大卫·柯特利，Helion能源创始人兼首席执行官

你当时筹了很多钱吗？

And did you raise a lot of money?

我正想说，投资者当时是怎么想的。

I was just gonna say, how did investors right.

我们当时收到了很多关注、很多关注，还有大量的投资意向。

We got a lot of interest, a lot of interest, and a lot of and and and a lot of offers at the time.

最终，我们发现——这又回到了专注的问题上——我们在Helion非常谨慎地控制了股东名单的规模和投资人数。

In the end, what we found was that, and this gets back to that focus part, is that what we have done at Helion is be very careful around how big the cap table is and how many people are are are invested in Helion.

因此，我们实际上并没有在那天或演示日接受任何投资意向，也没有就此达成任何交易。

And so we didn't actually end up, taking offers that day or during demo day, and and closing any closing around that.

直到九个月后，我们才与那些投资者中的几位，以及更多其他潜在投资者深入接触，对他们进行了尽职调查，确保他们真正理解这项技术，同时也让我们自己确信，他们对这项技术感到安心，因为这与以往演示日上人们见过的技术有些不同。

And so it wasn't until nine months later when we had worked with a number of those investors and then a a bigger pool also and done diligence with them, make sure they really understand the technology, gotten them comfortable with the technology, but also getting us comfortable that they're comfortable with the technology because this is a bit different than what what what maybe people had seen at demo day before.

因此，对我们来说，最重要的是：第一，拥有一小群投资者，因为我们知道，随着我们不断建造多个原型和可部署的聚变系统，我们将继续融资，目前已筹集了十亿美元。

And so it was really important to us to, one, have a have a smaller pool of folks because we knew we were we would go on to, we raised a billion dollars now, multiple rounds as we built built multiple prototypes and and deployable fusion systems.

第二，确保所有人对时间表、以及我们需要建造和部署的硬件与资本投入达成一致。

And then also make sure that everybody's on the same page about the timelines and, the amount of hardware and capital we're gonna have to to build and deploy.

我张着嘴坐在这里，因为这真是个大动作，大卫，你花时间深入了解投资者，甚至在这个过程中筛选掉那些可能不耐烦或制造麻烦的人吗？

I'm sitting here with my mouth open a little bit because this is such a power move, David, to, like, sort of, like, you know, spend time, really get to know the investors, and and sort of did you weed people out in the process saying, oh, they're just gonna sort of be impatient or cause problems?

还是说你真的深入了解了他们？

Or did did it did you really get to know them?

是的。

Yes.

没错。

Yep.

我们最终组建的这轮投资，只包括少数几位大型投资者，而且当时他们也通过了我们的筛选，现在回头看，我们是否还会做同样的决定，这确实是个好问题。

And and that and that the the round that we ended ended up building was of only a handful of larger investors, but also ones that were were spending the qualification for me at the time, and it's a good question of whether we we would make that decision now.

这可能带有一点事后诸葛亮的意味。

So this is maybe a little bit of hindsight.

他们愿意花时间去学习一些工程和物理知识，或者聘请懂工程和物理的人来做技术尽职调查。

Was that they were going to spend the time to either learn some of the engineering and physics or hire somebody that's gonna learn the engineering and physics and do technical diligence.

有意思。

Interesting.

是的。

Yeah.

对我来说，我当时的想法是，我从未谈论过这一点，不管怎么说。

And for me, the the thought process there I've never talked about this for what it's worth.

这真是个非常好的问题。

This is super this is a really great question.

我知道我们面临很多大问题需要解决，这需要在技术和设备上不断迭代，而我们当时就在埃弗雷特工厂。

Was that I knew that we had a lot of big problems to solve and that it was going to take iterations of both technology and machines as well as we're here in Everett Building.

我们的制造面积从5000平方英尺扩大到了现在的50万平方英尺。

We went from 5,000 square feet to 500,000 square feet now of manufacturing.

这需要大量的研发、时间和耐心。

And so and that was gonna take some real development and timeline and patience.

是的。

Yeah.

耐心才是关键。

Patience being the operative word here.

有意思。

Interesting.

我可以问一下，那次演示日之后的首轮投资人都有谁吗？

Who were can I ask who those investors were in that first sort of post demo day round?

萨姆当时参与了吗？还是他更早之前就投资了？

Was Sam in it or he had invested earlier?

那时候，萨姆并没有参与。

So so at that time, Sam was not.

所以，当时YC是领投方，我们的首轮投资由Mithril资本主导。

So Y Combinator was, so, our first round was led by Mithril Capital.

然后我们还有的投资方是

And then and then we had

哦，那是彼得·蒂尔。

Oh, that's Peter Thiel.

那是彼得·蒂尔。

That's Peter Thiel.

对吧？

Right?

好的。

Okay.

然后我们有了Capricorn投资集团，Capricorn主导了我们的B轮融资。

And then we had Capricorn Investment Group, and and Capricorn led our series b round.

我们已经公开谈论过那些参与这一轮的其他人士。

And we've already talked about publicly the other folks that that were in that.

我们还有另外六位在Demo Day期间投资的人。

We had another half dozen folks, that were at Demo Day that had invested.

好的。

Okay.

那么继续。

Moving on then.

所以听起来，你在Y Combinator的时间花得很值得。

So it sounds like your time with Y Combinator was well spent.

对我来说，对于业务来说，简直是极其值得的。

Oh, for me, for the business, fabulously well spent.

好的。

Okay.

所以，即使到现在，当我们为海利昂引进高管和领导者时，我们仍然在做这件事——这又回到了创始人心态，这也是为什么我抱歉没见到你，因为我觉得那本会是我们共度的一段绝佳时光。

And so we we have and we we, even now are, continue to as we bring in executives and leaders at Healion, how do we this comes back to founder mode too, which is why I'm sorry I missed you because I think that would have been a really fabulous time together.

快跟我们说说你的创始人心态吧。

Tell us quickly about your founder mode thing.

每当有新的高管和领导者加入时，我们首先会专注于招聘，确保我和克里斯都参与招聘流程，因为我们希望继续在公司内部培养这种文化。

We still, as new executives and leaders come on, one, focusing on on hiring to make sure I'm in Chris and I are in the hiring process because we're looking to continue to grow that at at the company.

我认为保持创始人心态、聚焦于真正重要的事情、尽全力推动进展、绝不找借口说事情应该慢一点、或者我们做不到、或者解决不了这个问题，是绝对至关重要的。

I think it's absolutely critical that keeping that founder mode mindset, focusing on what matters, absolutely driving to that as hard as you can, taking no excuses for why things should be slower or we can't do up do that or solve that problem.

然后，尽管有时对团队来说很烦人，但克里斯、我、乔治和其他创始人都会深入细节。

And then getting in the weeds, as annoying as it is sometimes to, the teams, Chris and I, George, the other founders, we are in the weeds.

当出现问题时，我们会亲自去解决，而不是仅仅指望它能自己解决。

When there's a problem to be solved, we're gonna go and and and help solve that problem and not just hope it gets done.

所以，我希望我的所有领导者都能这样。

And so and we I want all my leaders to be that way.

