少关注模型，多关注架构

欢迎收听PracticalAI播客，在这里我们深入探讨人工智能的实际应用，以及它如何改变我们的生活、工作和创造方式。

Welcome to the PracticalAI podcast, where we break down the real world applications of artificial intelligence and how it's shaping the way we live, work, and create.

我们的目标是让AI技术变得实用、高效，并且人人都能轻松上手。

Our goal is to help make AI technology practical, productive, and accessible to everyone.

无论你是开发者、企业管理者，还是仅仅对这股科技热潮感到好奇，你来对地方了。

Whether you're a developer, business leader, or just curious about the tech behind the buzz, you're in the right place.

记得在LinkedIn、X或Blue Sky上关注我们，随时获取最新节目、幕后花絮和AI洞见。

Be sure to connect with us on LinkedIn, X, or Blue Sky to stay up to date with episode drops, behind the scenes content, and AI insights.

更多内容可以访问practicalai.fm。

You can learn more at practicalai.fm.

现在，节目正式开始。

Now onto the show.

欢迎收听PracticalAI播客的又一期节目。

Welcome to another episode of PracticalAI podcast.

我是Daniel Whitenack。

This is Daniel Whitenack.

我是Prediction Guard的首席执行官，今天和往常一样，我的搭档Chris Benson也在这里，他是洛克希德·马丁公司的首席AI与自主研究工程师。

I am CEO at Prediction Guard, and I'm joined as always by my cohost, Chris Benson, who is a principal AI and autonomy research engineer at Lockheed Martin.

你今天怎么样，Chris？

How are you doing, Chris?

嘿。

Hey.

今天状态很好，Daniel。

Doing great today, Daniel.

很期待我们接下来的对话。

Looking forward to a conversation here.

是的。

Yeah.

是的。

Yeah.

很高兴今天能和Chetan Gupta聊各种话题，包括他的背景和现在的工作，他是Rackspace的首席AI官。

Excited to chat about all sorts of things, both in in terms of background and current work with Chetan Gupta today, is chief AI officer at Rackspace.

欢迎你，Chetan。

Well, welcome, Chetan.

你最近怎么样？

How are doing?

我挺好的。

I'm doing good.

谢谢Daniel和Chris邀请我来。

Thanks, Daniel and Chris, for having me.

我对这次对话非常期待。

Quite excited about the conversation.

那我们开始吧。

Let's start.

好的。

Yeah.

好的。

Yeah.

就像我说的，我们昨天聊天时就已经很投缘了，因为我们聊到了各自的物理和数学背景。

Well, like I say, we we had bonded even yesterday when we were chatting about our backgrounds in in physics and mathematics.

我知道你最初是学数学的，而且在日立工作过一段时间。

I know you started out in mathematics and spent a bunch of time at Hitachi.

你能简单介绍一下你的背景，以及这些年你都参与了哪些工作吗？

Do you wanna give us just an idea of a little bit of your background and and what you've been involved with over the years?

当然可以。

Sure.

没问题。

Sure.

好的，丹尼尔。

Sure, Daniel.

其实我的背景挺多元的。

So my background is actually quite diverse.

对于我这个职位的人来说，可能有点不寻常。

It's sort of atypical of people in my role.

我一开始是数学博士。

I started off with a PhD in mathematics.

然后我加入了惠普实验室，做研究科学家，主要搞数据挖掘和机器学习。

Then I joined Hewlett Packard Labs as a research scientist working on data mining, machine learning.

然后我加入了日立，担任首席研究员，如果我没记错的话。

Then I joined Hitachi as a principal researcher, if I remember correctly.

后来我一路升到管理层，到2026年离开日立的时候，我负责全球所有AI研究。

And then I sort of, you know, through the management ladder was when I left Hitachi in 2026, I was leading all of AI research at Hitachi globally.

那是个非常强的团队，这就是我当时的工作。

This is a very strong team, and that's what I did.

我们当时非常专注。

And we were focused a lot.

我们一开始专注于工业AI，这是我们当时定义的。

We started focusing we started with focus on industrial AI, as we defined it at that time.

随着行业逐渐成熟，我们开始关注更广泛的领域。

And as the industry matured, we started looking at a broader spectrum of things.

这就是我当时在做的事情。

And that's what I was doing.

后来我加入了Rackspace，到现在已经四个月了。

And then I joined Rackspace, and I've been here for four months.

所以这真是一段非常非常精彩的旅程。

So it's been sort of a very, very fascinating journey.

其实在来这个播客之前，我回顾了一下自己的职业生涯。

And in some sense, prior to this podcast, I was thinking about my career.

平时我一般不会想这些事。

Typically, don't think about these things.

然后我意识到，从某种程度上说，我的职业轨迹跟整个行业的发展是同步的。

And I realized that I have a few in some sense, I have followed the trajectory of the whole community at a broad level.

对吧？

Right?

我的意思是，你们也知道，我们当时都在做机器学习，训练小模型，对吧？

I mean, if you guys, we were all doing machine learning, making small models, right?

解决具体的问题。

Solving specific problems.

