VR。
Samsung Gear VR.
对。
Correct.
Gear VR的初衷是
Gear VR was to
当时无论出多少钱，他们都不同意为我们制造所需的显示屏，但我们和他们达成了一项协议，作为三星Gear VR产品的软件和硬件设计方。
get There the was really no amount of money that would convince them to do what we need to make our display, but we did a deal with them to basically be the software and hardware designers behind the Samsung Gear VR product.
所以，无论你是购买了Gear VR，还是随手机免费获得，都是Oculus促成了这件事。
So if you bought Gear VR or if you got it for free with your phone, it was Oculus that made that happen.
我们卖出了超过一千万台，但它们只便宜了3美元。
And we sold over 10,000,000 units, which was But they were $3 off.
它们便宜了6美元。
They're $6 off.
对吧？
Right?
没错。
They were Correct.
它们是
They were
便宜了3美元。
$3 off.
我从来就不喜欢便宜3美元，因为我觉得这根本不是真正的优惠。
I always didn't like $3 off because I felt like it wasn't the real thing.
我同意，这根本不是真正的优惠。
I agree it was not the real thing.
看吧。
Look.
人们忘了，我是个爱好者，15岁的时候就开始制作VR头盔。
People forget that I was a hobbyist who started building VR headsets as a 15 year old.
是的。
Yeah.
然后我让你们所有人都给了我一大笔钱来实现这个目标。
And then I got all of you guys to give me a bunch of money to do it.
这是一门生意。
It's a business.
所以我早就知道，减3美元根本不可能真正推动这场VR革命，但它确实让你们愿意掏钱，而且我们从DK1到DK2的进展非常快。
And so I knew that three Doff was not going to actually work to build this VR revolution, but it got you guys to give us money and like the but like we got from DK one to DK two very quickly.
还有个好消息。
And here's the good news.
DK1时期，我们卖出了大约5.5万台，当时成为了历史上最畅销的VR头盔。
DK one, we sold about 55,000 units, which made it the best selling VR headset in history at the time.
顺便说一下，之前销量最好的是美国的飞利浦Scuba。
By the way, the previous bestseller was the Philips Scuba in The US.
在日本则是Takara DinoVisor。
It was the Takara DinoVisor in Japan.
是的。
Yeah.
但这两款产品加起来，总销量正好是5万台。
But combined, that same design sold exactly 50,000 units.
所以我们特意卖了5.5万台，让DK1成为销量冠军。
So the 55, we specifically sold that much so that the DK one would be the best seller.
而我们卖出了25万台DK2。
And we sold 250,000 DK twos.
顺便说一句，我并不是在批评你。
I'm not criticizing you, by the way.
我只是觉得，你以前确实该批评一下。
I'm just like, you once should criticize.
不。
No.
但一旦你体验了第六代，你就会觉得，哇，这才是真正的体验，因为你可以在房间里自由走动。
But once you I just meant once you experience the sixth off, you're like, wow, that's the real thing because you can, like, walk around the room.
没错。
That's right.
所有那些著名的DNA。
All the infamous DNA.
尽快实现这一点至关重要。
It was critical to get to that as fast as possible.
我的意思是，我认为在DK一代的时候，我们就知道自己是在借时间过日子。
I mean, I would argue the DK one days, we knew we were living on borrowed time.
我们生活在一个人们看到它后，只能得到五分钟演示的世界，然后他们会说，哦，这简直太不可思议了。
We were living in this world where people would see it, and they'd get a five minute demo, and they'd say, oh, this is absolutely incredible.
但如果没有六代追踪技术，会让人感到眩晕和恶心。
But without six off tracking, it will make them sick, and it will make them nauseous.
网络晕动症由多种因素引起，但主要机制是当内耳感受到的与眼睛看到的信息不匹配时，身体会认为这种不匹配一定是某种毒素干扰了周围神经系统造成的。
Cybersickness is caused by a lot of things, but the primary mechanism is that when you have what your inner ear feels mismatched with what your eye sees, your body believes that mismatch must be due to a poison that is interfering with your peripheral nervous system.
那么，当你中毒时，你需要做什么？
And so what do you need to do when you're poisoned?
排出毒素。
Expel the poison.
是的。
Yeah.
所以我们知道，我们当时是在借来的时光里。
And so we knew that we were, you know, on borrowed time.
这是一款出色的演示硬件，但我们知道，必须在更多人发现真相之前，尽快实现高质量的六自由度追踪。
It was a great piece of demo hardware, but we knew we needed to get to really good six off tracking before too many people figured out what was going on.
我们做到了。
And we did.
对。
Yeah.
我们做到了。
And we did.
我能讲讲我们刚才提到的销售故事吗？
Can I tell the selling story we just talked about?
请讲。
Please do.
好的。
Okay.
我不确定这是否已经公开。
I don't if it's public.
好吧。
Alright.
这是我的帕尔默加密货币故事，我之前
So here's my Palmer crypto story, which is I hadn't
甚至没想过这一点
even thought about this
这么多年了。
for so many years.
有一次，大概是2013年，我们打了个电话，他说他想让整个商店只接受比特币。
Well, one was we had a call once where in like 2013, he's like, I wanted to make the store all Bitcoin.
对吧？
Right?
我当时说，我太爱你了。
And I was like, I love you.
这太棒了。
This is awesome.
事实上，你曾说过，你是Coinbase的第一个客户。
And in fact, were almost, you said, the first customer of Coinbase.
我昨天刚和布莱恩·阿姆斯特朗在一起。
So I literally was with Brian Armstrong yesterday.
是的。
Yeah.
我们翻阅了我的邮件，以唤起记忆。
And we were going back through my emails to refresh our memory.
我有一封我和布莱恩·阿姆斯特朗之间的邮件往来记录。
And I have an email chain of me with Brian Armstrong
是的。
Yeah.
在2012年底，我跟他讨论支持比特币的事，他说Coinbase即将推出一个支付系统。
In late twenty twelve talking about supporting he's like, Coinbase is going to be rolling out a payment system.
你可以在你的网站上接受比特币支付，而我们将成为他们的第一个客户，我当时不断向他提问。
You'll be able to take Bitcoin on your website, and we were gonna be their very first customer, and I was grilling him.
我说，我为什么要选你们而不是BitPay？
I said, why should I work with you instead of BitPay?
但我确实有一个Coinbase的钱包。
And but I did have a I had a wallet at Coinbase.
我是早期的比特币用户，早在任何交易所出现之前，我就在论坛上讨论比特币了。
I was an early Bitcoin guy, like, before there were any exchanges and the Bitcoin talked on our in forums.
我跟你说过，我在我网站上卖横幅广告赚了400个比特币。
Like, I told you I sold banner ads on my website for 400 Bitcoin.
我用8000个比特币买了一部三星Galaxy S Vibrant，那是T-Mobile的第一款Galaxy手机。
I bought a Samsung Galaxy s Vibrant, which was the T Mobile's first Galaxy phone for 8,000 Bitcoin.
你能想象如果我没买那部手机会怎样吗？
Could you imagine if I wouldn't have bought that phone?
那时候我本可以做更多比特币赌博。
I could have done so much more Bitcoin gambling at the time.
但确实如此。
But yeah.
但后来律师们介入了。
But then the lawyers came in.
这是最糟糕的部分。
This is the worst part.
我们获得了批准，然后来了新的首席运营官，他请来了真正的律师，然后他们叫停了我们。
We got approval, and then we got a new COO, and he brought in real lawyers, and they stopped us.
他们停止了。
They stopped.
他们问：你们怎么处理这些税务问题？
They said, how do you handle the taxes on this?
这是资产吗？
Is it an asset?
这是一种货币吗？
Is it a currency?
最好不要是，否则就违法了。
It better not be or it's illegal.
然后发生了什么？这是我的故事，但我刚刚跟你确认了这是真的。
And then what happened this is my story, but I just confirmed with you it was true.
那个家伙是从布兰登那里听说的。
That guy had heard it from Brandon.
布兰登是CEO，我经常跟他聊一些这类事情。
So Brandon was the CEO, and I sort of would talk to him more about some of this stuff.
所以他们收到了一个报价。
So they got an offer.
