Max Junestrand: The Willingness to Learn Faster Than Anyone Else
打开互动全文版（中英对照 + 朗读 + 问答）→Max Junestrand 分享 Legora 如何从被 YC 拒绝成长为服务全球超过 3% 律师的法律 AI 平台。
Max Junestrand shares how Legora grew from a YC rejection to a global legal AI platform serving over 3% of the world's lawyers.
这比我们第一次来 Y Combinator 时受到的欢迎要热情得多。
This is a much warmer welcome than what we had coming into Y Combinator the first time.
Legora 的 YC 之旅始于一次拒绝，但这个故事稍后再讲。
Legora's YC journey started with a rejection, but more on that story later.
在 YC，他们会教你一个东西，叫做两句话描述。就是用两句话描述你的公司。Legora 是律师的智能体式操作系统，它从头到尾处理复杂的法律工作，让律师能够取得前所未有的成就。现在全球超过 3% 的律师都是 Legora 的活跃用户。等你被 YC 录取后，你会了解更多。
In YC they teach you something and it's called the two sentence description. It is describe your company in two sentences. Legora is the agentic operating system for lawyers and it handles complex legal work from start to finish so lawyers can achieve more than ever before. And over 3% of all the world's lawyers are now active users of Legora. You will learn more about this when you get accepted into YC.
当我们开始关注法律行业时，感觉这是将大型语言模型应用于解决实际问题的最佳机会之一。而我们面对的软件看起来像是 90 年代构建的。作为软件工程师，我们被优秀的软件宠坏了，因为我们喜欢为彼此、为自己构建软件。但在法律行业，这就是软件过去的样子，也是我们起步时所面对的。
When we started looking at the legal industry, it felt like one of the biggest opportunities to apply large language models to solve real problems. And the software that we were up against looked like it had built in the '90s. As software engineers, we have been incredibly spoiled with great software because we love to build for each other and for ourselves. But in the legal industry, this is what the software used to look like and what we were up against when we started out.
现在，Legora 已经成长为一个全球品牌，把裘德·洛这样的名人带到了纽约华尔街。但我仍然觉得我们才刚刚开始。
Now, Legora has grown into a global brand, putting celebrities like Jude Law at the New York Wall Street. And still I feel like we are only getting started.
我在 Legora 最喜欢的 Slack 频道是“客户之爱”。我们的客户会在这里发布他们觉得没有我们的系统就无法实现的事情。我们从许多销售电话中挖掘出这样的引述。我想和你们分享几个，因为我认为归根结底，这才是最重要的。
My most favorite Slack channel in Legora is customer love. It is where our customers post things that they sort of truly feel like they couldn't have achieved without our system. And we mine many of our sales calls for quotes like these. And I wanted to share just a couple with you because I think at the end of the day, this is what it's all about.
在一次比赛中，一位家长问我如何兼顾工作和所有的高尔夫旅行。一个词：Legora。从想法到执行的道路从未如此短暂。从未，伙计。如果 Legora 明天消失，我就回去执教高中篮球。
A parent asked me at one of the tournaments how I juggle work and all of the golf travel. One word, Legora. Never has the path between an idea and execute been so short. Never, man. If Legora went away tomorrow, I just go back to coaching high school basketball.
当 Legora 进入 YC 时，我们甚至被质疑在法律行业销售和工作的想法。这是一个如此保守的行业，以前从未对软件宽容过。但从 2024 年 10 月正式发布到上个季度末，我们的 ARR 从 100 万增长到 1 亿，从瑞典的三名工程师起步，发展到全球超过 750 人的公司。这太棒了。而且我们从欧洲起步做到这一点，我认为更是锦上添花。
When Legora went into YC, we were questioned about the idea of even selling and working in the legal industry. It is such a conservative business that had never been sort of lenient to software before. But from the time that we went into GA of October 2024 until the last end of quarter, we've grown from 1 to 100 million in ARR and starting from just three engineers in Sweden to a company of over 750 people all over the world. This is awesome. And the fact that we've gotten to do this starting from Europe, I think just adds another flare in that hat.
但 Legora 的故事差点就没发生，而且实际上并不是从我开始的。事实上，它的开始几乎像一个笑话的开头。2020 年，在 GPT 之前，在 AI 变得流行之前，一位律师、一位物理学家、一位工程师和一位心理学家创办了一家名为 Judelica 的公司。他们观察到，法学院学生去实习基本上就是总结法庭案例。借助谷歌早期的 BERT 模型，他们开始探索这个领域可能是什么，以及可以构建什么。
But the Legora story almost didn't happen and it actually didn't start with me. In fact, it started almost like the great beginning of a joke. And so a lawyer, a physicist, an engineer, and a psychologist back in 2020, before GPT, before AI was cool, started a company called Judelica. And what they observed was that law school students were going into their internships basically to summarize court cases. And with the early models from Google called BERT, they started to explore what this space could be and what was possible to build.
我的故事始于瑞典群岛的一个岛屿，看起来像这样。我在一场排球比赛中遇到了我的两位联合创始人。我认为创业创始人的故事可以来自许多不同的地方，这绝对不是一个高概率的起点。但当我遇到 August 和 Siga 时，他们向我展示了他们构建的产品演示。我们的群聊实际上看起来就像这样。他们给了我项目的演示，用 GPT-3.5 他们构建了一个非常简单的产品，可以解释股票期权协议到底意味着什么。于是我们就这样开始了，我说这是一个非常酷的演示，一个非常酷的项目，我的联合创始人 Siga 说我很高兴你愿意帮忙。结果我帮忙的远不止这些。
My story started in the Swedish archipelago at an island which looks something like this. and I met my two co-founders at a volleyball game. I think startup founder stories can come from many different places and this definitely wasn't one, you know, high odds of a place to start. But when I met August and Siga, they showed me a demo of the product they had built. And our group chat actually looked exactly like this. They gave me a demo of the project and with GPT-3.5 they had built a very simple product that could explain what a stock option agreement really meant. And so we started with that and I said this is a super cool demo, a super cool project and my co-founder Siga said I'm glad you want to help out. And I ended up helping out a little bit more than just that.
当 GPT-3.5 出现时，那是我们这一代的互联网时刻。我从大学辍学，从未完成硕士论文，因为不构建的机会成本变得太大了。但当我们起步时，我们并不明显会成功。因此，为了了解法律行业是什么、意味着什么、律师实际上如何工作，我们想出了一个巧妙的计划：从律师网站上找到他们的邮箱，给他们发冷邮件，并在 LinkedIn 上给他们写信，邀请他们共进午餐，以便了解更多关于他们不同执业领域的信息。我们提出支付他们的小时费来与我们共进午餐。许多人接受了这个提议。但幸运的是，许多人没有让我们支付小时费，许多人甚至支付了午餐。
When GPT-3.5 came, that was the internet moment of our generation. I dropped out of college. I never finished my master thesis because the opportunity cost of not building had become too large. But when we started out, it wasn't obvious that we were going to be a success. And so to learn about what the legal industry was, what it meant, how lawyers actually worked, we came up with the ingenious plan of cold emailing lawyers on their emails that we could find on their websites and writing to them on LinkedIn asking for a lunch where we could learn more about their different practice areas. And we offered to pay their hourly fee to have lunch with us. And many of them took us up on that offer. But fortunately for us, many of them didn't make us pay for the hour and many of them even paid for the lunch.
当我们站稳脚跟后，我们联系了北欧最大的律师事务所之一 Mannheimer Swartling。但就在两年前，他们的管理合伙人在电视上说，AI 在他们开始使用时，实际上更多是人工而非智能。因此，我们必须非常努力地改变人们对技术和 AI 在法律领域可能是什么的看法。在法律行业，事情顺利时你不会得到报酬，事情出错时你会受到惩罚。该领域许多早期的 AI 解决方案需要大量的训练。你必须带来大量的用例和示例，才能得到非常非常少的结果。当 GPT-3.5 出现时，这真正改变了局面。于是我们与 Mannheimer Swartling 合作，最终搬进了他们的办公室，与他们紧密合作。
And once we had gotten our feet under ourselves, we approached one of the largest law firms in the Nordics named Mannheimer Swartling. But just two years before that, their managing partner had gone on TV and said that AI back when they were starting to use it was really more artificial than intelligence. And so we had to work really hard to change the perception of what both technology and what AI could be in the area of law. In law, you are not paid when things go right, you are punished when things go wrong. And many of the initial AI solutions in the space had required immense amount of training. You had to sort of bring a lot of use cases and examples just to get very very little results. And when GPT-3.5 came, that really changed the story. And so with Mannheimer and Swartling, we ended up moving into their offices and moving and working really close with them.
