本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
2015年，布罗克曼、埃隆·马斯克和萨姆·阿尔特曼共同创立了一家名为OpenAI的非营利组织。
In 2015, Brockman, Elon Musk, and Sam Altman started a nonprofit called OpenAI.
他们的目标是研发通用人工智能。
And the plan was to pursue artificial general intelligence.
当初还是非营利机构时，公司就在格雷格·布罗克曼的客厅里起步了。这些人坚信，实现与人类智力相当的通用人工智能是可能的。说实话，当时硅谷的大多数人觉得这不过是个有趣的副业，大家的关注点都集中在社交媒体上。
The company or the nonprofit when it was a nonprofit back then began in Greg Brockman's living room and these folks were convinced that achieving artificial general intelligence or AI on par with human intelligence was possible and to be honest, most people in the valley thought that that was an interesting side project, but most of the attention was on social media at the time.
转眼十一年过去，今天我们站在这里。
Well, fast forward eleven years and here we are.
OpenAI很可能在明年上市，估值将达到一万亿美元。
OpenAI is probably going to go public within the next year at a trillion dollar valuation.
他们很快就会宣布，ChatGPT的用户数已突破十亿——第三方数据已经显示出这个趋势。
They're going to announce likely because third party data is showing it a billion users in ChatGPT fairly soon.
他们还拿到了历史上金额最大的一笔风险投资，估值高达1220亿美元。
And they of course raised the largest venture capital around in history at a 122,000,000,000.
如今，他们站在了这项彻底改变世界、吸引全人类关注的技术最前沿。
So they are at the leading edge of a technology that has captured all of our attention and is changing the world.
当你听格雷格·布罗克曼说话时，即使回溯到很久以前的对话，也能清楚地感受到他对这项技术未来走向和产品发展方向的坚定信念。
And so when you listen to Greg Brockman, one of the things that you can see even from his conversations all the way in the past is a clear conviction and understanding of where this technology would lead and where the products would go.
真的很了不起。
It's amazing.
听听ChatGPT出现之前、GPT模型刚起步那会儿的播客，就能听到格雷格对如今模型和应用的发展方向有着无比清晰的判断。
You listen to the podcasts from pre chat GPT or really in the early days of the GPT models and you can hear Greg speaking with absolute clarity about where models and applications would go today.
所以，为了今天的内容画上句号，让我们和格雷格·布罗克曼一起展望OpenAI和前沿技术的未来。
So I think to close our day, let's take a look into the future of where OpenAI and the frontier is going in a conversation with Greg Brockman.
有请格雷格。
Let's welcome Greg.
面对持续的变革与机遇，TMT领域的领导者需要交付实实在在的成果，而不仅仅是想法。
In the face of ongoing disruption and opportunity, TMT leaders need to deliver tangible results, not just ideas.
当速度和执行力成为关键时，普华永道结合市场洞察、行业专长，以及人工智能、云计算和新兴技术，加速您的转型，从战略到落地实现可衡量的投资回报。
When pace and performance matter most, PwC combines market insights and deep sector experience with AI, cloud, and emerging tech to accelerate your transformation and drive measurable ROI from strategy to execution.
普华永道能帮您预见下一步，超越颠覆，赢得竞争。
PwC can help you anticipate what's next, outpace disruption, and compete.
如需更多信息，请访问 pwc.com。
For more information, visit pwc.com.
保险不是千篇一律的，选保险也不该像硬挤进一件根本不合身的衣服。
Insurance isn't one size fits all, and shopping for it shouldn't feel like squeezing into something that just doesn't fit.
所以多年来，车主们一直很喜欢 Progressive 的‘自定价格’工具。
That's why drivers have enjoyed Progressive's name your price tool for years.
使用‘自定价格’工具，你告诉他们你愿意花多少钱，他们会给你匹配符合你预算的方案。
With the name your price tool, you tell them what you want to pay, and they show you options that fit your budget.
别再到处找折扣、费劲算保费、反复调整保障范围了。
Enough hunting for discounts, trying to calculate rates, and tinkering with coverages.
不管你是第一次买保险，还是只是想找个更适合你和家人的方案。
Maybe you're picking out your very first policy, or maybe you're just looking for something that works better for you and your family.
无论如何，他们都能让你轻松看清所有选择。
Either way, they make it simple to see your options.
不用猜，没意外。
No guesswork, no surprises.
想不想看看买汽车保险原来可以这么简单又轻松？
Ready to see how easy and fun shopping for car insurance can be?
上 progressive.com 试试他们的‘你定价’工具吧。
Visit progressive.com and give the name your price tool a try.
告别繁琐比价，找到真正符合你生活需求的保障方案。
Take the stress out of shopping and find coverage that fits your life on your terms.
Progressive 保险公司及其关联公司。
Progressive Casualty Insurance Company and affiliates.
价格与保障范围受各州法律限制。
Price and coverage match limited by state law.
很高兴见到你，格雷格。
Great to see you, Greg.
谢谢你的邀请。
Thank you for having me.
你知道吗，格雷格，这是我们第四次聊天了，每次谈的都是 OpenAI 的产品方向。
You know, Greg, this is our fourth time speaking and and we've spoken every time about OpenAI's product direction.
我觉得我有点明白了。
And I think I'm starting to get it.
你知道吗，之前我们聊到‘超级应用’这个词，其实不太准确，因为你做的这个应用是把 OpenAI 的代码功能和 ChatGPT 整合在一起。
You know, was this conversation that a super app was the wrong term for what you were doing with the app that you're building bringing codecs which is coding side of OpenAI's product browser and chat GPT together.
当人们听到‘超级应用’，第一反应是：不对，超级应用应该是能直接嵌入其他所有应用的那个东西。
And when you use the word super app people would be like, no, super app is actually something that you can just use every other app within.
但现在看到这些功能逐渐融合，也许‘超级应用’这个说法反而更贴切了。
And now as we've seen these products come together, actually super app might be the correct term.
至少从我们外部人的角度看，现在确实越来越明显了。
You know, at least for us on the outside we're starting to see it.
当你需要做任何事，都会先在 ChatGPT 里输入一个指令，然后 OpenAI 的技术会自动调用你的浏览器或电脑帮你完成。
That when you need to do anything, it will start with a prompt in chat GPT and then OpenAI's technology will use either your browser or your computer to get that done for you.
这样理解对吗？
Is that the right way to think about it?
我觉得这个视角挺到位的。
I think it's a pretty good perspective.
没错。
Right.
我觉得如果从更宏观的角度看，我们真正想打造的是通用人工智能。
And I think to really zoom out, the thing we're actually trying to build is in AGI.
没错。
Right.
从聊天开始，人们一直在使用的其实是一个语言模型。
That if you think about what people have been using since chat, it's a language model.
没错。
Right.
这两者之间存在着巨大的差距。
There's a big gap between these.
太棒了。
Amazing.
你能和聊天机器人对话，它也会回应你。
You can talk to chat, talks back to you.
很好。
Great.
太棒了。
Wonderful.
但我们在2022年推出时，根本没有记忆功能，对吧？
But when we launched in 2022, there was no memory, right?
它没法连接任何工具，也没有上下文。
It's not hooked up to any tools, has no context.
所以，真正的对话智能只是人们完成工作、实现目标所需能力的一部分。
And so it really is that this conversational intelligence is only one part of what people really need to get work done to be able to achieve their goals.
而我们的方向是打造一个A。
And where we're going is to have an A.
I。
I.
它真的会时刻为你着想，对吧？
That's really looking out for you, right?
你能为它设定目标和方向，让它持续为你思考。
That you can provide the goals, the directions that it's constantly thinking about.
今天我能为亚历克斯做些什么，让它能帮你解决那些特别棘手或特别琐碎的问题？
What can I do for Alex today that it's able to go and solve super hard problems, very mundane problems?
你早上醒来，邮箱已经被整理好了；如果你在考虑健康计划，它也能帮你推进，比如查找医疗方案，或者来回给你提供相关信息，至少能做到这些。
You wake up, your inbox is organized, but also if there's like a health plan that you are thinking about, that it can help you help you achieve that, figure out medical treatments or, you know, sort of back and forth, provide you with that kind of information, at least.
我们花了很多时间思考的问题是：你希望的界面是什么样的？你想要的产品到底是什么？
And I think that the question of what's the interface you want, what is the product that you want, is what we spend a lot of time thinking about.
答案是，你希望几乎没有任何界面。
And the answer is you want almost no interface.
你希望根本感觉不到产品的存在。
You want no product.
对吧？
Right?
你希望这个东西就像你和我之间的自然互动一样。
You want this to be like, what's the interface between you and me?
没错。
Right.
就是能跟一个持续存在的实体交流，它能帮你完成各种目标。
Just being able to talk to a persistent entity of some form that's able to go and accomplish goals for you.
要构建这样的系统确实不容易。
And so building that is hard.
这需要时间，但我们已经有不少关键部件了，对吧？
It will take time, but we have a lot of the pieces, right?
我们正在不断整合产品层，努力提升模型能力，让整个系统越来越少依赖点击按钮、切换开关和切换模式这些操作。
And we're increasingly bringing together the product layer, trying to make the models better, trying to make the whole system just so there's less like clicking buttons and toggles and changing modes and all these things.
并不是说过程中完全不会有这些元素，但长期来看，趋势是简化和整合。
Not to say that there won't be some of those along the way, but the long term trajectory is towards simplification, unification.
是的，你说界面会慢慢消失，这点很有意思。
Yeah, it's very interesting that you say the interface, will melt away.
所以我想再深入问一下，今天很多使用像ChatGPT这样的产品的人会发现，机器人最后总会提个建议。
And so to go a little bit deeper with my question, many of us who use products like ChatGPT today will see that the bot will make a suggestion at the end.