你收到了扎克伯格的报价，去了洛杉矶，看了演示，然后觉得这是有史以来最棒的东西。
You got an offer from Zuck went down to LA, saw the demo, and was like, this is the greatest thing ever.
我觉得当时你就告诉我，他当场给你开了一十亿美元的价？
And I think maybe on the spot you tell me on the spot offered you a billion dollars or something?
他差不多是当场提出的。
He was more or less on the spot.
是的。
Yeah.
差不多是当场给他开了一十亿美元的报价。
More or less on the spot offers him a billion dollars.
然后我记得布兰登给我打了电话，他说我们要拒绝这十亿美元。
And then I remember Brendan called me, and he's like, we're gonna turn down the billion.
我说，好吧。
I'm like, okay.
为什么啊？
Like, why?
他说，帕尔默有很多比特币，根本不在乎钱。
And he's like, well, Palmer has a lot of Bitcoin and doesn't really give a shit about money.
那时候比特币才200美元。
And this was when Bitcoin was $200.
说实话，我直到刚才才问过帕尔默。
Like, literally, that was I I never actually asked Palmer until just now.
我之前在台上说的时候，想确认一下那是不是真的。
Was like, was that true before I say it on stage?
是真的。
It was
布兰登告诉我，原因就是这个。
Brandon told me was the reason
我正在检查我的钱包。
I checking I was checking I was checking my wallet.
说，比特币涨到208美元了。
Said, oh, Bitcoin at $208.
我稳了。
I am set.
是的。
Yeah.
但后来他又回来了，出价30亿美元。
But then but then he came back and offered 3,000,000,000.
但更重要的是
But but but the more important thing
当2014年《福布斯》给我颁发30位30岁以下精英奖时，他们要求每个人给年轻创业者一句建议。
When I when when when Forbes gave me the 30 under 30 award in 2014, they they they they had everybody had to give one sentence of advice to young entrepreneurs.
而且我还上了封面。
And I was on the cover.
上面写着：帕尔默·拉基，‘多买比特币’。
It said, Palmer Lucky, quote, buy more Bitcoin.
所以我是这里的元老了。
So I was I was the OG here.
所以后来你们最终达到了更高的价格。
So so then then you eventually got to they got to a higher price.
但更重要的是，先给点背景。
But more importantly, just just to give context.
对吧？
Right?
卖出去的真正原因不仅仅是个人赚钱。
The real reason for selling was not just, like, personal money.
而是因为我们所有人都看到了趋势，未来将会有更多的数十亿资金进入。是的。
It was the fact that we we all saw the writing on the wall, which is gonna be billions more Yep.
研发。
R and d.
嗯，你还记得另一面吗？
Well, you remember the other side of that.
你在这件事上是站在另一侧的。
You are the other side of the table on that.
所以对我来说，作为创始人，我对为了钱而出售并不感兴趣，而且我也相信Oculus的价值会远超十亿美元。
So, like, for me, the founder position was I I wasn't interested in selling for money, and I also believed that Oculus is gonna be worth a lot more than a billion.
为什么要为了十亿美元出售并放弃控制权呢？
Why would you sell out for a billion dollars and give up control?
让我们下定决心的是，马克回来后说：第一，我们在某些方面与微软和谷歌的立场一致。
The thing that convinced us is Mark came back and said, one, here's why we are aligned in ways that Microsoft and Google is not.
我们希望VR和AR成为下一代平台，而他们则普遍希望维持现状。
Here's why we want VR and AR to be the next platform, whereas they generally want the status quo to continue.
所以他让我们相信，他们真的想要我们想要的东西。
So he convinced us they really did want what we wanted
是的。
Yeah.
而不是把我们拆解出售，或者利用我们来促进Xbox的销售。
Which was not to sell us for parts or use us to sell more Xbox.
是的。
Yeah.
没有冒犯微软的意思。
No offense to Microsoft.
我喜欢那些人。
Love those guys.
但是的。
But Yeah.
真正起决定作用的是他们愿意为这一信念投入资金。
The the real thing that made the difference was that they were willing to put money behind that belief.
他们说，未来十年，我们每年都会保证向你们投入十亿美元用于研发。
They said, we will put a billion dollars in r and d behind you every year for the next ten years guaranteed.
是的。
Yeah.
依我回忆，我当时其实也参与了这笔交易。
I think I was actually in the deal too, as I recall.
那才是最重要的
It was it was it was that was the thing that was most important
为了
to
对我来说，因为你开始算这笔账
me because you start running the numbers.
我怎样才能每年拿到十亿美元来做这件事
How am I gonna get a billion dollars a year to do Yeah.
研发？
R and d?
而推出一个平台需要大量资金。
And you're launching a platform takes a lot of money.
Xbox当初推出第一代主机就花了大约50亿美元。
Xbox spent, like, 5,000,000,000 launching the first Xbox.
你...你...这...内容制作就是很烧钱。
You you you it just it's the content is expensive.
人员成本很高。
The people are expensive.
市场营销也很昂贵。
The marketing is expensive.
所以我们有像克里斯这样的人来告诉我们，我们要筹集比历史上任何公司都更多的资金。
And so we had people like Chris coming and saying, we're gonna raise more money than we've ever raised in any company in history.
你还记得吗，那轮融资是7500万美元。
You you remember, it was a $75,000,000 round.
是的。
Yeah.
我们进行的这轮融资是7500万美元。
Round we did was 75,000,000.
估值2.5亿美元。
$250,000,000 valuation.
那时候这可是个大事件。
It was a big deal back then.
如果我们筹集数十亿美元，我就无法掌控自己的公司。
And for us to raise billions, I'm not gonna have any control of my company.
如果市场下行，而你的烧钱率又依赖每年十亿美元的风投，那可能需要好几年才能熬过去。
It would have taken years if the market had downturned and you have a burn rate that relies on a billion a year in VC money.
我的意思是，你完蛋了。
I mean, you're toast.
我的意思是，那时候根本不存在每年十亿美元的风投。
I mean, there wasn't a back then, there wasn't a billion a year in VC.
我们真的有一篇《华尔街日报》的文章，说我们有一个十亿美元的基金，当时就像一场泡沫。
We we literally had a Wall Street Journal article that we had a billion dollar fund, and it was like this bubble
是的。
That's
对。
right.
泡沫现象。
Bubble thing.
没错。
That's right.
所以我们把基金的百分之七点五投了进去。
So we put seven and a half percent of our fund.
这就像有个上限一样。
It's like we could there's like a yeah.
上限在
Limit to
所以马克的这个提议，他们说他们是有承诺的。
so this this offer from Mark, they said that they would said they were committed.
我确实记得我说过的话。
I actually remember exactly the words I said.
这在一本叫《未来的历史》的书里有记载。
It's it's in a book called the history of the future.
作者获得了大量Facebook内部资料，这是一本不错的书。
The author got a bunch of Facebook internal it's a good book.
一大堆Facebook内部文件。
Bunch of Facebook internal documents.
但我发给创始人们的那封邮件，他们说：哦，是的。
But the email that I sent out to the founders, they're like, oh, yeah.
我觉得扎克伯格可能在忽悠我们。
I think Zuck might be might be BS ing us.
我说，如果扎克伯格在耍我们，那他就是范·海伦。
I said, if Zuck is playing us, he's Van Halen.
而这正是最终说服我们的关键点。
And and that that was really the thing that they convinced us in the end.
现在他们花了，你说是800亿美元？
And now they've spent, you said, 80,000,000,000?
官方公布的数字是600亿美元。
The official public number is $60,000,000,000.
我看待这件事的方式，也是我说服自己接受被自己创办的公司解雇这一事实的方式——我年轻时创建了这家公司，把整个成年生涯都投入其中——那就是，最终，Facebook并没有真正收购Oculus。
The way that I look at this and the way that I the way that I convince myself that it wasn't so bad that I got fired from my own company that I I built as a youth and put my whole adult life into is that in the end, Facebook didn't really acquire Oculus.
Oculus接管了Facebook。
Oculus took over Facebook.
我们成为了他们在支出、研发和产品创造引擎中压倒性的主导部分。
We became the overwhelming and dominant component of their spend and their r and d and their product product creation engine.
这相当不错。
That's that's pretty good.
这相当不错。
That's pretty good.
是的。
Yeah.
知道。
Know.
太惊人了。
It's amazing.
是的。
Yeah.
