这么多字看下来，全是从市场和投资的角度，我还以为在分析大模型的技术发展前景，没想到是炒股的前景，说实话有点失望。
或者说外国人的思路还是不一样，更关心投资市场而不是科技发展和应用。

What Would an AI Crash Look Like?
人工智能泡沫破裂会怎样？
Bubbles are hard to identify and even harder to time, but they aren’t hard to explain. A new technology or some other catalyst raises investors’ expectations, prices rise, and speculators pile in believing that they can resell the asset to someone else for even more money later.
泡沫很难识别，更难把握时机，但解释起来并不难。一项新技术或某种其他催化剂提高了投资者的期望，价格上涨，投机者蜂拥而入，相信他们以后能以更高的价格将资产转售给他人。

The key is that a bubble can form even if lots of investors are rational. So long as there are enough investors trading on momentum — or blind optimism — sophisticated traders will find it profitable to ride the bubble rather than to bet against it. And while bubbles pop, only some set off wider economic and financial crashes like the one the world endured in 2008.
关键在于，即使大量投资者理性，泡沫也可能形成。只要有足够多的投资者追逐动量——或者盲目乐观——那么精明的交易者就会发现，与其做空泡沫，不如顺势而为从中获利。而且，虽然泡沫会破裂，但只有一部分泡沫会引发像世界在 2008 年所经历的那样的广泛经济和金融危机。

That account comes from A Crash Course on Crises: Macroeconomic Concepts for Run-ups, Collapses, and Recoveries (Princeton University Press, 2023) by Markus Brunnermeier of Princeton and Ricardo Reis of the London School of Economics. At an admirable 111 pages, it’s a great resource — and one I’ve gone back to as talk of an AI bubble has returned in recent weeks. But Crash Course is about “macro-financial crises,” not bubbles per se. That makes it helpful not just for thinking through whether there’s an AI bubble, but for imagining what an AI crash would look like.
这一观点出自普林斯顿大学的 Markus Brunnermeier 和伦敦经济学院的 Ricardo Reis 合著的《危机速成课：宏观经济学概念在上涨、崩溃和复苏中的应用》（普林斯顿大学出版社，2023）。这本书篇幅仅 111 页，内容精炼，是一份极好的参考资料——近几周随着人工智能泡沫论的重现，我曾多次翻阅。但《危机速成课》关注的是“宏观金融危机”，而非泡沫本身。这使得它不仅有助于思考是否存在人工智能泡沫，还能帮助我们想象人工智能泡沫破裂会是什么样子。

News Stories Mentioning AI and Bubbles
提及人工智能和泡沫的新闻报道
Concerns have waxed and waned since ChatGPT launched in late 2022
自 2022 年底 ChatGPT 推出以来，担忧情绪时起时伏

Crash Course sketches three models of the run-up to a macro-financial crisis, and each one suggests things to watch for with AI.
《危机速成课》勾勒了宏观金融危机前三种上涨模式，每一种都为观察人工智能提供了启示。

The first is the classic speculative bubble already described, and in the model it’s not enough for everyone to be overly bullish about AI. The real bubble dynamic comes from the combination of naive investors extrapolating — line goes up — combined with investors who know valuations are out of whack but hang on for fear of selling too early. This is what a lot of the AI bubble worry has been about so far: As my colleague Edward Harrison wrote of the AI-heavy S&P 500 in August, “We’re in the unusual situation where fund managers almost uniformly say US stocks are overvalued, yet everyone is piling in.”
首先是经典的投机泡沫，正如前面所描述的，在模型中，仅仅是所有人都对人工智能（AI）过度看好是不足以形成泡沫的。真正的泡沫动态来自于天真的投资者（他们认为“股价只会涨”）的过度推断，与那些明知估值不合理但又害怕过早卖出而继续持有的投资者相结合。到目前为止，这正是许多人对 AI 泡沫担忧的原因所在：正如我的同事 Edward Harrison 在 8 月份撰文谈论 AI 权重股占比较高的标普 500 指数时所说：“我们正处于一种不同寻常的境地，基金经理几乎一致认为美国股市被高估了，但每个人都在涌入。”

That sounds bubblicious, and it’s the dynamic that “gets the party going,” says Reis. But it's not enough on its own to create the sort of crash that he and Brunnermeier write about.
Reis 表示，这听起来确实像个泡沫，并且是“让派对继续下去”的动态。但这本身不足以造成他和 Brunnermeier 所描述的那种崩盘。

The second model in the book involves misallocation — investors mistakenly backing the wrong firms in the hot sector because some policy or distortion biases what gets funded. The third concerns the role of shadow banks that are vulnerable to runs and amplify dips in asset prices.
书中提出的第二种模型涉及资源错配——投资者因为某些政策或扭曲因素影响了资金的流向，从而错误地投资于热门行业中的错误公司。第三种模型则关注那些容易遭受挤兑并放大资产价格下跌的影子银行的作用。

Applied to AI, Crash Course’s brief tour of crisis economics suggests three gauges of the risk:
将这些应用于 AI 领域，《崩盘课程》对危机经济学的简要回顾提出了三个风险衡量指标：

