马克·安德森谈人工智能寒冬与智能体突破

毫无疑问，人们会喜欢，毫无疑问，已经有一些公司正在出现。

There's no question people are gonna like there's no question there are gonna be companies it's already happening.

有些公司认为自己正在基于模型构建价值，却会被下一个模型彻底淘汰。

There are companies that think that they're building value on top of the models, and they're just gonna get blissed by the by the next model.

这种情况毫无疑问正在发生。

There's no question that's happening.

但我认为，毫无疑问的是，任何技术融入人类复杂现实世界的过程都注定是混乱而复杂的。

But I think there's no question also that just the process adaptation of any technology into the real messy world of humanity is just going to be messy and complicated.

这不会是简单直接的。

It's not going to be simple and straightforward.

这将会是混乱而复杂的。

It's going be messy and complicated.

将会涌现出大量公司和大量产品，事实上，还会诞生整个行业，专门帮助这些技术真正触达普通人。

And there are going to be a lot of companies and a lot of products, products, and in fact, entire industries that are going to get built to basically actually help all of this technology actually reach real people.

投入这些公司的资金规模有多大？我的意思是，Dario 在 DoorDash 播客中提到过，DoorDash 就曾说：你为什么不直接多买十倍的 GPU 呢？

The amount of capital going into these companies I mean, Dario talked about it on the DoorDash podcast, DoorDash was like, why don't you just buy 10x more GPUs?

他说，如果模型没有达到确切的性能水平，我就要破产了。

And he's like, because I'm going to go bankrupt if the model doesn't exactly hit the the performance level.

作为OpenAI、ThinkOmachines和WorldApps的投资人，你怎么看待这种风险呢？

How do you think about that also as a risk on you know, you guys are investors in OpenAI and ThinkOmachines and WorldApps.

我们似乎正在以相当高的速度利用规模定律。

It seems like we're leveraging the scaling laws at a pretty high rate.

你觉得，对于可能出现的下行情况，你有多大的把握？

Like, how comfortable, I guess, do you feel with the downside scenario?

比如说，如果增长放缓了，你觉得你能重新调整这些基础设施和资本投入吗？

Like, and say like things peter out, you think you can kind of like restructure these build outs and, you know, capital investments?

是的。

Yeah.

我应该先说一下，我经历过互联网泡沫破裂，我可以花几个小时跟你讲关于互联网泡沫破裂的故事，那真是糟糕透顶。

So I should start by saying, so I lived through the .com crash, and I can tell you stories for hours about the .com crash, and it was horrible.

不。

No.

那太糟糕了。

It was awful.

那简直是末日般的。

It was was it was apocalyptic.

顺便说一下，当时很多所谓的互联网泡沫破裂，实际上是一场电信崩溃。

By the way, the a lot of the .com crash was actually at the time, it was actually a telecom crash.

这是一场带宽危机。

It was a bandwidth crash.

真正导致资金全军覆没的是电信公司。

The the thing that actually crashed that wiped out all the money was the the telecom companies.

Global Crossing？

Global crossing?

Global Crossing，是的。

Global global yeah.

我来自新加坡，他们在那里铺设了大量海底电缆。

So I'm from Singapore, and they they laid so much cable over over our oceans.

哦，实际上，在互联网泡沫时期存在一条规模定律，美国商务部在1996年发布了一份报告，称互联网流量每季度翻一番。

Oh, actually, there was a scaling law in the .com era, and it was literally the The US Commerce Department put out a report in 1996, and they said Internet traffic was doubling every quarter.

而且在1995年和1996年，互联网流量确实每季度翻了一番。

And and, actually, in 1995 and 1996, Internet traffic actually did double every quarter.

于是这就成了所谓的规模定律。

And so that became the scaling laws.

因此，这些电信创业者们纷纷筹资建设光纤网络，预期带宽需求会继续每季度翻一番。

So what all these telecom entrepreneurs did was they went out and they raised money to build fiber, anticipating that the demand for bandwidth was gonna keep doubling every quarter.

每季度翻一番，就像国际象棋盘上的麦粒一样。

Doubling every quarter, though, is like, know, grains of chess on the chessboard.

到了某个时候，数字会变得极其庞大。

Like, at some point, the numbers become extremely large.

对吧？

Right?

事实上，互联网自诞生以来一直在持续增长。

And and and it really and really what happened was the Internet the Internet, by the way, continuously kept growing basically since inception.

它一直在持续增长。

It is you know, it's it's continuously grown.

它从未萎缩过。

It's never shrunk.

而且它的增长速度比人类历史上任何其他事物都要快，但到了1998年、1999年，它已经不再每季度翻一番了。

And it's grown really fast compared to anything else, you know, in in in human history, but it wasn't doubling every quarter as of 1998, 1999.

因此，他们对所谓增长规律的预期与现实之间出现了巨大差距。

And so there was this gap in the expectation of what they thought was a scaling law versus reality.

而这正是导致互联网泡沫破裂的原因——他们过度投资了像Global Crossing这样的公司，过度建设了光纤网络，顺便说一句，还有光纤电信设备、所有网络硬件，以及实际的数据中心。

And that's actually what caused the dot com crash, which was the they they way over companies like Global Crossing way overbuilt fiber, which is sort of the and by the way, fiber telecom equipment, you know, so all the all the networking gear, you know, and then and then, by the way, the actual physical data centers.

所以，这标志着数据中心建设的开始，随后是数据中心的过度建设。

So, like, that was the beginning of the of the of the data center build and then and then data center overbuild.

因此，你看到了这种情况，但确切地说，大约有2万亿美元的市值蒸发了。

And so you had that, but it was it was literally I think it was, like, $2,000,000,000,000 got wiped out.

对吧？

Right?

这简直像耶稣一样。

It was like Jesus.

这就像一场大戏，顺便说一下，其中另一个微妙之处是，互联网公司本身几乎从不负债，因为科技公司通常不依赖债务运营。

It was like a big it was and by the way, the other the other subtlety in it was the Internet companies themselves never really had any debt because tech tech companies generally don't run on debt.

但电信公司是靠债务运营的。

But the telecom companies run on debt.

物理基础设施公司是靠债务运营的。

Physical infrastructure companies run on debt.

因此，像Global Crossing这样的公司不仅筹集了大量股权资金，还筹集了大量债务。

And so the companies like Global Crossing not just raised a lot of equity, they also raised a lot of debt.

