Meta计算辩论 | Diet TBPN

Meta 正在退出计算业务，把多余的算力拿去卖了。

Meta is selling compute they're getting out of the computing business.

他们说，现在做我们需要的事，根本不需要自己囤电脑了。

They said, We don't need computers anymore to do what we need to do.

我们真用不着这些设备了。

We don't need them.

接下来我们要把这些设备对外出售。

We're going to be selling them.

Meta Platforms 正在规划一项云基础设施业务，打算把 AI 算力和模型租给客户，跟 AWS、GCP 这些行业巨头竞争。

Meta Platforms is developing plans for a cloud infrastructure business to sell access to AI computing power and models competing with industry leaders like AWS and GCP.

作为 Meta Compute 计划的一部分，公司正在考虑把现有 AI 基础设施上托管的各种模型，连同原始算力一起对外提供服务。

The company is considering selling access to various AI models hosted on its existing AI infrastructure as well as raw computing capacity as part of its Meta Compute initiative.

通过出售多余的算力，Meta 希望能收回在 AI 基础设施上的巨额投入——包括在数据中心和高端芯片上花掉的数千亿美元。

Meta plans to generate revenue from excessive computing power could help return its investment in AI infrastructure, which includes hundreds of billions of dollars spent on data centers and expensive chips.

所以这件事引发了各种反应。

And so lots of reactions to this.

新兴云市场正在下跌。

The neo cloud market is selling off.

有意思的是，Meta 手里有不少新兴云的合同。

Oddly enough, Meta has a bunch of neo cloud contracts.

有些公司现在之所以抛售，是因为他们既是买家，又是竞争对手。

Some of those companies are selling off because now they're a buyer and also a competitor.

大家对 Meta 如何站稳脚跟、找到能 justify 这笔巨额资本支出的业务，看法不一。

Lots of different takes about, you know, Meta finding its footing, finding something that justifies the massive CapEx.

当然，Meta

Of course, Meta

这确实让人很困惑，约翰。

is It's deeply confusing, John.

是吗？

Yeah?

真的吗？

Is it?

我是说，整件事。

I mean, the whole thing.

哪部分让人困惑？

The whole What's

这有什么好困惑的？

confusing about that?

我觉得他们这么做其实很务实。

I think it is practical what they're doing.

是的。

Yeah.

但作为一个一直以来大力支持Meta的人，

But as somebody that that, you know, I would say overall has been a big, you know, cheerleader for Meta.

在我看来，它确实是最好的。

Think it's truly the best in in my view.

这简直是个完美的生意。

It is a perfect business.

这让人对整体战略缺乏信心。

It doesn't give you a lot of confidence in like the strategy overall.

他们签下的这些新型云业务合同，价值高达数百亿美元。

They're signing these neo cloud deals worth tens of billions of dollars.

这些项目都是靠投入数千亿美元才建起来的。

They're built, you know, spending hundreds of billions of dollars.

当然，他们可以说，自己做这类云业务本身就是一笔好买卖。

And yeah, they can make the argument that these type of like doing any type of neo cloud deals themselves is just good business.

这确实是眼下实现投资回报的最佳方式。

It's it's just like how it's just the best way to get ROI today.

没错。

Yep.

但这就让人怀疑，Meta短期内是否真有能用上这些算力的产品问世，而这明明是他们的核心策略。

It doesn't give you a lot of confidence that there's near term products on the horizon for Meta that are gonna be able to utilize that capacity themselves, which is clearly been their strategy.

我知道马克和团队从来没说过想做云业务。

I know Mark and and the team have never have never said we wanna be in the cloud business.

他们提到过

They've talked about

是的

Yeah.

他们讨论过这种可能性。

They've talked about the possibility of it.

对

Yeah.

但MSL的明确目标是个人超级智能。

But the stated goal of MSL is personal super intelligence.

我们还不清楚

We don't know what

我以前挺看好这个方向的，我觉得你也非常看好。

Which I was a fan of, and I think you were a huge fan of.

你确实是。

You were,

就像是

like,

手机上有个助手，穿梭在你的社交网络里。

a manace on your phone going around your social networks.

这正是我最看好这一点的地方。

That's my biggest bull case for all of this.