展开剩余字幕（还有 480 条）
而他们只花了30亿美元就让我接手了
And it only cost them $3,000,000,000 for me to take
变得不错了。
one good.
它变得很好了。
It's gotten good.
Meta Quest很不错，我觉得，你知道的，就像，我相信
The Meta Quest is good, and I think the you know, it's like, I I believe
Quest是我被他们解雇前最后参与的项目。
Quest was the last thing that I worked on before they fired me.
是吗？
Is that right?
是的。
Yeah.
我的意思是，我仍然非常相信虚拟现实。
I mean, I I still very much believe in VR.
这比我想的要花更长时间。
It's taken longer than I expected.
我的意思是，就像你说的那样
I mean, as is, you know
我是个真正的
I'm a true
信徒。
believer.
加密货币和很多其他东西可能需要更长时间，但我我相信
Crypto and a bunch of stuff can take longer, but I I
它一定会成功的。
It it is going to work.
有些人看到它失败了反而感到高兴。
And there's there's people who take a kind of glee, and it's failing.
你必须理性地看待，它其实并没有失败。
You gotta actually look, like, like, rationally, it is not failing.
我认为它没有达到某些人——包括我自己——所期望的水平。
I think it's not living up to what certain people, myself included, believed that it would be.
但当你看到这些文章，比如现代科技媒体其实不是科技媒体，他们是反科技媒体。
But when you see these articles from from like, modern tech press is not the tech press, they're the anti tech press.
但当你看到现代反科技媒体的文章，你知道，为什么VR没有成功？
But when you see modern anti tech press articles, you know, why didn't VR succeed?
他们没有提到上一代Quest卖出了2000万台。
They're not mentioning that the last version of Quest sold 20,000,000 units.
他们提到的销量已经超过了Xbox，在
They're not mentioning more than Xbox, at
至少一年内。
least one year.
它的销量远超Xbox。
It sold way more than Xbox.
就像，Quest 2在其问世的前三年里销量比任天堂64还要好。
It's like, Quest two sold better for the first three years that it existed than the Nintendo 64.
所以这简直就像一场失败。
Like and so it's like, what a failure.
它只比我童年时拥有的每一款游戏主机卖得都好。
It only sold better than every game console I owned as a child.
这是因为他们在和智能手机做比较。
Like, it's it's it's this it's it's because they're comparing to smartphones.
是的。
Yeah.
是的。
Yeah.
但智能手机带来了前所未有的技术变革浪潮。
But smartphones are this unparalleled explosion in technological change.
对。
Yeah.
我们没能赶上历史上最重大的计算机使用方式变革。
We're not as good as the biggest change in the way we've used computers in history.
是的。
Yeah.
但我们只是比它低了一步而已。
But we're we're we're we're like the next we're only one step down.
顺便说一下，这还只是Quest 2。
So I I I'm and, like, by the way, that's just Quest two.
你后来还推出了Quest 3，和它同期销售了一段时间。
You also then had Quest three selling for the same time as a while.
当然，你还有PlayStation VR，卖了大约六百万台。
Like, you had, of course, PlayStation VR, which sold about about 6,000,000 units.
我们还没有PlayStation VR 2的销售数据，但显然它们的销量也是数百万台。
We don't have numbers for PlayStation VR two yet, but, like, they're definitely selling millions of units.
你还有HTC Vive，至少卖出了三到四百万台。
You had HTC Vive, which sold at least three or 4,000,000 units.
从任何角度来看，你都在面对一个拥有数千万用户使用这项技术的世界。
It's like you're looking at a world where there are, by any measure, low tens of millions of users of this technology.
这已经很不错了。
That's that's pretty good.
在iPhone 3GS之前，用iPhone的人并不多。
There weren't that many people using iPhones until the iPhone three GS.
是的。
Yeah.
就像他们必须等到iPhone的第三代才行。
Like like they it's like they like, you had to get to the third generation of iPhones.
2012年，Facebook才转向移动平台。
It's 2012 that Facebook did their mobile pivot.
人们都忘了这一点。
People forget that.
花了五年时间，人们才真正意识到这是全新的平台。
It was five years in that it actually was obvious to people that it was like the new platform.
没错。
That's right.
没错。
That's right.
就好像人们事后对发生的事情进行了重新解读。
It's like it's like people have this retrospect, like this retcon kind of of what happened.
嗯，我也喜欢把它和iPod比较一下。
Well, I also I I like to compare it to iPods too.
我有一段时间没做这个对比了，但我们曾经有个虚荣指标，指出Oculus Rift在第一年的销量超过了iPod前三年的总和。
It's been a while since I did this exercise, but we had some vanity metric where we pointed that the Oculus Rift in the first year outsold the first three years of the iPod.
而且，这确实是事实。
And, like, it's true.
iPod后来才开始呈指数级增长。
IPod took off in an exponential way later.
但到了那前三年的时候，它已经登上了主流杂志的封面。
But by the time it was like, those first three years, I mean, it was on the cover of major magazines.
它还赢得了年度最佳科技产品奖。
It was winning gadget of the year.
当时所有最热门的，这简直成了必备品，而我们的销量却超过了它。
All the hottest like, it it was an it thing to have, and we were we we were we were outselling it.
过去，成功就足够了。
It used to be that success was enough.
是的。
Yeah.
但现在这就不够了。
That's not enough anymore.
必须取得百亿级别的成功，否则你就是个无名之辈。
It needs to be megabillion success or you're nobody.
另一点是，你就是个失败者。
Well, the other thing is that there's You're a failure.
还有，比如人体工学、视觉效果这些方面。
They're obviously, like, the ergonomics, the the visuals.
显然，它们只会变得越来越好。
Like, they're just obviously gonna get better.
是的
Yep.
而且它们现在是个问题。
And they're an issue right now.
比如，你会容易出汗。
Like, it's kinda you get sweaty.
就是这样的，对吧？
It's kinda like right?
所以，我觉得有些事情很难预测，比如用户行为，但有些事情很简单。
And so, like, I feel like there's hard things to predict, like user behavior and like, but they're easy things.
当然，它们会变得越来越小、越来越轻，对吧？
Like, of course, they're gonna get smaller, lighter Right.
更快、性能更强、图形更好、游戏更棒。
Faster, more performant, better graphics, better games.
有一篇非常值得一读的好文章，写得非常好，发表在palmerlucky.com上，标题是《免费还不够便宜》。
There's a there's a great article that's definitely worth reading, very well written, on palmerlucky.com called Free Isn't Cheap Enough.
这是互联网上排名第一的Palmer Lucky博客。
It's the number one Palmer Lucky blog on the Internet.
免费还不够便宜？
Free Isn't Cheap Enough?
这篇文章是我写的。
It's written by me.
但这是一篇很棒的文章，我在多年前提出了一个名为‘免费还不够便宜’的假设。
But it's this great piece where I laid out this hypothesis years ago called free isn't cheap enough.
我认为阻碍VR发展的不是每台设备的价格。
And I argue that the thing holding VR back is not the price per unit.
不是因为它是300美元而不是200美元。
Like, it's it's not the fact that it's $300 versus $200.
是的。
Yeah.
而是你刚才提到的那些问题。
It's the it's what you're talking about.
它太重了。
It's too heavy.
它太出汗了。
It's too sweaty.
质量和内容的流水线根本跟不上，无法让普通人、不是游戏玩家、不是早期采用者，比如你妈妈
The quality and content pipeline is just not there to make an average person, not a gamer, not an early adopter, like your mom
是的。
Yeah.
来使用它。
To use it.
所以当人们说我对VR不感兴趣时，你可以给他们一个简单的假设。
And so when people say, I'm not interested in VR, you can give them an easy hypothetical.
好吧。
Okay.
等等。
Well, wait.
如果它是一副太阳镜，效果像《黑客帝国》那样，而且只要99美元呢？
What if it was a pair of, like, sunglasses, and it was as good as the matrix, and it were $99?
你会用吗？
Would you use it?
就连想象力贫乏的人也会说，是的。
And it doesn't even unimaginative people are like, yeah.
如果我能通过一副眼镜实现我想象中的任何事情，我当然会用，但那并不是VR现在的样子。
I like, if I could do literally anything that I could imagine in a pair of glasses, of course, I would do it, but that's not what VR is.