正如我一开始所说，这里的热情欢迎与我们第一次与 YC 互动时截然不同。我们在五月申请，承诺能够使用大型语言模型查询所有法律文件。当我们去面试时，我们被问到的第一个问题是：你们服务哪种类型的律师？你知道，当时我们对法律领域还不太了解。我们的回答是：“你什么意思？律师还有不同类型吗？”当时 YC 的合伙人对此并不满意。我痛苦地记得，当我们给出那个回答时，Tom Bloomfield 在面试中开始大笑，我们知道我们完蛋了。但这也给了我们坚持下去的毅力和动力。所以，在经历了很多失望之后，第二天我们没有工作，但第三天我们又重新开始构建，两个月后，我们以新的名字和不同的平台被录取了。我们做了功课，我们完成了工作。从瑞典来看，被 Y Combinator 录取感觉就像我们要去奥林匹斯山。我们要与神共餐，我们要与来自所有我们知道的伟大科技公司的其他创业创始人一起工作。当我们到达 Y Combinator 时，我们意识到我们实际上知道的比我们想象的要多得多。这是我最喜欢的照片之一。
And as I said in the beginning, the warm welcome here was quite different from the first time we interacted with YC. We applied in May with a promise of being able to query all legal documents using large language models. And when we came to our interview, one of the first questions that we got asked was what type of lawyers are you serving? And you know, not very knowledgeable about the legal space yet. Our response was, "What do you mean? Are there different type of lawyers?" And that did not land very well with the YC partners at the time. And so I actually remember very painfully that Tom Bloomfield started laughing in the interview when we gave that response and we knew that we were cooked. But that also gave us the perseverance and the drive to continue. And so with a lot of let down the day after we did not work but the day after that we got back to building and two months later we got accepted under a new name and with a different platform. We had done our homework. We had done the work and from Sweden the idea of getting accepted into Y Combinator felt like we were going to Mount Olympus. We were going to dine with the gods and we were going to get to work with other startup founders who came from all of the great tech companies that we knew of. And once we got to Y Combinator, we realized that we actually knew a lot more than we thought we did. And this is one of my favorite pictures.
这是我在第二次 YC 面试后打完视频电话时截的图，当时我们知道情况比第一次好多了。所以在我们被 Y Combinator 录取后，我们搬进了我之前提到的那家律师事务所。Mannheimer Swartling 是北欧最大的律师事务所，但他们真的不知道如何与软件和技术打交道。我们最终得到了一间没有真正窗户的会议室，空调在下午 5 点左右就会关掉，但离可乐冰箱很近。所以每天下午 6 点左右，会有一名工程师走过去，抓住门来回扇动，让更多氧气进入房间。那些会议椅至今还在我们所有人的背上留下了印记。我想 3 年后，我坐在车里，带着两个额外的屏幕，那时我们从律师事务所的会议室搬到了我们第一个真正的办公室。我认为从整个经历中得到的关键启示之一是，首先，我们很快学会了不要接受‘不’作为答案。但也要学会主动寻求帮助的力量。我认为在打造 Legora 的头三年里，一个重要的教训是，作为一个个人和一家公司，让别人希望看到你成功、希望帮助你，是非常强大的。
And it is the screenshot I took after we did the video call after our second YC interview when we knew that things had gone a lot better than the first time. So after we got accepted into Y Combinator, we moved into this law firm I mentioned earlier. Mannheimer Swartling was the biggest law firm in the Nordics, but they had very no real clue about how to work with software and how to work with technology. And we ended up getting a conference room with no real windows where the AC would turn off at about 5:00 p.m., but it was really close to the Coke fridge. And so at about 6 p.m. every day, one of the engineers would go up, grab the door, and start waving it back and forth to get more oxygen into the room. And those conference chairs still made an imprint on all of our backs. I think 3 years later and me in the car there with two more screens is when we make the move from the conference room of a law firm to our first real office. And I think one of the key takeaways from this entire experience is that first off, we quickly learned not to take no for an answer. But also the power of simply reaching out and asking for help. I think one big lesson from these first three years of building Legora has been that it is very powerful to be a person and to be a company that others want to see succeed and that others want to help.
所以我们来到了旧金山，在伯纳尔高地的一个 Airbnb 租了一个房间，开始了我们的建设。在 YC 期间，我们的 ARR 从零增长到一百万美元。我们也艰难地认识到，在旧金山和欧洲之间跨时区工作对团队造成了巨大负担。我找不到这张照片，但我买了一个网红环形灯放在笔记本电脑上。我之所以有那个环形灯，是因为我每天早上 1 点到 10 点都在做销售电话，从旧金山打给欧洲，因为那是我们最初的客户群所在地。
And so we came to San Francisco, we got a room at an Airbnb up in Bernal Heights and we started building. And during YC, we grew from zero to a million dollars in ARR. And we also learned the hard way that working across San Francisco and Europe took a big toll on the team. I couldn't find a picture of this, but I bought an influencer ring light that I put on my laptop. And the reason I had that ring light was because I was doing sales calls between 1:00 a.m. in the morning to 10:00 a.m. in the morning every single day in San Francisco to Europe because that was where our initial customer base was.
在 YC 期间经历了火箭般的增长轨迹后——现在两年后，与一些公司相比这已经算慢了——我们开始了融资。这是我第一次融资。我完全不知道会发生什么，但我们在一个半星期内安排了 80 场投资者会议。我们有欧洲投资者、美国投资者。那次经历给我的主要收获是，你需要成为一个非常会讲故事的人，才能激发兴趣并让人们加入你的团队。与一些欧洲投资者和与美国投资者交谈也是非常不同的体验。我觉得在美国，我们更多地考虑可能出什么好结果，而不是可能出什么错。我们最终与一家名为 Benchmark 的风险投资公司合作。这是我们的合伙人 Chetan 在我们签署投资条款清单时的照片。我记得在他的桌子上谈判那份条款清单，电脑开着，在 Excel 里精确到小数位来确定股份数量，他们最终投资了 951 万美元。
And after a rocket trajectory during YC, which now two years later is slow compared to what some of the companies are doing, we set off to fundraise. And this was my first time fundraising. I had no idea what to expect, but we had lined up 80 investor meetings in a week and a half. And we had European investors, American investors. And the main takeaway from that experience for me was that you need to become a really good storyteller in order to drive interest and to get people on your team. It was also a very different experience talking to some of the European investors versus some of the American investors. I feel like in the US we think more about what can go right rather than what could possibly go wrong. And we ended up working with a venture firm called Benchmark. This is a picture of Chetan, our partner, when we had signed the term sheet. And I remember negotiating that term sheet on his desk with a computer opened going on decimals in the Excel to determine the exact number of shares and they ended up investing $9.51 million.