所以你在2014年夏天参加了Y Combinator，你知道，你之前就已经开始创业了。

So you did YC in in summer fourteen, you know, you started a little bit before that.

你已经做这件事很久了。

You've been doing this for a while.

随着公司的发展，你是否需要学习如何成为一位领导者和管理者？

Did you have to train yourself on how to become a leader and a manage manager as your company grew?

当然，我必须不断成长。

Definitely, I have had to grow.

这绝对是毫无疑问的。

Like, that is that is absolutely for sure.

但其中一部分，我认为也在于我的个人成长经历了不同的阶段，曾经一度偏离了创始人模式，后来又回归了。

But part of it, think, has also been I've had different development periods in my personal growth where I actually strayed away from founder mode a little bit and have come back.

因此，当我们加入Y Combinator时，Helium的创始人们已经是一个团队了。

And so so at the time that we joined Y Combinator, the founders at Helium, we are we are a team.

那时我们不到十个人，现在却已超过五百人。

We are less than 10 people then, and we're well over 500 people now.

当时我们非常亲力亲为，即使还不知道这个词，我们确实处于一种高度参与的状态——每个人都深度投入业务， literally 亲手拧螺丝、组装机器，同时密切关注每一个细节，以便快速执行，让机器运转起来并实现聚变。

We we were very much, whether we knew the terminology at the time, in that mindset of very hands on, they're all very involved in the business, you know, literally hands on to the point of where we were turning wrenches and building machines, but also really paying attention to the details so that we could execute quickly and make the machines work and and do fusion.

把这些要素整合在一起，我们当时非常专注于这些，但并没有真正思考过框架。

And and and so putting all those pieces together, we were really focused on that, but we hadn't really thought about the framework.

总之，我们只是自然而然地在做这些事。

Anyway, we're just sort of naturally doing it.

在建造Helion、扩展规模、并建造多台机器、不断刷新聚变世界纪录的这些年里，这种状态有时会有所偏离。

Over the years of building Helion, scaling, building now now multiple machines that then set continual world world records for fusion, that is sometimes drifted.

很容易放松下来，进入一种更商业化的、传统的商业思维模式。

It it it's easy to relax into more of a an a business mindset, more of a traditional business.

松开缰绳，说：我就让这部分业务自己运行，不再那么专注地投入其中。

Take your hands off the reins and say, like, I'm just gonna let this part of the business do its thing and not really be really, you know, focused on it.

而某种程度上，我们后来对此感到后悔。

And then and then in some ways, we've we've come to regret that.

有些时候我回过头来看，简直不敢相信。

That there have been pieces where I look back and and I'm like, woah.

我们引入了官僚主义和流程。

We've put bureaucracy and process.

我们变慢了。

We slowed down.

我们没有加速。

We didn't accelerate.

我们变慢了。

We slowed down.

我们为什么会这样做？

Why did we do that?

多年来，有一位顾问一直问我：大卫，你是艾森豪威尔还是巴特？

And had to come back an adviser for many years who said, David, are you Eisenhower or Pat?

你是从幕后协调全局，然后让你的将领们放手去干、向前推进吗？

Do you lead from behind the scenes orchestrating the field and then letting your generals go off and and and and, you know, move move the the go forward?

还是你会深入前线？

Or do you get in the trenches?

你会钻进坦克、冲进战壕，亲自在前线带领作战吗？

Do you get in the tank and in the trenches and then actually go and and lead on the front lines?

他们花了多年时间告诉我，我应该做艾森豪威尔。

And then they and they spent many years telling me that we should I should be Eisenhower.

我应该躲在幕后，设定最佳方案，交出去，然后让它自动执行，根本不予关注。

I should be behind the scenes, set us the best set us specifications, hand it off, and let it just execute by itself and not not pay attention to it.

结果发现，这并不是我的风格。

It turns out that's not me.

结果发现，这也不是 Healing On 所有创始人的风格。

And it turns out that's none of the founders at Healing On.

而是要冲到前线，深入其中，解决问题。

That it's it's get to the front lines and get in there and solve the problems.

然后培训他人，激励他们，提供他们需要的资源，再让他们放手去干，因为这边又有个新问题、另一场战斗、另一场火要扑灭、另一条生产线要建造、另一个招聘渠道要打通、另一轮融资要进行——接着跳上去，全身心投入解决那个问题。

And then train folks, motivate them, give them the resources they need, and then let them go continue the thing because there's something over here, another another battle to fight, another fire to put out, another manufacturing line to build, another hiring pipeline to open, another fundraise to do to go and then and and jump onto that and put all your all your effort into that that thing and solve that problem.

然后组建团队去执行，之后他们就能自己运转了。

And then build the team to go do that and then they can go run.

然后就这样去构建和发展业务。

And then and then go and build and build the business that way.

结果发现，这种模式不仅让我这个人性上产生共鸣，而且我也见过它是最有效的。

And that turns out that one that resonates with me as a human, but it also has been I have seen the most effective.

我经常被人说这种方式无法规模化。

And I've been told it doesn't scale.

人们总是这样告诉我。

People tell me that all the time.

但我认为这并不对。

And I don't think that's right.

我认为，只要你离开去解决业务的某一部分时，留下的组织仍然保持着创始人的状态，以高速甚至加速前进，如果我做得好的话，你就可以暂时去专注于其他事情。

I think that as long as what you're leaving behind when you go solve some part of the business is a part of the organization that is still operating in founder mode and going forward at velocity at high speed and hopefully accelerating, if you've if if I've done a good job, then you can go off and focus on something else for a while.

然后隔一段时间检查一下，看看进展如何。

And then check-in every once in while, see how it's going.

你有没有想过，利用这项技术去做一些更简单、更快、更赚钱的事情？

Do you ever get tempted to take this technology and work on something that's, like, easier and faster and lucrative?

因为

Because

嗯。

Mhmm.

因为听起来你一开始也是这样，直到你真正专注于此。

Because it sounds like that's how kind of how you started until you really focused.

所以我只是想知道，这种诱惑有多大——可能来自你自己，也可能来自外部投资者之类的，去选择更容易的路。

And so I just wonder how much, like, temptation is there, maybe from yourself, but also from outside investors or whatever to, like, do the easy thing.

一直都有外部压力要你这么做。

There is constant external pressure to do that.

一个具体的例子是我们使用一种叫氦-3的燃料，这是一种稀有的氦同位素，安全且清洁，但非常稀少。

A a a specific one is we use a fuel called helium three, a rare isotope of helium, that is safe and clean, but pretty rare.

我们很早就发现，事实上，这并不是我们发明的。

And we very early found in fact, we didn't invent this.

使用这种燃料是聚变领域最早的想法之一，但如果不使用它，大多数人会选用其他燃料。

This is one of the earliest ideas in fusion is using this fuel, which is most people if you don't use this.

他们使用的是不同的燃料。

They use a different fuel.

人们不使用它，是因为它非常稀有且昂贵。

And and people don't use it because it's pretty rare and expensive.

因此，我们很早就开发出了一种在地球上生产它的方法。

And so what we developed very early is a way to make it, to actually make it here on Earth.