你问它关于营养的问题，它会反问：要我帮你制定一个健康计划，还是饮食计划？
You you ask it about nutrition and it says, should I make a health plan for you or make a diet plan for you?
你问它——抱歉，Ranjan，我们刚聊过旅行——但你一问旅行的事，它就能给你出一份行程安排。
You ask it, sorry Ranjan, we just talked about travel, but you ask it about travel and it will give you an agenda, for instance.
所以，我理解得对吗？在ChatGPT里接下来要发生的是——
And so, am I hearing you right that what's going to happen within ChatGPT, just
到
to
举个例子，你跟它聊你的健康决定，它可能会说：你可能得去看看某个专科医生。
give an example, is you talk to it about your health decisions and it might say, you know, you probably need to go to this specialist.
我来帮你预约一下。
Let me make an appointment for you.
然后它真的会替你去完成这件事。
And then it will go and actually take that action on your behalf.
所以，它不再只是个对话界面，而是能理解你的意图，并主动帮你把事情办成。
So it goes from simply a conversation interface to actually understanding your intent and then going out and accomplishing that for you.
没错，你说得完全对。
That that I that's exactly right.
而且我觉得，如果你用过柯达，顺便问一下，现场有多少人用过Codex？
And I think that if you've used Kodak and by the way, how many people in the room had used have used Codex?
很多人用过。
A lot of people.
是的，有不少人用过。
Yeah, decent number of people.
我们的目标就是让每个人都能享受到Codex的强大功能，对吧？
And that our goal is to really bring the power of Codex to everyone, right?
让智能代理普及到每个人。
To bring agents to everyone.
这项技术现在就已经存在了，对吧？
That technology exists right now, right?
你可以像我一样，把Codex连接到Slack、Gmail和日历上。
You can hook up like I hook up my Codex to Slack, to my Gmail, to my calendar.
在OpenAI内部，有很多非技术人员也在使用它。
And there are many people within OpenAI non technical users.
名字里有'代码'，但其实跟编程没太大关系。
You know, it's got code in the name, but it's not really about code.
它真正厉害的地方，是作为一个通用工具，通过智能代理帮你完成各种任务。比如我们公关团队有个人，之前在组织一场活动，她就让代理自动询问所有参会者的饮食偏好，安排座位表，把琐碎的事全办了，她自己就能专心思考活动的整体构想和目标。
It's really about having this general purpose tool using harness an agent and the kinds of things, for example, someone on our comms team does is she was organizing an event and it would just ask all of the event attendees for their dietary preferences, set up the whole seating chart, kind of did all of that work so that she could focus on the parts that she wanted to and really thinking about the vision of what she wanted to achieve.
我觉得这种趋势会很快普及到各个领域。
And I think that we're going to see this across the board.
现在再想AI连接各种工具，已经不是科幻了。
So the it's not sci fi anymore to think about an AI that's hooked up to these tools.
我记得我们最早在ChatGPT里尝试工具使用，是2023年的事。
And I remember with, you know, our very first attempt at tool use in chatty was 2023.
大概是三月或四月左右，我们推出了插件功能。
I think in like March or April or something, we released plug ins.
大家还记不记得早期ChatGPT时代的那些插件？
Do people remember plug ins back in early chat days?
那根本行不通。
That didn't work.
完全不行，因为当时的模型还没准备好。
It didn't work at all because the models the models weren't ready.
对。
Right.
这种使用方式的方向是对的。
The form factor is correct.
很明显，未来AI跟你的Gmail对话会像喝水一样自然。
Obviously, you're going to have an AI that's able to, like, talk to your Gmail like no question.
但那时候我们一次只能让模型接入三个连接器，而且上下文记忆特别短，可能就两千到四千个标记。
But we could only like have three different connectors exposed to the model at a time or start forgetting, you know, we had like two k, maybe four k token context.
根本没什么记忆能力，对吧？
Like there's just no memory, right?
这就像是六七十年代的早期计算机一样。
It's kind of like when you had early computers in the sixties or seventies or something, Right.
你知道，那点可怜的内存空间。
You know, the tiny little memory banks.
而今天，你的手机性能早就超过了当年的任何超级计算机。
And today you have your phone that's like better than any supercomputer from that era.
我觉得AI的发展方向也正是这样，对吧？
And I think that's where we're going with these models, right?
进步的速度实在太快了。
The rate of improvement has been so steep.
现在，你可以接入上百种不同的工具，甚至能连上整个文件系统。
So now you can have hundreds of different tools accessible that we have the ability to hook them up to whole file systems.
几乎能把整个互联网，还有你想要的任何应用，都放在模型的手边。
You can almost have the full power of the Internet and like almost any application you want at the model's fingertips.
而且它还很聪明，对吧？
And it's smart, right?
它的上下文容量能达到五亿一千万个token，看你从哪个角度看。
It's got $512,000,000 token context depends how you squint on it.
而且这些模型的能力也越来越强大了，对吧？
And the capability level is also getting so, so powerful, right?
现在这些模型都能解决那些一直悬而未决的数学和物理问题了，对吧？
These models are now solving unsolved math problems and physics problems, right?
真正帮人们实现了以前根本做不到的事情。
Really helping people be able to achieve things they couldn't otherwise.
我们正站在一个新时代的门槛上，智能代理将彻底改变我们工作的方式——无论是软件工程、金融、法律销售，还是个人生活方面。
Like we are on the air, on the edge of this era of agents really transforming how we all operate, whether it's in software engineering, finance, legal sales and in our personalized too.
所以，你刚才举的那个例子，我来具体说说。
So just to unpack that example that you
你的一个同事正在和ChatGPT聊一个活动的事，然后问：‘咱们该怎么联系参会者呢？’
were giving there, one of your colleagues is chatting with chat GPT about an event and then suggest, hey, how should we contact event attendees about something?
结果根本不用自己动手去翻活动系统，只要模型觉得这个主意不错，你一同意，它就会自动接管，连上你用的工具，直接帮你把事办了。
And instead of like saying, okay, I have to do that and going into an event program, basically what happens is the the interface will take over from there once it says it's a good idea and you agree and then hook into whatever tools you're using and then do it for you.
没错。
Exactly.
它会通过连接 Gmail，帮你翻遍收件箱，找出所有参会的人。
So it uses, you know, Gmail connector searches through your inbox to find all the people who are attending.
然后如果你设置了饮食限制之类的选项，它就会发现：哦，这些人已经填过饮食要求了。
And then if you're on the like, what what's different dietary restriction sees, oh, these people already have their dietary restriction.
至于那些没填的，它会根据你的设置，自动起草一封邮件。
These people I do not drafts an email depending on exactly how you have things set up.
它可能会说：我刚写好了这几封邮件。
It might say, hey, I drafted these emails.
我可以发出去吗？
Can I send them?
如果你用的连接器不允许它直接发邮件，那它就会等你手动确认后才发。
If you have a connector that doesn't even let it send emails that I drafted, you need to send them.
在另一个场景里，如果你对系统足够信任，它可能会直接说：邮件我写好了，已经发出去了。
And I in a different world, you could also imagine that it's you built enough trust with the system where it says I drafted the emails and actually sent them.
我觉得这其实凸显了智能代理时代一个非常关键的点——那就是信任。
And I think that this actually points to a really important aspect of the agentic era, which is trust, right?
我们需要真正学会如何在这些系统表现好和表现不好时，都建立起对它们的信任。
That we need to really learn how to build trust with these systems, where they're good, where they're not.
弄清楚哪些事情该交由它们处理，以及你愿意让它们承担多大的责任。
Figure out what you want to delegate to them and how you want to entrust them with responsibility.
而我们认为，这种信任是靠积累得来的，对吧？
And that's something we view as earned, right?
不是随便就能赋予的。
It's not something that we can grant.
但如果我们为操作者——也就是这个AI所服务的人——提供丰富的工具、控制权、监督和管理，我们认为这会非常关键。
But by providing lots of tools and control and oversight and supervision to the operator, to the person who that this AI is operating on behalf of, like, we think that that is going to be such an important thing.
所以，这是一项核心的产品功能和差异化优势。
So that's a key product feature and differentiator.
是的。
Yeah.
当你回看早期的一些尝试时，比如OpenAI曾让你在ChatGPT里直接叫Uber，这背后其实有一长串公司都想让你在聊天里完成行动，但都没真正成功。
And when you go back to some of the early attempts at this, there was this like move that OpenAI had to let you call an Uber within ChatGPT and it followed a long line of companies that have tried to get you to take action within chat, but it never really took off.
这里的不同在于，聊天机器人可以直接控制你的浏览器或电脑，这样你就不用再担心某个插件能不能用了。
And the difference here might be that the chatbot can take control of your browser or take control of your computer and then you don't necessarily have to worry about like is this plug in going to work.
它会直接接管你的设备，帮你把事情办妥。
It goes and accomplishes that for you by taking over your machine.
所以我在想，面对现在这些主流的用户界面——也就是其他所有应用和软件——ChatGPT要想真正有用，会不会遭遇阻拦？毕竟它得能自由地替用户执行各种操作。
So I wonder, you know, if you expect a fight from the user interfaces that we have today, aka like all the other apps, all the software, where to be truly useful ChatGPT will have to not be blocked to be able to go out and execute these actions on behalf of a user.
首先，我觉得这已经不是理论上的事了。
Well, look, first of all, I'd say that it's this is not theoretical at this point.
对吧？
Right?
早就有人在用代码助手了。
That people have been using codecs.
那是个独立的产品，独立的应用程序。
So it's separate product, separate app.
你得单独安装，而且它主要聚焦在软件开发上。
You have to install it separately, really started to focus on software engineering.
但非软件方面的工作量也在急剧增长，编码工具的使用量简直呈爆炸式上升。
But the amount of non software work that has been happening, codecs have been absolutely exploding.
