AI界的史蒂夫·乔布斯？, 科技巨头AI混乱阶梯, 2026预言水晶球

对吧？

Right?

他们推出了新的图像模型，叫什么Nano Banana。

They launched new image do like, Nano Banana.

Nano Banana不是这部纪录片的重要部分吗，你知道的

Isn't it is Nano Banana a big part of the documentary documentary of like, you know

我还没看到那里，但有可能。

I haven't gotten to it yet, but it could it could be.

但没错，你说得对。

But, yeah, you're right.

这确实很有趣。

It is it is it is interesting.

而且我认为，我们待会儿会谈到一些预测，但这些公司现在正利用自己的优势来确立定位，尤其是在技术日益商品化的情况下，这今年会非常重要。

And and I think that, I mean, we're gonna get into predictions a little bit, but these these companies leaning on their edges and sort of getting their positioning now, especially as things commoditize is gonna be a big deal this year.

好吧。

Alright.

让我们短暂休息一下，回来后聊聊这场AI浪潮如何重新洗牌，尤其是对大型科技公司地位的影响。

Let's take a quick break and come back, and we're gonna talk a little bit about what this AI moment has done in terms of the reshuffling of the deck of cards, as far as, like, where big tech stands.

实际上，情况并没有像几年前甚至去年看起来的那么乐观。

It actually isn't as as positive as, you know, it looked a couple years ago or last year even.

所以，我想谈谈这一点，然后我们可能在节目最后留点时间做一些预测。

So, I wanna cover that, and then we'll we'll probably have time for some predictions at the very end of the show.

我们稍后马上回来。

So we'll be back right after this.

让我介绍一下我的合作伙伴NordVPN。

Let me tell you about my partners at NordVPN.

如果你想要观看在你所在地区无法提供的体育赛事、电视节目或电影，可以通过NordVPN切换虚拟位置到正在播放这些内容的国家来实现。

If you ever wanna watch sporting events, TV shows, or films that aren't available in your region, you can do it by switching your virtual location to a country which is showing that content with NordVPN.

NordVPN还能在你旅行时，无论身在世界何处，保护你在使用公共Wi-Fi时的数据安全。

NordVPN also helps protect your data while you're traveling and using public Wi Fi wherever you are in the world.

它是世界上最快的VPN，流媒体时完全不会缓冲或卡顿。

It's the fastest VPN in the world with no buffering or lagging while you stream.

NordVPN 拥有超过7400台服务器，覆盖118个国家，支持轻松切换虚拟位置。

NordVPN has 7,400 plus servers across 118 countries with easy virtual location switching.

它支持最多10台设备，速度极快。

It supports up to 10 devices, and it's extremely fast.

要获得NordVPN计划的最优惠折扣，请访问 nordvpn.com/bigtech。

To get the best discount off your NordVPN plan, go to nordvpn.com/bigtech.

通过我们的链接，您在两年计划上还可额外获得四个月服务。

Our link will also give you four extra months on the two year plan.

NordVPN 提供30天无风险退款保证。

There's no risk with Nord's thirty day money back guarantee.

链接也在播客节目描述框中。

Link is in the podcast episode description box as well.

这就是NordVPN。

That's NordVPN.

我们回到《大科技》播客，今天邀请到每月做客的MG·塞格勒，和我们聊聊科技与人工智能的最新动态。

And we're back here on big technology podcast with MG Seigler who joins us once a month to talk about the, the latest in tech and AI.

你可以在 SpyGlass 上找到他的文章，网址是 spyglass.org。

You can find his writing at SpyGlass, which is at spyglass.org.

强烈推荐。

Highly recommend it.

好的。

Alright.

MG，你写过一些关于人工智能就像混沌的内容。

MG, so you wrote a little bit about how, AI is like a chaos.

混沌是一架梯子，而人工智能的混乱正为公司、尤其是大型科技公司创造机会，让它们能够攀上这架梯子，并重新洗牌商业和技术领域的权力格局。

Chaos is a ladder and the AI chaos is like creating opportunities for companies, big tech companies in particular, to move up that ladder and, and reshuffle the deck in terms of the power in in the business and the technology community.

我们稍后会讨论一些这些公司，但我的主要观点是：

We'll we'll go we'll we'll touch on some of these companies, but here's my big takeaway.

我观察了这些公司目前的状况与以往相比的变化，发现它们中有很多并没有像我之前预期的那样向上攀升，甚至有些在下滑。

I looked at, you know, where these companies stand compared to where they were previously, and I found more of them I don't want I don't wanna say falling down the ladder, but, like, less up the ladder than I, you know, previously expected.

我来简单列举几个例子。

Here, I'll just run through a couple.

比如微软，和OpenAI有大笔合作。

Microsoft, for instance, big OpenAI deal.

但现在他们好像陷入困境了。

Now they're kinda stuck in the mud, it feels like.

有人开始对Copilot提出疑问。

There are people asking questions about Copilot.

他们现在没有像当初把先进的GPT模型整合进Bing时那样 standout 的产品。

They don't have a standout product in a way that they did when they were shuffling state of the art GPT models into into Bing for instance.

Azure确实在快速增长，但这可能只是以一种其他人无法获得的方式转售OpenAI的技术，因为存在排他性。

So Azure is growing growing in a big way, so maybe that's and but that's probably just, I don't know, is that reselling, OpenAI in in a way that, you know, others don't have access to because of exclusivity.

亚马逊是另一个大玩家。

And Amazon was another big one.

对吧？

Right?

你提到的他们的重要合作是Anthropic，而Anthropic现在又和NVIDIA、谷歌，还有谁合作了？

Their big partnership that you mentioned was Anthropic, you know, and and Anthropic is now partnering with NVIDIA and Google and, and, who else?

还有微软。

And Microsoft.

还有

And

微软。

Microsoft.

是的。

Yeah.

你知道，Alexa Plus 如我们之前讨论的那样表现尚可，但 AWS 并没有大放异彩。

You know, Alexa Plus is doing okay like we talked about, but AWS isn't, isn't killing the world.

