Lyft CEO on Strong Growth & Autonomous Vehicles, How Dario Amodei Stirred Up Silicon Valley

欢迎大家收听Information TITV。

Welcome everyone to the Information TITV.

我是Stephanie Palazzolo，今天是8月7日，星期五。

My name is Stephanie Palazzolo and it's Friday, August 7.

今天节目里，Lyft公布了强劲的第二季度需求，但对今年剩余时间的业绩指引引发了一些疑问。

Today on the show, Lyft reported strong Q2 demand, but its guidance for the rest of the year is raising questions.

我们将采访该公司CEO，了解更多未来动向。

We'll talk to the company's CEO for more on what's ahead.

接着，我们会深入探讨华尔街和硅谷对Anthropic CEO Dario Amodei的看法，为公司的IPO做准备。

We'll then take a look behind the curtain at how Wall Street and Silicon Valley view Anthropic CEO, Dario Amodei, ahead of the company's IPO.

我们将与我的合著者Cory Weinberg一起解读我们最新的报道。

We'll unpack our latest profile with my coauthor, Cory Weinberg.

节目最后，我们会分析字节跳动如何规划其AI模型的未来，还有一期特别版的《编辑精选》，深入探讨软银的融资策略以及停滞不前的华纳兄弟。

And to close out the show, we'll examine how ByteDance is planning the future of its AI models, plus a special edition of the Editor's Cut, where we take a closer look at SoftBank's financing strategy and the stalled Warner Bros.

派拉蒙合并案。

Paramount merger.

今天的节目会很精彩，咱们直接开始吧。

It's going to be a great show, so let's get right on into it.

Lyft第二季度需求强劲。

Lyft reported strong demand in its second quarter.

总预订量增长了23%，这得益于其欧洲出租车业务以及通过合作伙伴关系获得的创纪录的乘车份额，提升了每趟行程的平均价值。

Gross bookings rose 23% as its European taxi business and a record share of rides through partnerships lifted the average value of each trip.

现在和我一起讨论本季度以及Lyft长期战略的是Lyft首席执行官David Risher。

Joining me now to discuss the quarter and Lyft's long term strategy is Lyft CEO, David Risher.

David，欢迎来到TITV。

David, welcome to TITV.

很高兴你能来。

It's great to have you on.

谢谢你，Stephanie。

Thank you, Stephanie.

David，Lyft又交出了一份非常强劲的季度成绩单，但你们的指引显示总预订量增速将从此前的水平放缓。

So David, Lyft delivered another really strong quarter, but your guidance calls for gross bookings growth to moderate from the pace we've seen recently.

跟我说说，是什么推动了这种预期？

Tell me kind of what's driving that outlook.

好的。

Yeah.

我觉得总体来看，正如你所说，这是一个创纪录的季度。

So I think the big picture is, so it was a record quarter, as you say.

我们的预订量创了纪录，出行次数也创了纪录，活跃乘客超过了3050万。

We had record bookings, we had record rides, we had over 30,500,000 active riders.

回顾几年前，那真是天壤之别。

And when I look back a couple of years, it's really sort of a night and day difference.

我们现在增长得非常快，无论是收入还是利润。

We are growing kind of like a weed, both at the top and the bottom line.

我们的收入，也就是我们所说的预订量，就是乘客支付给我们的金额，从来没有这么高过。

Our top line, which we call bookings, that's the amount that riders pay for us, has never, never been higher.

我认为这直接得益于我们以客户为中心的增长策略。

And I think that's a direct result of our customer obsession driving our growth strategy.

而且我们的利润也创了历史新高。

And our profits have never been higher as well.

说到前景，其实非常强劲。

In terms of the outlook, it's actually really strong.

我的意思是，我们今年有望完成十亿次出行。

I mean, we're on track to do a billion rides this year.

以我们的规模，即使很小的差异也会被放大，但我对我们在美国和海外市场的增长感到非常满意。

So at our scale, even small differences look kind of magnified, but I'm actually really happy with our growth both in The US and overseas.

对。

Yeah.

你知道，十亿次出行确实不容小觑。

You know, 1,000,000,000 rides is nothing to scoff at for sure.

我觉得你在昨天的财报中提到的一个话题是自动驾驶汽车。

I mean, I think one topic that you touched on yesterday during earnings was autonomous vehicles.

看起来整个网约车行业已经普遍认同，自动驾驶汽车是未来非常重要的一部分。

It does seem that the ride hailing industry broadly has come to agree that AVs are a really big part of the future.

你们显然有合作，Uber也有合作伙伴。

You know, you guys obviously have partnerships, Uber has partners.

你觉得Lyft未来在这方面怎么突出自己？

How do you see Lyft kind of standing out in this area moving forward?

对。

Yeah.

我觉得现在确实能看到两种很不一样的策略，但也有共同点。

So I think you're actually seeing two very different strategies play out, but maybe some common ground in this.

自动驾驶对网约车来说会非常好。

AVs are going be great for rideshare.

因为这是在现有平台上推出的新产品，乘客都很喜欢。

And the reason is because it's a new product on an existing platform that riders really like.

比如我们在亚特兰大和一家叫Main Mobility的公司合作测试，就能看到这一点。

And we can see that when they're being tested in places like Atlanta, where we're testing with a company called Main Mobility.

我们马上要在纳什维尔和Waymo开展下一步合作，那边需求很大，你也明白为什么，这个产品很酷，给人一种很可靠的感觉，能在自己的安静空间里，诸如此类。

We're just about to unveil the next step in our work with Waymo in Nashville, and we can see a lot of demand out there, and you can understand why, it's a cool product, it feels very reliable, you're in your own kind of quiet space and so forth and so on.

所以这一切都非常好。

So all of that is really good.

不过，这对整个行业来说也会是一个巨大的变化，对吧？

Now, it's also going to be a big change for the industry, right?

我们得把新的硬件引入平台，还要考虑政策问题，甚至还有房地产方面的考量。我给你举个例子来说明一下。

We're going to have to bring new hardware onto the platform, there are policy considerations, there are even real estate considerations, so let me give you an example of what that looks like.

在纳什维尔，就像我说的，我们和Waymo合作得非常深入，我们为他们做两件事。

In Nashville, where, as I say, we're partnered with Waymo in a very, very deep way, we're doing two pieces for them.

一方面，我们提供所谓的车队运营服务。

On the one hand, we're providing what are called fleet operations.

这些就是后台的工作，那些你从来不会看到的部分，比如维护、清洁、充电，但对自动驾驶汽车来说，这些非常重要。这些都是电动车，是非常复杂的设备。

So this is all the back of the house stuff, the stuff you never see, the maintenance, the cleaning, the charging, but it's a big deal for AVs, These are all EVs, they're quite complex pieces of equipment.