我们已经证明了可以生产它。

And we've shown we can make it.

我们现在有一整套流程来过滤和分离它供自己使用，但我们也完全可以拿去出售。

We have now have a whole process to filter it and isolate it for ourselves, but we could go sell it.

我们今天就可以去卖它。

We could go sell it today.

如果你仔细看的话，它的总潜在市场规模可能达到数亿美元。

And it's, you know, total total addressable market, if you really squint at it, is maybe several $100,000,000.

所以这是一个合理的市场，它很有用，对人类有益。

So it's it's a reasonable market, and it's and it's useful and would be good for humanity.

事实上，它被用作医疗成像产品，但这并不是我们的目标。

In fact, it's used for medical imaging as as as a as a product, and it's not our goal.

它并不是一个万亿美元的市场。

It's not the trillion dollar market.

它不是为全人类提供清洁、安全的电力。

It's not clean, safe electricity for all of humanity.

它非常小众，但我们今天就能做到，并且已经有客户主动来找我们寻求它。

It is very niche, But we could do it today and and and get and and have customers that have come to us asking for it.

但我们拒绝了。

And we say no.

我们拒绝是因为，第一，它并不能解决我们想要解决的问题。

We say no because, one, it's not it doesn't solve the problem we want.

我担心的是，我见过一些企业被早期阶段那种更小、更容易获取的产品所分散注意力，然后就此止步。

And I worry that I've watched businesses get diverted by that early stage, much smaller, easier to grab, product, and then stop there.

我们不想那样做。

And and we don't wanna do that.

另一方面，我也关注市场动态，看看这类市场是如何变化的。

The other part is I also look at market dynamics is that you look at how markets like that change.

是的。

Yes.

现在这非常赚钱。

It's super lucrative now.

很好。

Great.

你去解决那个问题。

You go solve that problem.

我降低成本，很快它就没那么赚钱了。

I drive costs lower, and now very quickly, it's not as lucrative.

我认为我们拥有这种自由。

And and I think we've had the freedom.

这让我们回到了之前讨论过的一个话题：因为我们选择了一些投资者，以及我们所做的工作，所以我们拥有说不的自由。

This gets to something we were talking to earlier, but it gets we've had the freedom to say no because of some of the investors that we've picked, because of the work that

我们是的。

we're Yeah.

是的。

Yeah.

如果他们只关注三年或甚至五年的回报，他们就会对我们施加很大压力，那样我就可能无法说不。

If they were interested in a three year return or even a five year return, then they would be pushing us very hard, and I may not be able to say no.

没错。

Right.

没错。

Right.

是的。

Yeah.

通过真正聚焦于那些将全球清洁能源部署视为市场的群体，这为我作为领导者和首席执行官提供了空间，让我能够带领团队专注于这一目标。

By by really focusing folks that really are looking towards that global scale deployment of clean electricity as as the market, that that has given me the space as a leader and as the CEO to be able to focus the team on that.

在Helion，这是一个每月进行的对话。

And and it's a monthly conversation at Helion.

是的。

Yes.

所以你说过，聚变这个想法早在三十年代就开始了。

So you were saying that, you know, the idea of Fusion was started back in the thirties.

他们当时在研究，并做了大量出色的工作。

They were researching it and did all this great work.

多年来，人们总是说，聚变还有二十年才能实现。

And over the years, it's always been like, oh, Fusion's twenty years away.

还有二十年。

Twenty years away.

是什么发生了变化，使得现在它不再还有二十年那么远了？

What has happened that now makes it not twenty years away?

事实上，根据你的说法，由于这笔微软交易，它现在只有三年之遥了，对吧？

In fact, according to you, it's three years away, right, with this Microsoft deal?

到底是什么发生了变化，才让这一切成为可能？

Like, what what has changed to make that happen?

没错。

Yeah.

今年早些时候，我们为微软的发电厂破土动工了。

We broke ground on our power plant for Microsoft earlier this year.

因此，我们正在全力推进，并相信我们能在2028年实现并网供电。

And so we are absolutely pushing forward and believe we can deploy electrons on the grid by 2028.

推动这一进展的是其他技术。

And what's enabled that has been other technologies.

在核聚变领域，已经取得了巨大的进步。

There's been a tremendous amount of progress in fusion.

今天我们对上亿度等离子体物理的理解，早已不同于二十年前，更不是四十年前的水平。

Our understanding of the physics of those 100,000,000 degree gases today is not what it was twenty years ago and was is not what it was twenty years before.

因此，这方面已经开展了大量卓有成效的工作。

So there's been a tremendous amount of that work that's being has been done.

我绝不能忽视这一点，即这种进展有多么伟大，以及支持这项基础科学的政府资金有多么重要。

And I cannot never discount how great that has been and how great some of the government funding that supported that, the basic science, has been.

但与此同时，摩尔定律也发生了作用。

But in parallel to that, Moore's Law happened.

计算成本或我们对计算能力的理解——即能够封装到芯片和硅中的功率——在过去十年中发生了翻天覆地的变化，这正是摩尔定律几十年来对我们产生的影响。

That the price of compute or the way we think about it is the amount of power that you can pack into a chip, into silicon, has so radically changed in in the intervening decade, literally decades of Moore's Law that that has applied to us.

我回想起上世纪五十年代和六十年代那些在聚变领域工作的人，我对他们的成就感到无比钦佩，因为那时计算机根本还不存在。

I think back to what the folks in the nineteen fifties and sixties did in Fusion, and I am, like, in awe of that work because you had work with the computer didn't exist.

那时他们只能依靠手工计算极其复杂的物理问题，而这些工作我们如今连尝试都不会用手算了。

So this was, like, hand calculations of really tough physics, things that we we we wouldn't even try to do by hand anymore.

他们打算建造能够实现聚变的机器——我们还没怎么讨论过聚变的机制，但需要巨大的电力。

And they were gonna build machines that did fusion by we haven't talked about the mechanisms of fusion too much, but large amounts of electricity.

事实上，我们现在的Polaris机器目前峰值电力输入已超过100吉瓦，同时从机器中输出大量电力。

In fact, they are able to take our current Polaris right now is running at over a 100 gigawatts of peak electricity into the machine and large amounts of electricity coming out of the machine.

他们在晶体管尚未发明的时代，就已经能够建造出在极短时间内切换大电流的系统。

And they were able to build systems before the transistor existed that could commute large amounts of current in very short periods of time.

而且是用真空管。

And it's with vacuum tubes.

这对我来说简直不可思议。

It's absolutely amazing to me.

而在这一期间，硅技术已经发展到我们可以——我这里给出一些具体数据。

And in the intervening time, silicon has now evolved to where we can I'll give you hard numbers here.

我们需要能够在纳秒级别内接通电力，具体来说是十纳秒的时间。

We need to be able to to turn on electricity in we measure it in nanoseconds, in ten nanoseconds of time.

我们必须能够做到这一点，而当我刚入行时，这非常困难。

We have to be able to and when I started my career, that was very hard to do.

十纳秒低于一吉赫兹。

Ten nanoseconds is less than a gigahertz.