说实话，我觉得元宇宙这一系列应用里，有太多可能性可以成为日常必备工具了。

Like, it's very there there are so many different applications that I can imagine being a daily driver of in the meta family of apps.

奇怪的是，据我所见，这些想法根本没人认真试过。

Oddly, none of that has really been even tried, in my opinion.

感觉确实有点这样。

It feels like a little bit Yeah.

所有真正重要的

All the really

到目前为止看到的，只有Muse和Spark。

seen so far is Muse, Spark.

在基准测试上表现不错。

Good on benchmarks.

还算可以，你知道的。

Like, decent, you know.

不过，真没什么值得特别兴奋的地方。

Again, not anything to that that anyone should really get that excited about.

作为一个，你知道的……

As a, you know

作为一家API提供商。

As a as an API provider.

没错。

Exactly.

正是如此。

Exactly.

他们确实宣布了会通过API发布这个功能。

They did announce that they were going to release it via an API.

我觉得他们还没发布。

I don't think they have.

也许有吧。

They might

这模型确实不错，但我觉得不会有多少人用。

still It's a good model, sir, but I don't think it will have a lot of demand.

对。

Yeah.

然后我们还看到了Meta Vibes，那是个类似Midjourney的东西。

And then we've seen Meta Vibes, which was a mid journey

是的。

Yeah.

说唱模型。

Rapper.

就算Muse Spark没达到顶尖水平，它能不能在Meta自家的应用里派上用场？

Even if Muse Spark is not on the super giga frontier, can it be good enough to get some work done inside Meta family of apps?

应该是这样的。

Like, should.

我觉得是的，但他们还没找到那个核心亮点。

I would imagine, yes, but they just haven't found that killer feature.

现在有很多应用都是靠人工智能驱动的。

There are plenty of applications that are AI powered.

有不少模型并没有走超级智能那条路，但也已经站稳了脚跟。

There are plenty of models out there that have found their footing without being on the, you know, super intelligence path or on that particular curve.

有意思的是，昨天我们还在聊那个故事，说谷歌告诉Meta：‘我们没这个算力。’

And it's interesting because yesterday we were talking about the story where Google had been telling Meta like, hey, we don't have the capacity Yeah.

给你们用的。

For you.

可现在Meta自己手头却有充足的算力。

And here, Meta is with with plenty of capacity themselves.

我觉得这事儿不能看得太重，毕竟只是彭博社一篇报道。

I don't think we can read too much into this because it's just one article from Bloomberg.

是的

Yeah.

我觉得谁将成为算力的主要买家，这一点会非常关键。

I think it will matter a lot who the potential buyers of compute are gonna be.

是的

Yeah.

如果是一些我認為市場會很感興趣的公司，那情況就不一樣了。

If it's there's a number of companies that I think the market would get excited about.

是的

Yeah.

但如果他们只是去跟别人打价格战，试图直接竞争的话——

But if they're actually just going and trying to compete as

是的

Yeah.

有点奇怪的是，这消息居然是通过泄露出来的，说是打算卖算力，而不是像SpaceX那样，直接拿下Anthropic的大单，一出来就引爆市场，为IPO造势。

It's weird that it came as a leak about around, like, a plan to sell compute as opposed to just what SpaceX did where it was just like, boom, huge contract with Anthropic, lots of excitement going into the IPO.

SpaceX 上市时讲的这个故事，简直完美得恰到好处。

Like, that was such a perfectly massaged story as SpaceX entered the public markets.

要是他们直接说：嘿，

That would have been great if they just said, hey.

我们有个 Frontier 实验室，现在每个月付我们十亿美元。

We have a Frontier lab that's paying us a billion dollars a month now.

而且这笔收入下个季度就会体现在财报里。

And, like, it's gonna show up in earnings next quarter.

大家做好准备吧。

Like, get ready.

但股市对此反应热烈。

But the stock market loved it.

股价涨了不少，我不确定是因为投资者觉得这是 Meta 的巨大增长点，

Like, the stock is way up, and I don't know if it's way up because they see it as a huge growth area for Meta.

还是说这想法其实没那么离谱？

Now, is it that crazy?

