Sam Altman on Trust, Persuasion, and the Future of Intelligence - Live at the Progress Conference (Ep. 259)
Hello, Sam.
你好，萨姆。
Happy to do this with you.
很高兴和你一起做这件事。
Excited to do it again.
很高兴能再次这样做。
Now, the last two months or so, there have been so many deals involving OpenAI, and I’m not even talking about the globetrotting.
现在，过去两个月左右，涉及 OpenAI 的交易非常多，我什至不是在谈论环球旅行。
A lot of them are local, or there’s new product features such as Pulse.
其中很多是本地的，或者有新的产品功能，例如 Pulse。
Presumably, you were productive, to begin with.
想必，你一开始就很有生产力。
How is it that you managed to up your productivity to get all this done?
您是如何提高工作效率来完成这一切的？
I don’t think there’s one single takeaway other than I think people almost never allocate their time as well as they think they do.
除了我认为人们几乎从来没有像他们认为的那样分配时间之外，我不认为有任何其他的收获。
As you have more demands and more opportunities, you find ways to continue to be more efficient.
当您有更多的需求和更多的机会时，您就会找到继续提高效率的方法。
We’ve been able to hire and promote great people, and I delegate a lot to them and get them to take stuff on.
我们已经能够雇佣和提拔优秀的人才，我将很多事情委托给他们并让他们承担一些工作。
That is the only sustainable way I know how to do it.
这是我知道的唯一可持续的方法。
I do try to make sure we increasingly, as what we need to do comes into focus — and there’s, as you mentioned, a lot of infrastructure that needs to be built out right now — I do try to make sure I understand what the core thing for us to do is.
我确实努力确保我们越来越多地关注，因为我们需要做的事情成为焦点 ——正如你提到的，现在需要建设很多基础设施 —— 我确实努力确保我了解我们要做的核心事情是什么。
It has, in some sense, simplified, and it’s very clear what we need to do.
从某种意义上说，它已经简化了，而且我们需要做什么非常清楚。
That’s been helpful.
这很有帮助。
I don’t really know.
我真的不知道。
I guess another thing that’s happened is more of the world wants to work with us, so deals are quicker to negotiate.
我想发生的另一件事是世界上更多的人希望与我们合作，因此交易谈判速度更快。
On hiring hardware people
关于雇用硬件人员
You’re doing much more with hardware or matters that are hardware-adjacent.
你正在用硬件或与硬件相关的事情做更多的事情。
How is hiring or delegating to a good hardware person different from hiring good AI people, which is what you started off doing?
雇用或委托给优秀的硬件人员与雇用优秀的人工智能人员（您一开始就是这么做的）有何不同？
You mean consumer devices or chips?
你指的是消费类设备还是芯片？
Both.
两个都。
One thing that’s different is that cycle times are much longer.
不同的一件事是周期时间要长得多。
The capital is more intense.
资本更加激烈。
The cost of screen up is higher, so I like to spend more time getting to know the people before saying, “Okay, you’re just going to do this, and I’ll trust that it’ll work out.” Otherwise the theory is the same.
筛选的成本更高，所以我喜欢花更多的时间去了解人们，然后再说：“好吧，你就这么做，我相信它会成功。”否则理论是一样的。
You try to find good, effective, fast-moving people.
你试图找到优秀、高效、行动迅速的人。
Get clear on what the goal is and just let them go do it.
明确目标是什么，然后让他们去做。
I visited Nvidia earlier this year.
我今年早些时候访问过 Nvidia。
They were great.
他们很棒。
They were great to me.
他们对我很好。
They’re super smart, but it just felt so different from walking the floor of OpenAI.
他们非常聪明，但感觉与在 OpenAI 上行走有很大不同。
I’m sure you’ve been to Nvidia.
我相信你去过 Nvidia。
People read Twitter less.
人们阅读 Twitter 的次数减少了。
At least on the surface, they’re less weird.
至少从表面上看，它们没那么奇怪。
What is that intangible difference in the hardware world that one has to grasp to do well in it?
要想在硬件世界中取得好成绩，必须掌握的无形差异是什么？
Look, I don’t know if this is going to turn out to be a good sign or a bad sign, but our chip team feels more like the OpenAI research team than a chip company.
看，我不知道这会是一个好兆头还是坏兆头，但我们的芯片团队感觉更像是 OpenAI 研究团队，而不是一家芯片公司。
I think it might work out phenomenally well.
我认为效果可能会非常好。
You’re extending the model —
你正在扩展模型 ——
We are.
我们是。
— of your previous hires to hardware.
— 您之前雇用的硬件人员。
With some risk, but we are.
虽然存在一定风险，但我们确实如此。
There’s this fellow on Twitter.
Twitter 上有这个人。
His name is roon.
他的名字叫罗恩。
He’s become quite well known.
他已经变得很有名了。
What is it that makes roon special to you?
是什么让 roon 对您来说如此特别？
He’s like a very lateral thinker.
他就像一个非常横向的思考者。
You can start down one path and jump somewhere completely else, but keep going down, stay on the same trajectory, but in some very different context that’s unusual.
你可以从一条路径开始，然后完全跳到其他地方，但继续向下，保持在相同的轨迹上，但在一些非常不同的环境中这是不寻常的。
He’s great at phrasing observations in an interesting, useful, whatever way, and that makes him fun and quite useful to talk to.
他擅长以有趣、有用的方式表达观察结果，这让他很有趣，而且与他交谈非常有用。
I don’t know.
我不知道。
He brings together a lot of skills that don’t often exist in one person’s head.
他汇集了许多通常不存在于一个人头脑中的技能。
How does that shape what you have him work on?
这对你让他从事的工作有何影响？
You see this about him, and then you think, “Ah, roon should do X.”
你看到他的这一点，然后你会想，“啊，roon 应该做 X。”
I very rarely get to have anybody work on anything.
我很少让别人做任何事情。
One thing about researchers is they’re going to work on what they’re going to work on, and that’s that.
研究人员的一件事是，他们将致力于他们将要从事的工作，仅此而已。
Someone put an essay online, and it said, in all the time they worked at OpenAI, they hardly ever sent or received an email, that so many things were done over Slack.
有人在网上发了一篇文章，里面说，他们在 OpenAI 工作的这段时间里，几乎没有收发过电子邮件，所以很多事情都是通过 Slack 完成的。
Why is that?
这是为什么？
What’s your model of why email is bad and Slack is good for OpenAI?
您认为为什么电子邮件对 OpenAI 不好而 Slack 对 OpenAI 有好处，您的模型是什么？
I’ll agree, email is bad.
我同意，电子邮件很糟糕。
I don’t know if Slack is good.
不知道Slack好不好。
I suspect it’s not.
我怀疑不是。
I think email is very bad.
我认为电子邮件非常糟糕。
The threshold to make something better than email is not high, and I think Slack is better than email.
做出比电子邮件更好的东西的门槛并不高，而且我认为 Slack 比电子邮件更好。
We have a lot of things going on at the same time, as you observed, and we have to do things extremely quickly.
正如您所观察到的，我们有很多事情同时进行，而且我们必须非常快地做事情。
It’s definitely a very fast-moving organization.
这绝对是一个发展非常快的组织。
There are positives about Slack, but there’s also, I dread the first hour of the morning, the last hour before I go to bed, where I’m just dealing with this explosion of Slack, and I think it does create a lot of fake work.
Slack 有积极的一面，但也有一点，我害怕早上的第一个小时，也就是睡觉前的最后一个小时，我正在处理 Slack 的爆炸式增长，我认为它确实创造了很多虚假的工作。
I suspect there is something new to build that is going to replace a lot of the current office, productivity suite, whatever you think of docs, slides, email, Slack, whatever, that will be the AI-driven version of all of these things, not where you tack on the horrible — you accidentally click the wrong place and it tries to write a whole document for you or summarize some thread or whatever, but the actual version of you are trusting your AI agent and my AI agent to work most stuff out and escalate to us when necessary.
我怀疑有一些新的东西要构建，它将取代许多当前的办公室、生产力套件，无论你想到什么文档、幻灯片、电子邮件、Slack，等等，这将是所有这些东西的人工智能驱动版本，而不是你附加可怕的 —— 你不小心点击了错误的地方，它试图为你写一个完整的文档或总结一些线程或其他什么，但你的实际版本是信任你的人工智能代理和我的人工智能代理来解决大部分问题并升级必要时给我们。
I think there is probably finally a good solution for someone to make within reach.
我认为最终可能有人可以找到一个触手可及的好解决方案。
How far are you from having that internally?
你距离内部实现这一点还有多远？
Maybe not a product for the whole world, that it’s in every way tested, but that you would use it within OpenAI?
也许不是一个适合全世界的产品，它在各个方面都经过了测试，但你会在 OpenAI 中使用它吗？
Very far, but I suspect just because we haven’t made any effort to try, not because the models are that far away.
很远，但我怀疑只是因为我们没有做出任何努力去尝试，而不是因为模型有那么远。
Since talent, time, human capital is so valuable within your company, why shouldn’t that be a priority?
既然人才、时间、人力资本在您的公司中如此宝贵，为什么不应该将其作为优先事项呢？
Probably we should do it, but people get stuck in their own ways of doing things, and a lot of stuff is going very well right now, so there’s a lot of activation energy for a big new thing.
也许我们应该这样做，但人们会陷入自己的做事方式，而现在很多事情进展顺利，所以对于一个大的新事物来说有很多激活能量。
On what GPT-6 will enable
GPT-6 将支持哪些功能
What is it about GPT-6 that makes that special to you?
GPT-6 有何特别之处？
If GPT-3 was the first moment where you saw a glimmer of something that felt like the spiritual Turing test getting passed, GPT-5 is the first moment where you see a glimmer of AI doing new science.
如果说 GPT-3 是你第一次看到一丝曙光，感觉像是通过了精神图灵测试，那么 GPT-5 就是你第一次看到人工智能在做新科学的那一刻。
It’s very tiny things, but here and there someone’s posting like, “Oh, it figured this thing out,” or “Oh, it came up with this new idea,” or “Oh, it was a useful collaborator on this paper.” There is a chance that GPT-6 will be a GPT-3 to 4-like leap that happened for Turing test-like stuff for science, where 5 has these tiny glimmers and 6 can really do it.
这是非常微小的事情，但到处都有人发帖，比如“哦，它解决了这个问题”，或者“哦，它想出了这个新想法”，或者“哦，它是这篇论文的一个有用的合作者。” GPT-6 有可能会是类似 GPT-3 到 4 的飞跃，发生在类似于图灵测试的科学事物上，其中 5 有这些微小的闪光，而 6 确实可以做到。
Let’s say I run a science lab, and I know GPT-6 is coming.
假设我经营一个科学实验室，并且我知道 GPT-6 即将推出。
What should I be doing now to prepare for that?
我现在应该做什么准备呢？
It’s always a very hard question.
这始终是一个很难回答的问题。
Even if you know this thing is coming, if you adapt your —
即使你知道这件事即将发生，如果你调整你的 -
Let’s say I even had it now, right?
假设我现在已经拥有了，对吧？
What exactly would I do the next morning?
第二天早上我到底要做什么？
I guess the first thing you would do is just type in the current research questions you’re struggling with, and maybe it’ll say, “Here’s an idea,” or “Run this experiment,” or “Go do this other thing.”