一吉赫兹的CPU现在已经是现货产品。

A gigahertz CPU is off the shelf.

你现在哪里都能买到。

You can get it anywhere now.

但以前这真的很难。

But it used to be really hard.

我们每次进行聚变时需要收集的数据量达到数十吉字节。

We have the amount of data we have to collect for every pulse where we do fusion is measured in tens of gigabytes of data.

而且这一点从未改变。

And and that hasn't changed.

一直以来都是这样。

That has always been the case.

我们一直都需要这么多数据。

We've always needed that amount of data.

但当我刚入行时，我们谈论的是千字节和兆字节，而现在已经完全不同了。

But, again, when I started my career, we talked in in kilobytes and megabytes, and and and that's just not the case anymore.

我们现在真的能够处理这些数据了。

It's we can we can actually handle that.

所以我们现在不再谈论吉字节了。

And so we're not talking about gigabytes anymore.

我们现在谈论的是太字节级别的数据，没错。

We're talking about terabytes of data Yeah.

这些数据是在Healion收集的。

That we collect at Healion.

这些都是我们现在能够做到的事情。

And that's that's things that we can do now.

另一个重大进展是光纤技术。

The other big one is fiber optics.

在这些系统中，比如在Polaris中，我们有数万个大型半导体器件在进行通信和切换，以处理电力。

Is that in these systems, we have, in Polaris, for instance, we have tens of thousands of of large scale semiconductors that are commuting, that are switching for the doing the electricity.

每个器件都由一根光纤控制。

And each one is controlled by a fiber optic.

而这根光纤必须在纳秒级的时间内做出响应。

And that fiber optic has to respond again in nanoseconds of time scale.

这在过去真的非常困难，但如今你可以从百思买买到一块10吉比特的光纤网络卡，其响应速度远低于一纳秒。

That used to be really, really hard, but you can go by a 10 gigabit, So that's way less than a nanosecond fiber fiber networking card from Best Buy today.

而且这彻底改变了我们诊断这些系统、理解其内部运行状况、控制和操作它们的能力。

And and and and it's just it has just totally transformed our ability to to diagnose these systems, to understand what's happening inside of them, to control them, to run them.

因为物理原理从未改变。

Because the physics has never changed.

这种高速的上亿度气体一直移动得如此之快，我们需要纳秒级甚至几十纳秒级的时间尺度来应对。

The gas, this high speed 100,000,000 degree gas has always moved so fast that we needed one nanosecond or tens of nanoseconds of time scale.

而现在我们可以做到了。

And now we can do it.

此外，高速计算本身也带来了变革，我们在机器各处分布了大量可编程逻辑单元。

And also high speed computing just in itself, we have a small programmable logic brain on every single distributed all over the machine.

过去我们需要大型超级计算机来实时理解这些系统中的情况，但现在我们不再需要了。

We used to need big supercomputers to understand what was happening real time in these, but we don't anymore.

我们只需在机器各处部署小型可编程逻辑单元，并通过以太网构成一个大型网络总线来控制整个系统。

We just put a little programmable logic all over the machine and there's a large, you know, network bus that that runs the whole thing over Ethernet.

这使我们如今能够实现二十年前非常困难且昂贵的任务。

And that has now enabled us to do what we we what was twenty years ago very, very hard and expensive to do.

二十年前，这根本是不可能做到的。

And twenty years before that was impossible to do.

哇。

Wow.

这太酷了。

That is so cool.

给我们讲讲你们最新的机器吧。

Tell us about your latest machine.

我们的第七代机器叫做北极星。

Our seventh generation machine is called Polaris.

我们以一颗进行氦聚变的恒星命名它，它同时也是北极星，指引之星。

We named that after a star that does helium fusion, also happens to be the north star, the guiding star.

北极星。

Polaris

而北极星正在运行

And Polaris lives

我们称这座建筑为大熊座，以小熊座命名。

in a building we call Ursa, named after Ursa Minor, the little dipper.

那就是北极星所在的星座。

That's the constellation that Polaris lives in.

我们就是这样为工厂和工厂内的发电机命名的。

That's how we name the plants and we name the generators themselves that go inside the plants.

我们去年完成了大部分机械建设，并于十二月开始为北极星产生等离子体。

We finished most of the mechanical construction last year and we're able to begin making plasmas in December for Polaris.

今年我们为北极星连接了更多的电力，不仅制造数百万度的聚变等离子体，现在还将其加热到数千万度。

We spent this year now hooking up more power to Polaris, not just making these fusion plasmas running at millions of degrees, but now superheating them to tens of millions of degrees.

在北极星项目中，它现在已经完全建成，我们已经以全功率运行聚变反应数月了。

And and in Polaris, we are now have it's fully complete, and we've been running now at full power doing fusion for for several months now.

现在进入优化阶段。

And so now it's an optimization.

这是微调。

It's tweaking.

正在提升功率。

It's getting the power up.

我们目前专注于输出功率。

It's getting the power out that we're focused on.

我有个问题。

I have a question.

听起来制造能量需要消耗大量的能量。

It sounds like it takes a ton of energy to make energy.

这有点像是从五万英尺的高度来看。

This is sort of like that 50,000 foot.

但这里的目的是，除非我完全误解了整个对话，你们的输出能量远大于输入能量，这个比例是多少？

But it the the goal here, unless I've really misunderstood this whole conversation, is that you make so much more in the output than you did in the what is the ratio?

融合领域大多数人试图实现的是产生巨大的能量输出与输入之比，也就是我们所说的产额。

What most folks in fusion are trying to do is make fusion reactions that are have tremendous, what we call yield, energy out per energy in.

我们的目标是尽可能高效地利用输入的每一度电。

What we're trying to do is a little bit Our goal is to be very, very efficient with every all all the electricity we put into it.

因此，我们在2014年Y Combinator期间证明了这一点：我们可以从墙上取电，输入到聚变电磁铁——也就是执行工作的部件——然后回收其中96%的电能。

And so this is what we proved, in 2014 during our Y Combinator time, is that we could take electricity, from the wall, put it into fusion electromagnets, the the parts that do the work, and then recover 96 of that electricity.

哦。

Oh.

在你进行任何聚变之前。

Before you've done any fusion.

所以现在，聚变部分只需要补上剩下的4%以及一点点额外的电量用于出售。

So now the fusion part only has to make up that last 4% and a little bit extra to sell.

大多数聚变领域的研究者会损失他们输入的大部分能量以及输出的大部分能量。

Most folks in fusion lose most of the energy they put in and most of the energy they pull out.

所以他们的效率，如果我客气一点说，大概只有10%。

So they're in the efficiency of maybe, if I'm being generous, 10% efficiency.

很多人效率甚至低于1%。

Most a lot of folks are less than 1% efficiency.

这意味着聚变过程必须弥补所有这些损失。

And what that means is the fusion process has to make up all that loss.

它必须更具能量。

It has to be much more energetic.

它必须更大、更复杂、更昂贵，建造时间更长。

It has to be bigger, more complex, more expensive, take longer to build.

