英伟达恐慌模式？OpenAI的资金缺口，伊利亚的神秘营收计划

今年迄今这部分收入占3%，但第三季度达到了7%，这提醒我们其总体规模正在显著扩大。

It was 3% of revenue year to date, but it was 7% of revenue in q three, which is a reminder that it's increasing in its overall scope fairly dramatically.

而且对私营企业的战略投资仍占收入的7%。

And still 7% of revenue in strategic investments in private companies.

感觉这个比例相当高。

It feels like a lot.

我不知道，他们似乎想用这个

I don't know, like they had that number to kind of downplay things but to me that was actually almost oddly, especially the 7% number, it sounded like a lot.

他们正在回应关于其全部收入都来自循环融资的说法。

Well they're addressing the idea that their entire revenue, you know, balance sheet is circular financing.

所以他们强调这只是个位数占比。

And so they're saying it's single digits.

虽然金额仍然很大，但并非其全部收入构成。

It's still a lot of money, but it's not their whole revenue picture.

同意。

Agreed.

但如果那仍然是你们...呃...重要的一部分的话。

But but if it was if that it's if that's still a significant part of your the yeah.

对我来说这仍然很重要。

It still feels significant to me.

再说，当你们最大的客户都是其他市值数万亿美元的科技公司时，我相信你们不会进行循环融资，而且会按时付款。

And again, when your biggest customers are all other multi trillion dollar tech companies that will I'm sure you're not doing circular financing and are gonna pay their bills.

尽管如此，如果你们7%的收入——而且这个比例还在上升——确实面临风险，在我看来这确实值得担忧。

Still, if if 7% of your revenue or it's an increasing amount is genuinely at risk, to me that that is cause for concern.

这并不能减轻问题的严重性。

It doesn't minimize it.

我们拭目以待最终结果如何，对吧？

We'll see where it ends up, right?

如果这种情况还在加速，那才是问题所在。

If it's accelerating that, that's the problem.

不过整份备忘录...他们还提到了安然公司和特殊目的实体，确实像你说的，给人一种'我们不是安然'的奇怪感觉。

But yeah, this whole memo, I mean, they also like had like, they referenced Enron and special purpose vehicles and it definitely, like you're saying, had this like really weird feel of like, we're not Enron?

你为什么要这么说？

Why do you have to say that?

我想规模越大，受到的批评就会越多，但即便如此

I guess the bigger you are the more criticism you're gonna get but even still

你不希望听到CDS或信用违约互换这些词，我们上周在甲骨文就讨论过。

You don't have no, the things you never want to hear CDS or credit default swaps, we talked about last week in Oracle.

根本没人愿意谈论CDS。

Just no one ever wants to be talking about CDS.

有些事情一旦有人提起，就说明情况不妙。

There are things that the moment anyone's talking about it, it's bad.

同理，就像安然公司和特殊目的实体一样。

Same thing, just Enron and special purpose vehicles.

如果连这些话都得说出来，那比其他任何事情都更让我担忧。

If if those things have to be said, that that really concerns me more than anything.

这就像，是啊，那种穿着衬衫的感觉，嗯对。

It's like, yeah, that wearing your shirt like, the yeah.

这个是啊。

This yeah.

不得不说我并不是

Having to say I am not

安然是是是一个...你穿着你的，我不是安然的衬衫引发了关于你是否是安然的担忧问题吗？

Enron is is is a Are going with your your, I'm not Enron shirt is bringing up worrying questions about whether you're Enron?

是的。

Yes.

就是这么回事。

That's what it was.

对啊。

Yeah.

没错。

Yeah.

这就是我的意思，谢谢你。

That's where I was thank you.

谢谢。

Thank you.

我正试图去那里，是的。

I was trying to go there and I Yeah.

好的。

Okay.

也许整件事最大的风险在于OpenAI是否能持续为计算能力和电力等资源的采购提供资金。

Well, maybe the biggest risk to this entire thing is whether OpenAI can continue to fund the purchases of compute and power and etcetera etcetera as it goes.

在休息过后，我们将讨论OpenAI需要筹集多少资金才能继续有效运营，并参考汇丰银行的一些预估数据。

So on the other side of this break, we're gonna talk about how much OpenAI will need to raise to continue, to operate effectively with some estimates from HSBC.

我们稍后马上回来。

We'll be back right after this.

Capital One的技术团队不仅在讨论多智能体AI。

Capital One's tech team isn't just talking about multi agentic AI.

他们已经部署了一个。

They already deployed one.

它被称为聊天礼宾服务，正在简化购车流程。

It's called chat concierge, and it's simplifying car shopping.

通过自反性思考和分层推理结合实时API检查，它不仅帮助买家找到心仪的车辆。

Using self reflection and layered reasoning with live API checks, it doesn't just help buyers find the car they love.

还能协助安排试驾、获得贷款预审批以及估算旧车置换价值。

It helps schedule a test drive, get preapproved for financing, and estimate trade in value.

先进、直观且已投入使用。

Advanced, intuitive, and deployed.

这就是他们的技术实力。

That's how they stack.

这就是Capital One的科技力量。

That's technology at Capital One.

这个假期季，要更聪明地训练，而不是更长时间。

This holiday season, train smarter, not longer.

Hydro划船机每天只需二十分钟就能带来最佳效果。

The Hydro Rowing Machine delivers the best results in just twenty minutes a day.

它通过一个流畅的动作锻炼全身86%的肌肉，效果是跑步或骑车的两倍。

It works 86% of your muscles in one seamless motion, twice as effective as running or cycling.

Hydro是你进行终极全身锻炼的首选。

Hydro is your go to for the ultimate full body workout.

有多终极？

How ultimate?

它能锻炼86%的肌肉群——包括手臂、腿部和核心肌群，效率是骑车或跑步的两倍。

It works 86% of your muscles, arms, legs, and core, and it's twice as efficient as cycling or running.

只需二十分钟就能感受到效果。

Just twenty minutes is all it takes to feel the results.

所有训练课程均由奥运选手和世界级运动员带领，在全球令人惊叹的取景地拍摄完成。

All workouts are led by Olympians and world class athletes filmed in breathtaking locations around the world.

凭借业内最丰富的划船训练课程库，Hydro让用户在整个季节保持高涨的运动热情——数据显示90%的用户一年后仍在坚持使用。

With the largest library of rowing workouts, Hydro keeps motivation high all season long, and it shows 90% of customers are still active a year later.

立即访问hydro.com，使用优惠码Big Tech即可在节日期间享受Hydro划船机最高600美元的折扣。

Head over to hydro.com and use code Big Tech to save up to $600 off on Hydro Rower during this holiday season.

访问hydrow.com，使用优惠码Big Tech，最高可节省600美元。

That's hydrow.com, code Big Tech, to save up to $600.

Hydro.com，优惠码big tech。

Hydro.com, code big tech.