我觉得呢

Well, I think that

超大规模客户那边

a hyper scale people have

一直在期待一个平台。

been wondering platform.

这几百亿美元的投入，回报到底从哪儿来呢？

Where's the ROI gonna come from for this hundreds of billions of dollars Yeah.

这笔支出？

Spend?

对啊。

Yeah.

到目前为止，还真没看出明确的回报来源。

And up until now, there's been no obvious place that it's gonna come from.

对吧？

Right?

是的

Yeah.

Manus那笔交易正在取消中。

The Manus deal, that's being unwound.

是的

Yeah.

他们跟MidJourney和Vibes那边也有合作，但情况不太明朗。

They have the the deal with, you know, mid journey and vibes.

这还不清楚。

That's, you know, unclear.

是的

Yeah.

是的

Yeah.

很明显，他们迟早能把AI技术整合到眼镜里。

There's it seems very obvious that they're gonna be able to integrate AI into their glasses over time.

是的。

Yeah.

但如今真正有市场契合度的智能眼镜，其实更偏向于摄像头功能，而不是把人工智能和眼镜结合起来。

But the glasses that have product market fit today are more of just like the I I think the product market fit is really with the camera, not the intelligence combined with a pair of glasses.

对。

Yeah.

令人惊讶的是，我们还没看到任何扩散模型或者图像生成模型的应用。

Which is also surprising that we haven't seen a diffusion model, an image model.

Instagram 整体来说内容质量挺干净的，至少我关注的动态是这样。

Instagram in general is pretty it's it's remarkably slop free, least my feed is.

我很少看到那些泛滥的病毒式低质图片。

I don't see a lot of viral slop images.

但你可以想象，AI 功能比如背景替换之类的东西，如果做得很自然，用户会非常欢迎，也很享受。

But you can imagine AI powered features, background replacement, a lot of that stuff being AI enhanced, and people receiving that very positively and enjoying that.

我最近在用 Adobe 的产品，比如 Photoshop，已经集成了 Gemini 和 Nano Banana，你可以接入各种图像模型，用起来特别有意思。

I was using Adobe products recently, and like in Photoshop, they have an integration with Gemini, Nano Banana, and you can plug in all the different image models, and you can actually use the tools in very interesting ways.

我老是回到那个想法——个人超级智能。

I keep going back to that idea of, like, the personal superintelligence.

我到底想要什么？

What do I actually want?

前几天我遇到一个挺矛盾的情况，我一直对Meta应用里AI功能太少感到失望，我知道我可能是唯一有这种感觉的人，因为Instagram上普遍对AI挺反感的。

I ran into an interesting conundrum the other day because I've been sort of disappointed by the lack of AI in Meta apps, which I know I'm like the only person that feels that way because the general vibe on Instagram is like extremely anti AI.

Meta其实掌握着我发的每一个短视频的详细数据。

Meta does have granular data about every reel I post.

有些表现好，有些就一般。

Some do better than others.

我去了Instagram里的Meta AI，问它：我该多做点什么才能让账号增长？

I went to Meta AI in the Instagram app, and I asked, what should I do more of in order to grow my account?

这其实挺有用的，要是能有个高级的、个性化的数据分析工具，专门帮我分析我的内容就更好了。

This would be useful, sort of a high powered analytics tool, personalized analytics tools for what I do.

但我得到的却是一个特别笼统的LLM回复。

But I got this very generic LLM response.

我们或许能找出来。

We might be able to pull it up.

我记得我把那张图发在动态里了。

I think it's in I shared the image in the timeline.

我看看能不能在这儿调出来。

Let me see if I can pull it up here.

它建议我专注于可持续的自然增长，而不是像互粉刷量这种短期套路。

It said, buffer recommends focusing on sustainable organic growth instead of quick hacks like follow trains.

坚持发短视频，因为它们的曝光量比图文轮播高出36%，而且还有120

Post reels consistently since they get 36% more reach than carousels and a 120

居然在引用一家社交媒体管理SaaS公司的博客文章，这真有点好笑。

It's so funny to be referencing it's referencing a blog post from a social media management SaaS company.

没错。

Exactly.