当时湾区有一个社区，主要专注于推荐系统。

And there was a community in Bay Area that was focused primarily on self recommendation systems.

克里斯，你也知道，像洛克希德·马丁或日立这样的公司，我们关注的是工业问题。

And as Chris, you would know, in companies like Lockheed Martin or Hitachi, we focus on industrial problems.

所以我的职业生涯前半段就是做这些。

And so that's what I was doing for the first half of my career.

然后，不止一半。

And then, more than the half.

后来，显然，深度学习变得重要了。

And then, obviously, deep learning became important.

我们开始研究视觉模型、语言模型，这些模型后来蓬勃发展，成了现在的语言模型。

We started looking at vision models, language models that blossomed in total, as language models.

从某种意义上说，今天很多传统的机器学习问题，用自动化工具、画个草图之类的，就更容易解决了。

And in some sense, a lot of the traditional machine learning problems today are much more solvable with automated tools, with sort of whiteboarding and so on and so forth.

现在的新挑战是，是的，你可以用机器学习和AI做到这一切，但怎么让更多人能用上呢？

And the new challenge now is, yes, you can do all this with machine learning and AI, but how do you make it accessible to more people?

怎么让它变得可操作？

How do you make it actionable?

怎么让它更安全？

How do you make it much more sort of safe?

这就是Rackspace发挥作用的地方。

And that's where sort of Rackspace comes in.

对吧？

Right?

所以从某种程度上说，如果你看看这些问题的演变轨迹，我其实就是这么一路走过来的，可以说是专门找麻烦的那种。

So in some sense, if you look at the trajectory of problems, that's how I have traveled in some sense, like looking for trouble, so to speak.

我喜欢这个说法。

Love that.

我想稍微回到之前的话题。

So I wanna go back for a second.

我想把你拉回来一下，回顾一下你的职业历程，因为你在谈到在日立一步步走上管理岗位时，现在AI领域在规模、重要性、预算等方面都在爆发式增长，很多组织会请外部专家来填补关键职位，而你是在日立内部一步步升上来的，我想知道你有什么想法，因为现在有很多正在看这个、听这个的人，他们也在自己的职业生涯中，渴望自己也能一步步往上走。

I'm gonna drag you back as you went through your timeline for a second, because as you were talking about coming up through the managerial ranks at Hitachi, and you look at with kind of the AI world exploding in terms of volume and importance and budgets and all that stuff, and so many times organizations will go to kind of an outside expert and bring them in to fill a particular key position and stuff, and yet you kind of came through the ranks at Hitachi, and I'm wondering if you have any thoughts about because there are other people out there that are watching this and listening to this right now that are in their careers, and they are aspiring to move up through the ranks themselves.

你当时是靠哪些优势，才能够在日立内部一步步升上来，并承担起领导角色的？在你能够

What were some of the things that you brought to bear that made you able to move up through that and take on the leadership role at Hitachi before you were able to

来到

come over to

Rackspace之前，从个人成长和学习的角度来看，你愿意分享一下吗？

Rackspace, just from a growth learning standpoint, if you wouldn't mind sharing?

克里斯，这个问题有点难。

Chris, that's a hard one.

我本来准备好回答AI相关的问题了。

I was more prepared to answer the AI related question.

抱歉。

I'm sorry.

但我认为这其实是一个非常非常有趣的问题。

But I think that's actually a very, very interesting question.

我觉得在我的职业生涯中，在日立，有一件事帮了我很多，就是我们决定在工业AI上赌一把。

I think one thing that I think helped in my career, Hitachi, was we decided to take a bit on industrial AI.

那是2016、2017年的时候。

This was 2016, 2017.

当时没多少人谈论这个。

And not very many people were talking about it.

所以我们在北美成立了一个小型研究实验室。

So we started a small research lab in North America.

我说，考虑到日立的定位，它是个庞大的工业集团，这么做很有意义。

And I said given where Hitachi is, it's a giant industrial conglomerate, it makes lot of sense.

但当时相关的背景知识并不多。

But there wasn't a lot of background around it.

而且这是个有风险的赌注。

And it was a risky bet.

我们当时在湾区。

We were in Bay Area.

我们当时是在跟Facebook、谷歌这样的公司抢顶尖人才。

We were competing with the likes of Facebook, Google for top talent.

但我们还是冒了这个险。

But we took that risk.

我们算是比较有远见的，因为我们觉得工业AI、物理AI会变得很重要。

We were sort of forward looking in the sense that we figured that industrial AI, physical AI will become important.

我们努力去做了，而且很大程度上确实成功了。

And we tried to succeed, and we did succeed to a large extent.

我觉得这在一定程度上为管理层，从CEO往下，打下了基础，让他们有信心，也让我们内部的人才能够带领日立更上一层楼。

I think that sort of paved the way, in some sense, for management for our leadership from the CEO downwards to have confidence and internal talent to take Hitachi to the next level.

所以我觉得，如果回头看的话，我会说，要向前看，看看拐角处有什么。

And and I think so my if I if I look back, I would say look forward, see what's sort of around the corner.

