前雷曼兄弟交易员：AI 无法为你带来优势

我觉得AI和智能体不会给你带来优势。

I think that AI and agents will not give you edge.

技术在不断进步。

Technology continues to evolve.

但要执行这些，还是需要人来操作。

But executing that requires execution by humans.

如果真是这样，那为什么所有交易公司都在计算上花那么多钱？

If that's the case, why are all the trading firms spending exorbitant amounts of money on compute?

因为他们已经是执行这些事情的专家了，你知道，创造智能、智能体、规则，还有发现优势。

Because they're already pros of executing those things, you know, creating the intelligence and the agents and the rules and finding the edges.

这就是他们能不断壮大的原因。

That's the reason why they are able to grow.

你看到现在系统化交易的人数在激增吗？

Are you seeing an explosion in the number of people who are systematically trading today?

月活跃开发者数量，六个月里从两万增长到了八万多。

The number of the monthly active developers, it's been growing from 20,000, 80,000 plus in six months.

这就是我喜欢资本市场的地方。

That's what I love about capital markets.

在我看来，资本市场就是街头格斗。

Capital markets, in my opinion, is street fighting.

我对风险很上心。

I'm partnered about risk.

我从雷曼兄弟开始做起。

I started the Lehman Brothers.

没有什么能永远持续下去。

Nothing is going to last forever.

Yoshi，非常感谢你来参加这个播客。

Yoshi, thank you so much for coming on the pod.

谢谢你邀请我，兄弟。

Thanks for having me here, man.

今天，零售量化交易者还有优势吗？

Is there edge for a retail quant trader today?

是的，当你创造出一个优势的时候。

Yes, when you create one.

这取决于你能不能创造出来。

It depends on if you can create one or not.

所以我觉得不管是散户还是专业交易员，其实都不重要。

So I don't think it really matters if you're a retail trader or pro traders.

关键在于你什么时候创造出来，你就有优势。

It really, depends on when you create it, you have an edge.

如果你创造不出来，你就没有优势。

If you cannot create it, you don't have an edge.

你从这个角度看待问题，确实很独特。

You're seeing this from a very differentiated angle.

你的API被很多这些人使用。

Your API is used by a lot of these guys.

你通常看到他们利用什么样的优势？

What sort of edges do you tend to see them exploiting?

所以我之前说，当你有优势的时候你就有优势，当你创造优势的时候你就有优势，我想解释的核心点是，即使是散户，只要你做得够多，或者你真的找到了那个优势，你也能有优势。

So the reason why I said that you have an edge when you have an edge and when you create one, the whole point that I wanted to try and explain is that even retail traders could have an edge when you do it enough or when you really find it.

我们的平台提供可编程的访问，也就是API访问。

So our platform is offering programmable access, API access.

然后有些人会创造自己的优势，找到自己的优势，然后持续赚钱。

And then there are people who create their edge, who finds their edge, and they keep making money.

确实有这样的人。

So there are people like that.

我不知道他们算专业还是散户。

I don't know if they are considered to be pro or retail.

我不清楚怎么分类，但确实有人通过我们的平台找到了优势。

I don't know the categorization, but it is true that some people do find edges using our platform.

我觉得这特别有意思，是因为我觉得，作为社会，现在的文化趋势几乎把每个人都推向这个方向，让大家更多地交易，去追求财务自由，有些人甚至想靠运气一路冲上去，对吧？

The reason why I find this so interesting is because I think, as a society, the cultural trends that exist today are pushing everyone in almost this direction to trade more, to, you know, access financial freedom, and in some cases even punt their way to the top, right?

我觉得这就是大家现在看到的。

I mean, that is what I think everyone is seeing.

你看那些永续合约交易所，简直火得不行。

You look at perps exchanges and they are absolutely booming.

我跟一个顶级永续合约交易所的创始人吃过午饭，他们的产品简直火爆。

I had lunch with, a founder of one of the top perps exchanges, and their product is just ripping.

所以我想问你，如果你今天处在他们的位置，我知道你在这方面有些背景，你会实施哪些流程来建立一个系统化的框架，在市场中赚钱？

And so my question for you is, if you were in those shoes today, and I know you have some background in that, processes would you implement in order to build a systematic framework for making money in the markets?

如果我是真正的参与者，是的，交易

If I were the actual participants on Yes, the trading

站在参与者这边。

on the side, participant side.

我的意思是，这就是AI和智能体话题发挥作用的地方。

I mean, that's where this AI and agents conversation kick in.

因为AI和大语言模型，确实大大减少了资源和时间。

Because of the AI and LLM, it really hugely reduces the resources and time.

即使你是个体，也可以尝试探索优势在哪里。

Even if you're an individual, try to explore where the edges are.

任何试图创造优势的过程，其实就是不断试错。

Any process of trying to create an edge, it's just trial and error.

这就像做研发一样。

It's an R and D.

没有AI的话，当然会花更长的时间，对吧？

And without an AI, of course, it takes much longer time, right?

关键就在于编码。

It's all about coding.

关键在于优化，然后根据市场变化的趋势不断微调。

It's all about the optimization and then constant fine tuning based on the trend of the market changes.

所以，我们如何把这种过程自动化并进行微调，这会让你的工作变得非常耗费资源。

So how we can do that kind of process and automating to fine tune it, it becomes much resource intensive in terms of what you do.

这意味着你可以做更多的试错。

That means that you can do more trial and errors.