你本来还以为会怎么样呢。

Like, you would think that Yeah.

如果你有，我的意思是，很明显他们并没有针对这个使用场景做优化。

If you had I mean, and clearly, they're not focused on this use case.

对。

Yeah.

是的。

Yeah.

但你可能会觉得，平台其实有机会为创作者提供更智能的个人化工具，让他们能更专注地持续创作。

But you would think that there would be an opportunity to give creators personal and super intelligence to just keep at creating on the platform.

对。

Yeah.

可以利用 Instagram 的数据分析，看看哪些内容更能把观众转化为粉丝，然后重点投入这类内容。

So use Instagram analytics to see which content converts viewers into followers and double down on it.

这不就是我之前让你做的吗？

Like, that's what I asked you to do.

我问过你，我该怎么改进，才能更好地增加关注者？

I said, what can I do better to do my following?

你手上明明有所有数据，知道什么内容表现好。

Like, you have all the data about what does well.

我不希望自己一个个去看，这条视频有五千播放，

I don't want to go and look at this reel got 5,000 views.

那条有五万播放，

That one got 50,000 views.

这两者到底有什么区别？

What's the difference here?

我希望你来帮我分析这个。

I want you to do that.

与其投付费广告，不如和你领域内的小博主合作，做真实的内容互推。

Collaborate with micro creators in your niche for authentic cross promotion rather than paid ads.

多互动，回复评论，慢慢建立信任。

Engage with responding to comments and testing out trust.

还提到要优化个人主页，在简介里用清晰的关键词，保持视觉风格的一致性。

It also said to optimize your profile with clear keywords in the bio and maintain a consistent visual brand.

这些我其实已经在做了。

A lot of that I'm already doing.

很多内容都过时了。

A lot of that is old.

所以我挺失望的，不管你怎么评价Muse Spark和它的标准之类的。

And so I'm just sort of disappointed that say what you want about Muse Spark and its benchmarks or whatever.

很明显，如果他们能把这个模型和用户的实际数据打通，本来是可以很好地整合的，但他们就是没把产品做扎实。

Clearly, if they wire that model up to the user's data, they should be able to integrate appropriately, and they just haven't productized it properly.

也许等我用上Manus，再接个API，能勉强做到，但我觉得这功能就该直接集成在Instagram的搜索框里，毕竟他们那儿已经有L了。

And maybe if I get Manus and I get an API integration, like, I could get there, but, like, it should just live in the Instagram search box, I imagine, since they already have an L on there.

这纯粹是个老功能了。

It's just a legacy one.

从产品角度看，光说‘App里AI需求很大’是不够的。

On the product side, it's just not enough to get to, okay, there's crazy demand for AI within the app.

亚当，你问过他们吗，股东价值要是三颗蛋，那那只鹅的价值是多少？

Adam, is ask it if the shareholder value is three eggs as the goose value?

这听起来就是一堆词胡乱堆在一起。

That just sounds like a like a jumble of words.

这话说得对吗？

Is that even a correct sentence?

我不确定。

I don't know.

听着挺通顺的。

Sounds correct.

对我来说。

To me.

说实话，现在最能引爆Instagram增长的，反而是亚当·梅塞里的那些正面镜头视频。

Mean, honestly, like the best it's place very ironic, but the best place for Instagram growth hacks these days is just Adam Messeri's front facing videos.

他经常发Instagram。

He takes to Instagram.

他特别会用Instagram。

He uses Instagram very well.

他直接发帖，拍了很多不同主题的短视频。

He goes direct, posts reels about all sorts of things.

所以今天我刷到一条他的视频。

So today I got surfaced one.

如果你发了内容但效果不好，该不该删掉再重新发一遍？

If you post something and it doesn't do well, should you delete it and then post it again?

他说不用，因为算法还是会给出同样的结果。

And he says no, because the algorithm will give the same result.

那些看过第一次的粉丝本来就不喜欢，第二次再看到只会更反感。

And your followers who saw the first one didn't like it, they're going to see it and like it even less the second time they see it.

他只是回答一堆常见问题，做问答，这反而是了解Instagram平台运作机制最好的方式，只是不够个性化。

He just answers a bunch of common questions, he does Q and As, and it's actually the best way to communicate and get insights into how the Instagram platform works, but it's not personalized.