然后还有苹果，我们谈过苹果的问题，还有 Meta，对吧？它似乎还没理顺头绪。

And then you have Apple, which we've talked about the Apple problems, and then Meta, right, which doesn't seem to have his act together.

它不像早期那样把事情打理得井井有条了。

It's act together in a way that, you know, it did in the early days.

所以，读完你的文章后，我的主要体会是——显然，谷歌表现不错。

So my big takeaway from reading your piece I mean, obviously, Google's doing well.

但读完你的文章后，我的主要想法是：等等，不对劲。

But my big takeaway after reading your piece is like, wait a second.

难道这些科技巨头不会以我们所有人预期的方式抓住这个机会吗？

Like, do do are are the big tech companies not gonna capitalize on this in a way that we all thought?

也许真正的势头其实正在转向人工智能实验室本身。

And and maybe the momentum is actually moving to the AI labs themselves.

是不是大科技公司获得的推动没那么大，而下一个巨头将是OpenAI和Anthropic这样的公司，它们的估值似乎每几周就翻一番？

Like, is it is it that big tech doesn't get as big of a boost and the next big companies are the OpenAI's and Anthropix whose valuations, you know, seem to double every few weeks?

你对此有什么看法？

What's your what's your perspective here?

是的。

Yeah.

我几周前写了这篇文章，也就是新年之前，去年年底的时候。

I wrote this a few weeks ago, you know, ahead of the New Year, so the end of last year.

我当时只是想退一步，看看主要玩家的现状，就像你刚才列举的那些公司。

And I was just, yeah, trying to take a step back and look at the current states of the major players, you know, as you as you rattled off a bunch of them.

而且，再次说一下，我的混乱阶梯类比，我只是在想，谁是更稳定的，谁是不太稳定的。

And, again, like, my chaos ladder analogy, I just think of, like, who are the stable ones versus who are the less stable ones.

我认为目前谷歌绝对是领先的，但情况会变化。

And I think definitely at the top is is Google right now, but it changes.

对吧？

Right?

就在去年年中，他们的股价还非常低迷，我当时还写过这篇文章，而其他玩家的情况则不同。

Like, was in the middle of last year, their stock was, like, super depressed, and I wrote about that at the time, versus where the other players were.

而且，为什么会出现这种情况？尽管他们似乎已经通过Gemini找到了方向，但市场似乎仍然低估了他们。

And, like, why was that even though it seemed like they were, you know, found their footing with Gemini, but it still seemed like they were sort of being undervalued by the market.

而今天，就在刚才，他们首次达到了4万亿美元的市值，加入了这个俱乐部。

Fast forward to today, literally, they just hit 4,000,000,000,000 for the first time, so they joined that club.

你知道吗，过去六个月里，他们的市值上涨了大约1.5万亿美元，这简直不可思议。

And, you know, they're up like a 1 and a half trillion, I think, in, you know, in the past six months in terms of market cap, which is just incredible, obviously.

太疯狂了。

Crazy.

但这恰恰说明了这种情况有多么不稳定，即使是如此规模的公司，也可能在如此短的时间内发生巨大变化。

But it just shows you, like, how how fluid this sort of situation is that it can sort of change that quickly for even a company of that size.

但我目前认为他们是最稳定的。

But I do view them right now as the most stable.

听起来，他们已经牢牢锁定了与苹果的合作关系。

They've got this Apple partnership sort of locked in now, it sounds like.

因此，我认为这种稳定性可能会持续一段时间。

And so I think that that stability is going to probably continue for a while.

我认为Anthropic也是这些更稳定玩家之一，尤其是在与OpenAI的现状相比时。

I do view sort of Anthropic as another one of these more stable players, certainly in, you know, Versus if you compare them to where OpenAI has been.

我认为OpenAI稍微更稳定一些。

I think OpenAI is a little bit more stable.

你说他们都正在下滑。

You say that they're all falling.

我认为他们比以前更稳定了，主要是因为他们至少基本敲定了与微软的交易，现在他们成为了公共利益实体

I think they're a little bit more stable than they have been, mainly because they finally have this Microsoft thing at least somewhat put to bed, right, the deal, and now they're they're the public benefit for

这肯定是稳定的。

the is is stable for sure.

它们更稳定了。

They're more stable.

是的。

Yeah.

相比一年前，那时一切都还不确定，现在肯定好很多了。

Then they certainly than a year ago, like, where everything was up in the air.

不过，还有很多问题。

There's there's a lot of questions, though.

对吧？

Right?

你听到了年底的事情，比如Notion，我们上次见面时可能也聊过OpenAI的API。

And, like, you heard it towards the end of the year, the Notion, you and I talked about it maybe in the last time we were together, the OpenAI API.

对吧？

Right?

抱歉。

Sorry.

OpenAI API。

The OpenAI API.

OpenAI 首次公开募股。

The OpenAI IPO.

他们会在 2026 年上市吗？

Do they go out this in 2026?

他们真的能做到吗？

Can they possibly do that?

如果是这样，像 Anthropic 明显也有传言说他们也会这么做。

And if that's the case, like, is Anthropic obviously, there's talk that they're also gonna do that.

如果 Anthropic 率先上市，这可能会给 OpenAI 带来很多问题。

And if Anthropic beats them out there, like, that causes potentially a lot of problems for OpenAI.

所以，尽管他们现在可能更稳定了，但看起来他们真的需要稳住阵脚，把业务理顺，以便在需要上市时（听起来他们很可能需要上市，因为他们需要筹集大量资金）能够拥有正确的叙事。

And so while they might be more stable right now, they really need to hunker down and get the business in line, it feels like, in order to have the correct narrative to be able to go out if and when they need to, which it sounds like they probably will need to given the capital that they need to raise.

因此，再次强调，这些公司目前的稳定性水平，我认为是一个动态变化的目标。

And so, again, the level of stability that these companies have right now, I think, is a moving target.

但对我来说，它们似乎正开始逐渐分化，形成一些更稳定和不太稳定的群体，虽然尚未完全固化。

But it does feel a little bit to me like they're starting to sort of not certainly not cement in place, but starting to coalesce around these, like, more stable and less stable pockets.

