内特·西尔弗谈人生的混合策略

《与泰勒对话》由乔治梅森大学莫卡特斯中心制作，旨在弥合学术思想与现实问题之间的鸿沟。了解更多请访问mercadis.org。每期对话的完整文字记录及附带实用链接，请访问conversationswithtyler.com。大家好，欢迎收听《与泰勒对话》。今天我在纽约与内特·西尔弗进行现场对谈。

One: Conversations with Tyler is produced by the Mercatus Center at George Mason University, bridging the gap between academic ideas and real world problems. Learn more at mercadis.org. For a full transcript of every conversation, enhanced with helpful links, visit conversationswithtyler.com. Hello everyone, and welcome to Conversations with Tyler. Today I'm chatting with Nate Silver, live in New York.

我们在此庆祝内特的《悬崖边缘：孤注一掷的艺术》平装版发行——这本书去年我曾由衷评价为'每一页都充满纯粹乐趣'。我们还将重温一年前和九年前的一些预测，并畅聊所有我想探讨的话题。内特，欢迎你。泰勒，总是很愉快。在当前边际下，你从研究期望效用理论中学到什么新东西了吗？

We are here to commemorate the paperback edition of Nate's On the Edge, The Art of Risking Everything, a book that last year I described sincerely as absolute fun on every page. We're also gonna revisit some of our earlier predictions from a year ago and nine years ago and talk about everything I want to talk about. Nate, welcome. Tyler, always a pleasure. At current margins, do you learn anything from studying expected utility theory?

嗯，我大概花十分之一时间玩扑克，在这方面你确实会明确计算这类数值，对吧？

Well, I mean, I probably spend a tenth of my time playing poker, and certainly in that respect, you're quite explicit about calculating things like that. Right?

但你是学到新观点、新理论、新概念，还是仅仅在运用比如十三年前学到的知识？

But are you learning new ideas, new theories, new concepts, or are you just applying what you learned, say, thirteen years ago, whenever?

不，我感觉它对我依然新鲜。当我和人们讨论这本书时，大家似乎能理解期望值部分，这对经济学入门者很容易解释清楚。

No. I feel it still feels fresh to me. I mean, I think so when I've talked about the book to people, people kind of understand the expected value part. Right? That's easily explicable to economics one zero one people.

但求解均衡点的部分人们理解较少——即在给定每个人都理性出牌动机下应该形成的均衡状态。这种思考甚至延伸到我分析网络言论时，比如人们为何在推特或X平台上表现出特定行为。

I think the equilibrium part that solve for the equilibrium is something that people don't get as much. It's like, what's the equilibrium that should emerge given everyone has their incentives to play this hand optimally, so to speak. Right? Or everyone has different incentives, but they're all being rational in some capacity. Like, that kind of thing I think about a lot when it comes to even like online discourse, for example.

你知道，为什么人们在推特或X上会有某种特定行为。

You know, why are people behaving a certain way on Twitter or X, for example.

说说随机纳什均衡。你认真看待这个概念吗？它在足球点球中似乎成立。

Say random Nash equilibrium. Do you take that seriously? It seems to hold with soccer kicks.

对吧？但在扑克中你会运用这个理念吗？当然。我有时会刻意随机化——比如看锦标赛计时器，数字大时就虚张声势，数字小时就打得保守些。

Right? But in poker, anything you do, you ever use the concept? Oh, sure. No, I will literally randomize sometimes, right? Where you look at the tournament clock, and if it's a high number, then you'll bluff, and it's a low number, then you'll play more passively, for example.

你还可以观察筹码旋转之类的。我是说...（思考中）

You can look at the rotation of your chips or things like that. I mean, look. And you think some of

其他人也这么做。

the other people do it too.

是啊。你能看出某种迹象。看到人们抬头看时钟，就像在发呆，他们在等待某个数字出现，对吧？你可以旋转你的牌。

Yeah. You can tell there's a certain look. You see people are like looking up at the clock, and it's like a kind of, I don't know, daydreaming, and they're and they're waiting for a number. Right? You can like rotate your cards.

人们直到真正解出扑克的纳什均衡时才意识到混合策略有多少。基本上，每一手牌都是某种混合策略，对吧？反过来讲，如果你能察觉到对手的任何破绽或解读出信息，就能从完全无差别策略转向在该情境下的优势策略。

One thing people didn't realize until they actually solved the Nash equilibrium for poker is how many mixed strategies there are. Basically, every hand is a mix of some kind. Right? Yeah. The flip side to that is that if you have any tell or read on your opponent at all at all, you move from literal indifference toward a dominant strategy in that context.

对吧？所以我其实很相信观察玩家的微表情，把握当下氛围，关注他人对你的看法以及你在牌桌上的形象变化。不过扑克确实是现实生活中博弈论最彻底的体现，甚至超过其他领域。比如在橄榄球等运动中也能看到。

Right? So I believe actually in, like, reading players tells, picking up on the moment and how you are perceived and how your table image shifts a lot. But no. I mean, poker has gone further than anything else in the literal manifestation of game theory in real life. I mean, you also see it in in football and things like that.

我正在做NFL的战术模型。我在想NFL球队是否在最优地混合他们的策略，比如在三档长码数时选择跑球战术，由于出人意料且本应选择传球，其期望价值反而更高，对吧？

I'm working on an NFL model now. Right? I wonder if NFL teams are optimally mixing up their strategies, for example, but you see it there. Right? Like a a draw play on third and long, right, has a higher expected value because it's not expected, and because you're supposed to always pass there, right?

经济学最神奇之处在于它能预测人类行为——在重复试验且参与者能获得良好反馈并有优化动机的情况下，我认为预测相当准确。

And economics is amazing in the sense that it predicts human behavior, I think fairly well when you have repeated trials and people that have good feedback and incentive to be optimal.

你解读微表情的能力，在现实生活中对你有帮助吗？如果有，具体是怎样的？

Your ability to read tells, does it help you at all in real life? And if so, how?

可能对某些人有帮助，但非常特定。比如我来的时候坐地铁没座位，我判断那个人要起身了，结果错了。微表情并非总是百分百准确。

There's probably some people reading, but it's very specific. Right? I was on the subway getting here. Right? I didn't have a seat on the train.

我当时觉得，哦，那个人要站起来了。我能看出来。但我错了。对吧？这些微表情并不总是完全可靠的。

I like, was oh, that person's gonna stand up. I can tell. And I was wrong. Right? The tells are not always a 100%

可靠的。如果有人想学习如何解读微表情，除了多打扑克或实际练习之外，

reliable. If someone wants to learn how to read tells other than just say playing a lot of poker or doing the thing,

他们是怎么做到的？我认为没有什么能替代它与现实世界情境的关联性，因为这是非常依赖语境和语义的。六月份大部分时间我都在世界扑克系列赛，参加了主赛事——一场1万美元买入的比赛。有个人的颈动脉（或者叫颈动脉？我不确定具体名称）跳动得非常剧烈，你能直接看到他脖子上的心跳。这通常是个明显的破绽信号，但有些人拿到强牌时反而会更紧张。

how do they do that? I think there is not a substitute for how it correlates to real world scenarios because it's so kind of contextual and kind of semantic. So I was at the World Series for most of June, played in the main event, which is a $10,000 tournament, and there was a guy who had a very rapid heartbeat that you could see his It's called the cartoid artery, or cartoid, I don't know what it's called, right? But his heart's beating in his neck, and so that's obviously often a tell. However, for some people, they get more nervous when they're bluffing.

有些人拿到强牌时会更紧张。对吧？所以你必须结合行为来综合判断，这需要具体情境分析。考虑到他只是个中等水平玩家，不是顶尖高手，也不是我们扑克里说的'鱼'（新手）。不，你需要在脑海中建立一种隐性的数据库，我觉得就是通过大量观察积累的。

Some get more nervous when they have a strong hand. Right? So you have to like kind of correlate that with with behavior, and that's contextual. The fact that he was an intermediate player and not a really good player, or a complete novice fish, we call him in poker. So, no, you need you're building up kind of an implicit database in your head, I think, just by watching a lot.

包括在你没参与牌局时也要观察。玩扑克时你经常要弃牌，大部分时间都在观察别人，我会做些小预测：'他这个动作看起来像有强牌'。如果你能有60%的准确率，相比50%的随机猜测就是巨大优势。

And including watching when you're not involved in a hand. In poker, you're folding, and most films are poker most of the time, but I'm watching people and kind of making a little prediction. Oh, that feels strong, something he did. Right? And then, you know, if you're right 60% of the time, 60 instead of 50 random, right?

这在赌博中是巨大的优势。任何55比45的优势都不得了，60比40或65比35的优势就能让你成为常胜玩家。

That's a huge edge. And in gambling, any 55, 45 edge is enormous. Right? And sixty forty or sixty five thirty five is will make you a very winning player.

看NBA比赛时，你能通过肢体语言解读出有效信息吗？比如'卢卡今晚情绪不好'这种？

And if you're watching an NBA game, is there any meaningful way in which you can read tells? Like, oh, Luca has a bad attitude tonight.

我不擅长这个。不过有时候能看出球队的整体状态。我今年NBA季后赛投注成绩不错，关键要看'这支球队今晚在状态吗？'理论上这应该具有操作性，虽然市场可能已经反映了这些信息。但确实存在'这支球队累了'之类的信号。

I'm not. I mean, you can tell from the body language of a team sometimes. I had a fairly good year betting the NBA playoffs, and does the team have it tonight? I mean, that should be in principle actionable, I would think, but you would think that markets might pick up on that. But yeah, there are things like this team is tired.

他们阵容不齐整。比如掘金队对雷霆的系列赛，掘金是强队，表现超出我预期，但他们没法跟对手持续抗衡四节。不过这更多是基于背景知识的判断，而非纯粹的肢体语言解读。

They don't have the personnel. The nuggets in that OKC series, they're a great team. They beat my expectations, but they're not gonna be able to keep up with this team for four quarters. But that's not like physical read so much as kind of Again, it's kind of a little bit of physical observation and a lot of priors in context.

你最近从政治学学术文献中学到什么有趣的东西？

What was the last interesting thing you learned from the academic political science literature?

哎呀，这个说出来可能显得有点冒犯...

Oh boy, this is going to seem like an insult to

别担心，我们这里欢迎各种冒犯性言论。

No, no. We're all for insults here.

我看了很多Substack之类的文章。天哪，我也不确定。

I look, I read a lot of substacks and things like that. Oh my gosh. I don't know.

我认为这可能是正确答案，明确地说。

I think that might be the right answer, to be clear.

不。你看，我觉得在某些方面，学术圈产生的东西挺有意思。比如有效利他主义就很有趣。批判理论或所谓'觉醒主义'的起源对更广泛的话语体系产生了很大影响。

No. Look, I I think in some ways, you know, what has come out of academic circles have been interesting. I mean, I think effective altruism has been interesting, for example. You know? I mean, certainly critical theory or whatever the origins of wokeness have had a lot of influence on on the discourse more broadly.

但说实话，我现在不怎么读期刊论文了。很多讨论都转移到Blue Sky上，而我实在受不了那个平台——感觉就像个自我陶醉的回音室。所以我觉得学者们在这方面可能失去了影响力，尽管我和学者们的交流经历很棒，每年也会在大学做几场活动。

But like, yeah, I don't I don't really read a lot of journal articles anymore. It's also a lot of it's gone on Blue Sky, I think, and I can't tolerate Blue Sky really. It's too much of a circle jerk, I think. And so, you know, I think academics have maybe lost influence in that in that respect. Even though I had great experiences talking to academics, and I'll do events at universities a couple of times a year.

他们还陷入了一种思维定式——对特朗普条件反射般的反对（虽然我也反对特朗普）。但这种倾向某种程度上削弱了学术应有的细致入微，那种'慢火烹饪'式的思考方式本应更从容，却不太适应当今新闻周期快速反应的需求——当然你是少有的例外。

And they also kind of fall in this trap where, like, it's so reflexively anti Trump, a lot of it, and I'm anti Trump, right, in most senses of that term. But something about it has kind of melted the brains a little bit of the nuance and the subtlety and the things that The slow cooking method of academia where you're supposed to take more time with things, think is not always a good match, you being one of the major exceptions, for kind of rapid fire reaction to the news cycle.

如果有一篇你希望有人写的论文，会是什么主题？

If there were a paper you wish someone would write, is there such a thing?

你是说某个你想了解但尚未有人研究，理论上学者可以完成的课题？也许现在论文本身就不太合适，像花几天写的Substack文章、博客或书籍可能更好。我依然相信书籍能比任何形式都更深入探讨结论。比如我来时听的播客里，那位揭露爱泼斯坦案的《迈阿密先驱报》女记者的调查报道。

Something you'd like to know, but you don't know now, and in principle, an academic could do it? Yeah. Maybe papers is kind of the wrong part of the curve right now, where like, yeah, like a Substack newsletter that you take a couple of days with or a blog post, right, or a book. I mean, I'm still a big believer in books where you kind of go into a conclusion more in-depth than anyone has before, right, or investigative journalism. You know, I was just listening on that on the way over to like the Rust Outhead podcast with the woman who kind of helped to break the story of Jeffrey Epstein for the Miami Herald.