所以，找到优势的过程，我觉得因为AI和大语言模型而缩短了。

So the process to get to finding an edge, I think it got shortened up because of this AI and LLM.

所以如果我现在做这件事，我就会这么去处理。

So if I were doing this today, so that's how I would approach it.

因为，你知道，不是说用了AI就一定能找到优势。

Because, like, you know, it's not gonna be, okay.

如果我用AI和智能体，我就能找到优势。

If I use AI and agents, I can find an edge.

对我来说，更重要的是让整个过程大大缩短，这样我就能做更多的试错，从而找到更好的优势。

It's really more for me to make the whole process much shorter so that I can do more trial and error, so I can get better edges.

所以这就是我的看法。

So that's how I'd frame it.

那你觉得，这种迭代过程的时间缩短，会不会让市场中的机会更加民主化？

And do you think that that shortening of the, call it, time horizon for iterating on certain processes, do you think that that democratizes opportunity within the markets?

还是说，这会让机会集中在那些对迭代循环非常精通的专家手里？

Or do you think that concentrates it in the hands of people who are absolute experts in those iterative loops?

在我看来，我们还得考虑到，有优势并不意味着你就能成功。

In my opinion, I think we also have to think about that having an edge doesn't mean that you're going to be successful.

有优势只是第一步，第二步是真正根据你的优势去执行和坚持，而这完全是另一种技能。人的行为不会改变，对吧？

Having an edge is a first step and the second step is actually execute and follow through based on your edge and that's a completely different skill set human behaviors don't change, right?

我们总是有贪婪，也总是有恐惧，正因为如此，有些人即使有很好的优势也永远不会成功，因为如果你对执行感到非常恐惧，你就永远无法执行这些优势。我觉得这就是交易或者我们生活中做任何事情的趣味所在。

We have always greed always fear and because of that reason, some people will never be successful even if you have amazing edges because you will never be able to execute the edges if you are so fearful of actually executing it and I think that's the interesting part of this trading or anything that we do in life.

我们可能恰好处于完全相同的情况，拥有相同的工具或相同的条件，但结果会因你、因我或因任何人而完全不同。

We may happen to be in the exact same situation, the same weapons, or like same features, but outcomes will be all different based on like you or based on me or based on anyone.

我认为在交易环境中也是如此，即使你有优势。

And I think that happens in this trading environment too, even though you have an edges.

你刚才提到了行为动态。

You touched on the behavioural dynamics just there.

在与你的用户交流、了解你的用户时，那些零售交易者、零售系统化量化交易者，他们往往有哪些你经常看到的行为？

In speaking to your users, in understanding your users, which retail traders, retail systematic quantitative traders, what sort of behaviors do they tend to have that you see come up often?

我看到那些更成功的人，他们往往是重复性的，基本上就是有优势并且真正执行。

What I see that tends to be more successful is that they're like, you know, the repetitive basically having an edge and actually executing.

然后当市场动态发生变化时，他们会完全停止交易，暂停一下，改进或微调以适应新的市场动态，然后再回来。

And then when the market dynamics changes, they actually completely stop trading and take a pause and improve it or fine tune it for the new market dynamics and come back.

这就是我看到的那些真正有优势并且能正确执行的人的模式。

And that's a pattern that I see for those people who actually have an edges and actually in addition to that being able to execute in the right way.

所以这就是我看到的模式。

So that's a pattern that I see.

关于这些成功的零售量化交易者，他们从财富创造的角度是如何看待市场这个问题的？

With regards to these successful retail quant traders, how do they approach the problem of the markets from a wealth creation perspective?

他们倾向于把它看作一个现金流生意，模型在赚钱，然后他们花掉这些钱吗？

Do they tend to see it as a cash flow business and that the models are making money on their money and they spend that?

还是说这是一个复利游戏，他们真的在努力积累和构建财富？

Or is it a compounding game where they're really trying to run and build wealth in the process?

我觉得，当然，这真的取决于个人。

I think, of course, it really depends on people.

每个人开始的方式可能都不一样。

How people start it, probably it's all different.

比如，有些人只是有闲钱和空闲时间，而且你会编程。

Like, some people just had to spare money and spare time, and, you know, you're able to code.

比如说到这个阶段，你其实不需要会编程。

Like, at this point, you don't have to be able to code.

我可以制定这些规则，然后开始尝试。

I can create these rules and start trying.

而且大多数时候，你知道，就是……

And most of the time, like, you know, the...

你只是在玩玩而已，对吧，就像个爱好。

You're just playing around, right, as in a hobby.

看看效果怎么样。

Let's see how it works.

咱们真的去把你脑子里想的执行到现实世界里。

Let's actually, like, execute what you're thinking into the real world.

这是验证你的想法到底能不能赚钱的最简单方式，或者，你知道，就是……

This is the easiest way to verify why you think actually makes money or, like, you know, the...

你知道，事情总会发生的。

You know, things happen.

反馈周期非常短。

Very short feedback loop.

做错了就亏钱，做对了就想赚钱，对吧？

When you do bad things and you lose money, and when you do it right, you want to make money, right?

所以我觉得确实是这样，但归根结底，很难看出来，因为进入这个领域的人，各自的动机都不一样。

So I think that happens, but at the end of the day, like, you know, it's very difficult to see that because all the people who come into this situation, it all has a different motivation.