对吧？
Right?
这波增长曲线太惊人了，完全符合我们的预期。
It's been this like incredible exponential curve, exactly the thing that you you would expect.
而有了OpenAI，现在ChatGPT的普及率和使用频率，就跟Slack一样了。
And with an open AI, we basically have the same level of penetration now and usage of Slack, right?
在OpenAI内部，几乎所有人都完全依赖Slack工作。
It's like everyone in open eyes like an entirely Slack based company.
我们平时基本不用邮件了。
We do not use email for the most part.
如果你不在Slack上，那就等于没法干活。
It's like really like if you're not on Slack, you're not going to do any work.
现在Codex应用也渐渐有了这种趋势，几乎每个人都在把Codex接入各种工具。
And it's kind of feeling that way now with Codex app as well and that everyone's Codex is hooked up to all of these tools.
生态系统会怎么发展，我觉得会是个很细致的问题，因为关键在于，我们坚信应该打造一个充满活力、蓬勃发展的生态，让每个人都能真正参与其中并获得回报。
How the ecosystem ecosystem evolves, I think it's going to be a very nuanced thing because I think one thing that is very important is that we believe that there should be an ecosystem that gets to be vibrant and thriving and that people can really build and see the benefits.
我们确实从合作伙伴那里看到了这种趋势，我记得有几家合作伙伴，我们跟他们说：我们很想训练我们的AI，让它能熟练使用你们的软件。
And so we've actually seen this from partner companies where, you know, that we I remember there's a couple of different partners where we said, hey, we really want to train our AI to be really good at using your software.
当时我们也不确定他们会怎么回应。
And we didn't know what they would say.
结果他们说，这是他们见过的最友善、最支持合作伙伴的沟通方式。
And actually, the response we got is this is the most partner friendly outreach we've ever had.
是的。
Right.
他们觉得，你们愿意专门优化AI来适配他们的工具，这简直是天赐的机会——因为他们的产品使用率会因此大幅提升。现在大家都在想，怎么才能不只在AI时代活下来，而是真正腾飞。
That the idea that you will make your AI specifically good at using our tool and they just see the opportunity because their tool be used just so much more as a result and that everyone is trying to think about how did they not just survive as a company into the air, but thrive?
怎么才能真正抓住这个活动量激增带来的优势？
Like, how do you really get the advantages of the fact there's going to be so much more activity?
如果你不拥抱AI，反而把它拒之门外，那你的处境只会越来越糟，不可能兴旺起来。
And if you don't have AI in there, if you shut it out, then you're actually going to be declining, not thriving.
没错。
Right.
这实际上让OpenAI站到了一个全新的位置。
This kind of makes OpenAI puts OpenAI.
首先，我们来聊聊人们使用Codex的情况。
So first of all, you're going to bring you talk about people using codex.
你的一位同事提到过，你也谈过，你们已经把ChatGPT整合进Codex，反过来也能把Codex嵌入ChatGPT——也就是说，作为ChatGPT的用户，你们之前说的那种体验：它不只是建议你下一步该做什么，而是直接帮你完成，这种事马上就要实现了。
So one of your colleagues shared and I think you've talked about this too that you've brought ChatGPT into codex so you can bring codex into ChatGPT, which is is basically like if we're users of ChatGPT, this experience that we talked about of ChatGPT not only suggesting what you might want to do next, but going to do it for you, that's going to happen.
这样一来，你就相当于成了一个操作系统。
And so it makes you effectively an operating system.
想想看？
Think?
但不是像iOS那样的操作系统，你不会打开手机再去点不同的应用。
But not the operating system like an iOS where you would like go open up the your phone and then tap different apps.
更像是所有应用的交互都会通过这个界面来完成。
It's almost as if all interaction with all apps will happen through this interface.
这就是你们的目标吗？
Is that the ambition?
我觉得这么说也可以，但我对它的理解有点不一样。
I think that you could describe it that way, but I think of it a little differently.
我想知道，一个通用人工智能的理想交互界面应该是什么样的？我们称之为个人通用人工智能。
Like, the way that I think about this is that what is the ideal interface to an AGI, what we call it kind of a personal AGI?
我觉得，其实答案就是我们现在正在用的这种界面。
And I think that it's again the same interface that you and I are using right now.
你只是想跟一个助手聊天，对吧？
You just want to talk to an assistant, right?
你希望有个东西能替你去办事、去操作。
You want to talk to something that can go and and work and operate on your behalf.
所以，没错，这个智能体、这个通用人工智能，它确实会拥有自己的计算机。
And so that, yes, like that agent, that AGI, that AI will have its own computer, right?
它还能访问各种资源，就像理想的同事一样，能直接过来帮你操作电脑。
It'll have its own access to things that maybe can, you know, like ideal coworker would be they can come over and type things on your computer, too.
所以需要一些对您系统的访问权限，以及授权它部分操作。
So some access and delegated access to your own own system.
有时候，您可能会把收件箱的访问权交给它。
And, you know, maybe you delegate access to your inbox sometimes.
它自己也可能有个收件箱，用来查看它需要处理的信息。
Maybe it has its own inbox with some sort of some sort of window into into the things that it needs.
您会把邮件转给它处理。
You forward emails to it.
其实仔细想想，这并不算新鲜事，对吧？
These are not actually if you think about it, this is not unprecedented, right?
这就像您和真人助理，或者任何同事合作的方式——我们早就习惯如何建立信任边界，确保彼此能顺利配合了。
It's like the way that you work with an assistant who's a person that we've we've actually or any coworker, really, we've spent a lot of time really thinking about how do you build these trust boundaries and make sure that you're able to operate together.
所以我觉得，这只不过是一种不同的形式。
And so I think of it as just a different thing.
它并不是什么全新的东西。
It's not it.
你可以把它看作一个操作系统，但操作系统这种东西，感觉像是上个时代的产物了。
You could think of it as an operating system, but an operating system is almost something from a different time.
这是栈的另一个层面。
Different layer of the stack.
这其实更关乎你如何与整个技术世界互动。
This is really more about how do you interface with technology broadly?
我觉得AI最棒的地方在于，它让机器更贴近人，而不是逼着我们去适应文件夹、路径这些不自然的玩意儿——那些东西反映的是机器的逻辑，不是人的思维方式。
And I think that the beautiful thing about AI is it's really about bringing the machine closer to the human rather than us having to contort ourselves into like files and folders and like all of these details that somehow are not natural, That are more about how the machine operates rather than how we operate.
是的。
Yeah.
说到个人智能，我也不太好说。
Talking about a personal intelligence, it's sort of I don't know.
你上周看WWDC了吗？
Did you watch WWDC last week?
没有。
No.
我没看。
No.
错过了。
I missed it.
我本来被禁了，但还是在电视上看了。
I was I was banned, but, I watched it on TV.
拜托，苹果公司。
Come on, Apple.
不管怎样，你和新版Siri看起来要正面竞争了。
Anyway, it does look like you and and and Siri, the new Siri are gonna come kind of into competition.
对吧？
Right?
因为Siri会作为一个智能体嵌入到你所有应用之上，让你直接操作；而ChatGPT只是iPhone上的一个普通应用。
Because they're an app that's going to sit or an intelligence that will sit on top of all of your apps, and let you take action and ChatGPT will be an app on the iPhone.
所以，你能不能谈谈，这种定位对OpenAI会不会构成挑战？你们是怎么从战略上考虑的？
So then talk a little bit about whether that positioning is gonna be difficult for OpenAI and how you're thinking about that strategically.
我只是想换个角度再说一遍。
Well, I just think, again, to give it a little differently.
我觉得我们正处在智能代理时代的开端。
Like, I think that we're in the beginning of this new agentic era.
在人工智能的发展中，每当出现新的能力水平，就意味着有机会重新思考一切。
And the way that this has always gone in AI is that when you have a new level of capability, it means you have an opportunity to rethink everything.
对吧？
Right?
重新思考人机交互的方式，以及技术到底能做什么。
Rethink how people interface, how like what the technology is capable of.
我觉得这次也不例外。
And I think that this is no no different.
没错。
Right.
在我看来，比如我所看到的未来趋势，首先是A。
In my mind, mind, like the kinds of things that I see on the horizon, for example, A.
我指的是
I.
用于解决科学问题。
For solving scientific problems.
对吧。
Right.
我觉得我们已经开始看到一些苗头了。
And I think we're starting to see the inklings of this.
比如今天，我们宣布有医生在同行评审的期刊上使用了三个或三个这样的系统。
Like, for example, today, we announced have in peer reviewed literature, people, doctors who are using three or three.
没错。
Yep.
那都是很久以前的事了。
It was like forever ago now.
对吧？
Right?
这是最早期的推理模型之一，用来诊断那些医生多年都束手无策的病人。
It was like one of our earliest reasoning models using that to find diagnoses for people who I had no no answers from doctors for many, many years.
比如有个人，二十年来一直被一种不明病症困扰。
You know, there's an example of someone who had spent twenty years with a mysterious ailment.
最终，靠着这项技术，终于找到了病因。
Finally, it's been diagnosed through the use of this technology.
如果你觉得，好吧，这些模型确实能做到这些，那又怎样呢？
And if you're like, Okay, you've got models that can do that, they can do that.
关键在于，同样的问题——你能不能用上这个应用？
And then it's really about like, you know, the same like distribution and like, you know, can you get access to an app?
对我来说，这种说法根本站不住脚。
You know, to me, it's it doesn't type check.
我们面对的，是一种根本性的新事物。
It's like we have something fundamentally new.
这并不意味着未来不会有竞争。
And so that's not to say that there won't be competition.
我倒觉得肯定会，而且这对每个人都会是好事。
I actually think that there will be and it's going to be great for everyone.