这可是涉及众多国际知名恶棍的大新闻。但人们总是懒惰的，当时没几个记者真正深挖这个案件的每个细节。

She's like, this was a major story featuring all types of people that are extremely well known and notorious internationally. Right? But people are kind of lazy. Right? And there were not like a lot of reporters really kind of pursuing every detail of that case.

所以我觉得90%的学术论文完全可以用博客文章替代。包括那种'我做了些回归分析'的腔调——与其用希腊符号之类的装腔作势，不如直接说出你的想法，

And so I think that 90% of academic papers would work perfectly fine as blog posts, right? Including the tone of like, okay, I ran some regressions instead of having Greek symbols and things like that, and the pretense of all of it. Just say what you think,

然后让世界去评判。

right? And let the world sort it out.

说出你的想法。另外，我喜欢写通讯的一点是，我能用自己的语气表达：什么时候我在稍加推测？什么时候我在开玩笑？什么时候我更严肃地呈现一些原创发现之类的内容？而这些在典型学术论文那种枯燥、刻板的语气中可能会丢失。

Say what you think. And also, one of the things I like about writing a newsletter is I can use my tone to say, When am I speculating a little bit more? When am I making a joke? When am I being more serious presenting some original finding and things like that? And that can be lost with dry, sterile tone of a typical academic paper.

那么当你考虑最大化期望值时，你该对长寿有多执着？你应该成为下一个布莱恩·约翰逊吗？是的，我经常问自己

So when you think about maximizing expected value, how obsessed should you be with longevity? Should you be the next Brian Johnson? Yeah. I I ask myself a lot

这些问题。我不知道。我认为如果你过度优化日常生活中的每一个参数，最终可能会过得相当痛苦。我的意思是，如果我说我非常关心长寿，那可能有点虚伪。到了四十多岁，无论是出于选择还是被迫，你确实需要对某些事情更加谨慎一些。

of these questions. I don't know. I think if you hyper optimize for every parameter in your day to day life, then you probably wind up being fairly miserable. I mean, I guess it'd hypocritical if I said I was terribly concerned about longevity. Mean, in your forties, are kind of Both by choice and by force, you kind of are required to be a little bit more careful about certain things, I suppose.

是的，我还没有完全加入那个行列。我相信以后会的。也许三年后我们再聊时，我会变成——谁知道呢——某个健康狂人什么的。

Yeah. I haven't gotten on that bandwagon quite as much. I'm sure I will. Probably if we do this again in three years, I'll be, I don't know, right? Some health nut or something.

但目前，我正在做严肃但渐进的改进。话说回来，你认真对待帕斯卡的赌注吗？我确实认真对待我们对宇宙本质极度不确定这一观点。

But for now, for now, I'm making serious, but incremental improvements, I guess. As a better, do you take Pascal's wager seriously? I take the notion that we're profoundly uncertain about the nature of this universe seriously right

为什么不相信呢？对吧？这似乎能最大化期望值。直接投入其中。你不必说你知道了答案，但你可以说：'这是我找到答案最可能的路径，从相信某事开始。'

Why not believe? Right? It would seem to maximize expected value. Just jump right in. You don't have to say you know the answer, but you would say, Well, this is the most likely path I have for getting to the answer, is to start by believing something.

嗯，我并不真的相信天堂这类概念，也不完全认同帕斯卡的观点。我是说，我给这类事的概率大概是p=0.0

Well, I don't really believe in heaven as such, right, or I don't believe in kind of like Pascals I mean, I put, you know, p equals point o

01。好吧，p=1，但总归存在某种可能性。

o one. Well, p equals one, but there's some chance.

对吧？确实存在可能性。但你看，在所谓的无限伦理问题中，现在就有关于虾类福利的大讨论。

Right? There's some chance. But like, mean, look, a lot of problems in, I guess it's called infinite ethics, you now have this big debate over shrimp welfare. Sure.

举个例子。为了A

Example. For A

很多这样的例子。我就想，我们已经有足够多在地球上可解决的问题了，所以我不会太担心那些事情。我认为作为思维实验，它们可能是有帮助的，这没问题。但如果你忽略那1%的边缘情况，可能让像威尔·麦卡斯基尔这样的人去操心那些事也挺好。

lot of them. And I'm like, We have enough problems that are in the terrestrial tractable realm, Well, I'm just not gonna worry about those things so much. And I think as thought experiments, they can be helpful and fine. Right? But like, I think if you kind of like, you know, ignore the 1% of edge cases, it's probably good to have some people, Will McCaskill or whatever, worrying about those things.

确实需要有人对这些事施加影响，但我更关注可操作的领域。我一直更像扑克玩家——重要的是你手上的牌，重要的是你的判断，理论固然重要，但具体、本地化且可执行的信息总是占主导地位。

There's shrimp well for people having some influence on that, but I'm trying to move in the realm of the tractable. I've always been more poker. What matters is what hand you have, what matters is the read that you have, and the theory matters, but information you have that's local and specific and actionable always predominates that.

所以你简而言之是摒弃了帕斯卡的赌注。

So you dismiss Pascal's wager, in short.

我不会为帕斯卡的赌注操心。就像我说的，我的立场已经从无神论者逐渐转变为更明确的不可知论者，但这并非因为帕斯卡赌注本身，而是因为这个世界的奇妙。确实，我认为某些人工智能的发展应该让人们多思考现实的本质。

I don't worry about Pascal's wager, I guess. Again, as I said, I've evolved toward being more affirmatively kind of agnostic as opposed to atheist, I suppose. But that's not because of Pascal's Wager per se, just because of the strangeness of the world. You know, I do think that like some of the AI stuff ought to have people asking a few more questions about the nature of reality, I suppose. Yeah.

你的书中有大量关于冒险者的内容，副标题就叫《孤注一掷的艺术》。你认为人们会在多大程度上区分不同类型的冒险？比如有人会说，泰勒你去过危险地区旅行，但我完全不敢深潜——我觉得这种风险隔离非常明显。

Now there's plenty of material in your book about people who take a lot of risk. Subtitle is The Art of Risking Everything. How much do you think people segregate areas of risk taking? So for instance, people might say, well, Tyler, you traveled to some dangerous locations, but I'm totally afraid to scuba dive. I feel I'm very strongly segregated.

我要么承担相当风险，要么几乎完全不冒险。这是你对大多数人的模型吗？这很正常。我书中用的例子是...

I either take a fair amount of risk or close to none at all. Is that your model of most humans? That's pretty normal. Yeah. Mean, the example I think I use

以西结·伊曼纽尔。他坚持不去餐厅，因为2022、2023年仍有新冠疫情，但他却骑摩托车——这明明是单位里程最危险的行为之一。人们在风险认知上通常缺乏元理性。

in the book is Ezekiel Emanuel. You Darn shouldn't go to restaurants because we still have some COVIDs like now in 2022, 2023, right? But he rides a motorcycle, which is like known to be one of the most dangerous things per mile that you can do. Yeah. People are are usually not super meta rational about risk.

对吧？

Right?

但他们应该这样吗？比如，应该保持统一的风险态度吗？这样更符合元理性吗？

But should they be? Like like, should you have one general risk attitude? Is that more meta rational?

听着，那些看似非理性的行为，往往在更高层面上是理性的。

Look. I think things that seem irrational are often kind of rational on on a higher plane. Right?

我吗？

Do I.

对，对。就像，你知道的，损失厌恶。懂吗？你大概不知道，我前几天刚被偷了一副耳机。

Yeah. Yeah. Like, you know, loss aversion. Know? You shouldn't know, I had a pair of headphones stolen from me the other day.

对吧？我为此感到特别内疚。我们没锁车门。某种程度上，你知道，算是我的错。

Right? And I felt very kind of guilty about that. We hadn't locked the door of our our car. It was kind of, you know, kind of my fault.

但这是在纽约发生的？

But like This is in New York?

其实是在纽约上州。

In actually, in Upstate New York.

嗯。

Yeah.

在纽约上州可得小心点。

You gotta be careful of Upstate.

盗贼真正的老巢啊。对对，没错。

Real home of thieves. Yeah. Yeah. Sure.

但你会想，好吧，我还买得起新耳机。不是什么大事。但如果你长期这样马虎，那就会造成更多麻烦，也是种浪费。所以像损失厌恶这类心理，未必...我是说，它们源于理性思维的缺陷迭代，可能在更高层面有其作用，或者在人类文明早期曾有过进化意义。

But like and you're like, oh, okay. Well, I can afford a new pair of headphones. Not that big a deal. But like but if you were chronically sloppy about things like that, then that would cause more hardship and would be wasteful. And so like, know, loss aversion, things like that aren't necessarily I mean, they come from flawed iterations of rationality that might serve a serve a higher purpose or might have served an evolutionary purpose earlier in kind of human civilization.

你能想到真正整合了风险态度的人吗？就是他们有某种程度的冒险倾向，并且在不同事情上大致均衡地应用。我是说，看看那些...或者那些人干脆都在监狱里？

Can you think of cases of people who have truly integrated attitudes toward risk? That is, they have some level of risk taking. They apply it more or less evenly to all the things they do. I mean, look at Or those people just in jail?

是的。不。我是说，看，我认为硅谷有些人确实有点那种态度。你知道，这往往与其他可能不太理想的特质相关。我觉得，计算风险的部分很难把握。

Yeah. No. I mean, look, I think some of Silicon Valley has that attitude a little bit. You know, it tends to be correlated with other things that might be less desirable. The calculated risk taking part is hard, I think.

对吧？如果你非常成功——我是说，玩扑克时你会发现，人们连胜时打法完全不同。自信时打法也截然不同。实际上他们通常会打得更好，明显更好，人类可能也是如此。但我在职业生涯中也有过飞得太靠近太阳的时刻，那时你开始收到越来越糟的反馈。这非常——非常有害，对吧？

Right? If you're very successful I mean, thing you see when you do play poker is like people play very differently when they're on a winning streak. People play very differently when they're confident. They usually play better, quite a bit better, in fact, and probably true for human beings too, but I've had moments in my career where I flew way too close to the sun and you stop getting you begin getting worse feedback from people. That's very that's very damaging, right?

当人们可能对你阿谀奉承，当你屏蔽批评时。我是说，埃隆似乎有点这个问题，尽管埃隆至少——比如他没有在关税和技术移民问题上向特朗普让步，所以他至少还有些原则。但这是个很大的问题，而且你身边围满了唯唯诺诺的人。

Where people are sycophantic around you potentially, where you shut off criticism. I mean, Elon seems to have this problem a little bit, although Elon at least has kind of Elon didn't concede to Trump on tariffs and skilled immigration, things like that, so he had some values at least, but that's a very big problem, think, and you're surrounding yourself with with yes men.

你为什么还没去过俄克拉荷马州？我想去

Why haven't you been to Oklahoma yet? I'd like

我本来计划——我希望尼克斯队能对阵雷霆队，这样我就能去那儿看几场比赛。飞往俄克拉荷马城，住在拉马达这类酒店，甚至可能是家不错的酒店，比去麦迪逊广场花园看季后赛便宜多了。

to go. I I had designs on I was hoping the Knicks were gonna play OKC, then I'd have gone for a couple games there. It's cheaper to fly to OKC, stay in a in a Ramada something, or maybe probably even a nice hotel than to go to Massasburg Garden for for playoff games.

我本来想尝试在俄克拉荷马城看场季后赛，不是总决赛。门票太便宜了，我都惊呆了，因为我习惯了湾区和高雄市的价格——虽然我不会买，但我会看看然后说，哦，太贵了。

I was thinking of trying to see a playoff game in OKC, not the finals. The tickets were so cheap. I was stunned because I'm used to, like, Bay Area and New York City prices, which I won't buy at, but I look at them, and then I say, oh, that's too much.

纽约市有点疯狂——我是说，人们发现一点：所有富人真正喜欢的高端消费，价格都暴涨了。这就像世界上最小的小提琴问题。比如高端寿司Omakase。

New York City is kind of crazy for I mean, I think one thing people figured out is like, all the kind of high taste things that rich people really like, there's been huge inflation. This is like the world's tiniest violin problem in like Yeah. High end sushi omakaze, for example.

我现在不这么做了。

I don't do it anymore.

是啊。不。这是——知道吗？就像美网公开赛，去法拉盛草地看比赛本来是件很有趣的事。对吧？

Yeah. No. It's it's it's know? Or like the US Open used to be something really fun going to the US Open in Flushing Meadows. Right?