所以有些人可能会发现，并且变得非常自信，觉得它一直运作得很好。

So some people may find and become very confident that it's continued to work well.

所以他们可能会投入更多的钱进去。

So they may be able to put that, put more money into it.

我觉得有些人会一直非常谨慎，有自己的规则，不会冒险投入超过一定金额的钱。

I think some people continue to be very, very tight and then have their own rules that they're not going to risk any more dollar amount than certain amount.

我觉得这又取决于每个人对恐惧和贪婪的程度。

I think, again, this is based on that each individual's amount of the fear and greed.

Alls On Open 由 Onex Capital Group 赞助，他们是石油衍生品领域的顶级做市商。

Alls On Open is sponsored by Onex Capital Group, the top market maker in oil derivatives.

他们刚刚发布了终端机的第四版，Flux，里面有两件事如果你有石油敞口的话值得了解一下。

They just released the fourth version of the terminal, Flux, and there are two things in it worth knowing if you have oil exposure.

第一，它实时流式传输800条石油曲线的公允价值，包括那些难以定价的曲线。

First, it streams live fair value on 800 oil curves, including the ones that are hard to mark.

第二，它包含他们自己的持仓数据，所以你可以看到市场是多头还是空头，以及他们入场的价格。

Second, it carries their own positioning data, so you can see how the market is positioned long and short and at what price they got in.

你可以在flux.live找到它，链接在描述里。

You can find it at flux.live, and the link is in the description.

当你看看你的用户群，我很好奇，使用API的客户中，散户和机构对冲基金的比例大概是多少？

When you look at your user base, I'm curious what portion tends to be your retail versus institutional hedge fund guys, you know, using the API.

我问这个问题是因为我认识一些用过Alpaca的人，他们在上面开展业务。

And I ask that question with the context of me knowing people who have used Alpaca, who have run businesses on top of it.

是的。

Yes.

很想听听你的看法。

Would love to hear it.

所以作为Alpaca，作为我们的业务，我们大概有两条业务线。

So as Alpaca, as our business, we have like two business lines.

一条呢，我称之为B2B2C。

One is, I would call it, like, B2B2C.

就是其他企业基于我们的基础设施来构建，这样他们就能成为自己的券商，向他们的终端客户提供交易应用。

So other businesses build on top like, businesses are built on top of our infrastructure so that they become their own broker dealer to offer that trading app to their end customers.

所以基本上，这算是B2C。

So basically, it's a B2C.

所以这些客户中有很多机构，遍布50个不同国家的数百家机构和券商，他们都在为终端客户打造类似Robinhood的投资应用和券商服务。

So those are a lot of institutions, hundreds of institutions, broker dealers in 50 different countries building Robinhood type of the investing apps and broker dealers for their end customers.

这就是我们其中一条业务线。

So that's one of the business lines we have.

另一条就是我们刚才聊到的，也是我们在2018年最初起步时做的。

And another one is like the one that we're talking about that we originally started with in 2018.

就是人们直接连接我们的API，用自己的资金进行交易。

Just people just directly connect to our API and deploy their own capital and trade.

我觉得有意思的是，当你说机构还是散户的时候，这两者之间的界限开始变得模糊了。

What I felt interesting is that when you say institutional or retail, it kind of starts the line between those two things getting blurred.

就像你自己一个人就能注册一个有限责任公司，不用通过什么Stripe Atlas之类的，一键就能搞定。

It was like you by yourself can create LLC without using, I don't know, Stripe Atlas or something and just one click and you can create LLC.

然后你突然就被当成机构投资者了，因为这是公司账户。

You're suddenly considered as institutional investor because it's an entity account.

所以这个问题问起来其实很微妙，因为你看，虽然我们开的是机构公司账户作为经纪账户，但实际上可能只有一个人在那儿交易，对吧？

So it's very, very, you know, nuanced based when you ask that question because, like, you know, okay, this is an institutional entity account that we're opening as a brokerage account, but there's actually only one person who's, you know, just doing the trading, right?

所以很难说清楚。

So it's kind of hard to see.

所以呢，公司账户这块，我们看到增长挺多的。

So like, by entity accounts, like, we're seeing a lot of increase.

我觉得原因还是，你看，现在全球范围内，独立创业者越来越多了。

I think the reason why is, again, you know, the rise of the solopreneurs in general when we look at the world.

人们靠自己就能赚更多的钱。

People are being able to make more money by themselves.

我觉得赚钱或者创造就业变得越来越独立了，因为有了YouTube博主、TikTok博主，还有那些智能体之类的。

I think the money making is becoming or job creation is becoming extremely independent because of the YouTubers or TikTokers or because of the agents and everything.

所以NTD账户的数量在增长，但我不会把每个人都当成机构投资者。

So the numbers of NTD accounts are growing, but I wouldn't consider everyone to be institutional investors.

这就是我的看法。

So that's how I think about it.

你这话说到点子上了，第一，界限确实越来越模糊了。

Now you've hit the nail on the head there with, number one, it's getting blurred.

而且说实话，这让我开始思考资产管理和对冲基金行业的未来会是什么样子。

And, honestly, that is making me think about what the future of asset management or hedge fund businesses look like.

因为你看，有个人，他交易一个账户，比如用Alpaca的API，而且他的回报还不错。

Because you have a guy, he trades an account, say he uses Alpaca's API, and his returns are good.

你懂吧？

You know?