Are investors buying or lending to AI firms without any conviction except the belief that they can unload later to the ‘greater fool’?
投资者是在毫无信念的情况下购买或借贷给人工智能公司，除了相信他们以后可以将它们转售给“更大的傻瓜”？

Are there barriers — like, say, VCs investing in anything labeled AI — that are preventing money from reaching the most promising firms?
是否存在一些障碍——比如风险投资公司投资任何打着人工智能旗号的公司——阻碍了资金流向最有前景的公司？

How much of the AI boom is funded by debt and and are those lenders vulnerable to runs?
人工智能热潮中有多少是靠债务融资的？这些贷款方是否容易受到挤兑的影响？

‘Yes’ to the first one would signal a bubble. ‘Yes’ to all three would warn of a crash.
第一个问题回答“是”将预示着泡沫。三个问题都回答“是”则预示着崩盘。

— Walter Frick, Bloomberg Weekend
— 沃尔特·弗里克，《彭博周末》

Predictions 预测
“Big AI will find itself fairly friendless in DC.” The left will be skeptical of giant corporations that use lots of resources and threaten jobs. The right will be wary of a black box created in a lab. — Joe Weisenthal, Odd Lots
“大型人工智能将在华盛顿举步维艰。”左翼将对那些消耗大量资源并威胁就业的大公司持怀疑态度。右翼将警惕实验室里创造出的“黑箱”。——乔·韦森塔尔，《Odd Lots》

The AI boom will require lots of nuclear power: “Soaring demand for electricity will drive a $350 billion nuclear spending boom in the US,” says Bloomberg Intelligence. — Will Wade, Bloomberg News
人工智能热潮将需要大量核能：“彭博情报公司称，飙升的电力需求将推动美国核能支出达到 3500 亿美元的繁荣。”——威尔·韦德，《彭博新闻》

Reactor Revival 反应堆复苏
US will spend more than $350 billion on nuclear power through 2050
美国将在 2050 年前在核能领域投入超过 3500 亿美元

The material of the 21st century will be… metal-organic frameworks, for which the Nobel Prize in Chemistry was awarded this week. The recipients’ work pioneered “creating molecular constructions with large spaces through which gases and other chemicals can flow… [with potential applications like] capturing carbon dioxide and harvesting water from desert air.” — Charles Daly, Bloomberg News
21 世纪的材料将是……金属有机框架，本周因此获得了诺贝尔化学奖。获奖者的工作开创了“创造具有大空间的分子结构，气体和其他化学物质可以从中流动……[潜在应用包括]捕获二氧化碳和从沙漠空气中收集水。”——查尔斯·戴利，《彭博新闻》

The risk of another World War is growing: “Memories of the last world war have faded, and the current generation of leaders and experts — from China and Russia to the US and elsewhere — is showing signs of waning humility and growing hubris, similar to European leaders in the summer of 1914.” — Andreas Kluth, Bloomberg Opinion
世界大战的风险正在增加：“上一场世界大战的记忆已经淡去，而当前一代的领导人和专家——从中国、俄罗斯到美国及其他地区——正表现出谦逊感减退和傲慢感增强的迹象，这与 1914 年夏天欧洲领导人的情况相似。”——安德烈亚斯·克鲁斯，《彭博社观点》

NATO may build a “drone wall” along Europe’s eastern border. It would take years, and no one is quite sure how it would work. But the idea is gaining momentum and would involve a system that would likely take cues from Ukraine. — Gerry Doyle and Jake Rudnitsky, Bloomberg News
北约可能在其欧洲东部边境修建一道“无人机墙”。这将耗费数年时间，而且没有人确切知道它将如何运作。但这一想法正获得动力，并将借鉴乌克兰的经验。——彭博社 Gerry Doyle 和 Jake Rudnitsky

The dollar isn’t going anywhere: It’s still on one side of 89.2% of all FX trades, according to the BIS. — Daniel Moss, Bloomberg Opinion (Although, counterpoint: Gold’s rally is helping China challenge the dollar.)
美元不会消失：根据国际清算银行的数据，美元仍占所有外汇交易的 89.2%。——彭博社评论员 Daniel Moss（尽管也有反驳观点：黄金的上涨正在帮助中国挑战美元的地位。）

What Are the Chances...
人工智能泡沫破裂会怎样？——彭博社
Q3 Prediction Markets Recap
第三季度预测市场回顾

Every quarter we look back at the markets cited in the newsletter: Here’s Q1 and Q2. The second quarter was rough in terms of accuracy, as lots of plausible-but-unlikely events kept happening — mostly involving Donald Trump.
每个季度，我们都会回顾新闻通讯中提到的市场：这是第一季度和第二季度的回顾。第二季度在准确性方面很艰难，因为许多看似可能但不太可能发生的事件接连发生——其中大部分都涉及唐纳德·特朗普。