但我认为，你使用这项技术的方式、它能让你做到的事情，跟我们以前见过的任何东西都完全不同。
But I just think that the ways in which you're going to use this technology, the things that will be capable of and what it'll make you capable of doing are just totally different from anything we've seen before.
我本来想问你，是不是得自己开发设备？我的想法是，你得通过苹果才能接触到用户，不知道这个假设靠不靠谱。
You know, I was going to ask you, well, does it mean that you'll have to, you know, create your own device, assuming that, like, my concept is is, you know, that you're going to to go through Apple to get to the user, assuming that's somewhat valid.
但事实是，你其实已经在做了。
But the answer is you already are.
对吧？
Right?
OpenAI 现在就在开发一款设备。
You're so OpenAI is working on a device right now.
这事儿已经公开报道过了。
Certainly has been publicly reported.
去年十二月我去你办公室的时候，萨姆就告诉我这事已经在进行了。
I I was in your office December and Sam told me that this is happening.
这不止是一个设备。
It's multiple devices.
所以，如果你想想你将来会怎么跟这些AI互动，这个设备——或者这一系列设备——扮演什么角色？
So, if you think about the way that again you're going to interface with, with these AIs, how does that device play in or series of devices?
说真的，咱们不妨退一步看，我觉得这只是一个全新时代的开端。我觉得最关键的转变其实根本不是关于设备本身。
Well, look, think again, I I would just step back and say that I think this is the beginning of something very new and that I think about the way that I think I want I want you to say, like, I think the biggest shift that has happened in terms of interface again, it's not even about devices and things like that.
真正重要的是从对话式智能的转变，就像聊天这种模式——现在AI已经足够了解你，让你愿意认真读它的回复。
It's really about the shift from conversational intelligence, like kind of the chat paradigm where it's like kind of you have an AI that's personalized enough to you that it's worth reading its output.
对吧？
Right?
你问它问题，它给你答案。
You ask it a question, you get an answer.
从这种能帮你回答问题的AI，变成能真正替你做事的智能代理，这才是巨大的变化。
It's something that's useful to you to agents where they're capable enough to actually do things for you.
这种转变意味着，你与它的互动方式也必须跟着改变。
Like, that is a big shift and that implies a difference in how you want to interact.
所以，你只会希望有一个能掌握你所有背景信息的智能代理。
And so you kind of are just going to want a single agent that has access to your context.
这一点在个人生活中也是如此。
And this will be true in personal life.
在工作场景中也同样适用。
This will be true in a business context.
对吧？
Right?
你可以想象一下，假设你有一位在每个领域都拥有博士学位的同事，
You imagine, for having a you know, imagine you have a Ph.
拿过好几个诺贝尔奖，
D.
你雇了这么一位，或者雇上一百位，但从来不让他们参加任何会议，
In every field coworker, you know, Nobel Prizes, multiple of them, and you hire one of these, you hire 100 of them and you don't invite them to any meetings.
那他们根本发挥不了什么作用。
They're not going be very useful.
所以问题来了，你怎么把上下文真正融入到AI里头？
And so there's something about how do you get context into the A.
I.
I.
不是静态地给，而是动态地更新，对吧？
Not just statically, but dynamically, right?
随着上下文变化，随着你的业务流程不断演进，你怎么构建一个能让AI随时访问的上下文层？
As context evolves, as your business processes evolve, how do you have a context layer that is accessible to an A.
I.
I.
让AI能充分发挥它原始的智能潜力？
That lets the AI operate to the extent of that raw intelligence?
所以，要想让AI在会议中随时可用，用起来顺手、触手可及，这背后需要一套全新的设计。
And so finding ways to make that AI be accessible, so available in your meetings to make it very ergonomics, it's very easy to get access to.
我觉得，这一切都得重新思考。
I think all of that's going to require a rethink.
但我认为，归根结底，关键在于先从智能代理的形式入手，再反过来思考怎么让它获得所需的上下文。
But I think again, it's just the core for me starts from thinking about the agentic form factor and then working backwards to how do you just make this have the context it needs?
而且信任问题，
And again, the trust is going to
将是让整个系统运转起来的核心。
be such a core part of making this whole equation work.
就像随身带着一个设备，随时想做点什么，它就能自动帮你搞定。
So it kind of like having this this device with you at all times and being like, I need to get that done and it goes and does it for you.
我觉得这确实是其中一部分。
And I think that that will be part of it.
但即使你没有这样的设备，也不意味着你就被排除在外了，对吧？
But I almost even if you don't have a device like that, it's not like you're going to be out of the game, right?
因为这个AI并不是说非得有一个单一的形态，好像只有设备本身才是AI。
Because it's this this AI is not because there's one thing, there's one version of it where you think of it where it's like the device is the AI.
你不会指望你的手机就是AI本身。
You want your phone to be the AI.
你希望的，不管是什么样的定制设备，都应该是那个AI本身。
You want, you know, whatever whatever, you know, custom device you're you're thinking about to be the AI.
但实际情况不会是那样。
But it's not going to be like that.
它更像是一个界面，就像你的手机只是你的一部分一样。
It's going to be like an interface like no more than your phone is you.
对吧？
Right?
它只是你与AI互动的接口。
It's an interface to you.
它让你能随时召唤我，想问问题的时候就能找到我。
It's a way that I can sort of, you know, call you up whenever I need you, whenever I want to ask you a question.
接入的方式其实有很多种，对吧？
And there's different ways of accessing, right?
比如同步的电话通话。
There's like synchronous phone call.
我可以给你发短信。
I can text you.
我可以给你发邮件。
I can email you.
我觉得我们会逐渐变得
And I think that we're going to
和我们与智能助手互动的方式非常相似。
be much the same with how we interact with our agents.
有报道称，OpenAI 正在开发双向语音模型。
There's been some reports that OpenAI is working on these like bidirectional voice models.
我们以前好像聊过这个。
I think we've talked about that in the past.
目标是打造一种 AI，你可以和它自然对话，它也能理解你的话，并以更自然的方式回应你。
Like the goal is to have like an AI that you can you can speak with and will be able to process that and speak back with you in a much more natural way.
关于这个，你能透露一些信息吗？
Can you share anything about that?
没有。
No.
不过说正经的，我是说，你看，我觉得这项技术的整体形态，也就是我们拥有的语音模型的方式，你知道的，这种非常酷的语音体验，到现在大概有一年半到两年的时间了。
But now more seriously, I mean, look, I think that the general shape of the of the technology, like the way that we had we had we've had voice models, you know, kind of a really cool voice experience for, you know, a year and a half, two years now.
你知道的，我们最早在2024年三四月做了演示，然后推向市场，大概是在那年年底。
You know, we first demoed it back in March, April 2024, brought it to market, you know, maybe late that late that year.
它的工作原理，以及所有模型的工作原理，基本上就是把它们串联起来。
And the way that it works and the way that everyone's models work is that you basically chained together.
其实最初这类系统的工作方式，是把语音转文字、文字转文字、再转语音这几个模型串起来用。
Well, the original way that these things worked was that you would chain together a text to a speech to text model.
先做文本到文本的处理，然后再用文本转语音，整个过程其实挺别扭的，对吧？
Then you do a text to text model, and then you would do a text to speech model horribleness, right?
就是把这三个环节硬生生地接在一起。
Like these three things chained together.
即便现在有了统一的模型，能同时接收输入并生成回复，但对话中的轮流发言问题依然存在。
It still has been the case that even if you have one unified model that's able to kind of take in input and then, you know, able to output a response, you still have this problem of turn taking.
对吧？
Right?
想象一下，我们得轮流说话，谁也不能插嘴。
Imagine that like we have this like you cannot overlap, you cannot interrupt.
你讲完一轮，必须等我完全说完才能接话——这根本不是人跟人聊天的方式。
It's just like once you you speak to me in a turn and then you gotta wait for me to finish my whole response, that is not how human conversation works.
所以我们现在用的其实是个临时凑合的办法：模型得判断‘哦，对方好像说完了’，或者‘哦，对方好像要开始了’。
And so that we we basically have, like, a hack where we have these models that determine, oh, it seems like the turn has ended, and, oh, it like seems the turn has started.
我们干嘛非得谈什么‘轮流’呢？
And we're like, why are we talking about turns?
没错。
Right.
这种轮流方式太不自然了。
Turns are getting so unnatural.
这其实是人在迁就机器，适应它的局限。
This is the humans contorting ourselves to the machine and its limitations.
所以，你自然希望打造一个AI模型，能像你我对话那样自然流畅。
And so the obvious thing that you want to accomplish is a model in AI that works much more like you and I do.
对吧？
Right?
它能一边接收输入，一边生成回应。
It's able to process input at the same time it's processing output.
当然，这个领域里很多人都在朝这个方向努力。
And all of that is, of course, something that many people in this field are trying to to run towards.
我觉得，一旦我们转向这种自然、流畅、像人一样对话的界面，那将会非常令人兴奋。
I think it's going to be very, very exciting as you move to these natural, very human fluid like conversational interfaces.
以前没人见过这样的东西。
No one's seen anything like it.
比如说，现在和ChatGPT语音交互的感觉，其实挺神奇的。
Like, one thing that that I think about is the the current interaction with, you know, chat GBT voice in many ways is magical.
对吧？
Right?
很多人在通勤时用它，能问各种问题，但同时也特别让人抓狂。
So many people use it on their commute, able to ask all these questions, but it also is so frustrating.
对吧？
Right?
一旦它打破这种魔法，你就发现：我想接着问个问题，它却一直插话，根本听不完我说的。
Whenever it breaks the magic because it's like you realize, oh, I want to, like, add some follow-up and it keeps talking over you and it didn't it's just like that.
这完全说不通。
It's just it doesn't make sense.