咱们就说夏普比率挺高的，交易各种不同的市场，因为有了这些新工具，他能很快地调整自己的策略。

Let's call it reasonably high sharp, trades a bunch of different markets, can iterate very quickly on his process because of all the new tools.

他完全可以开一家对冲基金。

He could whip out a hedge fund.

你说呢？

You know?

他有业绩记录。

He he has the the track record.

所以这是一个观点……

And so this is a take...

我很想听听你对这个行业未来走向的看法。

I'd love to hear your take more so on where you see the future of the business going.

你觉得整个生态系统会怎么演变？

I guess, how do you see that whole thing and ecosystem evolving?

然后第二个问题，不是想一下子问你太多，你觉得十年后对冲基金的数量会比现在多吗？

And then, I guess as a second one to that, not to throw too many questions at you, but do you think there are more hedge funds ten years from now than there are today?

我们经历过Y Combinator，就是YC，作为校友，我们可以追踪到申请加入这个批次的企业类型。

So we went through Y Combinator, YC, and as alumni, we can track, like, what kind of businesses are applied into the batch.

最近几批，对冲基金创业公司的数量在增加，因为大概十八个月前，我从来没在YC毕业班里见过这种创业公司，但最近十八个月，现在每批都能看到，比如AI对冲基金。

Last few batches, numbers of the, hedge fund startup is increasing because I never saw that kind of startups applied in the YC graduating before probably like eighteen months ago, last eighteen months, but now we see like every batch, I see there's like AI hedge fund.

所以这本身就说明，你知道，新……的数量在增加。

So that itself shows that, like, you know, amount of the numbers of the new...

我不知道该怎么称呼它，叫对冲基金肯定是在增加的。

I don't know how you call it call it hedge fund is definitely increasing.

但是，再说，什么是对冲基金的定义呢？

But, again, like, what is the definition of a hedge fund?

比如，如果你从朋友和父母那里筹了一些钱，那还算是对冲基金吗？

Like, if you raise some money from your, you know, friends and parents, is that still a hedge fund?

还是说，被称为对冲基金的门槛是管理资产超过2500万、1亿美元？这个界限也很模糊，因为以前运营这种机构要贵得多。

Or is that threshold of being called a hedge fund is assets under management more than 25,000,100 million dollars It's also very blurred because it used to be much more expensive to run that kind of institution.

你需要有真正的大型托管人，或者需要有很多支持系统、基础设施，一大堆东西，但现在起步变得极其便宜。

You need to have, like, real big custodian or you need to have, like, a lot of a support system, infrastructure, a bunch of the things, but everything's now becoming extremely cheap to start with.

所以我觉得，当你说对冲基金的数量时，这取决于定义。

So I think, like, when you say numbers with the hedge funds, it's based on the definitions.

但是，你知道，那些自认为是对冲基金的人，我觉得这个数量会继续增加，十年后会多得多。

But, like, you know, those people who think they're hedge funds, I think number of that will continue to increase, it's going be a lot more in ten years later.

YC，对冲基金，交易交易新实验室。

YC, hedge funds, trading trading neolabs.

你怎么看？

What's your take?

你知道，我……

You know, I...

对，我不想带个人偏见。

Yeah, not to insert my bias.

对我来说，想要竞争、拿风投的钱，然后去跟Jane Street、Sudell Securities、Jump Trading、Hudson River Trading这些公司竞争，这看起来不只是最不可能的任务，而且可能是在所有不可能的任务里，最不可能的那个。

To me, the prospect of trying to compete, raise VC money, and compete with the Jane Street, Sudell Securities, Jump Trading, Hudson River Trading, it seems like not just the most impossible task, but probably within impossible tasks, the most impossible within those impossible tasks.

你有什么想法？

What are your thoughts?

我其实没怎么深入想过这个问题，但我觉得你说得太对了。

I never really thought about that too deep about that, but I think that's so right?

而且，你知道吗，同时，那些你刚才提到的机构里的人，可能也在申请YC，对吧？

Like, you know, and at the same time, people who were at those institutions you just mentioned is probably maybe applying for the YC as well, right?

你懂我的意思吗？

You know what I mean?

所以做这件事的人可能是同一批人。但我认为最大的区别当然是，当你的规模足够大的时候，有些东西你能接触到，有些你接触不到。而且，这种情况会继续开放，门槛会越来越低。

So the people who's doing that maybe in the same individuals But I think the big difference, of course, is that when you're big enough in the scale, there are certain things that you can and you cannot access And again, that's going to continue to be opening up, and then the entry hurdle is going to go lower.

所以我觉得这种个人化的趋势，会导致实体数量或者对冲基金数量大幅增加。

So I think this phenomenon trend of the people becoming more individualized, which results in numbers of the entities or numbers of the hedge funds, I think, will increase significantly.

我其实和Robinhood的Vlad聊过这个，我们谈到代理商务或者MCP，B2C交易应用的接口，也必须要有这个MCP和代理层，就是因为这个原因。

And I was actually talking with Vlad of the Robinhood about this, and we're talking about how agentic commerce or, this MCP, the interface of the B2C trading app, also have to have this MCP and agentic layer is really because of that reason.

因为会有更多的个人创业者，这就需要更多自动化的金融服务操作。

Because it's going to be more solopreneurs, and then that requires more of the automated financial services actions and operations.

所以你需要让所有的金融服务工具都准备好迎接代理。

So you need to have all the financial services tools to be agent ready.