在那次投资之后，Redpoint 在三周后抢先领投了我们的 A 轮，然后我们开始遇到挑战。我们银行里有大约 3500 万美元。团队只有 10 个人。实际上，之后有一个月，我们从这 3500 万美元的利息中赚的钱比从客户那里赚的还多。当你变成一家银行时，这可不是什么好迹象。在第一次董事会会议上，我们有 Chetan 和 Redpoint，还有我和我的联合创始人。我们做出了一个非常艰难的决定，那就是冻结我们的销售动作。因为当你与律师合作时，你只有一次机会做好。如果你出现，产品不工作，或者系统延迟太高，或者当你的 Azure 实例流量太大时系统宕机，你就完蛋了。在那六个月的销售冻结之后，我们才真正开始爆发。那就是你看到从 100 万到 1 亿的曲线真正开始的地方。
Now, following that investment, Redpoint pre-empted our Series A just three weeks later, and we started to run into a challenge. We had about $35 million in the bank. We were 10 people in the team. And there was actually a month following that that we made more money from interest rates on those 35 million than we did from customers. And that is not a very good sign when you have become a bank. And going into our first board meeting, we had Chetan and we had Redpoint and it was me and my co-founders. And we made a very hard decision, which was to freeze our sales motion. Because when you work with lawyers, you only really get one chance to get it right. If you show up and the product doesn't work or if the lag in the system is too high or if the system goes down when there's too much traffic on your Azure instance, you are toast. And after that six-month sales freeze was when we really started ripping. And that's where you see that curve from 1 million to 100 million really starting.
在那开始时发生的事情是，你知道，从一家寻找 PMF 的小创业公司转变为真正建立我们的业务和产品。在 Legora 的早期，甚至在 Legora 之前，我们决定构建什么的方式是通过投票。我们会在整个团队中进行民主投票。对于任何曾经构建过软件的人来说，你知道如果厨房里厨师太多，通常做不出一道好菜。这正是当时的情况。我们在做太多功能。我们不得不为夏季后、销售冻结后的产品发布重建一切。我们必须构建一个平台，能够适应不断变化的基础，包括底层模型以及我们用来构建智能体工作流的框架，如 LangChain 等。所以在 2024 年 10 月，我们写下了我们称之为 Legora 产品宣言的东西。这基本上是把我们学到的一切收集起来，放进一份非常简单的文档中，与当时整个 25 人的团队分享，并以此团结大家。当时，我们的年度经常性收入约为 130 万美元，而我们许多竞争对手的收入是我们的 10 倍，但他们的产品只有这些功能中的一个。我也认为这是从瑞典起步的结果。市场不够大，无法容纳真正的独角兽或十角兽或更大的公司。所以我们从 YC 得到的最早的问题之一就是，你们什么时候搬到美国？什么时候跳过去？正是在这次产品重新聚焦之后，我们不仅在产品上超越了竞争对手，还积累了足够的势头跳到美国，并开始争夺我们市场中最大的客户。
And what happened at the start of that was, you know, turning from a small startup looking for PMF to really establishing what our business and what our product was going to be. In the very very early days at Legora, so even pre-Legora, the way that we would decide what to build was by voting. We would make democratic votes in the entire team. And for anyone who has built software before, you know that if you have too many chefs in the kitchen, that typically does not make one very good dish. And that was exactly what was going on. We were working on too many features. We had to rebuild everything for the product launch that we were going to have after the summer, after the sales freeze. And we had to build a platform where it was adaptable to foundations that were constantly changing both with the underlying models but also with the frameworks that we were using to build agentic workflows like LangChain etc. And so in October of 2024, we wrote down what we would call the Legora product manifesto. This was basically gathering everything we had learned, putting it into one very simple document that we shared with the entire 25 strong person team back then and that we would rally the troops around. At the time, we were doing about $1.3 million in annual recurring revenue, and many of our competitors were doing 10 times that with products that only had one of these features. And I also think that's a function of having started in Sweden. The market is not big enough to house a real unicorn or a decacorn or something even bigger than that. And so one of the earliest questions we got from YC was when will you move to the United States? When will you make the jump? And it was off the back of this product refocus that we were able to not only compete out the competition on product but build enough momentum to make the jump over to the US and start to compete for the biggest logos in our markets.
我认为作为工程师和技术人员，我们经常过分强调对产品和技术的依赖。但如果我在过去三年里学到了什么，那就是一切都在于人。构建产品是一回事，但构建一家公司是非常不同的事情。我们在这方面犯了许多早期错误。我们不得不摒弃的一种招聘模式是寻找简历上花哨的 logo，或者我们喜欢称之为 y 截距的东西。如果你考虑某人的技能曲线，它可能一开始很高，但如果他们没有良好的上升轨迹，他们在一家指数级扩张的公司里会很难工作。所以相反，我们开始寻找像我们一样的人，他们想疯狂工作，具有很高的增长潜力。我们一次又一次地加倍投资于这样的人才。我们今天的最佳销售员只有 23 岁，他卖出了超过 1000 万美元的 Legora。
And I think as engineers and as tech people, we often overemphasize the reliance on product and on the tech. But if there's anything I've learned in these past three years, it's that it's all about the people. And building product is one thing, but building a company is a very different thing. And we made many early mistakes around this. One of the hiring patterns that we had to unlearn was looking for fancy logos on résumés or what we like to call y-intercept. If you think about somebody's skill curve, it might start out really high, but if they don't have a good trajectory upwards, they are going to have a really hard time working in a company that is scaling exponentially. And so instead, we started to look for people like us who wanted to work insanely hard with a very high growth potential. And we doubled down on that talent over and over again. Our top seller today is 23 years old and he sold over more than $10 million worth of Legora.
他没有销售背景。他从大学辍学，通过不断加倍投入这类人才，我认为我们在 Legora 建立了一种非常独特的文化。其中之一就是引入和构建文化契合度。在我们大约 500 人之前，我面试了工程以外的每一位候选人。现在我面试每一位总监及以上级别的人。但我们所做的是，我们在业务中设定了文化基础，让它自我选择。我们在 Legora 有三个价值观：全力以赴、追求卓越、共同成长。它们合在一起就是 LFG。加油。这也向领导者和加入公司的其他人发出了一个信号，他们可能来自更成熟的背景，加入一家价值观中包含脏话的公司，很快就能告诉他们这是怎么回事。
And he had no sales background. He jumped out of university and by continuously doubling down on that type of talent, I think we have built a very unique culture at Legora. And one of those things has been about bringing and building cultural fit. Up until the point where we were around 500 people, I interviewed every single candidate outside of engineering. Now I interview every director and upwards. But what we have done is we have set the cultural foundation in the business where it self-selects for itself. We have three values at Legora: Lean in, fight for excellence, and grow together. And together they make LFG. Let's go. That also sends a signal to leaders and others who join the company who are coming from more perhaps established backgrounds where joining a company where the values includes profanity quickly sort of tells them what this is about.
在斯堪的纳维亚创业时，另一个非常复杂的文化现象是所谓的 Jantelagen 或詹代法则。这对我们的美国朋友来说非常陌生。但在瑞典文化中，这基本上意味着你不应该认为自己是个人物，或者你不比其他人更重要，你的想法也不比别人的更好。我认为这种谦逊的一部分是好的，它支持一种文化，让最好的想法胜出，让职业生涯早期的人感到有权力发言。但如果你认为你不是个人物，或者你不认为你比其他人更好，那么要建立世界上增长最快的企业公司之一是非常困难的。因此，我们必须真正创造一种美国、欧洲和亚洲文化的完美结合，以创建今天的 Legora 这样的全球公司。所以我真的认为，建立强大的公司文化是吸引和留住最优秀人才的最佳保证。
Another very complex cultural thing starting a company in Scandinavia is something called Jantelagen or the law of Jante. This is very foreign to our American friends. But in Swedish culture this basically means that you should not think that you are somebody or you are not more important than anybody else, and your ideas are not better than anybody else's. And I think a part of this humility is good and it supports a culture where the best ideas win and where people earlier in their careers feel empowered to speak up. But it's very hard to build one of the fastest growing enterprise companies in the world if you think that you aren't somebody or if you don't think that you are better than anybody else. And so we have had to really create a nice cocktail of US, European and Asian culture all in one to create a global company that is Legora today. And so I really think that building a strong company culture is the best guarantee to attract and keep the best people.