所以我觉得，我们真正需要的，正是这种能自然流畅互动的AI。
And so I think that part of what we need, part of, like, the whole point of this AI is to be something that you can interact with interact with fluidly and naturally.
顺便说一句，这不只是关乎个人使用场景。
And, by the way, I think it's not just going to be about the sort of use case.
我们常想到的是个人用途，但工作场景同样重要。
Like, we we kind of think about the the personal use case, but it's also really the work use case.
我觉得自己用Codex时最神奇的体验，就是像很多人那样通过语音来操作它。
And I think some of the most magical experiences that I've had with Codex have been when operating it through voice like many people.
我们内置了语音功能。
We have a voice thing built in.
有些人用第三方应用，但一旦你意识到，打个简短消息反馈挺方便，可要写一整段话、把所有想法都表达出来，那感觉就完全不一样了。
Some people use third party apps for it and you just get a very different experience when you start to realize that, like typing a quick message to give some feedback easy, but like writing out a whole paragraph and everything you want.
太糟糕了。
Horrible.
没人愿意这么干，对吧？
No one wants to do that, right?
你就是想随口说说，然后立刻得到反馈，这些都会实现的。
You just want to be like saying things and you want the real time feedback loop and all of that is going to happen.
那会非常棒。
It's going to be amazing.
我们简单聊聊模型的改进吧。
So let's talk about model improvement briefly.
几年前有人讨论过，大型语言模型快要遇到瓶颈了。
So there was a discussion a couple of years ago that large language models were about to hit a wall.
展开剩余字幕（还有 341 条）
那完全是错的。
That was wrong.
我现在一直在想，我觉得大家也都在想，这些模型还能进步到什么地步？改进什么时候才会停下来？
And something that I'm thinking about is I think we're all thinking about it is how much better can these models get and when will the improvement stop?
你们怎么看？
Any thoughts?
我觉得，当你在实际构建这些模型的时候，
Well, I think that this is a place where when you're kind
会有一种直觉，这种感觉外人很难体会到，因为我们能看到所有的数据，也能看到为这些改进所付出的努力。
of building these models, you get kind of a sense and an intuition that I think is harder to get from the outside because we see all the data points and we see also the work that goes into these improvements.
所以这个问题有两个方面。
And so that there's two parts to the answer.
一方面，我觉得这背后的基础科学，是我所知道的最神秘、最重要的科学发现和实证观察之一。
One is I think that the fundamental science is one of the most mysterious and important just scientific discoveries and empirical observations that I can that I that I'm aware of that I can imagine.
对吧？
Right?
我们居然真能建成这些模型，而且规模定律还在持续生效。
That we are able to actually build these models and that the scaling laws continue.
对吧？
Right?
只要你不断训练，增加数据、算力，优化架构，性能就一直能提升。
That it just is the case that you can just keep training these models, more data, more compute, better architectures.
这其中的改进其实非常多。
And there's a lot of improvements that go in.
但每次我们遇到预期之外的 scaling 问题，其实都是出了岔子。
But every time we've kind of run into a like, oh, this isn't quite scaling the way we expect, it's we have a problem.
要么是代码有 bug，要么是数学推导不够严谨，或者实现和理论对不上，诸如此类的问题。
We have a bug that our math wasn't quite right, that, oh, our implementation isn't isn't isn't quite matching the math, whatever the thing is.
我觉得这一点特别重要，得真正理解透。
And that is, I think, a very important thing to to sort of internalize.
事实上，如果我们回溯到这个领域的起点，就会发现神经网络早在上世纪四十年代就已提出，那时候连计算机都还没普及呢。
And, actually, if you we've done studies where you go back to the beginning of the field, right, that neural nets themselves were designed in, like, nineteen forties, right, before computers.
对吧？
Right?
作为一种可能解释大脑如何处理信息的模型。
As a model of maybe this is maybe this is how the brain processes information.
第一个硬件实现是1959年的感知机。
First hardware implementation was 1959 with the Perceptron.
如果你看看这个领域的里程碑式成果，就会发现它们都遵循着一条极其平稳、确定的路径——不断投入更多算力。
And if you look at landmark results in the field, that the landmark results follow this incredibly smooth deterministic path of more compute being poured into them.
所以七十年，甚至快八十年了，一直有人断言这些东西永远不可能成功、永远不可能扩展、迟早会撞上墙，可墙到现在都没出现。
And so seventy years of people, maybe eighty years now, of people saying this stuff is never gonna work, never gonna scale, going to hit the wall, hasn't hit the wall yet.
目前根本看不到任何障碍。
There's still no wall in sight.
所以我认为，根本原理是支持这一点的。
And so I think that the fundamentals allow it.
至于实际操作的难度，那是另一回事。
Now the practicality is hard.
对吧？
Right?
实际上，建造这些超级计算机非常困难。
Actually, building these massive supercomputers, it's hard.
成本很高。
It's expensive.
并不简单。
It's not easy.
对吧？
Right?
我们有团队在拼命解决这些极其复杂的技术难题。
That we have teams that just, like, work so hard to solve these incredibly hard technical problems.
我们甚至得自己设计网络协议，团队里有人会深入检查每一层架构，因为图上总会出现一些异常波动。
We have our own network protocol that we've had to design, that we have people who look at every single layer of the stack, that there's weird wiggles in the graph.
理解这些神经网络时，你要明白，根本没有什么抽象层可言。
And you the way to think about these neural nets is that there's like there's no abstractions.
对吧？
Right?
只要有一点出错，就可能引发连锁反应，最后才显现出来。
It's almost like any little piece that's wrong can have a ripple effect that only shows up down there.
所以你需要有人彻底理解整个系统。
And so you need people to deeply understand all of it.
但只要你凑齐了对的团队，给大伙儿明确的目标，大家肯埋头苦干，结果绝对值得。
And yet, if you get the right team together, put the right mission in front of people, and people do that grind, the outcome, it's worth it.
对吧？
Right?
这事儿办得到，也做得到。
And it's achievable and it's possible.
所以我觉得，基于这些原因，进步还会继续下去。
And so I think that for those reasons, the progress will continue.
那我想听听你的看法，模型离现在的水平还能再往前走多远？
So then I'd love to hear your perspective if if models can basically progress much further from where they are today.
假设OpenAI做出了最好的模型，就像一个拥有十五个博士学位、情商极高、从不抱怨、还能主动替你办事的人。
Let's say OpenAI builds the best model and it's the equivalent of like something with like 15 PhDs with excellent emotional intelligence that doesn't complain and goes out and does stuff for you.
接下来的模型制造者会做出差一点的，比如十三个博士学位，情商也不错，照样能帮你做事。
And then the next model maker will build a less good, but it has 13 PhDs and it's like pretty good EQ and will still go and do things for you.
但当所有模型都达到这种智能水平时，差别到底在哪里呢？
So where does the differentiation come in when you get to that level of intelligence?
我们看到这些模型制造者几乎总是步调一致。
We've seen the model makers kind of move in lockstep.
一个刚有突破，下一个马上就跟上，也做出同样的突破。
One makes an advance, the next one comes in and makes the advance.
结果就是，大家都变得一样聪明了。
So they all become that smart.
那还能不能做出区别来？有可能吗？
Do they is it possible to differentiate?
我觉得这个问题可以从好几个角度来回答。
Well, I think there are several dimensions to the answer.
首先，我认为确实存在一种供不应求的状态，从商业模式来看，每个提供商的计算资源都已经被抢购一空。
Number one is I do think there's a bit of an detractor state where just like from a business model perspective, every provider sells out all their compute.
是的。
Okay.
没错。
Right.
我觉得我们正朝着这样一个方向发展：计算资源根本无法满足所有需求。
I think that is just like the world that we're heading towards where there just is not going to enough compute to serve all the demand.
没错。
Right.
我们正在进入一个由计算驱动的经济，每个人都会持续使用这些模型来完成各种重要任务。
That we're heading to this compute powered economy, that everyone's going to be using these models all the time to be able to accomplish tasks of interest.
这一点我们已经看得清清楚楚了。
And we just see it.
现在我们讨论的，就是计算能力的瓶颈问题。
It's like right now we're talking about compute constraints.
现在使用这些智能代理的人数，大概有几千万，甚至上亿吧。
And like, the number of people using these agents is, like, order of 10,000,000, 20,000,000, maybe.
你明白吧？
You know?
我们离全球规模还差得远呢。
It's like, we're not a planet scale.
ChatGPT 都已经有十亿用户了。
ChatGPT is like a billion users.
对吧？
Right?
但我们还没把这种智能代理的能力普及到那个程度。
But we haven't brought the agentic power there yet.
所以你看这些因素，当前的使用深度，跟未来相比简直微不足道。
So you're just looking at these factors and the depth of usage also tiny compared to where we're going.
因此我觉得，不管是什么供应商、什么能力水平、开源模型，还是这些新兴云平台，未来真正稀缺的，终究还是计算资源。
And so I think that we're just going to be in a world where even if you have different vendors, different capability level, open source models, all these things, these neo clouds, like, I think the computers are just to going be the scarce resource.
我觉得这事儿肯定会迅速铺开。
And I think that it's going to go go go to use.
所以某种程度上，你觉得这行当值不值得进入？新玩家有没有机会？
So to some extent, I think that the the like, is this a good business to be in and for new entrants to come into and things like that?
我的答案其实是肯定的。
My answer is actually yes.
我觉得市场大得惊人，我们现在连皮毛都还没摸到，需要更多动力和投入。
I think that there is like a huge market that we are just not going to be able to address, and we need much more energy and momentum there.
但第二点是，人们忽略了智力并不是单一维度的。
But the second thing is that it also misses the fact that intelligence is not a unidimensional thing.
对吧？
Right?