Meta 明显不稳定，因为他们把一切都搞砸了。

Meta's obviously unstable because they they blew up everything.

我们拭目以待。

We'll see.

我们可能距离他们宣布新的 AGS 超智能计划成果只差几周，甚至更短的时间。

We're probably a few weeks away, maybe less from whatever they're going to, you know, announce in terms of coming out of their new AGS superintelligence efforts.

他们的牛油果模型？

Their Avocado model?

对。

Yeah.

那是他们的内部代号。

That's what the the codename was.

对吧？

Right?

嗯。

Yeah.

所以我们很快就会看到。

So we're gonna see that soon.

微软，你知道的，由于这种正式又非正式地与OpenAI切断关系，现在他们可以独立行事了。

Microsoft, as you know, like, because of this sort of official, unofficial severing of the ties with OpenAI, they're now free to do this on their own.

但感觉起来，嗯，这条路并不顺利。

But it feels like, you know, that hasn't been easiest going.

因此，在某些方面，他们被过去与OpenAI合作取得的成功拖住了脚步。

And so in some ways, they're sort of held back by the success that they saw with OpenAI.

对吧？

Right?

而且你知道，他们仍然保持着这种独家合作关系。

And as you know, they still have that exclusive partnership.

因此，在很多方面，他们面临着一种奇怪的局面：必须在多个领域继续与OpenAI合作，他们确实也在合作，但同时又试图独立发展、与OpenAI竞争，并投资Anthropic等公司，尽管这些行为彼此之间本质上是相互竞争的。

And so in many ways, they have this weird situation where they have to sort of obviously still partner with OpenAI on a number of fronts, and they are, but trying to do their own thing and compete with them and and make investments in Anthropic and do all these other things, like, even though, you know, it's it's all sort of competitive with one another.

所以他们现在处境很尴尬，亚马逊，你知道的。

And so they're in a weird spot, Amazon, as you know.

但同样，感觉正逐渐明朗，谷歌目前处于非常有利的位置。

But, again, it's it's sort of feels like it's coalescing that that Google's in a really good position right now.

我认为Anthropic的处境相当不错，因为他们的模型至少看起来能匹配他们想要实现的目标。

Anthropic, I think, is in a pretty good position because they it feels like they have the right sort of model at least to match, you know, what they're trying to do.

再往下看，情况就越来越复杂了，目前最混乱的玩家主要是那些曾经大张旗鼓地启动、后来又推倒重来的公司。

And then going on down the line, like, it goes from there and, you know, all the all the way down to sort of the real the real chaotic, players right now, which is the mainly the ones that have that have blown it up and try to and started over.

对。

Right.

但有意思的是，想想我们刚才列出的这些。

But, you know, it's interesting because think about what we just listed here.

所以，在大型科技公司中，处于不稳定或较不稳定状态的是亚马逊、微软、苹果和Meta。

So, of the big tech companies, the ones that are in to use your your system in unstable or less stable positions are Amazon, Microsoft, Apple, and Meta.

我的意思是，五大科技巨头中有四个处于不利地位，这让我很惊讶，因为按理说到了现在这个阶段，距离ChatGPT发布已经三年多了，他们本都应该处于绝佳位置来抓住这个机会，而他们却罕见地落在了后面，但看起来确实如此。

Mean, that's four out of the big five that are not in a good position, which is surprising to me because you would expect them to be at this point, right, we're now three years and change past ChatGPT's release you would expect them all to be in prime position to be capitalizing on this, and it's very rare for them to be behind the eight ball, yet that's where it seems like they are.

我的意思是，你列出的每家公司都有各自的原因。

I mean, I think that they they have their each individual company that you list there has their own reasons.

对吧？

Right?

我们谈过微软与OpenAI的合作关系是一把双刃剑。

We talked about Microsoft's partnership with OpenAI was sort of a double edged sword.

对吧？

Right?

在早期，这确实帮了他们很多，但现在因为这种关系不再是独家的，他们现在必须自己另起炉灶，结果他们在走出这段合作后反而落后了。

Like, it helped them a ton in those early days, but now because that's not an exclusive, like, thing that you know, relationship that they have anymore and they really feel like they have to do their own thing, like, they're sort of they were behind the eight ball when they came out of that.

某种程度上，我觉得亚马逊和Anthropic的情况也类似。

In some ways, I think Amazon is similar with Anthropic.

对吧？

Right?

他们依赖Anthropic，就像微软和OpenAI合作的方式一样。

Like, they relied on Anthropic in the in a similar vein to the way that that Microsoft and OpenAI were working together.

所以现在他们一直在开发自己的模型，并已推出多个模型，但同时也保留了与Anthropic的合作关系。

And so now they've been trying to work on their own models and have launched several of them, but and they still have the Anthropic relationship too.

所以这部分原因在于，这一切都归结于这些概念有多循环往复。

And so a part of it is is all goes to the, you know, the the ideas of, like, how just circular all of this stuff is.

对吧？

Right?

这些参与者都在彼此合作，因为云服务提供商必须与模型开发者合作，而一些云服务提供商同时也是模型开发者。

And all of these players are working with one another because there's the cloud players that have to partner with the model makers, and some of the cloud players are also model makers.

最终，一些模型开发者将成为云服务提供商。

And then eventually, some of the model makers are gonna become cloud players.

因此，他们所有人都必须相互合作。

And so they all but they all have to sort of work with one another.

所以归根结底，从他们在市场中的位置来看，真正的赢家只有英伟达。

And so, you know, at the end of the day, the only real winner, as you could see by the way that the where they are in the market is NVIDIA.

如果这种情况贯穿全年，我们将看到他们在芯片领域面临的所有竞争。

And if if that stands throughout the year, we'll see with all the competition that they have coming their way on the chip front.

但没错，我的意思是，像这些全球最大的公司，我认为它们仍在努力寻找自己的方向。

But but, yeah, I mean, you're right that the the players, like these biggest companies of the world, I think that they are sort of still trying to find their way.