我说，那我们只是在争论VR什么时候会爆发。
And I say, well, then then we're just arguing about when VR is gonna take off.
问题不是它会不会发生。
It's not if it will.
我们知道物理原理允许我们实现我所描述的那种东西。
We know the physics allows us to get to something like what I've described.
所以我认为，一味地追求让VR变得更便宜是毫无意义的，因为即使它是免费的，也还不够好，人们不会持续使用。
And so I argued that this relentless push on making it cheaper and cheaper is pointless because even if it was free, it wouldn't be good enough for people to keep using.
我会暂停它。
I would pause it.
你可以把如今市面上的VR头显送给每一个美国人，其中超过90%的人会在一个月内停止使用，因为体验质量还不够好。
You could give a VR headset as they exist today to every person in America, and upwards of 90% of them would stop using them within a month because it's just the quality is not quite there yet.
换句话说，免费也远远不够便宜，无法让VR成功。
In other words, free isn't cheap enough to make VR succeed.
是的。
Yeah.
也许价格更高、体验更好的产品反而会更成功。
Maybe even a higher price, better one would be even more successful.
那正是我的论点之一。
That was one of that was my argument.
不幸的是，人们没有听我的，他们继续推进，推出了售价99美元的Oculus Go，结果又回到了免费的误区。
Unfortunately, people didn't listen to me, and they kept going, and they launched the Oculus Go at $99, which went back to being freed off.
而且是的。
And Yeah.
那真是一场完整的误入歧途。
That was a whole misadventure.
结果，这成了我和约翰·卡马克真正意见不合的一件事。
And ended up this is one of the things where me and John Carmack were actually at odds.
他认为价格方面，是的。
He thought that price yeah.
约翰·卡彭特，非常
John Carpenter, very
设备很出色。
sharp the gear.
他一直很喜欢这个。
He always loved that.
是的。
Yeah.
他是个非常敏锐的人，而且接地气。
He is a he is a very sharp guy, and he's a man of the people.
他希望每个人都能拥有VR，但结果发现售价3.99美元的Quest销量远超99美元的Go。
He wanted everybody to have VR, but it turns out that Quest at $3.99 hugely outsold Go at 99.
我认为这说明了它的市场定位。
And I think that says something about the shape of it.
现在你看到苹果和三星在攻击高端市场。
Now you see Apple and Samsung attacking the high end of the market.
我认为这成功的可能性更高。
I think that has a higher chance of success.
嗯哼。
Mhmm.
我真心祈祷这一代能成功，是的。
I'm I'm I'm praying that it works this gen Yeah.
否则我们还得再经历一段寒冬。
Or we're gonna have to go through another bit of a winter.
是的。
Yeah.
那么，我们继续聊聊Andoril。
So you've done so moving on to Andoril.
你在Andoril上做得太棒了。
You've done an amazing job with Andoril.
我觉得让我印象深刻的几点是，首先，这是一家标志性的公司，恭喜你取得了如此巨大的成功。
I think some of the things that stand out to me, mean, first of all, it's just an iconic company, and congratulations on the incredible success.
但你也知道，一开始你们面临的是负面舆论，那种感觉，你知道的。
But also, like, you started off, you know, negative press, like, this this kind of sense, you know
我们曾获得《连线》杂志颁发的‘硅谷最差人物’奖。
We won the Wired Magazine worst person in Silicon Valley award.
彭博社称我们是科技界最具争议的公司。
Bloomberg called us the most controversial company in tech.
那是Andrew才25岁，带着25个人在研发无人机防御系统的时候。
So this was back when Andrew was 25 people building drone defense systems.
是的。
Yeah.
顺便说一下，那一年WeWork被联邦政府起诉了。
And and by the way, this is the same year that WeWork had been been indicted by the feds.
是的。
Yeah.
Uber刚刚撤换了所有高级管理层。
Uber had just ousted all of its senior leadership.
他们？不。
Them as No.
是的。
Yeah.
当时最有争议的科技公司是Homer Lucky和他的24人无人机团队。
The most controversial tech company was Homer Lucky and his 24 drone drone people.
是的。
Yeah.
这简直不可思议。
It was incredible.
所以
So
那你从那时到现在有了七千名员工？
you so you went from that to now 7,000 people, employees?
这要看你怎么计算了，但如果包括承包商的话。
Just Depends on how you count it, but if you include contractors.
是的
Yeah.
而且我认为，像这样一家标志性公司，我猜它曾经可能在招聘方面遇到困难，你来告诉我。
And and sort of, I think, an iconic company that is, I assume, has gone from maybe having trouble recruiting in the you tell me.
从招聘困难转变为如今的人才管理。
Trouble recruiting beginning to now being a talent manager.
恰恰相反。
The opposite.
你
Did you
看到我们最新的招聘活动了吗？
see our latest recruiting campaign?
它叫‘别在末日世界工作’。
It was called don't work at the end world.
因为你不希望错误的人加入。
Because you don't want the wrong people to
这已经成为一种现象了，当你在早期做一件疯狂而新颖的事情时，情况确实如此，Oculus 也是这样。
It is it has become when when so when you're when you're you're doing something early and it's really crazy and it's new this is true with Oculus too.
你只会吸引到真正的信徒，那些因为相信你所做的事情而加入的人，他们本可以在别处赚更多钱。
You only get the true believers, the people who are there because they want to and they believe in what you're doing, and they could be making more money elsewhere.
只要他们觉得，为什么你要去一家每个人都在嘲笑、觉得是笑话的初创公司工作呢？就像早期的 Oculus，对很多科技圈的人来说就是这样。
As long as they feel like what they're doing is like, why would you go work at this start that everyone's laughing about and thinks is a joke like Oculus was early on to a lot of people in tech?
如果连薪水都给不了多少，你为什么要去那里？
Why would you go there if if if it's not even gonna pay you well?
答案是因为你真的相信他们所做的事情。
And the answer is because you really believe in what they're doing.
这自然过滤掉了那些跳来跳去的人和所谓的思想影响者。
And that's a natural filter for the lily pad jumpers, the thinkfluencers.
你知道我说的是哪些人。
You know the people I'm talking about.
你在领英上就能看到他们。
You see them on the LinkedIn.
你知道，这些人现在自称是AI专家。
You know, these are the people who they're they're AI experts right now.
他们一年前还是AR专家。
They were AR experts a year ago.
再往前一年，他们又是VR专家。
They were VR experts a year before that.
我想他们之前还自称是社交、本地、移动领域的专家。
What I guess they yeah, they were like, you know, the social local mobile experts before that.
Web 2.0时代，他们肯定都是Web 2.0的信徒，还有Web 3.0。
Web two point o, yeah, guarantee, they all were Web two point o people as well and Web three point o.
别忘了Web 3.05G。
Don't forget Web three point o five g.
是的。
Yep.
你和参加过足够多科技会议的人，都清楚我在说什么。
You and people who have been to enough of tech conferences know exactly what I'm talking about.
这太好笑了。
That's hilarious.
所以问题在于，Androle现在已经到了没有任何筛选标准的地步。
And so the problem is Androle has now gotten to the point where there's no filter.
我们现在够酷了，能把那些人吸引过来。
We are now cool enough to get those people.
你现在太酷了。
You're too cool now.
所以，找到对的人很重要。
And so it's important to get the right people.
更重要的是要拒斥那些错误的人，无论他们把自己打扮得多么光鲜，隐藏得多么好。
It's even more important to repel the wrong people no matter how well they dress themselves up and and hide their true nature.
你该怎么发现呢？
How do you discover that?
我们拍了一个广告，展示人们必须到野外去从事脏活累活，你的工作时间不可预测，你可能会见到家人。
We did a commercial where we showed people how you have to go out in the field and work in the dirt, and you and your hours are unpredictable, and you might see your family.
但也可能见不到。
You also might not.
广告里有个男人，他整个广告都在抱怨这份工作有多糟糕。
And we kinda and and the the ad there's a guy in the ad, and he he complains throughout the ad about how awful it is.
然后在结尾，他说：是的，别去安德罗尔工作。
And then at the end, he says, yeah, don't work at Anderall.