他给的建议都是千篇一律的通用版。

Like, he's just giving generic one size fits all advice.

我想要的是加上Meta人工智能、专门针对我账号的Adam Messeri大脑。

I want the Adam Messeri brain enhanced with Meta's AI tailored to my account.

我就想要这样的东西。

That's what I want.

就像一个社交媒体的智能助手。

And, like a social media copilot.

不过也许我这种想法比较少见。

But maybe I'm in the minority here.

我不太确定。

I I don't know.

说不定真就我一个人想要这个。

May maybe I'm the only one that would want this.

我觉得，先从创作者工具入手会是个不错的切入点。

I feel like, tools for creators would be a great way to to launch this.

哪怕是个小用户，几乎都关心‘要是能多点播放量就好了’，就算他们主要只是刷着玩。

Even like a small user, they everyone almost always cares about, 'I'd love to get a couple more views on whatever,' even if they're just using it, being primarily a consumption tool.

还有智能购物这个方向，去年我们在Meta Connect上问过马克·扎克伯格，就是那种Meta Ray-Ban眼镜，有AR显示和AI功能，你看到一双鞋，直接说‘帮我下单’就行。

And then there's also agentic shopping, which we asked Mark Zuckerberg about at Meta Connect last year, this idea that they're Meta Ray Ban displays, they have the HUD, they have AI, you should be able to look at a pair of shoes and say, Order me those.

他当时比划了一下，说这可能是未来的一种方向。

And he sort of gestured towards that being one potential possible future.

但真让我觉得不可思议的是，我们连购物流程中至少减少一步点击都没做到，也没多存点用户数据，缩短转化路径，提升转化率。

But it's crazy to me that we haven't even seen them really try to remove at least one click from the shopping experience, store some more of your data, shorten the funnel, increase conversion rates.

这对品牌有好处。

That's good for brands.

对在Meta上做广告的公司也有好处。

That's good for companies that advertise on Meta.

对Meta本身也有好处。

It's good for Meta.

对用户也有好处。

It's good for users.

可偏偏在把AI融入整个应用家族的过程中，大家似乎都没往这个方向使劲。

And it just doesn't feel like that's where the energy has been in terms of productizing AI within the family of apps.

所以我觉得，电商领域才会真正看到这种智能购物的落地。

And so it feels like e commerce will see agentic shopping happening.

我们正在越来越接近了。

We're getting closer.

电脑的使用体验越来越好。

Computer use is getting better.

API都有了，MCP服务器也到位了，Shopify还提供了不少相关工具。

APIs are there, MCP servers, and Shopify has a bunch of tools for this stuff.

但奇怪的是，Meta居然连试都没试过，我们也没看到他们推出过什么功能，比如你看到广告后点个按钮，智能助手就能帮你完成结账，所有信息确认都在Meta应用里搞定，根本不用跳转到Safari页面。

But it's weird that Meta hasn't even been experimenting there, and we haven't seen like, oh, yeah, like they launched a thing where if you see an ad, you can click a button, and the agent will try and go check out with you and then just confirm the details within the Meta app, you don't actually have to open up the Safari window.

我也不确定。

I don't know.

也许这个办法行得通。

Maybe that works.

也许行不通。

Maybe it doesn't.

至少先做个实验吧。

At least run the experiment.

也许他们已经做了。

Maybe they have.

我不清楚。

I don't know.

也许我只是没注意到。

Maybe I just it.

对。

Yeah.

这感觉像是扎克伯格去年所描绘的超级智能野心正在逐渐收场。

This feels like a wind down of, like, the super intelligence ambitions that Zuckerberg was gesturing towards last year.

但我也不确定。

But I don't know.

也许这正是其中的一条路径。

May maybe this is in the path.

也许这只是一个暂时的现象。

Maybe maybe this is just a temporary thing.

是的。

Yeah.

而且，你看看SpaceX和谷歌、Anthropic的合作，根本不是五年期的长期协议。

And and and, you know, if you look at the SpaceX deals with Google and Anthropic, they were they weren't like five year deals.

对吧？

Right?