我想你要做的第一件事就是输入你正在努力解决的当前研究问题，也许它会说，“这是一个想法”，或者“进行这个实验”，或者“去做其他事情”。
If I’m thinking about restructuring an entire organization to have GPT-6 or 7 or whatever at the center of it, what is it I should be doing organizationally, rather than just having all my top people use it as add-ons to their current stock of knowledge?
如果我正在考虑重组整个组织，以 GPT-6 或 7 或其他任何内容为中心，那么我应该在组织上做什么，而不是仅仅让我所有的高层人员将其用作他们当前知识储备的附加组件？
I’ve thought about this more for the context of companies than scientists, just because I understand that better.
我更多地从公司的角度而不是科学家的角度来思考这个问题，只是因为我更了解这一点。
I think it’s a very important question.
我认为这是一个非常重要的问题。
Right now, I have met some orgs that are really saying, “Okay, we’re going to adopt AI and let AI do this.” I’m very interested in this, because shame on me if OpenAI is not the first big company run by an AI CEO, right?
现在，我遇到了一些组织，他们真的在说：“好吧，我们将采用人工智能，并让人工智能来做这件事。”我对此非常感兴趣，因为如果 OpenAI 不是第一家由 AI CEO 运营的大公司，我会感到羞耻，对吗？
Just parts of it.
只是其中的一部分。
Not the whole thing.
不是全部。
No, the whole thing.
不，是整个事情。
That’s very ambitious.
这是非常雄心勃勃的。
Just the finance department, whatever.
只是财务部门，无论如何。
Well, but eventually it should get to the whole thing, right?
好吧，但最终应该解决整个问题，对吧？
So we can use this and then try to work backwards from that.
所以我们可以使用它，然后尝试从它开始向后工作。
I find this a very interesting thought experiment of what would have to happen for an AI CEO to be able to do a much better job of running OpenAI than me, which clearly will happen someday.
我发现这是一个非常有趣的思想实验，关于人工智能 CEO 必须发生什么才能比我更好地运行 OpenAI，这显然有一天会发生。
How can we accelerate that?
我们如何才能加速这一进程？
What’s in the way of that?
这有什么问题吗？
I have found that to be a super useful thought experiment for how we design our org over time and what the other pieces and roadblocks will be.
我发现这是一个非常有用的思想实验，对于我们如何随着时间的推移设计我们的组织以及其他部分和障碍是什么。
I assume someone running a science lab should try to think the same way, and they’ll come to different conclusions.
我认为经营科学实验室的人应该尝试以同样的方式思考，他们会得出不同的结论。
How far off do you think it is that just, say, one division of OpenAI is 85 percent run by AIs?
您认为 OpenAI 的一个部门 85% 由人工智能运行的情况还有多远？
Any single division?
有单独的部门吗？
Not a tiny, insignificant division, mostly run by the AIs.
这不是一个微小的、无关紧要的部门，主要由人工智能管理。
Some small single-digit number of years, not very far.
一些小的个位数的年数，并不是很远。
When do you think I can be like, “Okay, Mr. AI CEO, you take over”?
你认为我什么时候可以说，“好吧，AI CEO 先生，你接手吧”？
CEO is tricky because the public role of a CEO, as you know, becomes more and more important.
首席执行官很棘手，因为如您所知，首席执行官的公共角色变得越来越重要。
Let’s say if I can pretend to be a politician, which is not my natural strength, an AI can do it, too.
假设我可以假装成为一名政治家（这不是我的天赋），那么人工智能也可以做到。
Let’s say I stay involved for the public-facing whatever.
假设我继续参与面向公众的任何事情。
Just actually making the good decisions, figuring out what to do.
只是真正做出正确的决定，弄清楚该怎么做。
I think you’ll have billion-dollar companies run by two or three people with AIs, I don’t know, two and a half years.
我认为你会拥有价值数十亿美元的公司，由两三个人工智能人员运营，我不知道，两年半。
I used to think one year, but maybe I’ve put it off a bit.
我曾经想过一年，但也许我已经推迟了一点。
I’m not more pessimistic about the AI.
我对人工智能并不更加悲观。
Maybe I’m more pessimistic about the humans.
也许我对人类更加悲观。
What’s your forecast?
你的预测是什么？
I agree on all of those counts.
我同意所有这些观点。
I think the AI can do it sooner than that.
我认为人工智能可以比这更快做到这一点。
I think this is a good thing for society and a good thing for the future, not a bad one.
我认为这对社会、对未来都是好事，而不是坏事。
People have a great deal higher trust in other people over an AI, even if they shouldn’t, even if that’s irrational.
与人工智能相比，人们对其他人的信任要高得多，即使他们不应该这样做，即使这是不合理的。
The AI doctor is better, but you want the human, whatever.
人工智能医生更好，但无论如何你都想要人类。
I think it may take much longer for society to get really comfortable with this and for people in the organization to get really comfortable with this.
我认为社会可能需要更长的时间才能真正适应这一点，组织中的人员也可能需要更长的时间才能真正适应这一点。
On the actual decision-making for most things, maybe the AI is pretty good pretty soon.
在大多数事情的实际决策中，也许人工智能很快就会变得非常好。
You’re hiring a lot of smart people.
你雇佣了很多聪明人。
Do you ask yourself, “What are the markers of how AI-resistant this very smart person will be?” Do you have an implicit mental test for that, or you just hire smart people and hope it’s all going to work out later?
你是否会问自己：“这个非常聪明的人对人工智能的抵抗力有哪些标志？”你是否对此有一个隐含的心理测试，或者你只是雇佣聪明的人并希望这一切都能在以后解决？
No, I do ask questions about that.
不，我确实会问相关问题。
People will just lie, right?
人们只会撒谎，对吗？
They know they’re talking to OpenAI.
他们知道自己正在与 OpenAI 对话。
What do you actually look for in them?
您实际上在其中寻找什么？
A big one is how they use AI today.
其中一个重要问题是他们今天如何使用人工智能。
The people who still are like, “Oh, yes, I’d use it for better Google search and nothing else,” that’s not necessarily a disqualifier, but that’s like a yellow flag.
人们仍然会说，“哦，是的，我会用它来更好的谷歌搜索，而不是别的，”这不一定是取消资格，但这就像一面黄旗。
People who are seriously considering what their day-to-day is going to look like in three years, that’s a green flag.
对于那些认真考虑三年后他们的日常生活会是什么样子的人来说，这是一个绿旗。
A lot of people aren’t.
很多人都不是。
They’re like, “Oh, yes, probably it’s going to be really smart.”
他们会说，“哦，是的，这可能会非常聪明。”
Do you think scientific labs might get GPT-6 this year?
您认为科学实验室今年可能会获得 GPT-6 吗？
Not this year.
今年不行。
On government backstops for AI companies
论政府对人工智能公司的支持
Not this year.
今年不行。
Here’s a very difficult question.
这是一个非常困难的问题。
As you know, both you and I were fans of nuclear power, but we also know, the insurance for nuclear power plants is provided by the government.
如你所知，你和我都是核电爱好者，但我们也知道，核电站的保险是由政府提供的。
The plants might be quite safe, but people worry.
这些植物可能相当安全，但人们担心。
They’re Nervous Nellies.
他们是紧张的内莉。
There’s a lot of parties involved.
有很多各方参与其中。
The federal government does the insurance.
联邦政府负责保险。
Do you worry that the future holds the same for AI companies, or the Feds are your insurer, and how do you plan for that?
您是否担心人工智能公司的未来会一样，或者美联储是您的保险公司，您对此有何计划？
Again, even if AI is pretty safe, as with nuclear power, people are Nervous Nellies.
再说一次，即使人工智能非常安全，就像核能一样，人们也是神经质的。
How will you insure everything?
你将如何确保一切？
At some level, when something gets sufficiently huge, whether or not they are on paper, the federal government is kind of the insurer of last resort, as we’ve seen in various financial crises and insurance companies screwing things up.
在某种程度上，当事情变得足够大时，无论它们是否在纸面上，联邦政府都是最后的保险公司，正如我们在各种金融危机和保险公司搞砸的情况中看到的那样。
I guess, given the magnitude of what I expect AI economic impact to look like, I do think the government ends up as the insurer of last resort, but I think I mean that in a different way than you mean that, and I don’t expect them to actually be writing the policies in the way that maybe they do for nuclear.
我想，考虑到我对人工智能经济影响的预期，我确实认为政府最终会成为最后的保险人，但我认为我的意思与你的意思不同，而且我不认为他们实际上会像核能那样制定政策。
There’s a big difference between the government being the insurer of last resort and the insurer of first resort.
政府作为最后保险人和第一保险人之间存在很大差异。
Last resort’s inevitable, but I’m worried they’ll become the insurer of first resort, and that I don’t want.
最后的手段是不可避免的，但我担心他们会成为首选的保险公司，而我不希望这样。
I don’t want that either.
我也不想要那样。
I don’t think that’s what will happen.
我认为这不会发生。
What we’re seeing with Intel, lithium, rare earths is, the government is becoming an equity holder.
我们在英特尔、锂、稀土方面看到的是，政府正在成为股东。
Again, not of last resort, sort of second or third resort.
同样，这不是最后的手段，而是第二或第三种手段。
I don’t mean this as a comment about the Trump administration.
我的意思并不是对特朗普政府的评论。
I think this is something we might be seeing in any case or see in the future after Trump is gone.
我认为这是我们在任何情况下都可能会看到的事情，或者是特朗普下台后我们将来可能会看到的事情。
How do you plan for OpenAI, knowing that’s now a thing on the table in the American economy?
知道 OpenAI 现在已成为美国经济中的重要议题，您将如何规划 OpenAI？
Look, there are all these ways where you can imagine that I put almost no probability mass onto the world where no one has any meaning in the post-AGI world because the AI is doing everything.
看，你可以通过所有这些方式想象我几乎没有把概率质量放在后 AGI 世界中，在这个世界上，没有人有任何意义，因为人工智能正在做一切。
I think we’re really great at finding new things to do, new games to play, new ways to be useful to each other, to compete, to get fulfilled, whatever, but I do put a significant probability that the social contract has to change significantly.
我认为我们真的很擅长寻找新的事情做，新的游戏玩，新的方式对彼此有用，竞争，获得满足，等等，但我确实认为社会契约必须发生重大变化的可能性很大。
I don’t know what that will look like.
我不知道那会是什么样子。
Can I see the government getting more involved there and thus having some strong opinions about AI companies?
我能看到政府更多地参与其中，从而对人工智能公司有一些强烈的看法吗？
I can totally see that.
我完全可以看到这一点。
We don’t live our lives that way.
我们不那样生活。
We just try to work with capitalism as it currently exists.
我们只是尝试与目前存在的资本主义合作。
I believe that that should be done by the companies and not the government, although we’ll partner with the government and try to be a good collaborator.
我认为这应该由公司而不是政府来完成，尽管我们将与政府合作并努力成为一个好的合作者。
I don’t want them writing our insurance policies.
我不想让他们写我们的保险单。
On monetizing AI services
关于人工智能服务货币化
Now, I did a trip through France and Spain this summer with my wife.