我们的想法是，对某些寻求意义、使命和冒险的人来说，当他在一艘船上和一群特种部队士兵在一起时，这种感觉就来了。
And and the idea was that to a certain type of person who's looking for purpose and mission and adventure, like, there's a point where he's like on a boat with a bunch bunch of Special Forces guys.
他们坐在一艘小小的橡胶艇上。
They're on a little rubber boat.
他四处奔波，前往某个前沿部署地点。
He's bouncing around and, you know, and going off to some, you know, some forward deployed location.
我们确实这么做。
We actually do this.
我们有一支数百人的前沿部署工程师团队，支持许多这些任务，我们的客户对此非常满意。
We have a team of hundreds of forward deployed engineers supporting a lot of these these missions, And our customers love it for it.
但对某种类型的人来说，看到这一幕会说：天啊。
But the idea was to a certain person, they see that and say, oh my gosh.
这看起来简直是太酷了。
This looks like the coolest thing ever.
而对于那些想踢足球、做瑜伽、五点或五点半就回家的普通人来说，他们会因此却步。
And to the normal person who wants to play soccer and do yoga and and be home by five, fifteen, they're gonna be repelled by that.
是的。
Yeah.
对。
Yeah.
我对投资者也是同样的看法。
So I I by the same thing with investors.
获得投资者很重要。
It's important to get investors.
在某个阶段，你会变得如此抢手，以至于每个人都想投资。
At a certain point, you become so hot that everyone wants to invest.
然后你就得想办法如何拒绝那些与使命不一致的投资者。
And then you have to figure out how will I repel investors who are not mission aligned.
所以你得说些疯狂的话。
And so have to say crazy stuff.
你必须真正突破极限。
You have to you have to really push the limit.
你得在有限合伙人会议上引发这样的讨论：他们为什么会投资这家公司？
You have to you wanna generate that conversation in the LP meeting where they're like, are why why are you investing in this company?
你需要他们明确地说：我们百分之百支持他们，原因如下。
And you need them to to to to to say, we are a 100% behind them, and here's why.
在从事有争议的业务时，你不能只拥有顺境时的朋友。
And you can't you can't have the you can't have the fair weather friends when you're in in controversial business.
你是否
And do you
你觉得你能在7000人中成功做到这一点吗？
think you may manage to do that with 7,000 people?
在你的面试和流程中，你也必须这样做，因为这些人会通过广告漏网，而且他们看起来……
You also must do, like like, in your interviews and your process, you must because they slip through the ads, and they Look.
这完全是你的事，你永远不可能做到完美。
It's it's it's all you you'll never be perfect.
是的。
Yeah.
但你所做的每一件事，都会让问题在天平上向更好的方向移动。
But everything you do makes the problem slide this way on the scale towards better.
所以，别把安多里尔当成一个广告活动来运作。
And so, like like, don't work at Andoril as an ad campaign.
我们的招聘数量，就流入量而言，翻了三倍。
Tripled our recruitment on like, in terms of incoming.
而且，我们也收到了更多合格的申请者。
And and also, we got a lot of more qualified applicants.
但大部分情况下，进来的是不合格的人。
Mostly, it brought in people who are unqualified.
但你是在试图推动进展，我觉得这一切都累积起来有作用。
But you're trying to move the move the needle, and I I think it all it all it all adds up.
我们当然会雇佣一些不合适的人，然后把他们辞退。
We, of course, hire people who are bad fits, and then and then we fire them.
是的。
Yeah.
对。
Yeah.
这个
The
他们有时候自己就辞职了。
They quit themselves sometimes.
是的。
Yeah.
我还想说一点，我只在这款应用上参与。
The other thing I'll say and I'm I'm not I'm only on the app.
我的意思是，这家公司是投资者，但我并没有直接参与。
I mean, the firm's an investor, but I'm not directly involved.
但我作为一个外部粉丝，真的感到非常惊讶。
But I I'm I'm a fan from the outside, and I am just amazed.
很多初创公司，比如，他们推出产品，尤其是硬件，每两年一次。
Like, so many startups, like, they launch a product, especially hardware, like, every two years.
但我觉得每三周就看到一次新东西，我上个月刚见过你。
And I feel like every three weeks like, I just saw you.
你那时推出了这款新的AR头盔，而三周前，你好像还发布了一款自主潜水艇。
You had this new AR helmet, and three weeks ago, had, I think, an autonomous submarine.
你就是不停地推出各种产品。
Like, you just launch a lot of stuff.
这简直令人难以置信。
It's kind of incredible.
你是怎么做到的？
Like, how do?
25款产品。
25 products.
关于Andoril，有趣的是我们正在做的事情——我们推出了非常多的产品，并且每个产品都有一个相对精干的独立团队，这反而让我作为公司管理者更容易运营，而不是像那些必须投入大量资源开发一两款产品并不断扩展以满足所有人需求的公司。
The interesting thing about Andoril is what we're doing, where we're launching many, many, many products and where we are having individual product teams on each of those that are relatively lean is actually easier for me to run as a company than the companies that have to invest in one or two and grow them to serve everybody.
这方面最典型的例子就是那些大型企业软件公司，它们有成千上万的人只专注于一个产品，而这个产品必须同时满足所有用户的需求，是的。
Like, the ultimate example of this is these large enterprise software companies where you have thousands of people working on one thing, and it has to serve everyone in Yeah.
你的商业客户、消费者、企业客户和政府机构都要照顾到。
Your commercial and consumers and enterprise and government.
要让成千上万的人一起开发一个产品，这真是个极其困难的问题。
And it like, that is a very hard problem to build a product with thousands of cooks in the kitchen.
我只需要不断推出新产品，然后让每个团队保持较小的人数。
All I have to do is keep making new products, and then I keep a small number of cooks in each kitchen.
这最终成了一种相当自我调节的策略。
And that that that that that that ends up being a pretty self regulating strategy.
人们是有群体倾向的。
People are tribal.
人们在一定人数范围内工作得最好。
People work well with a certain number of people.
你能真正信任、相信他们做事动机纯正的人，数量是有限的。
Like, there's only so many people you can have very high trust in, where you trust they're truly doing things for the right reason.
我的意思是，当你了解周围每个人在做什么时，这个规模不会比一个家庭建筑队或一个小游戏开发团队大太多；你会不断看到，在各种社会和历史中，都反复出现这种特定规模的群体。
You I mean, where you understand what everyone is doing around you, it's not that much bigger than the size of a home construction crew or a small game development team or that either you kind of see these natural themes emerge over and over again across societies and across history where it's these groups of a certain size.
你看，比如开国元勋们。
You look I mean, the founding fathers.
我努力让每个业务单元都尽可能接近这个规模。
And I'm trying to keep every business unit as close to that size as I can.
他们都是工程师和产品人员，而不是真正的工程师。
And they're all sort of engineers and product people, not Engineers.
我猜你试图消除中层管理，比如
I assume you try to eliminate middle management and like
在阿得拉尔公司早期，有一段非常短暂的时期，我们的律师和游说者比工程师还多。
There there was a very brief period, very early in Adderall's history, where we had more lawyers and lobbyists than engineers.
那段时期非常非常短，但当你在这样一个高度合规、监管严格、政府事务繁重的行业创业时，不得不在这些方面过度投入。
It was very, very brief, but you had we had to overinvest in that when you're starting a company in such a compliance heavy, regulatory heavy, government interface heavy side.
从那以后，我们再也没有遇到过这个问题。
We have never ever had that problem since.
我觉得我们大约85%是工程人员。
I'd say we're about 85% engineering.
没错。
That's correct.
所以，是的，我们的管理开销非常少。
So, yeah, we we have we have quite little overhead.
我们没有太多中层管理。
We don't have a ton of middle management.
我们有很多项目和团队负责人，他们通常是来自其他项目的工程师晋升上来的。
We have a lot like, we basically have project and team leads that are typically engineers who came up out of other programs.
我们的一个项目负责人，我们叫他‘实习生之王’，因为他一开始是安德雷尔的第一个实习生。
Like, one of our program leads is a guy we call him the king intern because he started he was Anderle's first intern.
他19岁时就在安德雷尔开始工作。
So he started when he was 19 years old at Anderle.
我们当时只是他的暑期实习。
We were his summer internship.
然后他退学，加入安德雷尔，十二个月内就负责了一个产品，现在他领导着整个业务部门。
And then he quit school, came to Anderle, and within twelve months was leading a product, and now he's leading an entire business division.