当代理准备好了，金融服务就无处不在。

When the agent's ready, financial services are everywhere.

你甚至不需要太多金融服务操作，就能运营一家公司，不管是什么类型的公司。

You even don't need much of the financial services operations when you are running a company, any types of the company.

可以是像对冲基金那样的公司，那也是公司的一部分，对吧？

It can be a company as a hedge fund that's a part of that, you know, company, right?

然后你可能在经营某种Shopify业务。

And then you're running some kind of a, you know, Shopify business.

你完全不用担心那些事情，对吧？

You you you don't need to worry about those things, right?

所以我觉得，我们就像……

So I think, like, we are...

我们都……

We all...

谁在运营金融服务，我……

Who's running the financial services, I...

我们确实看到这些实体和业务在增加和扩展，数量和规模都会大幅增长。

We definitely see this rise and then expansion of the entities and businesses, just numbers and quantities going to just expand significantly.

在深入探讨你如何构建未来之前，先问个问题。关于对冲基金行业，你觉得你的视角是否让你产生了一些与众不同的或者逆向的观点？

Real quick, before we tap into what are the ways you're building towards that future, On the hedge fund industry, do you think that your vantage point has led you to any differentiated or contrarian takes?

我想问的是，你对对冲基金行业有什么看法？

I guess what I'm trying to ask is, what do you believe about the hedge fund industry?

还有，更广泛地说，关于量化交易、竞争优势和这个行业的商业模式，你有什么看法是大多数人会反对的？

And, I guess, broadly quant trading, about competitive advantage, about the business of it, that most people would disagree with you on?

我认为AI和智能体不会给你带来优势。

I think that AI and agents will not give you edge.

很多人问我，既然智能体和AI能找出赚钱的方法，那对冲基金还有什么意义？或者每个人都能赚钱了吗？

And I think a lot of people ask me, Okay, because agents and AI will figure out how to make money, so what are the points of hedge funds or, like, everyone can now make money?

我完全不同意这种观点。

That's what I completely disagree.

AI智能体、大语言模型，这些都只是技术。

AI agents, LLM, it's just technology.

技术在不断进步。

Technology continue to evolve.

但执行这些需要人来操作，因为钱最终是属于你这个人的。

But executing that requires execution by human because the money belongs, at the end of the day, you as a human.

你得交税。

You have a tax.

智能体不交税。

Agents don't pay tax.

你才是那个要负责交税的人。

You're the one who's liable for tax.

所以……

And so...

这意味着必须有人来执行。

Which means a human have to execute.

而这个因素对策略运行结果的影响被大大低估了。

And that element is underestimated so much about how much of it impacts the results of running strategies.

所以我不同意那种认为AI和智能体能给你带来赚钱优势的说法。

So I I I disagree about the, you know, AI and agents will give you an edge over how to make money.

这会是个...

This is gonna be a...

这是个很难的问题。

This is a hard question.

但如果是这样的话，为什么所有交易公司都在花巨额资金在计算和数据中心上，比如Jane Street和那些pod shops。

But if that's the case, why are all the trading firms spending exorbitant amounts of money on compute, on data centers, Jane Street, the pod shops.

看起来他们就像打开银行账户，把钱往这些上面扔。

It seems like they're just opening up their bank accounts and throwing cash at these things.

这是为什么呢？

What's the reason for that?

因为他们已经是执行这些事情的专家了。

Because they're already pros of executing those things.

所以继续投资于创造智能、代理、规则和寻找优势，这才是他们持续投入的方向，对吧？

So creating the intelligence and the agents and the rules and finding the edges, that's what they're continuing to invest in, right?

因为他们有严格的纪律和固定的路径。

Because they have a tight root and discipline.

这就是他们能够成长的原因。

That's the reason why they are able to grow.

就像普通业余爱好者，大多数普通人，我觉得你知道，我们很难做到有纪律地执行。

And just like regular amateurs, most of the regular humans, I don't think you know, we are able to execute with a discipline.

在我看来，这本身就是一门专业。

That itself is a profession in my opinion.

就像拳击一样。

It's like boxing.

比如，你知道的，那个……

Like, you know, the...

我觉得读，你知道的，那个……

I think reading, you know, the...

我觉得就像《创业维艰》，本·霍洛维茨的那本书。

I think it's, like, Hard Things About Hard Things, Ben Horowitz book.

当企业家或CEO并不是人类天生的姿态。

Being an entrepreneur or CEO is is not a natural posture of the human.

就像你得学着去摆出那种姿态。

Like, you have to kinda learn to take the posture of it.

就是那种……

It's that's...

作为人类来说，这太不自然了。

It's so unnatural as a human.

他讲的是拳击，你怎么摆出拳击的架势。

And he was talking about the boxing, How you, you know, do the form of boxing.

这对人类来说太不自然了，就像动物一样。

This is so not natural as a human being, like, an animal.

我也有同感，因为执行那种策略，或者按照那些看似行不通但统计上、逻辑上应该有效的规则去做，你能不能继续坚持执行，即使胜算不大……

And I I feel the same way because, like, executing that strategy or, like, rules by things that's not working but statistically, logically, it should be working can you can you like you know continue to with the discipline execute even the odds...

所以就像赌博，对吧？

So same as gambling, right?