现在，今年夏天我和妻子一起去了法国和西班牙旅行。
Every hotel we booked, other than the first one, well, we found it through GPT-5.
我们预订的每家酒店，除了第一家，都是通过 GPT-5 找到的。
We didn’t book it through GPT.
我们没有通过 GPT 预订。
Almost every meal we ate, right?
我们几乎每顿饭都吃过，对吧？
You didn’t get a dime for this.
你没有为此得到一毛钱。
I’m telling my wife, “Well, this just seems wrong,” right?
我告诉我的妻子，“嗯，这似乎是错误的，”对吗？
What is the new world going to look like soon enough?
新世界很快会是什么样子？
How will that work?
这将如何运作？
To zoom out even before the answer, one of the unusual things we noticed a while ago, and this was when it was a worse problem, ChatGPT would consistently be reported as a user’s most trusted technology product from a big tech company.
甚至在答案出现之前就缩小范围，这是我们不久前注意到的不寻常的事情之一，而且当时这是一个更严重的问题，ChatGPT 一直被报告为用户最值得信赖的来自大型科技公司的技术产品。
We don’t really think of ourselves as a big tech company, but I guess we are now.
我们并不真正认为自己是一家大型科技公司，但我想我们现在是了。
That’s very odd on the surface because AI is the thing that hallucinates, AI is the thing with all the errors, and this was when that was much more of a problem.
从表面上看，这很奇怪，因为人工智能会产生幻觉，人工智能会带来所有错误，而这正是问题更大的时候。
There is a question of why.
有一个问题是为什么。
Ads on a Google search are dependent on Google doing badly.
谷歌搜索上的广告取决于谷歌表现不佳。
If it was giving you the best answer, there’d be no reason ever to buy an ad above it.
如果它给了你最好的答案，就没有理由购买上面的广告。
You’re like, “That thing is not quite aligned with me.”
你会说，“那件事不太适合我。”
ChatGPT, maybe it gives you the best answer, maybe it doesn’t, but you’re paying it, or hopefully, all are paying it, and it’s at least trying to give you the best answer.
ChatGPT，也许它给了你最好的答案，也许它没有，但你正在付费，或者希望所有人都在付费，它至少试图给你最好的答案。
That has led to people having a deep and pretty trusting relationship with ChatGPT.
这使得人们与 ChatGPT 建立了深厚且相当信任的关系。
You ask ChatGPT for the best hotel, not Google or something else.
你向 ChatGPT 询问最好的酒店，而不是 Google 或其他东西。
If ChatGPT were accepting payment to put a worse hotel above a better hotel, that’s probably catastrophic for your relationship with ChatGPT.
如果 ChatGPT 接受付款以将更差的酒店置于更好的酒店之上，那么这对您与 ChatGPT 的关系可能是灾难性的。
On the other hand, if ChatGPT shows you it’s guessed the best hotel, whatever that is, and then if you book it with one click, takes the same cut that it would take from any other hotel, and there’s nothing that influenced it, but there’s some sort of transaction fee, I think that’s probably okay.
另一方面，如果 ChatGPT 向你显示它被猜测为最好的酒店，无论那是什么，然后如果你一键预订，就会得到与任何其他酒店相同的折扣，没有什么影响它，但有某种交易费用，我认为这可能没问题。
With our recent commerce thing, that’s the spirit of what we’re trying to do.
就我们最近的商业活动而言，这就是我们努力做的事情的精神。
We’ll do that for travel at some point.
我们会在某个时候为旅行做到这一点。
I’m not worried about the payola issue, but let me tell you my worry, and that is there may be a tight cap on the commission you can charge, because we’re now in a world, say where there’s agents, and someone finds the best hotel through GPT-7 or whatever, and then they just talk to their computer or their pendant, and they go to some stupider service, but the stupider service is an agent that books very cheaply, and they only really have to pay OpenAI a commission equal to what the stupidest service will charge.
我不担心工资问题，但让我告诉你我的担心，那就是你可以收取的佣金可能有一个严格的上限，因为我们现在在一个世界里，比如说哪里有代理，有人通过 GPT-7 或其他什么找到最好的酒店，然后他们只是与他们的计算机或他们的挂件交谈，然后他们去一些更愚蠢的服务，但更愚蠢的服务是预订非常便宜的代理，他们只需要向 OpenAI 支付相当于最愚蠢的服务会收费。
One thing I believe in general related to this is that margins are going to go dramatically down on most goods and services, including things like hotel bookings.
我认为与此相关的一件事是，大多数商品和服务的利润率将大幅下降，包括酒店预订等。
I’m happy about that.
我对此很高兴。
I think there’s a lot of taxes that just suck for the economy, and getting those down should be great all around.
我认为有很多税收对经济来说是不利的，降低这些税收应该是一件好事。
I think that most companies like OpenAI will make more money at a lower margin.
我认为大多数像 OpenAI 这样的公司会以较低的利润赚更多的钱。
Do you worry about the discrepancy between the fixed upfront cost of making yours the smartest model compared to the very cheap cost of all the competing agent has to do is book it for someone, and how do you use the commissions to pay for making the model smarter, in essence?
您是否担心使您的模型成为最智能模型的固定前期成本与竞争代理必须为某人预订的非常便宜的成本之间的差异，以及您如何使用佣金来支付使模型变得更智能的费用？
I think the way to monetize the world’s smartest model is certainly not hotel booking.
我认为世界上最聪明的模式的货币化方式肯定不是酒店预订。
But you want to do it, nonetheless.
但尽管如此，你还是想这么做。
No. I want to discover new science and figure out a way to monetize that that you can only do with the smartest model.
不。我想发现新的科学并找到一种只能通过最智能的模型才能实现的货币化方法。
There is a question of, many people have asked, “Should OpenAI do ChatGPT at all?
很多人问过一个问题：“OpenAI 到底应该做 ChatGPT 吗？
Why don’t you just go build AGI?
你为什么不去构建 AGI？
Why don’t you go discover a cure for every disease, nuclear fusion, cheap rockets, the whole thing, and just license that technology?”
你为什么不去寻找治疗每一种疾病的方法，核聚变，廉价火箭，所有这些，然后就授权这项技术呢？”
It is not an unfair question because I believe that is the stuff that we will do that will be most important and make the most money eventually, but my story, my most likely story about how this works, how the world gets dramatically better, is we put a really great superintelligence in the hands of everybody.
这不是一个不公平的问题，因为我相信这是我们要做的最重要、最终赚最多钱的事情，但我的故事，我最有可能讲述它如何运作、世界如何变得更加美好的故事，是我们把一个真正伟大的超级智能交到每个人的手中。
We make it super easy to use.
我们让它超级容易使用。
It’s nicely integrated.
它集成得很好。
We make you beautiful devices.
我们为您打造精美的设备。
We connect it to all your services.
我们将其连接到您的所有服务。
It gets to know you over your life.
它会在你的一生中了解你。
It does all the stuff for you.
它为你做所有的事情。
We invest in infrastructure and chips, and energy, and the whole thing to make it super abundant and super cheap.
我们投资基础设施、芯片、能源以及所有这些，以使其变得超级丰富和超级便宜。
Then you all figure out how the world gets way better.
然后你们就会明白世界如何变得更好。
Maybe some people will only ever book hotels and not do anything else, but a lot of people will figure out they can do more and more stuff and create new companies and ideas and art and whatever.
也许有些人只会预订酒店而不做其他任何事情，但很多人会发现他们可以做越来越多的事情，创造新的公司、想法、艺术等等。
Maybe ChatGPT and hotel booking and whatever else is not the best way we can make money.
也许ChatGPT和酒店预订之类的东西并不是我们赚钱的最佳方式。
In fact, I’m certain it’s not.
事实上，我确信不是。
I do think it’s a very important thing to do for the world.
我确实认为这对世界来说是一件非常重要的事情。
I’m happy for OpenAI to do some things that are not the economic-maxing thing.
我很高兴 OpenAI 能够做一些不追求经济最大化的事情。
Now, you have a deal in the works with Walmart that people can use GPT, ask it questions, “What should I buy at Walmart?” and then they can buy it at Walmart, and you and Walmart have some arrangement.
现在，您与沃尔玛达成了一项协议，人们可以使用 GPT，向其提问：“我应该在沃尔玛购买什么？”然后他们可以在沃尔玛购买，你和沃尔玛有一些安排。
Do you think Amazon will fold and join that, or are they going to fight back and try to do their own thing?
你认为亚马逊会折叠并加入其中，还是会反击并尝试做自己的事情？
I have no idea.
我不知道。
If I were them, I would fight back.
如果我是他们，我就会反击。
You would fight back?
你会反击吗？
I think so.
我想是的。
Yes.
是的。
Ads.
广告。
How important a revenue source will ads be for OpenAI?
对于 OpenAI 来说，广告作为收入来源有多重要？
Again, there’s a kind of ad that I think would be really bad, like the one we talked about.
再说一次，我认为有一种广告非常糟糕，就像我们谈到的那样。
There are kinds of ads that I think would be very good or pretty good to do.
我认为有些广告做得非常好或相当好。
I expect it’s something we’ll try at some point.
我预计我们会在某个时候尝试这一点。
I do not think it is our biggest revenue opportunity.
我不认为这是我们最大的收入机会。
What will the ad look like on the page?
广告在页面上会是什么样子？
I have no idea.
我不知道。
You asked a question about productivity earlier.
您之前问了一个有关生产力的问题。
I’m really good about not doing the things I don’t want to do.
我真的很擅长不做我不想做的事情。
That’s something you don’t want to do?
这是你不想做的事吗？
We have the world expert thinking about our product strategy.
我们有世界专家思考我们的产品策略。
I used to do that.
我曾经这样做过。
I used to spend a lot of time thinking about product.
我曾经花很多时间思考产品。
Now she’s much better at it than me.
现在她比我擅长得多。
I have other things to think about.
我还有其他事情要考虑。
I’m sure she’ll figure it out.
我相信她会明白的。
Whether or not you agree with it, what is the best this-is-not-a-bubble argument.
无论你是否同意，最好的“这不是泡沫”论点是什么。
Is it just the insatiable demand for compute?
这仅仅是对计算的永不满足的需求吗？
There’s a lot of arguments I’m tempted to give, but I think the intellectually most interesting one is we have no idea how much past human-level intelligence can go and what you can do with it as it does.
我很想提出很多论点，但我认为从智力上来说最有趣的一个是我们不知道人类水平的智能可以过去多久，以及你可以用它做什么。
There’s all the arguments that everyone has made.
这是每个人都提出的所有论点。
The one I would like to see people talk about much more is “How are you even supposed to think about vastly superhuman intelligence and the economic impacts of that?”
我希望看到人们更多地谈论的是“你应该如何思考超人的智慧及其对经济的影响？”
On AI’s future understanding of intangibles
论人工智能未来对无形资产的理解
Now, OpenAI is in talks with Saudi Arabia, with UAE.
现在，OpenAI 正在与沙特阿拉伯和阿联酋进行谈判。
Let’s take the most optimistic scenario for how all that goes.
让我们以最乐观的情况来说明这一切的进展情况。
What is it that top OpenAI management needs to know or understand about those countries, and how is it you learn it?
OpenAI 的高层管理人员需要了解或了解这些国家的什么？您是如何了解的？
Well, it would depend on what we were doing with them.