所以它们的杠杆率非常高。

So they're highly levered.

然后你就这么做了。

And so then you just do the thing.

这就像是，好吧。

It's just like, okay.

你有一个高杠杆的项目，正在过度建设产能。

You have a highly levered thing where you're you're just you're overbuilding capacity.

需求在增长，但没有你预期的那么快，然后突然就破产了。

Demand is growing, but not as fast as you hoped, and then boom, bankrupt.

对吧？

Right?

就像人们说的酒店业那样，总是酒店的第三个主人能赚钱。

And and then it's and then it's like they say about the hotel industry, which is it's always the third owner of a hotel that makes money.

对吧？

Right?

它必须破产两次。

It has to go bankrupt twice.

对吧？

Right?

你必须把所有过度乐观的狂热都洗掉，才能达到一个稳定的状态，然后才能赚钱。

You have to wash out all of the overoptimistic exuberance before it gets to actually a stable state, and then it makes money.

顺便说一下，所有这些数据中心和所有的光纤，现在都已经被使用了，但那是二十五年之后的事了。

So by the way, all of those data centers and all of those all the fiber that they're in use, it's all in use today, but twenty five years later.

但实际上，所花的时间是十五年。

But it it took and actually, the elapsed time was it took fifteen years.

从2000年到2015年，花了十五年才真正填满所有容量。

It took fifteen years from 2000 to 2015 to actually fill up all the capacity.

警示在于，过度建设是可能发生的。

The cautionary warning is the the overbuild can happen.

而且你知道，当你陷入这种情况时，基本上所有拥有机构资本的人都会说：天啊。

And and and and, you know, you you get into this thing where basically everybody everybody who basically has any sort of institutional capital is like, wow.

我不太知道该怎么投资这些疯狂的软件项目。

It's just I I don't know how to invest in these crazy software things.

但我肯定可以建数据中心，肯定可以买GPU，可以部署计算网格和这些东西。

But for sure, I can put build data centers, and for sure, I can buy GPUs, and I can deploy, you know, compute grids and and all these things.

所以，如果你是个悲观主义者，你可以看看这种情况，然后说：天啊。

And and so, you know, if you're a pessimist, you can look at this and you can say, wow.

这简直就是在复刻我们在2000年经历过的那种情况。

This is, like, really set up to be able to basically replicate, you know, what we went through what we went through in 2000.

显然，这会是坏事。

Obviously, that would be bad.

相反的观点，也就是我认同的观点，主要有几点。

The counterargument, which is the one I I agree with, which is the counter on the other side is a couple things.

第一，投资的这些公司都是最顶尖、最可靠的公司。

One is the companies that are investing all the the companies that are investing the money are like the bluest chip of companies.

而当年像Global Crossing这样的公司，更像是一个创业者。

And so back back in the in the doc like, Global Crossing was like an it was like an entrepreneur.

它是一个全新的风险项目。

It was like a a new venture.

但现在大规模投入的资金，来自的是微软、亚马逊和谷歌这样的公司。

But, like, the money that's being deployed now at scale is Microsoft and, you know, and Amazon and Google.

还有Facebook。

Facebook.

还有Facebook、英伟达，以及现在这些公司，比如OpenAI和Anthropic，它们现在已经发展成规模非常大的企业，拥有可观的收入。

And Facebook and Nvidia and, you know, these these these and and now, you know, by the way, OpenAI and Anthropic, which are now at, like, you know, really serious size, you know, as companies with, you know, very serious revenue.

这些是非常大规模的公司，拥有大量现金，而且它们从未使用过的债务额度也很多。

These are very large scale companies with, like, lots lots of cash, lots of debt capacity that they've they've never used.

因此，如今的这种投资方式在性质上与当年完全不同。

And so this is institutional in a way that that really wasn't at the time.

其次，至少目前来看，每一笔投入用于运行GPU的资金，都能立即转化为收入。

And then the other is, at least for now, every dollar that's being put into anything that results in a running GPU is being turned into revenue right away.

就像这样，你们都清楚这一点。

Like so and you guys know this.

每个人都迫切需要算力资源。

Like, everybody starved for capacity.

每个人都急需计算能力，以及相关的内存、互联设备和其他一切，比如数据中心空间。

Everybody starved for compute capacity and then, you know, all the associated things, memory and and and interconnect and everything else, data center space.

因此，现在每一笔投入都在立即转化为收入。

And so every dollar right now that's being put in the ground is turning into revenue.

事实上，我认为正在发生一件有趣的事：由于所有人都严重缺乏算力，我们如今能够使用的模型，实际上是受到供应限制影响后的劣化版本。

And in fact, I actually think there's an interesting thing happening, which is because everybody's starved for capacity, the models that we actually have that we can use today are inferior versions of what we would have if not for the supply constraints.

确实如此。

That's true.

没错。

If if right.

假设一个假设的世界，在这个世界里，GPU便宜十倍、丰富十倍，那么模型会好得多，因为你能投入更多资金用于训练，构建出更优秀的模型。

Suppose a hypothetical universe in which GPUs were 10 times cheaper and 10 times more plentiful, the models would be much better because you would just allocate a lot more money to training, and you'd just build better models, and they would be better.

所以我们现在得到的，其实是被压缩过的版本的

And so we're we're actually getting the sandbag version of

技术。

the technology.

不。

No.

我们使用的所有东西都被量化了，因为实验室必须保留完整的版本。

Everything we use is quantized because the the labs have to keep the the full versions.

对。

Right.

我们连好东西都得不到。

Like We're not even getting the good stuff.

是的。

Yeah.

但即使技术进步停滞，一旦GPU制造产能和内存等所有未来五到十年必须提升的基础设施到位，即使现在的技术也会变得好很多。

But but getting the good stuff, it's it's just even if technical progress stops, once there's, a much bigger build of, like, GPU manufacturing capacity and memory, you know, all all the things that have to happen in the course of the next five or ten years, once it happens, even the current technology is gonna get gonna get much better.

而且你知道，这种技术的用途多得数不清。

And then as you know, like, there's just like a million ways to use this stuff.

这种技术有上百万种应用场景。

Like, there's just like a million use cases for this.

这可不是简单地在网上传输数据包，然后指望人们自己找到用途。

Like, it you know, this isn't just sending packets across a a thing whatever and hoping people find something to do with it.

这是要把智能应用到人类活动的每一个领域，然后它会表现得极其出色。

This is just like, oh, we apply intelligence into every domain of human activity, and then it works, like incredibly well.