公司文化在 Legora 的另一种表达方式是通过像这篇发表在 Sifted 上的文章。对我们来说，只有赢。在赢家通吃或赢家拿大头的市场中，其他一切都是输，公司必须像没有第二名的地方一样运行。我认为可以说，当 Legora 成立时，它真的没有存在的理由。我们是众多法律 AI 公司之一。但由于这种心态，以及即使我们以任何正常标准衡量都取得了非凡的数字，我们仍然不觉得安全，我们继续像职业游泳运动员一样，低头看着自己的泳道，而竞争对手则侧目看着我们在做什么。在 Legora，我们必须建立一种文化来真正支持这种雄心。
Another way that company culture gets expressed at Legora is through articles like this one that ended up in Sifted. For us there is only winning. Everything else is losing in a winner takes all market or winner takes most market and the company has to run like there is no place for number two. I think it actually would be fair to say that when Legora was founded, it really had no reason to exist. We were one of many legal AI companies. But because of this mentality and because of not feeling like we're safe even though we have achieved extraordinary numbers by any normal measures, we continue acting like professional swimmers who are looking down in our lane while the competition is looking sideways at what we're doing. And at Legora we have had to build a culture to really support that ambition.
所以，我们比一些最大的竞争对手晚起步数年，必须在一个大多数人想在下午 3 点去 fika（意为吃肉桂面包和喝咖啡）或 5 点回家的国家建立这种强度。做到这一点的方法之一是创造文化价值观和文化时刻，真正将我们与许多其他人区分开来。所以在瑞典语中有一句话，你可以因为对某件事如此兴奋而醒来时尝到血腥味。我当时在用瑞典语写一篇文章，我用了那句话，我不知道采访者会将那篇文章以英文发布，所以当这篇文章出来时，它引起了很多有趣的反应。其中之一是 Legora 是由吸血鬼组成的吗？第二个是你们需要更好地使用牙线。很多好主意，但这现在已经变成了内部的“蓝烟”这个词，意思是，我们继续坚持下去。
So starting years behind some of our biggest competition and having to build this intensity in a country where most people want to go for fika at 3:00 p.m., which means having a cinnamon bun and a coffee, or go home at 5. One of the ways that that has been done is to create cultural values and cultural moments that really distinguish us from many others. And so in Swedish there is a saying you can wake up with a taste of blood because you're so excited about something. And I was doing an article in Swedish and I used that saying and I didn't know that the interviewer was going to release that article in English and so when this came out it got lots of fun reactions. One of them was is Legora made up of vampires. The second one was you guys need to floss better. Many good ideas, but this has now turned into the word blue smoke internally, which means that and we just keep going at it.
我认为将总部设在斯德哥尔摩有助于我们保持这一点。公司全球每个人都在斯德哥尔摩进行入职培训，夏天比冬天好得多。但这在全世界任何 Legora 办公室都创造了同样的氛围，无论你走进哪里。我认为一个很好的例子是，我们有一个美国非常大的律师事务所，我们的竞争对手正在与之合作。但那家律师事务所在根据收到的条款清单起草某些类型的文件方面有一个具体问题。我们的竞争对手曾承诺他们会解决这个问题。但他们没有。所以我们的一位美国工程师、法律工程师、律师了解了这个用例，坐上飞往斯德哥尔摩的航班，与我们的工程师共度一周，在瑞典解决了问题，然后回到美国并将其交付给客户。我认为瑞典基地已成为 Legora 真正的超能力之一。
And I think part of having our HQ in Stockholm has worked to help us keep this. Everybody globally from the company makes their onboarding in Stockholm, which is a lot nicer in summer than in the winter. But that has created the same vibe in any Legora office regardless of where you walk in all over the world. And I think one great example of this was we had a very big law firm in the US that our competition was working with. But that law firm had a specific problem around being able to draft certain types of documents based on incoming term sheets. And our competition had promised that they were going to solve that problem for them. And they hadn't. And so one of our US-based engineers, legal engineers, lawyers, learned about this use case, got on a flight to Stockholm, spent a week with our engineers, solved the problem in Sweden, and then went back to the US and delivered it to the customer. And I think that Swedish base has become one of Legora's real superpowers.
所以，在进入问答环节之前，我想总结一下。这个周末我与那些已经建立了规模的人同台，我还没有接近那个规模。所以我会把许多重要的经验留给他们。但我可以告诉你的是，三年时间比那要小一些。到目前为止，公司和 Legora 中真正重要的大部分事情并不是围绕签下一个大客户或完成新一轮融资或在曼哈顿开设新的大办公室。它发生在你知道的，一个没有窗户、氧气不多的房间里，或者在我们去机场的路上打电话，凌晨 2 点在 Slack 频道上试图在周一的重要演示之前修复一个 bug。没有人会把这些放在 pitch deck 里，但这就是公司的本质。你不应该等待那些大时刻。你必须热爱忙碌。我想如果我能给你们留下什么，那就是尽管我们始于瑞典的一家小律师事务所，我出生在瑞典群岛，班上只有 10 个人，我曾以为我的人生有一个大计划，我打算放学后去麦肯锡。但当 GPT-3.5 出现时，我们真正把它变成了我们的事业，我们抓住了我认为是我们这一代人的互联网时刻。所以你可以直接去做事情。Legora 和我自己就是活生生的证明。那么，我们将进入问答环节，我想邀请一位非常好的朋友，我们的 Y Combinator 合伙人 Gustav Armstr，我们现在真的变得专业了。这不是我们第一次这样做。
And so to round things off before we jump into the Q&A, I'm sharing this stage this weekend with people who have built things at a scale that I haven't yet gotten close to. And so I will leave many of the big lessons for them. But what I can tell you is that three years in is a little bit smaller than that. Most of what actually mattered in the company and in Legora so far did not revolve around signing a big client or closing a new round of financing or opening a new big office in Manhattan. It happened in, you know, a windowless room without much oxygen or on a call on our way to the airport, a Slack channel at 2:00 a.m. trying to fix a bug before a big presentation on Monday. And nobody puts that in a pitch deck, but that is what makes the company. And you shouldn't wait around for those big moments. You have to love the hustle. And I think if I can leave you with anything, it's that even though we started in a small law firm in Sweden, I was born in the Swedish archipelago with a class of 10 other people, I had I thought I had a big plan for my life and I was going to go to McKinsey after school. But when GPT-3.5 dropped in our knees, we truly made that into our thing and we captured what I believe to be our generation's internet moment. And so you can just do things. Legora and myself is living proof of that. So with that, we're going to jump to Q&A and I want to invite a very good friend, our partner from Y Combinator, Gustav Armstr, we're really becoming professionals at this now. This is not the first time we're doing this.
不，这很棒。
No, it's great.
很高兴见到你。
It's good to see you.
我也很高兴见到你。
You, too. It's good to see you.
谢谢你的演讲。太棒了。
Um, thanks for the talk. That was awesome.
真的很好。
Really good.
谢谢。
Thank you.
也总是感谢你回来在 Startup School 向 YC 和创始人们演讲。
And always thank you for coming back and speaking to YC and to founders here at Startup School.
当然。
Of course.
回来。
To be back.
我想问第一个问题，这是我收到很多创始人问关于你的问题，那就是你们是三位创始人，或者曾经是三位不是律师的创始人。
I want to ask the first question which is um one that I've gotten from a lot of founders who ask about you, which is you guys are three founders or were three founders who were not lawyers.
是的。
Yeah.
而你们创办了世界上最好的法律 AI 公司。领域专业知识不再那么重要了吗？
Uh and you started the best legal AI company in the world. Does domain expertise not matter that much anymore?
所以实际上有一个故事，一位瑞典风险投资家非常看好做 Leia 的 preede。
So there's actually a story of a Swedish VC who was very bullish on doing the preede of Leia.
但他们最终没有做，因为我们团队里没有任何律师。我最近在机场遇到了他们的一位合伙人，他们非常非常苦涩。Leia 或 Lora 对他们来说成了一个教训：或许领域专业知识并不是成功的必要条件。我认为你需要的是愿意了解市场，而我们很早就做到了这一点，通过让自己暴露在市场里，尽可能多地与律师相处。
But they ended up not doing it because we did not have any lawyers in the team. I met one of the partners from there at the airport quite recently, and they were very, very bitter. Leia or Lora turned out to be the lesson for them that perhaps domain expertise is not what you need to succeed. I think that you need a willingness to learn about the market, which we did very, very early by exposing ourselves and spending lots of time with lawyers, as much as we possibly could.