嗯，这取决于我们对它们做什么。
Putting data centers in a country or taking investment from a country or deploying commercial services would be very different than a set of other collaborations we could imagine.
将数据中心放在一个国家或从一个国家接受投资或部署商业服务将与我们想象的一系列其他合作有很大不同。
But generally speaking, to put data centers in a country, what we need to understand is who’s going to run it.
但一般来说，要将数据中心放在一个国家，我们需要了解谁来运营它。
We don’t operate our own data centers, but Microsoft or Oracle, or somebody else.
我们不运营自己的数据中心，而是运营微软或甲骨文或其他公司的数据中心。
What workload are we going to put there?
我们要在那里放置什么工作量？
What model weights are we going to put there?
我们要在那里放什么模型权重？
What are the security guarantees going to look like?
安全保证是什么样的？
We do want to build data centers around the world, with lots of countries, but for this question, which is the main thing we deal with other countries for, those are the kinds of questions.
我们确实想在世界各地建立数据中心，有很多国家，但对于这个问题，这是我们与其他国家打交道的主要问题，这些就是这样的问题。
If we were, which we don’t have current plans, if we were developing a custom model for some country, we’d have a whole bunch more questions.
如果我们是，我们目前没有计划，如果我们为某个国家开发定制模型，我们会有更多问题。
They have different legal codes, different expectations from a deal.
他们有不同的法律法规，对交易的期望也不同。
I’m not saying it’s in a bad way.
我并不是说这不好。
It’s just quite different, right?
完全不同，对吧？
You do the Jared Kushner thing?
你做贾里德·库什纳的事吗？
“Here’s the 25 books I read?” You sit down and you ask GPT-6 how to understand this culture, or you bring in three experts?
“这是我读过的 25 本书？”你坐下来问 GPT-6 如何理解这种文化，或者你引入了三位专家？
We bring in experts.
我们引进专家。
You bring in experts.
你请来了专家。
We talk to the US government a lot.
我们与美国政府进行了很多交谈。
We bring in experts.
我们引进专家。
Again, if we’re building a data center that a very trusted partner is going to operate, we know what the workload is, and it’s being built like a US embassy or US military base, we have a very different set of questions than if we were doing other things, which we have not yet decided to do, and we’d bring in more experts for.
再说一次，如果我们正在建设一个由非常值得信赖的合作伙伴来运营的数据中心，我们知道工作量有多大，而且它的建设就像美国大使馆或美国军事基地一样，那么我们会遇到一系列与我们正在做其他事情截然不同的问题，而我们还没有决定这样做，我们会为此聘请更多专家。
Those are quite intangible forms of knowledge, often.
这些通常是无形的知识形式。
How good do you think GPT-6 is at teaching you those things, or you still need the human experts to come in?
您认为 GPT-6 在教您这些事情方面有多好，或者您仍然需要人类专家的介入？
Because you could just ask your own model, right?
因为你可以问你自己的模型，对吗？
I don’t think GPT-6 will have those intangibles.
我认为 GPT-6 不会拥有这些无形资产。
It might surprise us, but that’d be very unexpected if I was like, “Oh, don’t need to talk to experts anymore.”
这可能会让我们感到惊讶，但如果我说“哦，不需要再和专家交谈了”，那就太出乎意料了。
Do you have an evaluation for that in the works?
您对此作品有评价吗？
Actually, for something very close to that, we do.
事实上，对于非常接近的事情，我们是这样做的。
I don’t want to pre-announce it.
我不想提前宣布。
That class of stuff, yes, we do have an evaluation.
是的，我们确实对这类东西进行了评估。
You do?
你做？
Yes.
是的。
Yes.
是的。
How good will GPT-6 be at poetry?
GPT-6 在诗歌方面的表现如何？
How good do you think GPT-5 is at poetry?
您认为 GPT-5 在诗歌方面有多出色？
Not that good.
没那么好。
It’s not what I want it for, so that’s not a complaint.
这不是我想要的，所以这不是抱怨。
My guess is, in a year, you’ll have some model that can write a poem as good as the median Pablo Neruda poem, but not the best.
我的猜测是，一年后，你就会有一些模型可以写出一首与巴勃罗·聂鲁达的中等诗歌一样好的诗，但不是最好的。
I was going to say, I don’t want to say GPT, whether it’s 6 or 7, but I think we will get to something where you will say, “This is not a long way to the very best, but this is like a real poet’s okay poem.”
我想说，我不想说 GPT，无论是 6 还是 7，但我想我们会达到这样的程度，你会说，“这距离最好的路并不远，但这就像一首真正诗人的好诗。”
In my view, there’s a big gap between a Neruda poem that’s a 7 on a scale of 1 to 10 and one that’s a 10.
在我看来，一首在 1 到 10 的范围内得分为 7 的聂鲁达诗和一首得分为 10 的聂鲁达诗之间存在很大差距。
I’m not sure you’ll ever reach the 10.
我不确定你能不能达到第 10 名。
I think you’ll reach the 8.8 within a few years.
我想几年之内你就会达到8.8。
I think we will reach the 10, and you won’t care.
我想我们会达到第十名，你不会在意的。
Who won’t care?
谁会不关心呢？
You won’t care.
你不会在意的。
I’ll care.
我会关心的。
I promise.
我保证。
You’ll care in terms of the technological accomplishment, but in terms of the great pieces of art and emotion and whatever else produced by humanity, you care a lot about the person or that a person produced it.
你会关心技术成就，但就伟大的艺术作品和情感以及人类创造的其他任何东西而言，你非常关心这个人或一个人创造了它。
It’s definitely something for an AI to write a 10 on its technical merits.
对于人工智能来说，在其技术优点上打 10 分绝对是一件好事。
My classic example of this is, the greatest chess players don’t really care that AI is hugely better than them at chess.
我的经典例子是，最伟大的国际象棋棋手并不真正关心人工智能在国际象棋方面比他们好得多。
It doesn’t demotivate them to play.
这不会影响他们玩游戏的积极性。
They don’t really care that they are.
他们并不真正关心自己是不是。
They really care about beating the other human, and they really get obsessed with that dude sitting across from them.
他们真的很想打败另一个人，他们真的很着迷于坐在他们对面的那个家伙。
The fact that the AI is better, they don’t care.
事实上，人工智能更好，但他们不在乎。
Watching two AIs play each other, not that fun for that long.
看着两个人工智能互相对战，持续那么久并不是那么有趣。
Let me tell you my worry about reaching the 10.
让我告诉你我对达到10名的担忧。
Evaluations rely a lot on these rubrics.
评估在很大程度上依赖于这些准则。
The rubrics will become good enough to produce very good poems, but maybe there’s something about the 10 poem that stands outside the rubric.
这些标题将变得足够好，可以创作出非常好的诗歌，但也许第 10 首诗有一些东西超出了标题。
If you’re just training on rubrics, rubrics, rubrics, it might in a way be counterproductive for reaching the 10.
如果你只是在评分标准、评分标准、评分标准上进行训练，那么在某种程度上可能会对达到 10 分起到反作用。
Evals can rely on a lot of things, including when you call upon the 10 and when you don’t.
评估可以依赖很多事情，包括何时调用 10 以及何时不调用。
Sure.
当然。
You can read a bunch in the process and provide some real-time signal.
你可以在这个过程中读取一堆数据并提供一些实时信号。
Say we have no human poets today writing 10s, and we’re asking those same people to judge and grade the GPTs.
假设我们今天没有人类诗人写 10 分，我们要求这些人对 GPT 进行评判和评分。
I’m worried.
我很担心。
Again, I think it will be fine.
再说一遍，我想一切都会好起来的。
To me, we’re talking about a 9, not a 10.
对我来说，我们谈论的是 9，而不是 10。
You don’t have William Wordsworth working for OpenAI.
你没有威廉·华兹华斯 (William Wordsworth) 为 OpenAI 工作。
This gets to a very interesting thing, which is, let’s say you can’t write a 10, but you can decide when something is a 10.
这就涉及到一件非常有趣的事情，也就是说，假设你不能写 10，但你可以决定什么时候某件事是 10。
That might be all that we need.
这可能就是我们所需要的。
Maybe humanity only decides collectively what’s a 10, and there’s something a little mysterious and history-laden about that process.
也许人类只能集体决定什么是 10，这个过程有点神秘且历史悠久。
Okay, but still, we can do it.
好吧，但我们仍然可以做到。
Now, maybe our decision is not very good because it is history-related and it does drift over time, and some things we all agree are great, the next generation decides they’re not, whatever, but if whatever process humanity has to determine what poem is a 10, you could imagine that providing some sort of signal to an AI.
现在，也许我们的决定不是很好，因为它与历史相关，而且确实会随着时间的推移而发生变化，有些事情我们都同意很棒，但下一代决定它们不是，无论如何，但如果人类必须通过任何过程来确定哪首诗是 10，你可以想象向人工智能提供某种信号。
Now that, again, if you know it’s an AI, maybe you don’t care.
现在，再说一遍，如果你知道它是人工智能，也许你就不在乎。
We see this phenomenon with AI art, but yes.
我们在人工智能艺术中看到了这种现象，但是是的。
On chip-building
芯片制造方面
To the extent you end up building your own chips, what’s the hardest part of that?
就您最终构建自己的芯片而言，最困难的部分是什么？
Man, that’s a hard thing all around.
伙计，这真是一件很难的事情。
That’s the hardest.
这是最难的。
There’s no easy part of that.
这并不容易。
Yes, no easy part of that.
是的，这并不容易。
Well, Jonathan Ross said it’s just keeping up with what is new.
乔纳森·罗斯说，它只是跟上新事物的步伐。
No, I’ll tell you why.
不，我会告诉你原因。
People talk a lot about the recursive self-improvement loop for AI research, where AI can help researchers, maybe today, write code faster, eventually do automated research, and this thing is well understood, very much discussed.
人们谈论了很多关于人工智能研究的递归自我完善循环，人工智能可以帮助研究人员，也许今天，更快地编写代码，最终进行自动化研究，这件事很好理解，也有很多讨论。
Very little discussed or relatively little discussed are the hardware implications of this: robots that can build other robots, data centers that can build other data centers, chips that can design their own next generation.
很少讨论或相对较少讨论的是这种情况的硬件影响：可以建造其他机器人的机器人，可以建造其他数据中心的数据中心，可以设计自己的下一代的芯片。
There’s many hard parts, but maybe a lot of them can get much easier.
有很多困难的部分，但也许其中很多都可以变得更容易。
Maybe the problem of chip design will turn out to be a very good problem for previous generations of chips.
也许芯片设计的问题对于前几代芯片来说会是一个非常好的问题。
The stupidest question possible: Why don’t we just make more GPUs?
最愚蠢的问题是：我们为什么不生产更多 GPU？
Because we need to make more electrons.
因为我们需要制造更多的电子。
What’s stopping that?
是什么阻止了这一点？
What’s the ultimate binding constraint?
最终的约束约束是什么？
We’re working on it really hard.
我们正在非常努力地努力。
If you could have more of one thing to have more compute, what would the one thing be?
如果你可以拥有更多的一件事来获得更多的计算能力，那一件事会是什么？
Electrons.
电子。
Electrons.
电子。
Just energy.
只是能量。
What’s the most likely short-term solution for that?