去早期场次，你可以随便去17个球场中的任何一个。过去挺划算的，50美元就能进场，现在要200美元，这算是合理价格。但现在你看到达美航空用AI工具进行更多价格歧视，我觉得我们可能——

Go You to like one of the early sessions. You can go like to any of the 17 courts or whatever. It used to be a pretty good deal. You get in for like $50, and now it's like $200, right, which is kind of the fair price. But like, you you see now Delta using AI tools to price discriminate a little bit more, which I think we probably

那对你来说会很糟糕。

should That's gonna be bad for you.

是啊，很可能。我们或许该制定些法律。不过在纽约，我们可以选择多家航空公司飞往几乎所有地方。对，不，我觉得——

Yeah, probably. We should probably have some law. Although in York, have In New York, we can fly on multiple carriers to almost everywhere. Yeah. No, I think

但如果它们都用同一套算法进行价格歧视，虽不算严格意义上的串通，但某种程度上类似共谋。这合法吗？你比我更清楚。目前有悬而未决的诉讼案，我不确定判决结果如何，但这将成为日益重要的问题。

it's But if they all price discriminate using a common algorithm, it's not literally collusive, but it's in a way like collusion. Is that legal or not? You would know more than me. There are pending court cases. I'm not sure how they're all being decided, but it will be an increasingly important issue.

确实。我认为其中有些做法是短视的——很可能会引发强烈的反弹。而且直觉上就让我觉得不对劲。

Yeah. I mean, I think there'll be I think some of this is short sighted the sense that we'll probably have a pretty profound backlash to it, I would think. Yeah. And I mean, it rubs me the wrong way intuitively.

如果你在赌NBA比赛或只是想享受观赛，现场看球队比赛有多大帮助？比如去看几场俄克拉荷马队的比赛，能获得什么有用信息吗？

If you're betting on NBA games or just trying to enjoy them, how much does it help you to have seen the team live? Like, if you go to see Oklahoma a few games, do you learn anything?

考虑到体育博彩的边际收益通常很小——你可能要么亏损1%的投资回报率，要么盈利1%——那么任何细微优势都有帮助。有时候市场反应会滞后，如果你在现场（这叫临场观察），就可能捕捉到某些信息。所以确实有用。

I mean, given that the edges are pretty small in general in sports betting, you're probably either losing by 1% ROI or winning by 1% ROI, then any of that is helpful, I think. And also sometimes, you know, the markets are a little bit behind. If you're kind of there in the arena, it's called court signing, you can potentially pick up on things. So it is helpful? I think so.

没错。哪怕只能将投资回报率提高0.5%，在体育博彩中也意义重大。排序复选投票制真的重要吗？它真能改变选举结果吗？潜在影响是存在的。

Yeah. I mean, again, if it improves you by like oh point 5%, your ROI, that becomes pretty meaningful in the context of a sports bet. Does ranked choice voting really matter? Does it really matter in affecting outcomes? It potentially can matter.

特别是我们看到候选人为此调整策略，比如互相背书。不过我现在对排序复选制有些保留意见，原因有二：首先它不符合孔多塞最优原则——候选人被淘汰的顺序会影响最终胜者。

I mean, particularly, you've seen candidates adopt different strategies for it by cross endorsing, for example. Although, I think I've kind of turned mildly anti RCV for a couple of reasons. Why? One is that it's not Condorcet optimal, right? The order in which candidates are eliminated.

希望我没念错这个术语吧？比如纽约市长选举中，布拉德·兰德算是曼达尼和科莫之间的折衷人选。按理说他在任何一对一较量中都能胜出，但他却倒数第三个（或者说倒数第二个）被淘汰了。

Hopefully, I'm saying that name correctly, right? Yeah. The order in which candidates are eliminated can affect who wins. Brad Lander, for example, is kind of the compromise candidate between Mamdani and Cuomo in the New York mayoral race. It's plausible he would have won any head to head matchup, but he was eliminated third from last, or I guess second from last, so he didn't

但常规投票规则下他会有单挑对决的机会吗？

get the But would he have had a head to head matchup under a normal voting rule?

嗯，我是说，你可能会...我不知道。或者我猜传统的决选方式可能也不会更好。不，我是说，有道理。你也差点让曼达尼在第一轮投票中胜出。

Well, I mean, it's plausible you could I I don't know. Or I guess traditional runoff wouldn't be any better potentially. No. I mean, fair enough. You also almost had Mamdani wind up winning the first choice vote.

尽管在2021年，埃里克·亚当斯在倒数第二轮遥遥领先的情况下差点输给凯瑟琳·加西亚，而加西亚在前一轮几乎被玛雅·威利淘汰。对吧？所以这比想象中更容易被操纵。而且，计票和统计所有票数要花很长时间。就在昨天——我们录制这段时是7月21日？

Although in 2021, Eric Adams nearly lost to Catherine Garcia despite being way ahead in the penultimate round, and Garcia nearly was eliminated by Maya Wiley in the round before that. Right? So it's more gameable than you might think. Also, takes a really long time to count the votes and all the tallies. Just yesterday, we're taping this, what is it, July 21?

所以选举结束三四周后，你才能最终得到所有选票转移的官方统计。结果并没有那么接近，但我认为我们需要在美国选举中更普遍地在24到48小时内完成计票。

So three or four weeks after the election, you finally get the official tally of how all the ballots transferred. It wound up not being that close, but I think we need it more of a norm in American elections toward counting votes within twenty four to forty eight hours.

巴西就是这么做的。对吧？

Brazil does it. Right?

印度也是。我之前在研究印度选举，那在不同方面都很吸引人。他们在喜马拉雅山区设有实体投票站之类的。我不知道他们使用多少种语言，那个国家在各方面有多么多元——但他们计票非常快。其他国家的出口民调也做得更好，在美国，出口民调机构试图做这种包含100个问题的、类似政治科学的、按人口统计加权的调查。

India does. I was doing a little work on the Indian election, which was which was fascinating in different ways. And they, you know, they have a physical polling place up in the Himalayas and things like that. I don't know how many languages they speak and how diverse and every respect that country is, and and they count their results very quickly. They also have better exit polling in international countries where in The US, the exit pollsters are trying to ask this like 100 item long proper kind of political science like demographically weighted survey.

对吧？而在欧洲和拉丁美洲这些地方，他们只想知道谁会赢，为了让电视节目更精彩。所以只问一个问题：你投给了谁？他们以非常有组织的方式进行。所以国外的出口民调相当准确，其他地方则很糟糕。

Right? Whereas in Europe and Latin America, places like that, it's like, just wanna know who's gonna win, right, to make for more exciting TV. So just ask them one question, who do vote for? They do it in a very organized way. And so exit polling is quite good in foreign countries and quite bad elsewhere.

但没错，我认为...共和党人不应该把‘我们希望快速计票而不是拖上数日数周’当成他们的论点。

But yeah, I think, you know, I don't think there should be a Republican talking point that we want the vote counted quickly and not taking days and weeks.

如果你能挥动魔法棒，将比例代表制引入美国国家层面的选举，你会这么做吗？

And if you could wave a magic wand and bring proportional representation to The United States at the national level, would you do it?

我想我...作为对两大党派足够多的批评者，我大体上必须说是。不过，我不知道。我是说，美国制度中有更多的否决点，这可能对资本主义、经济增长以及某种自由是有利的，对吧？

I think I I'm enough of a critic of the two major parties that I'd have to say yes for the most part. Although, I don't know. I mean, you know, we have more veto points in the American system. That's probably been good for capitalism and economic growth and and maybe a certain type of liberty, I suppose. Right?

我是说，我更希望看到更多探讨的问题是——也许有些好书在讨论——我看到美国和欧洲在经济上越来越分化。欧洲的寿命越来越长，我们却没有。我们仍有不错的经济增长，而他们没有...这其中有多少是...

I mean, I think a question I'd like to see addressed more, maybe there's some good books, is like, I see The US and Europe as diverging more and more. Economically, Europe is getting longer and longer life spans. We're not really right. We still have fairly good economic growth, and they don't really how much of that is kind

关于文化？但我们的欧洲人寿命很长。对吧？嗯，这正是我说的。或者如果你是新泽西出生的日裔美国女性，那不就是世界上最长寿的吗？

of culture? But our Europeans have long life spans. Right? Well, that's what I'm saying. Or if you're a Japanese American woman born in New Jersey, isn't that the world's longest life span?

我是说，新英格兰几乎像斯堪的纳维亚之类的地方，对吧？包括在某些方面略显乏味，对吧？食物有点平淡。但没错，我是新英格兰的大力支持者。虽然我不想住在那里，但新英格兰确实有其独特之处。

Mean, New England is almost like Scandinavia or something, right? Including being slightly boring in certain ways, right? The food's a little plain. But yeah, I'm a big proponent of New England. It's not where I wanna live, but New England exceptionalism.

他们在疫情期间表现好得多。波士顿早期疫情很严重，对吧？佛蒙特、新罕布什尔、缅因州的病例远低于美国其他任何地方。他们显然做对了什么。

They did much better under COVID. Boston had a very big outbreak early, right? Vermont, New Hampshire, Maine had much lower COVID cases than anywhere else by far in The United States. They're doing something clearly right.

如果我看看一些预测市场，有时我看到特朗普有6%到7%的可能性成为下一位共和党总统候选人。这是否因为市场难以压缩而导致的长赔率偏差？你认为这个数字准确吗？还是只是表达性赌客想再次看到特朗普，愿意亏钱传递这个信息？你怎么

If I look at some of the prediction markets, at times, I've seen Trump at six to 7% likely to be the next Republican nominee to run for president. Is that long shot bias that it's hard to shorten those markets? Do you think it's the right number? Is it just expressive bettors who wanna see Trump again, and they're willing to lose some money to send that message? What do you

看这个问题？预测市场——我为Polymarket提供咨询——已经改善了很多。流动性大幅提升，但你仍能看到明显的错误定价，比如Zoran在Polymarket市场启动时居然有7%的概率成为民主党候选人，而他出生在乌干达。我认为这无法规避。在PAPL投注市场上，教皇利奥的赔率也没超过1%之类。

make of that? So the prediction markets, and I consult for Polymarket, have gotten quite a lot better. They've gotten quite a lot more more liquid, but but you still see some, you know, obvious mispricings like Zoran actually had, like, a 7% chance of becoming the Democratic nominee when that market launched at Polymarket and he was born in Uganda. So there's no workaround to that, I don't think. I think on on the PAPL betting markets, didn't have Pope Leo priced higher than 1% or things like that.

我认为他们对一个没有内幕消息的系统过度自信，外部观点可能仅 marginally有用。老实说，我不知道。我们之前聊过一些学术历史学家、政治学家如何沉迷于反特朗普情绪。特朗普竞选第三任期的情景——我不说概率为零。

And I think they got overconfident about a system where they had no inside knowledge and the outside view was maybe only marginally useful and things like that. Look, I don't know. Yeah. Mean, we talked a little bit before about how some of the academic historians, political scientists have gotten enraptured by this anti Trump thing. And I do think the scenario where Trump runs for a third term I mean, I don't say it's zero.

对吧？如果概率是2%，那仍然高得惊人，但很可能是2%而非6%或7%。我的问题是：在那个世界里，我们还会举行选举吗？如果你公然无视宪法到那种程度。

Right? And I'd say if if if it's 2%, then that's still alarmingly high, but it's probably probably two and not and not six or seven. I mean, you'd have to just totally my question is, in that world, are we even having election in the first place? Right? If you're just kind of flagrantly ignoring the constitution that much.

我是说，

I mean,

我认为很多奇怪的威权政府也会举行某种形式的选举。对吧？

I think there are It's common that weird nondemocratic governments have elections of some kind. Right?

甚至苏联也举行过所谓的“选举”。但美国要发生这种转变...况且特朗普并不特别受欢迎。他的支持率早期因恐怖主义（我认为是主因）快速下跌后又回升，现在又跌了。不知道多少是恐怖主义因素，还是像爱泼斯坦事件之类的影响。

Most of the even the Soviet Union had, quote unquote, elections. They they do, but, like, for that transition to occur in The US of and also, you know, Trump is not particularly popular. His ratings kind of fell pretty fast early because of I think predominantly terrorists and then recovered. Now they've fallen again. Don't know how much of that is more terrorist versus, you know, if I could like the Epstein stuff, for example.

是的，很多情境下，民粹主义者容易受到欢迎。美国确实有足够的制衡机制。人们不会被强迫从单一来源获取信息。所以，虽然我对特朗普有很多担忧，但这可能不是最首要的问题。

Yeah, a lot of the scenario mean, you know, helps for populists to be popular. We certainly have enough kind of checks and balances in The US. People are not being forced that information from, like, from any one source. And so, yeah, I worry about a lot of things with Trump. That's probably not the foremost concern.