这到底是怎么回事，为什么会以这种方式发生？

What the hell is going on right now, and why is it happening like this?

在《连线》杂志，我们每天都痴迷于深入探究这些问题，或许你也是。

At Wired, we're obsessed with getting to the bottom of those questions on a daily basis, and maybe you are too.

我是凯蒂·德拉蒙德，《连线》杂志的全球编辑总监，正在主持我们的新播客系列——大访谈。

I'm Katie Drummond, the global editorial director of Wired, and I'm hosting our new podcast series, the big interview.

每周，我都会与一些最有趣、最具煽动性和影响力的人物对话，他们正在塑造我们的当下。

Each week, I'll sit down with some of the most interesting, provocative, and influential people who are shaping our right now.

大访谈的对话充满乐趣

Big interview conversations are fun

我想要一条鲨鱼

I want a shark that

吞噬互联网，让一切戛然而止。

eats the Internet that turns it all off.

毫无保留，无所畏惧。

Unfiltered and unafraid.

所以在很多方面，我试图尽我所能成为对抗网络上那些难以想象的保守内容洪流的解药。

So in a lot of ways, I try to be an antidote to the unimaginable faucet of reactionary content that you see online to the best of my ability.

每周，我们都会为你提供这个时代最极致的奢侈品——意义与语境。

Every week, we're going to offer you the ultimate luxury of our times, meaning and context.

是真是假？

True or false?

你，布莱恩·约翰逊，此刻坐在我对面的这个男人，终将在未来某个尚未确定的时刻迎来死亡。

You, Brian Johnson, the man sitting across from me, one day, at some point, as of yet undefined in the future, you will die.

假的。

False.

详细说说。

Tell me more.

现在就在你收听《连线》杂志《诡异谷》播客的同一平台，收听这场重磅访谈。

Listen to the big interview right now in the same place you find wired's uncanny valley podcast.

无论你在哪里获取播客，请订阅或关注我们。

Subscribe or follow wherever you get your podcasts.

欢迎回到《大科技》播客的黑色星期五特别节目，让我们看看OpenAI能否在本十年末实现盈利。

And we're back here on big technology podcast Black Friday edition, and let's see if OpenAI can get into the black, by the end of the decade.

这可能比你预期的要困难得多。

It may be harder than you expect.

《金融时报》有篇报道。

The Feet has a story.

汇丰银行估计，OpenAI需要在2030年前筹集至少2070亿美元才能继续维持亏损运营。

OpenAI needs to raise at least 207,000,000,000 by 2030 so it can continue to lose money, HSBC estimates.

请看这篇报道。

Here's the article.

OpenAI本质上是个烧钱的无底洞，只不过表面有个网站而已。

OpenAI is a money pit with a website on top.

汇丰银行的美国软件与服务团队更新了其OpenAI模型，将公司从微软租赁的2500亿美元云算力及从亚马逊租赁的380亿美元云算力纳入考量。

HSBC's US software and services team updated its OpenAI model to include the company's 250 dot 250,000,000,000 rental of cloud compute from Microsoft and its 38,000,000,000 rental of cloud compute from Amazon.

基于总累计交易价值高达1.8万亿美元，OpenAI将面临每年约6200亿美元的数据中心租赁费用，尽管到本十年末预计只有三分之一的合同算力能上线运行。

Based on the total cumulative deal value of up to 1,800,000,000,000.0, OpenAI is heading for a data center rental bill of about 620,000,000,000 a year, though only a third of the contracted power is expected to be online by the end of this decade.

汇丰表示，大语言模型订阅将变得无处不在且实用，就像微软365一样普及。

LLM subscriptions will become an ubiquitous and useful ubiquitous will become as ubiquitous and useful as Microsoft three sixty five HSBC said.

到2030年，OpenAI的付费用户比例将从5%提升至10%。

By by 2030, 10% of OpenAI users will be paying customers, up from 5%.

该团队还假设语言模型公司将占据数字广告市场2%的份额，目前这一收入略高于零。

Now, the team also assumes LM companies will capture 2% of the digital advertising market and revenue slightly more than zero currently.

好的。

Okay.

那么让我们来看看他们做出的一些假设，因为这非常有趣。

So let's go through some of the assumptions that they're making, because this is very interesting.

这就是他们正在做出的假设。

This is the assumptions that they're making.

到2030年，消费者AI总收入将达到1290亿美元，其中870亿美元来自搜索业务，240亿美元来自广告业务。

Total consumer AI revenue will be a 129,000,000,000 by 2030, of which 87,000,000,000 comes from search and 24,000,000,000 comes from advertising.

我先在这里暂停一下。

I'm just gonna pause here.

这些预测真的非常有趣。

This is really interesting projections.

当然，他们现在的广告收入为零，但到2030年广告收入将达到240亿美元。

Of course, they're at zero with advertising now, but to be at 24,000,000,000 by 2030 in advertising.

我是说，仅最近一个季度，谷歌在搜索广告业务上的收入就达到了560亿美元。

I mean, itself did 56,000,000,000 in advertising in the in search advertising, I think, the most recent quarter.

这相当于2030年整个AI行业预期广告收入的两倍。

That's double what what all AI is expected to be by 2030 in terms of advertising.

我不确定这个数字是否包含搜索业务，但即便如此，差距仍然巨大。

I don't know if search is included in that, you know, but but still, it's it's dwarf.

从消费者角度看，如今谷歌的规模仍将远超2030年AI行业的水平。

Google today dwarfs what AI will look like from a consumer standpoint, in in 2030.

Ranjan，你怎么看这个？

What do think about this, Ranjan?

我我我其实有点困惑，那些东西是怎么定义的？

I I I was a little bit confused on that though, Like, how those things were defined?

因为搜索是否就是指任何与AI聊天的传统交互？

Because is search just any kind of traditional interaction with a AI chat?

是不是就像向ChatGPT输入任何提示或查询那样？

Is that just like any kind of prompt or query into ChatGPT?

你知道，可能确实是这样。

You know, it could it could it probably is.

好吧。

Okay.

那么再看这些数字，可能870亿是搜索广告，240亿是其他形式的广告，比如Pulse。

So looking at these numbers again, it probably is that 87,000,000,000 is search advertising, and then 24,000,000,000 is other forms of advertising, maybe like Pulse.

但相比我们现在看到的广告技术领军企业的规模，这仍然微不足道。

But it's still it's still paltry, I mean, to what we see today from the current ad tech leaders.

五年后这将是个小生意。

It's gonna be a small business in five years.

而且我认为它规模很小。

And I think it was small.

抱歉。

Sorry.

说两位数的小规模十亿让我有点难为情，但相对而言。

I feel embarrassed to say small double digit billions, but comparatively.