因此，通过专注于它的效率，我们究竟关心什么？

And so by focusing on efficiency of it, of of what do we care about?

产品是什么？

What's the product?

不是聚变。

It's not fusion.

而是电力。

It's electricity.

所以我必须尽可能节省每一度输入的电能，尽可能多地回收输出的电能。

And so I better be saving every elect electron I put in, I should pull out and save as much as I can.

因此，对我们而言，聚变的目标现在是输出大约是输入电力的1.5到2倍。

And so therefore, our goal is now fusion, for us, we wanna make on the order of one and a half to two times the electricity we put in as output.

大多数聚变领域的人士，能量倍数在30倍到300倍之间。

Most people in fusion are what, 30 times to 300 times the energy.

因此，这正是我们追求的目标。

And so that that's the goal for for what we do.

而在Polaris项目中，今年早些时候我们启动它时，已经能够回收我们输入电能的90%以上。

And in Polaris right now, when we turned it on earlier this year, we're already recovering over that 90% of that that electricity that we are putting.

这太惊人了。

That's amazing.

哇。

Wow.

我感觉从头到尾这场对话中，我的嘴都张着合不拢了。

I feel like I've been sitting here with my mouth hanging open this whole conversation.

我学到了很多，大卫。

I'm learning so much, David.

我太兴奋了。

I'm so excited.

我觉得自己对聚变的了解比以前多了一点，而以前我几乎一无所知。

I feel like I'm gonna know just that a little bit more about Fusion than I used to, which was not much at all.

所以这是真实的。

So this is real.

你和微软签了这份协议，我还注意到里面加入了违约条款。

You signed this deal with Microsoft with I might add a penalty clause in

是的。

it.

嗯。

Mhmm.

也许你不想回答这个问题，但如果从现在到2028年之间出了什么问题，你认为最可能出错的是什么？

Maybe you don't wanna answer this question, but if something were to go wrong between now and 2028, what do you think is gonna be the thing that's go goes wrong?

你能预测吗？

Could you predict?

能。

Yeah.

我的意思是，我想告诉你的是，几乎可以肯定的是，并不是每件事都会按照你预期的方式发展。

I mean, what I'll tell you is it's it's almost a thing you know, is that not everything is gonna work the way you think it will.

所以当我们做前人从未做过的事情时，真正需要明白的是，无论是在商业上还是工程上，都会出现意想不到的延误。

And so what we like, that's that's the thing you really know when you're doing something that no one's ever done before, is there will be unexpected delays whether they're business or they're engineering.

你会学到一些新东西。

There'll be new things you learn.

有些会加速你的进程。

Some will accelerate you.

有些则会拖慢你的进度。

Some will decelerate you.

因此，我们建立这个企业是为了应对问题，而这正是创始人模式的一部分。

And so what we've done is we've built the business to solve problems, to be and that's that's part of founder mode.

并不是说：‘哇，’

It's not saying, wow.

出现了一个问题。

There's a problem.

我解决不了。

I can't solve it.

但我要怎么解决呢？

But it's how do I solve it?

或者我认为这可能是个问题。

Or or I think that might be a problem.

我该如何在今天进行投资，以便如果明天出现问题，我能在明天解决它？

How do I invest today so that if it's a problem tomorrow, I will be able to solve it tomorrow?

因此，我们之所以能成功打造出六代机器，就是试图尽可能远地预见未来并为此做好规划。

And so we've now that's how we've been successful of building six generations of machines is trying to see as far ahead as we can and plan for that.

因此，我担心的事情正是我们刚才讨论的很多内容。

And so the things I worry about are the things a lot of the things we're talking about.

如果我们不是95%的效率，而是94%的效率呢？

What if we're not 95% efficient but we're 94% efficient?

如果我们只有93%的效率呢？

What if we're 93% efficient?

我们如何构建能够实现这种演化的系统？

How do we build the systems to be able to evolve that?

如果我们从未遇到过这个问题，而我对此一直非常幸运，但如果我们遇到许可延迟怎么办？

What if, we haven't had this this problem and and and I've been very, fortunate for it, but what if we have delays in permitting?

如果我们实际施工过程中出现延误怎么办？

What if we have delays in actually doing the building?

在购电协议中存在财务罚款。

There are financial penalties on the on a PPA.

我认为，你与客户（比如微软）签订的任何真实协议都应包含这一点。

I think any real agreement that you have with with your with your customer, and Microsoft's our customer, should have that.

就像我们应当让企业对其承诺的交付内容负责。

Like, you did like, we should be holding businesses accountable for delivering what they what they say.

因此，对我们来说，这是一个要求：我们希望找到一个客户，其合同中明确包含这一条款。

And so that was a requirement for us, is that we wanted a customer that would that would that would have that as part of the contract.

因此，现在需要建立相应的系统。

And so it's now building the systems in place.

好的。

Okay.

我们的效率是93%。

We're 93% efficient.

我们该怎么做？

What do we do?

我们现在已经建成了比最低需求更大的制造能力，因为我们还需要更多的电力电子设备。

Well, we've built now the manufacturing capacity that's bigger than what we needed for the bare minimum because we're gonna need some more power electronics.

我们需要进行更多的聚变。

We're gonna need to do more fusion.

我们必须再加把劲。

We're gonna have to squeeze a little bit harder.

我们必须达到更高的温度。

We're gonna have to get to a little bit higher temperature.

所以现在就让我们开始建设这些系统，以便能够应对这些需求，并在学习过程中建立相应的制造和工程能力。

So let's now build those systems already so that we can handle that and build the manufacturing and and the engineering to be able to support that as as we learn.

因此，很多工作在于尽可能多地思考那些可能不会完全按计划进行的事情，以便你能设计出可以调整的旋钮，从而进行修正。

And so a lot of it is starting to think as much as you can around what are the things that may not work out exactly as you as you planned so you can build the knobs to tweak that, so you can adjust.

然后保持一种以解决问题为导向、充满热情且乐观的公司心态。

And then keeping a company mindset that's problem solving based and excited and optimistic about solving problems.

在某些方面，这种激励因素甚至比制造环节更难应对。

And in some ways, that motivational piece is almost the harder piece than the manufacturing piece.

所以你刚才说，你是个解决问题的人。

So you just said, like, you're a problem solver.

你会不断迭代。

You iterate.

你会不断构建。

You build.

但有时候，你会遇到一些完全无法控制的问题，比如政府或监管方面的因素。

But sometimes there's problems that you have literally no control over, government, you know, regulatory, that kind of thing.

你有没有遇到过那种让人惊呼‘天哪’的挫折？

Did you have any kind of setbacks where it's like, oh my gosh.

他们不会让我们做这件事，而且我没有任何应对办法？

They're not gonna let us do this thing, and I have no workaround?

你有没有过这样的时刻？

Did you ever have a moment like that?

我认为在我们创立公司时，确实有几件让我们非常非常担忧的事情。

I think when we founded the company, there were a few of those we were really, really worried about.

那些关乎公司生死存亡的问题，我几乎完全无法控制。

The the existential company killers that I don't have any that that I don't have much control over.