This quarter markets did better. Seven markets that we’ve cited closed, five to ‘Yes’ and two to ‘No.’ At the time we cited them, the average forecast for the Yes outcomes was 67%, and for the No outcomes 36%. Six out of the seven leaned in the direction of the actual outcome. Traders expected the Fed to cut in September and Francois Bayrou’s ouster as France’s prime minister. Way back in November of last year they expected Trump to implement “large” tariffs.
本季度市场表现更好。我们提到的七个市场已经收盘，其中五个为“是”，两个为“否”。在我们提到它们时，对“是”结果的平均预测为 67%，对“否”结果的平均预测为 36%。七个结果中有六个与实际结果方向一致。交易员预计美联储将在九月份降息，并预计弗朗索瓦·贝鲁将被罢免法国总理职务。早在去年十一月，他们就预计特朗普将实施“大规模”关税。

As for misses, Polymarket put just a 35% chance on Jimmy Kimmel returning to air by mid-October and that happened.
至于失误，Polymarket 对吉米·坎摩尔在十月中旬前重返荧屏的几率仅为 35%，但事实却发生了。

Here are a few other markets we’ve cited and where they are now:
以下是我们提到的一些其他市场及其现状：

How Prediction Markets Have Changed Since We Cited Them
预测市场自我们引用它们以来发生了哪些变化
A Google breakup is less likely, tech layoffs and a US-Venezuela conflict are more likely
谷歌分拆的可能性较小，科技行业裁员和美委冲突的可能性较大

Amateur Forecasters vs. Traders
业余预测者 vs. 交易员

Twice this year we’ve cited forecasts about a potential ceasefire between Russia and Ukraine — in May and in August — and both times we compared the real-money trading platform Polymarket to the no-money crowd forecasting platform run by the think tank Rand. (Disclosure: I was once a paid forecaster and freelance writer for the platform before it was absorbed by Rand.)
今年我们曾两次引用关于俄罗斯与乌克兰之间可能停火的预测——分别在五月和八月——两次我们都将真金白银的交易平台 Polymarket 与兰德智库运营的无资金预测平台进行了比较。（披露：我曾是该平台付费预测员和自由撰稿人，后被兰德收购。）

In May, Polymarket put the chances of a ceasefire this year at 45%, while Rand’s consensus forecast was just 26%. In August, Polymarket was at 35% while Rand was at 11%.
今年五月，Polymarket 将停火的可能性定为 45%，而 Rand 的共识预测仅为 26%。八月，Polymarket 的预测为 35%，而 Rand 为 11%。

Today, Polymarket is much closer to Rand: Polymarket puts the chances at 11% while Rand is down to 4%. We won’t have an official outcome until the end of the year, of course, but right now it looks like the amateur forecasters were ahead of the traders.
如今，Polymarket 的预测已非常接近兰德的预测：Polymarket 将概率定为 11%，而兰德则降至 4%。当然，我们直到年底才能得知官方结果，但目前看来，业余预测者似乎比交易员更胜一筹。

Keep an Eye On
密切关注
Prediction Markets vs. AI
预测市场 vs. 人工智能

The big news in prediction markets this week was that Intercontinental Exchange Inc., owner of the New York Stock Exchange, announced that it was investing $2 billion into Polymarket. (You can read more about what that means for prediction markets and for finance from Bloomberg Opinion’s Matt Levine, John Authers and Aaron Brown.)
本周预测市场的一则重磅新闻是，纽约证券交易所的所有者洲际交易所公司（Intercontinental Exchange Inc.）宣布将向 Polymarket 投资 20 亿美元。（您可以阅读彭博社评论员 Matt Levine、John Authers 和 Aaron Brown 的文章，了解这对预测市场和金融意味着什么。）

Perhaps just as important, though, was a report from the Forecasting Research Institute, a think tank — where, disclosure, I used to be a contributing editor.
但或许同样重要的是，一份来自智库“预测研究所”的报告——在此披露，我曾是该研究所的一名特约编辑。

FRI reports that large language models now outperform the median human forecaster, a reversal from a year ago. The very best human forecasters still were more accurate than AI, but FRI’s researchers project that will reverse by 2026 as AI continues to improve.
据 FRI 报道，大型语言模型现在的表现已超过了人类预测者的中位数水平，这与一年前的情况恰恰相反。最顶尖的人类预测者仍然比人工智能更准确，但 FRI 的研究人员预计，随着人工智能的持续改进，到 2026 年这一情况将发生逆转。

Other forecasting organizations have seen similar trends. Last month the forecasting website Metaculus reported that the startup ManticAI had secured 8th place in its summer competition, the highest result by an AI system to date.
其他预测机构也看到了类似的趋势。上个月，预测网站 Metaculus 报道称，初创公司 ManticAI 在其夏季竞赛中获得了第八名，这是迄今为止人工智能系统取得的最高成绩。

Prediction markets can be a great way of harnessing the dispersed wisdom of many human minds. But they may be headed to the mainstream just as AI is proving its ability to exceed the wisdom of crowds. In the meantime, as more investment flows to prediction markets, expect traders there to rely more and more on AI to make their predictions.
预测市场是汇集众人智慧的绝佳方式。但它们可能正要进入主流，而此时人工智能却已证明其超越群体智慧的能力。与此同时，随着越来越多的投资涌入预测市场，可以预见，那里的交易者将越来越依赖人工智能来做出预测。