我们有一个八万平方英尺的停车场，正在为他们在那里开放，这是我们在这个领域能提供的真正服务。

We have an 80,000 square foot depot that we're in the process of opening up for them right there, and that's a real service that we can add in this space.

然后在需求方面，今年晚些时候，乘客将有机会直接在Lyft平台上使用Waymo。这有点技术性，但我们称之为供应共享。实际上，这意味着并不是只有一批专门给Lyft的Waymo车辆，我们和Waymo会在我们两个应用之间动态共享这些供应。

And then on the demand side, later this year, riders will have the opportunity to be master the Waymo right on the Lyft platform, and this is kind of technical, but we call it supply sharing, and what it really means is it's not just a special sequestered group of Waymos that's only available for Lyft, we and Waymo are going to be sharing that supply dynamically across both of our apps.

这真的很有创新性，我为团队能达成这个安排感到非常自豪。我们认为，结合我们的一百五十万司机，这是进入这个全新领域的最佳方式。

So really innovative, really proud of the teams for working out this arrangement, and we think it combined with our one and a half million drivers is the best way to enter this whole new space.

你刚才提到了车队管理。

And you just talked a bit about fleet management.

跟我聊聊，为什么这个业务这么有吸引力？它是不是不仅仅是另一种物流服务？

Talk to me a little bit about why is that such an attractive business and is it more than just another logistics service?

是的。

Yeah.

因为它是一个非常专业的部分。

It is because it's a very specialized piece.

为了保持车辆运行，也许我们先退一步看，因为我们这里还有点时间。

So in order to keep cars again, maybe let's zoom out because we've got a little bit of space here to do this.

在今天的网约车世界里，实际上是司机负责保持车辆维护、正常运行和随时可用，你知道，尽可能频繁地。

In today's rideshare world, it's really the driver that's responsible for keeping that car maintained and up and running and available, you know, as often as as need be.

在自动驾驶汽车领域，至少在短期内，将由像我们Lyft这样的公司来负责这些。

In the autonomous vehicle space, at least in the near term, it'll be companies like us, like Lyft, that's responsible for doing that.

因为世界上没有那么多的专业人才来做这种分散式的工作。

Because there's just not enough expertise in the world to do this kind of in a distributed way.

不过我们在这方面很幸运。

Now, we're really lucky there.

我们做车队管理已经快十年了。

We've been doing fleet management for almost a decade now.

我们有一个子公司叫FlexDrive。

We have a subsidiary called FlexDrive.

我们多年前就成立了这个公司，帮助司机解决租车、保养等问题，如果他们不想用自己的车的话。目前这个平台上已经有大约一万五千辆车了，所以我们在这方面相当有经验。

We set it up years ago to help drivers with car kind of leasing and maintenance and so forth if they don't want to use their primary car, and it's something, we've got about 15,000 cars on that part of the platform already, so it's something we're quite accomplished at.

但在自动驾驶汽车领域，正如我所说，这变得更加重要，因为这是一个全新的领域。

But in the AV space, as I say, it turns out to be even more important because this is new ground.

没有人知道如何大规模地做这件事。

No one knows how to do this at scale.

但如果车不能上路跑，那就没人能赚钱了。

But if the cars aren't available to drive, then no one's making money.

所以这就是为什么这件事很重要，也是为什么我们认为自己在这方面有真正的竞争优势。

So that's the reason it matters, that's the reason why we think we've got real competitive advantage there.

你知道，我也不想提竞争对手，但你知道这迟早要来。

You know, hate to bring up a competitor, but you knew it was coming.

跟我说说，Uber的自动驾驶策略和你们有什么不同，你们打算怎么在这一点上做出差异化？

Talk to me about how Uber's maybe AV strategy differs from yours and how you're planning to differentiate there.

好吧，我可能还是多说说我们的策略，而不是他们的。

Yeah, maybe I'll focus more on our strategy than on theirs.

我们决定非常紧密地与我们认为是全球自动驾驶领域领袖的公司合作。

We've decided to work very, very closely and deeply with the people we believe are the world's leaders in AV.

他们确实是最成熟的。

They're certainly the most developed.

这些公司包括美国的Waymo，还有中国的百度，我们在伦敦和他们有合作。

This would be Waymo in The United States, Baidu, which is a Chinese company, and we're working with them in London.

我觉得如果我们看看这个策略，它其实和我们之前如何与其他公司做出差异化是一致的。

And I would say if I look at that strategy, it's very consistent with, again, kind of how we've maybe differentiated from other folks.

我们是深度合作伙伴。

We're very, very deep partners.

我们擅长合作，我这么说很自豪，但也是有依据的。

We're good at partnering, and I say that with pride, but I also say that with evidence.

我几年前刚开始的时候，我们大约20%的行程，我说的是行程，都跟合作伙伴有关。

When I started a couple years ago, about 20% of our rides, I'm talking about rides now, were tagged to a partner.

比如和Chase的Sapphire Reserve合作，或者DoorDash，这是最近的一个合作，还有美联航。

That might be a Chase with the Sapphire Reserve partnership, or DoorDash, is a more recent partnership, United Airlines.

这些都是我们在需求端的合作，增长非常快，非常快。

These are all partnerships that we have on the demand side that have grown like crazy, like crazy.

我们刚和DoorDash一起进入加拿大，那里的关联行程增长了大约40%。

We just entered Canada with DoorDash, linked rides are up about 40% there.

美联航，我刚看了一下，通过我们的美联航合作，人们已经积累了超过5亿英里的里程，你可以通过乘坐Lyft来获得里程。

United Airlines, I was just looking at this, people have accumulated over half a billion miles with our United Airlines partnership where you can get miles by taking Lyft.

所以，我们为擅长合作感到自豪，因为这些都是双方都能成功的合作，双方都能成功，这也是它们能长期成功的原因。

So anyway, we pride ourselves at being very good at partnerships because they're the types of partnerships where both companies succeed, both companies succeed, and that's why they tend to be successful over the long term.

再说，我看其他那些公司，感觉他们的风格不太一样，可能在合作方面也有不同的理念。

Again, I sort of look at the other guys and I see a different kind of vibe with them, and maybe different DNA around partnerships.

但我们坚信，进入这个市场最好的方式就是和少数真正的领导者建立深度合作，然后创造出一种全新的东西，让整体的效果大于各个部分的总和。

But we feel very, very strongly that the best way to sort of enter this market is in deep partnership with a small number of true leaders, and create something that is frankly new and where the hole is greater than some of the parts.

你刚才提到，Waymo今年早些时候通过Flex Drive在纳什维尔开始运营了。

And you did just mention, you know, Waymo obviously operating in Nashville that started earlier this year through Flex Drive.