我知道的是这些。

Here's what I know.

我知道的是这些。

Here's what I know.

在未来三到四年左右，基本上所有产品都会售罄。

In the next three or four year it's like somewhere between three or four years out, basically everything is selling out.

所以整个供应链都在售罄或即将售罄。

So like the entire supply chain is sold out or selling out.

因此，我们注定在未来多年面临持续的供应短缺。

And so there's no like, we're just going have chronic supply shortage for years to come.

市场将做出回应，这种回应正在发生——大量投资将涌入新建晶圆厂产能及其他相关领域，以应对这一局面。

There's going to be a response from the market that's going to result in an enormous it's happening now, an enormous flood of investment in a new fab capacity and everything else to be able to do that.

在某个时候，供应链的瓶颈至少会在一定程度上得到缓解。

At some point, the supply chain constraints will unlock, at least to some degree.

当这种情况发生时，这将成为行业增长的另一个推动力，因为产品会变得更好，所有东西都会变得更便宜。

That will be another accelerant to industry growth when that happens because the products will get better and everything will get cheaper.

所以我知道这一定会发生。

And so I know that's going to happen.

我知道这些部署和实际应用场景都非常吸引人。

I know that the deployments, the actual use cases are, like, really compelling.

就像我之前说的，随着推理和智能体等技术的发展，我知道它们会变得越来越好。

And then, like I said, you know, with reasoning and agents and so forth, like, I know they're just gonna get, like, much, much better from here.

所以我知道这些能力真的非常强大。

And so I I know the capabilities are, like, really, really serious.

我也知道技术进步不会停止。

I also know that the technical progress is not going to stop.

它正在加速发展。

It it is accelerating.

这些突破非常巨大。

Like, the breakthroughs are are tremendous.

我的意思是，甚至只是月复一月，这些突破都极其显著。

I mean, even just month over month, the breakthroughs are really dramatic.

所以，你知道，如果你是个愤世嫉俗的人——确实也有人这样——你可能会从2000年看到一些相似的影子，但我根本无法想象，未来几年内这种趋势会让人失望。

And so, you know, I think if you were a cynic and there there are cynics, you can look at 2,000, you can find echos, but I can't even imagine betting on that this is gonna, like, somehow disappoint in, you know, at least for years to come.

我认为下这样的赌注几乎是自杀行为。

I think it would be essentially suicidal to make that bet.

那是迈克尔·伯里吗？

Is that Michael Burry?

哦，那是个有趣的人。

Oh, that's an interesting guy.

是的。

Yeah.

我们挑个目标吧。

We'll pick on a guy.

我们挑一个具体的人来讨论吧。

We'll pick let's pick on one guy.

毕竟，他确实这么做了。

Well, because he did.

他公开了，他不在乎。

He he came out with it was it was He doesn't mind.

他是做英伟达空头的。

He was the NVIDIA Short.

对吧？

Right?

他公开了做英伟达空头的事。

He came out with the NVIDIA Short.

然后你们可能讨论过这一点：现在的模型在软件层面提升得如此之快，以至于如果你今天还在用三年前的英伟达推理芯片，你现在的收益反而比三年前还高，因为软件的进步速度超过了芯片的折旧周期。

And then you guys probably talked about this, which is the the analysis now that, like, the current models are getting better faster at such a rate that if you are running an NVIDIA if you're running an NVIDIA inference chip today that's three years old, you're making more money on it today than you did three years ago because the pace of improvement of the software is faster than the depreciation cycle of chip.

据我了解，谷歌可能在使用非常老旧的TPU——我不确定这是传闻还是公开信息，但我认为谷歌用这些老芯片赚得盆满钵满。

Then my understanding is Google is running don't know exactly what these are rumors that I've heard or maybe it's public, but I think Google's running very old TPUs, very profitable.

而且利润丰厚。

And profitably.

所以，据我所知，实际情况恰恰与伯里观点相反。

And so it actually turns out, as far as I can tell, it's actually the opposite of the Bury thesis.

他实际上完全错了。

He was actually 180 degrees wrong.

旧的NVIDIA芯片正在变得更有价值，这在历史上是前所未有的。

The old NVIDIA chips are getting more valuable, which is something that's literally never happened before.

从来没有任何时候，旧型号的芯片会变得更有价值，而不是更不值钱。

Like, it's never been the case that you have an older model chip that becomes more valuable, not less valuable.

而这再次体现了软件进步的惊人速度，以及能力回报的惊人速度，是的。

And and again, that's an expression of that just ferocious pace of software progress, ferocious pace of capability payoff Yeah.

你正在从这一面获得回报。

That you're getting on the other side of this.

所以，我就是觉得，押注反对这一点，是的。

And so I just the idea of betting against that, like Yeah.

是的。

Yeah.

嗯，其中一个

Well, one of It's

就像是邀请别人把你脸撕掉。

like an invitation to get your face ripped off.

我其中一个

One of my

我早期的一个成功案例是，模拟了H1月和H2月的使用寿命，通常建议是四到七年，但你可能现实地认为可以缩短到两到三年，但实际上使用年限反而在延长，而不是缩短。

early hits was, like, modeling the lifespan of the h January and h February and going like, you know, usually they advise like four to seven years and it was, you know, maybe you sort of realistically care cut it down to two to three, but actually it's going up and not down.

而且，我想，这正是我们的梦想。

And and, that's I mean, that's I think that's the dream.

我们正在提高利用率，我认为利用率能解决所有问题。

We are finding utilization and I think utilization solves all problems.

比如，你甚至能为端口找到用途，就连内存我们也都短缺。

Like you can find use cases for even the port like even memory, we're having a shortage.

就连我们现有的更差版本的内存，我们也在为它们找到用途。

And even like the shittier versions of memory that we do have, we are finding use cases for it.

所以这很棒。

So that's great.

在面临三年供应短缺的世界里，开源AI和边缘推理有多重要？

How important is open source AI and kind like edge inference in a world in which you have three years of supply crunch?

比如，如果你展望五年后，你会如何看待数据中心推理与边缘推理的区别？

Like do you think in the Like if you fast forward like five years, like how do you think about inference in the data center versus at the edge?

好吧，先从这一点说起，是的。

Well, so just to start yeah.

我认为开源在很多方面都非常重要。

So I think I think open source is very important for a bunch of reasons.