你还会看到各种报道提到，它们正在努力弄清楚究竟该采用哪种模型来利用人工智能。

And you see various reports to this to this idea that they're trying to figure out exactly what the right model should be of how they're leveraging AI.

而且它们都说一切都很顺利。

And and they all say that everything's great.

对吧？

Right?

比如，Meta用它来定制广告，微软用它来推销Azure，一切都很顺利，财报数据也看得出来，诸如此类。

Like, oh, Meta's using it to tailor ads and and, yeah, Microsoft's using it to sell Azure and it's and it's going great and you can see it in the bottom line and yada yada.

但它们之所以都在拼命忙乱、不断重组团队，是有原因的。

But, like, there's a reason why they're all scrambling and why they're constantly reshuffling their teams.

我认为到了2026年，随着它们争相在这一阶梯上抢占位置，这种情况会更多。

And I do think there will be a lot more of that in 2026 as they all try to jockey for position on this ladder.

对。

Right.

我的意思是，我可能会因为这么说而收到一些邮件，甚至后悔说出这些话。

I mean, I might get some emails from this, and I might regret saying this.

但除了通过Azure销售AI服务外，微软还有什么 standout 的AI产品？

But what standout product does AI product does Microsoft have outside of selling AI services from Azure?

除了通过AWS销售一些AI服务外，亚马逊还有什么 standout 的AI产品？

What standout product does AI product does Amazon have outside of selling some AI with AWS?

苹果有什么 standout 的AI产品？

What standout AI product does Apple have?

我们都清楚，他们并没有这样的产品。

We both know that they don't have one.

那Meta有什么 standout 的AI产品？

And what standout product does Meta have?

没有。

None.

AI产品。

AI product.

他们没有任何产品。

They have none.

没有任何。

None.

在消费端肯定没有，即使在企业端也是如此。

Certainly none on the consumer front and even on the enterprise front though.

对吧？

Right?

就像，这长期以来一直是微软所说的，你知道的，那一直是他们的备选方案。

Like, that's long been what Microsoft said would be, you know, sort of that's always been their fallback.

对吧？

Right?

如果他们在消费端做不成，显然他们会尝试在每一种新技术上都介入，但他们在这些领域并没有太多成功。

If they couldn't do it on on consumer side and obviously, they try across the board for every single different technology that comes out and they, you know, they don't have that much success on the fronts.

但你会觉得，他们至少可以依靠庞大的企业销售网络来翻身，就是他们一直以来通过的那些销售渠道。

But, like, you would think that they could fall back to the the, you know, massive enterprise just sales, you know, sell through that they sell through that they have.

而且，正如你所说，他们会说一切看起来都很棒。

And, yeah, to your point, like, they'll say that everything it looks great.

但同样，你得看看这些公司的实际行动。

And but, again, just watch the actions of these companies.

我觉得你说得完全正确。

I think you're exactly right.

我认为对于微软来说，这一点尤其明显，因为你一针见血地指出来了。

And I think for Microsoft in particular, it's it's glaring because you hit on it.

他们不仅有OpenAI的合作关系，还有Bing，作为最早推出的那个产品。

They had not only the OpenAI relationship, they had Bing as, like, the original, like, product out there.

它本应让谷歌坐立不安。

It was gonna make, like, it was gonna make Google dance.

现在谁在跳舞呢？

And who's dancing now?

是谷歌。

It's Google.

微软根本没在跳舞。

Like, Microsoft is not dancing.

你知道吗？

They you know?

完全没有节奏。

So No rhythm.

他们完全没有节奏。

They got no rhythm.

这是一次巨大的错失机会。

And it was a huge a huge missed opportunity.

他们本可以让必应成为真正能与谷歌竞争的对手。

They could have made Bing a real potential competitor to Google.

在某些方面，他们可能太早了，肯定撞了南墙，然后就退缩了，你知道的，那些明显奇怪的对话，没错。

And in some ways, they they, you know, they were probably too early, I'm sure, and they stubbed their toe and they pulled back, you know, the whole, obviously, weird conversations, yeah, going on.

这是历史上多么重要的一刻啊

What a moment in history that's

人们甚至离开了自己的妻子之类的事情。

People leaving their wives and stuff like that.

但尽管如此，他们当时就在那里，本来完全可以拥有像ChatGPT那样的成果，但他们就是没有聚焦在正确的关键点上。

But but still, they were there, and they they could have basically had what ChatGPT became, and, you know, they just didn't sort of focus on the right elements of it.

我知道。

I know.

我在这个话题上说得比预想的要长，但我真的认为这非常值得讨论。

I'm going a little longer on on this than I anticipated, but I do think that we're really this is really worth talking about.

当你看看真正的焦点在哪里时，

It's like, then you look at where the action is.

ChatGPT，从零开始。

ChatGPT, ground up.

Cloud Code，从零开始。

Cloud Code, ground up.

是的。

Yep.

Gemini 也是从零开始构建的。

Gemini, effectively, ground up.

是的。

Yep.

你知道，当我跟萨姆交谈时，有一件事让我印象深刻，他说你不能只是把AI简单地附加到现有产品上。

It's, you know, one of the things that's just sort of stuck out with to me when I was, you know, speaking with Sam was he was like, you can't just bolt AI on the existing products.

没错。

Yeah.

你需要从零开始构建。

You need to build ground up.

当我们进行这场讨论时，这些话显然显得非常有道理。

And, you know, as we have this discussion, you know, those words are certainly, like, seeming like they're ringing true.

就像现在，商业格局看起来是两方面的。

Like, it's the businesses the business picture right now is looking like twofold.

一是向那些试图基于你的API构建产品的公司销售AI服务，二是从一开始就打造原生的AI优先产品并逐步发展壮大。

One, you sell AI services to companies who are trying to, like, build on your API, or two, as you build these, like, native AI first products and you grow them from the very start.

是的。

Yeah.

这个故事并不新鲜。

And and this story is nothing new.

对吧？

Right?