到目前为止你们学到了什么？在它成为Lyft业务的重要贡献者之前，你们在关注哪些里程碑？

What are you learning so far and kind of what milestones are you looking out for before this becomes a meaningful contributor to Lyft's business?

当然。

Sure.

好，我来告诉你我们正在关注哪些指标。

Well, I'll tell you some of the things that we're holding ourselves accountable for.

首先是车辆的可用性。

Certainly availability of cars.

这一点怎么强调都不为过。

And again, you can't overstate this.

你想想看，比如说你去机场的时候。

Just imagine, for example, when you go to an airport.

一家航空公司能不能赚钱，关键就在于飞机能不能快速周转、尽快飞上天。

The difference between a profitable airline and a not profitable airline is, are the planes being turned around quickly and in the air?

要是飞机一直停在那儿，或者处于维修状态之类的，那谁也赚不到钱。

If they're just sitting around or kind of in maintenance mode or whatever, again, no one's making money on it.

汽车也是这个道理。

So very similar with cars.

我们必须让这些车一直保持运营状态。

We have to keep the asset utilized.

然后除了基本的可用性之外，你还得有强劲的需求，并且要用非常高效的方式把需求和合适的车、合适的地点匹配起来。

And then on top of that, aside from just basic availability, let's say, you've got to have strong demand and you have to match that demand in a very, very efficient way to the right car at the right place and so forth.

在这方面我们有多年经验，但这次是个新课题，因为我们把它叫做混合网络。

Now this is something we've had years and years of experience doing, but this is a new thing because this is what we're calling a hybrid network.

司机和自动驾驶汽车在同一平台上运行，全天候工作。

Drivers plus robot driven cars on the same platform working 20 fourseven.

所以我们确实在服务质量上对自己要求很高，确保乘客总能获得最好的体验。

So we're really holding ourselves to sort of a high bar in terms of service quality, making sure that riders are always getting the best possible experience.

在司机方面，同样重要，确保我们的司机也能充分被利用，对吧？

On the driver side, equally important, making sure our drivers are feeling fully utilized themselves, right?

我们不想让这以牺牲他们为代价。

We don't want to sort of have this come at the expense of them.

所以，你知道，我们能不能创造这个新的混合市场，让它对我们来说成功，当然，也对乘客、司机和Waymo都成功。

So it'll be that sort of, you know, can we create this new hybrid marketplace in a way that really succeeds for us, course, but also succeeds for riders, drivers, and for Waymo.

我是说，在这个过程中有没有出现什么意想不到的挑战？

I mean, were there any surprising challenges that have come up through this process?

比如，有没有遇到什么你完全没预料到的障碍？

Like anything, Maybe obstacles you ran into that you really weren't expecting?

这可是个全新的领域。

This is uncharted territory.

是的，我们还在很早期，现在谈大惊喜可能有点太早了。

Yeah, we're so early, it's probably a little too early to talk about big surprises.

我可以告诉你，稍微透露一点幕后情况，就连选择停车场的位置，这虽然有点技术性，但如果你把停车场放在离市区太远的地方，虽然地价便宜，但会有很多空驶里程，你得来回跑，还没接到任何乘客呢。

I can tell you, to go a little bit behind the scenes, even where you choose to put the depot, again this is technical stuff, but choose to put the depot, if you put it too far outside of town, the real estate is cheap, but then you have all these deadhead miles, You've got to go back and forth before you have any customers.

如果你把停车场放在铁路轨道的错误一侧，那你可能会发现自己的车十五分钟都动不了。

If you put it on the wrong side of railroad tracks, then you might find yourself in a position where you can't move your car for fifteen minutes.

如果你没有足够的停车位，那听起来好像很简单，对吧？

If you don't have enough parking for the cars, well that sounds easy, right?

但你可能需要为几百辆车准备停车位，因为有时候它们可能都在那里，要么在充电，要么在维护。

But you have to have parking for potentially hundreds of cars because they might at some point all be there, either napping or in maintenance mode.

所以你必须非常小心地展望未来两三年、四五年，决定你需要多大的空间。

And so you to be very careful about looking forward two, three, four years and deciding how much space you're going to need.

所以我们目前就处于这种物流规划阶段。

So we're very much at that sort of logistic.

还有，我再告诉你最后一件事。

Even, I'll tell you one last thing.

就是车辆的清洁。

The cleaning of the car.

人们会说，哦，洗车不是什么大事。

People say, oh, cleaning of a car is not big a deal.

大家都知道怎么洗。

People know how to do that.

但是这些车上有传感器。

Well, these cars have sensors on them.

这些传感器得用特别的方式清洗。

Those sensors have to be cleaned in a very particular way.

我是说，得用特定的方式擦拭，确保不留水痕之类的，因为那会影响车的运行。

I mean, wiped in a very particular way to make sure that they don't get streaks and so forth because that can degrade the operations of the car.

所以这听起来很技术性，而且现在还很早期，但正是这类事情我们非常专注，确保为我们的合作伙伴和乘客创造非常好的体验。

So this sounds very technical and it's very early, but it's that kind of thing that we're really laser focused on to make sure we create a great, great experience for our partner but also for our riders.

对。

Yeah.

听起来是个非常技术性的过程。

It sounds like a very technical process.

另外，这是不是有点奇怪，对吧，自动驾驶和Lyft多年来一直经营的业务之间，存在这种紧张关系。

I mean separately, is this kind of weird tension, right, between AVs and the business that Lyft has obviously run for many years.

你知道，你非常强调的一点是支持司机，但这也和未来可能形成一种矛盾，你知道，未来我们提供的出行服务中，人类司机会少很多。

You know, one of the things that you really emphasize is supporting drivers, but there is this tension between maybe that philosophy and also a future where, you know, we're going to be taking rides that involve many fewer, you know, human drivers.

那你怎么样把这两者协调起来呢？

So how do you kind of jive those two things together?

对，对，我觉得这是个很好的问题，也非常重要。

Yeah, yeah, I think it's a great question and really important.

我想稍微从哲学角度说一下，我确实认为企业认真对待这种转型是非常重要的。

And I'll maybe philosophize for just a second and say, I actually think it really is important for businesses to take these kinds of transitions quite seriously.

好，那我们具体在做什么呢？

Okay, so what are we doing?

我们做的第一件事，就是一如既往地支持我们的司机，对吧？

The first thing we're doing is we're supporting our drivers in the way that we always do, right?

我们有一个所谓的收入上限，或者说费用上限，抱歉，就是限制Lyft收取的费用。

We have what we call an earnings cap, sort of caps, or a fee cap, excuse me, that caps Lyft's fee.