那都是双方短期合作、各自脱身的机会。

They were shorter term opportunities for both sides to get out.

我觉得Meta完全可以效仿，这其实是个很实际的决定——我们确实有多余的算力，远超当前需求。

And I think that Meta could easily do something like that where they could make again, this is like the practical decision and say like, hey, we actually do have way more capacity than we need.

对。

Mhmm.

我们长远来看当然希望把资源都用满，但现在得先把产品推起来。

We plan on being able to utilize it fully over time, but we need to get our products sort of ramped up.

所以这段时间，为什么不把多余产能卖出去？也能向市场表明，我们不是完全不讲道理。

And so in the meantime, why not sell that capacity and signal to the market that we're not entirely irrational?

是的。

Yeah.

对。

Yeah.

我觉得Meta其实可以等，等到他们需要实现的功能变得非常明显为止。

I think Meta can actually wait be until, like, the the features that they need to implement are, like, extremely obvious.

明白。

Mhmm.

对吧？

Right?

比如苹果就等了很久才推出这些功能，尽管当时大家都快求着他们做了。

Like, Apple waited a long time to actually implement these things even though everyone was, like, basically begging.

我不确定人们是不是真在迫切期待Instagram上出现AI功能，也许将来会吧，那他们就可以先等等，反正现在也不太确定该怎么实现这些功能。

I I don't know if people are, like, begging for AI features in Instagram, and maybe that'll happen in the future, and they can just basically just wait while, like, oh, we're not really sure how to implement this stuff.

是的。

Yeah.

你可以在搜索栏里输入内容，或者别的地方，但实际体验提升并不大。

You can put in the search bar or whatever, but it doesn't improve the experience that much.

不过也许一两年后，会出现某个功能，让所有人都觉得Instagram非得有不可。

But maybe, you know, in a year or two, there's gonna be some feature that, like, wow, everyone really wants this in Instagram.

到那时候，他们肯定已经具备足够的计算能力了。

By then, you know, they'll have all the compute necessary.

他们可以在正式上线前九十天，才去准备服务器之类的资源。

They can you know, ninety days before they implement it, they they can, you know, get out of the the lease or whatever.

是的。

Yeah.

对。

Yeah.

没错。

Yeah.

这挺有意思的。

It's interesting.

我觉得他们需要多做些产品实验。

I feel like they just need to do more product experimentation.

比如，Instagram的核心团队，他们推出过那个叫什么来着？

Like, the core the core Instagram team, they launched the the what's it called?

Glimpses，还是什么来着？

Glimpses or something?

就是比动态消息更短的一种形式，更轻量。

It's like a shorter even shorter version of Stories, like, lower.

它就放在聊天界面的侧边栏里。

It it, like, sits in, like, this little sidebar of the chat.

他们确实在整个应用家族里推各种新功能，但感觉就像有个研究团队，却没有一个专门快速把AI功能产品化的团队。

Like, they are launching different features across the family of apps, but that culture like, it feels like they have a research organization, but they don't have an AI product organization that's actively productizing things as quickly.

比如Meta Vibes，虽然本质上就是MidJourney的白标版本，但至少上线得挺快。

I mean, Meta Vibes for you know, even though it was, like, this white label of mid journey was well, at least it was launched fast.

至少他们拿到了用户反馈。

At least they got the feedback.

我觉得他们从来就不是那种真正擅长产品创新的公司。

I don't think they've ever been, like, the the company that, like, really innovates on on product.

对吧？

Right?

没错。

Yeah.

总会冒出某个功能，让人眼前一亮。

There's gonna be some some feature that everyone's, wow.

这玩意儿用在AI上真是太棒了。

This is really great for AI.

然后他们就能

And then they can then

照搬过来。

Copy it.

再把这功能整合到所有应用里。

Just integrate it into all their apps.

去买吧。

Buy it.

到那时，他们完全有能力提供这项服务。

They have all the capacity to serve it by that point.

是的。

Yeah.

我不确定。

I don't know.

我不确定。

I don't know.

我不知道它会走向哪里。

I don't know where it goes.

阿米特在那里进行了投资，分享了他对Meta新闻的两个看法。

Amit is investing over there, shares his perspective, two perspectives on the Meta news.