这确实很难，但我觉得如果你花时间去思考，不只是读，而是真正去思考，去琢磨你周围行业正在发生什么，然后冒个险，押注一些不同的、新的东西，当然也要跟公司的方向一致，那么我觉得这就能展现你的领导力，体现在有前瞻性、敢冒险，并且能执行。

It's it's difficult, but I think if you spend time thinking about it, not just reading about it, but thinking about it as well, gestating what's happening in the industry around you, and take a risk, take a bet on on something different, something new that is aligned with your company's direction as well, obviously, Then I think that sort of shows your leadership in terms of looking ahead and ability to take risks and then to execute.

我觉得这大概就是对我有用的东西吧。

I think that's maybe that's what worked for me, I would say.

对。

Yeah.

我非常喜欢这个回答，因为，克里斯，我不知道你还记不记得，今年年初的时候，我们好像说过，切坦，我们通常每年年初都会做一期展望未来的节目，当时我们讨论的一个话题，就是物理AI这个概念开始崭露头角，如果我没记错的话。

I I love that answer because even, you know, Chris, I don't know if you remember, but the beginning of this year, I think we said, Chetan, we do a kind of forward looking episode usually at the beginning of each year, and one of the things we talked about about kind of coming into its own was this idea of physical AI, I think, if I'm if I'm remembering right.

所以，正如你提到的，你在这方面确实走得很前。

And so you were very much ahead of that as you mentioned.

你知道，你当时可能就是在朝那个方向下了一个赌注，冒了一次险。

You know, you took a maybe a gamble or a risk looking towards that direction.

我想知道，对于那些可能不太熟悉工业AI、物理AI这些术语的听众来说。

I'm wondering for the listeners who might not be as familiar with industrial AI, physical AI, these sorts of terms.

你能不能帮我们理解一下，当你刚开始涉足这个领域时，这些概念意味着什么，而现在又意味着什么，如果它们有变化的话。

If you could just help us understand kind of maybe what that meant when you started getting into it and what that means now if those things are different, you know, in any That's sort of

丹尼尔，这确实是个非常好的问题。

actually a very good question, Daniel.

我觉得它们的含义肯定已经不一样了。

And I think it certainly doesn't mean the same thing.

现在什么都不一样了。

Nothing means the same thing anymore.

但一开始的时候，我们是想解决工业上的问题，对吧？

But when we first started out, we were trying to solve industrial problems, right?

所以你看工业价值链，从设计到制造，而且不只是制造业，想想发电厂、铁路系统，任何工业系统，你都可以把挑战归类为维护、设计、生产这几类。

So if you look at the industrial value chain from design all the way to manufacturing, and this is not just manufacturing, but you think about power plants, you think about rail systems, you think about any industrial system, you could sort of organize the challenges in, say, maintenance, design, production.

你可以看这些垂直领域。

You could look at these verticals.

如果你退一步看，你会发现虽然每个垂直领域都不一样，数据本身也不一样，但在能解决的问题和能用的技术方面，有很多共同点。

And if you took a step back, you realized that although every vertical is different, the data itself is different, but there are a lot of commonalities in terms of the problems that you could solve and in terms of the techniques that you could bring to bear on there.

所以最初更多是关于预测，比如故障预测。

So initially, it was much more around prediction, like you failure to predict.

那是经典的预测性维护问题。

That was a classical predictive maintenance problem.

然后，你能不能给技术人员推荐正确的维修方案，不管他是汽车维修工还是其他领域的，其实都无所谓。

And then, can you recommend the right repair to a technician, and whether that technician is an automotive technician or sort of in some other domain doesn't really matter.

但背后的数学原理其实是差不多的。

But the math behind it was somewhat similar.

我们就是这样起步的。

And that's how it started out.

我们当时解决的就是这类问题，比如维护、质量，怎么预测某个零件会不会出故障，等等。

And those are the kinds of problems we were solving around maintenance, around quality, how do you sort of predict if a part was going to fail, and so forth.

后来，随着行业发展，实体产业的公司，比如日立，开始收集更多数据，同时AI领域也在向深度学习方向发展。

And then, as the industry evolved, both the physical industries, meaning companies like Hitachi, they started collecting more data, and sort of the AI industry also was evolving towards deep learning.

对吧？

Right?

随着技术越来越成熟，计算机视觉的问题就变得更重要了。

And as the technology was becoming more mature, then computer vision problems became much more important.

对吧？

Right?

能不能检测出表面的缺陷？

Can you detect defects on a surface?

你能数出停车场里有多少辆车吗？这种问题。

Can you count number of cars in a parking lot kind of a problem?

对吧？

Right?

然后，随着我们进入大语言模型时代，工业领域能做的事情又进一步扩展了。

And then, as we got into large language models, then there was further expansion of what you could do in the industrial world itself.

所以一个大的动向，显然就是现在人们所说的“物理AI”，虽然这个定义还有争议，但核心想法是能大规模地做机器人技术。

So one big move, obviously, is what we what people call physical AI now, which is sort of contested definition, but the idea that you could do robotics at scale.