就像移动互联网时代一样，许多公司崛起是因为他们从零开始打造了移动优先的应用，比如Airbnb，当然他们起步稍早一些，但正是这一点帮助他们迅速腾飞。

Like, this is what happened with mobile too, where, like, a bunch of players rose because they built these mobile first applications from Airbnb and, you know, they started a little bit before that, obviously, but that's what really helped them take off.

但Uber是一个绝佳的例子。

But Uber Uber is a great example.

对吧？

Right?

为什么Avis没有出现在App Store里呢？

Like, why why didn't Avis where was Avis, in the, in the App Store and

购买特斯拉。

Buying Teslas.

所以，本质上，你必须真正专注于掌控整个体验，并以原生的方式去做。

And so, like but the it's the same story of, like, yeah, you you've gotta really go after sort of the owning the whole experience and sort of doing it in a native way.

我甚至回想起Facebook的时候，我的意思是，我最近经常想到这一点，因为我觉得这在Meta最近的状况中经常被提及：当他们犯了关于HTML5的错误，认为我们不需要原生移动应用。

I even think back to when when Facebook I mean, I I often think back to this these days, because I feel like it comes up a lot with meta and and what's been going on there recently, where it's like when they've sort of made their mistake around HTML five and thinking like, oh, we don't need native mobile apps.

觉得这不会是什么大事。

Like, it's not gonna be a big thing to be a big deal.

我们要做的是一个可扩展的方案，让开放网络主导一切，但结果并非如此。

We're going to do this extensible thing where, you know, the open web is going to take over and it just didn't play out that way.

扎克伯格曾表示，这至少是他们在上一个时代最大的错误之一。

Zuckerberg has said that that was one of their biggest mistakes in the previous era, at least.

所以，再次强调，你必须围绕技术本身来构建产品，而不是简单地把技术硬加进去。

And so, you know, again, like the notion of you got to build these products around what the technology is and not just try to bolt on the technology.

你总能看到人们一再犯同样的错误。

You see everyone make the same mistake over and over again.

说到你的观点，我认为这个说法很好，那就是微软的100%原生AI产品到底在哪里？

And to your exact point, I think it's a good framing of it that like, where is Microsoft's a 100% sort of, you know, native AI product?

他们可能会说那是Copilot，但其实它更像是各种东西的拼凑。

Like, they'll say it's Copilot, but it's sort of like, you know, it's it's sort of like an amalgamation of a bunch of things.

首先，Copilot这个品牌命名一团糟，因为企业版叫Copilot。

First of all, Copilot, the branding is a mess because it's the enterprise thing is called Copilot.

消费版也叫Copilot。

The consumer thing is Copilot.

他们还有各种可爱的角色。

They've got characters, cutesy characters.

他们还在Office里搞了各种不那么可爱的AI功能。

They've got, you know, all sorts of of non cutesy things for doing it within Office.

他们根本理不清品牌定位。

They can't figure out the brand.

你上office.com，就会被直接推到Microsoft Copilot应用里，当时围绕这个糟糕的品牌命名还引发了一轮舆论风波。

You go to office.com, and you're you're shoved into Microsoft March app, whatever, like, the the press cycle was around that, like, their awful branding.

所以他们面临着各种奇怪的品牌问题，但亚马逊会说，显然是Alexa，而且看起来他们正在Alexa Plus上取得某种程度的成功。

So they've got all sorts of of weird branding issues, but Amazon would say, obviously, it's Alexa, and and they are having starting, it sounds like, to have, you know, some level of success with Alexa Plus.

但某种程度上，他们也被拖累了，因为Alexa属于上一代产品，对吧？虽然去掉了‘Plus’，但Alexa已经部署在数百万甚至上千万台设备和家庭中。

But in some ways, they were held back too because Alexa is from the old generation, right, of, you know, drop the plus, but Alexa was on millions and tens of millions of devices and tens of millions of homes.

因此，他们再次被自己在这一领域取得的成功所束缚。

And so they were, yeah, again, held back by the success of what they had in in that front.

所以他们不想彻底推翻重来。

And so they don't want to blow it up and start over.

因此，正如我所说，Alexa Plus确实如此。

And so as a result, Alexa Plus, I agree.

它比以前更好了，但仍然感觉像是后来加上的。

It's like it's better than it was, but it still feels a bit tacked on.

我不认为它对未来的AI发展方向会特别有利，而像Chechi BT这样的公司则不断推出新想法和功能，因为他们从一开始就构建在原生AI环境中。

And I don't think it's particularly going to be good for, you know, sort of these go forward future initiatives in AI, whereas something like Chechi BT, they just keep rocking and rolling, rolling out these new ideas and features because, again, they've been building in this AI native environment from the get go.

是的。

Yeah.

我的意思是，这场对话让我意识到，当我们开始思考这个问题时，也许人工智能如果成功了，其颠覆性会比很多人想象的还要大。

I mean, this conversation just suggests to me, like, as we start thinking about this is maybe AI is you know, if AI works, it's gonna be more disruptive than I think a lot of people imagine.

因为如果它真的会对这些 incumbent 科技公司构成威胁——顺便说一句，这些公司正是在这一领域投入巨额资本支出的——那它对其他所有事物又意味着什么？

Because if it's going to pose this threat to the incumbent tech companies, who by the way, they're the one putting all these this CapEx into into the equation, then what does it do for everything else?

关于这一点，我想补充最后一句，因为你的观点让我突然有了一个想法。

And the last thing I would say about this, because this you you your point jogged a thought in my head.

我觉得，关于人工智能，人们经常说，它就像当年的网站一样。

Like, I feel like with AI, you've often heard it said that it's basically it's like a website.

对吧？

Right?

每个公司都必须整合人工智能，就像当年每个公司都必须拥有网站，后来又必须有移动应用一样，这已经成了基本要求。

It's like every company is gonna have to have AI integrated, and it's just like every company had to have a website and eventually a mobile app, and, you know, it's like table stakes stuff.

对吧？

Right?

但如果事实并非如此呢？

But what if that's not the case?

对吧？

Right?

我的意思是，我认为在某种程度上确实会这样。

I mean, I think that it will be the case to a degree.

我确实认为所有这些网站和服务都会内置AI。

I do think that all these sites and and services will have AI baked into them.