最有可能的短期解决方案是什么？
Short-term, natural gas.
短期，天然气。
Easing, not full solution, but easing of the constraint.
放松，不是完全解决，而是放松约束。
Short-term, natural gas.
短期，天然气。
In the American South?
在美国南部？
Or wherever.
或者无论在哪里。
Long-term, it will be dominated, I believe, by fusion and by solar.
我相信，从长远来看，它将由核聚变和太阳能主导。
I don’t know what ratio, but I would say those are the two winners.
我不知道比例是多少，但我想说他们是两个赢家。
You’re still bullish on fusion?
您仍然看好融合吗？
Very much, and solar.
非常多，还有太阳能。
Do you worry that, as long as it’s called nuclear power, even if it works —
你担心吗，只要叫核电，就算有用 ——
Did I say the word nuclear?
我有说过“核”这个词吗？
No, you didn’t, but other people will.
不，你没有，但其他人会。
That people just won’t want it, getting back to the irrationality point in the insurance?
人们就是不想要它，回到保险的非理性点？
You’re the economist, not me, but I think there is some price point at a given level of safety where the demand for this will be overwhelming.
你是经济学家，而不是我，但我认为在给定的安全水平上存在某个价格点，对此的需求将是压倒性的。
If this is the same price as natural gas, maybe it’s unfortunately hard.
如果这与天然气的价格相同，不幸的是可能会很难。
If it’s one-tenth the price, I think we could agree it would happen very fast.
如果价格是十分之一，我想我们可以同意这会很快发生。
I don’t know what the cut point is between.
我不知道两者之间的切点是什么。
Do you ever worry there’s some scenario where, ultimately, superintelligence doesn’t need that much compute, and in some funny way, by investing in compute, you’re betting against progress over a 30-year time horizon?
您是否担心在某些情况下，最终超级智能不需要那么多计算，并且以某种有趣的方式，通过投资计算，您押注在 30 年的时间范围内不会取得进展？
The related thing, in the same way that people always want more energy if it’s cheaper, I think people will always want more compute if it’s cheaper.
相关的事情，就像人们总是想要更多更便宜的能源一样，我认为如果更便宜的话人们总是想要更多的计算。
Even if you can make incredibly smart models with much less compute, which I’m sure you can, the desire to consume in all sorts of new ways and do more stuff with more abundant intelligence, I’ll take that bet every day.
即使你可以用更少的计算量来制作令人难以置信的智能模型（我相信你可以），以及以各种新方式消费并用更丰富的智能做更多事情的愿望，我每天都会赌这个赌注。
The related thing I worry about is that there is a huge phase shift on how we do compute, and we’re all chasing a dead-end paradigm.
我担心的相关问题是，我们的计算方式发生了巨大的相变，而且我们都在追逐一个死胡同。
That would be bad.
那会很糟糕。
What would that look like?
那会是什么样子？
I don’t know.
我不知道。
We all switch to optical compute, like full-on optical compute or something.
我们都转向光学计算，比如全面的光学计算之类的。
And just have to spend a lot of money all over again?
还得重新花很多钱吗？
Yes.
是的。
Well, not on all of it.
嗯，不是全部。
The energy is the energy, but yes, on everything else.
能量就是能量，但是，是的，还有其他一切。
Now, I love Pulse.
现在，我喜欢脉冲。
Why don’t I hear more about Pulse, or do you think there is a lot of chatter out there?
为什么我没有听到更多有关 Pulse 的信息，或者您认为那里有很多讨论？
People love Pulse, but it is only available to our Pro users right now, which is not that many.
人们喜欢 Pulse，但它目前仅适用于我们的专业用户，而且数量并不多。
Also, we’re not giving much per day to users.
此外，我们每天为用户提供的服务并不多。
We will change both of those things.
我们将改变这两件事。
I suspect when we roll it out to Plus, you will hear about it a lot more, but people do love it.
我怀疑当我们将其推广到 Plus 时，您会听到更多关于它的信息，但人们确实喜欢它。
It gets great reviews.
它得到了很好的评价。
What do you use Pulse for?
你用脉冲做什么？
There are only two things in my life right now.
现在我的生活中只有两件事。
There’s my family and work.
那里有我的家人和工作。
Clearly, this is what I talk to ChatGPT about because I get a lot of stuff about that.
显然，这就是我与 ChatGPT 谈论的内容，因为我得到了很多相关内容。
I get the odd “New hypercar came out” or “Here’s a great hiking trail” or whatever, but it’s mostly those two things.
我听到奇怪的“新超级跑车问世了”或“这是一条很棒的远足小径”之类的东西，但主要是这两件事。
It’s great for both of those.
这对两者来说都很棒。
On Sam’s outlook on health, alien life, and conspiracy theories
论山姆的健康观、外星生命和阴谋论
I’d just like to do a brief interlude on your broader view of the world and just see how I should think about how you think.
我只想简单介绍一下您对世界的更广阔的看法，看看我应该如何看待您的想法。
People in California, they have a lot of views on their own health, some of which to me sound nutty.
加州人对自己的健康有很多看法，其中一些对我来说听起来很疯狂。
What do you think is your nuttiest view about your own health?
您认为您对自己的健康最疯狂的看法是什么？
That you’re going to live forever, that the seed oils are bad, or what is it, or do you not have any?
你会永远活下去，种子油不好，或者是什么，或者你没有种子油？
When I was less busy, I was more disciplined on health-related stuff.
当我不那么忙的时候，我在与健康相关的事情上更加自律。
I didn’t have crazy views, but I was like, I kind of ate healthy.
我没有疯狂的观点，但我想，我吃得健康。
I didn’t drink that much.
我没喝那么多。
I worked out a lot.
我锻炼了很多。
I tried a few things here and there.
我到处尝试了一些事情。
I once ended up in a hospital for trying semaglutide before it was cool, that kind of stuff.
我有一次因为在索马鲁肽还没有凉之前就尝试过这种药物而住进了医院。
I now do basically nothing.
我现在基本上什么都不做。
You just live family life.
你只是过着家庭生活。
I eat junk food.
我吃垃圾食品。
I don’t exercise enough.
我运动量不够。
It’s like a pretty bad situation.
这就像一个非常糟糕的情况。
I’m feeling bullied into taking this more seriously again.
我感觉自己被欺负了，不得不再次更加认真地对待这件事。
Yes, but why eat junk food?
是的，但是为什么要吃垃圾食品呢？
It doesn’t taste good.
味道不太好。
It does taste good.
味道确实不错。
Compared to good sushi?
与好的寿司相比？
You could afford good sushi.
你买得起美味的寿司。
Someone will bring it to you.
有人会把它带给你。
A robot will bring it.
机器人会带来它。
Sometimes late at night, you just really want that chocolate chip cookie at 11:30 at night, or at least I do.
有时在深夜，你真的很想在晚上 11:30 吃一块巧克力曲奇饼，或者至少我想。
Yes.
是的。
Do you think there’s any kind of alien life on the moons of Saturn?
你认为土星的卫星上存在外星生命吗？
Because I do.
因为我愿意。
That’s one of my nutty views.
这是我的疯狂观点之一。
I have no opinion on the matter.
我对此事没有意见。
No opinion on the matter.
对此事没有意见。
I don’t know.
我不知道。
Yes, that’s a way of passing the test.
是的，这是通过测试的一种方式。
What do you think about UAPs?
您对 UAP 有何看法？
Do you think there’s a thing?
你觉得有事吗？
I think something’s going on there.
我认为那里发生了一些事情。
I don’t know.
我不知道。
You think something is going on there.
你认为那里正在发生一些事情。
Again, I have an opinion that there is something that I would like an explanation for.
再次，我有一个观点，我想对某些事情做出解释。
I kind of doubt it’s little green men.
我有点怀疑这是小绿人。
I extremely doubt it’s little green men, but I think someone’s got something.
我非常怀疑这是小绿人，但我认为有人有东西。
How many conspiracy theories do you believe in?
你相信多少种阴谋论？
Because I believe in close to zero, at least in the United States.
因为我相信接近于零，至少在美国是这样。
They may be true for Pakistani military coups, but I think mostly they’re just false.
对于巴基斯坦军事政变来说，它们可能是正确的，但我认为它们大多是错误的。
Like true conspiracy theory, not just an unpopular belief?
就像真正的阴谋论一样，而不仅仅是一种不受欢迎的信仰？
Correct.
正确的。
I have one of those X-Files shirts, like “I want to believe.” I still have one of those shirts from when I was in high school.
我有一件《X档案》衬衫，上面写着“我想相信”。我至今还保留着一件我高中时穿的衬衫。
I want to believe in conspiracy.
我愿意相信阴谋。
I’m predisposed to believe in conspiracy theories, and I believe in either zero or very few.
我倾向于相信阴谋论，而且我相信的阴谋论要么为零，要么很少。
Yes, I’m the opposite of that.
是的，我恰恰相反。
I don’t want to believe, and I believe in very few.
我不想相信，而且我相信的人很少。
Like maybe the White Sox fixed the World Series way back when, right?
就像白袜队很久以前就解决了世界大赛一样，对吧？
Yes, stuff like that, I don’t quite count that.
是的，诸如此类的事情，我不太数。
Like a true massive global government cover-up, that requires a level of competence to people I just rarely ascribe.
就像真正大规模的全球政府掩盖行动一样，这需要我很少认可的人具备一定的能力。
Now, some number of years ago, this was before even GPT-4, I asked you if you were directing a fund of money to revitalize St. Louis, which is where you grew up, how would you invest the money?
现在，几年前，甚至在 GPT-4 之前，我问你，如果你正在领导一个基金来振兴圣路易斯，这是你长大的地方，你会如何投资这笔钱？
Now it’s a quite different world from when I asked you last time.
现在的世界与我上次问你时完全不同了。
If I ask you again to revitalize St. Louis, how would you spend the money?
如果我再问你振兴圣路易斯，你会怎么花这笔钱？
Say it’s a billion dollars, which is not actually transformational, but it’s enough that it’s some real money.
说是十亿美元，这实际上并不算什么变革，但有一些真金白银就足够了。
A billion dollars, and I’m willing to go spend personal time on it?
十亿美元，我愿意花私人时间吗？
You have free time.
你有空闲时间。
The universe grants you free time.
宇宙给予你自由的时间。
You don’t take time away from anything else you’re doing.
你不会从你正在做的其他事情中抽出时间。
You’re in charge.
你负责。
This is not a deeply incisive answer because I think this is not a generally replicatable thing, but unique to me, what I could do.
这不是一个深刻精辟的答案，因为我认为这不是一个普遍可复制的事情，而是对我来说独一无二的，我能做的。
I think I would try to go start a Y Combinator-like thing in St. Louis and get a ton of startup founders focused on AI to move there and start a bunch of companies.
我想我会尝试在圣路易斯创办一个类似 Y Combinator 的公司，让大量专注于人工智能的初创公司创始人搬到那里并创办一堆公司。
That’s a pretty similar answer to last time.
这和上次的答案非常相似。
I didn’t remember what I said last time, so that’s a good sign.
我不记得上次说了什么，所以这是一个好兆头。
You said the same thing, but you didn’t mention AI.
你也说了同样的话，但你没有提到人工智能。
AI to me seems quite clustered where we are in the Bay Area.