就像你相信统计，逻辑上应该有效，但大多数时候它不奏效，也许它会回来，但谁知道呢，这太不自然了，所以这意味着，在我看来，能够那样执行本身就是极致的专业。

Like you believe in the statistics logically it should work but most of the life it doesn't it may come back but who knows that's so unnatural so which means that that itself is, in my opinion, extreme profession to be able to execute that way.

所以这就是为什么我刚才说的那些机构，他们能把这当成一门专业来做。

So that's the reason why, like, all those institutions I just said, they know how to do that as a profession.

所以，你知道，现在他们得继续努力，在竞争中保持和提升自己的优势。

So, like, you know, now they have to continue to fight and improve their edges against their competitors.

所以你会说，因为他们已经……

So you would say that because they've...

因为他们在运营上虽然不是完美，但已经很出色了，他们会寻找任何能改进流程的小细节，投资于前沿模型，也就是所谓的基础设施，这就是他们正在做的。

Because they are operationally not perfect, but excellent, They look for any little bit that can improve certain processes and investing in frontier model, call it infrastructure, is what they're doing.

然后，当然，他们的竞争对手也在做同样的事情。

And then, of course, their competitors are doing the same thing.

这会影响市场的运作方式，然后他们还得理解这会对他们自己那些小优势产生什么影响，然后他们就得在新的市场动态中继续应对，但这一切都取决于那个基础，也就是你确实能按照那套规则去执行的预期。

That's going to impact how the market works and then they also have to understand how that's going to influence their own certain small edges and then they're just going to kind of have to do it against the new market movement as well but again it all depends on this foundation that expectation that you can actually execute based on that rule.

我觉得我们大多数人，虽然我们心里知道从统计上看这行得通，但我不认为我们能真正做到。

I think most of us, again, I don't think we can, even though we kind of know we know statistically it's gonna work.

执行完全是另一回事。

Like, execution is a very different game.

我现在想听听关于……

I wanna hear now about...

你刚才谈了一些你对对冲基金行业未来的看法，你觉得对机构、散户等等会怎么发展。

So you've talked a little bit about how you think the future of the hedge fund industry is gonna look like, how you think it plays out for institutions, for retail, etcetera.

你最期待以哪些方式去构建那个未来？

What are the ways in which you're most excited to be building towards that future?

对我们来说，我个人，我相信的是什么呢？

For us, I think personally, me, I believe in what is that?

自由在很多方面都非常重要，我知道这听起来很抽象，但我们公司的使命是向地球上的每个人开放金融服务。我知道这听起来有点无聊，但对我来说真的很重要，意义重大。我在日本和美国之间长大，你所在的位置决定了你能接触到某些东西，而如果你不在那里，你就接触不到。比如Robinhood推出他们的应用时，我好像总是提到Robinhood，但我真的很喜欢那个应用，它在2013年或2015年推出时，我当时在日本，我在等待名单上，但我甚至无法安装那个应用，因为那时我不是美国居民。我知道这是件小事，但当我想到我的生活时，比如即使是Spotify或Apple Music，你也需要一张信用卡。

Like freedom is very important in many ways I know it's so abstract conversation but our company's mission is open financial services to everyone on the planet and I know it sounds kind of boring without thinking about it but to me it's really important like it means a lot I grew up between Japan and The US and based on where you are as a location you have certain access to certain things and you don't if you're there for example when Robinhood launched their app again, like, I feel like I keep bringing up Robinhood but, like, you know, I love the app and then launched in 2000, I think, thirteen or fifteen I was back in Japan I was on the waitlist but I couldn't even install the app because I was not an American resident at that time I know it's such a small thing but when I think about my life for example even Spotify or Apple Music you needed to have a credit card.

账单地址必须在美国。

The billing address was in The United States.

如果你不是那里的居民，你没有那种信用卡，所以你不能使用Spotify或Apple Music。

And if you're not resident there, you don't have that kind of credit card, so you can't use Spotify or Apple Music.

这本身就是一种不自由。

That itself is so not freedom.

这太不公平了。

It's so not fair.

所以对我们来说，我希望我们能消除基于地理位置、出身、身份地位等带来的不公平。只要基础设施到位，任何人都能用，这样就能给每个人真正公平的准入和机会。

So like for us is that I want to I want us to remove the unfairness based on the location or based on like where you're from like you know what kind of person you are status anything as long as infrastructures are there if you can use it it's going to give really fair access and opportunities for everyone.

所以回到你的问题，这对我们来说非常重要。

So going back to your question, that's really important to us.

我们要确保所有基础设施都不是排他性的。

So making sure all the infrastructure is not really exclusive.

不管你在哪里，处于什么环境，都应该让尽可能多的人使用。

It can be used by as many people as possible regardless of the locations or, like, environment where you are in.

这还包括你最初能存入多少钱。

It it includes, like, how much money you can deposit initially.

所有这些都会造成排他性。

All those things create exclusivity ness.

所以我们想完全反过来，让服务变得人人可用。

So we we want to be completely opposite of that, to be, you know, usable.

然后，当然，像大型机构也可以使用……

And then, of course, like, big institutions can be use...

它们也能用，但这必须得尽可能……

Using that too, but, like, that has to be extremely as...

你知道，尽量降到最低。

You know, get lower and lower as possible.

所以这就是……

So that's...

我个人很期待的事情。

I'm personally excited about.

因为个体创业者也能成为真正的大型对冲基金，但必须有一个起点，让人不需要非得有某种背景，比如‘我在Citadel干了十年’这样的经历。

Because solopreneurs can be real big hedge fund, but there has to be some starting point that the person doesn't have to come with a certain background of, okay, I worked at the Citadel for ten years.