我们做了一些非常基本的事情，希望能让司机们感到被公平对待。

It does some very basic things to make sure the drivers feel, we hope, treated fairly.

顺便说一句，我们比开其他平台的司机有大约30%的偏好优势。

And to brag for a second, we have about a 30 advantage, preference advantage, versus drivers who drive for the other guys.

当我们问那些同时为两家平台开车的司机，你们更喜欢哪家？

When we ask drivers who drive for both, who do you prefer?

超过50%的人说我们，然后大约20%多的人说他们。

Over 50% say us, and then some 20 some percent say them.

所以这是一个很好的起点。

So that's a good place to start.

好的，那么在这样一个转型期间，你能做些什么呢？

Okay, then what can you do during a transition like this?

有几件不同的事情。

A couple of different things.

首先，你可以确保司机们确实感到被支持，并且明白这不是二选一的问题。

First, you can make sure that drivers really do feel supported and understand this is not either or.

会有很多人不想坐自动驾驶汽车，也许他们需要帮忙搬行李，也许他们想在一天结束时有人聊聊天。

There will be plenty of people who don't want to take an autonomous vehicle, maybe they want help with their luggage, maybe they want a nice conversation at the end of the day.

也许他们担心不能及时赶到机场，想跟司机说，你能稍微开快点吗，因为跟电脑说这个没用。

Maybe they are worried they're not going to get to the airport fast enough and want to be able to say to the driver, can you kind of step on a little bit because you can't say that to a computer.

所以这是其中一点。

So that's one thing.

第二点是我们其实在招司机。

The second thing is we're actually hiring drivers.

我之前提到过这个站点，这个站点技术性很强。

So I've mentioned this depot, this depot is quite technical.

我们在这个站点雇的人，大部分都是以前的Lyft司机。

The majority of the people we have hired to work in this depot have been former Lyft drivers.

这不是偶然，而是我们非常刻意这么做的。

And that's not an accent, that's something we're doing very, very deliberately.

我们就是在给以前的Lyft司机开后门。

We're putting our thumb on the scale for former Lyft drivers.

然后第三点是，我们想扩大市场，对吧？

And then third is, we were trying to grow the market, right?

所以这不是那种零和游戏，不是说每辆机器人车就会抢走一个人类司机的工作。

So this is not a thing where we sort of say, it's zero sum, every robot takes a human drive away.

不，不，恰恰相反。

No, no, it's the opposite.

记住，我们每年提供十亿次出行，这很棒，但美国人用私家车出行次数是一千六百亿次。

Remember, we give a billion rides, which is wonderful, but people do in The US 160,000,000,000 rides in their private car.

所以未来对乘客、司机和自动驾驶汽车来说都有巨大的机会，我们非常专注于把这块蛋糕做大，这样就不会让人觉得是零和游戏。

So there's a huge amount of opportunity in the future for riders, excuse me, for drivers and for ABs, and we're very focused on expanding that pie so it doesn't feel zero sum.

所以很明显，我们之前聊过Waymo在纳什维尔的情况。

And so obviously we've talked a bit about, you know, Waymo in Nashville.

你也提到过通过Free Now把百度的Apollo Go无人驾驶出租车带到欧洲。

You also did mention bringing kind of Baidu's Apollo Go robotaxis to Europe through Free Now.

显然，在政策制定者越来越关注中国科技的时候，这是一个相当引人注目的决定。

Obviously that's, you know, a pretty notable decision at a time where policymakers are increasingly, looking closely into Chinese tech.

你怎么看待这些地缘政治风险？

How do you kind of think about those geopolitical risks?

是的，这个问题问得很有见地。

Yeah, again, perceptive question.

而且每个国家的态度不一样，对吧？

And it's different country by country, right?

不同国家对街头中国技术的接受程度不同。

Different countries have different tolerances for Chinese technology on the street.

拿伦敦来说，伦敦街头其实已经有不少中国技术了。

If you look at London, London has a fair amount of Chinese technology on the street already.

比如我们熟悉的经典伦敦出租车，其实是吉利制造的，那是家中国公司。

In fact, the London cab, which we all know and love and kind of is iconic, is actually made by Geely, a Chinese company.

所以从某种程度上说，伦敦和英国已经习惯了中国汽车制造商的车在马路上跑了。

So to a certain extent, London and England have already gotten comfortable with Chinese OEMs, car manufacturers having their cars on the road.

那我们可以扮演什么角色呢？

So what's the role we can play?

我们能做的角色就是确保我们在尽最大努力处理好所有的数据隐私和安全问题，这些问题相当严重。

The role we can play is making sure that we are doing our level best to take care of all of the data privacy and security issues, and these are quite serious.

比如，我们在把任何数据送出英国之前都会先匿名化处理，这背后是一整套技术流程。

Mean, example, we anonymize all the data before we send it to anything outside of The UK, and that's a whole sort of technical thing.

我们还做出了一系列政策承诺，以确保我们真正为这些车辆承担责任。

There are a bunch of policy commitments we've made to make sure that we're effectively taking responsibility for these cars.

所以，我觉得这是一个非常有趣的合资运营的例子，中国公司百度——就像中国的Alphabet——拥有一个非常非常好的技术产品，自动驾驶表现非常出色。

So, I think it's a really interesting example of a kind of a joint operation, where the Chinese company, Baidu, which is kind of the alphabet of China, has a very, very good technical product that does a very good job driving itself.

但我们承担了更多面向客户和政府方面责任，而且希望整体能大于各部分之和。

But we take a lot of the responsibility for the more customer and government facing side of things, and hopefully, again, the whole will be greater than the sum of the parts.

现在还是早期阶段，我们刚开始用那个叫ByteDu RT6的产品绘制道路地图，这是个很棒的产品，乘客们都很喜欢，我自己也坐过，确实很棒。

Very early days, where we just started mapping out roads with that, it's called the ByteDu RT6, it's a great product, people love, riders love the product, it's actually, I've ridden it myself, it's a great product.

希望政策制定者会继续认为它对伦敦的交通系统是有益的，截至目前的情况确实如此。

Hopefully policymakers will continue to think it's additive to the London transportation system, and that's certainly what we've seen so far.

我想最后一个问题，再多跟我聊聊国际扩张的事吧。

And I guess, last question, just talk to me a little bit more about international expansion.

你们在七月份完成了对Free Now的收购。

You guys closed the Free Now acquisition in July.

这是Lyft在北美以外的第一个重大举措。

That's Lyft's first major step outside of North America.

你觉得为什么现在是进行国际扩张的好时机？

You know, why is this the right time for international expansion?

从现在起五年后，你如何看待这项业务在你整体业务中的位置？

Where do you see that business sitting in your overall business breakdown, you know, five, five years from now?

当然，是的。

Sure, yeah.