监管就是其中之一。我的意思是，我和萨姆肯定十年前就讨论过这个问题：即使我们能解决所有技术难题，打造出最终实现核聚变的原型机，也可能被归入传统核裂变电站的监管路径，导致无法快速部署——因为一台设备的审批要花十年，而在开始建造之前，就得投入十亿美元做各种研究，尽管技术和物理原理根本不需要这么繁琐。

And regulatory was one of those where, I mean, I I think we Sam and I, I'm sure, talked about this ten years ago, which was even if we were able to solve all the technology problems and build those prototypes that we eventually did that did fusion, you might get regulated in a path of, like traditional nuclear fission power and not be able to deploy quickly, that it takes ten years to license a machine, that it takes a billion dollars in in studies before you're ready to go build something, even though technology and the physics doesn't doesn't warrant that.

所以我们开始投入资源。

And so we started investing in it.

我们开始真正地建立关系，与监管机构会面。

We started doing building the real honestly, building relationships, meeting those regulators.

我记得，是的。

I remember Yeah.

2018年，我联系了负责监管我们的华盛顿州卫生部。

2018, I called the Washington Department of Health who handles who now is our regulator.

但当时，他们监管的是医院中的粒子加速器。

But at the time, they regulated particle accelerators in hospitals.

我说，你们还不了解聚变，但你们就是我们的监管机构。

I said, don't you don't know about fusion yet, but you're our regulator.

这是我们的第五代装置提案，以及我希望你们如何监管它，这是支票。

Here is our proposal for our our fifth generation machine and how I want you to regulate it, and here's the check.

让我们一起启动这个流程吧。

Let's go start the process together.

我们花了多年时间与他们合作，理解他们的关切点，同时让他们也理解我们真正的安全顾虑。

And it took years of us working with them to to understand what drives them, then them to understand us of what are of the real safety concerns.

这是大型工业电力。

This is big industrial power.

有一些事情

There's things

我只是想问一下。

Well, I was just gonna ask.

我们都熟悉核能的担忧，而且我们一开始就已经讨论过这个问题。

We are all familiar with what the concerns are with nuclear power, and we actually talked about it, at the beginning.

但人们对聚变能源担心的是什么？

But what's the what's the thing that people are worried about with fusion?

比如，最大的灾难会是什么？

Like, what's the big disaster?

是因为它又重又热而发生爆炸吗？到底是什么？

Is it just an explosion because it's so heavy and hot and like, what is it?

我们现在运行Polaris时最大的担忧主要是高压电。

The biggest concerns we have in running Polaris today is mostly high voltage.

是电子设备的问题。

It's the electronics.

因此，当我们运行这些设备时，会涉及大量的脉冲功率，大量储存的电能会在极短时间内——高达100吉瓦——输入到聚变装置中。

And so when we run these machines, we have large amounts of of pulse power, large amounts of stored electricity that then very quickly, 100 gigawatts worth goes into the fusion machine.

然后，大量的电能会迅速从设备中释放出来。

And then very quickly, large amounts of electricity come out of the machine.

因此，我们不希望人类处于这种环境中。

And so, we don't want humans in that environment.

所以，当我们运行Polaris时，那27,000平方英尺的建筑里一个人也没有。

And so so so that's the biggest one where we actually when we're running Polaris, no one's in the building in that 27,000 square feet.

但这些还不是最难应对的。

Those aren't the hardest ones though.

这才是最大的实际安全风险。

It's probably the biggest real safety concern.

这仍然是一种核反应过程。

This is still an atomic process.

你正在将原子融合在一起。

You're fusing atoms together.

就像医院里的粒子加速器一样，你会产生电离辐射。

And much like a particle accelerator in in a hospital, you do make ionizing radiation.

会产生X射线和中子。

X rays, you make neutrons.

这些会从机器中释放出来。

Those come off of the machine.

因此，你需要屏蔽人类，使其远离这些辐射。

And so you wanna shield humans from that.

实际上，当我去卫生部门时，我告诉他们：这是我们的屏蔽许可。

It's actually, an interesting story of when I went to the Department of Health and I said, this is our shielding permit.

这是我们保护人类免受机器辐射的方法，这些机器就像产生电离辐射的粒子加速器。

This is how we're gonna protect humans from the machines that are that are making like your particle accelerators that are making ionizing radiation.

他们说：很好。

They said, great.

病人在哪里？

Where is the patient?

因为我们的计算总是假设房间内有病人与机器在一起。

Because our calculations always have a patient inside the room with the machine.

我说，没有病人。

And I said, no patient.

他们都在房间外面。

They're outside the room.

就像在医院里一样，我们会有一个我们称之为屏蔽室的地方。

Like, we're gonna have just like in a hospital, we'll have what we call a vault.

我们会有一个带门的混凝土房间，门可以关闭。

We'll have a concrete room with a door that closes.

操作员在外部，但我们只有操作员。

And you have operators on the outside, but we only have operators.

他们说，哦，这彻底改变了情况。

And they said, oh, this radically changes this.

我们可以推进并快速开展这项工作。

We can go forward and go we can we can do this quickly.

这对我们来说非常有帮助，使我们能够构建可监控和操作的系统。

And that's been really enabling for us is to be able to build systems we can monitor and operate those.

但当你运行这些机器、进行聚变时，你并不在聚变机器所在的房间里。

But while you're running these machines, when you're doing fusion, you're not in the room with the fusion machine.

就像你不会在那些大型粒子加速器旁边，除了病人之外，你也不会在房间里一样。

Just like you're not you you don't other than the patient, you're not in the room with those big particle accelerators.

说到摩尔定律，你将来或许可以使用机器人，对吧？如果需要的话。

Speaking of Moore's Law, you could use robots at some point, right, for if you needed

没错。

Yep.

你知道的，机械装置，或者在波拉里斯房间里的生物？

You know, mechanical things, beings in the room with the with the with Polaris?

没错。

Yep.

全世界都在开展关于远程访问和远程操作的工作，就是使用机器人来完成这些任务。

There's work being done all all over the world on we call remote access and remote handling for using robots that do that.

我有个问题来自保罗，因为我之前说我在采访你。

I have a question from Paul, because I said that I was interviewing you.

我说：‘大卫，你会问什么问题呢？’

And I said, what question would you ask, David?

他说：‘我来读一下。’

And he said I'm gonna read it.

presumably，你已经做过实现净功率的模拟了。

Presumably, you've already done simulations where you achieve net power.

你能制造出和软件一样完美的东西吗？

Can you manufacture things that are as perfect as the software?

最难做到像软件一样完美的东西是什么？

What's the hardest thing to make as perfect as the software?

哦。

Oh.

这说得通吗？

Does that make sense?

嗯。

Yeah.

这是个好问题。

That's a good question.

这是个很棒的问题。

That's a great question.

我认为有两种东西在制造精度上和软件一样难以达到完美。

There's sort of two classes of things I think about that are as hard to to make as perfect as the software.

一是这些物理机器本身尺寸很大。

One is that the physical machines themselves are are large.

它们有40英尺长。

They're 40 feet long.