我们在这方面做得相当不错，建立了一种 meritocracy（能者上）的机制，大家看到的领导者都是在安德雷尔内部成长起来、经受过考验的工程师。
And, like, we've done a pretty good job of making a meritocracy where people see the people in leadership roles are themselves engineers who have cut their teeth and proven themselves inside of Andoril.
是的。
Yeah.
当你快速成长时，这要容易得多。
And when you're growing quickly, it's a lot easier.
对于一家稳定公司来说，这是一种更难实施的策略，我不推荐这样做。
It's like, this is a harder strategy for a stable company, and I can't recommend it.
但当你以高效的速度推出新产品时，总有一个位置可以让你从中挑选这些人，告诉他们：‘你来负责这个新产品，用你对组织运作方式的经验，以及大家对你的了解和信任，来让它成功。’
But when you're introducing products at as efficiently high rate, there's always a place for you to take these people to pluck them out and say, you're gonna be in charge of this new product, and you're gonna make it work using your experience of how to work inside of an organization and the fact that people know you and trust you.
静态的组织在这一点上会遇到困难，因为他们会觉得：根本就没有足够的职位。
Static organizations struggled with this because they're like, there just aren't enough roles.
这就是高级副总裁职位数量有限的问题。
It's the it's the how many senior VPs are there problem.
快速成长是件好事。
It's it's it's nice
快速成长真好。
to be growing fast.
你是怎么想到这些的？我想就这个问题问几个问题。
How do you come up like, I guess a couple questions on this.
比如，第一，你是怎么做到好像对硬件和软件的每件事都很了解的？
Like, one, how do you you seem to follow, like, everything in hardware and software.
我尽力了。
I try.
我看你的视频。
Watch your videos.
所以，我很想知道你是怎么做到的。
And so, like, I'm curious how you do that.
还有，你是怎么想出这么多产品点子的？
And then also, how do you come with all the ideas for the products?
是自下而上的吗？
Like, is that is that bottoms up?
是你自己想的吗？
Is it you?
还是说，你是怎么保持这种活跃的研发状态的？
Is it, like, how do you keep it like an active R
还有D？
and D?
我以前想出所有点子。
I used to come up with everything.
我现在只想出我们一半的产品点子。
I now come up with about half of our products.
另一半则来自客户需求以及团队成员提出的解决方案和问题，这很好。
The other half come through a combination of customer demand and people in the teams coming up with solutions and problems, and that's good.
换句话说，在最终阶段，我提出的每一个想法都应该比我的员工提出的更差。
Like, the the the term in the terminal stage, every idea I have should be worse than the ones my employees are generating.
是的。
Yeah.
我的意思是，今天可能也是如此。
I mean, that could even be the case today.
但我才是负责人。
But I'm the guy in charge.
我是那个不会被解雇的人。
I'm the one guy that can't be fired.
所以有一半的想法还是我的。
So half the ideas are still mine.
希望随着时间推移，这个比例能越来越小。
Hopefully, that shrinks over time.
抱歉。
Sorry.
那个关于‘他们’的问题的另一部分是什么来着？
What was the other bit of that question that was in terms of they
我只是很好奇，你是怎么跟上这些事情的。
Well, just like I'm just curious, like, how you kinda keep up with the stuff.
我该怎么跟上
How do I keep up with
那些奇怪的东西？
weird stuff?
我从不在公共场合承认这一点，但我确实一直在关注学术文献。
I would never admit so in in public, but I do actually keep up with the academic literature.
哦。
Oh.
我并不是
I'm not
比如，是计算机科学方面的哪些内容？
What like, like, of the the computer science?
只要是我觉得有趣的领域，我都会关注。
In any space that I'm interested in.
所以，我会阅读所有提到合成长链碳氢化合物的期刊。
So, like, I I am reading all the journals that mention synthetic long chain hydrocarbons.
我非常相信这一点。
I'm a huge believer in that.
这将改变一切，并让整个电池电动技术显得像它本来就是的那样——一场价值数万亿美元的闹剧，尤其是在航空领域。
It's gonna change everything, and it's gonna make this whole battery electric nonsense look like the multi trillion dollar farce that it is, especially in aviation.
这是一个很长的话题，但顺便说一句，我并不是说这个领域表现最好的人一定是学术界的，但只要紧跟学术前沿，你就永远不会落后太多。
This is a long topic, but, like, I'm not saying by the way, the best performers in that area are not necessarily academic, but by staying on top of the academic side, you'll never be too far behind.
学术界的人并不一定需要去创造、生产或销售东西。
People people in academia don't have to create or produce or sell necessarily.
他们的工作只是站在新事物的最前沿，因此大多数时候他们都不会落后太多。
Their job is just to be on the front lines of what's new, and so they'll never be too far behind that most of the time.
VR也是如此。
Same thing with VR.
光学也是如此。
Same thing with optics.
类似的情况还有，比如……
Same thing like There's like
VR和这些领域里有好的学术研究吗？
good academic stuff in VR in some of these areas?
没有，没有。
No, no.
没什么好的内容，但他们什么都会聊。
There's nothing good, but but they talk about everything.
好的。
Okay.
但这我想说的是，是的。
It's it's it's But I think of Yeah.
只是别去想。
Just don't think
换个角度想。
Think of it this way.
比如，假设你想要跟上用于制造叠层光学器件的新式曲面偏振片材料的最新进展，这种技术是将光路层层折叠。
Like, suppose that you are going to try to Like, you want to keep up to date with new curved polarizer materials that are used for making pancaked optics, where you're folding optical paths on top of themselves.
我想紧跟这方面的动态。
Like, I want to stay on top of this.
那你打算怎么做？
Well, what are you going to do?
你会去阅读那些宣传这类技术的公司发布的新闻稿吗？
Are you going to read press releases that of all the companies that are, you know, touting this stuff?
这些制造商根本不会宣传这些东西。
The company the the the manufacturers don't tout this stuff.
那你打算怎么跟上这些新进展的存在呢？
Like, how are you going to stay on top of the fact that there even are new advancements?
结果发现，一个相当不错的方法就是关注学术文献，那里有人在研究这些内容，他们会提到，比如中国某个团队在做这个，还有其他公司采用了这种方法。
And it turns out a pretty good way is to just follow the academic literature where you have people that are doing research on these things, and then they'll like, jobs say, you know, oh, and, you know, such and such group in in China is doing this, and there's other company that has this approach.
我们提出了一种针对X、Y和Z的全新方法。
What we propose is a novel approach to x, y, and z.
现在他们可能根本做不到，也永远不会做出什么有意义的成果，但他们毕竟还是做了其他相关的工作。
Now they probably can't actually do it, and they're never gonna pull off anything of interest, but they've still done, like, the the the other work.
我基本上把他们当作我的外包研究团队，只要我想了解某项技术的最新进展，我更愿意相信学术界，而不是新闻媒体。
I I am basically using them as my outsource research team on anything like if I need to find out what the state of the art on something is, I can trust the academics somewhat more than the press.
是的。
Yep.
因为媒体什么都讨厌。
Because the press just hates everything.
是的。
Yeah.
他们什么都讨厌。
They hate everything.
而且他们也很懒。
And so and also they're lazy.
他们不，他们不深入挖掘。
They don't they they they they're they're not they're not digging deep.
过去，媒体只是正面地转述新闻稿中的内容。
They they are they're basically re the the press used to regurgitate what was given to them in press releases positively.
现在他们只是转述，但语气已经变了。
Now they just regurgitate it, and and the and the the the tenor has changed.
但本质上还是同样的行为。
But it's it's the same it's the same activity.
只是重复新闻稿里的内容，而围绕它的评论发生了变化。
It's just repeating what's in the press release, and the commentary around it is what's changed.
所以，是的，实际上我通过阅读学术论文和期刊来保持更新，这些论文和期刊专注于我感兴趣的领域。
So, yeah, actually, I stay up to date a lot by reading academic academic papers and journals that are focused on the things that I'm interested in.
学者们也很喜欢抱怨产业界的人，比如核能领域。
The academics, also love when they complain about people in industry, like nuclear.