是的。
Yeah.
但我不认为你需要领域专业知识。这很可能取决于你所追求的市场。比如，如果你在量子计算或核聚变领域创业，那可能是个好主意。但法律对我们来说足够容易，我们边做边学。
But uh I don't think you need domain expertise. It probably, you know, depends on the market that you're going after. Like maybe if you're building in quantum computing or in fusion, that's a good idea. But legal was easy enough for us to pick up as we went along.
嗯，我昨天和一些创始人聊过，我们谈到了下一个类别，即 LMS 还不够好但将来会变好的领域。这让我想到你刚才说的，当你早期在玩 Bard 的时候。嗯，是什么给了你信心，在一个还不够好的模型上构建法律 AI 产品？它算是好，但还不够好。在座的许多其他创始人可能也在思考同样的问题，但领域不同，比如机械设计或其他模型还不够好的地方。所以我认为，我们刚开始时的那种部落知识是：你必须微调一个模型。
Um, I spoke to some founders yesterday and we spoke about sort of like the next categories where the LMS are not sufficiently good yet but they will be. And it made me think of what you just said when you were toying around with Bard in the early days. Um, what would you, what gave you the confidence to build a legal AI product on a model that wasn't, it wasn't good enough? It was like good but not good enough. And there's many other founders probably in this audience who are thinking about the same problem but in a different domain, maybe like I don't know mechanical design or some other part of somewhere where the models are just not good enough. So I think the sort of tribal knowledge when we got started was that you have to fine-tune a model.
是的。
Yeah.
对。在 2023、2022 年，那确实是游戏规则。我甚至记得，我想是彭博社花了数百万、数百万、数百万美元构建了一个彭博法律模型。我们的观点，部分是因为我们没有足够的钱，部分是因为我们真心相信那是对的，你知道模型会不断改进，对吧？就像我确信 Sam 今天晚些时候会上台这么说。
Right. In 2023, 2022 that was really like the name of the game. I even remember I think it was Bloomberg who spent you know millions and millions and millions of dollars building like a Bloomberg law model. Our view was partly because we didn't have enough money and partly because we truly believe that to be right, you know the models will keep improving, right? Like I'm sure Sam is going to come up on stage and say that later today.
是的。
Yes.
所以，只要押注模型会不断改进，而你应该关注的是如何将模型产生的价值交付给你的市场。这对我们来说是关键。即使第一代 ChatGPT 在法律领域不可用，部分是因为模型不够好，但部分也是因为它实际上不支持私人对话、欧洲数据托管。我记得我第一次销售时，我们基本上说我们像 ChatGPT，但我们在欧洲合规。那就是我们的推销词。那不是很有力的推销词。嗯，你知道，显然 ChatGPT 后来会解决这个问题，但当时已经足够好了。所以我认为我们认识到，你不必为这里的世界构建，你需要为今天的世界构建，也许领先一步。
And so by just betting that the models will keep improving and what you should be focused on is how can you deliver that the value that the models generate to your market. That was the crux for us. Even with the first generation of ChatGPT that was unusable in law, partly because the model wasn't good enough, but partly because it actually didn't support private conversations, European data hosting. I remember the first sales I ever did, we basically said we are like ChatGPT but we are compliant in Europe. Like that was the pitch. That was not a very strong pitch. Um, and you know, clearly ChatGPT would figure out how to do that later on, but at the time it was good enough. And so I think we recognized that you don't have to build for the world over here, like you need to build for the world today and maybe like one step ahead.
是的。
Yeah.
这有点像梯子，你应该最大化现在的机会，确保与客户建立非常好的关系，这样你也能和他们进行对话，因为如果每个人都押注于模型能力的指数级提升，Lora 平台将继续提供越来越多的价值。
And it's kind of like this ladder where you should maximize the opportunity now, make sure that you get really good relationships with your customers and so that you can also have that conversation with them because if everybody's betting on this, you know, exponential uplift in model capability, the Lora platform will continue delivering more and more value.
我认为我们只构建了公司历史上将要构建的所有软件的 1%，几年后产品可能会与今天大不相同。
I think that we have built 1% of all the software that we will build over the company's history and probably the product will look very different from what it does today a few years down the line.
是的。嗯，本周或过去两周有很多关于模型的新闻。嗯，对于制作纪录片的人来说，这将是一个很棒的剧集。我希望现在有人正在制作关于一切进展的纪录片。我不确定是否有人在做，但我希望有。嗯，开放权重小模型是当前的热门话题。嗯，几个月前我们在台上交谈时，你谈到你不依附于某个特定模型。你可以与许多不同的模型合作。嗯，基于本周的情况，你有没有更新你的观点？嗯，但比如你怎么看，我确信一年半前有一个时刻，似乎只有一两个模型真正足以驱动 Lora，而现在似乎有很多。
Yeah. Um, a lot of news on models this week or last two weeks. Um, it's going to be like a great episode for whoever makes the documentary. I'm hoping that someone is making a documentary right now about everything that's going on. I'm not sure if someone is, but I hope they are. Um, Open Weight Small is uh very much um the topic right now. And um when we spoke u on stage a couple months ago, you talked about that um you're not attached to one specific model. You can work with many different models. Have um have you updated your view based on this week? Um but like how do you think about so like I'm sure there was a moment like a year and a half ago when it seemed like there was just one or two models that were really sufficient to do to power legor and now there seemed like a lot more.
是的。我认为这很棒。是的。我的意思是，显然不同模型会在不同用例中有应用，所以如果你做客户支持，你优化的是周转速度、对客户的延迟以及解决特定工单的成本。嗯，我认为 Finn 曾经对每个解决的客户工单收费约 1 美元。在法律领域，你实际上往往想要最多的智能，因为 token 支出或软件支出相对于应用于问题的人类专业知识来说非常小。如果你在解决复杂的诉讼，你会想把最多的脑力投入到那个问题上。所以我们实际上有两类不同的用户。一类用户要求能够循环使用非常复杂的模型来长时间处理问题，比如累积大量的 LLM 成本，另一类用户则说，我们能否获得开源模型，这样我们的消费会更便宜。我认为答案介于两者之间。嗯。
Yes. Which I think will be great. Yeah. And and I mean you know clearly there will be uh applications of different models in different use cases and so if you take customer support um you're optimizing for you know speed of turnaround um you know latency to the customer and cost on you know solving a particular ticket. Um I think Finn used to charge like $1 per you know solved customer ticket. In law, you actually want the you know most amount of intelligence quite often because the fraction of token spend or software spend compared to like you know human expertise applied to the problem is really tiny. If you're solving you know complex litigation, you're going to want to throw the most brain that you can on that problem. And so we have really sort of two different camps. Like one group of our users are asking for the ability to loop really complex models to to work on problems for a long period of time like racking up big LLM costs and another group of our users are saying you know can we get access to open source uh you know that is going to be cheaper on our consumption and I think the the answer is like somewhere in between. Mhm.
我认为最终，我鼓励尽可能多的人构建的核心 IP 和能力之一是评估用例的能力，因为那是超能力，能让你有效地路由。嗯。
I think ultimately uh one of the core IPs and muscles that I encourage as many of you as possible to build is uh the ability to evaluate use cases because that is the superpower that then allows you to uh route things effectively. Mhm.
我们早期做了很多这方面的工作，因为我们雇佣了很多律师加入 Lora，我们让他们的一部分工作面向客户，但另一部分工作是构建用例，我们会把评估放在上面。成本、智能的前沿以及那个前沿在哪里，你知道，还有待观察。我认为还有一个算力问题，当我们与实验室和算力提供商锁定大合同时，那实际上是一件大事。我们不能让系统宕机，当世界上一定比例的律师都在使用它，他们的客户依赖它时。
We did a lot of that early on because we hired a lot of lawyers into Lora and we made part of their job customer-facing but part of their job was also to build out use cases that we would throw our evals on and the the frontier of like you know cost uh intelligence and where that sort of uh frontier lies you know yet to be seen. I think there's also a compute problem where when we lock in big contracts with the uh with the labs and with our compute providers that is actually a huge thing. We cannot have our system go down when you know a a number of percentage points of all the lawyers in the world are working with it and their clients rely on it.