我看到ESPN体育评论员斯蒂芬·史密斯——你可能也认识他——在民主党提名人的市场预测中高达9%。你觉得他实际的支持率有多少？

I've seen Stephen Smith, the ESPN sports commentator. You probably even know him. But as high as 9% in the market to be Democratic nominee. What do you

可能只有2%吧。

think his actual number is? Maybe 2%.

我认为是4%或5%。

I think it's four or 5%.

好吧。

Okay.

我觉得这不算离谱。

I don't think it's crazy.

我们曾列出20位最有可能的民主党提名人选，斯蒂芬·史密斯只是第九轮的选择。想象一下他获胜的情形？那只能是民主党人自我感觉良好的时候。我之所以没你那么乐观，是因为我认为会有其他更有影响力的人物出现。

We had a draft of the 20 most likely Democratic nominees. I Stephen Smith was a ninth round draft pick. Right. And what does that scenario look like where he wins? I think it's where Democrats feel really well, the reason why I would not be as optimistic as you are, right, is I think that other people with the platform will emerge.

说实话，一年前我们录制上次对话时，我都没听说过佐兰·马姆塔尼这个名字，现在他可能是党内知名度前五的人物了。虽然他来自乌干达无法参选总统，但会有其他人崭露头角。左翼可能会涌现更多能量，那些明显带有党派倾向的人——甚至可能有点偏执。

Right? I had not frankly heard of Zoran Mamtani's name when we taped our last conversation a year ago, when he might now be one of the five most recognizable names in the party. The fact that he's from Uganda prevents him running for president. You'll see other people emerge. I think there is probably some indication that you're going to have more energy on the left, more energy from people who are kind of explicitly partisan, maybe to the point of being a little bit paranoid even.

比如某些民主党派系也存在选举否认主义，认为特朗普从卡玛拉·哈里斯那里窃取了选举。虽然目前只是萌芽状态，但党派斗士正在筛选...人们会困惑：为什么民主党选民基础如此？因为库莫在纽约市全州选举中总是表现很好，

Right? You have you know, there is a little bit of election denialism among certain threads of Democrats, for example, that Trump stole the election from Kamala Harris. It's it's, you know, it's nascent for now, but but I think I think a partisan fighter And also, you're filtering out. I mean, people get confused because they're like, Okay, well, why is Democratic electorate? Because Cuomo did very well always in New York City when running statewide.

希拉里·克林顿在纽约市也很受欢迎，这某种程度上分流了2020年伯尼·桑德斯的选票。但这是因为中间派已逐渐远离民主党，所以这两种现象同时发生。我确实认为民主党可能高估了全国对进步治理的容忍度。不过话说回来，我认为J.D....

Hillary Clinton did very well in New York City. It kind of diverted Bernie Sanders' campaign in 2020. But it's because people in the center have migrated away from the Democratic Party, so you're having these two things that go on at once. I do think there's a chance that Democrats will overestimate how tolerant the rest of the country is for progressive governance, But with that said, I think J. D.

万斯也是个相当有缺陷的候选人。我认为任何非万斯的候选人都很难摆脱特朗普的阴影。因此，我认为民主党在2028年有55%对45%的胜算，这考虑了他们可能会提名一个从选举角度看我认为是次优的候选人。明白吗？我并不是在告诉人们应该根据谁的价值观来提名谁。

Vance is also a pretty flawed candidate. I think any other non Vance candidate is gonna have a lot of trouble escaping the shadow of Trump. So I regard Democrats as like a 55, 45 favorite for 2028, pricing in some chance they'll nominate what I would think of as a suboptimal from an electoral standpoint candidate. Right? I'm not telling anybody who they should nominate in terms of whose values they have.

你知道，我认为人们忽视了初选制度的重要性，以及选民的选择权和显示出的偏好。初选活动是一个相当好的代理，可以模拟大选竞选，因为它覆盖了所有50个州，你要筹款、发表演讲等等。但确实，我认为那种认为外部救世主是个中间派的观点，比我一年前所想的可能性要低一些。

You know, I think people neglect the importance of the primary system and the optionality and the revealed preferences of voters. It is a pretty good proxy for running a general election campaign, is running a primary campaign that goes through all 50 states, and you're raising money, you're giving speeches, things like that. But yeah, I think the idea of external savior who's a centrist is becoming a bit less likely than I would have thought a year ago.

你认为大型科技公司是否拥有专有信息，能比任何人或任何东西更好地预测选举？他们可能甚至懒得去做，但原则上是否存在这种可能性？我倾向于怀疑这一点。我是说，

Do you think the major tech companies have proprietary information where they can predict elections better than anyone or anything. They may not even bother to do it, but is it there in principle? I tend to doubt it. I mean,

民调有一个好处，尽管它有很多缺陷，而且我认为民调正处于——我不会称之为危机，但正在走向危机的路上——就是你有一个很好的基线可以比较。对吧？你可以说，我们要回到1936年，我想那是盖洛普第一次发布总统选举民调的时候，然后从经验上看这些民调在选举前x天的准确性如何，而AI类型的模型则没有任何校准的依据。对吧？我的意思是，你或许可以预测一个选民会如何投票。

one nice thing about polling for all the flaws that polling has, and I think polling has I wouldn't call it a crisis, but on the route to a crisis, is you do have a good baseline to compare with. Right? You can say, we are gonna go back to 1936, I think, was the first time Gallup issued a presidential election poll and look empirically at how accurate those polls are at x number of days out from the election, whereas there's nothing to kind of calibrate AI type models. Right? I mean, you could probably predict how an individual voter will vote.

所以如果他们收集——我是说，我告诉他们的东西很多。但如果你花了

So if they're gathering I mean, the amount of things I tell But if you spent a

很多钱，并且你拥有每个人的，比如说，Facebook帖子，然后你把它输入到一个足够复杂的模型中，

lot of money and you have everyone's, say, Facebook posts, and you just fed it into a sufficiently complicated model,

那不就是最好的预测模型吗？嗯，如果我只知道你的名字和邮编，我可能就能以90%的准确率预测你会投票给谁。从名字可以推断性别、种族、年龄。从邮编可以推断社会经济地位。这些信息已经能让你走得很远了。

wouldn't that be the best predictive model? Well, if I know just your name and your zip code, I can probably predict with 90% accuracy who you're gonna vote for. From the name, can infer gender, race, age. From the zip code, you can infer socioeconomic status. That gets you pretty darn far.

选举的问题在于它们是在1%或2%的边际上运作的。民调会出大问题，因为一个候选人得到51%而不是47%。这仍然相当不错。如果你在估计一个罐子里有多少颗软糖，你说‘哦，51%是紫色的’，而实际上是47%，那也不算太糟，对吧？但选举民调需要很高的精确度，你有许多方法可以得到一个粗略的近似观点，但我认为这些方法不会比民调等其他方法更优越。

The issue with elections is that they operate on this 1% or 2% margin. The polls will have a disaster because one candidate gets 51% instead of 47%. That's still pretty good. If you're estimating how many jelly beans there are a jar, and you're like, Oh, 51% of them are purple, and it's really 47, then that's not too bad, right? But you need a lot of precision for election polling, you have plenty of ways to get to a kind of blunt approximate view, but I think those would not be any superior to the other ways, polling namely.

你对最可能的民主党提名人有预测吗？我们可能

Do you have a prediction for most likely Democratic party nominee? We may

会再做一次这样的讨论。

be doing another of these in

三年了，对吧？AOC是我在模拟选秀中的首选，我认为纽约梅里尔选区的表现会强化这一点。我是说，她是个非常聪明的人，在某些领域展现出了策略性。听着，重申一下，我不是社会主义者或民主社会主义者，对吧？但即便是我也喜欢看到政党中出现新面孔，对吧？

three years, right? AOC was the first pick in the draft I did, and I think that would be reinforced by the New York Merrill race. I mean, I think she's a very bright person who has demonstrated some tact in in certain areas. And look, again, I am not a socialist or Democratic socialist, right? But even I am someone who likes seeing new faces in the parties, right?

那些对克林顿王朝或总爱抢占风头的科莫王朝、拜登家族毫无好感的人。所以我认为她足够聪明、足够新颖，在社交媒体上也足够出色。你看，虽然目前没人支持率超过15%，但如果让我免费下注，我会选她。

Who thinks who had no affection for like the Clinton dynasty or the Cuomo dynasty Bidens who can't seem to stay out of the spotlight. And so I think she's smart enough and fresh enough, good enough on social media. So, you know, no one's higher than 15%, but if you gave me a free bet, I think I'd pick her.

说到预测，一年前我们讨论过AI需要多久才能媲美人类超级预测者，当时你预测至少需要十到十五年。现在一年过去了，你想重新审视并修正这个预测吗？

Speaking of predictions, a year ago we talked about how long will it take AIs to be as good as human super forecasters, and you made a prediction where you said at least ten to fifteen years. Yeah. Now a year later, do you want to revisit that and revise?

我认为原预测基本正确。相比一年前，AI的进展大约在我预期的第40个百分位。我很好奇——你会怎么想？

I think it's probably about right. I mean, I I would say relative to a year ago, AI is about at the fortieth percentile of progress I would have expected. I'm curious I'd be curious what you would think. Right?

一年前我说两到三年。现在我要说一到两年——这相当于同样的预测区间。所以我觉得你的时间表太悲观了。

Well, a year ago, I said two to three years. Yeah. And right now, I'm gonna say one to two years, which is the same prediction. So I think you're way too pessimistic in your timetable.

这取决于竞赛的性质，对吧？

It depends on kind of how competitive the exercise is. Right?

如果是像数学奥林匹克这样的比赛。我知道它们刚达到金牌水平——我去年就说过，一年内它们能拿金牌。而一年前它们连'strawberry'这个词有几个'r'都搞不清楚。

If it's like a Math Olympiad tournament. I knew they just did gold medal performance. I said this last year. Said in a year, they're gonna do gold medal. And a year ago, they weren't sure how many r's were in the word strawberry.

所以你不认为在超级预测领域它们能...

So you don't think on super forecasting they can

处理静态问题和动态系统完全不同——后者输入参数时刻在变化。目前AI实验室...不对，是大语言模型在扑克方面非常糟糕。它们没受过扑克数据训练。当然如果专门训练，有些叫求解器的程序能做得很好。但它们无法仅从平庸的文本数据或业余的手牌分析中推导通用模式。

I think it's very different when you're dealing with a static problem as compared to a dynamic system where the inputs are changing all the time? So currently, AI labs, or not the AI labs, but the large language models are very bad, very, very bad at poker. They're not trained on poker data. I'm sure if you did train them, there are these things called solvers that are trained on poker data that do very well, right? But they cannot quite impute the general patterns just from mediocre text data or amateurish kind of hand analysis.

但若追问它们表现差的原因，它们会说：'是啊，面对复杂的演化博弈论动态，需要快速制定剥削策略时，我们确实吃力。'顶尖牌手能在几秒内完成隐含计算，而计算机求解最小化损失的牌局可能需要高性能电脑运算数分钟。数学奥林匹克方面我有点担心应试训练——因为设定这个目标后，实现目标就会带来巨大声望。

But if you probe them on why they're bad, they're like, yeah, maybe it's tough for us when you have a complicated evolving kind of game theory dynamic, and you have to develop exploitive strategies very quickly. I mean, the amount of like it is amazing poker players where if you have a computer solve a poker hand to get to one where there's enough loss minimization, it can take a very strong computer can take minutes, whereas poker player makes those calculations implicitly in a handful of seconds, for example. I worry with the Math Olympiad stuff, there's a little bit of teaching to the test where because you set this as the goal that a large language model should have, that therefore there's a lot of prestige when you meet that goal,

但某种程度上，针对考试教学不正是我们应该做的吗？即便是对人类而言也是如此，对吧？考试内容代表你认为重要的东西，那正是你该教授的。不过，就像

But isn't teaching to the test what we should do, even with humans in a sense, right? The test is what you think is important, and that's what you ought to teach. Well, but like

扑克或象棋的例子。据我所知，AI模型在象棋方面表现非常糟糕。

the poker example, right, or chess. I think AI models are very poor at chess from everything that I've heard, for example.

但其他AI模型会下象棋啊。没错。不，你看，这涉及到

But other AI models play chess. Great. Correct. No, look, this gets to

我对通用智能定义的疑问。我们可能会建立多层模型架构，逐步修补不同功能。比如现在ChatGPT还不能有效预订机票，但若专门开发代理模型——这类具有自主行动能力的模型正逐渐融入系统——很快就能实现

me the question of what's it mean to be generally intelligent? I mean, we'll probably have scaffolding of model on top of model, and you'll now patch different things. Right now, you can't really very effectively make a plane reservation using ChatGPT, but I'm sure if you dedicate a resource to that, then you have these agentic models now, agent models are just creeping into the system a little bit. But those will work

我觉得用不了一年。现在已经有专门的代理模型实验室

in less than a year, I think. Mean, there's an agentic model lab

这让我思考粗糙AGI与超级智能的区别。相比从前，我现在更不认为会出现智能爆炸。其实我从未真正相信过这种说法。就是那种通过简单数据训练就能超越数据集范畴的涌现型超级智能。

where I get over to like rough AGI versus super intelligence. I am less convinced that we're gonna have some intelligence explosion than I would have been maybe. I don't think I was ever convinced of it. Right? But like, you know, this emergent super intelligence where you train it on relatively simple data and extrapolates beyond the dataset.