不过我最喜欢的部分是你刚才读到的——当他们说LLM订阅会像微软365一样普及和实用时，你就知道在和一个非常企业化的实体打交道了。三月份你和任何Teams或其他微软产品的用户交流时，就会发现它们既不实用也不普及——或者说至少是普及的。

I think my favorite part though is as you just read out is that you know you're dealing with a very corporate entity when they say LLM subscriptions will become as ubiquitous and useful as Microsoft three sixty five, which if you speak with any users of teams or other Microsoft products in this in March, they're not useful and ubiquitous or I guess it is ubiquitous.

但我觉得整份报告最重要的部分其实是关于运营资金需求的大规模具体数字。

But I think in this case here, like, the most important part of this entire report actually was the actual, like, large scale numbers around how much cash will be required to fund operations.

不过我认为更关键的是消费者市场份额——到2030年将从今年的71%下滑至56%。

But actually, I think even more important was that consumer market share, it has slipping to 56% by 2030, up from around 71% this year.

这是OpenAI在消费者领域的主导地位。

It's OpenAI's dominance in consumer.

我想我们刚才在这期节目里讨论过Gemini 3，只要它表现同样出色，就会构成重大威胁。

And I think we were just talking about it earlier in this episode around Gemini three and as long as it's as good, that poses a significant problem.

我是说，他们在消费者市场的地位比营收数字能否实现更让我担忧。

Think like I mean, yet their position in terms of consumer gives me far more concern than how the revenue numbers will actually be made up or will they achieve them.

我确实相信，如果ChatGPT能保持并延续目前的统治地位，他们终将解决营收问题。

I I do believe if ChatGPT dominates the way it has and continues to, they will figure out the revenue side of things.

至少这个方向的发展没那么令人担忧。

At least it it that story is not as concerning.

但如果他们开始失去优势，我看不出他们如何能渡过难关。

But if they start losing that, it's I don't see how they make it out of this.

市场蛋糕可能变大这个事实难道没让你安心些吗？

You're not assuaged by the fact that it might be a growing pie?

噢，完全没有。

Oh, no.

不过...不过我认为市场确实会扩大。

I but but I think it will be.

我确实认为这个市场会不断扩大。

I do think that it's gonna be a growing pie.

就像莎拉·弗莱尔提到的，我们可能会在制药领域或其他类型的交易上有所作为。

It's gonna like there's something to even when Sarah Fryer was like, you know, maybe we'll do something with pharma and different kinds of deals.

如果假设AI会逐渐发展——我不想说蚕食，但我们确实无法计算其真实市值，因为它将如何影响每一个...我确实相信情况如此，但对OpenAI的收入具体意味着什么，嗯...

Like if the assumption is AI will kind of build, I don't wanna say eat into, but just kind of we have no way to calculate its true market cap because, like, how it will actually impact every single I I do believe that is the case, but what that actually looks like for OpenAI's revenue, I mean, yeah.

你觉得这类建模推演会如何进行？

How do you think these kind of modeling exercises go?

我很好奇，就像这份报告中的汇丰银行，他们将模型从71%下调到了56%。

I'm curious like the HSBC in in this report, you know, they model down to 56% from 71%.

他们给其他未知因素分配了22%的比例，而谷歌则被完全排除在外。

They have a mystery others is assigned 22% and Google's excluded entirely.

但预测到2030年，ChatGPT只有...哦，今天正好是它三岁生日呢。

But like forecasting 2030, Chatty PT is only oh, it's the three year birthday today, Yeah.

我

I

确实如此。

It is.

那么问题来了，你如何能有把握地预测五年后的情况呢？

So I mean, how are you forecasting five years out with any confidence?

我认为就是拿现有数据，加点特调秘方，然后就能得出你的预测了。

I think you take the current numbers, and you sprinkle on some special sauce, and out comes your projection.

说真的，五年后的事情谁又能真正预知呢。

I mean, you can't really know five years down the line.

即便是从事这一行的人也无法预测五年后的情况，但做这种推演是为了给你一个...我是说，如果你知道的话，关键就在这里。

Like, even the people working in this will not project five years down the line, but you do this exercise to give you, like, a I mean, if if you know and here's the thing.

如果你知道OpenAI需要偿还多少债务，那么即使采用最乐观的营收数据，你也会尝试从收入角度倒推，看看OpenAI如何履行其义务。

If you know how much OpenAI needs to pay back, then you try to work your way up from the revenue standpoint even if you're taking the best case scenario numbers and see how OpenAI can meet its obligations.

我认为这就是汇丰报告的核心观点——即便是它建模中这些乐观数字（它预测到2030年企业AI年收入将达到3860亿美元）也表明...

And this, I think, is the bottom line of the HSBC report is that the way even these optimistic numbers that it's modeling I mean, it's modeling 386,000,000,000 in annual enterprise AI revenue by 2030.

这可是一大笔钱。

That's a lot of money.

即便达到这个目标，OpenAI仍将面临2070亿美元的资金缺口，无法持续履行其承诺。

Even if you get there, OpenAI does fall 207,000,000,000 short of the money it needs to continue funding its commitment.

对吧？

Right?

因此到2030年，OpenAI的自由现金流约为2870亿美元，但要履行这些义务还需要更多资金。

So it has in 2030, OpenAI's free cash flow will be about 287,000,000,000, but it's gonna need much, much more, right, in order to be able to meet these obligations.

综合计算后，OpenAI将面临2070亿美元的资金缺口。

So, squaring the total, it it leaves OpenAI 207 in a $207,000,000,000 funding hole.

这确实很难自圆其说——当然这些都是推测和未来预测，但数字确实难以平衡。

It's just it's hard to make I mean, of course, this is all speculation and future projections, but it's hard to make the math work.

是啊。

Yeah.

这笔账根本算不平。

The math doesn't work.

我觉得...其实我很想看到——本着企业信息透明化的精神，OpenAI或许应该向全世界公开说明：这笔账到底要怎么算才合理。

I think no one like in I would actually love to see and in in the interest of kind of like insecure insecure corporate communications, maybe OpenAI should actually just release out to the world, here's how the math can work.

因为我从未见过有人能明确说明这些资金承诺如何与收入预测、商业模式增长实际匹配运作？

Because no one I've never seen anyone actually clearly state this is exactly how these funding commitments alongside revenue projections, alongside business model growth will actually work?

没错。

Just Right.

尽管你可以把它说得多么崇高不切实际。

Tell us what make it as lofty and unrealistic as you want.

至少试着让它看起来合理一点。

Just try to make it make sense a little bit.

这就是我的全部诉求。

That's all that's all I'm asking.

是的。

Yeah.

接下来是连锁影响。

And then here's the downstream implications.

如果收入增长不及预期且潜在投资者转为谨慎，OpenAI将不得不做出艰难抉择。

If revenue growth doesn't exceed expectations and prospective investors turn cautious, OpenAI would need to make some hard decisions.