但如果真正成功的公司是那些积极利用AI的公司呢？

But what if the really successful companies are the ones that are doing this, like leveraging AI?

我过去写过、谈过这一点，感觉AI确实让一些公司借助这种新兴技术变革，迅速跃升为全球最大的公司之一。

And I and I've written about this and talked about this a little bit in the past where it does feel like AI, there are companies that jump up in sort of the echelons into the biggest companies of the world by leveraging sort of these new technological shifts that come into place.

我认为大多数人会同意，AI就是这样的技术之一。

And I think most everyone would agree that AI is one of those.

所以，如果你只是简单地把AI附加到现有产品上，这不会让你的公司市值达到一万亿美元、两万亿美元或三万亿美元。

And so if you're just really tacking it on to your current products, that is not going to be what gets you to over the trillion dollar market cap pump or 2,000,000,000,000 or 3,000,000,000,000.

对吧？

Right?

这些目前市值低于一千亿美元的科技公司，只是试图将AI附加到现有产品上。如果我是他们，至少会内部设立一个臭鼬工厂项目，专门基于AI开发全新的产品，而不是仅仅做表面整合。

These companies that are out there right now that are sort of the sub $1,000,000,000,000 tech companies that are just trying to bolt this stuff on, like, if I were them, in some ways, I would at least have a Skunk Works project internally to, like, build entirely new products based off of AI and not just trying to tack it on.

因为历史已经证明，唯一能让你进入顶级公司行列的方法就是这么做。

Because that's again, history has shown that's the only way to actually get into the conversation and become one of the biggest companies in the world is by doing that.

没错。

Yep.

我认为谷歌在这方面做得很好。

And I think that's what Google has done well.

是的。

Yes.

是的。

Yes.

有几家公司做到了这一点，而谷歌正是其中的领军者。

There's a few companies that have done that, and Google is right there at the forefront of that.

没错。

Yep.

好的。

Okay.

所以，我们还剩十分钟，我 definitely 想聊聊。

So, I definitely wanna get into we have, like, ten minutes left.

我想谈谈一些预测。

I wanna get into some predictions.

我认为，尽管苹果遇到了一些问题，但今年它仍将取得巨大成功。

I think you and I both believe despite the the issues that Apple's had that it's gonna have a big year.

你今年的一个预测是，苹果的折叠屏手机将会大获成功。

One of your predictions for this year is that the folding phone is gonna be a big hit for Apple.

我觉得会的。

I think so.

我的意思是，这有点逆潮流而行，因为人们可能对它持怀疑态度，但我知道我们之前讨论过这个问题。

I mean, again, this is this is a little bit going against the grain because it does feel like that people are may maybe not skeptical of it, but I know you and I have talked about it.

对吧？

Right?

比如，我有一部 Pixel Fold 作为我的安卓备用机，用来测试各种东西。

Like, you were I I like the I have a Pixel Fold as my sort of Android backup device that I test things on.

我总体上非常喜欢它，但我觉得它让我看到了折叠版 iPhone 将会如何运作的方向。

And I do generally like it a lot, but I I think that I like it in that it it shows me the way forward of how the I the folding iPhone will work.

但我确实认为，如果他们能解决屏幕没有折痕的问题——从 CES 上的一些泄露信息来看，三星似乎已经做出了不会起折痕的优秀屏幕。

But I do think, like, if they nail the screen without a crease, which seems like out of CES, there were some leaks that maybe Samsung has a great screen that doesn't crease anymore.

所以感觉各种关键要素正开始逐步到位。

So it feels like the the pieces are starting to fall into place.

然后你再想想，我认为它未必会像大家普遍假设的那样，只是把标准 iPhone 展开变成一个更大的屏幕。

Then you go into, like, I don't think it will just necessarily be what everyone just assumes, which is just the standard iPhone that you unfold, then it's a bigger screen.

如果最终它的使用场景和营销方式有点不一样，我一点也不惊讶。

I wouldn't be shocked if it's it's going to end up being sort of like a little bit different of a use case and the way that they market it isn't a little bit different.

关于正面的长宽比，目前没人确定，但可能和普通手机不一样。

There's there's a whole weird no one knows for sure, but there might be a different aspect ratio to the front of the thing.

折叠起来时，它可能会更小、更宽一些，人们或许会喜欢这种更小巧、便于放入口袋的设备。

It might be a little, like, smaller and fatter when it's when it's folded, and people might appreciate having a smaller sort of device in their pocket.

而当它展开时，可能更像 iPad mini，这是我非常喜欢的设备，但它并没有像更大的 iPad 那样获得广泛的共鸣。

And then when it's unfolded, maybe it's more like an iPad mini, which is a device that I love, but it hasn't really resonated as big as as, you know, sort of the bigger iPads have.

但无论如何，也许这些不同的定位方式最终能带来成功。

But so anyway, so maybe there's these different these different ways to sort of frame it that does end up being a success.

如果它能与谷歌的新合作——让 Gemini 驱动 Siri——时机恰到好处，如果它成为苹果首款真正拥有强大 AI 功能的设备之一呢？

And if it's timed well with this new partnership with with Google to get Gemini powering Siri, what if it's one of the first devices with, like, really good AI finally from Apple?

这无疑会是一个非常有吸引力的卖点。

That could be a very compelling selling point, obviously.

当然。

Definitely.

我期待着蒂姆·库克出场，发布 iPhone Origami，让你能把 iPhone 变成一个 I

I'm looking forward for, to Tim Cook coming out there and introducing the iPhone, origami, and you're able to make your iPhone into a I

就是这部分。

like that part.

变成一只鸭子。

Into a duck.

是的。

Yeah.

这是蒂姆·库克，

It's Tim Cook,

然后是蒂姆·库克，老兄。

then Tim Cook, man.

可能是任何人。

Might be any number of people.

他可能

He might

会有一个相当棒的iPhone退休方式。

be one hell of an, of an iPhone, retirement.

他把它变成一个小算命工具，然后说：好了，我搞定了，各位。

He turns it into, like, a little fortune teller thing and then he goes, alright, I'm done everybody.