我认为边缘推理在很多方面也非常重要。

I think edge edge inference is very important for a bunch of reasons.

从实际角度来看，如果我们未来三年真的面临根本性的供应短缺——你们也知道，只要把未来三年的需求与供应进行对比，就能得出一个主要预测：核心区域的推理成本在未来三年会如何变化。

I I think just practically speaking, if we're just gonna have fundamental supply crunches for the next I mean, you guys know, if you just project forward demand over the next three years relative to supply, one of the main predictions you can do is what's gonna what what's gonna happen to the cost of of inference in the core, over the next three years.

而且，它可能会大幅上涨。

And, like, it may rise dramatically.

对吧？

Right?

就像，你知道，目前大型模型公司正在大力补贴。

Like so so what and then as as you know, like, the the the big model companies are subsidizing heavily right now.

对。

Right.

对。

Right.

那么，三年后，普通用户每天、每月使用AI服务所需的token成本会是多少呢？

And so so what's the what will be the average person's, you know, per day per month token cost, you know, three years from now to do all the things that they wanna do?

我不知道。

And I I don't know.

我觉得，你们身边肯定也有朋友。

It's gonna be I mean, I have you guys probably have friends.

我今天就有朋友，每天为运行OpenClaw支付一千美元的Claw token费用。

I have friends today who are paying a thousand dollars a day for OpenClaw for claw tokens to run OpenClaw.

对？

Right?

所以，好吧，每月三万美元。

And so, okay, $30,000 a month.

对吧？

Right?

顺便说一下，那些朋友还有上千个想法，想让他们的Claw去做更多事情。

And and by the way, those those friends have, like, a thousand more ideas of the things that they want their claw to do.

对吧？

Right?

因此，你可以想象，对于一个完全部署的个人代理，潜在的需求可能高达每天五万到十万美金的令牌成本。

And so you could imagine there there's, like, a latent demand of up to, I don't know, 5 or $10,000 a day of of tokens for a fully deployed, you know, personal agent.

显然，消费者无法承担这笔费用。

And obviously, consumers can't pay that.

对吧？

Right?

所以，这也让你对未来的潜在需求规模有了一个概念。

And so and so but it gives you a sense of the of the of the future scope of demand.

对吧？

Right?

所以，即使价格性能提升十倍，那也还是达到每天100美元，仍然远远超出人们的支付能力。

And so so even even if there's a 10 x improvement in price performance, that still, you know, goes to a $100 a day, which is still way beyond what people can pay.

所以，需求将会非常猛烈。

So there's just gonna be, like, ferocious demand.

顺便说一下，关于智能代理，另一个有趣的是，到目前为止，很多限制都是GPU限制。

By the way, the agent thing the other interesting thing is I think the agent thing so up until now, a lot of the constraints have been GPU constraints.

我认为现在智能代理也带来了CPU方面的限制。

I think the agent thing now also translates into CPU constraints.

对吧？

Right?

CPU和内存。

CPU and memory.

是的。

Yes.

CPU和内存。

CPU and memory.

对吧？

Right?

所以，整个芯片生态系统都将

And so, like, the entire chip ecosystem is just gonna

摆脱网络瓶颈。

get away from network constraints.

那个那个

That that

那才是关键。

would be the killer.

这可能会在多年内成为瓶颈。

That's all bottlenecked potentially for years.

所以我觉得，布拉德，我认为这实际上是可能的。

And so so I I think that, Brad and and I think it's actually possible.

我的意思是，一般来说，推理成本还会继续下降，但我想这么说吧。

I mean, generally, inference costs are gonna keep coming down, but I think the let's put it this way.

我认为，由于这些供应限制，成本下降的速度可能会在此处暂时放缓。

The rate of decline, I think, may level out here for a bit because of these supply constraints.

然后在某个时候，也许实验室不再提供如此多的补贴，这又会成为一个问题。

And then at some point, maybe the lab stops subsidizing so much, that again will be an issue.

因此，对推理的需求将远远超过集中式模型所能满足的范围。

And so there's just gonna be so much more demand for inference than can be satisfied with the centralized model.

然后你们也知道，苹果芯片在实现推理方面所取得的惊人创新确实非常了不起。

And then you guys know this, but like all the just the dramatic innovations that have happened in the Apple silicon to be able to do, inferences is quite amazing.

开源社区正在投入巨大的努力，去实现一种反复出现的模式：大型模型一开始根本跑不了，但六个月后，哦，它居然能在PC上运行了。

A level of effort being put, like, open source guys are putting incredible effort into getting you know, this recurring pattern where the big model will never run on a PC, and then six months later, oh, it runs on a PC.

对吧？

Right?

这太惊人了。

It's amazing.

而且有很多聪明人在研究这个问题。

And there's very smart people working on that.

所以这一切都存在。

So there's all that.

此外，还有其他的驱动因素。

And then look, there's also other motivators.

还有其他驱动因素，比如：大型中心化模型提供商究竟有多值得信赖？

There's other motivators, which is just like, okay, how much trust are the big centralized model providers?

它们在市场中建立了多少信任，而相比之下，在某些情况下，有些人对于某些用途，却并不愿意把一切完全交出去。

How much trust are they building in the market versus how much are at least in certain cases with some people, for certain use cases, people being like, well, I'm not willing to just turn everything over.

所以，还有所有这些信任问题。

So there's all the trust issues.

顺便说一下，还有纯粹的价格优化问题。

By the way, there's also just straight up price optimization.

有很多AI应用场景，并不需要云端的爱因斯坦。

There's many uses of AI where you don't need Einstein in the cloud.

你只需要一个智能的本地模型。

You just need a smart local model.

还有性能问题，比如你希望门把手能内置一个AI模型，以便能够进行门禁控制。

There's also performance issues where you're going to want your door knob to have an AI model you know, to be able to, you know, do, you know, to be able to do access control.

显然，所有带芯片的设备都会内置AI模型，其中很多将是本地运行的。

Obviously, like, everything with a chip is gonna have an AI model in it, and a lot of those are gonna be local.

所以是的。

And so yeah.

不。

No.

我觉得你会拥有，顺便说一下，还有可穿戴设备，你知道，你不想进行完整的往返传输。

Like, I think I think you're gonna have and then you're gonna by the way, also wearable devices, you know, you you don't wanna do a complete round trip.

你希望，不管你的智能设备是什么，都具备极低的延迟。

You want you know, you you whatever your smart devices are, you want it to be, like, super low latency.