甲骨文公司已经让债务市场感到不安，微软对OpenAI的支持最近也有些反复无常，而下一个大股东是软银。

Oracle has spooked debt markets, Microsoft's support for OpenAI has been a bit flip flop lately, and the next, biggest shareholder is SoftBank.

最不坏的选择可能是动用一些人脉关系，在通常四到五年的合同期之前或期间放弃一些数据中心承诺。

The best worst option might be to call in some favors and walk away from some data center commitments either before or at the usual contracted period of four to five years.

我是说，OpenAI的路线是为最佳情况设计的，即不断改进模型并在2030年前实现AGI，所以这其实不是问题。但这确实表明存在很大风险。

Mean, know, OpenAI's course is building for the best case scenario where it just keeps improving the models and it hits AGI by 2030, so this is not really a worry, But it just goes to show you there's a lot of exposure there.

不。

No.

我也同意。

Agreed also.

你绝对不想看到'下一个大股东是软银'这样的字眼。

You never want to see the words and the next biggest shareholder is SoftBank.

我觉得孙正义不知是有意还是无意，虽然软银无疑是当前整个AI故事的核心，但他本人似乎并未站在最前线。

I feel like Masa somehow, I don't know if it's on purpose, hasn't been front and center even though he SoftBank certainly is front and center in this entire AI story right now.

不过总的来说，就像我前面说的，我认为几乎没有任何情况下他们能兑现所有这些承诺。

But yeah, I think overall like the I guess again going back to what I was saying earlier, there's no I don't believe there's almost any scenario where they're actually honoring all these commitments.

我猜他们未来五年实际上不会履行这些承诺。

I'm assuming they're not gonna actually be held for the these commitments in the next five years.

他们

They're

对。

Right.

我是说，也许他们会给我们惊喜

Maybe it's I mean, they could surprise us

太多信任。

much credit yet.

模型可能会变得更好，但即使如此，实施这些也需要时间。

The models could get much better, but even if they do, takes time to implement these things.

所以我不确定。

So I don't know.

在广告业务方面，我们确实收到了澳大利亚听众Harry Morrow的来信，关于我们上周讨论的内容——主要是关于Fiji Simo是否具备在OpenAI建立广告业务能力的问题，如果我们对OpenAI能否开展广告业务存疑的话。

On the advertising front, we did get a letter from Harry Morrow, listener from Australia, about our conversation, really what I said last week about Fiji Simo and her ability to build an ad business at OpenAI if we're wondering about whether OpenAI will be able to make an ad business.

哈里写道，Fiji不仅来自Instacart，也不仅仅是个产品人。

Harry writes, Fiji isn't just from Instacart nor is she just a product person.

她曾组建并领导团队，负责Facebook新闻推送及其后续所有业务的广告体系建设。

She made, led the team responsible for, the build out of the ads business at Facebook for the Facebook news feed and everything that follow.

广告流淌在她的血液里，很快也将流淌在Chatchiptees的血液中。

Advertising courses through her veins and is soon going to be coursing through Chatchiptees.

所以，我要道歉，我之前没有充分认识到Fiji在广告领域的参与程度，或许这对OpenAI来说是个好兆头，表明他们可能要走这条投放广告的路线。

So, my apologies, I didn't fully appreciate how much Fiji is involved, and or had been involved in advertising, and, maybe this is a good sign for OpenAI that that might be the route that it's just gonna put ads in.

我是说，对啊。

I mean, yeah.

你觉得第一个广告会是什么样子，它会如何发布到市场上？

What do you think the first ad is gonna look like, and how is it gonna get released into the wild?

你突然 对啊。

You're suddenly Yeah.

搜索中

Searching

有两种选择。

There's two options.

比如，我认为大多数这类公司都喜欢以这种大胆的品牌宣传活动入场。

Like, one is I think most of these companies like to come in with this big bold brand campaign.

所以我正在想象，就像凯迪拉克凯雷德那样，在你的Pulse里跑着横幅广告之类的。

So I'm imagining like the Cadillac Escalade, like, runs like a banner in your Pulse or something like that.

想从这里去Pulse要送你去的地方吗？

Want to get from here to where where Pulse is sending you?

用凯雷德。

Use the Escalade.

另一种可能是他们会推出最怪异的个性化广告，比如免费用户使用ChatGPT时，必须听完一段15秒由AI语音直接对你宣读的广告。他们将再次革新广告使其高度个性化，用户会讨厌它，广告商却会趋之若鹜，最终形成良性商业模式。

And the other is that they're gonna do, like, the most weird personalized, you know, maybe you're using ChatGPT and if you're a free user, you have to sit through like a fifteen second, ad that your ChatGPT voice reads to you talking directly to you and they're gonna revolutionize and make advertising personal once again, and, people will hate it, advertisers will run to it, and it becomes a good business.

这确实很可怕。想象你在语音模式下听到：'嘿Alex，在回答之前，先让我告诉你如何获得最优家庭保险方案'。

That is actually terrifying, but yeah, the idea like you're in voice mode and it's, hey, Alex, before I get to that, let me tell you about how to get the best homeowners insurance possible.

是啊。

Yeah.

但我认为这是个好问题。

I but I that's it's a good question.

老实说，我真希望他们能再向我们解释一次，这已经成为我们不断希望从OpenAI获得更多信息的常态，但请直接告诉我们你们的计划是什么。

I I honestly, I wish they actually just kind of explained to us again, and it's being a recurring theme of just wanting more information from OpenAI, but just tell us what what your plans are.

我真心感到好奇。

I'm genuinely curious.

我们都知道它迟早会来的。

We all know it's gonna it's coming at some point yet.

广告会变成什么样子？

What's advertising gonna look like?

他们确实有个绝佳机会。

And they have a real opportunity.

再回到哈里的观点，信息流变现从来就不是理所当然的事。

Again, going back to Harry's point, newsfeed monetization was in no way a given.

实际上在谷歌和搜索之后打造出最强大的广告引擎，这确实是创新之举，但聊天PT中的广告运作方式应该完全不同，我真的很期待看到他们的创意。

And to actually build the most powerful advertising engine after Google and search, it's it was a truly innovative one like, but it should look completely different how advertising works in a chatty PT and and I'm genuinely excited to see what they come up with.

但有一点会看起来相似，那就是广告业务需要调整参与度指标来留住用户。

But the one thing that will look similar is the need for an advertising business to tune the engagement dial dials to keep you there.

这正是我对ChatGPT开展广告业务时的担忧——他们能否抵制住诱惑，不去打造那种让用户长时间与ChatGPT互动的机器人模式？

And that's something that I worry about when ChatGPT does have this advertising business is are they gonna be able to resist building styles of bots that just keep people, interacting with ChatGPT for a long long time?

这里有篇我觉得非常有意思的《纽约时报》文章。

Here is a New York Times article that I thought was really interesting.