如果你仔细看，擅长不同领域是另一回事，就算你天资再高，也不代表能立刻上手。
That that if you really zoom in, being being good at different domains is something where even if you're you have a lot of raw intelligence getting good.
如果你从来没练过，比如从来没做过产品演示，第一次肯定做不好。
If you've never practiced, like, you've never actually done a pitch or something, like, you're not gonna be good at it your first time.
对吧？
Right?
而且很多人根本没操作过电子表格。
And that there's lots of different never operate a spreadsheet.
对吧？
Right?
你不可能在做复杂建模这类事上取得成功。
You're not going to be able to succeed at doing some complex modeling.
所以我觉得，我们一直潜移默化地接受了一个观点：面对不同行业和领域，我们必须有所取舍。
And so I think that there is something that we have been internalizing, which is that we look across different industries and different domains, and we have to prioritize.
我们不可能同时在每一个领域都做到顶尖。
We can't possibly be great at every single area at once.
确实有很多人会说，
There is definitely a lot of, like, hey.
只要把通用智力提升上去，这些事自然就能应付了。
You just get the general intelligence up and it'll experience a lot of these things.
但要真正成为某个领域的专家，真正达到博士的水平，
But to really become a domain expert, to really be that Ph.
并且真正能够推动像这样的领域向前发展，这很难。
D.
另外，我还想说的是，我认为理解当你成功做到这一点时会发生什么，
And to really be something that can help push forward the ambition of a field like that's hard.
这一点也很重要。
And by the way, one thing I also want to say is that I think understanding what happens when you successfully do that.
我觉得建立一个好的思维模型很重要，你可以回顾一下AlphaGo的例子。
I think that having a good mental model of that's important, which is you look at something like rewind to AlphaGo.
对吧？
Right?
还记得第37手吗？那步棋彻底改变了人们对围棋的理解。
Remember move 37, this move that, like, changed people's understanding of the game.
而现在下围棋的人比以往任何时候都多。
And then now more people play go than ever.
没错。
Right.
它实际上激励了更多人去探索更远。
And it actually inspired people to do even more.
我觉得我们只会看到更多这样的情况。
I think we're just going to see that.
所以我认为，这种深度永远不会停止。
And so I think that the depth is never going to stop.
没错。
Right.
科学能深到什么地步？
How deep can you go on science?
没错。
Right.
有些人曾经以为，嘿，我们已经把物理学都搞明白了。
I think that people have thought sometimes that, hey, we found out all the physics.
一切都挺好的。
It's all good.
我们已经搞定了。
We're all done.
我不觉得这就是我们真正想要的未来。
And I don't think that that's the future we're signed up for.
我觉得我们注定要面对这样的情况：每解开一个谜题，就会引出十个新的谜题。
I think we're signed up for one where we got to keep find every time you unlock one mystery, every time you solve one mystery, unlocks like 10 more.
没错。
Right.
所以我觉得还有海量的工作要做，不同公司之间也有很大的空间可以做出差异化。
So I think that there's just going to be so much more to do and tons of room for differentiation across different companies.
所以我觉得我理解得没错，你的观点是，也许每个人都能扩大这些模型的规模，但最终拥有最多算力的公司会胜出。
So I think I'm reading you right and that your belief is maybe there's a way that everybody can scale up these models, but ultimately the company with the most compute is gonna win.
几个月前我们聊过，你提到过公司内部有人问过该买多少算力。
And, you know, we spoke a couple months ago and you had mentioned that like you were asked internally how much compute should we buy.
你说要全部买下。
And you said all of it.
他们又问：不，认真说，到底该买多少？
And they said no really, how much should we buy?
你回答：对，就全部买下来。
And you said no, buy all of it.
OpenAI 在采购算力方面绝对是领头羊。
And OpenAI is definitely the leader in buying compute.
看看他们花出去的钱吧。
Mean, see the money going out.
当然，他们通过投资也进账了不少钱。
Obviously, a lot of money coming in through investment.
你现在有了客户，业务也起来了，但花的钱依然惊人。
Now you've built a business with customers, but there's a lot of money going out.
你有没有想过，会不会有一天，这些钱根本还不起？
Do you ever wonder, hey, are are like, do you ever wonder maybe we're not gonna be able to pay all this money back because it's a
一个全新的领域？
brand new category?
我是从基本面来看这个问题的。
Well, the way that I look at it is on the fundamentals.
对吧？
Right?
你得明白，算力的部署往往要好几年才能落地，这一点是肯定的。
You you need to really look at the fact that the way that the compute goes is that it's multiple years out before compute actually arrives, right?
具体要看你做什么。
Depending on exactly what you're doing.
比如，我们已经连续好几年在自研芯片上投入，进展非常令人振奋。
For example, we've been investing in our own chip program now for multiple years and super exciting progress.
实际上，我们很快就会公布更多消息。
Like, you know, we'll we'll have more to announce actually pretty soon.
但能这么做，确实非常独特，对吧？
But the fact that we're able to do that is something very unique, right?
真正从供应链的全流程整合角度来思考。
Really thinking about the full vertical integration of the supply chain.
我觉得我们正走向一个世界，在那里，全球的算力根本满足不了所有需求。
And think I that the world we're heading towards is one where, again, there's just not going to be enough compute in the world to satisfy all the demand.
这一点我们看得非常清楚。
And we see this very concretely.
你就看聊天这一块的指数级增长，回头看看它的走势。
Like you look at the exponential, I mean, rewind to the exponential of chat.
再看看现在各种指数级发展的应用。
Look at exponentials run now.
想想我们能解决的问题，其实挺有意思的是，我们前天刚公布了一个新成果，是在化学领域，能合成新的反应路径。
You think about the problems that we that we are able to solve, you know, and it's it's actually kind of interesting that, you we just yesterday announced it was two days ago, announced a new result in, I I basically, chemistry and being able to synthesize, you know, new new improvement reaction.
而这一切都没怎么引起关注。
And all of this is without much attention.
我说的这一点，如果你深入一个领域，真的能彻底改变它。
You know, the thing I just said of if you go deep in a domain, you can really transform it.
但我们连皮毛都还没碰到。
And we're not even scratching the surface yet.
所以你要这么想，这个经济规模实在太庞大了。
And so the way to think about is the economy is so massive.
对吧？
Right?
我们自己增长的情况、人们愿意支付的费用，还有整个行业的规模和增速，都清楚地说明了这一点。
And we see it very concretely in terms of our own growth, in terms of what people are willing to pay and kind of the size and growth of this whole industry.
所以我最常思考的问题是：我们该怎么满足这种需求？
And so I think that the thing that I think about the most is how do we meet the demand?
怎么才能提供真正能支撑起人们想做的所有工作的工具？
How do you actually have something that can help support all of the work that people wanna do in the economy?
我觉得这事儿规模大得超乎想象。
And I think that is such a vast thing.
我们谁都还没真正意识到它有多巨大。
I don't think any of us have internalized it yet.
大家好。
Hi, everyone.
我是亚历克斯·坎托维奇。
Alex Cantowicz here.
我想跟大家聊聊我和Gravity合作制作的一部纪录片，探讨AI代理的安全未来。
I wanna tell you about a documentary I've made with Gravity to explore the future of AI agent security.
为了弄清楚我们是否真的准备好迎接自主代理，我采访了麻省理工学院的拉梅什·拉加拉教授、前白宫首席信息官特蕾莎·佩顿、米其林集团首席数据与AI官安比卡·拉贾戈帕尔，以及前阿里巴巴高管莎伦·盖伊。
To find out if we're truly ready for autonomous agents, I sat down with MIT professor Ramesh Rasgar, former White House CIO Theresa Payton, Michelin's Group Chief Data and AI Officer Ambika Rajagopal, and Sharon Guy, a former executive at Alibaba.
他们各自对这个不断变化的领域都提供了独特的见解。
They each offer unique insights into this evolving landscape.
最后，我们还对话了Gravity的首席执行官罗里·布伦德尔，一起探讨未来的方向。
We conclude with Rory Blundell, CEO of Gravity, to discuss the path forward.
在Gravity的引领下，欢迎加入我们的探索之旅。
With Gravity leading the way, join us on this journey.
完整纪录片可在节目说明中的链接观看。
You can watch the full documentary at the link in the show notes.
是的。
Yeah.
不过，容我提一句，一场价格战正在酝酿。
But if I may, there is a price war brewing.
至少报道是这么说的。
I mean, at least that's according to the reports.
很高兴你能来聊聊这个话题。
It's great to have you here to talk about it.
《华尔街日报》最近报道，OpenAI即将推出的新模型可能会大幅降价。
The Wall Street Journal recently had a report that an upcoming OpenAI model might have significant price cuts.
既然提供服务需要大量资源来应对不断增长的需求，如果这时候还要降价，那该怎么算这笔账呢？
And so again, like how can, you know, if it requires so much resources to serve this growing demand, in an environment where there might be price cuts.
你打算怎么让这个模式成立？
How do you make that math work?
嗯，我还是觉得
Well, again, I look at
换个角度来看。
it from a different angle.
如果我们回顾一下我们一路走来的发展，实际上我们一直在提升智能水平的同时降低价格，针对的是同等智能水平的服务。
So if you look at the whole history of what we've done, we actually have been increasing the intelligence, cutting price, right, for a fixed amount of intelligence.
但不知怎么的，杰文斯悖论一直在上演。
And people somehow just like like the Jevons paradox just keeps happening.
所以我认为，前沿的智能始终会是最昂贵的东西。
And so I think frontier intelligence will always be something that is going to be, you know, it's always going to the priciest thing.
但我相信，再过一年，这种级别的智能就会变得很普通，变得容易获得得多。
But I think that a year from now, that level of intelligence is going to feel pretty mundane and like, you know, going to be much more available.