对我来说，人工智能似乎非常集中在我们湾区的地方。
Is trying to get AI into St. Louis the right way to do that?
尝试将人工智能引入圣路易斯是正确的方法吗？
Isn’t that, in a way, working at cross-purposes?
从某种程度上来说，这不是在相互矛盾吗？
This is why I said it’d be like a “unique to me” thing.
这就是为什么我说这对我来说是“独一无二的”。
I think I could do it.
我想我能做到。
Maybe that’s hopelessly naive.
也许这是无可救药的天真。
On regulating AI agents
关于监管人工智能代理
Yes.
是的。
Should it be legal to just release an AI agent into the wild, unowned, untraceable?
将人工智能代理释放到野外、无人拥有、无法追踪是否合法？
Do we need some other AI agent to go out there and tackle it down, or is there minimum capitalization?
我们是否需要其他人工智能代理来解决这个问题，或者是否有最低资本要求？
How do you think about that problem?
您如何看待这个问题？
I think it’s a question of thresholds.
我认为这是一个门槛问题。
I don’t think you’d advocate that most systems should have any oversight or regulation or legal questions or whatever, but if we have an agent that is capable with serious probability of massively self-replicating over the internet and sweeping all the money out of bank accounts or whatever, you would then say, “Okay, maybe that one needs some oversight.” I think it’s a question of where you draw the threshold for where it should not be.
我不认为你会主张大多数系统应该有任何监督、监管或法律问题或其他什么问题，但如果我们有一个代理，很有可能通过互联网进行大规模自我复制，并将所有钱从银行账户或其他什么地方清除，那么你会说，“好吧，也许那个人需要一些监督。”我认为问题在于你在哪里划定不应该划定的门槛。
Say it’s hiring the cloud computing from a semi-rogue nation, so you can’t just turn it off.
假设它正在从一个半流氓国家租用云计算，那么你就不能直接将其关闭。
What actually should we do, or will we be able to do?
我们实际上应该做什么，或者我们能够做什么？
Just try to ring-fence it somehow, identify it, surveil it, put sanctions on the country that’s sponsoring it?
只是尝试以某种方式围堵它，识别它，监视它，对赞助它的国家实施制裁？
What do we do for people that do that today?
今天我们为那些这样做的人做什么？
Well, there are a lot of cyberattacks that come from North Korea, and I think we can’t do that much about them, right?
嗯，有很多网络攻击来自朝鲜，我认为我们对此无能为力，对吧？
I don’t know what the right answer is yet, but my naive take is we should try to solve this problem urgently for people using rogue internet resources, and AI will just be a worse version of that problem.
我还不知道正确的答案是什么，但我天真的想法是，我们应该努力为使用流氓互联网资源的人们紧急解决这个问题，而人工智能将只是这个问题的一个更糟糕的版本。
It will have better defense, also.
它也会有更好的防御。
For sure.
一定。
On new ways to interface with AI
与人工智能交互的新方法
Yes.
是的。
Now, if I think about social media and AI, here’s one thing I’ve noticed in my own work.
现在，如果我考虑社交媒体和人工智能，我在自己的工作中注意到一件事。
I’m so, so keen to read the answers to my own queries to GPT-5, but when people send me the answers to their queries, I’m bored.
我非常非常热衷于阅读我自己对 GPT-5 的疑问的答案，但是当人们向我发送他们的疑问的答案时，我感到很无聊。
I don’t blame them.
我不怪他们。
I know it’s super useful for them, but that makes me a little skeptical about blending social media and AI.
我知道这对他们来说非常有用，但这让我对社交媒体和人工智能的融合有点怀疑。
Am I missing something, or would you try to talk me out of that somehow?
我是否遗漏了什么，或者你会试图以某种方式说服我吗？
No, I’ve had the same observation.
不，我也有同样的观察。
I don’t want to read your GPT-5 queries, either.
我也不想阅读您的 GPT-5 查询。
Yes, but they’re great for me.
是的，但它们对我来说很棒。
I’m sure, and I’m sure you don’t want to read mine, but they’re great for me.
我确信，而且我确信你不想读我的，但它们对我来说很棒。
ChatGPT, I think, is very much like a single-player experience.
我认为 ChatGPT 非常类似于单人游戏体验。
I don’t think that means there’s not some interesting, new kind of social product to build.
我不认为这意味着没有一些有趣的、新型的社交产品可以构建。
In fact, I’m pretty sure there is, but I don’t think it’s the “Share your ChatGPT queries.”
事实上，我很确定有，但我不认为这是“分享你的 ChatGPT 查询”。
Videos or any sense of what that would look like?
视频或任何感觉会是什么样子？
That one’s doing well.
那个人做得很好。
People, clearly, they love making their own, but they also like watching other people’s AI-generated videos.
显然，人们喜欢自己制作，但他们也喜欢观看其他人的人工智能生成的视频。
No, I think none of this stuff is the really interesting kind of things you can imagine when you and I and everybody else have really great personal AI agents that can do stuff on our behalf.
不，我认为当你和我以及其他人拥有真正出色的个人人工智能代理可以代表我们做事时，这些东西都不是你能想象的真正有趣的事情。
There’s probably entirely new social dynamics to think about.
可能需要考虑全新的社会动态。
Just the physical form of ChatGPT on my screen or on my smartphone, is that more or less going to stay the same, but the thing will be better, or 13 years from now, it will physically just be an entirely different beast?
只是 ChatGPT 在我的屏幕或智能手机上的物理形式，或多或少会保持不变，但情况会更好，或者 13 年后，它在物理上将只是一个完全不同的野兽？
Because I can talk to it now.
因为我现在可以和它说话了。
Now it does video.
现在它可以做视频了。
Is it just a better version, or somehow it morphs?
它只是一个更好的版本，还是以某种方式发生了变化？
We are going to try to make you a new kind of computer with a completely new kind of interface that is meant for AI, which I think wants something completely different than the computing paradigm we’ve been using for the last 50 years that we’re currently stuck in.
我们将尝试为您打造一种具有全新界面的新型计算机，专为人工智能而设计，我认为人工智能需要与我们目前所使用的过去 50 年所使用的计算范式完全不同的东西。
AI is a crazy change to the possibility space.
人工智能是对可能性空间的疯狂改变。
A lot of the basic assumptions of how you use a computer and the fact that you should even be having an operating system or opening a window, or sending a query at all, are now called into question.
关于如何使用计算机的许多基本假设，以及您甚至应该拥有操作系统、打开窗口或发送查询的事实，现在都受到了质疑。
I realize that the track record of people saying they’re going to invent a new kind of computer is very bad.
我意识到人们说他们要发明一种新型计算机的记录非常糟糕。
If there’s one person that you should bet on to do it, I think Jony Ive is a credible, maybe the best bet you could take.
如果你应该押注一个人来做这件事，我认为乔尼·伊维是一个可靠的人，也许是你可以押注的最佳人选。
We’ll see if it works.
我们会看看它是否有效。
I’m very excited to try.
我很高兴尝试。
Haven’t you already been surprised, how robust it is that people love typing text into boxes?
您是否已经感到惊讶，人们如此喜欢在框中输入文本？
This sort of shocks me in the bigger picture.
从更大的角度来看，这让我感到震惊。
People are still texting all the time.
人们仍然一直在发短信。
It’s one of the most robust forms of internet anything.
它是最强大的互联网形式之一。
Maybe that will just stick forever, and it’s a sign of our own limitations.
也许这种情况会永远存在，这是我们自身局限性的标志。
How do we get past that?
我们如何克服这一点？
Texting, command lines, search queries, that’s my favorite interface.
短信、命令行、搜索查询，这是我最喜欢的界面。
You like it, I like it.
你喜欢，我也喜欢。
Maybe we’re just going to keep it.
也许我们只会保留它。
Well, a lot of people use it.
嗯，很多人都用它。
People love to text.
人们喜欢发短信。
People like ChatGPT.
人们喜欢 ChatGPT。
I remember when we were thinking about the interface for ChatGPT, I was very set that this was something people would be familiar with and want to use.
我记得当我们考虑 ChatGPT 的界面时，我非常确定这是人们会熟悉并想要使用的东西。
I think I grew up as a child of the internet with a lot of conviction that that was the right kind of internet.
我想我是作为一个互联网的孩子长大的，我坚信这才是正确的互联网。
Texting was my life as a teenager.
发短信是我青少年时期的生活。
On how normies will learn to use AI
关于普通人如何学习使用人工智能
If you have some kind of ideal arrangement, partnership with an institution of higher education, say, within two to three years, what does that look like?
如果你有某种理想的安排，与高等教育机构合作，比如说，在两到三年内，那会是什么样子？
You get to write the whole thing.
你可以把整个事情写下来。
I suspect that the whole model should change, but I don’t know what to.
我怀疑整个模型应该改变，但我不知道该怎么做。
I think the ideal partnership would look like we try 20 different experiments.
我认为理想的合作伙伴关系应该是我们尝试 20 种不同的实验。
We see what leads to the best results.
我们看看什么会带来最好的结果。
I’ve been watching these AI schools pop up with great interest.
我一直饶有兴趣地看着这些人工智能学校的出现。
It seems like a lot of them with very different approaches are all showing positive results.
看起来他们中的许多人采用了截然不同的方法，但都显示出了积极的结果。
I think the first few years of the ideal partnership would look like we run 20 wildly different experiments.
我认为理想合作关系的最初几年看起来就像我们进行了 20 个截然不同的实验。
Sometimes I have the fear that these institutions don’t have enough internal reputational strength or credibility to make any major change, forget about AI, and that to do a partnership with an institution like that is maybe intrinsically frustrating, and for the next 10 years, the actual model is privatized AI use on the side by faculty, by students, by labs, and in a sense there is no partnership other than actually just marketing your product to these people.
有时我担心这些机构没有足够的内部声誉实力或可信度来做出任何重大改变，忘记人工智能，与这样的机构建立合作伙伴关系可能本质上令人沮丧，在接下来的十年里，实际的模式是教师、学生、实验室私有化人工智能的使用，从某种意义上说，除了向这些人推销你的产品之外，没有任何合作关系。
Do you ever think that might be true?
你有没有想过这可能是真的？
Yes, and it wouldn’t super upset me if that’s what happens.
是的，如果发生这种情况，我也不会感到非常不安。
Yes.
是的。
What do you think will happen to the returns to a college degree?
您认为大学学位的回报会发生什么变化？
Not Harvard, not Stanford, but like a quite good state school, five, 10 years out.
不是哈佛大学，不是斯坦福大学，而是一所相当不错的公立学校，五年、十年后。
What’s the historical rate of decline of the value of that?
其价值的历史下降速度是多少？
Recently, it’s gone down, though for a long time it was going up quite a bit.
最近一段时间以来，它有所下降，但很长一段时间以来，它一直在上升。
Yes, I mean the last decade.
是的，我的意思是过去十年。
Oh, gone down.
哦，掉下去了。
I don’t know how much.
我不知道有多少。
I would guess that it goes down at a slightly higher rate than the last decade, but it does not collapse to zero as fast as it should.
我猜想它下降的速度比过去十年略高，但它并没有像应有的那样快地崩溃到零。
Then it’s the returns to doing what other than learning AI that go up?