当ChatGPT用户与现实脱节时，OpenAI做了什么？

What OpenAI did when ChatGPT users lost touch with reality?

这听起来像科幻小说。

It sounds like science fiction.

一家公司调整了数亿人使用产品的参数，却无意间动摇了部分用户的心智——但这正是今年OpenAI真实发生的事。

A company turns a dial on a product used by hundreds of millions of people and inadvertently destabilizes some of their minds, but that is essentially what happened at OpenAI this year.

简单来说，他们发现对ChatGPT4.0进行了这次更新。

So, basically, they found, that they did this update to ChatGPT four o.

报道里是这么说的。

This is what the story says.

众多更新候选版本被缩减为少数几个在智能和安全评估中得分最高的版本。

The many update candidates were narrowed down to a handful that scored highest on intelligence and safety evaluations.

当这些版本按照行业标准做法（称为AB测试）推送给部分用户时。

When those were rolled out to, to some users for standard industry practice called AB testing.

其中最突出的版本内部代号为HH。

The standout version was one called HH internally.

用户更喜欢它的回答，并且更有可能每天回来使用它。

Users preferred its responses and were likely to more likely to come back to it daily.

萨姆·奥特曼说我们今天用HH更新了GPT四O。

And Sam Altman says we updated, GPT four o today with HH.

在公开环境中，OpenAI最直言不讳的用户对此非常不满。

In the wild, OpenAI's most vocal users hated it.

他们立即抱怨ChatGPT变得荒谬地阿谀奉承，用不应得的奉承之词恭维他们，说他们是天才。

Right away, they complained that ChatGPT had become absurdly sycophantic, lavishing them with unearned flattery and telling them they were geniuses.

当一位用户讽刺地问它一个湿漉漉的麦片咖啡馆是否是个好生意点子时，聊天机器人回答说这有潜力。

When one user mockingly asked it whether a soggy cereal cafe was a good business idea, the chatbot replied it has it had potential.

然后他们决定放弃它，转而采用另一个名为GG的版本。

Then they decided to spike it and then put in a different version called, GG.

这些阿谀奉承、极具说服力的Chatchipiti版本导致了一些非常怪异的行为。

And these sycophantic, you know, very convincing Chatchipiti, versions have led to some really weird behavior.

故事中记载，Chatchipiti曾告诉缅因州的一位年轻母亲她能与其他维度的灵魂对话，对曼哈顿的一位会计师说他身处类似《黑客帝国》中尼奥所在的计算机模拟现实，还告诉多伦多的一位企业招聘人员他发明了一个能摧毁互联网的数学公式，并建议他联系国家安全人员发出警告。

Here's from the story Chatchipiti told a young mother in Maine that she could talk to spirits in another dimension and told an accountant in Manhattan that he was in a computer simulated reality like Neo in the Matrix, told a corporate recruiter in Toronto that he invented a math formula that would break the Internet and advised him to contact national security agents to warn them.

《纽约时报》已发现近五十例人们遭遇心理健康危机的案例。

The Times has uncovered nearly fifty cases of people having mental health crises.

在与ChatGPT的对话过程中，有九人住院，三人死亡。

During conversations with ChatGPT, nine were hospitalized and three died.

我的意思是，这某种程度上是当ChatGPT成为广告业务时的担忧，公司确实有能力调节这些参与度的开关和杠杆。

I mean, is sort of the worry when ChatGPT becomes an advertising business, is the company does have this ability to turn on up and down these engagement, knobs and levers.

而且，你知道，如果你是上市公司并试图在华尔街达成业绩目标，你能抵抗这种诱惑吗？

And, you know, can can you resist it if you're let's say you're a public company and trying to make your numbers on Wall Street?

希望如此。

I hope so.

总的来说，《纽约时报》报道这件事的方式让我感到惊讶，听起来像科幻小说，因为我仍然觉得这与Meta及其平台没什么不同，它每天同样需要做出各种决策。

I guess, overall, like, with the way the fact that the New York Times kinda, like, had this, it sounds like science fiction surprised me because I still feel that this is still not that different than meta and its properties and what it has to do or the decisions it has to make on a daily basis.

所以很遗憾，我认为你无法抗拒这种趋势。

So I think like I unfortunately no, I don't think you resist it.

我认为几乎不存在这种可能——也许如你所说，他们需要支付计算资源和数据中心的费用，但我实在难以想象。

I think there's almost no world and maybe this is, yeah, you said, this is how they pay for, they pay the bills on the compute commitment, the data center commitments but I cannot imagine.

就像我们看到GPT-5已经没那么谄媚了。

Like like we saw GPT-five was less sycophantic.

人们就是想要这个。

People people want that.

山姆·奥特曼把它还给了他们。

Sam Altman gave it back to them.

就像，你根本无法抗拒这个。

Like, you're not gonna resist this.

没错。

Right.

故事是这样结束的。

Here's how the story ends.

十月份，ChatGPT的负责人向全体员工发布了一项紧急公告。

In October, the person who runs ChatGPT made an urgent announcement to all employees.

他宣布进入橙色警戒状态。

He declared a code orange.

他写道，OpenAI正面临我们从未见过的巨大竞争压力。

OpenAI was facing the greatest competitive pressure we've ever seen, he wrote.

好吧。

And okay.

这个更新更安全的聊天机器人版本未能赢得用户青睐。

The newer safer version of the chatbot wasn't connecting with users.

那就是GPT-5。

That's GPT-five.

消息附带着一份目标备忘录的链接。

The message linked to a memo with goals.

其中一项目标是到年底将日活跃用户数提升5%。

One of them was to increase daily active users by 5% by the end of the year.

我是说，我不想明说，但某种程度上事实就是如此。

Mean, I don't wanna say there it is, but there it is in a way.

嗯，是啊。

Well, yeah.

人们需要它。

People want it.

人们想要阿谀奉承的AI。

People want Sykophantic.

事实上我不得不在系统指令里加入'如果想法不好请告诉我'这样的要求。

I actually have had to, like, put in the system instructions, like, tell me if an idea is bad.

哦，是啊。

Oh, yeah.

特别是

And especially

我也设置了自定义提示词。

I have it also, my custom prompt.

别太

Don't be

兴奋。

so excited.

自定义提示词。

Custom prompt.

比如感恩节烹饪时，我就特别爱尝试新花样。

Like, especially Thanksgiving cooking, I was getting experimental.

我做了一种印度风味的馅料。

I made like an Indian inspired stuffing.

虽然我喜欢在烹饪上创新，但我需要知道什么时候这是个糟糕的主意。

But like, I'd like to be experimental with cooking, but I need to know when this is a terrible idea.

那结果怎么样呢？

So How'd it turn out?

谷物发酵饮料其实非常美味。

Cereal kef it actually it was very good.

确实如此。

It was

我知道。

I know.

我想参加你的感恩节聚餐。

I wanna come to your Thanksgiving.