他说，看空。

He says, bearish.

看空的观点是，如果Meta愿意通过新云业务出售多余的算力，那是不是说明我们其实并不缺算力？

The bearish take is if Meta has excess compute that they're willing to sell via a new cloud business, doesn't that mean we aren't compute constrained?

这对新兴云服务商来说难道不是个坏消息吗？

Isn't this really bad for neo clouds?

如果Meta自己就能卖算力，那为什么还要给Core Weave或Iron优惠价？

Why would Meta give a deal to Core Weave or Iron if they just sell the compute themselves?

而且，既然闲置算力能支撑新业务，他们就没必要买那么多设备了，资本支出应该会下降才对。

Furthermore, wouldn't they cut CapEx because idle compute as the basis for a new business means they don't need as much compute as they bought, which means CapEx should come down.

这样的话，整个半导体行业都会受打压。

That would be bearish for all semis.

看多的观点是，即使Meta用的是闲置算力，只要他们真要做云业务，为了跟GCP、AWS和Azure竞争，反而可能加大资本支出。

The bullish take is if Meta is building a cloud business, even if they are using idle compute, which means they aren't compute constrained, they might end up spending more on CapEx to compete with GCP, AWS and Azure.

如果他们发现，靠云服务赚钱比靠广告更划算，那他们不就得像谷歌、微软、亚马逊那样，砸钱把云平台彻底建起来吗？

Like, if they realize that selling computing services on top of a Meta cloud is better than just ads, then wouldn't they end up having to spend in the same way that Google, Microsoft, Amazon do in order to build out a full cloud business?

他们在快速部署数据中心方面确实很强，可能赶不上SpaceX和AWS那么快，但在帐篷里快速部署GPU这方面，绝对算得上顶尖水平。

They do have a lot of capabilities in terms of spinning up data centers quickly, maybe not quite as quickly as SpaceX and AWS, but they're certainly like near the frontier of that capability in terms of putting up GPUs in tents.

所以资本支出增加对半导体行业是好事。

So more CapEx would be good for semi.

大家怎么看？

So what do people think?

我们到底怎么看待这个问题？

Where do we land on this?

市场对Meta反应积极，对新兴云服务商则持负面态度，毕竟现在多了一个新对手。

The market is certainly reacting positively to Meta and negatively to the Neo Cloud because there's a new competitor in town.

无论如何，扎克伯格和萨提亚·纳德拉的团队都在拼命推进。

Anyway Zuckerberg and the Satya Nadella team are going pretty hard.

真是让人笑死。

Say LMAO.

扎克伯格终于要扛下这波压力了。

Zuck finally takes the l.

明白了。

Okay.

所以如果这只是一个元宇宙风格的支线任务，而Meta训练的超级智能模型只是这一轮的VR版本，就像Meta Quest和VR战略被搁置那样，这些模型也被冷处理了。

So if this is a if this is a metaverse style side quest and the and the superintelligence meta trained models are the VR of this cycle for Meta's attempt at a new business creation, and the meta trained models get sort of mothballed in the way that the Meta Quest and the VR strategy got mothballed.

那Meta在MSL上的‘Ray Dance’会是什么样子？

What is the meta Ray Dance of MSL?

我的意思是，最后还能剩下什么？

Like, what would what would remain?

因为他们一旦关停某个项目，有时候会留下像Ray-Ban这样的产品——生意不错，还在增长，挺酷的，而且有持续开发，但方向要窄得多。

Because when they wind down one of these projects, sometimes you get a meta Ray Bans, which is pretty good business and growing and and cool, and, like, there's development there, but it's a much more narrow focus.

我觉得可能是图像模型，但他们似乎……

And I would say image model, but they they seem

但难道不会只是这个Neo Cloud业务吗？

But wouldn't it just be this, like, Neo Cloud business?

也许吧。

Maybe.

对。

Yeah.

这个会留下来。

That's what's gonna stay on.

但没那么

That's not as

有吸引力，不如Ray Bans。

fun as Ray Bans, though.

我想要一个面向消费者的产品。

I want a consumer product.

这是一家消费类公司。

It's a consumer company.