那么除了学习人工智能之外，做其他事情的回报会上升吗？
Like being on the college football team or what?
比如加入大学橄榄球队或者什么？
I don’t think the massive returns will accrue to doing AI for a small set of people, but I think the returns to using AI super well will be surprisingly widely distributed.
我不认为为一小部分人做人工智能会带来巨大的回报，但我认为很好地使用人工智能的回报将令人惊讶地分布广泛。
I think the most important thing AI will do is discover new science for all of us, and a lot of people will benefit from that, and people will start companies or get jobs doing that.
我认为人工智能所做的最重要的事情是为我们所有人发现新的科学，很多人将从中受益，人们会为此创办公司或找到工作。
I’m not a believer that that is the only thing that eventually makes money.
我不相信这是唯一最终能赚钱的事情。
I think people will just use AI for all sorts of new kinds of jobs or to do existing jobs better.
我认为人们只会将人工智能用于各种新的工作或更好地完成现有的工作。
Maybe the starkest example of this in 2025 is the day-to-day of how the average programmer in Silicon Valley did their workflow at the beginning of this year versus the end of this year.
也许 2025 年最明显的例子是硅谷普通程序员在今年年初和今年年底如何执行工作流程的日常情况。
Extremely different.
极其不同。
Extremely different.
极其不同。
You don’t really have to know how to program AI to do that, but you can get more done, and you probably have much more value, and the world is going to get much more software.
你实际上不必知道如何对人工智能进行编程来做到这一点，但你可以完成更多工作，并且你可能拥有更多价值，而且世界将获得更多软件。
I think we’ll see things like that for a surprising number of industries.
我认为我们将会在数量惊人的行业中看到类似的情况。
Say five years out, there’s a so-called normie person.
假设五年后，就会有一个所谓的正常人。
They’re not a specialist.
他们不是专家。
They want to learn how to use AI much better.
他们想学习如何更好地使用人工智能。
What will they actually do that will give them a high return to acquiring that skill?
他们实际上会做什么才能让他们获得该技能的高回报？
To learn how to use AI specifically?
具体要学习如何使用AI？
Yes.
是的。
Not program, not the inner guts, just actually in their job.
不是计划，不是内在的胆量，只是他们实际的工作。
I’m smiling because I remember when I was a kid and Google came out there, I had a job teaching older people how to use Google, and it felt like I just couldn’t wrap my head around it.
我微笑是因为我记得当我还是个孩子的时候，谷歌就出现了，我有一份工作教老年人如何使用谷歌，但感觉我就是无法理解它。
It was like, “You type the thing in, and it does this.” A thing that I’m hopeful about for AI is that, I think one of the reasons ChatGPT has grown so fast, is it is so easy to learn how to use it and get real value out of it.
就像是，“你输入一些东西，它就会这样做。”我对人工智能充满希望的一件事是，我认为 ChatGPT 发展如此迅速的原因之一是学习如何使用它并从中获得真正的价值是如此容易。
We don’t need startups to do that?
我们不需要初创公司来做这件事吗？
To teach people how to use AI?
教人们如何使用人工智能？
To teach people, yes.
教导人们，是的。
There is such a startup, or what’s the institution?
有这样的初创公司，或者机构是什么？
My school will teach me, that’s hard to believe.
我的学校会教我，这很难相信。
10% of the world will use ChatGPT this week didn’t exist three years ago.
本周全球 10% 的人将使用 ChatGPT，这在三年前还不存在。
I suspect a year from now, maybe 30% of people will use ChatGPT that week.
我怀疑一年后，那周可能有 30% 的人会使用 ChatGPT。
People, once they start using it, do find more and more sophisticated things to use it for.
人们一旦开始使用它，就会发现越来越复杂的用途。
This is not a top-of-mind problem for me.
这对我来说不是首要问题。
I think I believe in human creativity and adoption of new things over some period of number of years.
我想我相信人类的创造力和在一段时间内对新事物的采用。
You might just want to support or invest in the startups that will do this because if you’re bullish on AI, presumably you’re bullish on those startups, and it will help your business in turn.
你可能只想支持或投资能够做到这一点的初创公司，因为如果你看好人工智能，那么你大概也看好这些初创公司，而这反过来又会对你的业务有所帮助。
It seems odd not to have a theory of how we’re all going to learn to use AI better.
没有一个关于我们如何学习更好地使用人工智能的理论似乎很奇怪。
You can go to dog trainer school, and they teach you how to train a dog.
你可以去训狗学校，他们会教你如何训练狗。
Okay.
好的。
Maybe I have a blind spot here, and I promise I’ll go think about this more.
也许我在这里有一个盲点，我保证我会多考虑一下。
If you ask ChatGPT, “Teach me how to use you” —
如果你问 ChatGPT，“教我如何使用你” —
Yes, maybe that’s it.
是的，也许就是这样。
— it’s pretty good.
—— 相当不错。
Yes, so maybe you’re the school.
是的，所以也许你就是学校。
Maybe.
或许。
Yes.
是的。
Let’s say when your kids are old enough that they’re grown, they can go out on their own.
假设当您的孩子足够大时，他们就可以自己出去了。
In that future world, which is not so far off, do you think you’ll still be reading books, or you’ll just be interacting with your AI?
在那个并不遥远的未来世界，你认为你还会读书，还是只会与人工智能互动？
Books have survived a lot of other technological change.
书籍在许多其他技术变革中幸存下来。
I think there is something deep about the format of a book that has persisted.
我认为一本书的格式有一些深刻的东西一直存在。
It’s very mindy, whatever the current word for that is.
它非常用心，无论现在用什么词来形容它。
I suspect that there will be a new way to interact with a cluster of ideas that is better than a book for most things.
我怀疑将会有一种新的方式来与一组想法进行交互，在大多数事情上这比一本书更好。
I don’t think books will be gone, but I would bet they’re a smaller percentage of how people learn a new or interact with a new idea.
我不认为书籍会消失，但我敢打赌，它们在人们学习新事物或与新想法互动的方式中所占的比例较小。
What’s the cultural habit you have that you think will change the most?
您认为最需要改变的文化习惯是什么？
Like, “Oh, I won’t watch movies anymore, I’ll create my own,” or whatever, for you?
比如，“哦，我不会再看电影了，我会自己创作电影”，或者其他什么，对你来说？
AI will obliterate what you did when you were 23.
人工智能会抹掉你23岁时所做的一切。
This is boring, but I think the way I work, where I’m doing emails and calls and meetings and writing documents and dealing with Slack, that I expect to change hugely, and that has become a real, and I have a cultural habit and a rhythm of my workday at this point.
这很无聊，但我认为我的工作方式，我处理电子邮件、打电话、开会、编写文档和处理 Slack，我期望发生巨大的改变，而这已经成为现实，我现在有一种文化习惯和工作日节奏。
Spending time with my family, spending time in nature, eating food, my interactions with my friends, that’s stuff I expect to change, almost not at all, at least not for a very long time.
与家人共度时光、在大自然中度过时光、吃食物、与朋友互动，这些都是我希望改变的事情，但几乎不会改变，至少在很长一段时间内不会改变。
On AI’s effect on the price of housing and healthcare
人工智能对住房和医疗保健价格的影响
Do you think San Francisco will remain the center for AI?
您认为旧金山仍将是人工智能中心吗？
Putting aside China issues, I just mean for the so-called West.
抛开中国问题，我指的是所谓的西方。
Yes, I think that’s the default.
是的，我认为这是默认的。
It’s the default.
这是默认设置。
You think the city is just absolutely making a comeback?
您认为这座城市绝对会卷土重来吗？
It looks much nicer to me.
对我来说看起来好多了。
It seems nicer.
看起来更好看。
Am I deluded?
我被迷惑了吗？
I love the city.
我爱这座城市。
I love the whole Bay Area.
我喜欢整个湾区。
I particularly love the city, but I love the Bay Area.
我特别喜欢这座城市，但更喜欢湾区。
I don’t think I’m a fair person to ask because I so want it to be making a comeback and to remain the place.
我不认为我是一个公平的人来问，因为我非常希望它卷土重来并保持在这个位置。
I think so, I hope so, but very biased.
我想是的，我希望如此，但非常有偏见。
AI will improve many things very quickly, but what’s the time horizon for it making rents or home prices cheaper?
人工智能将很快改善很多事情，但它使租金或房价变得更便宜的时间范围是多长？
That seems like a tough one.
这似乎是一项艰难的任务。
Not the fault of AI, but land is land, and there’s a lot of legal restrictions.
这不是人工智能的错，但土地就是土地，而且有很多法律限制。
Yes, I was going to push back on the land is land.
是的，我要反驳土地就是土地。
There are a lot of other problems that I don’t think AI can solve anytime soon.
我认为人工智能短期内还无法解决很多其他问题。
There could be these very strange second-order effects where home prices get much cheaper, but sadly, I don’t think AI has a direct attack on solving anytime soon.
可能会出现这些非常奇怪的二阶效应，房价会变得便宜得多，但遗憾的是，我认为人工智能不会很快对解决问题产生直接攻击。
Food prices?
食品价格？
I would bet down.
我会打赌。
In the short run, energy might be a bit more expensive.
从短期来看，能源可能会更贵一些。
How long does it take for food prices to go down?
食品价格需要多长时间才能下降？
If they’re not down in a decade, I’d be very disappointed.
如果他们在十年内没有下降，我会非常失望。
If we think of healthcare, my sense is we’re going to spend a lot more on healthcare.
如果我们考虑医疗保健，我的感觉是我们将在医疗保健上花费更多。
We’ll get a lot for it because there’ll be new inventions, but a lot of the world will feel more expensive because rent won’t be cheaper.
我们会得到很多好处，因为会有新发明，但世界上很多地方会感觉更贵，因为租金不会便宜。
Food, I’m not sure about.
食物，我不确定。
Healthcare, you’ll live to age 98, but you’ll have to spend a lot more.
医疗保健，你会活到98岁，但你要花更多的钱。
You’ll just be alive more while you’re spending, right?
当你花钱的时候，你会更活跃，对吗？
Are people just going to think of AI as this very expensive thing, or will it be thought of as a very cheap thing that makes life more affordable?
人们会认为人工智能是一种非常昂贵的东西，还是会被认为是一种非常便宜的东西，可以让生活变得更加便宜？
I would bet we spend less on healthcare.
我敢打赌我们在医疗保健上的支出会减少。
I bet there are a lot of diseases that we can just cure or come up with a very cheap treatment for that right now.
我敢打赌，有很多疾病我们现在就可以治愈，或者想出一种非常便宜的治疗方法。
We have nothing but expensive chronic stuff that doesn’t even work that well.
我们除了昂贵的慢性病之外一无所有，而且效果也不太好。
I would bet healthcare gets cheaper.
我敢打赌医疗保健会变得更便宜。
Through pharmaceuticals, devices?
通过药品、设备？
Through pharmaceuticals and devices, and even delivery of actual healthcare services.
通过药品和设备，甚至提供实际的医疗保健服务。
Look, housing is the one to me that just looks super hard.
听着，住房对我来说看起来非常困难。
There will be other categories of things that we want to get more expensive, and of course, those will — status goods or whatever.