我要把这个交给约翰·特尼斯。

I'm gonna give this over to John Turnis.

祝你有个愉快的一天。

Have a have a good one.

嗯。

Yeah.

说到蒂姆·库克，《纽约时报》又发了一篇文章。

Speaking of it, of Tim Cook, the, the Times had an article that talked again.

我的意思是，这更像是无端猜测，根本不足以说明这里有什么大事发生。

I mean, this is this is more smoke than than you need to say that something's going on here.

《纽约时报》有篇报道说，蒂姆·库克告诉朋友他感到疲惫了。

He's the Times had this story that said Tim Cook has, told friends he's tired.

这 guy 要退休了。

This guy's gonna retire.

我觉得这可能不是，这可能是蒂姆·库克的最后一年了。

I think he's this is not this is not the, this is the last year, I think, for for Tim Cook.

只是，你知道的，不是基于任何确凿证据，只是看了这些报道，我觉得他真的要走了。

Just just, you know, not based off of any anything more than reading the reports, but I I think he's done.

我也这么认为。

I think so as well.

我的意思是，显然一直都有反对声音。

I mean, obviously, there's been a bunch of there's always the pushback.

对吧？

Right?

好吧。

Like, okay.

所以这份报告说这样，但谁会真的泄露蒂姆·库克告诉别人他很累的消息呢？

So this report says this, but, like, who would actually, you know, leak that Tim Cook is telling people that he's tired?

我认为格鲁伯对此稍微提出了质疑。

And I think Gruber, you know, pushed back on that a little bit.

而经常报道此事的古尔曼也对整个事件的时间点提出了质疑。

And then, you know, Gurman, who's done a bunch of the reporting on this, has pushed back on, like, the timing of the whole thing.

但我认为现在迹象太多了。

But I think there's too much smoke now.

很明显，各种迹象都在为他提供一种可能性，让他能在想退出时抽身，而且感觉今年晚些时候就会发生。

Like, clearly, the pieces are being put in place for him to at least have the optionality to step back when he wants to, and and it does feel like it's going to be at some point in this coming year.

你注意到了吧，新闻里提到他们让现任董事长亚瑟·莱维森参选连任，尽管他已经75岁了，而75岁通常是苹果董事会成员非正式的退休年龄。

You saw there was, you know, the news cycle around they're letting Arthur Levison, who is the current chair, stand for reelection even though he's 75, which is the unofficial official time of retirement for Apple board members.

你可能会认为，这显然预示着库克最终会接任董事长一职。

And you would think that that has to play directly into the notion that they're gonna have Cook obviously step into that chairman role eventually.

但他还尚未准备好辞去CEO职位，所以这整个阶段都属于过渡期。

But because he's not ready to step down from the CEO role yet, you know, this is all sort of the period of transition.

但确实感觉一切都在逐步就位，让库克能够接任董事长职位，或许也让约翰·特恩里斯接替他。

But it does, feel like that everything is is moving into place to have Cook be able to do that, to step into that chairman role, and perhaps to let John Turnis come in.

你知道，在同一则报道中，也有人提出疑问：是否还有其他潜在候选人？

You know, there was some push in that same story about, like, oh, are there other candidates that that could be in the offering?

你知道，那些被提及的名字都是些老生常谈的人选。

You know, a lot of the it's all the standard names that are thrown out there.

但大多数人的年龄都和库克相仿，只有克雷格·费德里吉例外，我认为他也是胜任这一职位的真正候选人。

The problem with most of them is that they're almost the same age as Cook, except for Craig Federighi, who is another one who I would view as a legitimate contender to do the job.

而且我觉得，这必须在特努斯和费德里希之间做出选择。

And it's I you know, it has to be sort of between those two, between Ternus and Federighi.

但所有的报道都指向特努斯，所以你不得不相信这就是趋势。

But everyone, you know, all the reporting points to Ternus, and so you have to believe that that's the thing.

我认为这其中的一个变数是，我很想知道他们会聘请谁来接替约翰·詹德里亚，担任人工智能部门的负责人。

The one thing that I do think that's a wildcard in this, I'm very curious to see who they hire to replace John Giandrea as, you know, the head of AI.

目前所有的报道都显示，来自VisionPro的迈克·洛克威尔已经接管了该部门，并向负责整体战略的费德里希汇报。

The the all the reporting right now has, of course, like, Mike Rockwell from VisionPro has sort of taken over the org, and then he reports up to Federighi who is sort of overseeing the entire, you know, strategy of that.

因此，他们要么让他继续领导，但你认为他们必须引进一位专注于人工智能的技术专家来真正统筹整个项目。

And so either they let him continue to run it, but you think that they have to get a sort of an AI focused technologist in there to actually sort of oversee this entire thing.

所以最终人选是谁，我们可以回到去年我们一直在讨论的话题：苹果会收购哪家公司？

And so who that ends up being, we can go back to you and I talking all of last year about like, who does Apple acquire?

也许他们会收购某家公司，以引进合适的人才来领导这个部门。

Like, maybe they do acquire someone to bring in to bring in the right talent in order just to lead that division.

我不会说这个人会是伊利亚·苏茨克弗。

I'm not gonna say it's like Ilya Sutskever.

我不认为他们会进行一笔像300亿美元、304亿美元或500亿美元这样的交易，仅仅为了购买安全的超级智能。

I don't think that they're gonna do a deal that's like, you know, the $30,000,000,000, $30.40, $50,000,000,000 that it would take to buy safe superintelligence just to do that.

但他们需要一个像那样的大人物来领导这项事业，因为目前他们只是用一个临时方案来应对——你可以称之为创可贴，而且是一个非常有效的创可贴，也就是和谷歌的合作。

But, like, they need a name like that, of that caliber in order to lead this and give people they they now have the sort of Band Aid stop gap, if you wanna call it that, a very good Band Aid, in the Google deal.