我们差不多在一年前完成了收购。

So we closed it almost exactly a year ago.

那是去年七月，过去一年我们确实变得更加全球化。

It was last July, and we've really spent the last year becoming a much more global company.

Free Now的预订量大约有十亿美元，所以相对Lyft的200亿美元来说规模较小，但在希腊、爱尔兰、西班牙以及欧洲大约九个不同的国家，它仍然有非常重要的存在感。

Free Now does about a billion dollars in bookings, so it's relatively small compared to, let's say Lyft's $20,000,000,000 but still a very significant presence in countries like Greece, in countries like Ireland, in Spain, and across about nine different countries in Europe.

这也是让我们能提前进入伦敦街头，帮助在那里商业化自动驾驶汽车的原因。

It's what's given us access to the earlier point, to the London streets to help kind of commercialize AVs there.

但日复一日，我们正在做的是把我们美国研发的技术和专业经验，以非常大的规模，真正带到欧洲，努力提升乘客和司机的体验。

But day by day, what we're doing is we're bringing the technology and the expertise we've developed in The United States at a at a very, very big scale, and really bringing that to Europe to try to kind of level up the rider experience and, the driver experience as well.

如果我想两三年、五年后，你看，这两个市场规模差不多，但Lyft传统上是一家美国本土公司，只在美国发展，最近才开始进入加拿大市场，而且在那里做得不错。

If I think, you know, two, three, five years out, you know, look, the two markets are roughly the same size, but Lyft kind of traditionally grew up as a domestic company, a US company only, we've more recently gotten started in Canada and are doing quite well there.

所以我当然希望欧洲能发展得很快，甚至可能比美国还快，因为那里的机会太多了。

So I would certainly hope that Europe will actually grow frankly, very, very fast, even maybe compared to The US, because there's so much opportunity there.

潜在市场同样巨大，但我们的渗透率远低于美国。

The TAM is equally big, but we're so much more of under penetrated.

所以敬请期待，我们看吧，现在还在早期阶段，刚刚一年，但最后我要说的是，几周前我有幸去了西班牙的巴塞罗那，测试我们的新整合，让你可以打开Lyft应用的测试版，叫一辆Free Now的出租车，我试了四次，每次都完美运行。

So stay tuned, we'll see, it's still early days there, just a year end, but I'll finish up by saying I was actually lucky enough to be in Barcelona a couple of weeks ago in Spain, testing out our new integration, which allows you to open up the Lyft app in beta mode, and call a free now taxi cab, And I did it four times and it worked perfectly every time.

所以从2027年开始，你将能够在美国

So starting in 2027, you'll be able to kind of roam across The U.

漫游。

S.

在欧洲，只要打开Lyft应用，就能随时随地叫到车。

And Europe with your Lyft app and be able to get a ride anywhere you want.

希望这会是一次很棒的体验。

Hopefully it'll be a great experience.

真让人兴奋。

Exciting stuff.

好，期待下次财报后还能请到你。

Well, looking forward to hopefully having you on after the next earnings.

再次感谢你参加我们的节目，David。

Again, thank you so much for joining us, David.

刚才是Lyft的首席执行官David Risher，在TITV接受采访。

And that was David Risher, CEO of Lyft here on TITV.

他是宗教领袖还是CEO？这是我同事Cory Weinberg和我在最新一篇关于Anthropic CEO Dario Amodei的人物特写中试图回答的问题。

Religious leader or CEO, that was the question my colleague Cory Weinberg and I were trying to answer in our latest profile on Anthropic CEO Dario Amodei.

今天来和我们讨论这篇报道以及我们昨天发布的另一篇关于Canva增长放缓的文章的，是我们的资深记者Cory Weinberg。

Here to discuss that story as well as another piece we published yesterday on Canva slowing growth is our senior reporter covering Anthropic, Cory Weinberg.

科里，欢迎来到节目。

Cory, welcome to the show.

嘿，斯蒂芬。

Hey, Steph.

那么，退一步来说，我们当初为什么要做这个人物专访呢？是什么让达里奥这个人这么有意思，值得我们深入挖掘？

So, you know, taking a step back for a second, let's talk a little bit about why did we even wanna do this profile in the first place and, you know, what made Dario such an interesting person for for us to really dig into?

对。

Yeah.

我觉得达里奥之前已经被报道过好几次了，但从来没有像现在这样，他基本上站在商业世界和科技世界的顶端。

I mean, I think Dario has been profiled, you know, several times before, but never during a moment like this where he is essentially on top of the business world, on top of the tech world.

Anthropic 今年早些时候在销售额上超过了 OpenAI。

They've Anthropic passed OpenAI in sales earlier this year.

它完全主导了 AI 编程工具这个最赚钱的市场。

It totally dominates the most lucrative market in AI coding tools.

而达里奥·阿莫德伊的崛起之路非常不寻常，从顶尖到顶尖，我觉得硅谷其实并没有完全理解他这个人。

And Dario Amodei has taken a really unusual ascent from the top to the top, and he is really a figure that I think Silicon Valley doesn't fully understand.

我们见过很多CEO被媒体用宗教般的词汇来形容，说他让追随者对他的领导力充满敬意，而且总体上他也能在员工中维持很高的忠诚度。但从来没有人像他这样，通过做出艰难决策和权衡取舍来赢得大量忠诚，比如在商业决策和坚守原则之间如何取舍，以及在这些事情上接受公众审视。

We've seen a lot of CEOs described as, you know, sort of in sort of religious terms that he inspires devotees to his his leadership and and and just generally is someone who retains a lot of loyalty among employees, but never someone like this who's retained a ton of loyalty largely by making hard choices and difficult trade offs about what the business decisions to make versus how to stand by its principles and and sort of be under the microscope on things like that.

我很好奇你的一些看法。

And I'm curious for you.

很明显，我觉得这篇报道里有很多让人意外的细节。

I mean, obviously, I think there were a lot of really like surprising pieces to come out of the profile.

但在整个采访和报道过程中，你了解到的最让人意外的事情是什么？

But what was like the most surprising thing that you learned coming out of this reporting process?

对。

Yeah.

有些关于达里奥核心圈子里到底有哪些人的细节，真的很有意思。

There were some really fun details to learn around who actually is in Dario's immediate circle.

那其实是我在报道中试图搞清楚的一个问题，目的是帮助我了解他到底是个什么样的人。

That was actually one question that I was sort of tackling in the reporting to kind of help guide me around who is this guy?

你知道，我们读过他的文章。

You know, he he we've read his essays.

我们最近也在电视上看到他，比如Anthropic和五角大楼起冲突，或者美国政府要对Anthropic实施出口管制的时候。

We've seen him more recently on on TV around, you know, sort of controversies when Anthropic is fighting with the Pentagon or the the US government is putting export controls onto Anthropic.