如果你想了解核能领域的最新动态，并想获得一些反主流的观点，那就关注学者们对现代小型模块化反应堆的评论。
If you wanna stay on top of what's what's what's going on in nuclear and you want kind of the contrarian view, follow what the academics are saying about modern, small, modular reactors.
你会看到他们谈论新进展，然后抱怨这些技术行不通。
And you'll see them talking about new developments and then complaining about how they don't work.
有时候他们的批评很有道理，有时候却完全毫无意义。
And sometimes their complaints are good, and sometimes they make absolutely no sense.
但是，就像
But, like It
它告诉你该关注什么。
tells you what to look at.
它告诉你该关注什么。
It tells you what to look at.
而且，我真的没有足够的时间去成为这方面的真正专家。
And and, like, I I I don't have enough time to become a real expert on this stuff.
就像，我只是觉得人们并不了解。
Like, I I just people didn't know.
我曾经是新闻专业的学生，但后来退学了。
I used to be a journalism major till I dropped out.
我曾经是加州州立大学长滩分校《每日49人报》的在线编辑，那是当时全美第四大的学生运营报纸。
I was the online editor of the daily 49er at Cal State Long Beach, which is the fourth largest student run paper in the country at the time.
现在它不再叫《每日49人报》了。
And it's not the daily 49er anymore.
它现在叫《每日微风》，因为‘49人’指的是淘金热，而淘金热现在被认为具有种族歧视色彩。
It's now the daily breeze because 49er is a reference to the Gold Rush, and the Gold Rush is racist racist now.
我不是在开玩笑。
That's I'm not kidding.
这就是他们这么做的原因，这很疯狂，因为球队名字还是49人队。
That's why they did it, which is crazy because the team is still the 40 niners.
那这怎么可能呢？
So how could that be?
不管怎样，这份报纸认为自己的名字具有种族歧视意味，于是决定在一百多年后改名。
Anyway, the the the the the the paper decided its name was racist, and so they would rename it after a hundred plus years.
但作为一名记者，我意识到你真正能亲自从第一手来源跟进的事情是有限的。
But you I what I realized as a journalist is you can only really follow so many things yourself directly from first sources.
比如，你可以选择一个报道领域并持续关注，但最终你必须把研究工作外包给其他人，让他们整合并把这些信息汇总起来。
Like, if you you could choose a beat and you could stay on top of it, but you have to outsource at some point the research to other people to consolidate and and bring these things together.
我非常看好在我的指导下使用AI来完成这项工作。
I'm very bullish on using AI at my direction to do this.
我希望能让AI去抓取所有内容——新闻稿、社交媒体上的讨论、学术资料，然后给我一个不受任何商业利益或意识形态影响的总结。
I wanna be able to send an AI and say, scrape literally everything, the press releases, the social media chatter, the academic stuff, and then give me give me a summary that is not contaminated by some profit interest or ideological interest.
是的。
Yeah.
你不需要说服在场的人，让他们相信媒体对科技有偏见。
You don't have to convince this crowd that the press is anti tech.
加密货币可能是最令人难以置信的了。
Crypto has been the most I guess it's unbelievable.
我们十二年来一直试图发表一篇关于实际发生事情的理性文章，但始终没能成功。
Like, for we've tried for twelve years to have, like, one reasonable, like, article printed about, like, the actual stuff going on.
AI这块确实太难了。
Well, and the AI stuff is so hard.
现在，随着大量虚假水军文章的出现，AI领域已经开始感受到压力了，广告方面的热度还会继续上升。
Well, the AI start they're starting to feel it now with all these, like, fake water articles and things, and it's gonna just heat up more on the ad.
当人们看到Andoril，说‘瞧啊，Andoril也想蹭AI的热点，假装他们所有东西都是AI’的时候，真是太好笑了。
It's so funny when people, they look at Andoril and say, oh, look at Andoril trying to jump on to the AI grift, trying to pretend that all their stuff is AI.
我们已经做了八年了。
We've been doing it for eight years.
我公司的名字就叫Andoril Industries。
The name of my company is literally Andoril Industries.
这个缩写词就是AI。
The acronym is literally AI.
我们以前根本没谈过这个，因为AI那时候不酷。
Like, it's it is so fright we we didn't talk about
在早期阶段。
it in the
早期我们没谈过，因为AI那时候不酷，是的。
early days because AI wasn't cool Yeah.
是的。
Yeah.
差不多九年前了。
Almost nine years ago.
那时候就像冷聚变或者虚拟现实。
It was like cold fusion or VR.
是的。
Yeah.
是的。
Yeah.
对未来是胡说八道，对现在则完全不是。
Nonsense for the future, never in the present.
但没错，媒体把我说成是蹭AI热点的写手，这真的很可笑，事实上，在大多数人的眼里LLM还只是个模糊的念头时，我就已经身处其中了。
But, yeah, it's it's so funny to be painted by the media as a as a as an AI, you know, a coattail writer when in fact, I've I've been been there before LLMs were were were were a glimmer in most people's eye.
是的。
Yeah.
只是因为约翰·卡马克告诉过我。
Only because John Carmack told me.
他告诉我AI一定会成功，所以我才相信了他。
He told me AI was gonna work, and that was why I believed him.
听他的话确实是个不错的选择。
That's that's a good person to listen to.
他是个值得信赖的人
He's a good person to
所以，一个重要的前提假设，我想听听你的看法，安德鲁，我这么说对吗？就是中美之间日益加剧的竞争。
listen So a big a big play kind of premise, think, of Andrew, tell me if I'm wrong, is is the growing rivalry between The US and China.
我听你在播客里谈过这个话题，但你觉得，以当前国防承包商和科技行业的发展趋势来看，我们并没有为这种竞争做好充分准备。
And I've heard you talk about this on podcasts, but you don't think that on the current trajectory with the defense contractors and technology industry that we're well set up for that.
对。
Right.
所以，也许你能稍微谈一谈。
And so maybe you could talk a little bit.
我非常想听听你对这种竞争的看法，特别是你如何看待这场科技竞赛中的劣势。
I would love to hear your thoughts on that rivalry and then specifically on how you see kind of handicap the technology race.
比如人工智能，还有国防、机器人技术。
So AI and, you know, defense, robotics.
我们花的钱太多了。
We're spending way too much.
但我们得到的回报太少了。
We're getting way too little.
我们的发展速度远不及地缘政治对手。
And we're not progressing at the rate that our geopolitical adversaries are.
如果这是和平时期，且我们或我们的盟友完全没有冲突的迹象，那或许还问题不大。
And if it were a perfect if it were peacetime and there was no inkling of conflict for us or our allies, that maybe wouldn't be a problem.
但这并不是我们所处的世界。
But that's not the world that we live in.
是的。
Yeah.
因此，与中国的竞争非常激烈。
And so the rivalry with China is is a big one.
它比我们刚开始创办安德罗尔时的范围更广；如果你回看最初的融资演示文稿，还记得那一页画着一只熊和一页画着一条龙。
It was a little broader than I mean, when we started Anderol, if you look back at the first pitch deck, you remember there's this page with a bear and a pair with page with a dragon.
我们列出了中国的路线图，也列出了俄罗斯的路线图。
And we lay out what the Chinese road map looks like, and we laid out what the Russian road map looked like.
我们当时明确指出，俄罗斯将入侵乌克兰。
And we literally lay out that Russia is going to invade Ukraine.
他们打算用武力夺取。
They intend to take it by force.
普京有一句非常精彩的话，他说，在人工智能领域获胜的国家将成为整个世界的统治者。
We there's a great quote by Putin where he says the country that wins in the sphere of AI will become the ruler of the entire world.
我对这句话非常钦佩。
I have a huge amount of respect for that quote.
首先，这种表达方式简直太有气场了。
First of all, like, what an incredible amount of of style there.
我并不是说我喜欢普京。
I'm not saying that I like Putin.
我非常不喜欢他。
I super don't.
我认为我们创办这家公司的目的就是为了对抗普京。
I argue we started a company to fight Putin.
我个人在俄罗斯和白俄罗斯都受到了制裁。
I'm personally sanctioned in Russia and Belarus.
我 literally 一去那里就会被俄罗斯安全部门逮捕。
I literally cannot go there without being arrested by Russian security forces.
所以别把这当成我会把帕尔默钉在墙上那样。
So don't take this in like the, I'm gonna nail Palmer to the wall.