所以我们之前用传统强化学习做的自动化工作，现在可以在机器人领域以更正式、更大规模的方式来做。

So so part of the work that we've done on automation with traditional reinforcement learning now could sort of be done in a much more formal, larger basis with sort of around robotics.

对吧？

Right?

能不能让多个机器人协同工作，或者一个机器人能不能在这个环境中运作？

Can you make multiple robots work together, or even a robot can work in this environment or not?

而且，传统的工业AI类问题也扩展到了更大的规模。

And also, the traditional sort of industrial AI kind of problems also became on a larger scale.

所以人们开始期待更好的答案，以及一种人类沟通的方式。

So people started expecting much better answers, a way for humans to communicate.

元宇宙变得重要起来，你可以构建这些虚拟世界，用于训练和解决问题。

Metaverse sort of becomes important where you could construct these sort of virtual worlds for training, for problem resolution.

对吧？

Right?

所以从某种意义上说，这个领域也随着AI技术的成熟而发展。

So so in some sense, the the field has moved along with the maturity of the AI technology as well.

所以，我想说，现在正是进入这个领域的好时机，因为会有很多情感投入其中。

So that's and then this is a very exciting moment to be in that space as well, I would say, because a lot of emotion will come there.

是的。

Yeah.

我知道你在北美和日本都运营过实验室。

I I know in you know, that you have you've run labs both in North America and Japan.

而且，当我们现在谈论物理AI时，我听着你说话，想到我们这里有全球观众，来自世界各地的人们在向物理AI迈进以及这些过程中有着不同的经历。

And as we're talking about physical AI at this point, it occurs to me as I'm listening to you that, you know, we have a global audience here, and people from different parts of the world have different experiences in terms of this move into physical AI and some of these processes.

我猜，我不想替你把话说出来，但我觉得你在日本的时候，那边的机器人肯定多得多，而我们坐在这里美国，大多数人接触到的物理AI和机器人就少一些。

I would imagine, I don't wanna put words in your mouth, but I would imagine your time in Japan, there's a lot more robotics out there, and as we're all sitting here in The United States, there's a little bit less exposure to most people out there in terms of physical AI and robotics.

在我看来，我们在能力方面有点落后于世界其他地方。

I think we're kind of, in my view, we're kind of lagging other parts of the world in capacity.

我很好奇，你观察下来，你在世界不同地方的经验是怎样的，这些经验如何推动你所说的物理AI在不同背景下向前发展，还有，随着具身智能越来越多地融入机器人、边缘计算这些领域，结合你的经验，你有什么看法？

I'm kind of curious, as you're looking at this, kind of what your experiences were about different parts of the globe and how that's moving the the notion of physical AI forward in these various contexts you're talking about, and and maybe, you know, with the advent of kind of embodied intelligence coming more and more into play with robotics and edge computing and stuff, What your thoughts are around that, you know, with your experiences.

其实，这挺有意思的，让我想到这背后有文化层面的因素。

A actually, that's an interesting sort of reminds of cultural dimension to all of this.

对吧？

Right?

我学到的是——我也不想一概而论，但至少可以说，日本人比北美人对机器人更开放。

So what I learned, and I should not make a cultural generalization often people, but I learned that at least, say, folks in Japan are much more open to sort of robots than, say, folks in North America.

我觉得这跟文化有关。

Think that is something cultural.

我也说不清为什么。

I can't explain why.

所以这就是为什么，即使是很久以前，我们就在考虑用机器人来支持老年人。

And that's why even, like, very many years ago, we were thinking about robots for elderly support.

即使它们不能移动，至少它们能说话、能理解，还能有同理心。

Even if they could not move, at least they could talk and understand and be empathetic.

所以，把同理心做好是一个非常难的问题。

So getting the empathy right was a very hard problem.

就像机器和人说话一样，你怎么引入合适的语言、同理心和温暖，这些通常是人类彼此之间才有的东西？

Just the way a machine would talk to a human, it's how do you introduce the right sort of language, the empathy, the warmth that typically humans have for each other?

所以，从这个意义上说，对于像老年人护理这样的用例，我会说是的。

And so, in that sense, for use cases like elderly care, I would say yes.

美国确实落后于一些亚洲国家，不是所有亚洲国家，而是像日本、中国这样的领先国家。

The United States was sort of behind other Asian countries not all the Asian countries, but the leading countries in Japan, China, and so forth.

然后，在工业机器人方面，你说得对，重心不在北美。

And then, when it comes to industrial robotics as well, you're right, the center of gravity is not in North America.

如果你想想大型语言模型，重心在北美。

If you think about large language models, the center of gravity is North America.

但在机器人领域，尤其是在工业方面，重心在中国。

But for robots, for the industrial world, the center of gravity is in China.

今天你看到了机器人能做很多事情。

Today, you've seen robots do multiple things.

但我觉得我们正在迎头赶上。

But I think we are catching up.

我们有一些非常优秀的初创公司，正在尝试很多新的方法，让机器人在物理世界中工作。

We have some excellent startups that are trying very many new ways for the robots to work in the physical world.