那么，你们会关注哪些具体的基准测试来评估……我相信你们基本上是……
So are there specific benchmarks you look at for how... I'm sure you are basically...
嗯，我们看的是 Lora 基准。我们实际上上周五宣布并发布了它，这是过去 3 年的工作。我们一直将其保密到周五，原因是当时基本上只有 OpenAI 或 Anthropic 的模型，所以发布基准测试并不那么有趣。我们只会在系统中提供这两种模型，用户可以选择。现在可用的选项多了很多，而且我们惊讶地发现，SpaceX 和 Grok 在我们的基准测试中表现最佳，尤其是考虑到成本。我们甚至还没有与他们签订数据处理协议，所以实际上我们还不能向客户提供它们，但现在我们会开始提供。
Well, we look at the Lora bench. So we actually announced and released that on Friday last week, and it has been work for the past 3 years. We've kept that internal up until Friday, and the reason for that was it was basically OpenAI or Anthropic models, and so it wasn't that interesting to publish benchmarks on. And we would only make those two models available in our system, and the user could pick. Now the smorgasbord of available options is a lot larger, and actually we found to our surprise that SpaceX and Grok was one of the best performant models, especially given the cost, on our benchmarks. And we don't even have them on our data processing agreement, so we actually can't yet offer them to our customers, but now we will start to.
那么，你的客户对于使用哪些模型有强烈的偏好，还是他们只想要最好的，无论哪个最好？
And do you see your customers having strong opinions about which models they want to use, or is it sort of like they just want the best, whatever is the best?
有些客户有强烈的偏好，尤其是银行，你知道，大型律师事务所……他们不想要中国模型。
Some of them have strong opinions, especially like banks, you know, big law firms have a... they don't want Chinese models.
是的。有道理。
Yeah. Makes sense.
嗯，我想稍微回溯一下，比如当你决定要创办一家初创公司时。在座的很多人……过去两天我和 Startup School 的参会者吃了两顿晚餐。我们讨论了什么时候应该创办一家初创公司？有人在 DeepMind 工作，有人正在实习或工作。通常你已经工作了一两年。嗯，你认为触发你决定“我就要做这个”，而不去麦肯锡或其他选择的原因是什么？
Um so I want to go back to just a little bit earlier, like when you decided you want to build a startup. A lot of people in the audience here... I've had two dinners in the last two days with attendees at Startup School. We talked about when should you start a startup? People are working at DeepMind, people working at internships or jobs right now. Often you're one or two years into it. Um what would you say was the trigger for you to say I'm just going to do this and not do McKinsey or the other options that you had?
嗯，我父亲在互联网泡沫期间有两家公司，两家都在筹集了过多风险投资后归零。所以也许一些……
So my dad had two companies during the dot-com boom, both of which went to zero after raising too much venture capital. So maybe some...
他对你来说是个好导师吗？
Has he been a good mentor for you?
是的。当然。
Yes. Yeah, absolutely.
你们多久聊一次……
And how often do you speak about...
哦，我们可能每天或每隔一天都会通话。
Oh, we speak probably every day or every other day.
太好了。是的，聊很多事情。
Great. Yeah, about lots of stuff.
我以前……现在经常太累了，不想走路回家。所以我可能会坐火车或打车，但以前我会走路回家，然后给他打电话，那会很有趣。而且嗯……
I used to... now I'm like too tired to walk home often. So I would like take the train or like take a cab, but I used to walk home then I would call him and that would be a lot of fun. And um...
有时我们称他为高级顾问。嗯，实际上我们有一个最大的客户推介。我们资源紧张。所以他进来为那个客户做了 PowerPoint。
Sometimes we call him senior adviser. Uh we have actually one of our biggest customer pitches. Um we were stretched for resources. So he came in and did the PowerPoint for that client.
那时他什么都知道了。
He knew everything at this point.
是的。是的。他知道大部分。
Yeah. Yeah. He knows most of it.
所以他在互联网泡沫期间有两家公司，都归零了，然后他在我长大的群岛有一家船务公司。所以我一直 surrounded by 创业精神，我一直觉得我想在那个领域证明自己。我非常有竞争力。我以前玩电子游戏。我以前玩 Dota 2，嗯，比上学还多。然后我认为创办公司就像是最佳的……这是你能进入的最大的竞技场，你要么胜利离开，要么失败离开，要么还行。我非常坚定地想要……我可能应该更多地了解世界，我认为麦肯锡是一个很好的起点，我仍然这么认为。我认为它真的教会你如何工作。你在那里学到职业道德，但我做过实习，所以我想我在学习期间也学到了一些。
So he had two companies in the dot-com boom, both went to zero, and then he had a boat business in the archipelago when I grew up. And so I've always been surrounded by entrepreneurship and I always felt like I wanted to prove myself in that arena. I'm very competitive. I used to play video games. I used to play Dota 2, um, like more than I went to school. And then I think building a company is like the optimal... it's the biggest arena that you can enter and you'll either leave victorious or you'll leave defeated or you'll be okay. And I was very decided in that I want... I should probably learn more about the world, which I thought McKinsey was going to be a great place to start at, which I still think it is. I think it really teaches you like how to work. You learn work ethic there, but I did an internship and so I think I got a bit of that even during my studies.
然后我们有了 GPT 这个关键转折点，所以我们谈论这个，就像几年前在 GPT 之前，你对投资并不那么兴奋，因为没有很多……
And then we had this pivotal moment with GPT and so we were talking about this like you weren't as excited about investing a few years back like pre-GPT because there wasn't a lot of...
没有很多进展，就像……我认为在 GPT 之前的一两年，没有很多新想法。
There wasn't a lot going on like... I think the last year or two before GPT it was not a lot of new ideas.
对，我记得在 GPT 之前坐下来，列出我能想到的所有创业想法，大多数都相当无聊。我认为大多数都不会成功。
Right, and I remember sitting down pre-GPT like here are all the startup ideas I can think of, most of them were pretty boring. I don't think most of them would have worked out.
是的。是的。
Yep. Yep.
嗯，然后小公司相对于大公司获得了巨大的优势。
Um and then like small companies got a really big advantage versus big companies.
我的意思是，我认为你从微软构建 Copilot 的方式中看到了这一点。它很长时间都没有成功。是的。
And I mean I think you saw this in just like the way that Microsoft would build co-pilots. Like it didn't work for a really long time. Yeah.
我记得当它出来时，我们就像，“天哪，我们完蛋了。”你知道，每个律师已经在 Word 和 Outlook 中工作，现在他们只需要使用 Copilot。结果它没有成功，而我们仍然能够创造大量价值的能力是巨大的。我认为现在回想起来，你必须自己弄清楚，这让你学到了很多，我真的认为你学到的东西的数量是你愿意忍受的不适程度的函数。
And I remember when that came out and we were like, "Oh my god, like we're so screwed." You know, every lawyer already works in Word and Outlook and now they're just going to use Copilot. It turns out like it didn't work and the ability for us then to still build a lot of value was huge. I think now reflecting back on that you learn so much by having to figure it out and I really think that the amount of things you learn is a function of the amount of discomfort that you are willing to endure.
是的。嗯，当我们创办 Legora 时，我非常内向。我真的不习惯与客户交谈。我想尽可能多地编码。你知道，当你第一次不得不与员工分道扬镳时，所有这些都太难了，但你意识到初创公司是你在短时间内学到最多东西的地方之一。我认为，你知道，如果我再创办一家公司或加入一家公司，我仍然觉得在其他初创公司工作让我学到了很多。实际上，在创办自己的公司之前，我曾在初创公司工作过。
Yeah. Um, when we started Legora, I was like super introverted. I really didn't feel comfortable talking to customers. I wanted to code as much as possible. You know, when you have to go separate ways with an employee for the first time, like all of this is so hard, but you realize that startups is one of the places where you can learn the most in a short amount of time. I think, you know, if I were to build a company again or join a company, I still feel like I learned a lot working at other startups. I actually worked at startups before starting my own.