它们确实具备推理能力，有时相当聪明。我现在会直接说'ChatGPT认为如何'，不再回避'思考'这个词。但桌面工作的近似通用智能，与超级智能或体力劳动的AGI之间存在巨大鸿沟。人们过早做出跨越性假设，比如数学奥林匹克竞赛——可能答案本就隐含在训练数据中。

I mean, I think they do reason sometimes. Think they're kind of quite smart, and I no longer am bashful about saying, Oh, Chatchipity thinks this. He used to avoid that term think. But there's a big gap between approximate general intelligence for desk jobs, and then super intelligence on the one hand, or then AGI for physical labor on the other hand. I think people are much too quick to make that leap, and I think the Math Olympiad in part because maybe the answers are somewhere latent in the training data.

即便不是如此，当你尝试解决卢卡斯批判之类的问题时——这是个很好的数学法则——我认为AI模型也需要遵循类似原则。

But even if they're not, if you try to solve what's the Lucas critique or whatever else? It's a good math law. A version of that, I think, for AI models.

我的直觉是，如果让五位超级预测者为今年夏天将发布的GPT-5撰写五页提示词，我们可能已经达到目标了。这算不上超级智能，你也可以说不是AGI，但人类超级预测者本身并不惊人——他们不是爱因斯坦。

My intuition is that if you took five super forecasters and just had them write a five page prompt for GPT-five, which will be out this summer, that we'd be there already. I don't think it would be superintelligence. You could say it's not AGI, but the human super forecasters, they're not that impressive. Right? They're not Einsteins.

他们只是方法得当且训练有素。

They just have good methods and they're disciplined.

这需要大量的常识积累。我认为既需要理解世界运行规律的普遍知识，也需要启发式思维。可能难度会稍大一些。我是说，你能训练思维方式吗？我有时在现实生活中尝试这样做，对吧？

It requires so much on general knowledge. I think both general knowledge of how the world works and heuristics. I think it might be a little tougher. Mean, can you train on a thought process? I try to do that in real life sometimes, right?

我的伴侣是艺术家，对艺术有很好的鉴赏力，而我没有。我大概只有B或B+的水平？我可以向他学习，观察他喜欢什么样的艺术作品。但这样可能最多让我达到B+的理解程度。

My partner is an artist and has a very good eye for art, and I don't. I maybe have a B, B plus, right? I can train myself on him and say, what does he like in art? And so you know? But that probably caps me at like a B plus understanding potentially.

而在金融市场中，B+级别的交易员往往处于交易的亏损方。

And in markets, a B plus trader is often on the losing side of a trade.

你觉得自己在艺术鉴赏方面能达到多高水平？假设你投入两到三年时间，虽然不是全职，但把它作为篮球之外的两三件重要事情之一。我...

How good do you think you could become at appreciating art? Let's say you put in two to three years, not full time, but made it one of your two or three big things after basketball. I'd

可能会非常认真地学习。听着，我有信心当我全心投入某件事时，能做得相当好。我认为这方面没有生理限制，但问题是，我现在有四五个占据全部时间的爱好，其中有些已经变成职业了。或许建筑更适合我。在Silver家族的DNA里，确实有些建筑天赋的基因。

probably be very studious about it. Look, I have confidence that when I devote myself to things, can get quite good at them. I think there's no physical limitation there, but still, mean, I have four or five hobbies I occupy all my time, some which have already become professions. I think maybe architecture more. In the silver DNA, there are some architecture genes.

我们与弗兰克·劳埃德·赖特没有血缘关系。我祖母是建筑师，设计了我现在在西切斯特郡偶尔居住的房子。我叔叔也是建筑师。所以这更像是我会发展的方向。我喜欢空间感强的事物。

We're related not by blood to Frank Lloyd Wright. My grandma was an architect, designed the home I live in part time now in Westchester County. My uncle is an architect. So that would kind of be more the direction where I would go out. I like things that are spatial.

我有非常好的空间记忆力。如果我去过某家餐厅，即使是十年前完全不同的餐厅，我也能清晰记得那个空间布局。而艺术方面，我的鉴赏力只能说一般。

Have a very good spatial memory. If I've been to a restaurant before, even if it was a different restaurant ten years ago, I remember that space very well. Whereas art, I have just an okay eye for it.

九年前我们第一次对话时，讨论过为什么没有更多职业运动员公开出柜。当时我们都认为在未来一段时间内，会有更多NBA球员公开同性恋身份。我问了ChatGPT，它说杰森·科林斯十二年前出柜后，至今再无人跟进。

Now nine years ago, in our first conversation, we discussed why more professional athletes had not come out as gay. And you and I both thought then that over the forthcoming time period, more players from the NBA would come out as gay. Yeah. I asked ChatGPT. They said Jason Collins came out twelve years ago, and since then, it's been zero.

为什么会这样？我们哪里判断错了？

Why is that? What did we get wrong?

我觉得体育界出现了些保守主义回潮，可以看到更多直言不讳的特朗普支持者。但这种情况...

I mean, there's been a little bit more of a conservative backlash in sports where you see kind of more outspoken Trump supporters. But that's

就在最近，对吧，有一个九年周期。2020年就是其中一部分。极端‘觉醒’现象属于这个时期。同性婚姻在美国社会已经相当稳固，而且

very recent, right, where you have a nine year period. There's 2020. The extremes of woke are part of this period. Gay marriage is really quite solidified in American life, and

没有人公开出柜。从很多方面看这挺让人惊讶的。听着，我并不是那么政治正确。我的意思是可能存在某些选择机制。我不确定这些机制是否与基因有关，但也许这让同性恋者在高中时期更容易出柜。

no one comes out. It is surprising, I think, a lot of way. Look, I am not so PC. I mean, there may be some selection mechanisms. I don't necessarily know that they're genetic, but maybe it becomes easy for gay people to come out in high school and things like that.

他们被体育之外的领域吸引，尽管男同性恋者通常比同年龄段异性恋男性身材更好、更健壮。对吧？所以这不是身体能力问题，而是对体育的选择倾向。随着体育越来越职业化——高中阶段可能用这个词不太准确——但你会被分流到库珀旗之类的项目。他并非整个高中都在缅因州度过。

They're attracted to fields apart from sports even though gay men typically are quite in shape and quite athletic relative to straight men of the same demographic cohort otherwise. Right? So it's not like a physical ability thing, but selecting into sports. And as sports becomes more professional, I guess not the right term at the high school level, but you're tracked into Cooper Flag, whatever. He didn't spend all his high school in Maine.

他去了蒙特沃德学院之类的学校。所以任何让你分心的事，无论是关于身份认同的疑问还是其他什么，都会让你偏离轨道，后期可能更难弥补。GPT还声称WNBA有44名球员公开了非传统异性恋身份。

He went to Monteverde Academy or whatever else it was. And so if anything distracts you from that, whether it's questions about your identity or anything else, you get off track, maybe it's harder to make that up later on. GPT also claims that 44 players in the WNBA have come out as something other than standard hetero.

是啊。差距很大对吧？

Yeah. It's a big gap, right?

差距确实很大。去年我看过凯特琳·克拉克在印第安纳狂热队对阵纽约自由人队的比赛。WNBA非常有意思。那是我参加过的人口结构最独特的体育赛事。

It's a big gap for sure, right? And I saw Caitlin Clark in Indiana Fever against New York Liberty last year. The WNBA is fascinating. Right? It's a very, very, very different demographic than any sporting event I've ever been at before.

我觉得这很棒。这项运动正在获得更多文化关注度。

I think it's kind of great. It's getting more cultural salience, I think.

博彩业对它的相对发展有帮助吗？

Does betting help it in relative terms?

两年前我在女子NCAA篮球锦标赛下注赚了不少，当时市场还没调整过来。但后来人们发现这些赔率设置得很宽松。在任何博彩领域，只要你表现好，连续盈利后就会被限制投注——甚至当你的期望值为正时，赌场也会很快限制你。不过现在可能有些聪明的大学小子能靠女子体育博彩大赚

Two years ago, had a really good time betting the women's NCAA basketball tournament before the markets had adapted, but I think people figured out, okay, these are pretty soft lines. So, yeah, no. I mean, the issue is in any type of betting, if you're any good, you'll get limited if you have repeated success, or even waitress at Look EV plus EV to the casino, you'll also get limited quite quickly. But yeah, probably betting women's sports probably is there's probably some smart kid out there, some smart college kid who's gonna make a lot

一笔

of money

赌WNBA、女子大学篮球，实际上任何女子比赛都赌。

betting WNBA, women's college basketball, women's anything really.

他是在建立模型还是仅凭直觉，或者在网上阅读

And he's building a model or just using intuition or reading a

大量资料？他可能怎么做呢？我猜是模型结合对联盟的了解能让你走得很远。比如伤病数据，女子运动比男子运动差很多。她们没有相同的报告规范。

lot on the internet? How might he be doing that? My guess is the model combined with knowledge of the league can get you pretty far. So injury data, for example, is much worse for women's sports than for men's sports. They don't have the same reporting protocols.

甚至NCAA锦标赛，我们实际上试图在女子NCAA模型中使用伤病数据，你必须一个一个地通过谷歌搜索来获取，对吧？所以，掌握别人没有或很少有人拥有的知识仍然有回报。人们基本上很懒，不一定愿意做这些基础工作。你之前提到库珀·弗拉格。如你所知，他被达拉斯小牛队以第一顺位选中。

Even for the NCAA tournament, we actually were trying to use injury data in our women's NCAA model, and you have to do one by one individual Google searches for it, right? And so, there's still a reward to having knowledge that nobody else has, or that few other people have. And people are basically pretty lazy and don't necessarily do the legwork on that. You mentioned Cooper Flagg before. As you know, he was drafted number one by the Dallas Mavericks Yeah.

他们相当

Who are quite

不错的球队。休斯顿和费城分别以第二和第三顺位选秀，奇才队是第六顺位？

a good team. And you have, what, Houston and Philly drafting two and three, and the Wizards draft number six?

是的。我们需要重新设计NBA选秀抽签吗？之前人们抱怨说摆烂太多，现在他们又说抽签对摆烂球队奖励不够。不，我认为现在可能已经相当优化了。

Yeah. Do we need to redesign the NBA draft lottery? People complained before that there was, like, too much tanking, and now I guess they're saying that it doesn't reward tanking teams enough. No. I think it's probably it's probably fairly optimal.

我是说，今年有点奇怪，前三名的球队在某些方面都很成功，对吧？或者有其他优势。但我其实挺喜欢迈克·扎里安之前提出的那个‘轮盘’提案——每支球队每三十年获得一次状元签，而且你提前知道什么时候轮到你。

I mean, this year was was weird in the sense that, like, all the top three teams are accomplished in certain ways. Right? Or other things going for them. But no. I, you know, I, you know, I kind of like the old was it the Mike Zarian proposal for, like, the wheel where every team gets the first pick once every thirty years, and you kind of know ahead of time which one it is.

对。我...我不喜欢在任何情况下鼓励失败，我也不认为球迷会介意支持一个联赛里有一支强队和一支永远的弱旅。世界上最受欢迎的运动当然是足球。在那里，有些球队几乎每年都能赢，有些情况下持续了近一个世纪。

Yeah. I I I don't like I don't like incentivizing failure in any context, and I don't think fans mind like rooting for leagues where one team is great and the other is a perpetual underdog. Right? The most popular sport in the world, of course, is is soccer. And there, you know, you have teams that are are winning teams every year for a century almost in some cases.

人们似乎并不太在意。很奇怪的是，美国体育（这是个老生常谈）有点社会主义色彩，而欧洲体育则更资本主义。

It doesn't seem to bother people as much. It's it's very weird how American sports, it's a cliche, are kind of socialist, and European sports are more capitalist, basically.

NBA常规赛的赛制设计是否存在危机？

Is there a crisis in NBA regular season design?