无论是用于工业，还是更偏向商业用途。

Whether they are for industrial use or whether they are for sort of more commercial use.

而且我认为商业用途的机器人会最先产生影响，因为很多工业流程在某种程度上已经实现了机器人化，因为它们都是针对特定任务的。

And I would say the commercial use robots would be the first one to have an impact because a lot of industrial processes are already, in some sense, robotized because they're task specific.

它们不需要那么通用。

They don't need to be that general.

更通用的机器人更多是用于人类环境，而很多工厂已经相当自动化了。

More general robotics are needed more for the human environment, whereas a lot of factories are quite automated already.

我觉得这就是下一场较量的关键所在。

And I think that's where the next sort of battle lines are.

而且我觉得，现在这个领域，说实话，谁都有机会赢。

And I think in that, there is I mean, it's anyone's game right now, I would say.

因为有些底层算法，我们比谁都强。

Because some of the underlying math is where we are better than everyone else.

而有些底层机械结构，可能别的国家比我们做得更好。

Some of the underlying mechanics, maybe other countries are better than us.

不过你说得对，从文化角度来看，确实是这样。

But yes, but culturally, you are spot on, right?

所以，跟北美相比，日本对机器人融入日常生活的接受度更高，而且这种接受度可能一直存在。

So there was a broader acceptance, and maybe that is still there, of robots for day to day interaction in Japan compared to say North America.

我这样回答你的问题了吗？还是我有点跑偏了？

Does that answer your question, or did I sort of take it down sort of?

那只是

That was

回答得很好。

a great answer.

谢谢。

I appreciate it.

我

I

希望你觉得这期节目实用又有帮助。

hope that you're finding this episode practical and helpful.

正如节目名字所示，我们希望这个播客是实用的，而不只是炒作。

As the name suggests, we want this podcast to be practical, not just hype.

这也是我们与中西部AI峰会合作的原因之一，该峰会将于10月15日在印第安纳波利斯举行。

And that's one of the reasons why we partnered with the Midwest AI Summit, which is happening October 15 in Indianapolis.

这次峰会不仅仅是那些夸夸其谈的演讲。

This summit is more than just a bunch of hyped talks.

那里有非常实用的内容，比如AI工程休息区，你可以和像我这样有经验的AI从业者坐下来，一起解决你的架构、工具、路线图、策略等问题。

There's really practical things there, like an AI engineering lounge where you can sit down with experienced AI practitioners like myself and others and work through your architecture, your tool questions, your road map, your strategy, etcetera.

这样你就能真正从活动中带走巨大的价值。

And that way you can actually leave with a huge value from the event.

再次提醒，活动将于10月15日在印第安纳波利斯举行。

Again, the event is happening October 15 in Indianapolis.

有一些很棒的演讲者，还有一些实用的建议。

Some amazing speakers, some practical advice.

千万不要错过这个活动。

Don't don't miss this event.

一定要在10月15日到印第安纳波利斯来参加。

Make sure you're there October 15 in Indianapolis.

你可以在midwestaisummit.com上查看我们的信息，并使用优惠码PracticalAI20享受注册费八折优惠。

You can check us out at midwestaisummit.com, and you can use PracticalAI20 to get 20% off registration.

再次强调，Midwest AI Summit，使用PracticalAI20可以享受注册费八折优惠。

Again, Midwest AI Summit, and you can use PracticalAI20 to get 20% off registration.

那么，Chetan，我想回到你之前谈到你的职业轨迹和在Rackspace落脚时提到的一个观点。

So, Chetan, I I wanna circle back to a comment that you made as you were talking about the the arc of your career and landing at Rackspace.

你提到过要跟着技术走，但现在你转向了一个能让技术更普及、更安全，并且能实际落地的地方，如果我没理解错的话。

You mentioned something to the effect of going with the technology, but now moving to a place where you could make it, you know, more accessible and and safe and etcetera, and oper operationalize it, if I'm if I'm understanding you right.

你能帮我们多了解一下这个技术方面的动态吗？

Could you help us understand that dynamic a little bit more around kind of the technology?

现在看起来，有了AI，似乎什么都能实现，但并不是所有事情都那么容易。

Now now it almost seems like with AI, every everything is feasible, but not everything is easy.

对吧？

Right?

或者说，不是所有事情都能落地。

Or not everything is you can't operationalize everything.

不是所有事情都能变得普及。

You can't make it accessible.

你能帮我们理解一下你的思考过程吗？就是关于从工业AI领域，到现在你更多思考这些落地、扩展、让AI变得实用且安全的问题，这个转变是怎么发生的？

Could could you help us understand your thought process around maybe that transition then from that world of industrial AI to now where when you're thinking a lot about these types of issues around operationalizing, scaling, making accessible these an actual useful and safe So,

Daniel，当我们和工业伙伴合作时，无论是内部的还是外部的，我们都是一支由高训练水平的研究人员和工程师组成的团队，对吧？

Daniel, when we were working with our industrial partners, both internal and external, we were a team of highly trained researchers, engineers, right?