但这个人到底在听谁的？

But who is this guy actually listening to?

他的圈子里都有谁？

Who is in his circle?

我一开始跟Anthropic的很多投资人聊的时候，有个印象特别深，就是他们觉得他的圈子其实不大，这让我挺意外的，因为我以前采访CEO的时候，投资人通常都会很快说，哦，他听谁的。

And sort of a distinct impression I was getting initially when I was talking to a lot of investors in Anthropic is they could they didn't really think he had that wide of a circle, which was surprising to me because when I profiled CEOs before, investors are usually sort of quick to say, oh, he's listened.

比如，他跟某某商界大佬关系很近。

He's, like, gotten really close to x y z business luminary in the past.

他把这个人当导师之类的。

He has, you know, taken on this person as a mentor.

但Dario这边我完全没听到这些，这让我特别意外，尤其是他还是个第一次当CEO的人。

I got none of that with Dario, which was really surprising to me, especially for a first time CEO like Dario is.

大家说他会跟联合创始人聊很多，其中一个是他妹妹，他们关系很亲密。

People said he listens to his co he talks to his cofounders a lot, one of whom is his sister, who he's very close to.

还有，他也会和他妻子交流。

And he also talks to his wife.

他把妻子视为重要的顾问。

He considers his wife a close counselor.

她叫卡米拉·克拉克。

Her name is Camilla Clark.

在行业活动上，她经常陪伴在他身边。

She is very often at his side at industry events.

在太阳谷活动上她也去了。

She was there at Sun Valley.

两人还一起去了趟迈阿密，会见贾里德·库什纳和伊万卡·特朗普。

The two of them took a trip down to Miami together to meet with Jared Kushner and Ivanka Trump.

卡米拉在公众视野中几乎没有存在感。

And Camilla has really no public profile whatsoever.

她非常低调，但对达里奥来说，她是个非常重要的人，尤其是在他处境混乱的那段时间。

She is she is very much low profile, but she was sort of a very important person to Dario, especially during such a chaotic time for him.

所以，这是我发现的人物报道中一个挺重要的新方面，我觉得有必要让读者知道。

And so that was that was kind of one new aspect of the profile that I found sort of important to bring to readers.

嗯。

Mhmm.

对。

Yeah.

我知道你当时真是费了好大劲去挖她的信息。

I know you were like really deep in the trenches trying to find anything about about her.

而且特别有意思的是，看到文章下面的评论，或者人们在社交媒体上讨论这件事。

And it was super funny just like seeing the comments on the story or people talking about it on social media.

我看到好几个人说，达里奥结婚了？

I saw several peep several people that were like, Dario's married?

我完全不知道。

Like, I had no idea.

所以她确实很低调。

So she's very much under the radar.

完全同意，完全同意。

Totally, totally.

我觉得从我的角度来看，有一件事让我很惊讶，我一直在琢磨的是，最近几个月好像有个争论，就是达里奥说的那些话有多少是营销炒作，比如他说那些关于神话的疯狂言论，是不是只是为了给模型造势，其实是一种超级聪明的营销策略。

I mean, I think from my end too, I think one thing that surprised me that I was trying to figure out is I feel like in the last couple of months, there's been this debate of like how much is what Dario says kind of like marketing hype, oh, did he say all those crazy things about mythos because he just wants to hype up the model and it's actually this like super insanely smart marketing strategy thing.

但有意思的是，经过这次采访，我真的觉得他就是很真诚。

But it was interesting because I really feel like coming out of this, I just really believe that he's just genuine.

他就是非常真诚，而且这些确实是他真心相信的东西。

He's just very genuine and these are things that he truly believes.

我和他的一些同事聊过，甚至在他最早在OpenAI的时候，那时候他完全不知道自己会站在聚光灯下，他们就说，你知道吗，他在家里用一台隔离的电脑工作。

I mean, talking to some of his colleagues, even from like his earliest days at OpenAI where he very much did not know he was going to be in the limelight, They're saying things like, you know, he works on this air gapped computer from his home.

他会打印一些备忘录，那些内容其实并不是什么大家会觉得是机密的信息。

He's like printing out these memos of things that are like, you know, not really people- not really something people would consider to be kind of top secret information.

他告诉同事他不敢去中国，因为他怕被绑架。

He is telling his colleagues he's afraid to go to China because he's afraid of getting kidnapped.

他就是特别多疑，即使是在最早的时候，那时候很多人说，你看GPT1、GPT2，虽然挺酷的，但连个完整的句子都写不好。

Like he was just hyper paranoid even in the earliest days whenever a lot of people were saying like, you know, GBT1, GBT2, they were like, all right, this is cool, but like it can barely string together a sentence.

而且那时候Dario就说，不，这太吓人了，我们得把它锁起来。

And even then Dario was like, no, this is like so scary and we need to like lock this down.

所以我觉得，这真的帮我理清了一些事情。

So that just, I think, really helped to clear up some of that stuff for me, I think.

对，我同意。

Yeah, I agree.

我是说，作为记者，我们得有点阴谋论思维，才能追查故事线索。

I mean, think like as reporters, we have to be somewhat conspiracy brained, just pursuing leads for stories.

比如，还有什么比‘这家伙其实不信自己说的话，只是为了炒作’更大的阴谋呢？

Like, you know, what a bigger conspiracy than like, this guy doesn't actually believe what he says, and he's just saying it to hype it up and and and all of that.

但我觉得你得面对报道所揭示的事实，那就是这家伙在很多方面一直很一致，他真的深信这项技术可能非常危险，而且可以说，他对中国这样的国家行为体试图获取这项技术有点偏执。

But I think you have to be real about what the reporting shows, which is this guy has been consistent in a lot of these traits around really believing deeply that this technology can be very dangerous being, you know, sort of almost somewhat paranoid, you could argue, about state actors like China trying to get their hands on the technology.

你在他在OpenAI的最初时期就能看到这一点。

You saw that from his earliest days at OpenAI.

正如我们报道的那样，他的同事们也都指出这是他一个关键特质。

That was what as you were you know, as we reported, like, all of his colleagues were sort of pointing to as, like, a key trait of his.

所以那确实非常令人震惊。

So that was definitely incredibly striking.

完全同意。

Totally.

而且我觉得，文章中我们还想触及的另一个话题是，随着Anthropic规模越来越大，它不得不面对如何在核心原则与赚钱、打败竞争对手之间取得平衡的问题。

And I think, you know, another topic that we tried to touch on in the piece is this idea that as Anthropic has gotten bigger, it's had to deal with balancing these kind of core principles with, you know, making money and beating out competitors.