对。

Yeah.

问题是，我们关心是谁做的吗？

The question, do we care who makes it?

本周最大的新闻之一是AI2（艾伦研究所）的倒闭，这是美国一家真正的开源模型实验室。

One of the biggest news this week was the collapse of AI two, the Allen Institute, one of the actual American open source model labs.

是的。

Yep.

我对美国的开源项目没那么乐观。

And I'm not that optimistic on American open source.

是的。

Yeah.

你们投资了Miss Charles，而Miss Charles在中国以外的地方表现得非常好。

Like, you you guys invested in miss Charles, and miss Charles is doing extremely well outside of China.

就这些了。

That's about it.

是的。

Yeah.

我们走着瞧。

We'll see.

我们走着瞧。

We'll see.

听好了。

Look.

首先，我认为我们确实关心是谁制造的。

Number one, do think we care I do think we I do think we care who makes it.

我想说这一点。

I would say this.

前一届总统政府想在美国扼杀它。

The the the previous presidential administration wanted to kill it in The US.

就像想把它淹死在浴缸里，所以他们想除掉它。

Like, wanted to drown in the bathtub, and so they wanted to kill it.

所以至少我们现在有一个真正希望它成功的政府。

So at least we have a government now that actually, like, actually wants it wants it to happen.

你也在委员会里吗？

And you're in the council?

我已经问过了。

And I've asked.

是的。

Yeah.

所以，这个政府，不管人们对其它政治问题有什么看法——确实有很多——但我觉得这个政府对人工智能，特别是开源人工智能，有着非常开明的看法。

So the the you know, this admin for whatever other political issues people have, which are many, you know, this administration has, I think, a very enlightened view, and in particular, an enlightened view on AI and in particular on open source AI.

因此，他们非常支持。

And so they're very supportive.

我的理解是，中国各大公司有着非常明确的理由推动开源，因为它们从根本上认为，目前无法在中国以外，特别是无法在美国市场销售商业人工智能产品，原因多种多样。

My read is the the Chinese have a very the various Chinese companies have a very specific reason to do open source, which is they they they don't fundamentally, they don't think they can sell commercial AI outside of China right now, at least specifically not not in The US for a combination of reasons.

所以，它们似乎把开源人工智能视为一种亏损导向策略，用以对抗国内的付费服务以及一些附属产品。

And so they they kind of view, I think, open source AI as a bit of a loss leader against basically domestic, you know, paid paid services and then kind of, you know, kind of ancillary products.

你知道，它们对此非常热衷。

You know, they're but they're very excited about it.

顺便说一下，我觉得这很棒。

By the way, I think it's great.

我觉得他们这么做真是太好了。

I think it's great that they're doing it.

你知道，我觉得深度求索就像是送给世界的礼物。

You know, I think DeepSeek was like a gift to the world.

我认为开源的影响体现在两个方面。

I think the great thing about open source open source the the the impact of open source is felt two ways.

一方面，你可以免费获得软件，另一方面，你可以学习它是如何工作的。

One is you you get the software for free, but the other is you get to learn how it works.

对吧？

Right?

所以，比如那篇论文。

And so, like The paper.

那篇论文。

The paper.

论文和代码。

The paper and and the code.

对吧？

Right?

还有代码。

And the code.

所以，比如说，我觉得这太棒了。

And so, like, for example, thought I this was amazing.

所以OpenAI发布了一个，这是一项了不起的技术突破，简直太棒了。

So OpenAI comes out with a one, and it's an amazing technical breakthrough, and it's just, like, absolutely fantastic.

但当然，他们没有详细解释它是如何工作的。

But, of course, they don't explain how it works in detail.

而且，当然，他们隐藏了推理过程。

And then, of course, they hide the they hide the reasoning traces.

对吧？

Right?

然后大家就都说，好吧。

And and then and then and then everybody's like, okay.

这太棒了。

This is great.

但谁能复现这个呢？

But, like, who's gonna be able to replicate this?

其他人能做得到吗？

Are other people gonna be able to do this?

你知道的。

You know?

里面有什么独家秘诀吗？

Is there a secret sauce in there?

然后R1发布了，代码和论文都公开了。

And then r one comes out, and it's just like, there's the code and there's the paper.

现在全世界都知道怎么做了。

And now the whole world knows how to do it.

然后，你知道吗，三个月后，其他所有AI模型都开始加入推理功能。

And then, you know, three months later, every other AI model is is adding reasoning.

所以你会看到这种双重效应：即使中国自己的模型没有被广泛使用，但由此带来的全球教育和信息传播却极其强大。

And so so you get this kind of double like, even if the Chinese models themselves are not the models that get used, the education that's taking place to the rest of the world, the information diffusion, you know, is incredibly powerful.

所以这种情况确实发生了。

So that happens.

然后我也说不准。

And then I don't know.

我们走着瞧吧。

We'll we'll see.

你知道，有一大批美国的开源AI模型公司。

You know, there are a bunch of American, you know, open source, you know, AI model companies.

我的意思是，你看。

I mean, look.

肯定会非常巨大，实际上已经如此了。

There's gonna be tremendous you know, there already is.

当然，主要的模型公司之间将会存在巨大的竞争。

There's, you know, there's gonna be there's tremendous competition among the primary model companies.

你知道，根据不同的统计方式，现在有四五家大型模型公司正在以不同方式齐头并进。

You know, there's depending on how you count, there's, like, four or five, you know, big co model companies now that are of neck and neck in different ways.

显然，Axon、MetaWribe 和 Evolve 都在全力推进超越对手的计划。

Then obviously both Axon and MetaWribe and Evolve, both have huge attempts to of leapfrog underway.

此外，还有一大群初创公司，包括许多我们之前提到的公司，正在尝试不同的技术路径。

And then you've got, you know, a whole fleet of start ups, new companies, including a whole bunch that we're back in that are, you know, trying to come out with different approaches.

然后还有其他各种情况。

And then you've got whatever it is.

我不确定。

I don't know.

目前中国有多少家主流的基础模型公司？

How how many how many, like, mainline foundation model companies are there in China at this point?

大概是六家。

It's probably six

是五只老虎

It's five tigers

他们就这么叫的。

is what they call it.

因为权利问题，Quinn的情况存疑。

Quinn is in questionable because of Right.

领导层发生了变化。

There's change in leadership.

对。

Right.

但确实如此。

But yeah.