那么你会说，在所有的……

And so would you say that across all...

在你业务的每个部分，嗯。

Across each part of your business Mhmm.

你是想把世界打开吗？

You're trying to open the world up?

所以比如说交易方面，就是给零售交易者提供基础设施来执行策略，可能达不到顶级量化基金的水平，但至少能在一定程度上进入市场。

And so say for trading, it's retail traders giving them infrastructure to execute strategies, probably not the same level as a top quant fund, but at least access to the market to some extent.

这就是你对自己正在打造的业务的想法吗？

Is that how you think about the business you're building?

是的，因为即使我们在2018年推出这个股票交易API时，我们这么做的原因就是没有一个真正可以一键使用的交易API。

Yeah, because even when we launched this stock trading API 2018, The reason why we did it was that there was no trading API that's actually usable just with a click of the button.

如果你……

If you...

我觉得当时Interactive Brokers可能有API，但它不是真正的REST API。

I think they had I think at that time, Interactive Brokers may have had an API, but it's not really REST API.

有一些固定的API，然后你每天晚上都得刷新。

There's, like, some fixed API, and then you have to refresh it every night.

然后还有，你知道，其他的……

And then there was, like, you know, other...

我不记得了。

I don't remember.

比如有些其他券商，你得先申请，因为要填一大堆文件才能拿到API文档。

Like, some other brokers, you have to apply for the application because, like, you have to fill in this whole documents to just get this API document.

如果是API的话，API文档应该直接就在那里。

Like, if it's API, API document should be there.

你不应该需要申请才能拿到API文档。

You shouldn't have to apply to get the API document.

这样就造成了排他性，意味着你得有足够的资源才能去申请。

That creates exclusivity, and which means that you have to have a certain resource enough to apply for it.

这本身就让我挺烦的。

That itself, to me, kind of frustrates me.

为什么不直接给人们用，让它被使用呢？而且API交易不是那样的，所以这是我们想要改变的。当然，那时候会碰API的人非常非常少，因为你得会写代码，而且你得有非常明确的使用意图。

Like why don't you just give it to people just let it let it be used and the API trading was not like that so that's something that we wanted to change of course like at that time people who's going to touch the API is very very limited because you have to be able to code, and you have to have a very intentional way that you want to use it.

现在完全不一样了，因为有了大语言模型，每个人都是开发者。

Now it's completely different, because everyone is a developer because of the LLM.

是的，所以我们就是这样开始的，也是我们思考问题的方式。

Yeah, so that's how we started it and how we think about it.

你看到现在系统化交易的人数激增了吗？

Are you seeing an explosion in the number of people who are systematically trading today?

哦，是的，绝对没有。

Oh yeah, definitely not.

嗯，跟Anthropic和Alpuana AI的增长相比，你知道，他们的收入增长了十倍，疯狂得很。相比之下，我觉得很尴尬，因为你知道，月活跃开发者数量，我想是从……

Like, well, compared to, like, you know, how the Anthropic and Alpuana AI is growing, what is it, 10x of the revenue, like every Crazy crazy Compared to that, I feel, like, you know, very embarrassed because, like, you know, the number of the monthly active developers, I think it's been growing from...

那是多少？

What is that?

嗯，是的，大概两万月活跃开发者，六个月后超过八万。

Like, yeah, like, 20,000 monthly active, you know, developers, like, 80,000 plus in, like, six months.

六个月翻了四倍。

Like, 4x in six months.

那真是……

That's...

对我来说，这数字真是大得惊人。

To me, like, it's crazy big number.

说实话，这比Anthropic还多。

Honestly, that's more than anthropic.

你知道吗？

You know?

但Anthropic是多少？

But anthropic is, what?

十倍，十倍。

Ten ten x.

像这样，几年时间，跟那个一比，我有点不好意思。

Like, this is, like, like, like, few years, like, always compared to that, I feel a little bit embarrassed.

但对我们来说，哇，这增长真是爆炸性的。

But to us, it's like, wow, it's an explosive amount of the growth that we're seeing.

我很好奇，这些增长是不是集中在某些特定的地区？

I'm curious, are there any particular geographies where that's concentrated?

目前来看，主要还是在美国。这其中的一个原因是，在美国注册公司比较容易，即使你不是美国人，也可以在美国注册公司，然后被视为美国公司、美国企业，对吧？

Right now, I think it's still The United States, and that's also a part of the course the part of the reason is that it's much easier to spin up the entity and even though you may not be an American but like you know you can spin up the entity in The United States and you could be considered American entity American business, right?

就像我一样，我仍然是日本人，所以……

So the same as me, like, I'm still Japanese, so...

我问这个问题的原因是，跟人交流时，我觉得整个世界都在朝着一个方向发展，就是人们想要冒险、想要交易、想要抓住那个右尾的机会，对吧？

The reason why I asked that question is because speaking to people, Overall, I think the world is headed in the direction where people want to punt, people want to trade, people want to hit that right tail, right?

有趣的是，某些文化似乎更

And it's funny, certain cultures seem to be

倾向于

more

追求这个。

predisposed to pursuing that.

我举个例子，比如韩国。

And I'll give the example of Korea.