我原本对此持怀疑态度，直到观看了今年的季后赛，发现比赛节奏更快、观赏性更强，从头到尾都令人愉悦。听着，我认为这是个动态目标。如果把赛季缩短到66场或72场，人们依然会抱怨场次太多。我曾住在麦迪逊广场花园附近，你知道，普通尼克斯比赛的入场价可能高达200美元，所以我理解他们为何不愿改变。不过，我觉得或许可以尝试66场或72场左右的赛制，看看效果如何。

I had been kind of a naysayer of this a little bit until I watched this year's playoffs and kind of how much higher octane they were and how enjoyable it was to watch a game from from start to finish. Look, I think it's a moving target. If you shorten the season to 66 games or 72, people will still complain about about that being too many. And I used to live near Madison Square Garden, and they're you know, for regular Knicks game, they're the get in price can be $200, so I understand why they might not want to change it. But, yeah, I think you probably wanna go to maybe not 66 or maybe 72 or something like that and see how that goes.

今年出现了太多决定性伤病。我记不清完整名单了，但似乎近半数重要球员，包括第七场总决赛，实质上都是由伤病决定的。

There were so many decisive injuries this year. I forget the whole list, but it seems like almost half the important players, including the seventh game of

看看其他运动——橄榄球比赛中，你大约只参与半数的进攻回合；冰球里除了守门员，球员们分三到四组轮换上场；

the finals, were decided in essence by injury. Well, look, in every other sport, right, in football, you're only playing half the downs roughly. Right? Right. In hockey, apart from the goalie, you're playing you know, there are three or four lines at shift.

棒球比赛时，你只有一半时间在防守。而NBA篮球持续高强度对抗——这完全是不同的运动。回看七八十年代的比赛录像，你会感叹这根本是另一种运动。

Right? In baseball, you're only in the field half the time. The continuous strain on NBA basketball played at such a high I mean, it's totally different sport. Right? You go back and like watch clips from like the seventies or the eighties, and you're like, it's a totally different sport.

我认为现在这项运动好多了，但我也说不准。看到这些球员如此出色真让人沮丧，比如杰森·塔图姆就是个非常优雅的球员。哈利伯顿动作虽然有点不协调，但步伐间自有一种优雅。我觉得联盟必须...我是说，你在棒球界也见过类似情况，比如汤米·约翰手术之类的。

I think it's a much better sport now, but like, I don't know. I mean, it sucks to see these guys like, you know, so great. Jason Tatum is a very graceful player. Halliburton is a little herky jerky, but has a certain grace in how he steps. I'm like, yeah, I think the league has to I mean, you've seen it in baseball too with like Tommy John surgeries and things like that.

而且，你知道，将事物优化到我们未曾预料的程度并推向极限总会带来副作用。现在的平衡点可能是球员们总在挑战极限，对吧？他们竞争心强，报酬极高，对教练和训练师的高强度要求也很合理。短期激励也是个因素。

And and, you know, there are always downsides to achieving optimization to a degree we didn't necessarily expect and and pushing things to the brink. Now, the equilibrium probably is that players always are pushing things to the brink, right? They're very competitive, they're extremely well compensated, right? There is appropriately demand for coaches and trainers and everything else for high effort, I think. You also have short term incentive.

理论上，球队不会太关心合同即将到期的明星球员会怎样。

Teams don't really care that much about what happens to their star player if he's on the end of his contract in theory.

他们可能还乐见其成。

They might be happy.

没错。但我觉得必须...或许应该缩减几场比赛，或者设定规则比如单个球员最多打75场。每位球员保证至少7天休息时间？可能还得提前公布这些安排。

Yeah. But I think you have to have I think you probably have to trim a few games off schedule or have something where maybe you say, okay, the individual player limit is 75 games. Right? Every player is guaranteed at least seven days rest. Maybe you are required to publish them in advance.

所以如果我要去看扬尼斯为密尔沃基打球，我对没有扬尼斯的雄鹿队毫无兴趣。

So if I'm going to see Giannis play for Milwaukee, have no interest

以他们目前的状态，没有扬尼斯的雄鹿队确实如此。比如说，这可能是一个中间步骤。假设你在给勒布朗·詹姆斯提建议，而接下来这一年可能是他的最后一年，或者肯定是倒数第二年。湖人队在接下来的一年里实力不足以真正走得更远。你可能也同意这一点。

in the Bucks without Giannis in their current form. For example, maybe something like that could be an intermediate step. Let's say you're advising LeBron James, and this coming year possibly is his last or certainly his next to last. The Lakers are not good enough to really go anywhere in the next year. Probably you agree with that.

但他签的合同是多少？大约5000万美元？是的。你会给勒布朗什么建议？作为一个普通人，勒布朗，你应该最大化什么？

But he has a signed contract for what? About $50,000,000? Yeah. What do you advise LeBron? Just as a human being, LeBron, what should you maximize?

你会

What do

怎么告诉他？勒布朗现在处于一个有趣的位置，因为从职业生涯的指标来看，他现在显然是NBA历史上最好的球员。我同意。但我不认为他会赢得人们的心。我的意思是，你可以说勒布朗在巅峰时期不如乔丹。

you tell him? LeBron's in a funny place because I think he's clearly, by the career metrics, now the best NBA player of all time. I agree. I don't think he'll ever win the hearts of people. I mean, you could argue LeBron was not as good as Jordan at the peak.

对吧？你也可以说卡里姆的巅峰更高，也许是在他职业生涯的早期。所以勒布朗几乎没有什么可以做的，除非再赢三个总冠军，才能让他在与乔丹的讨论中领先。我会想去一支年轻的球队，既能赢得总冠军，又能指导一些年轻人。比如，去马刺队。

Right? You could argue Kareem's peak was also higher, maybe very early in his career. So there's kind of nothing LeBron could do short of winning three more titles that would kind of put him ahead of Jordan in that conversation. I'd wanna go to a young team where I can win a title and mentor some guys. I mean, go to the Spurs, for example.

那可能会很有趣。对吧？

That might be a lot of fun. Right?

如果他们今年无法和勒布朗一起赢得总冠军。哦，我

If they're not gonna win a title with LeBron, not this year. Oh, I

认为马刺队今年如果有勒布朗·詹姆斯，可能会赢得总冠军。真的吗？

think the Spurs with LeBron James this year could potentially win a title. Really?

那会让我大吃一惊。

That would shock me.

我会给这个概率，大概4%或5%吧。文班亚马确实非常非常出色。当时他打出的水平，在Darko NBA实时追踪表现的指标中——这个指标对30岁正值生涯中期的球员通常意义不大，但对维克多这样的新秀却很能说明问题——他赛季报销前可能已经是联盟第五强的球员了。

I I would give that, I don't know, four or 5%. I mean, Wemby is very, very good. I mean, he was playing at a level where there's this Darko NBA metric that kind of tracks performance in real time, which is usually not irrelevant for like a 30 year old mid career player, but for like someone like Victor is relevant. He was probably the fifth best NBA player by the time he kind of shut down his season.

但他坚持打完整个赛季的概率呢？这么年轻的球员，看看历史上的真正巨星：贾巴尔很早就有成就，但多数人需要时间，比如迈克尔·乔丹就花了几年。

But the chance that he's still on the court by the end of the year, and as such a young player, you look at the other true greats over time. Kareem won something very early, but usually people need you know, Michael Jordan needed some years.

勒布朗...勒布朗当年也需要...而且他们现在...

LeBron LeBron needed would needed be and, you know, and and they have little

让我们看看。我敢打赌，不过我更想押他们输。当然，可能要等到明年？对吧？

let's see. I bet, but I'd love to bet against them. Well, well, next year maybe, though. Right?

他们还有一年成长时间。到时候迪伦·哈珀可能成为正资产，而勒布朗可以在圣安东尼奥签个底薪合同之类的。

They have one more year of maturity. They probably feel maybe Dylan Harper's a positive value asset at that point, and LeBron can have his, like, you know, sign a minimum contract in San Antonio.

反正我们都知道这是你的建议。你怎么看这个观点：每个赛季只有几支球队能夺冠，而且你提前就知道是谁？比如去年可能是波士顿、俄克拉荷马，或许还有丹佛，其他队都是冷门。你觉得呢？

Anyway, we know that's your advice. What do you think of the hypothesis that in a given season, there's only a few teams that can win it all, and you know in advance who they are? So maybe last year, it was Boston, Oklahoma, possibly Denver, and you should just bet on those teams. Anything else is a true long shot. What do you think?

听着，每个篮球迷都相信某种理想化的版本，但我认为现在有点夸大其词了。很多球队——比如掘金，在约基奇被视为季后赛大杀器之前，人们根本不信他们能在季后赛走远。猛龙几年前也不符合这个模式。

Look. I think every basketball fan believes in some, like, stylized version of that, but I think it's become, like, a little bit overstated. Many of the teams now, you know, the nuggets, people until Jokic was seen as, like, this ultimate beast of a postseason player, people were very skeptical about their ability to thrive in the postseason. Right? You know, Toronto didn't fit the paradigm a couple of years ago.

就连字母哥在成功前也经历过季后赛挣扎。NBA有参考价值的历史并不长，可能人们过于关注公牛和勇士的王朝案例了...

Even Giannis had struggled in the postseason before they found success later on. So, yeah, look, you don't have that many years of relevant NBA history. Think people maybe over indexed a bit to like the the Bulls and Warriors dynasty. Why are

你最近在专栏里说奇才的年轻核心很平庸，为什么？

the core young players of the Wizards mediocre, as you suggested in a recent substack?

他们在最有价值的选秀年里没拿到多少高顺位签。说实话他们还算可以——这话主要是我合著者的观点，不过我也同意就是了。

I mean, they haven't had that many elite picks in the in the most valuable drafts. Right? I mean, they're probably doing okay. That was my kind of, I think, co authors saying that they were mediocre. Although, I agreed with him to be fair.

迈克·哈利伯顿排第几来着？第12名。斯蒂芬·库里是第七名。我记不清所有名次了，但你知道的，

Mike Halliburton was what? Number 12. Steph Curry was number seven. I forget all the number, but, you

并非所有优秀球员都是高顺位新秀。我还没研究过NBA选秀中技能持续性的问题。在我看来，像圣安东尼奥、俄克拉荷马城甚至迈阿密这样持续挖掘球员的能力似乎相当稳定，

know, not everyone's a top pick who's great. I haven't looked at how persistent skill in NBA drafting is. To me, it seems fairly persistent for San Antonio or OKC to consistently pick up players or even Miami, I mean,

这确实像是一套相当稳定持久的能力组合。你现在对费城76人队有什么最新看法？去年你对他们挺乐观的，而我持悲观态度。没错，这次我想宣布我的预测胜利了。

that seems to be a pretty consistent, persistent skill set. And what's your current update on the Philadelphia 76ers? You were kind of bullish last year, and I was down on them. Yeah. I wanna claim victory on that one.

是啊，达里尔·莫雷，你和泰勒聊过他吗？

Yeah, I mean, Daryl Morey, Have you had him on conversation with Tyler?

没有，但他会接受采访吗？当然会。

No, but would he do it? Yeah, of course.

好吧。达里尔是个超级极客，他痴迷国际象棋这类东西对吧？他会告诉你他的策略波动性很大。

Okay. Daryl's a big geek. He loves chess, loves a lot of things like that, right? He will tell you that, like, his strategies are very high variance. Right?

你看，如果恩比德保持健康，如果埃奇科姆新秀年表现出色，如果保罗·乔治...我是说，他们有可能...比如下赛季赢56场——56场很多了，52、53场会让人震惊吗？也不至于。东部现在格局很开放。

You know, if Embiid is healthy and if Edgecombe is good in his first year and if Paul George is I mean, they have, like you know, if they won 56 games next year, would it be 56 is a lot. 52, 53, would that be shocking? I mean, not totally. Right? The the the East is pretty wide open.

他们现在这种第二时间线的构想稍微乐观了些。不，我们刚做了个专题，按未来十年夺冠概率给30支NBA球队排名，在我看来他们正好居中，大概第15名左右。

You know, they have this kind of, like, second timeline that now is a little bit more optimistic. No. So we we have this piece now where we're ranking all 30 NBA teams by their long term ten year chance of winning a championship, and they're right in the in the middle for me. I think, like, fifteenth out of 30. Who are

你人生中最重要的导师有哪些？包括篮球领域的。

the most important mentors in your life, including for basketball?

就体育写作而言，比尔·詹姆斯可算是我的导师。

I mean, for sports writing in general, you know, Bill James is somebody who's a mentor. You know, I can say And

除了阅读之外，你从他本人那里学到了什么？他确实是个很棒的作家。

what did you learn from him personally other than reading? Which he's a great writer.

我是说，比尔有点小聪明，对吧？但重要的是明白，关于技术主题的好文章本身就可以是好文章。如果你作为作家有八分水平，作为统计学家也有八分，这可能比在写作上十分而在另一领域只有两分更有价值，尤其是在我们现在所处的这个Substack盛行的领域。理查德·泰勒是我在芝加哥时稍有接触的人，我视他为导师之一。

I mean, Bill is kind a wise ass. Right? But understanding that good writing about technical subjects can still be good writing, period. That if you're an eight out of 10 as a writer, and eight out of 10 is a statistician that might be more valuable than being a 10 out of 10 and a two out of 10 in the other area, potentially in this area where we're well of sub stacks and things like that. Richard Thaler is somebody I got to know at Chicago a little bit, and someone I'd consider a mentor.