但那不包括像月之暗面这样的公司。

But that does not include that that includes, like, Moonshot.

是的。

Yes.

Kimi、DeepSeek、ZAI、Quinn，其中有一个在里面。

Kimi, DeepSeek, ZAI, Quinn o one is in there.

对。

Right.

然后，字节跳动，还有你。

And then, and ByteDance and and then you've

看到字节跳动会是下一个梯队。

seen ByteDance would be, like, the next tier.

字节跳动没那么突出。

ByteDance weren't as prominent.

他们之前没有，但现在你知道的，

They weren't have a But now, you know,

是的。

it's yeah.

但它们，你知道，字节跳动非常令人鼓舞，而且 presumably 会有更多产品推出，腾讯大概也会有更多东西发布，等等。

But they're you know, CDance is very inspiring, and presumably, have more stuff coming and Tencent probably has more stuff coming and and so forth.

所以，你看。

And so so so, like, look.

这里有一件你可以预见的事：目前并不存在这些市场，它们不会只是在美国和中国之间展开；现在已经有大约十几家主要的金融模型公司，达到了某种规模或临界点。

Here here would be a thing you can anticipate, which is there are not these markets, they're not going to be between The US and China right now, there's, a dozen primary financial model companies that are, like, at scale at at some level of, like, critical mass.

三年后，不会还有十几家。

It's not gonna be a dozen in three years.

对吧？

Right?

因为这些行业根本容纳不了十几家。

Like, just because these industries don't bear a dozen.

最终可能只剩下三到四家大赢家，或者一两家大赢家。

It's it's gonna be three you know, there's gonna be three or four big winners or maybe one or two big winners.

所以，会有大量这样的公司不得不寻找替代策略。

And so there's gonna be, like, a whole bunch of those guys that are gonna have to figure out alternate strategies.

我认为，开源就是其中一种策略。

And I think, like, open source is one of those strategies.

所以我认为你可能会看到，整个问题在于，谁会去做开源呢？

And so I I think you could see, like, a whole I think the question's, like, who's gonna do open source?

我认为这是一件非常动态的事情。

I I think that that that's a very dynamic thing.

我觉得很难预测会发生什么。

I think it's very hard to predict what happens.

我认为这非常重要。

And I think it's very important.

英伟达正在做很多工作。

NVIDIA is doing a lot.

你想要

You want

尝试一下吗？

to try?

我只是说，没错，就是这样。

Well, I was just saying, well, exactly.

然后你去到英伟达，你知道，在商业战略中有一个老概念叫‘使互补品商品化’。

Then you go to NVIDIA and then and then, you know, just to get an industrial there's there's an old thing in business strategy, which is called commoditize complement.

对。

Right.

所以如果你是黄仁勋，这其实很明显。

And so if you're Jensen, it's just kind of obvious.

当然，你希望把软件商品化。

Of course, you wanna commoditize the software.

是的。

Yeah.

而且值得称赞的是，他为此投入了巨大的资源。

And he's and to his enormous credit, he's putting enormous resources behind that.

所以也许这真的是英伟达在做，我觉得这会很棒。

And so maybe it maybe it's literally NVIDIA, and I think that would be great.

对。

Yeah.

是的。

Yeah.

早期欧洲项目出现了叙事上的偏差。

Narrative violation to European projects in the beginning.

该死。

Damn.

我很快要举办我的欧洲大会，我邀请了他们两位。

I'm hosting my Europe conference soon and I got both

都来了。

of them.

他们把我们给套住了。

They got us.

把我们给套住了，马克。

Got us, Marc.

把我们给套住了。

Got us.

等一下。

Wait a minute.

彼得当时在哪里？斯图伊伯格做那件事时在哪里？

Where was Peter so where was Stuyburger when he did a

他在奥地利。

He was in Austria.

是的。

Yeah.

对。

Yeah.

他在哪儿？

He was in what?

他在维也纳。

He in Vienna.

哦，他在维也纳。

Oh, he was in Vienna.

那他在哪儿？

Then where is he

他要搬去旧金山。

he's moving to SF.

好的。

Okay.

好的。

Okay.

好吧。

Alright.

好的。

Okay.

就这样了。

There we go.

然后，是的，那个Pi的人好吧。

And then, yeah, the Pi guy alright.

Pi团队是欧洲人。

The Pi guys are European.

是的。

Yeah.

他们在莫斯科是好朋友。

They're they're buddies in Moscow.

马里奥也在那里。

Mario is also there.

是的。

Yeah.

对。

Right.

而且他们确实。

And are they yeah.

他们还没有宣布任何变动，是吧？

They haven't announced yet any sort of change changed or have they?

不。

No.

他们有

They're they have

他们在那儿有一家公司。

a company there.

好的。

Okay.

明白了。

Got it.

好的。

Okay.

很好。

Good.

是的。

Yeah.

很好。

Good.

总之，我认为PI和OpenCloud

Anyways, I think PI and OpenCloud

是非常重要的软件技术，我只是想让你谈谈你的看法。

are very important software things, and and I just wanted you to just go off on what you think.

是的。

Yeah.

所以我认为，这两者的结合是最重要的十大软件之一。

So I think in the the combination of the two of them, I think, is one of the 10 most important software.

OpenCloud获得了所有关注。

OpenCloud got all the attention.

但没错。

But Right.

谈谈PY。

Talk about PY.

PY是一种理念。

PY is kind of the idea.

PY是一种架构上的突破。

PY is kind of the architectural breakthrough.

对于我们这些年纪较大的人来说，从大约1970年到我不确定的某个时间，软件界一直有一个非常重要的概念。

For those of us who are older, there was this whole thing that was very important in the world of software basically from, like, 1970 to I don't know.

它至今仍然非常重要，但从1973年到Linux诞生为止，我们曾经称之为Unix思维的东西。

It it still is very important, but, like, '19 from 1973 to, like, basically the creation of Linux, which is basically this this thing we used to call, like, the Unix mindset.

因为当时存在各种不同的理论。

Like, so so because there were all these different, know, theories.

有各种不同的操作系统、大型机，还有Windows、Mac等等。

There are all these different operating systems and mainframes and and then, you know, all these Windows and Mac and all these things.

而在这一切背后，其实一直存在着一种Unix思维的理念。

And then there was this kinda behind it all was this idea of kind of the Unix mindset.