而且在某些情况下，它确实不得不做出妥协。

And, you know, in some cases it's had to really compromise.

比如，它之前说过不会接受中东投资者的钱，后来却接受了。

For instance, it took money from investors in The Middle East after previously saying that it wouldn't.

然后很明显，就像你写的，今年早些时候它可能更新或回退了一些负责任的扩展政策的内容。

And then obviously as you wrote, it may be updated or rolled back some aspects of its responsible scaling policies earlier this year.

你觉得这些妥协和Dario看似极其讲原则的性格一致吗？

I mean, do you kind of jive those compromises with Dario seemingly, you know, like ultra principled nature?

对，我觉得他们面临一个非常艰难的平衡，他们声称他们的商业目标，想要在AI竞赛中领先的愿望，本质上是为了让他们能够冲到前沿，确保他们已经制定了所有适当的安全协议和策略，以避免灾难，无论是网络安全、生物武器，还是任何强大AI可能带来的坏事。

Yeah, I think they have this very you know, sort of difficult balancing act where they've stated that they their commercial goals, their desire to be on top of the AI race is essentially a way for them to race to the frontier to be able to ensure that they have developed all the proper safety protocols and, you know, sort of tactics to avoid catastrophe, whether it be cybersecurity, bioweapons, anything bad that can happen with really powerful AI.

但我听到的最触动我的批评，是跟这种方法还有 Dario 有关的，特别是当他们不得不不断做出权衡的时候，Anthropic 的成功可能反而加剧了 AI 领域的竞争态势，这让事情变得危险得多。

But I think the critique that I hear that most resonates strongly with me relates to this approach and with Dario, especially as they have to make these constant trade offs, anthropic success has perhaps only intensified the race dynamics within AI that can make things a lot more dangerous.

他们五年前创立 Anthropic 时，就是明确冲着不信任 Sam Altman 和 OpenAI 那帮人，不放心让他们来负责任地掌控这项强大的技术。

They they explicitly went out to start Anthropic five years ago because they didn't trust Sam Altman and a lot of the OpenAI crew to be the ones to be responsible stewards sort of of this powerful technology.

我觉得，从某种程度上说，Anthropic 的成功，加上 OpenAI 的成功，反而加剧了这场竞争。

And I think in some ways, Anthropic success on top of OpenAI's success has only intensified the race, I would say.

你看，现在中国也在非常努力地发展这项技术。

Like, you know, you now have, like, very and a very intense effort within China to develop this technology as well.

当然，所有大型 AI 和科技公司，从 Meta 到 Alphabet，都是如此。

Obviously, all the big AI and tech companies, you know, from Meta to Alphabet too.

所以，我也不知道。

And so I don't know.

对我来说，我还在纠结的问题是，Anthropic 的创立本身，是不是就是 AI 竞赛的"原罪"。

I think, like, to me, like, the the question that I'm still trying to wrestle with is, was Anthropic's founding itself sort of an original sin of the AI race?

如果这些是他们想要避免的态势，因为把资本主义和这种非常危险的技术混在一起，可能会变得非常危险。

If these are sort of dynamics that they're trying to avoid because it can get really dangerous to mix, like, capitalism and this this this really dangerous technology.

现在Anthropic正处于资本主义热潮的顶端。

Well, Anthropic is sort of at the the top of the capitalistic heat right now.

所以他们只会继续面临更多这样的权衡，比如在商业可行性和安全性之间如何取舍。

And so they're gonna just keep facing more of these trade offs around what do we do to sacrifice commercial viability versus safety.

原罪。

Original sin.

这里有很多宗教隐喻，我很喜欢。但我也觉得，这种“起飞”的概念，就是那种我们迟早会开发出下一代更好的AI，这确实让竞争更加激烈了，对吧？

Love the there's so many religious references in this, but, I do think also this idea of takeoff, which is like, you know, at some point we're going to develop AI back and develop the next better version of AI is like heightening this race so much more, right?

如果Anthropic相信自己能做好这件事，他们就必须在其他人之前达到这种自我改进的AI水平，因为谁先达到，谁就会走上一条很难停下来的轨道。

Like if Anthropic believes that they are the ones that could do this correctly, they need to reach this sort of self improving AI before anyone else because whoever reaches that first, it's really just going to send them on this trajectory that's going to be hard to stop.

对。

Yeah.

而且我觉得，当我和硅谷的消息源聊天时，至少那些有朋友在Anthropic工作的人，你知道吗？

And I think that the, when I'm talking to sources around the valley that at least just, like, have friends at Anthropic, you know?

我觉得大家都注意到，这家公司——我们在报道里用过这个说法。

Like, I think everyone observes that this company we used the phrase in this story.

他们有一种，一种堡垒心态。

They have, a bunker a bunker like mentality.

这试图传达几个不同的意思。

And that sort of is trying to convey a couple different things.

你知道，一个是他们非常封闭，很少泄密，非常注重安全，但他们也极其坚守自己的文化，以及达里奥预测未来的能力。

You know, one is just they are very locked down, very few leaks, very heavy on security, but they also are incredibly adherent to their own culture and to Dario's ability to predict the future.

我和Anthropic的人聊过很多次，都听到这个说法。

I heard that again and again talking to Anthropic people.

达里奥能看到未来，而Anthropic开放的文化，可以说是公司最大的财富。

Dario sees the future, and Anthropic's culture of openness and is sort of the greatest treasure that the company has.

我觉得Anthropic的风险不在于它没有原则，或者它，你知道，有点，怎么说呢，说话不算数，它的风险可能是过于自信了。

And I think the risk to Anthropic isn't that it's sort of unprincipled or that it you know, sort of is, you know, sort of talking out of both out of its mouth, its risk is maybe being high in its own supply.

我觉得它的风险在于，它有一种非常忠诚的文化，相信像达里奥这样的标志性领袖，并且真心相信这项技术，相信克劳德，它的产物，能够自己创造下一个前沿，克劳德不仅是白领行业的未来，也是硬工程、机器人技术的未来，基本上整个就业格局都会被颠覆，五年到十年后，你我都认不出这个世界了。

I think its risk is sort of having this combination of a very sort of loyal culture that, you know, believes in a signature leader like Dario and really believes in this technology that Claude, its its creation will be able to create the next frontier itself, that Claude is the future of, you know, not only sort of a white collar business, but also the future of hard engineering, of robotics, that essentially the entire, you know, sort of employment landscape is going to be blown up, and neither you nor I will recognize the world five to ten years.

我觉得这是Anthropic很多人内心深处的信念，也是驱使他们如此努力工作、取得如此成功的原因。

I think that is like sort of a deep belief within a lot of people at Anthropic and is what drives them to, you know, continue to work as hard as they do and to be as successful as they are.