泰勒，我关注你很久了。不过我也是那种——虽然老套——但总喜欢自己开辟道路的人。在选举预测方面，哥伦比亚大学的罗伯特·埃里克森教授早期就有不错的选举模型，但总体上这些做得都不太好。所以我就想，既然我想要的产品不存在，那我就得自己动手创造。

I've followed you for a long time, Tyler, well. But yeah, I'm also someone who's kind of It's a cliche, but kind of always blazed my own trail a little bit. I think in terms of election forecasting, there's some people Robert Erickson's a professor at Columbia who had pretty good election models early on, but in general, all this was done badly. So I'm like, okay, the product I want doesn't exist. I'm to have to go ahead and make

那你现在觉得自己还需要导师来做什么呢？

it myself. What do you feel you need mentors for now?

我不确定自己是否需要导师。我只是想要一些对此感兴趣的人——

I don't know if I have mentors for. I just want people that are interested-

比如我需要导师来了解AI领域的新动态，对吧？我可以自己跟进，但需要很多帮助。

Like, I need mentors to learn what's new in AI, Right? I can follow it myself, but I need a lot of help.

也许‘导师’这个词不太适合AI方面的事。读兹维的文章——是马肖维茨吗？对——他就像是跟进AI发展的导师，因为他对此非常冷静且全面。基本上每周他都会写一篇关于AI的长篇大论。

Well, maybe mentor is not quite for AI stuff, reading Zvi, is it Mashowitz? Yes. Is a mentor for following AI developments, right? Because he's very level headed about it and very comprehensive. He'll write a novel every week basically on AI.

但他认为AI会毁灭我们所有人。你称他为冷静还挺有意思的。你可能觉得他是对的，不过——

But he thinks it's going to kill us all. It's funny you would call him level headed. You might think he's correct, but

听着，我确实认为AI安全理性主义者群体有点像个泡沫。比如如果你去参加Manifest大会——你去过吗？

Look, I do think that AI safetyrationalist community, it is kind of a bubble, right? If you go to the Manifest Conference, for example Have you been

没去过，但我知道那是什么。

No, I know what it is.

我该如何委婉地表达呢？我见识过许多怪人，但这些人即使在那个群体中也显得格外另类，我认为这种特质让他们对从多角恋到其他任何事物都保持高度实验性和开放态度。不过我觉得他们可能没有意识到：一、政治约束；二、人类潜在的适应能力。正如我之前所说，从初级通用人工智能跃升至全面通用人工智能，再到物理领域的全面通用人工智能，最终演变为超级智能，这一路径对我而言绝非显而易见。记得在Manifest时，我曾对AI 2027的预测者们表示，我们的时间线已经延后了几年。

How would I put this diplomatically? I've been around a lot of weird These people are weird even relative to the other weird people in that cohort, I think, in ways that allow them to be very experimental and open minded about everything from polyamory to whatever else. I do think they maybe don't recognize, A, political constraints, B, the adaptability of human beings potentially. See, as I talked about before, I think it's not at all obvious to me that you leap from rough AGI to comprehensive AGI, and then to comprehensive AGI for the physical realm, and then to a SI. I think the AI twenty twenty seven forecasters when I was at Manifest, I said, Okay, our timelines have moved out by a couple of years.

这个更新挺有意思的。比如我重写了《悬崖边缘》的前言，大概是在二三月完成的——那时马斯克对白宫影响力很大，他们正加速推进AI事务，关税政策也开始实施。我在新版前言中提到，相比成书时，我对末日降临的预期概率提高了。

That seemed interesting, an interesting update, for example. So I wrote a new preface to On the Edge, and I wrote this back in, I guess, February and March, right? At a time when Elon had a lot of influence in the White House, when they were pressing the accelerator on AI stuff more, when tariffs were going into effect, and I wrote then that my pay doom had increased relative to the book.

你只是不认为它能

You just don't think it can

成为超级预测者，但它可能毁灭我们所有人。我担心人类会利用AI更便捷地制造武器或实施某些恐怖活动，比如危险的化学合成——这些我都忧心忡忡。

be a super forecaster, but it can do us all in. Well, I worry about humans killing other humans with with AIs that make making weapons easier or certain types of terrorism easier, right, or chemical compounds could be dangerous. I worry about all of that.

但那不算末日，只是些恶性事件。如今无人机每天都能造成严重危害。

But that wouldn't be doom. Those would be some bad events. Right now, drones can be very bad, and they are every day.

听着，我们与技术共处的时间——你也提过这点——我们与核武器共处的世代并不长，这依然令人不安。就像尼克·博斯特罗姆论述的：当某项技术具有不对称性时，任何一个疯子只要获得无法完全防护的系统，都可能毁灭世界。

Look, we have not lived with techno I mean, you've this before too. We have not lived with nuclear weapons for that many generations. Right? That's still a little bit frightening. You get to get some of the Nick Bostrom stuff where eventually when we have some technology, we invent where it's asymmetric, where any one crazy person can maybe destroy the world if they get access to the right systems that can't be that well protected and things like that.

但从某些方面看，AI比人们预期的更接近人类，这在我看来是个积极进展。如果你考虑...

But, look, in some ways, I think the AIs are more human like than people expected, and that to me seems like a positive update. If you think about

根据你书中的框架，那些多元类型的人如何看待风险？他们是否普遍更厌恶风险或更焦虑？还是说他们都自认为掌握了他人不知的秘密？你如何归类他们？仅仅是开放性高吗？还是别的特质？

the manifold types in terms of the framework in your book, how they think about risk, is there a common feature that they're more risk averse or that they worry more? Is their common feature that they like the idea that they hold some kind of secret knowledge that other people do not have? How do you classify them? They're just high in openness or what is it?

他们的经验开放性很高，尽责性也很强——至少部分人是这样。

They're high in openness to experience, right? I think they're very high in conscientiousness. Are they? I don't know. Some of them are.

确实部分人是。我认为他们在争议性上呈现高方差特征

Some of them are, yeah. I think of them as high variance in contentiousness

我认为，比起高效利他主义者（EAs），理性主义者群体更具高波动性。某种轻信倾向确实是个问题。显然，EA因Sam Bankman-Fried事件声誉受损，但说实话他们本该承受更多舆论压力——那起事件实在太恶劣了，而且早有征兆，包括他接受你和他人采访时显露的端倪。

rather The than high in EAs are and the rationalists are more high variance, I think. I I think there can be a certain type of gullibility is one problem. I I think obviously EA took a lot of hits for Sam Eichmann Fried, but I think if anything, they probably should have taken more reputational damage. Right? That was really bad, and there were a lot of signs of it, including his interviews with you and other people like that.

没错。这与扑克玩家形成鲜明对比，虽然表面特质相似，但他们更为多疑且深谙世故。另外湾区很怪异，我感觉西海岸正与全美其他地区加速分化。

Yeah. Because it contrasts with poker players who have similar phenotypes, right? But are much more suspicious and much more street smart. Also, the Bay Area is weird. I feel like the West Coast is diverging more from the rest of the country.

就像渐行渐远那样。连行为习惯都不同了，对吧？举个小事例：在湾区参加家庭派对，可能酒水寥寥；而在纽约，主人若不备足酒精饮料简直是大不敬。

Kind of like a long way away. Just the mannerisms are different, right? Small thing, right? You go to a house party in the Bay Area, there may not be very much wine, for example. In New York, if the host isn't drinking, then it'd be considered sacrilege not to have plenty of booze at a party.

诸如此类的细微文化差异。西雅图给我的感觉几乎像加拿大。这些差异正在不断扩大。

Little things like that, little cultural norms. You go to Seattle, it feels like Canada to me almost. So these things are diverging more.

为什么末日信念与多伴侣关系实践存在关联？我认为确实如此。如果问Ayla，她大概会说：既然我们终将灭亡或迎来奇点，不如及时行乐。

And why is belief and doom correlated with practice of polyamory? And I think it is. I mean, if you ask Ayla, I guess, she might say, Well, if we're

确实存在这类享乐主义倾向，虽然整体上不算极度纵欲的运动。在我看来太过经济理性化了——即便作为经济学家的我，也不认为人们会如此精于算计。更可能是心理上对两种观点都存在 predispositions。

all gonna die or go to whatever singularity there is, we might as well have fun in the meantime. There's some of that kind of hedonism, although in general it's not a super hedonistic movement. It seems too economistic to me. Even I, the economist, I don't feel people think that economistically. That there's more likely some psychological predisposition toward both views.

不。理论上可以说社会以多伴侣关系组织会更高效——人们其实已在许多方面隐性实践，包括LGBTQ群体对此的不同态度，以及生育率下降可能产生的影响。但关键在于他们不受传统社会武装观念的约束——那个群体对此极为重视，对吧？

No. Look, I guess you could argue that society would be better organized in a more polyamorous relationship. People do it implicitly in a lot of ways anyway, Including in the LGBTQ community has different attitudes toward it potentially, and if there's not as much childbearing that can have an effect potentially. Yeah. But I think it's kind of like Look, they're not being constrained by the armed society thing that is taken very seriously in that group, right?

足够的疏离感让他们能更彻底地实践这种理念。这种风气也略微渗透到硅谷，比起我认识的华尔街类型，这里更异想天开得多。

There's enough disconnectedness and aloofness where they're able to play it out in practice more. That creeps into a little bit into Silicon Valley too, which can be much more whimsical and fanciful than the Wall Street types I know, for example.

为什么加拿大球队赢不了斯坦利杯？

Why can't Canada win the Stanley Cup?

记住。

Remember.

上次是1992年，对吗？

It's 1992 the last time. Is that correct?

我记得我忘了具体时间。2012年在《纽约时报》写过这个。2013年，2013年，对吧？

I remember I've forgotten when. Writing about this in the New York Times in 2012. 2013, 2013, right?

是的。现在依然如此。

Yeah. It's still true.

是啊。我去我最喜欢的运动酒吧之一。把那篇文章打印出来贴在洗手间里。让人们上厕所时能看到文章。你可以读点东西。

Yeah. I go in one of my favorite sports bars. Had the article printed out in the bathroom. Let people see articles in the bathroom while you're taking a shit. You can read some article.

我在洗手间读自己的文章。听着，税法确实有点影响对吧？佛罗里达没有州所得税，在设有工资帽的联盟里——其实我有个好朋友是佛罗里达黑豹队的助理总经理桑尼·梅塔。在硬工资帽联盟里实际多赚10%到15%，这差别可大了。

I'm reading my own article in the bathroom. Look, tax laws matter a little bit, right? Where Florida has no state income tax, in a league with a salary cap Actually, a good friend of mine is the assistant general manager of Florida Panthers, Sonny Mehta. And yeah, if you're making de facto 10 or 15% more in a hard capped league, that matters quite bit.

如果在加拿大打球，代言收入会不会更高些？还是说并非如此？

If you're playing in Canada, isn't your endorsement income higher maybe, or is that not true?

整体市场更小。比我们穷些对吧？作为加拿大NHL巨星在美国打球——我不觉得西德尼·克罗斯比会有什么问题。不知道他有没有接蒂姆霍顿斯的广告之类的。嗯，我不清楚。

It's a smaller market overall. It's kind of less wealthy than us, right? You're a Canadian NHL superstar playing in The US, I don't think Sidney Crosby is in any trouble. I don't if he's on Tim Horton's ads or whatever, right? Yeah, I don't know.

全球范围来说，我们过去常讨论哪些国家能免疫民粹右翼情绪。现在日本也开始出现这种苗头。加拿大在国家层面没有，但某种程度上差点发生。

Around the world, we used to talk about which countries are immune to populist right sentiment or whatever you want to call it. Now Japan is flirting with populist right sentiment. Canada, not at the national level, but it almost happened in some way.

几年前渥太华的卡车司机抗议事件。

The truck driver thing a couple years ago in Ottawa.

对。所以自我们上次聊过后有什么新进展？这会蔓延到所有地方吗？还有哪里能幸免？爱尔兰也出现了迹象。

Yeah. So what's your update since the last few times we've spoken? Is it just gonna come everywhere? Is anywhere immune? In Ireland, you see signs of it.

他们可能赢不了下次选举，但几年前这种势头似乎完全不存在。现在你看，麦格雷戈参选了，虽然我觉得他不会赢，但总会产生些影响。

They may not win the next election, but it seemed to be totally absent a few years ago. Now, you know, MacGregor is running, and I don't think he'll win, but it'll have some kind of impact. You know,

我认为欧洲部分地区，可能还有加拿大这些反美情绪较强的地方，对民粹主义思潮的抵抗力会更强。比如我有个爱尔兰朋友，他是同性恋，20岁移居到这里。他说在爱尔兰长大时，社会对同性恋相当排斥。

I think parts of Europe that are and maybe Canada are anti US may become more resistant to populist sentiment, for example. Right? I have a friend who is Irish actually and gay and moved here when he was 20. He's like, yeah, when I was growing up in Ireland, it was quite anti gay. Right?