但某些最小部件的精度我们以千分之一英寸来衡量，也就是千分之一英寸。

And but the accuracy of some of the smallest parts, we measure in thousandths of an inch, one thousandth of an inch.

因此，你需要能够制造出这样的组件：当它们组装在一起时，整体长度达到40英尺，同时精度达到千分之一英寸。

And so you have to be able to manufacture components that are both can have the scale when they're all put together of 40 feet long and be as accurate as one thousandth of an inch.

这比百万分之一还要精确。

This is well better than one part in a million.

这很难做到。

And that that's hard to do.

这真的非常难，我们为此付出了大量努力。

That's really really hard to do and it's taken us a lot of work to do that.

在Helion早期，这真的很难。

And early in in Helion, it was really hard to do.

这是个非常好的问题，因为我们现在使用了一种技术，对我来说有时就像魔法一样，那就是激光追踪器。

We This is a really great question because we have a technology we use now, which seems like magic to me sometimes, in that, laser trackers.

这和你们现在在电动汽车上可能看到的扫描激光雷达类似。

Is that is that and this is some of the things like scanning lidar you might have for electric cars now.

在建造Polaris时，我们可以在整个建筑和机器上安装许多小型反射镜。

And you can build in in Polaris, in the building of Polaris, have little retro reflectors, little mirrors all over the building and all over the machine.

我们实时进行监测。

And real time we monitor.

我们使用激光追踪器，能够在整栋建筑内实时监测精度，精确到千分之一英寸。

We have laser trackers that can monitor the precision location within a thousandth of an inch over the whole building.

而且在实时监控中，当天气变冷或变热时，建筑会发生移动和变形，我们现在能够检测到这些变化。

And real time as we watch when it gets cold out or hot outside, the building moves and shifts, and we can detect that now.

这已经变得

And that's been

天啊。

Oh my god.

如此具有赋能性。

So so enabling.

当我们建造第三代机器Trenta时，我们花了将近一个月的时间来校准所有部件。

When we built Trenta, our sixth generation machine, it took us almost a month to align all the parts.

我们将所有部件安装在大型导轨上。

We put all the parts on large rails.

我们先校准导轨，然后需要在三个维度上进行微调。

We align the rails, then we had to tweak them in three dimensions.

我们每天都有人拿着卷尺，甚至激光测距仪，进行调整，使用小锤子、垫片等各种工具。

And we had people out every day with tape measures and even laser tape measures, adjusting it, little hammers, shims, all kinds of things.

而现在我们只需在所有部件上安装回射器，就能在几小时内完成校准，虽然可能比这还快。

And now we just put retro reflectors on everything and we're able to align it in in, let's say, hours, though it might have been faster than that.

哇。

Wow.

哇。

Wow.

相比之下，之前需要几周甚至一个月时间。

Versus versus weeks and a month.

而且，在我们的进度预算中，我确实预算了一个月的时间来对齐这个。

And it it was, like, in our schedule budget, I had budgeted a month to align this.

然后他们说，大卫，已经完成了。

And they're like, David, it's done.

我就问，你说完成了是什么意思？

And I'm like, what do you mean it's done?

就像，这不可能完成。

Like, It can't be done.

我要下来检查一下测量数据，因为你刚告诉我，时间从一个月缩短到了一小时。

I'm gonna come down here and I'm gonna, like, check the measurements because you just told me you accelerated from a month to an hour.

这就是技术发展或技术演进的力量。

And that's that's the power of some technology development or technology evolution.

另一个类似的时间尺度问题是，高温气体移动得非常快，因此你必须将其控制在十亿分之一秒的极高精度内。

The other is the same thing as a time scale is that we have the the the a hot gas moves quickly and so you have to control it to this very high speed one billionth of a second.

模拟十亿分之一秒非常容易。

And it's very easy to simulate one billionth of a second.

这台吉赫兹级别的计算机可以轻松完成。

This gigahertz computer can go do it, no problem.

但在现实中实现却非常困难。

But doing it in real life is hard.

就像我提到的光纤电缆，虽然很好，但根据长度不同，光在光纤中的传播时间是固定的。

It's things like the fiber optic cable I mentioned is great but depending on the length, the speed of light in that fiber optic cable has a given time.

而这个时间比我们关心的时间还要长。

And so and that's longer than the time we care about.

因此，你必须测量并了解光纤电缆的长度，因为光速会限制你达到特定的时间尺度。

And so you have to now measure and understand the length of that fiber optic cable because the speed of light will limit you to a given time scale.

因此，在现实世界、真实空间中，光速的物理限制、导线中电子的移动速度、实际半导体的开启时间与计算机模型之间的差异，

And so doing it in real life, in real space, the physical limits of the speed of light, the the speed of electrons moving in wires, the turn on time of a real semiconductor versus a computer model.

这些都是我们花费大量时间去理解并确保准确的方面。

Those are a lot of the things that that that we spend a lot of time on is understanding those details and and getting them right.

许多出错的地方，恰恰是因为没有把这些细节处理好。

Some of the things that have gone the most wrong have been not getting those right.

在Trenta项目上，校准机器花了整整一个月。

On Trenta, it took a month to align the machine.

我当时并没有计划要花一个月的时间。

I didn't plan for a month then.

在Polaris项目上，我预估了

On Polaris, I planned for a

一个月，因为我想

month because I

吸取了教训。

had learned my lesson.

不是在Trenta上。

Not on Trenta.

所以我们只是在那上面浪费了时间，因为我们没有做好。

And so that we just burned schedule on that because we didn't get it right.

而在Polaris上，我们进展得更快，这很棒。

And on Polaris, we it it went faster, which is great.

但同样地，在一些时序细节以及如何让所有东西尽可能精确和及时方面。

But similarly, on on some of the details of timing and how do you get everything to to be as accurate and time as possible.

我们在那方面投入了大量时间，并且这实际上是一些最艰难的挑战，因为只有当你构建了整个机器，拥有了这些大规模系统后，你才能吸取这个教训。

That has we've spent a lot of time on that and and it's been some of the some of the hardest challenges in in practice because you don't learn that lesson until you build build the whole machine, until you have these large scale systems.

在实验台上的小规模测试运行完美。

The small scale on the bench works perfect.

没问题。

No problem.

到了大规模层面，光纤变得更长了，你就错过了。

Go on the big scale when now the fiber optics are much longer and you missed it.

所以你必须在计算机系统中加入校正机制来弥补这一点。

And so you gotta go build that in the computer system to correct for that.

我们可以做到，但这确实带来了一些非常有趣的工程挑战。

And we can do that, but that's that's been that's been some really interesting engineering challenges.

你手头的事情真多啊，大卫。

You have a lot on your plate, David.

我知道。

I know.

你是怎么完成这一切的？

Like, how do you do all this?

是啊。

Yeah.

这真是

This is

很多。

a lot.

我们忙不过来。

We're overwhelmed.

但我

But I

我确实好奇，我认为这 probably 不是你的99个问题之一：竞争。

do wonder, one thing that I'm imagining probably isn't one of your 99 problems is competition.