我们还会希望获得更昂贵的其他类别的东西，当然，这些东西会是地位商品或其他什么。
I would take the healthcare goes down a bit.
我认为医疗保健会下降一点。
On reexamining freedom of speech
关于重新审视言论自由
With all the blizzard of new ideas coming, patent law, copyright law, those are based on earlier technologies and earlier models of how the world would work.
随着大量新想法的出现，专利法、版权法，这些都是基于早期的技术和早期的世界运作模式。
Do we need to reexamine or change those radically for an AI-drenched world, or we can just keep what we have and modify it a bit?
我们是否需要重新审视或从根本上改变这些，以适应人工智能充斥的世界，或者我们可以保留现有的并进行一些修改？
I really have no idea.
我真的不知道。
I’m a big free speech advocate, but I can imagine the world saying, “Well, with all this AI-driven content, we need to reexamine the First Amendment.” Do you have a view on that?
我是言论自由的大力倡导者，但我可以想象世界会说：“好吧，有了所有这些人工智能驱动的内容，我们需要重新审查第一修正案。”您对此有何看法？
Without thinking much, I put out a tweet recently about how we’re going to be allowing more freedom of expression in ChatGPT.
我最近没有多想，就发布了一条推文，内容是关于我们将如何在 ChatGPT 中允许更多的言论自由。
This is the famous erotica tweet.
这就是著名的色情推文。
It’s funny what people get upset about.
人们感到不安的事情很有趣。
It is funny what animates people.
有趣的是，是什么让人们充满活力。
Because all you’re saying is you’re not going to stop people, right?
因为你只是说你不会阻止别人，对吗？
Well, we used to not a long time — well, that’s not totally fair.
好吧，我们过去的时间并不长——好吧，这并不完全公平。
We’re going to allow more than we did in the past.
我们将允许比过去更多的事情。
A very important principle to me is that we treat our adult users like adults and that people have a very high degree of privacy with their AI, which we need legal change for, and also that people have very broad bounds of how they’re able to use it.
对我来说，一个非常重要的原则是，我们像对待成年人一样对待我们的成年用户，人们对他们的人工智能拥有非常高的隐私权，我们需要为此进行法律改变，而且人们对如何使用它有非常广泛的界限。
To me, this should be one of the easiest things to agree on by most people in the tech industry or even most people in the US government.
对我来说，这应该是科技行业大多数人甚至美国政府大多数人最容易达成一致的事情之一。
I kind of dashed this tweet off and closed my computer, and it didn’t even hit my mind that it was going to be, really, a firestorm.
我把这条推文扔掉，关上了电脑，我什至没有想到这会是一场真正的风暴。
It was that we made a decision, which I also think was a fair one, over the summer, that because there were new problems, and that particularly because we wanted to protect teenage users, we were going to heavily restrict ChatGPT, which is also always a very unpopular thing to do, and along with the rolling out of age-gating and some of these mental health mitigations, we were going to bring back and in some cases increase freedom of use for adults.
我们在夏天做出了一个决定，我也认为这是一个公平的决定，因为出现了新问题，特别是因为我们想保护青少年用户，我们将严格限制 ChatGPT，这也始终是一件非常不受欢迎的事情，随着年龄限制和其中一些心理健康缓解措施的推出，我们将恢复并在某些情况下增加成年人的使用自由度。
I was like, “Yes, I’ll tell people that’s coming because the first model update is shipping soon, but this should be a nonissue.” Boy, did I get that one wrong.
我当时想，“是的，我会告诉人们即将到来，因为第一个模型更新即将发布，但这应该不是问题。”男孩，我是不是听错了。
I think maybe it’s just, people don’t believe in freedom of expression as much as they say they do.
我想也许只是人们并不像他们所说的那样相信言论自由。
That was my opinion, yes.
这是我的意见，是的。
Everyone thinks, “Okay, my own freedom of expression, I can handle it.
每个人都想：“好吧，我自己的言论自由，我可以处理。
I need it.
我需要它。
My ideas are all right, but yours — ”
我的想法没问题，但你的 —— ”
For greater privacy rights, is it subpoena power that needs to be changed, or something else in addition?
为了获得更大的隐私权，是传票权需要改变，还是其他什么？
Subpoena power.
传票权。
Well, let me just say, I believe that we should apply as much protection as when you talk to your doctor, your human doctor or your human lawyer, as you do when you talk to your AI doctor or AI lawyer.
好吧，我只想说，我相信我们应该像你与你的医生、人类医生或人类律师交谈时一样提供尽可能多的保护，就像你与人工智能医生或人工智能律师交谈时一样。
Right now, we don’t have that.
现在，我们还没有那个。
Correct.
正确的。
Do you think there’s enough trust in America today for people to trust the AI companies the way we sort of trust doctors, lawyers, and therapists?
你认为今天的美国有足够的信任让人们信任人工智能公司，就像我们信任医生、律师和治疗师一样吗？
By revealed preference, yes.
根据显性偏好，是的。
On humanity’s persuadability
论人性的说服力
Yes, by how many people talk to AI.
是的，取决于有多少人与人工智能交谈。
LLM psychosis, everyone on Twitter today is saying it’s a thing.
LLM 精神病，今天 Twitter 上的每个人都在说这是一件事。
How much a thing is it?
东西多少钱啊？
A very tiny thing, but not a zero thing, which is why we pissed off the whole user base, or most of the user base, by putting a bunch of restrictions in place.
这是一个非常小的事情，但不是零的事情，这就是为什么我们通过设置一系列限制来激怒整个用户群或大部分用户群。
The “treat adult users like adults” includes an asterisk, which is “treat adults of sound mind like adults.” Society decides that we treat adults that are having a psychiatric crisis differently than other adults.
“像成年人一样对待成人用户”包含一个星号，即“像成年人一样对待心智健全的成年人”。社会决定我们对待患有精神危机的成年人与其他成年人不同。
This is one of these things that you learn as you go, but when we saw, the kind of like put ChatGPT into role-playing mode, or pretend like it’s writing a book, and have it encourage someone in delusional thoughts, 99 point some big number percentage of adults, totally fine with that, some tiny percentage of people, also, if they talk to another person who encourages delusion, it’s bad.
这是你边走边学的事情之一，但当我们看到时，就像把 ChatGPT 置于角色扮演模式，或者假装它在写一本书，并让它鼓励处于妄想思维中的人，99 分，大部分成年人对此完全同意，还有一小部分人，如果他们与另一个鼓励妄想的人交谈，那就不好了。
We made a bunch of changes, which are in conflict with the freedom of expression policy.
我们做了很多改变，这些改变与言论自由政策相冲突。
Now that we have those mental health mitigations in place, we’ll again allow some of that stuff in creative mode, role-playing mode, writing mode, whatever, of ChatGPT.
既然我们已经采取了这些心理健康缓解措施，我们将再次允许 ChatGPT 的创意模式、角色扮演模式、写作模式等中的一些内容。
The thing I worry about is not that there will be a few basis points of people that are close to losing grips with reality, and we can trigger a psychotic break, and we can get that right.
我担心的不是人们会在某些基点上接近失去对现实的把握，我们可能会引发精神崩溃，我们可以做到这一点。
The thing I worry about more — it’s funny, the things that stick in your mind.
我更担心的事情 —— 很有趣，那些在你脑海中挥之不去的事情。
Someone said to me once, “Never ever let yourself believe that propaganda doesn’t work on you.
有人曾经对我说过：“永远不要让自己相信宣传对你不起作用。
They just haven’t found the right thing for you yet.” Again, I have no doubt that we can’t address the clear cases of people near a psychotic break.
他们只是还没有找到适合你的东西。”再次强调，我毫不怀疑我们无法解决那些接近精神病的人的明显案例。
For all of the talk about AI safety, I would divide most AI thinkers into these two camps of “Okay, it’s the bad guy uses AI to cause a lot of harm,” or it’s, “the AI itself is misaligned, wakes up, whatever, intentionally takes over the world.”
对于所有关于人工智能安全的讨论，我将大多数人工智能思想家分为这两个阵营：“好吧，是坏人利用人工智能造成了很大的伤害”，或者是“人工智能本身错位了，醒来了，无论如何，故意接管世界。”
There’s this other category, third category, that gets very little talk, that I think is much scarier and more interesting, which is the AI models accidentally take over the world.
还有另外一类，第三类，很少有人谈论，我认为更可怕也更有趣，那就是人工智能模型意外地接管了世界。
It’s not that they’re going to induce psychosis in you, but if you have the whole world talking to this one model, it’s not with any intentionality, but just as it learns from the world in this continually coevolving process, it just subtly convinces you of something.
并不是说它们会引起你的精神病，而是如果你让全世界都在与这个模型交谈，它并不是有意为之，而是就像它在这个不断共同进化的过程中向世界学习一样，它只是巧妙地让你相信某件事。
No intention, it just does.
无意，只是这么做。
It learned that somehow.
它以某种方式了解到这一点。
That’s not as theatrical as chatbot psychosis, obviously, but I do think about that a lot.
显然，这不像聊天机器人精神病那么戏剧化，但我确实经常思考这一点。
Maybe I’m not good enough, but as a professor, I find people pretty hard to persuade, actually.
也许我不够好，但作为一名教授，我发现实际上很难说服别人。
I worry about this less than many of my AI-related friends do.
我比许多与人工智能相关的朋友更担心这一点。
I hope you’re right.
我希望你是对的。
Yes.
是的。
Last question, on matters where you can speak publicly.
最后一个问题，关于您可以公开发言的问题。
At the margin, if you could call in an expert to help you resolve a question in your mind of substance, what would that question be?
顺便说一句，如果你可以请一位专家来帮助你解决你心中的一个实质性问题，那么这个问题会是什么？
I have an answer to this ready to go, but only because I got asked to — well, maybe I’ll tell the story right after.
我已经准备好了这个问题的答案，但只是因为我被要求“——”好吧，也许我会马上讲这个故事。
This is like I take this spiritually, not literally.
这就像我从精神上而不是字面上理解这一点。
There will come a moment where the superintelligence is built.
超级智能终将建成的时刻到来。
It is safety-tested.
它经过安全测试。
It is ready to go.
一切准备就绪，可以出发了。
We’ll still be able to supervise it, but it’s going to do just vastly incredible things.
我们仍然能够监督它，但它会做非常令人难以置信的事情。
It’s going to be self-improving.
它将自我完善。
It’s going to launch the probes of the stars, whatever.
无论如何，它将发射恒星探测器。
You get the opportunity to type in the prompt before you say okay.
在说“好”之前，您有机会输入提示。
The question is, what should you type in?
问题是，你应该输入什么？
Do you have a tentative answer now?
现在你有初步答案了吗？
No, I don’t.
不，我不。
The reason that I had that ready to go in mind is, someone was going to see the Dalai Lama and said, “I’ll ask any question about AI you want.” I was like, “What a great opportunity.” I thought really hard about it and that was my question.
我之所以做好这样的准备，是因为有人要见达赖喇嘛并说：“我会问你想要的任何有关人工智能的问题。”我当时想，“多么好的机会啊。”我认真思考了一下，这就是我的问题。
Sam Altman, thank you very much.
萨姆·奥尔特曼，非常感谢你。
Thank you.
谢谢。
Photo Credit: Jeremi Rebecca
照片来源：杰里米丽贝卡