Unix思维就是这么一种理念：在早期，让计算机行业真正运转起来的操作系统是IBM开发的OS/360，这是一个庞大的操作系统，旨在运行一切。

And the Unix mindset was this thing where basically you don't have these, like like, in the old days, like like, the operating system that, like, made the computer industry really work, like, in the nineteen sixties was this thing called OS three sixty, which was this big operating system IBM developed that was supposed to basically run everything.

这就像一个巨大的、悬浮在空中的单体架构。

And it was this, like, giant monolithic architecture in the sky.

它就像一座巨大的软件城堡。

It was like a giant castle of software.

它确实运行得非常好，他们在这一点上非常成功。

And by the it worked really well and they were very successful with it.

但它是一座高高在上的巨大城堡。

But it was this huge castle in the sky.

但这个系统几乎让人望而却步，你必须是IBM内部的人，或者与IBM关系非常密切，才能真正理解系统的每一个细节。

But it was this thing that was almost unapproachable, which is like you had to be kind of inside IBM or very close to IBM, and had to really understand every aspect of the system worked.

然后，来自AT&T、后来来自伯克利的Unix团队站出来，说：不。

And then the the Unix guys originally out of AT and T and then out of out of Berkeley, you know, came out and they said, no.

我们要采用一种完全不同的架构。

Let's have a completely different architecture.

架构的工作方式是：我们会有一个提示符和一个shell。

And the way architecture is gonna work is we're gonna have we're gonna have a a prompt and a and a shell.

然后，所有的功能都会以这些独立的模块形式存在，你可以将这些模块串联起来。

And then and then we're gonna all the functionality is gonna be in the form of these discrete modules, and then you're gonna be able to chain the modules together.

所以，操作系统本身几乎就像一种编程语言。

And so, like, the the the it's almost like the operating its operating system itself is gonna be a programming language.

这导致了 shell 的核心地位，进而催生了 Unix 工具的串联使用，随后出现了 Perl 这类脚本语言，让你可以非常轻松地实现这种操作。

And then that led led to the the the sort centrality of the shell, and then that led to sort of, you know, basically chaining the other Unix tools, and then that led to the emergence of these these scripting languages like Perl where you you could basically kind of very easily do this.

然后，shell 变得更加复杂。

And then the shells got more sophisticated.

接着，你看，第一点，这种方式是行得通的。

And then and then and then look, like, you know, that that that number one, that worked.

那就是我成长起来的世界。

And that that was the world I grew up in.

我是个 Unix 用户，从大约1988年开始，一直到我的整个职业生涯都是如此。

Like, I was I was a Unix guy, you know, sort of from, call it, 1988 to, you know, kind of all the way through my work.

而且它真的非常有效。

And it worked really well.

它在后台运行。

It it's in the background.

你知道，普通人并不需要了解它。

You know, normal people don't need to didn't need to necessarily know about it.

但如果你从事系统架构或应用开发，你对它了如指掌。

But, like, if you were doing, like, system architecture or application development, you you you knew all about it.

从那以后，它就一直躲在后台了。

And then, you know, it's been in the background ever since.

你看，你的Mac里仍然有Unix壳层，你的iPhone里也 somewhere 隐藏着一个Unix壳层。

And, you know, look, your Mac still has a Unix shell, you know, kind of in there, and your iPhone still has a Unix shell kind of buried in there somewhere.

所以它们都在那里，而Windows的壳层则是这种体系的一种奇怪衍生物。

So they're kind of in there, and then, you know, the Windows shell is kind of a, you know, sort of a weird derivative of that.

但你知道，互联网是基于Unix运行的，而智能手机，无论是iOS还是Android，都是Unix的衍生系统。

But, you know, but look, the Internet the Internet runs on Unix, and then smartphones actually, both iOS and Android are Unix derivatives.

所以，Unix最终还是胜出了。

And so, you know, kind of Unix did end up winning.

但不管怎样，我们后来就开始对此习以为常了。

But but anyway, and then we just started taking that for granted.

所以，总的来说，我认为Pi和OpenClaw发生的事情是，那些人发现的：我总是说，伟大的突破在事后看来都是显而易见的。

Then and then so so basically, the the way I think about what happened with Pi and then with OpenClaw is basically what those guys figured out is I always say the the great breakthroughs are obvious in retrospect.

对吧？

Right?

那也是最好的一种。

Which is The best kind too.

最好的一种。

The best kind.

在当时它们并不明显，否则早就有别人做到了。

They weren't obvious at the time or somebody else would have done them already.

所以这确实是一次真正的概念飞跃。

And so there is a real conceptual leap.

但当你回头去看时，就会觉得，哦，当然如此。

But then you look at it sort of the backwards looking and you're just like, oh, of course.

对我来说，这些永远是最好的突破。

To me, those are always the best breakthroughs.

所以实际上，语言模型本身就是这样。

So actually language models themselves are like that.

就是，哦，下一个词预测。

It's just like, oh, next token completion.

哦，当然了。

Oh, of course.

是的。

Yeah.

还有什么其他目标重要吗？

What other objective mattered?

对。

Yeah.

没错。

Exactly.

对。

Right.

但她甚至说，直到有人真正做出来之前，这并不明显。

But she's even saying it wasn't obvious until somebody actually did it.

对吧？

Right?

因此，这个概念上的突破是真实、深刻、有力且极其重要的。

And so the conceptual breakthrough is real and deep and powerful and and very important.

所以，我看待PI和OpenClaw的方式是，它本质上是将语言模型的思维与Unix壳层提示的思维结合起来。

And so the way I think about PI and OpenClaw is it's basically marrying the the language model mindset to the to the Unix basically shell prompt mindset.

那么，什么是智能体呢？

And so it's it's basically this idea that what what so what is an agent?

对吧？

Right?

正如你所知，多年来，聪明的人们一直在试图弄清楚什么是智能体，提出了许多架构来构建智能体。

And as as and as you know, like, smart people have been trying to figure out what an agent is for for for decades, and they've had many architectures to build agents in the whole thing.

结果，什么是智能体？

And it turns out what is an agent?

所以，我们现在知道，智能体就是以下这样。

So it turns out what we now know is an agent is the following.

它是一个语言模型，而在其上层，它是一个 bash shell。

It's it's a language model, And then above that, it's a bash it's a bash shell.

所以它就是一个 Unix shell。

So it's a it's a Unix shell.