但最终，苹果希望自主掌控这项技术，而得益于与谷歌的合作，他们有了数年的缓冲时间，可以暗中开发自己的项目，因此他们需要找到合适的人选。

But eventually, Apple's gonna wanna own this and do this themselves, and so they need the right person that they now have years of runway, thanks to this Google deal, to be able to work behind the scenes on their own stuff.

是的。

Yeah.

一个热门人选是Mistral，我们几天后会邀请Mistral的首席执行官亚瑟·门施来做客节目，谈话快结束时，我突然想起来问了他一句：‘你有没有考虑过把公司卖给苹果？’

One of the popular names is gonna be Mistral, and we do have Mistral CEO Arthur Mensch coming on the show in a couple of days, and, at the end of the conversation I said, oh, I forgot to ask you, are you gonna sell to Apple?

他回答说，他之前在谷歌深脑工作，他说：‘我离开大公司创业，就是为了不再回到大公司。’

And, and he said, I He was at I think he was at Google DeepMind, and he was like, I didn't leave a big company to start a company to get acquired to go back to a big company.

不过，钱能说话。

Now, money's money talks.

是的。

Yes.

但钱总是说话的。

But Money always talks.

但欧盟也说话。

But the EU talks too.

我不知道欧盟想要多少。

I don't know how much the EU wants.

对。

Right.

它

It

可能达不到那里。

might not be to there.

被一个美国的大债务人收购。

Be bought by an American big debtor.

没错。

That's right.

但我们先看看。

But let's see.

嗯。

Yeah.

好吧。

Alright.

所以你还有一个预测，就是有人会收购Perplexity。

So so one more prediction that you have is someone is going to buy Perplexity.

你不觉得会是苹果。

You don't think it's going to be Apple.

会是谁呢？

Who could it be?

我认为最有可能的是三星，因为很明显他们和他们有广泛的合作。

I think the most likely one is Samsung for the obvious reasons that they have a big partnership with them.

不过三星似乎正在更深入地整合Gemini。

Though Samsung is in the process, it sounds like, of of more deeply integrating Gemini as well.

你知道，他们长期以来一直是安卓领域的领导者，所以你会觉得，谷歌做的任何事情，他们都会紧密配合。

You know, obviously, they've they've been a long time leader in the Android space, and so you'd think, like, anything that Google does, they're going be closely aligned with.

但没错，他们确实和Perplexity有合作关系。

But, yeah, again, they have the partnership with Perplexity.

我不觉得收购价格会低。

I wouldn't be shy it's going be a big price.

对吧？

Right?

像这家公司，我记得上次估值是150亿。

Like that company, I think, was last valued at 15 no.

现在可能有200亿了。

Maybe 20 now.

200亿美元。

20,000,000,000.

所以，要拿下这家公司，价格一定会高得惊人。

So anything to take them off the table will be immense in terms of price.

所以这缩小了谁能做这件事的范围。

So it narrows down the options of who could do it.

我认为不会是苹果，原因我们已经讨论过了。

I don't think it'll be Apple for the reasons that we've talked about.

我认为苹果已经试过这条路，但决定不走下去。

I think that Apple kicked those tires and decided not to go down that path.

我认为也不会是谷歌，原因有很多。

I don't think it'll be Google for a host of reasons.

我觉得他们其实并不需要它。

I don't think they really need it.

你可能会看到微软之类的企业试图入场，达成一笔交易，或许有助于重新建立或确立某种面向消费者的立足点。

You could see a Microsoft or someone like that trying to come in and and do a deal that, again, maybe helps reset or establish some sort of consumer more consumer oriented foothold.

类似这样的事情不会让我感到惊讶。

Something like that wouldn't shock me.

当然，你还可以列出其他一些玩家，比如亚马逊之类的公司。

And there's a you know, you can go down the list of a handful of other players, the Amazons of the world and whatnot.

但我确实觉得，音乐业务的增长可能正在放缓，以至于他们不得不为这家公司找出未来的方向。

But I do feel like the music is maybe slowing to the point where they might have to figure out the path forward for that company in particular.

是的。

Yeah.

我对Perplexity的看好程度远不如以前了。

I'm much less bullish on Perplexity than I was.

我只是觉得，他们的功能迟早会被Gemini和ChatGPT这类产品吞并，你真的还需要一个AI搜索引擎来做聊天机器人就能完成的事情吗？

I just think that, like, their functionality is just gonna be swallowed up by the Geminis and the chattypetees of the world, and it's like, do you really need an AI search engine to do the same thing that you can do within chatbots.

是的。

Yeah.

他们曾以为浏览器会是一个巨大的突破口，还说要收购Chrome。

And they thought they thought, like, the browser thing would be, you know, a huge beachhead, and they they said they were gonna buy Chrome.

对吧？

Right?

还有TikTok。

And and TikTok too.

他们还留着胡子，

And They had a beard,

不是吗？

didn't they?

所以他们一直在寻找策略，我认为他们正逐渐耗尽接下来的选项，无法再找到维持所需增长势头的策略，以继续融资。

So they they've been looking for there were these strategies, and, I think they're sort of running running out of options of what the what the next strategy will be in order to get to maintain that momentum that they need to in order to keep fundraising.

网站是 spyglass.org。

Website is spyglass.org.

每当涉及科技和人工智能时，这绝对是我的必读内容。

Definitely one of my must reads whenever it comes to tech and AI.

MG，和你交谈总是很棒。

MG, it's always great to speak with you.

非常感谢你来参加节目。

Thanks so much for coming on the show.

谢谢你，亚历克斯。

Thank you, Alex.

一如既往地聊天真好。

Great to chat as always.

好了，各位。

Alright, everybody.

感谢收听和观看。

Thank you for listening and watching.

正如我提到的，我们周三会回来，和Arthur Mensch一起，然后周五，抱歉，是Mistral的首席执行官Arthur Mensch，周五我们将回顾本周的新闻。

We'll be back on Wednesday, as I mentioned, with Arthur Mensch and then on Friday, sorry, Arthur Mensch, the Mistral CEO, and then on Friday to break down the week's news.

再次感谢收听和观看，我们下次再见，欢迎收听《大科技》播客。

Thanks again for listening and watching, and we'll see you next time on big technology podcast.