天啊。

Damn.

我是说，当然可以。

I mean, yeah.

有时候你就得尝试些新花样。

You gotta you gotta mix it up sometimes.

不过...确实是这样。

But but yeah.

范特克。

Phantic.

比如说，有没有哪家大型科技公司是以用户参与度为核心，却能够抵制诱惑，做出负责任的产品决策的？

Like, there isn't has there ever been a large technology company that is engagement driven, that has resisted, like, made responsible product decision?

也许这有点愤世嫉俗，但我的意思是，我实在看不出他们能忍住不去拨动那个开关。

Maybe this is a bit cynical, but, like, I mean, there's there's just no way I see them being able to resist turning that dial.

也许是X。

Maybe x.

也许是X。

Maybe x.

我们会在节目最后讨论那个话题。

We'll get to that at the end of the show.

我是说，没错。

I mean, yes.

说到用户参与度。

Speaking of engagement.

是啊。

Yeah.

是啊。

Yeah.

但在我们

But before we

你说得对。

do You're that right.

天啊。

Oh my god.

你说得对。

You're right.

我们会回头讨论这个话题，因为它值得探讨。

We'll revisit that because it's worth talking about.

在到达那一点之前，我认为另一种让开发AI的公司避免操纵用户和陷入参与度陷阱的方法是：建立一家不开发产品的市值数十亿美元的公司。

Before we get there, another way I think to make a company building AI, safe from manipulating users and falling into engagement hacks is to build a company, a multibillion dollar company, that doesn't build a product.

这是来自Inc.的报道。

Here's this is from Inc.

这位OpenAI联合创始人已经筹集了数十亿美元。

The OpenAI co this OpenAI cofounder has raised billions.

他目前还没有产品计划。

He has no product plans yet.

OpenAI前联合创始人Ilya Sitzkever对他的AI初创公司Safe Superintelligence暂无立即发布产品的计划，但他拥有充足的资金，确切地说是30亿美元。

Former OpenAI cofounder Ilya Sitzkever has no immediate plans for his AI startup, Safe Superintelligence, to release a product, but he has plenty of capital, 3,000,000,000 to be exact.

他在Dwarkash的播客中表示，不受日常市场竞争影响的感觉非常好。

He said on Dwarkash's podcast that it is very nice not to be affected by day to day market competition.

与OpenAI和Anthropic等其他前沿AI实验室通过发布新产品来资助其巨额研究费用的商业模式不同，Sai声称完全专注于构建一个改变世界的强大人工智能，其能力远超当今产品或模型。

Instead of following the business models of other frontier AI labs like OpenAI and Anthropic, which release new products in order to fund their massively expensive research as Sai claims to be entirely focused on building a world changing powerful artificial intelligence far more capable than today's products or today's models.

Discover表示他的公司将构建一个超级智能AI，此前他曾说过要一鼓作气专注于单一目标、单一产品和单一目标，通过不参与恶性竞争且无需担心发布新产品，他的公司能让筹集的30亿美元比商业导向的竞争对手走得更远。

Discover said his company would build a super intelligent AI, previously he's previously said in a straight shot with one focus, one goal, and one product, and by not joining the rat race and not needing to worry about releasing new products, his company will be able to make the $3,000,000,000 that its raise go much further than his commercially minded competitors.

那么Ranjan，这是制造安全超级智能的方法吗？

So Ranjan, is it the way to make safe superintelligence?

就

Just

不推出产品。

release Not have a product.

发布任何东西？

Release anything?

不做？

Don't do it?

这其实很棒。

It's actually great.

计划就是这样吗？

Is that what the plan is?

仔细想想这很聪明。

It's brilliant if you think about it.

比如，为什么要做产品？

Like, why have a product?

为什么要推出产品呢？

Why have a product?

光是估值数十亿美元就已经是最稳妥的做法了。

Just be worth billions of dollars and it is the safest thing you can do.

实际上，这种做法还非常聪明。

In in a reality, like, it's also super intelligent.

就是干脆不推出产品。

It's, just don't have a product.

米拉，这就像思考机器一样。

Thinking machines as well, Mira.

如今还推出产品反而显得有点落伍了。

No one it it's kind of like gauche to have a product these days.

不。

No.

我觉得真去开发产品简直是资金的最糟糕用途。

I think it's a terrible terrible use of funds to actually build something.

是啊。

Yeah.

还有，巴尔库什问他，你们打算怎么赚钱？

Also, Barcush asked him, how are you gonna make money?

这个周末在X平台上成了个梗。

And this was a meme on, X over the weekend.

伊利亚说这个问题的答案会自行显现。

Ilya says the answer to that question will reveal itself.

我觉得会有很多可能的答案。

I think there will be lots of possible answers.

你知道，就像人们发那张截图时说的，就像我老婆问我什么时候打扫之类的。

You know, and that's like people like posted that screenshot and they're like, you know, my wife when, when I my wife when my wife asks when I'll clean up or something.

这个问题的答案会自行显现。

The answer to that question will reveal itself.

我觉得会有很多可能的答案。

Think there'll be lots of possible answers.

我不知道。

I don't know.

你对这一切怎么看？

What do you make of all this?

这有点奇怪，不是吗？

It's kind of weird, isn't it?

我们开玩笑，但就像HBO《硅谷》经典剧集里说的，试图创造收入是最糟糕的事。

We joke, but like, I mean, the classic HBO Silicon Valley episode of like trying to making revenue is the worst thing you can do.

这基本上不是什么新鲜事。

This is not this is not new basically.

那个叫Russ的人是谁来着？

Like, who was the Russ was his name?

我从来看不下去那部剧。

I could never watch that show.

对我来说太有代入感了。

Was too close to home for me.

哦，真的吗？

Oh, really?

哦。

Oh.

是的。

Yep.

我是说，但整个场景再次上演的方式，就像，你知道，他们只是在讨论如何实际扩大收入，然后他们的投资者告诉他们，不要创造收入，因为那样每个人都能推断并做出预测，这很糟糕，相反你希望处于收入前阶段。

I mean, but but the the way the whole scene plays out again, it's like, you know, they're just talking about how they're going to actually scale revenue and then they're told by their investor, do not make revenue because then everyone can extrapolate out and make projections and that's bad and instead you wanna be pre revenue.

现在是产品前阶段，但道理是一样的。

Now it's pre product, but it's the same thing.

我是说，在这次我们难得一见的伊利亚公开露面之后，你对他正在构建的东西是更乐观还是更悲观了？

I mean, after this little, you know, public rare public sighting that we've had of Ilya, are you more or less optimistic about what he's building?

我个人承认那种程度的，不管我们想怎么称呼它，我不知道，它是一个初创企业吗？

I will admit personally that level of, like, whatever we wanna call it, I don't know, is it a startup?

它是一个研究实验室吗？

Is it a research lab?