我认为我们所处的世界，正逐渐让人们开始认真思考价值。
And I think that the world that we're in is one where people are starting to really think about value.
这种转变其实非常有意思——在过去，至少今年第一季度到现在的这段时间，人们只关注AI代理这些新玩意儿。
And it's actually a very interesting shift where over the past, you know, first quarter, maybe up until now, people have just been like this AI agent stuff.
一切都还很新鲜。
It's all new.
我们得把这套体系引入企业内部。
We need to bring into our enterprise.
我们可不想被甩在后面。
Like, we don't wanna be left behind.
我们该怎么融入这个未来？
How do we be part of this this future?
现在大家开始意识到，行了。
And now people are like, okay.
得确认这东西真能带来投资回报和实际价值。
Like, let's make sure this is actually delivering ROI and value.
我觉得这其实是个很好的转变。
And I actually that's a great place to be.
对吧？
Right?
因为大家都在问关键的问题。
Because people are asking the right questions.
今天我开了几场客户会议，听到大家说的正是这些话。
And I hear this I I had some customer meetings today where people were saying exactly this.
他们问，我们怎么才能做好基本的支出管控呢？
They're like, how can we have even just like good spend controls?
我们怎么才能实现可观测性？
How can we have observability?
说实话，我们今天才真正开始重视支出管控这个问题。
And I think we literally today just really spend controls.
所以，没错，就是这样。
So, you know, it's like, Okay, exactly.
我们正在全力投入企业级产品的准备，以及客户明确告诉我们需要的工具。
We're really investing hard in enterprise readiness and the tools that our customers are telling us that they need.
对我来说，这代表了我们公司也在经历一场转变——不再只是想着发布模型，而是真正开始思考整个端到端的流程。
And I think that that for me is the shift that we've also been going through as a company is really not just thinking about, hey, we're just going to release models and, you know, have a model really thinking about
端到端的
the end to end of
整个商业生态。
the business.
我们怎么把这些应用到真正解决客户的问题上？
How do we bring this into solving problems for real customers?
这种变化正在每个行业飞速发生。
And that is happening so quickly across every single industry.
还有那么多公司还在琢磨该怎么最好地利用这些模型，而我们也在同步学习。
And the number of different companies that still feel like they're wrapping their mind around how to best make use of these models, we're learning at the same time.
我觉得这整个领域才刚刚起步。
I think it's just so early in this whole game.
对我来说，市场规模增长得如此之快，我们自己的收入也在迅猛上升。
To me, the the, like, the absolute size of the market growing so quickly, our own revenue ramp grown grown so quickly.
我觉得现在还远远看不到尽头。
I think it's still just like none of
我们谁都没想到这一切会发展得这么迅猛。
us are anticipating how steep that's all gonna go.
你们会降价吗？
Are you gonna cut prices?
我的回答始终是肯定的。
I so, again, answer is always yes.
没错。
Right.
但我认为，未来持续发生的是，我们会不断推出前沿模型。
But it's about like, I think that what's going to keep happening is that we're going to have frontier models.
短期内不太可能出现巨大的转变。
I don't think there's going to be like a massive shift in in the short term.
我觉得这种事不会发生。
I don't think that that is the kind of thing that's going to happen.
但你应该预期的是，用一年左右的时间，达到今天这种顶尖水平的智能，成本会低得多。
But I think the thing you should anticipate is that over a year long time horizon to get to today's level of intelligence that feels very premier, it's going to be much cheaper.
但同时，会出现一种全新的、好上很多的东西。
But there's going to be a new thing that is going to be so much better.
你就会想，我干嘛还要用那个别的呢？
And you're like, why would I ever use this other one?
对。
Right.
这本来就是以后一直会这样发展的趋势。
It's just it's just how it's always going to be.
最近我看到一些挺有意思的推文和采访。
So had some interesting tweets and interviews recently.
他最近说，模型本身正在变成一种大宗商品，真正有价值的其实是公司本身——也可能是个转述。
He recently said the model is becoming a commodity and the valuable asset is company or this might be a paraphrase.
真正有价值的是那种能持续从你的数据中学习的、公司专属的AI系统。
The valuable asset is a company specific AI system that continually learns from your data.
你怎么看这个说法？
What do you think about that?
现在跟微软竞争，感觉奇怪吗？
And is it weird to be competing with Microsoft now?
你看，我认为在这个技术栈的任何一层，都不会
Well, look, I don't think that there's any layer of the stack here that is going
完全从价值链条中被剔除出去。
to just kind of be removed from the value chain.
我觉得这些因素是相乘的关系。
I think that these things multiply together.
如果你想想最基础的算力层，对吧，那就是没算力就没AI。
And if you think about the most the base layer of compute, right, that is something where it's just like no compute, no AI.
某种程度上，你可以说算力已经变得像大宗商品一样了。
And to some extent, you could say, oh, compute is commoditized.
说白了就是算力的堆砌。
It's just flops.
谁还在乎这个呢？
Who cares about it?
但实际情况是，你看看现在的芯片股，再看看那些卖算力的公司，市场给它们的估值就能看出来，这里头确实有个极其关键的底层资产。
But in reality, like, you look at you look at today's chip stocks, you look at the, you know, people who are selling compute, kind of what what is what the market is valuing people at, and they see that there's fundamental asset here that is just so critical.
我认为这是因为它是收入来源。
And I think that is because it is a revenue center.
任何做AI的人都得依赖它，而且在提升效率、压缩利润空间这些方面，还有很多有意思的现象。
It is something that anyone who's building AI has to rely on and that there's a bunch of very interesting dynamics in terms of the efficiencies that you can squeeze out in the margin, all these things.
但从根本上说，就算你眯着眼看，觉得它好像已经同质化了，可价值并不会消失。
But fundamentally, even though it's like you kind of squint it and say it's commoditized, it's not it's it's it's not that the value goes away.
利润空间也不会消失。
It's not that the margins go away.
市场之所以愿意给它高估值，是因为它有根本性的价值，而且它的关键性会随着时间越来越强。
It's like something that the market will reward because it has fundamental value and the importance of it's going to go up over time.
从大家抢购H100的价格上，你就能看出来。
You can see that with some of the prices that people are paying for H100s.
没错。
Right.
Hopper芯片可没过时。
Hoppers are, you know, kind of, you know, not not obsolete.
没错。
Right.
但它们是上一代的芯片。
But they're they're they're a previous gen chip.
在正常情况下，我们根本不会面临供应短缺。
And nor in any normal situation, we're not totally supply constrained.
根本没人会买它们。
No one would be buying them.
但现在的市场价格却比以前高多了。
But instead, the market prices are up relative to to where they were before.
所以这里出现了一种反常现象。
So there's this inversion that's happening.
而且我觉得这种情况还会持续，因为需求暴增，价格、利润，还有整个产业链各环节的指标都会继续上涨。
And again, I think it's going to keep happening where because everyone has this avalanche of demand that you're going to see prices and margins and all of these things continuing to increase at various levels of the stack.
我觉得模型领域也是同样的道理，虽然现在模型之间的竞争非常激烈。
I think the same kind of applies for models where the models themselves are also again, they're not there's there's a lot of competition there.
如果真是这样，对企业来说绝对是好事。
And if that's very good, I think it's good for the enterprise.
对客户和消费者来说也是好事。
I think it's good for customers, consumers.
但我觉得在很多领域，我们的模型一直是最聪明的，对吧？
But I think that there's a lot of areas where, for example, our models have always been the sort of smartest ones, right?
那些能解决极其复杂问题的模型。
The ones that are able to solve these incredibly hard problems.
我觉得我们现在才刚刚开始看到这种变革性的影响。
I think we're just starting to reach a phase where you're to see the transformative impact from that.
没错。
Right.
如果我们真能通过模型加速科学进展，模型越智能，进展就越快。
It's like if we're really able to speed up science through models, the smarter the model, the faster it's going to go.
这和那种能帮你订机票、安排日程的对话型模型完全不同。
And it's very, very different from a model that has a conversational interface that you're able to, you know, is able to book your travel, right, or organize your calendar.
所以这也是我们能做得非常出色的一个方面。
So that's also dimension I think we're going to do a very good job in.
但我只是想说，这是另一个领域。
But I'm just saying it's different area.
接下来的问题是，你该怎么把这种智能真正连接到你的客户身上？
And then I think that the question of, well, how do you actually connect the intelligence to your own customers?
没错。
Right.
要让这些在不同领域建立了卓越业务的企业真正获得价值，这很关键。
To real value to you have all these enterprises that have built incredible businesses in different domains.
这是一件大事。
And it's a huge thing.
如果没有相关领域的专业知识，你是根本做不到的。
And it's not something where if you don't have domain expertize that you're just going to be able to do.
没错。
Right.
其中一部分原因是，你需要考虑受监管的行业。
And part of it is that you need you think about regulated industries.
想想教育领域吧，这里有家长、老师、学生，这些不同的角色需要以非常细致的方式互动。在这些领域里，有很多价值可以挖掘，关键在于思考工作流程该怎么设计，这些模型又该如何协同运作。
You think about any area where there's like, you know, I think about education where you have a parent, you have a teacher, a student, you have these different parties that need to interact in very thoughtful ways for all of these areas, all of these domains that there is a lot of value to be built by being in that area and thinking about how the workflow should work, how these models should be orchestrated.
所以我真的觉得，机会多得是，我们必须作为一个完整的生态系统通力合作，才能实现这些系统真正可能带来的价值。
And so I I really think that there's more than enough to go around, and I think that we have to work together as a whole ecosystem in order to deliver the kind of value that I think is possible from these systems.
明白了。
Okay.
我再回到萨提亚的观点上，多说一句。
Just to go back to the Satya point one more time.
他说模型已经变成了一种大宗商品。
He's called models a commodity.