对。

Yeah.

而且有意思的是，你可以说Meta完全可以打造一个出色的推理业务。

And it's interesting because you can you can make the case that Meta could build an amazing inference business.

对。

Yeah.

对吧？

Right?

他们已经为数百万家企业提供服务了。

They already serve millions of businesses Mhmm.

在全球范围内。

Globally.

是的。

Yeah.

他们完全可以有力地说明，比如我们帮助商家获取客户，还帮助他们交付产品和服务。

They could make a pretty compelling case for how, like, hey, we help people acquire customers and we help people deliver products and services over here.

对。

Yeah.

我们以前就讨论过这个。

We've talked about that before.

我不太认同，他们拥有所有移动游戏公司和DTC电商企业，这并不意味着用户自然就会来我们这儿买代币。

I don't know that I buy that that the fact that they have like, you know, every single mobile gaming company and d to c e commerce business on is actually flows over to, well, now get your tokens from us.

是的。

Yeah.

我只是觉得，人们脑子里总觉得，Meta 就是个面向消费者的公司。

I just I just I just think that, like, people have it in their head, like, okay, Meta is a consumer business.

是的。

Yeah.

他们拥有规模巨大的销售团队和客户管理能力，非常擅长接触客户。

They have incredibly massive sales teams, account management They know how to get in front of customers.

是的。

Yeah.

他们已经不再只是做一种普通商品生意了，对吧？

They've got to the point where, you know, again, they they're not operating a commodity business, right?

是的。

Yeah.

他们拥有自己的社交网络。

Having their, you know, social networks.

云计算更接近第一种情况。

Cloud is is closer to one.

对。

Mhmm.

不过，我挺想知道他们会怎么发展。

But anyways, I'm interested to see.

我觉得他们很快就得就这件事发布自己的官方说法，不然总这么流传着一些未经证实的消息，让他们陷入被动。

I'm I'm I imagine they'll have to come out with their own kind of news around this pretty quickly so that they're not sitting in limbo with just one kind of rumored article floating out there and

对。

Mhmm.

大家就会开始各种猜测。

And people are speculating.

Meta的Ray-Ban智能眼镜就是个好例子，像这种带AI功能的消费电子产品，比如Ring之类的，只要在语音或图像技术上做得特别扎实。

The Meta Ray Ban display is a good example of, like or Meta Ray Ban's good example of, like, there's there's probably a piece of consumer hardware that's AI enabled, like the Ring or something where or just getting really good at voice models or just getting really good at image models.

如果他们能把范围收得更窄、更聚焦，完全有可能在这个领域占据绝对主导。

And if they're a little bit more narrow and constrained, they could probably completely dominate that.

但想同时搞超级智能和编程代理，方向有点分散。

But trying to do super intelligence and coding agents, and it's a little scattered potentially.

Jae Hoon说，我们目前的算力仍然严重不足。

Jae Hoon says, We are still massively short compute.

Meta和XAI正在出售算力，因为他们的模型根本没有实际推理需求。

Meta and XAI are selling compute because there's no inference demand for their models.

这本质上是个算力分配问题。

It's a compute allocation problem.

算力都集中在那些自己用不上的公司手里，而不是算力过剩。

Too much compute in the hands of players with no internal use for it, not a compute surplus problem.

再来点Meta的新闻。

More meta news.

什么新闻？

What is that?

据NPR的Bobby Allen说，

Apparently, according to Bobby Allen over at NPR,

哦，他们

Oh, They

考虑过收购Kalshi，是的。

had consider buying Kalshi Yeah.

之前他们还在开发自己的预测市场应用，这算是Meta的经典操作。

Before it's developing its own prediction market app that is sort of a classic meta

明白。

Mhmm.

操作手册。

Playbook.

这基本上推翻了你之前的理论，也就是他们可能只是在做那种靠名气驱动的东西。

This sort of puts to bed your theory that they might be just making a a, you know, a Clout based.

一种靠名气来预测的市场，大家比拼的是谁能……我看……

A clout based prediction markets where you compete for your ability to I see the

好吧，那个 Manifold 平台就是这样的。

Well, that's I'm still Manifold is that.