然后我会说，你应该深入思考你和谁一起创业以及问题领域。我认为我们选择的是一个领域，而不是一个问题，我认为这是两种不同的方式。法律和 AI 肯定会成为一件事，这完全显而易见。
And then I would say you should think really deeply about who you started with and the problem space. And I think we picked a space more than we picked a problem which I think is like two different ways of going about it. It was completely obvious that legal and AI was going to be a thing.
是的。
Yeah.
它将如何成为一件事非常不确定，所以我们只是说让我们朝着这个大致方向前进，然后我们会弄清楚。这是一种方式，我想另一种方式是找到一个具体问题，非常清晰地解决它，产生 ROI，然后从那里扩展。
How it was going to be a thing was like very uncertain and so we just said let's march in this general direction and then we'll figure it out. That's one way of doing I guess the other way of doing it is like you find a specific problem, you solve that very clearly, you generate ROI and then you expand from that.
是的。
Yeah.
嗯，你之前提到，我知道你非常有竞争力。嗯，这在 Legora 内部是如何体现的？比如 Legora 的员工如何知道 Max 是一个非常竞争的 CEO？
Um you mentioned earlier, well I know that you are extremely competitive. Um how is that expressed inside Legora? Like how does a Legora employee know that Max is a very competitive CEO?
我认为 Legora 的每个人都是这样自我选择的。
I think everybody at Legora is like that self selects.
如何如何表达？是像周五会议吗？这是我们的目标，然后……
How how does how's expressed? Is it like a Friday meeting? Here's our goals and like...
嗯，是的。所以我们让每个人都知道每月的目标。我认为设定季度目标有点棘手。
Well, yes. So we do everybody knows every monthly goal. I think setting quarterly goals is like a little bit tricky.
从一开始只有你们三个人的时候就是这样吗？
Is that how it was from the very beginning with just three of you guys?
嗯，不，那时没有目标。那时只是每天你试图最大化结果。
Um no then it was no goals. Then it was just like every day you like try to maximize the outcome.
好的。
Okay.
嗯，然后我们有了，你知道，当你有 750 人时，你需要更好地规划，并有一些更好的内部沟通。
Um and then we got you know when you're 750 people you need to plan a little bit better and like have some better internal communication.
实际上，沟通是规模扩张时真正会出问题的地方。我想很多人已经谈过这一点了。但你真的能感受到工程团队的那种驱动力，他们想打造最好的产品和最好的功能。如果你把我们的表格审查和市场其他替代方案比较，真的可以排个名次。
Actually, communication is like the thing that really breaks when you scale. I think a lot of people have talked about that already. But you really feel this drive from engineering that they want to build the best product and the best feature. And if you compare our tabular review versus the other alternatives in the market, it's really sort of you stack rank it.
是的。
Yeah.
现在斯德哥尔摩有很多人在休假，因为是七月。大多数休假的工程师比以往任何时候都更有效率，因为他们稍微休息了一下，但然后他们也能编写代码来解决他们想打败、想赢的问题。
And we have lots of people on vacation now in Stockholm because it's July. And most of the engineers who are on vacation are more productive than they have ever been because they're taking a bit of time off, but then they also get to code now on problems that they like want to beat and they want to win.
这是一项团队运动。我认为
It's a team sport. And I think
你可以有竞争力，但要以团队的方式。所以你真的需要巩固这种同赢同输的策略。当我们取得大胜时，会真正庆祝；当我们失败时，会一起哀悼。
you can be competitive but be it in a team way. And so you really need to solidify this winning or losing together strategy I think. And when we get a big win, that is really celebrated, and when we have a loss, we really mourn it together.
是的。
Yeah.
然后我们立即制定计划，如何扭转局面或如何实现逆转。我觉得人们处于解决模式，而不是责备模式。我认为这对我们是一个很大的动力驱动。但我们也一直处于追赶模式，对吧？当我们起步时，有其他法律 AI 公司比我们更大，所以我们总是觉得，至少在开始时，我们在追赶。现在我们大得多了，
And then we immediately make a plan of how can we flip it or how can we make a land and turn it around. And I feel like people are in solution mode rather than blame mode. And I think that has been a big momentum driver for us. But also we've been in chasing mode, right? There were other legal AI companies that were bigger than us when we started, and so we always felt, at least in the beginning, like we were chasing. Now we are much bigger and
假设你在某个时候很快成为最大的
let's say you were the biggest one at some point very soon
是的
yes
那会影响你的动力或你的
is that going to impact your motivation or your
我想那时，这很有趣。我思考了很多关于那一刻会发生什么。
I think at that time it so that's interesting. I thought a lot about what happens at that moment.
是的。
Yeah.
我认为我们需要要么挑出另一个敌人，说好吧，我们现在必须比他们更大，要么我们必须真正找到这种比昨天的自己更好的方式。我认为有竞争对手是好事。你不应该过度关注他们在做什么，但这绝对像一根胡萝卜。而且这是一个容易团结起来的事情，因为你不想要的是很多支线任务。
I think we need to either pick out another enemy and go like okay we have to be bigger than them now, or we have to really find this way of being better than ourselves yesterday. And I think that having competitors is a good thing. You should not be overly focused on what they're doing, but it's definitely like a carrot, I think. And it's an easy thing to rally around because what you don't want is you don't want a lot of side questing.
是的。
Yeah.
你希望每个人都专注于同一件事。
You want everybody focused on the same thing.
所以，我从观众那里收到的一个问题是，你如何变得更有野心？当我问这个问题时，我想到了我现在合作的一家公司，他们说：“哦，我们想在欧洲做这个国家，然后我们会在欧洲做另一个国家，然后也许明年我们会做美国。”你们在批量中就在欧洲做了很多国家，然后
So, one question I got from the audience is how do you become more ambitious? And when I ask this question, I'm thinking of a company I'm working with right now who's like, "Oh, we want to do this country in Europe and then we're going to do this other country in Europe and then maybe next year we'll do US." You guys did a lot of countries in Europe just you're in the batch and then
现在我们在大约 50 个国家
and now we're at like 50 countries
现在你们无处不在，而且规模扩张得非常快。这是野心，还是产品奏效了，还是
and now you're everywhere and you scale up really quickly. And is that ambition or is that the product was working or is
不，这绝对是野心。我的意思是，我认为这就像
No, it's definitely ambition. I mean I think it's like
一个人如何学会这一点，或者一个人如何变得更有野心？这可能吗？
how does someone learn that or how does someone become more ambitious? Is that possible?
是的，我认为这是可能的。我实际上认为变得更有野心的最好方法是拥有一群其他非常有野心的同龄人。所以在我高中的最后几年，我相当懒。我玩了很多电子游戏。我在学校过得挺轻松，没有太努力，也不知道我的圈子之外世界有多大。
Yeah, I think it's possible. I actually think the best way to become more ambitious is to have a peer group of other very ambitious people. So in my final years of high school, I was pretty lazy. I was playing a lot of video games. I had a pretty easy time in school and I didn't work that hard and I didn't know how big the world was outside of my sphere.
嗯。
Mhm.
大学头两年也是一样。
And the first two years in college was the same.
是的。
Yeah.
然后我交了一群非常有野心的新朋友，这让我豁然开朗。我从零实习变成一年做八份工作。我实际上认为这是 YC 最好的部分之一。我想 PJ 在一篇文章中谈到了这一点，就像你来到这里，然后你意识到哇，每个同龄人都真的很有野心，所以你们一起变得更有野心。
And then I made a new friend group of very ambitious people and that made it click for me. I went from doing zero internships to doing eight jobs in one year. And I actually think that's one of the best parts about YC. And I think PJ talked about that in one of the essays, which is like you come here and then you just realize that wow every peer like everybody's really ambitious and so you become more ambitious together.
是的。
Yeah.
我认为现在我们在 Lora 内部就是这样。
And I think now we are like that inside of Lora.