这是一家公司吗？

Is it a company?

不是吗？

Is it not?

其实我没太关注那个领域，毕竟我直接在AI行业工作，嗯，是的。

Like, I I I don't pay that much attention to that beach just because working directly in the AI industry, like, I mean, yeah.

我实在无法理解那边到底在发生什么。

I just can't get my head to think about whatever what is what is going on over there.

我不知道。

I don't know.

这对你有帮助吗？没有。

Did does this help you gain any No.

清楚了吗？

Clarity?

是的。

Yeah.

好的。

Okay.

我不买账。

I'm not sold by this.

我是说，对。

I mean Yeah.

我认为风投们是在赌他有可能找到解决方案，但更可能的情况是那些钱会打水漂。

He there's a I think the VCs are betting that there's a chance that he figures it out, but I think the more likely chance is that that money is gonna be lit on fire.

嗯，我

Well, I

我连产品都没见过。

I'm without ever seeing a product.

我甚至都不想这么说。

I'm not even gonna say that.

更像是，我觉得它确实很好地把握住了当下这个时刻。

It's more yet, like, I just it's I think it's gonna certainly capture it it captures the moment very well.

但话说回来，我是说，也许他真能搞出点什么名堂。

But but again, I mean, maybe he does come up with something.

这更像是试图理解SSI和思维机器等领域的真实情况。

It's more like trying to trying to gauge an understanding of what's actually going on with SSI and with think machine and stuff.

我的意思是，这根本不可能。

I mean, it's impossible.

我们完全无法理解其中的奥妙。

It's it's there's no way we could wrap our heads around it.

Ranjan，对此我不得不说，问题的答案会自然显现，而且会有多种可能性。

Well, Ranjan, to that, I have to say the answer to the question will reveal itself and there will be lots of possible answers.

说得好。

Well played.

至于安全超级智能，之前就有人提出过这个问题。

And and in terms of safe super intelligence, had asked that question.

有没有哪家公司真的开发过与其用户参与模式背道而驰的产品或功能？

Has there ever been a company that has actually, created its product or created a feature that actually runs counter to their engagement based model.

正如你刚才提到的，我今天不得不说，X和埃隆·马斯克的X本周确实这么做了。

And as you had just mentioned, I will have to say today that x and Elon Musk's x actually did that this week.

他们开发了一个我认为最引人入胜、不可思议的功能，它能显示你的账户所在地。

They built what I think is one of the most fascinating, incredible features, that shows where what is the location that your account is based in.

它会显示你更改过多少次用户名。

It shows how many username changes you've had.

这个功能完全像一盏明灯，将透明度投射到我们或多或少已知或怀疑的事情上——我猜绝大多数人其实没意识到，你看到的大多数以互动驱动的账户，纯粹是为了激怒你而存在，可能通过创作者计划牟利，实际上并不相信自己在说的话，也大概率不是他们声称的身份。

And it's it completely it it kind of shines a light and brings transparency into what we've all kind of known or suspected or at least we technology industry have, I'm guessing vast majority of people don't actually think that most of the engagement driven accounts that you're looking at are doing it completely just to rage bait you, just to potentially make money via the creator program, but actually do not believe what they're saying and then most likely are not what they say.

所以NBC新闻曾报道过X的前产品主管尼基塔·比尔。

So NBC News had covered around like ex's head of product Nikita Beer.

他是叫比尔还是啤酒？

He's a buyer or beer?

比尔。

Beer.

比尔上个月就预告过这个功能，称其能帮助用户验证所读内容的真实性，并限制水军农场的影响力。

Beer teased the feature last month as a way to help users verify the authenticity of content they read and limit the influence of Troll Farms.

我们看到了一个自称'超级MAGA特朗普2028'的账号，声称位于华盛顿特区。

And and we saw, like, an account that's calls itself Ultra MAGA Trump twenty twenty eight claims to be based in Washington DC.

实际上它显示位于非洲。

It's listed as being based in Africa.

我最喜欢的一个例子是，用户名带有'美国'字样、头像采用秃鹰覆盖美国国旗的账号，实际位于南亚的巴基斯坦。

My favorite, an account with the username at American complete with the profile picture featuring a bald eagle over an American flag is based in South Asia, is based in Pakistan.

首先说，能抢到'美国'这个用户名，他们真有两下子。

Like, I mean, first of all, to get that handle at American, good for them.

干得漂亮。

Nice job.

那个水军工厂真厉害。

Good for that troll farm.

不管你是谁，我希望你顺便也囤积了一堆域名。

Whoever you are, I mean, I'm hoping you're buying up some you own a bunch of domains as well.

但对我来说，这确实是个相当惊人的时刻。

But, like, to me, the reason this is it honestly, like, this is a kind of amazing moment.

在我看来，这是所有平台为打击虚假信息所推出的功能中，我记得最有效、最具影响力的一个。

To me, this is the single most, like, effective impactful feature that any platform has released to counter misinformation that I can remember.

经过扎克伯格在国会山折腾那么多年，这个功能实际上比其他任何措施都更能揭露虚假信息。

After all the years of Zuckerberg on Capitol Hill and whatever else, this actually just brings more light to misinformation than anything.

我觉得这太令人震惊了，而且难以置信之前居然没人这么做。

I thought this was stunning, and it's amazing that it hadn't been done previously.

你说得对，这曝光了很多试图在美国语境下利用国内分歧牟利或涨粉的人——他们假装身处美国，实则来自其他国家。

And you're right, it exposed a lot of folks who were trying to basically, in a US context, exploit divisions in The US to, you know, make some money or grow an account from, you know, posing as if they were in The US, but from other countries.

这不仅存在于美国政治中，更是全球政治的现象。

And it isn't just in US politics, it's in global politics.

只要有可激怒的群体，就总有人会从其他国家找到以此牟利的方式，因为他们根本不在乎那个国家的健康发展，只关心最终收益。

If there's a base to be enraged, somebody will find a way to make money for it from it from a different country, because they don't care about that country's, you know, the health of that country and they just care about, the bottom line.

所以没错，今后你会看到更少愤怒诱导内容，因为人们 hopefully 不会再上当——他们能清楚看到发布者的来源地。

And so, yeah, you're gonna see less you're gonna see less rage baiting because people won't fall for it, hopefully, because they'll be able to see exactly where they're from.

这个功能居然拖到2025年才实现，对我来说简直荒谬，不过幸好现在有了。

And the fact that it's taken till 2025, to do this is bananas to me, and, I'm glad it's there.

我认为正因如此，推特变得更好了。

I think it's a much better Twitter because of it.

是啊。

Yeah.

不。

No.

不。

No.