他正在打造自己的前沿智能系统。
He's trying to build his own frontier intelligence.
他实际上是在对客户说：你们得跟我们合作，我们会帮你们构建能从你们数据中学习的闭环。
He's telling potentially your customers, hey, you gotta come work with us because we're gonna help build these loops that will learn from your data.
我觉得他能一直用到2032年的你的知识产权。
He's got access to your IP, I think, till 2032.
那这让你觉得怎么样？
So how does it
听到萨提亚这么说，你是什么感受？
make you feel to hear this coming from Satya?
我觉得眼下最重要的是，AI正在推动经济转型，让所有人都能从中受益。
Look, I think that the most important thing that is happening right now is the usage of AI in the economy to really transform the economy and to uplift everyone.
这正是我最关注的方向。
And so I think that that is something that I'm really focused on.
越多的人致力于实现这一点，对大家就越有利。
And the more that people are trying to make that happen, I think that that's better for everyone.
听说GPT-5.6快来了。
GPT 5.6 is rumored to be on its way.
这只是一个推特上的传言，但我还是念给你听听。
It's supposed to be so this is just a Twitter rumor, but I'm gonna read it to you.
谣言总是最精彩的。
Always the best rumors.
成本只有Fable的三分之一，支持高达一百五十万令牌的上下文，智能编码流程更强大。
Three times cheaper than Fable up to 1,500,000 token context, stronger agentic coding workflows.
这些说法有几分是真的？
How much of that is true?
那我们对GPT 5.6该有什么期待？
What should we expect for GPT 5.6?
我的意思是，你当然期待它更好、更快、更聪明，方方面面都升级。
I mean, look, you should always expect better, faster, smarter, the whole thing.
所以所有消息都得到了证实。
So everything confirmed.
Twitter上看到的，别全信。
Definitely believe everything you read on Twitter.
也许吧。
Maybe not.
这确实给我个人生活带来了一些麻烦。
That has actually been a source of problems in my personal life.
明白了。
Okay.
那我最后想聊聊健康的话题。
So I want to end on health.
你之前提过好几次了。
You brought it up a couple of times.
其实早前观众里有人问过这个问题。
We actually had a question in the audience about it earlier.
有时候你会看到一个故事，读完心里就嘀咕：我知道这人是在对媒体说话，他说的听着好像有道理，但总觉得哪里不对劲，而且后续再也没了下文。
Sometimes there's a story, and you read it and you say to yourself, I know this person is speaking to the media and I know that what they're saying sounds like maybe it's true, but there's something wrong with the story and we're not going to see more of it.
最近我就读了几个这样的故事。
And I've read a couple of those recently.
其中一个，我记得好像是你朋友、GitLab 的首席执行官，他姓贺，得了癌症，把能做的检查都做了一遍。
One is, I think, is it your friend, the GitLab CEO, He Sid got cancer and used, he got all the diagnostic testing he could have.
所以他直接疯狂地做了各种检测，把数据全都喂给ChatGPT，还请了一些人帮忙开发了专门的应用程序，结果居然把癌症控制住了。
So, he just went out and tested like crazy and fed that data into ChatGPT with the assistance of some people who had built purpose built application for it and was able to beat back the cancer to a degree.
还有澳大利亚一只叫Rosie的狗。
There was also this dog Rosie the dog in Australia.
你们听说过Rosie的故事吗？
You guys heard of Rosie?
有个人给他的患癌狗做了活检，把基因突变数据输入AlphaFold，然后靠着聊天机器人帮忙设计出一种mRNA疫苗，注射到狗体内，结果狗不仅能跳上桌子了，肿瘤也明显缩小了。
Like the craziest story where this guy biopsied his dog which had cancer, ran the mutations across AlphaFold and then was able to design an mRNA vaccine that he injected into the dog with assistance of chatbots to build this thing, which ended up being able to jump over tables again and the tumor shrunk.
当我们思考AI和医疗的未来时，能不能帮我们理清这个问题：这些到底是少数碰巧上了头条的特例，背后其实有我们没听到的隐情，还是说这将来会成为常态？
When we think about the future of AI and health, help us sort out the truth with this question.
这些案例是偶然登上新闻的例外，还是说未来会变成标准做法？
Are these a couple of outliers that made good headlines, but there was something about the story we weren't hearing or is this gonna become standard in the future?
肯定会的。
Absolutely.
一定会成为常态。
Gonna become standard.
当然了。
Absolutely.
我身边就有不少朋友，他们都会把自己的健康检测数据拿去，用Codex这类工具来分析。
And it's I I personally have a number of friends who have done very similar things of get the data, right, like your health diagnostics and use codex.
对吧？
Right?
利用这些模型从中获取有价值的见解。
Use these models to get insights from them.
我觉得这样的人其实很多。
And I think that there are many people.
据说每周有大约两亿三千万人会用ChatGPT来咨询健康相关的问题。
Like, I think that there's about 230,000,000 people each week who use ChatGPT for health queries.
对吧？
Right?
这个规模真的太惊人了。
And that's been that's this is like a staggering scale.
对吧？
Right?
这些情况经常发生，比如你上传一份影像报告。
And this is people these are people sometimes you upload a scan.
有时候医生给你的意见还互相矛盾。
Sometimes you have doctors who are telling you conflicting information.
我觉得过去患者一直处在被动状态。
And I think that we've been in a world where patients are not empowered.
对吧？
Right?
患者得自己当医生。
Patients have to be the doctor.
对吧？
Right?
最终做决定的只能是你自己。
You're the decider.
你要对自己的健康负责。
You are accountable.
对吧？
Right?
医生要是犯了错，你这辈子都得为此买单。
You know, doctor makes a mistake and you're gonna be paying the price for the rest of your life.
这完全是另一种激励机制。
Like, it's just it's a very different kind of incentive.
这对我来说特别重要。
And this is very personal for me.
我妻子有好几种健康问题，说实话，要不是靠聊天工具，我真不知道我们该怎么应付她这么多状况。
You know, my wife has a number of health conditions and I think that we've just been we've not like, I don't even know how we'd be able to manage many of her conditions right now without the use of of chat.
我觉得我们才刚刚踏上这段旅程。
And I think we're just at the beginning of this journey.
对吧？
Right?
我觉得，哪怕你有最顶尖的医疗团队、最好的资源、最权威的专家，能做的事情也是有限的。
That I think that the degree to which even if you have the best medical team, the best access, the best best experts, there's only so much that can be done.
对吧？
Right?
你想想，有些事情根本超出了人类的能力范围，有时候甚至只是有人没仔细看病历。
You think about the things that are just outside of the reach of humanity or even just sometimes it's like someone didn't even read the chart.
对吧？
Right?
结果漏掉了关键细节。
And kinda missed the detail.
这些方面，我们完全可以通过这些工具大幅提升。
All of that, we should able to improve massively through these tools.
所以我觉得，个性化医疗有时候体现在药物研发上，可能是面向大众市场的，但有时候也针对那种‘单人定制’的情况，就像我今天早些时候提到的那些诊断。
And so I think that the personalized medicine and sometimes it's going to be about drugs and drug discovery that are of you know, for for mass market, but sometimes it'll be even for the kind of n of one things like the the disease, diagnoses that I mentioned earlier today.
有时候，只是为了更好地理解某种疾病，寻找新的潜在疗法。
Sometimes it will be for just, like, trying to understand conditions and trying to come up with with new potential therapeutics.
所有这些，我们现在正亲眼目睹它们在发生。
All of that, we're seeing it happening right now in front of our eyes.
这可不是理论上的东西。
It's not theoretical.
确确实实正在发生。
It's really happening.
我认为，人工智能最令人惊叹的可能性之一，就是它能大幅改善我们的健康状况。
And so one of the most I think it's like one of the most astounding possibilities of AI is how much it can improve our health.
你想想整个医疗体系的连锁反应，对吧？现在医疗开支已经大得惊人。
And you think about the ripple effects of the system, right, where so much spending on the health care system happens right now.
医疗是经济中极其重要的一部分。
It's a massive part of the economy.
如果你真能帮助人们预防问题，提前干预潜在的健康风险，那会带来巨大改变。
That if you're actually able to help people prevent issues, right, to to get ahead of potential, you know, health health problems.
这能真正减轻大量的负担和压力。
That's something that actually then alleviates a lot of burden, a lot of strain.
我们现在所处的世界里，医生和护士都精疲力尽了，这可不是小事，而是正在我们眼前发生的真正危机。
And we're in a world where doctors are burned out, nurses are burned out like there's like a real crisis that's happening in front of us.
我觉得人工智能能帮上这些忙。
And I think AI will be able to help with all of that.
只要我们能明智地部署和使用它，这种潜力是实实在在存在的。
Like, we have that potential if we deploy it and use it wisely and well.
所以，把人工智能应用到医疗领域，这确实是我个人投身于OpenAI这项事业的核心动力之一。
And so I think that applying AI to medicine, like, that's something that I is really a personal motivation for me in in thinking about this whole journey of of what we're building, what we're trying to do with OpenAI.
我坚信，作为全球社会和共同体，我们能够充分利用这一机遇。
And I'm hopeful that we as a as a world and a community can make the most of that.
但愿如此。
Let's hope.
我觉得我们一定能做到。
I think we will.
我非常非常有信心。
I'm very, very confident.
格雷格，真的非常感谢你。
Greg, thank you so much.
谢谢。
Thank you.
非常感激。
Appreciate it.
太感谢了。
Thank you so much.
天哪。
My god.
谢谢大家。
Thank you everyone.
今天过得还开心吗？
You have a good time today?
谢谢。
Thank you.
咱们明年再办一次吧？
Should we do it again next year?
对。
Yes.
你来吗？
Gonna come?
对。
Yes.
好嘞。
Alright.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。