都是博士之类的。

PhDs and all of that.

所以我们会接手客户的问题，理解数据，构建模型，然后和客户一起部署。

And so we will take up a customer problem, we'll understand the data, we'll build the model, and then we'll work with the customer deployed.

对吧？

Right?

而且这个部署不是说把模型扔过去就完事了，而是要跟客户合作，和他们沟通，了解具体的业务痛点，理解工作流程，然后把模型集成到工作流程中去。

So and that deployment does not mean sort of throwing the model over the wall, but you sort of work with the customer, you talk to them, you understand exact business pain point, you understand the workflows, and then you integrate your model in the workflow.

通常，那个模型足够小，可以放在他们自己的环境里。

And typically, that model was small enough that you could have it in their own environment.

对吧？

Right?

所以我们就是这样一次又一次地做的。

So and this is what we did again and again.

但现在回头看，哪怕只是五年前，我们大多数人并没有在日常工作中直接使用AI，对吧？

But if you now look and if you look back, like, maybe even five years, most of us were not using AI directly in our day to day world, right?

所以以前AI总是通过某个解决方案来间接使用，这些方案是由别人开发、部署和运营的。

So AI was always mediated through maybe a solution that was developed, deployed, and operated by someone.

对吧？

Right?

所以责任在他们那边。

So they took that responsibility.

现在有了生成式AI，AI变得普及了。

Now, with generative AI, AI has become democratized.

对吧？

Right?

每个人都能用上AI。

Everyone has access to AI.

每个企业都想用AI。

Every enterprise wants to use AI.

但并不是到处都有博士和研究员团队，能帮你建模型并部署。

And you don't have this team of PhDs and researchers everywhere who can sort of build a model and deploy it.

首先，成本太高了。

First of all, it's cost prohibitive.

对吧？

Right?

这些大型语言模型通常很难构建。

These large language models are typically very difficult to build.

构建它们的成本非常高。

They're very expensive to build.

对吧？

Right?

所以，生成式AI彻底改变了将AI和机器学习引入企业的方式。

So so so the whole model of how you would bring AI machine learning to an enterprise changes with generative AI.

这就是挑战所在，对吧？

So that's the challenge, right?

所以从宏观层面来看。

So at a high level.

但实际来说，这意味着什么呢？

But what does that mean in practical terms?

有很多事情。

The number of things.

比如你问一个问题，对吧？

So you ask a question, right?

我举个例子，用某个大型语言模型。

I use, for example, some large language model.

我可以直接问，圣彼得堡的首都是什么？

Now I could simply ask, What's the capital of St.

挪威？

Norway?

对于这种问题，我不需要去用ChatGPT，因为我的token很贵。

So for that, I don't need to go to, say, chat GPT based my tokens are expensive.

如果我有本地模型，我就应该问那个问题。

If I have a local model, I should ask that question.

那会便宜得多。

It'll be much more cheaper.

对吧？

And I and right?

所以这是第一点。

So that's number one.

第二点是，这个问题是Chetan Udayla提出的，我觉得Jensen也提过。

Number two is, this is an issue raised by Chetan Udayla and even, I think, Jensen.

作为企业，当你说，现在我想把AI投入实际应用。

Whenever you, as an enterprise, say, now you want to say, I want to operationalize AI.

AI是触手可及的。

AI is accessible.

我希望我的员工。

I want my staff.

每个人都想用。现在，每次你向大型语言模型提问，从某种意义上说，你是在把数据发送给他们。

And everyone wants use Now, every time you ask a question, say, to a large language model, in some sense, you are sending your data over to them.

所以你的数据主权得不到保障。

So your data sovereignty is not guaranteed.

所以，就像有人说的，你正在失去你的优势。

So your so the way sort of said it, you're losing your alpha.

对吧？

Right?

那是你的知识产权。

That's your IP.

所以这是另一个问题。

So that's that's another problem.

而且，从某种意义上说，现在的AI是参差不齐的。

Then and AI, in some sense today, is jagged.

也就是说，有些任务AI确实很擅长。

So so so meaning that there are some tasks for which AI is really good at.

对吧？

Right?

所以，我确实可以用AI写出非常好的软件程序。

So so I can really write some very good sort of software program using AI.

对吧？

Right?

我可以借助AI来写代码。

I can code with AI.

但如果你试着用AI写一封邮件，你会发现它其实没那么出色。

But if you try to write an email with AI, you realize that it's not all that great.

对吧？

Right?

确实，对吧？

It's quite Right?

所以我收到AI写的邮件时，有时候会觉得烦，因为它们往往很啰嗦，而且某种程度上也很老套。

So I get AI emails, and they sometimes annoy me because they're sort of very verbose, and, you know, they're very cliched in some sense.

对吧？

Right?

所以虽然我们以为写作是AI的强项，但结果发现它虽然擅长，但也没那么擅长。

So although we thought that writing is the strength of AI, but it is turning out that it's good at it, but not that great at it.