他说了一些关于普京的好话。
He said something good about Putin.
但让我们坦诚地说吧。
But let's, like, let's be honest.
这真是相当有邦德反派级别的气场。
That's some really good, like, Bond villain level.
是的。
Yes.
他们会成为整个世界的统治者。
They'll become the ruler of the entire world.
你确实得佩服这一点。
You gotta respect that.
就像你说的，还挺不错的。
Along like you said, was pretty good.
那是对一群技术高中的学生讲话。
That speaking to a bunch of technical students at a technical high school.
你说的是很久以前吗？
A long time ago, you're saying?
是的。
Yeah.
那是2007年。
Back in 2007.
是'07年吗？
Like '7?
好吧。
Okay.
这人当时很紧张。
This guy was wet.
就像你说的，普京一定读了很多科幻小说。
Like, Putin must read a lot of science fiction.
是的。
Yeah.
那就是他
That's He
他起得特别早。
was up this really early.
我的意思是
I mean
记住，我们在2017年初的年度融资演示文稿中引用了这句话，那时AI还一点都不火。
Well, remember, we are quoting it in an annual pitch deck in early twenty seventeen before AI was cool at all.
所以，天啊，我真希望美国领导人能有这么超前的思维。
So like, man, I would kill for a US leader to be that far ahead.
你能想象吗，如果有人站出来？
Could you imagine if you had if you had somebody come out?
我甚至不会点名是谁，但如果你有个人从白宫走出来，说我们必须控制拉格朗日点。
I won't even name who, but if you had somebody come out in the White House, he we have to hold the LaGrange points.
必须有人占据轨道上的制高点。
Someone has to have the high ground in orbit.
如果我们打算从小行星带开采资源，地月空间的插入和转移站点必须得到保护。
The cislunar insertion and transfer pads must be protected if we're gonna extract resources from the belt, asteroids.
所以你知道我在说什么。
So you know what I'm talking about.
想象一下，如果这些想法也来自他自己的头脑，而不是别人写出来的，那会怎样？
And imagine if that was also could you imagine if that was like from his own mind and not not something that somebody wrote?
那就是普京。
That that is Putin.
他在这些方面是个有能力的对手。
He's a competent adversary in these ways.
所以，是的，我同意。
And so I I yeah.
这确实是竞争的重要组成部分。
It's like that was that was very much part of the rivalry.
问题是，当然，现在这种竞争已经进入了最终阶段。
The problem is, of course, that now that rivalry has kind of reached its terminal stage.
对吧？
Right?
他已经对乌克兰采取了行动。
He's made his move on Ukraine.
大家都看到了这件事正在发生。
Everyone's seen that it's happening.
因此，我们还能施加影响的只剩下中国了。
And so the remaining one that we can influence is China.
你提到了所有的无人机相关事情。
And you called out all the drone stuff.
我的意思是，你知道这些都会
I mean, you you knew that was all gonna all
这些东西。
the stuff.
没错。
That's right.
比如乌克兰的无人机。
Like, the Ukraine drone.
我们当时讨论的完全是这样，我的意思是，我们做的演示文稿准确预测了将要发生的事，其中指出，未来的战争将主要围绕这些广泛部署的系统展开。
We we we we were talking about literally I mean, our deck laid out exactly what was going to happen, and it said that warfare is gonna be about a lot of these proliferated systems.
是的。
Yeah.
很多情况下，战争将不再依赖那些需要数十年时间积累的高端精密装备。
A lot like, it's gonna move away from exquisite things that you build up capacity over decades.
你可以制造一些低端系统，其生产速度足以应对持续的冲突。
You can set lower end systems that you can actually manufacture at a rate that is relevant to an ongoing conflict.
当然，这正是俄罗斯目前正在发生的情况。
And, of course, that's exactly what's what what what's what's exactly what's happened in Russia.
但总的来说，2017年的时候，我们生活在一个许多人相信历史已经终结的世界里，认为不会再有重大的地缘政治变动，联合国只要发几封信就能阻止扩张主义行为。
But, you know, the the very big picture here is we were living in a world in 2017 where a lot of people believed we lived at the end of history, that there was no more geopolitical movement of significance to happen, and that United Nations could write mean letters to dissuade people from expansionist actions.
而我们现在看到的，正是那种更 cynical 的现实——我们正在见证那些 cynical 之人的胜利。
And that what we're seeing playing out is kind of the more cynical version of of of we're seeing the cynics.
那些 cynical 的人是对的。
The cynics were right.
那些说‘不’的人是对的。
The cynics who said, no.
战争仍然是当下的现实，希望它不是未来的常态，像Androl这样的公司能早早起步，这是一件好事。
War is still a thing of the present, hopefully not the future, and it's a it's a good thing that companies like Androl, I think, were were getting started early on this.
是的。
Yeah.
想象一下，如果Androl是在欧洲冲突爆发时才成立，或者更糟——是在台海冲突爆发时才成立，那会怎样？
Because imagine if Androl was started right as conflict kicked off in Europe or worse, being started right as conflict kicks off over Taiwan with China.
那就太晚了。
It's it's too late.
你不能在战争爆发后再开始研发武器，还指望能产生威慑作用。
You you you can't start working on bombs after the war has started and expect to have any deterrent impact.
你只会沦为战争的参与者，而不是阻止战争的人。
You're just gonna be part of fighting wars instead of preventing them.
你该怎么评估实力差距呢？
How do you, like, handicap?
我一谈到人工智能，总是听到中国和美国。
So I always I like, with AI, I always hear about China, US.
我很少听到关于俄罗斯的消息。
Not I don't hear Russia as much.
也许他们在这方面很厉害，我不知道。
Maybe they're good at I don't know.
但话又说回来，你该怎么评估实力差距呢？
But but then, like, how do handicap it?
嗯，事实上，今天确实如此。
Like Well, see, that's that that that's true today.
是的。
Yeah.
普京并没有那样想，他不是在说人工智能会偏向赢家，然后我赌美国会赢过我。
It's Putin didn't say that thinking they well, he wasn't saying, I think AI is gonna accrue to the winner, and I bet The US is gonna beat me.
他们投入了很多。
They were investing a lot.
只是他们没有走上获胜的道路。
It just they they didn't take the winning path.
对吧？
Right?
比如，他们没有去做一些这类基于注意力的模型、聚焦型模型。
Like, they they they didn't do some of these like, the the the kind of attention based models, focus based models.
只是他们没走这条路。
Like, it just that wasn't the path they went on.
他们没走上获胜的道路。
They didn't get the winning path.
所以你是对的。
And so you're right.
现在主要是美国和中国。
It's mostly US and China now.
俄罗斯正在努力追赶。
Russia's trying to catch up.
是吗？
Are they?
但他是在认为这将是他们的非对称优势时这么说的。
But he he he said this at a time where he thought that was going to be their asymmetric advantage.
他以为他们会远远领先于其他人，率先掌握所有那些致命无人机技术，并在别人还没反应过来之前就投入使用。
He thought they were gonna figure out all the all the killer drone stuff way ahead of everybody else, and he would have put it to work before anyone else even knew how to respond.
想象一下，如果普京推出了类似ChatGPT的东西，作为我向国会议员们解释的简化版本。
Imagine if Putin had made chat GPT is kind of the the the the simplified version I use for members of Congress.
是的。
Yeah.
但就像我所说的，美国拥有这种分散的资本主义体系。
But, like, with the the way I sort of like, America has we have this sort of decentralized capitalist system.
中国则是明显集中的命令与控制体系。
China has a centralized, obviously, command and control.
他们的人口更多。
They have more people.
他们非常聪明。
They're very smart.
深海技术以及其他这些方面都显示他们迅速赶上了。
The deep sea can all these other things show they caught up very quickly.
他们在制造业方面领先。
They're ahead of manufacturing.
是的。
Yep.
比如，另一方面，我认为美国的这种超级大国优势恰恰在于，现在媒体上人人都在抱怨的那些问题——比如超大规模企业刚刚投资了一万亿美元，这到底是个泡沫吗？
Like, yeah, we you know, on the flip side, like, I think this superpower of The United States, right, is the very fact that every I think everyone's complaining about in the press right now, which is we just like, the hyperscalers just invested a trillion dollars and, like and is it a bubble?
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。