这不一定只是为了钱。

It's not necessarily just money.

我觉得达里奥就是这样的。

And I think Dario embodies that.

我觉得人们有时候就是不喜欢这样。

I think like people just don't always like that.

我觉得人们会因此感到威胁。

I think people, you know, feel threatened by that.

所以我觉得这仍然是一家非常引人入胜的公司。

So I think it continues to be a really fascinating company.

完全同意。

Totally.

我的意思是，我们在文章里还提到了另一个话题，就是Anthropic和其他AI实验室正在更多地转向开发应用程序，这讽刺地在某些情况下最终会和它们最大的客户竞争。

I mean another topic we touched on in the piece is this idea that Anthropic and other AI labs are moving more into building apps, which ironically ends up competing with their biggest customers in some cases.

这其实和我们昨天发表的那篇关于Canva的文章有关，Canva显然是OpenAI和Anthropic模型的大用户，而且最近确实遇到了一些瓶颈。

This really ties into another piece that you and I published yesterday about Canva, which is obviously a big user of models from OpenAI and Anthropic and these speed bumps that it's really run into recently.

其中一个原因是因为他们看到越来越多的客户转向了ChatGPT。

One reason for which is because they're seeing more of their customers go to ChatGPT.

那跟我说说，Canva那边到底发生了什么，导致它的增长放缓了？

Tell me kind of what, what happened with, with Canva that led to its slowing growth and, yeah, what, what can, what can you look Canva

Canva这周早些时候发布了最新的季度业绩，这让投资者非常震惊，而且Canva是一家私有公司。

really shocked investors, earlier this week when it delivered its latest quarterly results, and and Canva is privately held.

所以这是一份发给投资者的内部备忘录。

So this was a a sort of private memo sent to investors.

里面说，我们年初承诺的增长目标，现在预计会慢很多。

That said, we what we told you where we told you we were gonna grow earlier this year, yeah, we're expecting a lot slower growth.

今年的增长率大概会在20%左右。

We're going to grow about 20% this year.

部分原因是我们没能及时开发出自己的模型，没法向更广的用户群提供AI功能。

And it's in part because we've struggled to really develop our own models in time and be able to offer kind of AI features to a wider set of our user base.

Canva现在的处境挺棘手的，他们还在想办法解决这个问题。

And Canva is like a really interesting sort of pickle or has been in a really interesting pickle that they they are trying to figure out.

他们的大多数用户都是免费使用Canva的。

They most of their users use Canva for free.

他们采用的是免费增值模式，而且你知道，当这家成立13年的公司只是提供基于云的设计软件工具时，服务用户的成本，当你需要支付推理令牌的费用时，就变得高多了，你知道，就是那些我可能提到的所有昂贵的AI术语。

They they have a freemium model and, you know, sort of the cost to serve that user when, you know, this 13 old company was just offering, you know, sort of cloud based, you know, software design tools is a lot more expensive when you're needing to pay for inference tokens, you know, sort of the, all the, all the expensive AI words that I could say.

所以当他们基本上是通过像ChatGPT、Claude、各种Gemini工具这样的工具来提供AI服务时，成本太高了。

And so when they were essentially offering AI through, you know, tools like ChatGPT, Claude, various Gemini tools, it was too expensive.

比如今年早些时候，他们不得不放慢AI的推广速度。

Like earlier this year, they had to essentially slow a rollout of AI.

结果，你知道，收入增长就放缓了。

And, you know, as a result, have seen like slowing revenue growth.

当然，如果他们不能给用户提供免费的AI，用户就会开始尝试并转向像ChatGPT这样提供免费AI的工具。

And of course, like if they're not able to offer their users free AI, users will start experimenting and turning to tools that do, like ChatGPT.

这种动态在很多软件公司中都有体现，但我觉得在Canva身上特别值得关注，因为它的商业模式很独特。

And so those dynamics are kind of echo across a lot of software companies, but I think are really interesting to kind of look at at Canva, which, you know, sort of has this unique business model.

在你走之前，我还有最后一个想法。

And one last thought before I let you go.

我听过一个很有意思的理论，就是说你看像OpenAI和Anthropic这样的公司，有些人认为，我们可能在未来一年内就会达到一个点，他们可能不再通过API公开发布模型了，对吧？

I have heard this really interesting theory, which is that when you look at companies like OpenAI and Anthropic, there is this theory that some people have, which is that, you know, we might reach a point even in the next year where they might stop kind of publicly releasing models through an API, right?

其实我们已经看到一些最新的模型有这种情况，就是发布是分批的，或者政府要先审批客户，然后他们才能用这些模型。

Like we already kind of see this happening with some of the most recent models where they have like a staggered release or the government has to, you know, approve customers before they can access these models.

我的意思是，如果我们真的相信这些模型会变得这么危险，我觉得很有可能他们会停止提供这些模型。

I mean, if we really believe that these models are going to become so dangerous, I mean, I think there's a very real possibility that they stop maybe offering them.

那样的话，OpenAI和Anthropic怎么赚钱呢？

And in that case, how is OpenAI and Anthropic going to make money?

他们就得做应用了，对吧？

They're going to have to make apps, right?

所以我觉得有个论点就是，应用初创公司应该比现在更担心，因为谁知道那个业务会是什么样子，他们是否还能有一两年后的API业务。

And so I think there is this argument that like app, know, app startups should be even more worried than maybe what they are right now because who knows what that business might look like and if they're going to have an API business even in, a year or two.

我觉得，是的，每个风投，每个在这个领域的初创公司高管都在屏息以待，等着这个问题的答案。

I think, like, yeah, every everyone, every VC, every startup executive that is in this space is holding their breath sort of deeply, sort of as we wait sort of where this question goes.

Anthropic和OpenAI怎么往上游走，可以这么说吧？

What does how does Anthropic and OpenAI move up the stack, so to speak?

而且我觉得每个人都喜欢告诉自己，我们有自己的专长，你懂的。

And I think everyone likes to tell themselves we have our own, you know, sort of specialty.

可以说，我们有自己的护城河。

We have our own moat, so to speak.

而且我觉得这对每家公司来说都不一样，但是，你知道，你一直有那个优势，直到你不再有。

And I think it's it's different for every company, but, you know, you you you have that advantage until you don't.

而且，你知道，所以我觉得对于Canva来说，他们对未来展现出了很大的信心。

And and, you know, so I think with Canva, I think they're conveying a lot of confidence about their future.

他们在说，看，我们已经开发了自己的内部模型。

They're saying, look, we've developed our own in house models.

我们仍然在盈利。

We're still profitable.

我们有十三年的品牌记录。

We have a thirteen year brand track record.

那就走着瞧吧。

And we'll see.