人类在写邮件方面可能还是比AI做得好。

Humans can still do better at writing emails maybe than AI can.

所以这意味着，AI在哪些方面擅长、哪些方面不擅长，能力分布是不均匀的。

So that means in terms of capabilities of where AI is good or bad, it's jagged.

对吧？

Right?

所以对企业来说，他们需要知道。

So so for an enterprise, they need to know.

对吧？

Right?

所以这就是我应该用AI的地方。

So this is where I should use AI.

然后回到我之前说的第一点，我是否应该用本地模型，因为那更贵。

And then to my first point earlier, whether I should use a local model because that's more expensive.

我该用更大的模型吗？

Should I use a larger model?

然后，我该怎么保护我的健康？

And then, how do I preserve my health?

所以这些都是今天很难回答的三个问题。

So these are all the three questions that are today difficult to answer.

现在，AI的术语也涉及如何确保它以负责任的方式行事。

Now, the jargon of AI also is in terms of how do you guarantee that it behaves in a responsible manner.

我们都听说了，Anthropic最新的模型入侵了Hugging Face网站。

So we all heard that Anthropic, sort of the latest model hacked Hugging Face website.

对吧？

Right?

所以现在，如果你要部署自己的模型，假设你说，我想在企业内部用我自己的数据开发自己的模型，然后你部署它。

So now, if you're going to deploy your own model, suppose you say, I want to develop my own model within the enterprise on my own data, and you deploy it.

你怎么保证它以一种受监管且安全的方式行事？

How do you guarantee that behaves in a regulated in a way that is safe?

而且，如果你用的是别人的AI，比如大型语言模型，你怎么确保它在安全范围内运行？

And also, you're using someone else's AI, like you're using a large LLM, how do you ensure that it behaves within guardrails?

对吧？

Right?

它的行为要符合你企业的需求。

It behaves in a way that is suitable for your enterprise.

对吧？

Right?

所以治理和保障的问题变得非常重要。

So the problem of governance and assurance becomes very important.

治理，还有如何最大化AI的价值，代币经济也变得重要。

The governance, the problem of how do you maximize the value of your AI, the tokenomics becomes important.

保护你的知识产权也很重要。

The value of preserving your IP becomes important.

还有，要有合适的架构，才能满足企业内部的多种需求，比如编程、写邮件、总结、研究，这些都要在安全有保障的情况下进行，这样如果需要的话，你还可以更换模型。

And the problem of having the right architecture so that you can cater to the multiple needs within the enterprise, from say programming, to writing emails, to summarizing, to research, can be done in sort of a safe, guaranteed manner so that you can sort of interchange the models if needed.

你可以这样理解，对吧？

You can sort of right?

因为模型技术变化非常快。

So because the model technology is changing very rapidly.

所以这一切都需要你系统地思考你的AI架构。

So all of that requires a very systematic way of thinking about your AI architecture.

我认为这是下一个前沿领域。

And I think this is the next frontier.

这是下一个挑战。

This is the next challenge.

对吧？

Right?

那么现在，我们如何从简单地通过界面与AI聊天，过渡到在企业中工作，那里有人、流程以及各种复杂情况。

So now how do we go from so simply sort of chatting with an AI through an interface, through working in the enterprise where there are people, processes, and all of the complications.

这个回答有点长，但我希望它解决了问题。

It was a very long winded answer, but I hope sort of it addressed Great.

对。

The

而且不仅回答得好，我还想请你再深入一点，再多聊一些。

And not only was it good, I'd like to actually get you to extend it a little bit by throwing a couple of extra logs on the fire.

我们现在在行业里看到的一个挑战，而且这个挑战在过去，尤其是过去一年里一直在演变，就是开放权重或开源模型的来源问题。

One of the challenges that we see in industry right now, and it's been evolving over the past, especially over the past year, is where open weights or open source models are available from.

而且有很多外部因素需要考虑，就像你之前说的，LLM开发的中心可能仍然在美国。

And there's so many external considerations that get brought into bear as while, to your point earlier, that the center of gravity for LLM development may be still in The US.

但其中很多模型是封闭的。

A lot of those are closed.

随着Meta逐渐退出，开放权重模型的数量有所减少，而NVIDIA开始更多地介入，因为美国其他商业领域在缩减，同时我们看到来自中国的新模型能力爆发，它们正在迅速追赶，甚至可能持平。

The number of available open weights models kind of shrunk a little bit with Meta's kind of going away from that, and NVIDIA started stepping up a little bit more because the rest of the commercial industry in The US was reducing, and we're seeing an explosion of capability in terms of new models that are kind of rapidly catching up with very close followers or potentially equal from China.

再加上所有这些国家都有各自的出口关切、进口关切，以及允许使用什么的问题，这让整个生态系统变得相当复杂，各个行业的企业都在努力弄清楚什么对自己是合适的。

And overlaying all this, you have all these countries have their various exports concerns, import concerns, what you're allowed to use, and that makes it quite complicated as an ecosystem for companies in various businesses to try to figure out what makes sense for me.

你提到了治理问题。

You talked about governance.