在韩国，有一种社会契约，你很难丢掉工作，工作保障很高，但可能买不起很好的房子。

Korea, you have a social contract where very hard for you to lose your job, a lot of job security, probably won't be able to buy a very nice house.

所以你可以放手一搏，风险几乎为零。

So you can punt it, the downside is near zero.

最坏的情况，你还是有工作。

Worst case, you still have your job.

你还是有人照顾的。

You're still taken care of.

所以我想问的是，我很好奇Alpaca在那里是不是增长很快，因为对我来说这很合理。

And so the point I was getting at was I was curious if Alpaca was seeing lots of growth there because it would make sense to me.

我觉得总体来说……

I think that's in general that...

不过顺便说一句，这真的很有意思，我刚才在想……

But by the way, it's so interesting, like, I was just thinking that...

所以工作保障会让人均交易量增加。

So job security creates more trading volume per person.

这是你的类比，对吧？

That's your analogy, right?

对。

Yes.

我确实想通过问我的GPT或量化模型来获取真实数据。

I definitely want to, like, get the real data by asking my, you know, GPT or quant.

因为那会非常有趣。

Because, like, that would be so interesting.

因为我觉得土耳其也是以大量交易闻名的。

Like, because I think Turkey is also known for a lot of trading too.

但那里的人有工作保障吗？

But, like, is that job security there?

我不知道。

I don't know.

日本，我们以前有。

Japan, we used to.

我们有个说法，字面意思是渡边太太，对吧？

We had the which is called which literally means like it literally means it's a Watanabe son's wife, right?

但它的意思是，日本的家庭主妇因为太无聊了，而且她们的丈夫白天要上班，所以她们用高杠杆交易美元日元，曾经因此影响过市场。

But what it meant was that housewives in Japan because they are so, I don't know, bored and their husbands, you know, work during the daytime, so they trade dollar yen effects with a massive leverage, so they used to move the market as a reason.

太疯狂了。

It's crazy.

再说，考虑到现在的社会讨论，这真的是个很糟糕的例子，因为日本总是男人工作，妻子待在家里做别的事情，对吧？

And again, this is such a very bad example considering current social conversation because Japan is always like men work, and then wife just stays at home and just do some other stuff, right?

然后她们被认为很无聊，所以就成了渡边太太。

And then they're considered bored, so they become Mrs.

去交易美元日元。

Watanabe and trading dollar yen.

实际上，做家务比在办公室上班喝啤酒还要辛苦。

In reality, housework is even more harder than working at the office and drinking beer.

但这真的很有趣。

But that's super interesting.

我觉得交易量的增长当然来自某些特定领域。

I think trading volume growth comes from the, of course, certain areas.

不过具体来说，那些量化交易，我倒没觉得有哪个国家特别突出。

Like specifically though, however, those quant trading, I don't really see there's specific countries that grow specific.

因为我觉得你提到韩国，那里的交易量很多是散户在频繁买卖。

Because I think where you're talking about South Korea, trading volume is a lot of retail click click trade, buy and sell.

对。

Yes.

但说到自动化交易，我觉得情况就不太一样了。

In terms of like more of the automated trading, I think it's a little bit different story.

你知道吗，在录这期播客之前，我走在纽约去录音室的路上，就在想我想问些什么问题。

You know, before this podcast, I was walking to the studio in New York, and I was thinking about the sort of things I'd like to ask.

我想到的一个问题就是，你是日本人，却在做这种业务。

One thing that genuinely came up is the fact that you're Japanese building this sort of business.

我去过日本很多次。

I've been to Japan many times.

我很高兴。

I'm glad.

那是个很棒的地方。

It's a wonderful place.

我研究过教育体系，也研究过人们成长的方式，按理说他们不应该做你现在做的事情。

When I study the education system, study the way people are brought up, they are not supposed to be doing what you're doing.

你知道吗？

You know?

这太不可思议了。

It's wild.

这怎么会发生呢？

How did that happen?

嗯，想想看，我其实不喜欢……

Well, think, like, I didn't like to be...

嗯，我想我尽量对自己的感受和那种挫败感保持诚实。

Well, I guess I try to be honest with my feeling and that frustration.

我觉得那真的是我精力的来源，而且那种感觉就是这不公平。

I think that's really motivation of my energy, and that's really feeling this is not fair.

我觉得这真的一直推动着我，我就是不想待在人多的地方。

I think that really drove me always, and I just didn't want to be in the place where a lot of people are.

所以就是这样……

So that's just...

既然如此，很明显你的思维方式很不一样，你觉得是什么因素导致了这种不同？

And within that, because clearly you think in a very differentiated way, what are the inputs that you think led to that differentiation?

我问这个问题是因为，虽然我们是一个做对冲基金量化交易的播客，但我喜欢用优势和竞争力来框架一切。

And I ask the question because, we are a hedge fund quant trading podcast, but I like framing everything in terms of edge and competitive advantage.

即使在故事里也一样。

Even within story.

明白了。

Got it.

甚至有人说，比如乔布斯，对，有道理，他是个智商150的嬉皮士，显然他会想出一些与众不同的东西，不管是好是坏，是天才还是巧妙，因为他的输入和经历都太独特了。

Even people say like Steve Jobs Yeah, makes sense He is a 150 IQ hippie, and it's like clearly he was going to come up with something differentiated, good or bad, genius or ingenious, because his inputs were so differentiated and his experiences were so differentiated.

所以我也问你同样的问题。

And so I'm asking the same question to you.