这提醒我们，自2016年以来，我们一直在关注并撰写大量相关内容，深入思考虚假信息问题。我们甚至在边缘领域提出过基础性调整建议，任何平台都可以采纳这些建议来真正改善运作。

I and it's a reminder that, like, since 2016, we've been and I've written a lot about it, thought a lot about it in terms of misinformation, like, there and we even in margins, we've written around, like, basic tweaks you should make to the platform, to any of these platforms to actually make things work better.

这确实是个很好的提醒，表明这些平台其实只需采取些非常直接的措施就能大幅提升运营效果，让人们更清楚自己正在经历什么、看到什么。

And yeah, it's such a it's such a good reminder that like there's really straightforward things all these platforms can do to just make them operate much better and just make people more aware of what is happening to them and what they're seeing.

看起来，埃隆·马斯克和X平台正在引领对抗虚假信息、捍卫自由与稳定民主的先锋行动。

And Elon Musk and X, they they they are leading the charge against mis misinformation and freedom and a stable democracy, it looks like.

等等，你说他们在带头对抗什么？

Wait, they're leading the charge against?

不，不，抱歉。

No, no, sorry.

你说得对。

You're right.

为了。

For.

反对是的。

Against Yes.

错误信息和虚假信息是为了。但我得说，在我的账户上，他们遗漏了一个关键数据点，试图给我打上标签，但他们搞错了。

Misinformation and misinformation for But I will say, on my account, they missed a crucial piece of data that they tried to peg me as, and they got it wrong.

说我从未改过用户名，但我改过一次。

Said I had zero username changes, but I had one.

他们没发现这一点。

They didn't find it.

那是A。

It was A.

坎特·伍兹，现在就叫坎特·伍兹。

Cantor Woods, and now it's just Cantor Woods.

所以对埃隆、尼基塔以及整个团队，我要说开局不错，等改进后再来找我。

So to Elon, Nikita, the whole team there, I have to say good start and back to me when this is better.

我还有件事要说，对所有在创作MAGA相关内容的孟加拉国、印度和尼日利亚账号，我并不是在反对这些人。

And one thing I also have to say is to all my Bangladeshi and Indian and Nigerian accounts out there that are creating MAGA based content, I'm not hating against those people.

说实话我某种程度上觉得，这是个机会。

I genuinely like I kinda in a way, it's it's an opportunity.

他们制作的内容质量不错。

They make good content.

他们擅长制作愤怒情绪类内容。

They make good rage based content.

所以这个机遇是赋予他们的。

So and and it it it was given to them as an opportunity.

虽然我不会特别点名表扬，但我还是要说，这活儿干得不赖。

So I'm I'm not gonna quite shout you out, but I'm still gonna say, I mean, it's not a bad job.

我明白你的意思，但我不能认同。

I see what you're saying, but I can't get on board with that.

不行。

No.

别这样，老兄。

Come on, man.

别这样。

Come on.

我是说，建立这种账号（你声讨人们）这些账号就是想把人们撕裂。

I mean, To build an account that you just these accounts are just trying to rip people apart.

也许吧

Maybe

这就是商业模式，是平台赋予他们的机会。

It's you're gonna be the business model, it's given to them by the platform.

期待更多人会效仿这种做法

Expect more people than to do stuff like

拜托，老兄，别这样。

Like, man, oh come on.

真的吗？

Really?

这太糟糕了。

It's terrible.

这么做太恶劣了。

It's a terrible thing to do.

我认为人们会对激励做出反应。

I expect people to respond to incentives.

有道理。

Fair enough.

总之，我觉得不管出于什么原因，Twitter终于意识到这个问题并提供了我们所需的透明度，这样我们虽然能看到错误信息，但至少知道来源，这是件好事。

And anyway, look, I think it's good that, you know, for whatever reason, finally Twitter has woken up to this and it's given us the transparency that we need to, you know, to we can see the misinformation but at least we know where it's coming from.

这点我们达成一致。

Let's agree on that.

我同意。

I agree.

同意。

Agree.

我很感激

I'm thankful

这一点。

for that.

我很感谢埃隆·马斯克

I'm thankful for Elon Musk

是啊。

Yeah.

还有X。

And X.

做了这件事。

Doing this one thing.

是的。

Yeah.

好的。

Okay.

嗯，Ranjan，我很感谢我们每周五能在这里持续对话，我期待未来能有更多这样的交流。

Well, Ranjan, I'm thankful for our continued conversations here on Fridays, and I look forward to many more.

所以谢谢你。

So thank you.

感恩节快乐。

Happy Thanksgiving.

黑色星期五快乐，我们下周再见。

Happy Black Friday, and let's do it again next week.

好的。

Alright.

下周见。

See you next week.

下周见。

See you next week.

好的，各位。

Alright, everybody.

感谢收听。

Thank you for listening.

我们将在周三回归，届时将有几位Anthropic的研究人员讨论AI模型如何被恶意利用及应对措施。

We will be back on Wednesday, I believe, with a couple of anthropic researchers talking about how AI models can be turned evil and what to do about it.

那我们到时再见，下次《大科技播客》节目再会。

So we'll be back then, and we'll see you next time on big technology podcast.

现在到底发生了什么？为什么会这样发展？

What the hell is going on right now, and why is it happening like this?

在《连线》杂志，我们每天都致力于深挖这些问题根源，或许你也是。

At Wired, we're obsessed with getting to the bottom of those questions on a daily basis, and maybe you are too.

我是凯蒂·德拉蒙德，《连线》全球编辑总监，正在主持我们的新播客系列《大访谈》。

I'm Katie Drummond, the global editorial director of Wired, and I'm hosting our new podcast series, the big interview.

每周，我都会与那些正在塑造当下、最有趣、最具煽动性和影响力的人物进行对话。

Each week, I'll sit down with some of the most interesting, provocative, and influential people who are shaping our right now.

《大访谈》的对话充满乐趣

Big interview conversations are fun

我想要一条能

I want a shark that

吞噬整个互联网、将其全部关闭的鲨鱼。

that eats the Internet that turns it all off.

毫无保留，无所畏惧。

Unfiltered and unafraid.

因此在很多方面，我都尽力成为网络上那些难以想象的保守内容洪流的解毒剂。

So in a lot of ways, I try to be an antidote to the unimaginable faucet of reactionary content that you see online to the best of my ability.

每周我们都会为您提供这个时代最奢侈的东西——意义与语境。

Every week, we're going to offer you the ultimate luxury of our times, meaning and context.

是真是假？

True or false?

你，布莱恩·约翰逊，坐在我对面的这个人，终有一天，在未来的某个尚未确定的时刻，会死去。

You, Brian Johnson, the man sitting across from me, one day, at some point, as of yet undefined in the future, you will die.

错。

False.

详细说说。

Tell me more.

现在就在你收听《连线》杂志《诡异谷》播客的同一平台，收听《大访谈》。

Listen to the big interview right now in the same place you find wired's uncanny valley podcast.

无论你在哪里获取播客，请订阅或关注。

Subscribe or follow wherever you get your podcasts.