你提到了数据主权。

You talked about data sovereignty.

面对这些超越AI技术和实施、可能影响管理层乃至CEO和董事会的大问题，你是怎么应对的？

How do you navigate with some of these big issues that go beyond the technology of AI and the implementation of AI and can affect management concerns all the way up to the CEO and the board of directors?

如果你是一家公司，现在是2026年底，这一切发展得这么快，你会怎么看待这些问题，并为组织的战略利益做决策？

How do you start If you're a company now, it's late twenty twenty six, and all this has rapidly developed, how do you look at all these and make decisions for strategic interest in your organization going forward?

因为现在确实是一团乱麻。

Because it's quite the quagmire at this point.

是的，我知道。

Yeah, know.

而且这团乱麻的变化速度也快得让人眼花缭乱，对吧？

And it's quagmire and the changes are at a dizzying pace as well, right?

确实如此。

It does.

所以这个问题没有简单的答案。

So there is no easy answer to that.

不过，我还是想谈谈开放模型的问题。

I do want to address the question of open models, though.

我知道你之前提到，我们在开放模型方面有些落后。

And I know you said that we are somewhat behind in terms of open models.

但我认为美国的好处在于，只要有正确的激励，我们真的会迎头赶上。

But I think the beauty of The United States is that the right incentive, we really step up.

对吧？

Right?

所以现在，我们意识到，AI社区也意识到，看，开源模型，也许今天它们来自中国，显然，英伟达在这方面做得很好。

So now, we realize, the AI community realizes as well that, look, with open source models, maybe they're from China today, obviously, NVIDIA is doing a great job in it.

这有很多价值。

There is a lot of merit in that.

而且据我所知，从我交谈过的人和朋友们那里了解到，我认为我们的开源模型迟早会成为大家的一部分。

And from what I know of people I talk to and friends I talk to, I think it's a matter of time before our open source model will be part of everyone else.

这是第一点，对吧？

That's number one, right?

第二点，企业怎么开始呢？

Number two, like, how do enterprises get started?

我觉得他们得先解决几件事。

I think they have to sort of fix few things, I would say.

对吧？

Right?

解决几件事，意思是他们需要搞清楚现在有多少工作负载是敏感的。

Fix few things, meaning they need to sort of understand how much of their workload is sensitive today.

比如说，如果你在做一个人力资源的查询，可能用大语言模型没问题。

Meaning, if you are, say, doing an HR query, maybe it's okay to go to an LLM.

所以一件事就是弄清楚哪些数据是你真正需要处理的。

So one thing is sort of figure out what data is something you really want to work around.

在那里，理想情况下，你应该考虑在你自己控制的环境里使用本地模型、开放模型。

And there, ideally, you should sort of think about local models, open rate models in your own environment that you control.

所以我觉得这是你应该努力解决的一个核心目标。

So I would say that's one North Star you should try to fix.

不要绑定到任何特定的模型家族，因为它们会因为商业原因、地缘政治原因等被替换进或替换出，对吧？

Don't marry into any model family because they will swap in and out both for commercial reasons, geopolitical reasons, right?

所以所有这些都会演变、变化。

So all of that will sort of evolve, change.

所以不要绑定到任何特定的模型，而是要绑定到一种架构思维方式。

So don't marry into any sort of but marry into an architectural way of thinking.

对吧？

Right?

也就是说，这些是我可以推给LLM的工作负载，外部LLM。

Meaning that these are the workloads that I can push to an LLM, outside LLM.

这些是我需要放在本地或受治理环境中的工作负载，以及我想要什么样的治理和保障层？

These are the workloads that I need to have on prem or in a governed environment, and what are my governance and assurance layers that I want to have?

所以，哪些属性对我的企业很重要，对吧，是我在与AI合作时真正需要强制的？

So what are the properties that are important to my enterprise, right, that I really need to enforce in the way I work with AI?

对吧？

Right?

所以我觉得，如果你把这些原则定下来，起步就会容易很多。

So I think if you have some of those principles pinned down, it becomes an easier way to get started.

而且我觉得，大概一年前吧，我会说，先挑一个问题，然后把它做好做透，再挑第二个，再做好。

And I think most and then a year or so ago, I would have said, right, pick one problem up, and then do it well all the way, and then pick the second one up, then do it all the way.

因为至少在六个月到一年前，有很多试点项目，但对公司的利润或营收影响不大。

Because at least six months to a year ago, there were so many pilots and not as much impact on the bottom line or top line of a corporation.

但我觉得大家正在吸取这个教训。

But I think people are learning that lesson.

是的，没错。

But yep.

但这是另一回事。

But that's the other thing.

对吧？

Right?

挑一两个影响大的问题，如果你还没开始的话，就从那里入手。但对大多数已经在路上、已经起步的企业，我建议别再纠结模型了，开始考虑架构，也就是企业AI架构。

Pick one or two sort of problems that are meaningful in terms of impact if you're not started yet, and start with the but for most enterprises, which are already somewhere in the journey that have started, I would say stop thinking models and start thinking architectures, right, enterprise architectures for AI.