然后智能体可以访问 shell，当然，希望是在一个沙盒环境中，也许是在沙盒中。

And then it's in then the agent has access has access to to the shell in, you know, hopefully hopefully in a sandbox, maybe maybe in a sandbox.

所以它就是这个模型。

So it's it's the model.

它是 shell，还有一个文件系统。

It's the shell, and then it's a it's a file system.

然后状态就存储在文件中。

And then the state is stored in files.

然后，你知道，文件本身还有 Markdown 格式。

And then, you know, there's the markdown format for the, you know, for for the files themselves.

接着，就像 Unix 里所谓的定时任务一样。

And then and then there's basically what in Unix is called a cron job.

有一个循环，还有一个心跳机制，这个东西基本上会定时唤醒。

There's a loop, and then there's a heartbeat for the there's heartbeat, and and the thing basically wake wakes up wakes up.

所以这本质上就是大语言模型加上壳层、文件系统、Markdown 和定时任务。

So it's basically LLM plus shell plus file system plus markdown plus cron.

结果发现，这就是一个 HS。

And it turns out that's an HS.

而除了模型之外，所有这些部分我们其实都已经完全了解和掌握了。

And and and every part of that other than the model is something that we already completely know and understand.

事实上，Unix 壳层所蕴含的潜在能力是极其惊人的。

And in fact, it turns out the, like, the latent power of the Unix shell is, like, extraordinary.

因为基本上，壳层里就存在着巨大的潜在能力。

Because basically, like, all like, there's just like an there's just enormous latent power in the shell.

Unix命令的数量非常庞大。

There's enormous numbers of Unix commands.

已经有海量的命令行接口可以访问各种功能，我的意思是，你的整个系统——只要你用的是Mac或者手机，你的电脑本质上就已经运行在一个壳层上了。

There's enormous number of command line interfaces into all kinds of things already in the you know, your entire I mean, your entire just to start with, your computer runs on a If you're running a Mac or or or a phone, your computer your computer is running on a shell already.

因此，你电脑的全部功能都可以通过命令行来使用。

And so, like, the full power of your computer is available at command line level.

而且，事实证明，将其他功能以命令行接口的形式暴露出来非常简单。

And then it turns out it's really easy to expose other functions as a command line interface.

所以，像我们还需要MCP以及这些复杂的协议之类的整个想法，

And so, like, this whole idea where we need, like, MCP and these, like, fancy protocols, whatever.

其实不用。

It's like, no.

我们不需要。

We don't.

我们只需要一个命令行工具就够了。

We just need, a command command line thing.

所以这就是架构。

So that's the architecture.

然后你会发现，你的智能体是什么？

And then it turns out, what is your agent?

你的智能体就是存储在文件系统中的一堆文件。

Your agent is a bunch of files stored in a file system.

而当我真正理解这一点时，有一件事彻底震撼了我，那就是：好吧。

And then there's the thing that just, like, completely blew my mind when I wrap my head around it as a result of this, which is like, okay.

这意味着你的智能体现在实际上独立于它所运行的模型，因为你完全可以更换底层的另一个大语言模型。

This means your agent is now actually independent of the model that it's running on because you can actually swap out a different LLM underneath your agent.

你的智能体的个性会因为模型不同而略有变化，但所有存储在文件中的状态都会

And your your agent will change personality somewhat because the model is different, but all of the state stored in the files will

被保留。

be retained.

不同的指令集，

Different instruction set,

但你只需要编译它。

but you just compile that.

对。

Right.

没错。

Exactly.

而且这没问题。

And it's alright.

这就像是换了一艘船然后重新编译。

It's like, right, swapping out a ship and recompiling.

但它仍然是你的代理，保留了它所有的记忆和能力。

But it's it's still it's still your agent with all of its memories, and with all of its capabilities.

顺便说一下，你也可以更换外壳。

And then by the way, you can also swap out the shell.

所以你可以把它迁移到另一个也是 bash shell 的执行环境。

So you can move it to a different execution environment that is also is also a bash shell.

顺便说一下，你也可以替换文件系统。

By the way, you can also switch out the file system.

对吧？

Right?

你可以替换心跳机制、定时任务框架、循环、代理框架本身。

And you can and you can and you can swap out the the heartbeat for the the cron framework, the the loop, the the agent framework itself.

所以归根结底，你的代理其实就是它的那些文件。

And so your agent basically is basically, at the end of the day, it's just it's just its files.

然后就是你的

And then and then there's your

简单的调用。

Simple call.

是的。

Yeah.

它其实就是那些文件。

It's it's basically it's it's just the files.

另外，作为这一结果，代理本身会带来一些重要的特性。

And then by the way, as a consequence of that, the agent and then the agent itself, it turns out a couple important things.

首先，它能够自我迁移。

So one is it it's it it can migrate itself.

对吧？

Right?

因此，你可以指示你的代理：迁移到不同的运行环境，迁移到不同的文件系统，或者迁移到不同的语言模型等。

And so you're you you can instruct your agent, migrate yourself to a different, runtime environment, migrate yourself to a different file system, migrate yourself to a different you know, like, we swap out the language model.

你的代理会为你完成所有这些操作。

Your agent will do all that stuff for you.

最后还有一件非常惊人的是，代理具备完全的自省能力。

And then there's the final thing, is just amazing, which is the agent is the agent actually has full introspection.

它实际上了解自己的文件，并且能够重写自己的文件。

Actually it actually knows about its own files, and it can rewrite its own files.

对吧？

Right?

顺便说一句，这在历史上几乎是没有任何广泛部署的软件系统具备的特性——你所使用的系统竟然能完全了解自身的工作原理，并能如此自我修改。

Which, by the way, is basically no widely deployed software system in history where the the the thing that you're using actually has full introspective knowledge of how it itself works and is able to modify itself like that.

我的意思是，虽然曾有一些玩具系统具备这种能力，但从未有过一个广泛部署的系统拥有这种功能。

That I mean, there have been toy systems that have had that, but there there's never been a widely deployed system that has that capability.

这进一步引出了一个能力，当我真正理解它时，简直让我震惊不已——你可以告诉代理为自己添加新的功能和特性，而它真的能做到。

And then that leads you to the capability that just, like, completely blew my mind when I wrap my head around it, which is you can tell the agent to add new functions and features to itself, and it can do that.

扩展你自己。

Extend yourself.

对。

Right.

扩展你自己。

Extend yourself.

就是，扩展你自己。

Like, extend yourself.

给自己赋予新能力。

Give yourself a new capability.