并不是说没有其他人或公司在做这件事，但我就是觉得，这并不是让你夜不能寐的事情。

And not because there aren't other people, companies working on this, but I just get the feeling that that's not something that you lose sleep about at night.

是这样吗？

Is that right?

一般来说，不用担心聚变领域的竞争，甚至也不用担心天然气和其他电力系统的竞争。

Generally, don't worry about fusion competition or even natural gas and and other other power system competition.

我认为全球对能源的需求如此巨大，尤其是在人工智能数据中心和电动汽车的电力需求方面。

I think the need is so great in the world, especially with AI data centers and the power need there in electric vehicles.

随着电动汽车越来越便宜、数量越来越多，对清洁基荷电力的需求变得如此巨大，真正的竞争其实是与我们自己竞争。

And as we've gotten cheaper and cheaper electric vehicles and more and more of them, the need for clean base load electricity is so big that the competition is with ourselves.

关键在于我们如何尽可能快地建造。

It's just how do we build as absolutely as fast as possible.

我们确实会从技术共享的角度来思考，比如公开设计图，以及我们如何谈论这个问题。

We do think about it from the point of view of of giving away the technology, sort of publishing schematics and and how we think about how we talk about it.

但归根结底，核心问题仍然是我们如何尽可能快地建造。

But really, at at the end of the day, it comes back to how how do we just build as absolutely fast as possible.

嗯，如果聚变能改善其形象，它可能成为一种有竞争力的技术。

Well, guess vision, if it cleans up its image, could be a competitive technology.

我不知道这是否正确，但是……

I don't know I don't know if that's true or not, but

我认为最终还是取决于时间表。

I think it it it will always come down to timelines.

我们的目标是实现聚变装置的流水线生产，建立一个聚变发生器的超级工厂，让它们源源不断地从生产线上下线。

Our goal is to get to where we're building fusion on assembly lines, a gigafactory of fusion generators, where they're coming off the assembly line.

它们也被迅速部署了。

They're being then deployed very quickly as well.

有一件事让我感到惊讶，它比预期的还要好，因为我们一直被所有专家告知这是不可能做到的，这属于创始人模式的一部分。

One of the things that I was surprised by that went better than expected and and, because we were told by every expert that you couldn't do this is part of founder mode.

专家们告诉我们，这是做不到的。

We're told by experts, you can't do this.

对于我们的发电厂Orion，我们获得了环境许可的非重大性认定，而这是从未有人做过的事情。

It's not a thing that's ever been done, which is for Orion, our our power plant, we were granted a determination of non significance for the environmental permit.

这简直是一个法律术语。

And this this is like in the Such a legal word.

在环境许可运作的细节中，但这种分类有时会给予某些太阳能电站。

In in the weeds of how, environmental permits work, but it's a a classification given to some solar farms.

大多数太阳能电站都得不到这种分类，因为它们的环境影响可能非常大。

It's not even given to most of them because because the environmental impact can be so big.

他们却说不行。

And and they said no.

核能永远不可能获得这个评级。

No nuclear power can ever have it.

聚变能源太新了。

Fusion power is too new.

这永远不会发生。

It'll never happen.

我们花了多年时间与这些专家合作，探讨聚变技术真正的担忧。

We went through multiple years of working with the experts of these are the real concerns of fusion.

这些是聚变技术的危险。

These are the hazards of fusion.

他们说：不行。

And they said, no.

这就是你的分类。

Like, this is your classification.

这意味着我们可以从今天开始建设。

And what that means is we could start building today.

我们可以立即破土动工。

We and we could break ground.

我们获得了这样的决心。

And we we we got that determination.

我们获得了租赁批准，就在租赁签署后的几个小时内就开始了土方工程。

We got a lease approved and then had had we're moving earth within the hour of the lease being signed.

由于聚变的影响规模如此之小，所以能够如此迅速地推进。

Because of of the the scale of the impact of fusion is so low that you can move that fast.

我不知道基于铀的核能是否能够做到这一点。

And I don't know that that, uranium based nuclear power will ever be able to do that.

永远做不到。

Will ever do that.

是的。

Yeah.

所以目标是无关紧要，这想想还挺讽刺的。

So the goal is non significance, which is kind of funny when you think about it.

环境影响如此之低，以至于我们可以完成研究并立即部署。

The environmental impact is so low that we can do the study, deploy it.

而我们目前的目标仍然是逐个地点进行部署。

And then our goal is right now, we still deploy site by site.

你会说：我要去这个社区。

You say, I'm gonna go to this community.

我要在这里建一座工厂。

I'm gonna build a plant here.

你要花一年时间与社区、当地环保机构和公用事业公司合作。

We you spend a year working with the community and the local, environmental agencies, the utility.

你要花上好几年。

You go multiple years.

然后你才能部署一座工厂。

You now then go deploy a plant.

我们的目标是改变这里的标准，不再像建造机场那样去建设场地，而是像制造飞机一样。

Our goal is to change the the metric here and make it not like building airports where you go and work at building the site, but like building airplanes.

我们就在埃弗雷特这里，紧邻巨大的波音工厂，每天看着飞机从厂房里生产出来。

We're up in here in Everett and right by the big Boeing plants watching the planes come out of the building every day.

而且

And

而这就是你希望实现聚变的地方。

and and that's where you want fusion.

如果你要部署太瓦级、价值万亿美元的发电厂和发电机，你就必须以这种方式进行。

If you were gonna deploy terawatts and trillions of dollars worth of power plants and generators, you have to do it in that way.

现在，我们就在埃弗雷特这里、在我身后这条传送带上制造设备。

So now we build machines right now on conveyor belts behind me here in Everett.

这对聚变来说很不寻常。

And and that's that's that's weird for fusion.

大卫，你在仅有的一点空闲时间里，怎么放松和娱乐呢？

David, what do you do to relax and have fun in in your minimal amount of time that you have?

我有一个九岁的儿子，现在迷上了冰球，加入了冰球队。

I have a nine year old son who is into ice hockey now, joined an ice hockey team.

所以我已经成为那种热爱周六冰球比赛的父母了。

So I have become one of those parents at the at the I love my Saturday ice hockey games.

这真的很有趣。

And so that that's fun.

然后，我会尝试把我在工程、科学和迭代开发中学到的东西带回家。

And then and then a lot of, you know, trying to take some of the engineering, the science, the iterative development things that I have done and bring that home.

现在是万圣节季节。

It's Halloween season.

这是我最喜欢的季节。

It's my favorite season.

所以，我会思考如何制作机械动画，如何把工程和计算机编程的一些东西带回家。

And so, it's a lot of how do you build an, animatronics and how do you take some of that engineering, the computer programming, and and bring that home.

因此，沿着这些方向有很多有趣的项目。

And and then so lots of fun projects along those lines.

我猜也是。

Oh, I bet.

你儿子喜欢这些吗？

Does your son like those?

大部分喜欢。

Mostly.

大部分喜欢。

Mostly.

有时候会觉得有点尴尬，就是

Some sometimes it's cringe is the

我用过、也学会了一个词。

word that I've I've used I've learned.

是的。

Yeah.

对这个很熟悉。

Familiar with it.

是的。

Yeah.