那时爱尔兰是个非常宗教化的国家。现在几乎变成激进支持LGBTQ+的国度了。

It was a very religious country. Now it's like almost aggressively pro LGBTQ plus right, I think.

但可能两者并存。你知道德国选择党的党魁是女同性恋，伴侣还是斯里兰卡人，对吧？

But could be both. As you know, the head of AFD in Germany, she's a lesbian with a Sri Lankan partner. Is that correct?

差不多是这样。

Something like that. Yeah.

但她属于民粹右翼。

Yeah. But she's populist right,

同时她公开自豪地支持同性恋权利。我觉得政治就像恒温器效应，人们会越来越朝现状的反方向跑。疫情后反建制和民粹情绪高涨有其原因，但现在可能已经见顶了。

and she is pro gay rights, and openly and proudly so. Look, I I mostly believe in kind of like thermostatic effects in politics where people run the opposite direction of the kind of status quo more and more. I I think there were reasons why anti incumbent and populist spirit rose from the pandemic onward, but we might have been at a high watermark for it potentially Totally,

看起来还没到顶峰，其他不少国家这种势头还在上升。

it seems we're not at a high watermark. It's rising in quite a few other nations.

我主要考虑加拿大、西欧等高收入国家的情况。有件事一直困扰我，甚至问过首相们——许多欧洲关键国家的选民显然希望大幅减少移民（不论你我怎么看），但为何始终未能实现？右翼民粹政党却持续崛起。

I guess I'm thinking about Canada, Western Europe, other high income countries. Maybe I'm wrong. Something that's puzzled me, and I've even asked prime ministers this, I've never gotten a good answer, it seems to me really a lot

为什么他们的诉求得不到满足？

of European voters in some key nations want much lower immigration. Whatever you and I might think. They clearly want it, and they don't get it. The populist parties on the right continue to rise. Why don't they get it?

为什么掌权的中间派政客不干脆采取行动呢？即使他们不喜欢做这件事，也可以捏着鼻子去做，然后继续保住职位？

Why don't the centrist politicians in power just do something, hold their noses if they don't like doing it, but just do it and stay in office?

我是说，你在丹麦已经看到了一些这样的例子。比如移民数量大幅减少。就连贾斯汀·特鲁多也采取措施削减了合法移民和非法移民的数量。

I mean, you have seen some of that in Denmark. For example, you've seen much less immigration. Even Justin Trudeau has somebody to cut down the amount of legal immigration as well as illegal immigration.

是啊。但看看德国没这么做。英国没这么做。爱尔兰没这么做。荷兰现在开始做了，但考虑到我对其他多数议题上民主问责制的理解，这个过程似乎过于缓慢了。

Yeah. Mean, look- Germany's not doing it. UK's not doing it. Ireland is not doing it. Netherlands is now doing it, but it just seems to be a very slow process given how I think about democratic accountability on most other issues.

我是说，美国左翼过去在这方面更有节制，对吧？比如伯尼·桑德斯在移民问题上就更谨慎些。如果你有一个慷慨的福利国家，那么接纳新成员可能会让所有人付出代价。不过我觉得这些可能在几年内就会逆转——当世界意识到我们普遍面临年轻人口短缺、生育危机（如果你想这么称呼的话）以及人口老龄化时，就会渴望那些愿意努力工作且具备技能的人。所以他们可能正在顺应未来趋势。

I mean, there used to be more restraint of it on the left in The US, right? Where Bernie Sanders, for example, would have issues like immigration where he's a little bit more cautious, right? If you have a generous welfare state, then admitting somebody new to it is potentially gonna cost everybody. Look, I think all this might reverse within a few years where the world realizes we have a shortage of young people in general, a fertility crisis, if you want to call it that, right? And aging the population, and will want to have people who are willing to work hard and are skilled in particular, and so maybe they're skating to where the puck is going.

我认为美国在某些方面还拥有特别高效的政治成果。人们认为这些大致五五开的政党联盟理所当然——虽然特朗普在很多方面重塑了共和党，但每次选举结果仍基本持平，只是联盟构成不同。也许我们的选举技术更成熟。目前美国相比世界多数国家更亲移民。而由于反弹效应，特朗普至少在名义上又开始转向亲移民立场，这本来会很有趣。

The US also has, I think, particularly efficient political outcomes in some ways. People take for granted the fact that you have these coalitions that are fiftyfifty approximately very consistently, that Trump remade the GOP in a lot of ways, but still every election is roughly fiftyfifty with a different coalition than you had before. So maybe our electoral technology is better. Right now, America is quite pro immigrant compared to most of the world. And because of backlash, Trump becoming at least nominally more pro immigrant again, it would have been interesting.

但我不确定这是事实。我知道人们引用的所有民调，但当我观察选民对ICE（移民海关执法局）活动的容忍度时，我的感觉是他们只是...

I'm not sure that's true, though. I know all the polls that people cite, but when I look at the electorate's willingness to tolerate, say, ICE activity, my sense is they'll just

勉强忍受罢了。听着，共识是我们确实有些坏人——就像特朗普称呼的那样，人们对他们没什么同情心。我认为大众总体上对 migrant worker（外来务工人员）还算宽容，对技术移民也普遍支持。人们想要加强边境管控。

put up with it. Look, I think the consensus is that we have some bad guys, as Trump would call them, and people are not particularly sympathetic to them. I think people are broadly tolerant toward the migrant worker class for the most part. I think people are broadly supportive of skilled immigration. People want more border enforcement.

这其实存在共识，只是两党似乎无法达成一致。不过我也同意民调存在些问题。如果你问人们'如何看待这份关于移民的10个问题清单'，他们会倾向左翼立场。但问'你更支持谁'——

There is a consensus there that the two parties can't seem to actually reach a consensus on. But yeah, look, I agree with the polling issues a little bit. If you ask people, How do you feel about this 10 item list of questions of immigration? They'll side with the left. Who do you like more?

特朗普还是民主党？他们仍会说更信任特朗普一点。民主党确实存在这个问题——你给他们一寸，他们就要一英里。人们有理由担心这种过度行为，比如疫情期间——说实话前三个月我非常谨慎，

Trump or Democrats? They'll still say, We trust Trump a little bit more. I mean, Democrats do have this problem a little bit where you give them an inch and they'll take a mile. I mean, I think people correctly are worried about those excesses, right? Where I think during COVID, for example, I mean, I was kind of very cautious in the first three months really of COVID.

记得我第一次去户外餐厅是六月左右？当时吓坏了。老实说几个月都没接触过除几个朋友外的其他人。但后来发现半数公共卫生专家希望彻底禁止社交直到疫苗问世，甚至之后也可能继续。所以人们担心一旦给进步派某些权力，事情就会失控——尽管表面上看他们可能同意自由派的渐进主张。

Remember the first time I went outdoors to a restaurant, it was June or something, right? I was terrified. I hadn't been around other people, apart from a couple of friends, for being honest, in months, right? But then you realize that half of the public health people wanted no real socialization until we had vaccines, and maybe not even after that, right? And so people are worried that if you let progress in some power, then things do get taken too far, even though nominally the incremental stuff, they might agree with the liberal side.

最后两个问题。首先，你对美国有什么出人意料的预测吗？我是说，美国还能持续多久？我我我不确定。我不确定是否能给出确切预测。

Two last questions. First, do you have a surprising prediction for The United States? I mean, how long will The US last? I I I don't know. I don't know if I have a prediction there exactly.

要做出人意料的预测很难啊，泰勒。

It's hard to have surprising predictions, Tyler.

确实，但你对NBA有什么惊人预测吗？比如我认为湖人队今年会表现糟糕。虽然不算意外，但我的记录显示湖人队实际排名还挺高的。对了，上次对话你说卢卡被正确评价了。

It is, but do you have a surprising prediction for the NBA? Like, I think the Lakers will do poorly this year. I'm not sure it's surprising, but In my record, said the Lakers rank relatively highly, actually. Yeah. Our last session, you said Luca was properly rated.

现在我发现连他自己的球队都不想要他。勒布朗也不太乐意和他共事。之前杰伦·布伦森也不愿和他打球。在我看来这...

I now see his own team didn't want him. LeBron is not crazy about having him around. Earlier, Jalen Brunson didn't wanna play with him. To me, that's a

都是负面信息。所以我认为湖人队会表现不佳。我更关注长期品牌无形资产，这在NBA似乎很重要。他们现在有新老板可以随意支付奢侈税。说实话，去年我看卢卡时，他状态不太对劲。

lot of negative information. So I think the Lakers will do poorly. I'm looking more kinda like a long term brand intangible, which I think does seem to matter a lot in the NBA. They have new ownership now that can spend all the luxury tax that they want. Look, when I saw Luca last year, he didn't seem quite like himself.

但这正是你该押注超常发挥运动员的原因吗？不，他们已经超常发挥了。而一个有问题的运动员——比如酗酒或不够自律，只要改过自新就有更大成长空间。完全健康的卢卡加上长期薪资结构良好，可能打造强队。拉里·伯德是你这边的例子，但我倾向于...

But it's actually a reason why you wanna do you wanna bet on the overachieving athlete? No, they're already overachieving. Whereas an athlete who is a problem, right, has a drinking problem or not drinking, having too many beers or being a little bit lazy, he has much more potential to grow if he gets his act together. So fully healthy Luca, and their salary cap situation's pretty good in the long run, could be a powerful team. Larry Bird is an example on your side, but I tend

在心理层面押注这类人失败。也许这就是人与人之间的差异...

to bet against such people mentally when it comes to it. Maybe that's a difference between People

你是指当前处于巅峰状态的人？

who are maximizing currently?

不。是指那些有问题的人，比如酗酒、不够尽责或态度恶劣的。

No. People who have some problem, like they drink too much or they're not conscientious enough or bad attitude.

但这里需要高波动性对吧？如果卢卡用了Ozempic之类的，说不定突然就能让湖人夺冠了。

But you high variance here. Right? You want high variance here, and maybe if Luca takes Ozempic or something, then all of a sudden, like you have a Lakers championship.

再次推荐一下你的新平装版作品《边缘：孤注一掷的艺术》。最后一个问题是，你接下来有什么计划？

Again, a plug for your the new paperback edition, On the Edge, The Art of Risking Everything. And the final question is, what will you be doing next?

目前我正在做一个NFL（美国职业橄榄球大联盟）的模型项目。确实，泰勒，我挺怀念写书的日子——即便遇到写作瓶颈或采访受阻，这些长期项目对我而言始终意义重大。

I'm working on an NFL model currently. Yeah. I miss having a book project, Tyler, because every day, even days when you're struggling to write or an interview doesn't come through, whatever else, having these big projects that I work on, I find to be very important.

又没人拦着你，所以具体是什么计划？还是说仍在构思中？

No one's stopping you, so what's it gonna be? Or you're still looking?

我不想透露太多...首先，我的电子报《银色简报》去年表现远超预期，算是短期项目。关于下一本书——可能几年后才会动笔——我有几个体育题材的创意，有些可能会让人意外。德州扑克也算是个长期课题，把这类技艺练到极致很有成就感。

I don't wanna give I have So first of all, the newsletter, Silver Bulletin, did quite a bit better than I expected last year. That's kind of like the short term project. I have a couple of ideas for books when I write the next book, which might be in a few years. I have ideas for books about sports I don't want to give away too much that might surprise people a little bit. I still think of poker as a little bit of a project, getting it satisfying to be really good at something like poker.

但现阶段重点是经营电子报，这是为下次大选准备的三年半计划。之后或许终于能跳出选举周期——虽然我每次都这么说，28年这次可能是最后一次了。

But for now, the newsletter building toward that is kind of the three and a half year plan here until the next election, then maybe I finally won't escape the election treadmill maybe after that. I said that every time. This might be the last one, twenty eight.

让我们拭目以待。内特·西尔弗，非常感谢你。谢谢泰勒。感谢收听《与泰勒对话》，您可以在苹果播客、Spotify或任何播客平台订阅本节目。

We'll see. Nate Silver, thank you very much. Thank you, Tyler. Thanks for listening to Conversations with Tyler. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.

若喜欢本期节目，请给予评分和评论，这能帮助其他听众发现我们。我的推特账号是TylerCowen，节目账号是cowenconvos。下期节目之前，请持续收听与学习。

If you like this podcast, please consider giving us a rating and leaving a review. This helps other listeners find the show. On Twitter, I'm TylerCowen, and the show is cowenconvos. Until next time. Please keep listening and learning.