是的。
Yeah.
因为我不觉得我是老板。我觉得我是运营公司的团队中的一员。你知道，我们的 CFO 有时会挑战我们的营销团队要更有野心，而你知道，CFO 通常可能不会关心这个。
Because I don't feel like I'm the boss. I feel like I'm a peer in the team that is running the company. And you know, our CFO sometimes challenges our marketing team to be more ambitious and you know, like a CFO maybe typically wouldn't care about that.
是的。
Yeah.
但我们都非常积极进取，这真的很有趣。然后你真的可以在整个公司组建这样的团队。不仅仅是在高层，对吧？就像无处不在。是的。然后你可以看着你的创造像螺旋一样变得越来越有野心，这真的很有趣。我也认为你需要有大的偶像和大的公司在你前面。我的意思是，在开始时
But we're all just like so forward leaning and it's just so fun. And then you really get to assemble this group across the entire company. It's not just at the top, right? It's like everywhere. Yeah. and then you get to watch your creation like spiral and become more and more and more and more ambitious and that is really fun. I also think you need to have big idols and big companies ahead of you. I mean in the beginning
谁是你的偶像？
who was who was your idol?
嗯，所以我们会把 Spotify 和 CLA 视为斯德哥尔摩有史以来最大的成功科技故事。
Well, so we would look at Spotify and CLA as these are the biggest success tech stories from Stockholm that have, you know, ever happened.
然后我们意识到，哇，我们的市值比 CLA 还大。
And then we realized, wow, we have a bigger market cap than CLA.
那是一个有点奇怪的时刻。
And that was like kind of a weird moment.
是的。是的。
Yeah. Yeah.
嗯，因为我们觉得我们已经完成了从零到一。
Um because we feel like we have done zero to one.
是的。
Yeah.
而从一到十，然后从十到一百仍然存在。那就在我们面前，所以是的
And 1 to 10 and then 10 to 100 still remains. that's in front of us and so yeah
嗯，你会说，嗯，所以这里的观众想知道，现在你在学校或即将毕业时，哪些技能更有价值，嗯，相对于不那么有价值的，基本上就是，你需要创业的技能或你需要进入职业的技能，它们是否在很短的时间内发生了变化？
um would you say uh so so people in the audience here are wondering what skills are either more valuable to have now as you're in school or you're graduating from school um versus not more valuable like basically is is the skills that you need to start a company or skills that you need to get into like like basically your career have they are they changing in a very short amount of time now
我认为讲故事至少对 CEO 来说很重要。
I think storytelling is at least for the CEO.
是的。
Yeah.
我完全低估了这有多重要，以及这会是多好的一项技能，因为最终发生的是，你需要把公司卖给自己。
I had completely underappreciated how important and how good of a skill that would be because what ends up happening is you need to sell the company to yourself.
是的。
Yeah.
就像如果你不相信自己的故事，这么努力工作真的很难。你需要把故事卖给员工。就像现在我们在招聘，你知道我们在和实验室竞争人才，你知道我们需要推销在 Lora 工作会比在 Anthropic 工作更好的故事。你需要把故事卖给投资者。你需要把故事卖给客户。我认为在大学或高中或任何地方，花时间学习这一点是有用的。我不知道如何学习它。我实际上不知道我是怎么学会的，除非我只是有很多接触。
Like working this hard if you don't believe in your own story is really hard. You need to sell employees. Like now we are hiring people you know we are competing with the labs for talent and you know we need to sell the story of how working at Lora is going to be better than working at Anthropic. You need to sell the story to investors. You need to sell the story to customers. And I think that in college or in high school or wherever that is, like spending time to learn that is useful. I don't know how to learn it. I actually don't know how I learned it unless I just had a lot of exposure.
是的。
Yeah.
我一直喜欢故事。我喜欢电子游戏。我享受好故事。但我认为那是最重要的事情。
I always liked stories. I like video games. I enjoy a good story. But I think that is the most important thing.
然后也许其他真正的技术技能。
And and then maybe the other really tech technical skills.
好的。
Okay.
在进入技术话题之前，我再说一点。我以前是个很糟糕的队友，也是个很糟糕的合作者。我从小到大打了很久排球。在队里，我是那种队友发球失误时就会冲他们吼“你太烂了，这种球怎么会发丢？”的人。那样做，下次比赛就没人会叫上你了。
I'll say one more before we go to technical. I used to be a really bad teammate and a really bad collaborator. I played volleyball for a long time growing up. I would be the type of person on the team who, when my teammate missed the serve, I would yell at them and go, "You suck. How could you miss that serve?" That doesn't get you invited to play the next game.
嗯。
Yeah.
那样既不会赢得追随者，也带不来好的氛围。
It doesn't invite a followership and it doesn't invite a good vibe.
对。
Yeah.
所以到了高中，我真正明白了这一点，不得不改变那种工作方式。我认为技术能力、迭代速度，以及根据客户反馈快速迭代的能力，仍然是最重要的起步技能。
So in high school I really understood that and had to pivot that way of working. I think technical skills, the velocity and your ability to iterate on customer feedback, is still probably the most important skill to get started.
嗯。
Yeah.
即使现在我们规模更大了，我们也经常会在客户电话里听到客户遇到的某类问题。我会把它丢到产品频道里，然后问：我们能多快解决这个问题并让客户满意？在一家千人体量的公司里这么做有点奇怪，但我们仍然是这样做的。同时你得确保不要过度来回摇摆，要继续朝着正确的愿景建设。但我仍然认为，小公司相对于现有竞争对手的速度能力就是超能力，而当你拥有超能力时，就必须全力发挥。对我来说仍然如此：我应该做我特别擅长的事，公司里的其他人做他们特别擅长的事。
Even now when we're bigger, very often we will hear about a particular type of problem a customer is having on a customer call. I will dump that in the product channel and I'll go, how quickly can we turn this around and delight the customer? That's a bit strange to do at a thousand-person company scale, but it's still how we do it. And you need to make sure that you don't overly zigzag and that you continue building towards the right vision. But I still think that your ability as a smaller company to be fast versus your incumbent competition is the superpower, and when you have a superpower you need to lean into that. That's still true for me: I should be doing the things that I'm uniquely good at, and other people in the company do the things that they're uniquely good at.
嗯。
Yeah.
我觉得技术能力可能如今也不一样了。技术这个词的含义变了。
I think also technical skills, maybe they're different these days. Technical means something else.
是的。但它的意思仍然是能够构建大规模系统，对吧？仅仅因为你能用 AI 生成大量代码，并不能成为你不是一名优秀软件工程师的借口。
Yes. Well, but it still means being able to build systems at scale, right? Just because you can produce a lot of code with AI is not an excuse for not being a very good software engineer.
对，对。我仍然认为我们需要很多优秀的软件工程师。我们在 Legora 正在招聘一百名。
Right. Right. And I still think we need many good software engineers. We are hiring a hundred of them at Legora.
人们加入 Legora 后能接触到的最有趣的问题是什么？
What's the most interesting problems that people get to work on when they join Legora?
哦，目前最有趣的问题之一是从反应式智能体向主动式智能体的转变。
Oh, one of the most interesting problems right now is this transition from reactive agents to proactive agents.
能举个例子吗？
So what's an example?
过去三年使用这个产品时，你会给 Lora 一个提示或一组指令，它就会去执行。现在我们基本上是把 Lora 连接到不同的上下文片段，当它收到触发信号时，就会开始做事情。这意味着，如果销售团队收到一份合同，它会被路由给一个 Lora 智能体。该智能体会并行开始处理。如果需要升级给律师，它就会升级。否则，它可能直接执行合同。或者你把整个数据室连接到一个 Lora 智能体，智能体会自动开始整理数据室并生成尽职调查报告。这就是让智能体在你不告诉它该做什么的情况下自行做事。
So for the last three years of working with the product, you know, you will give Lora a prompt or a set of instructions and it's going to go off and do that thing. Now we are basically connecting Lora to different pieces of context and when it gets a trigger it will start to do something. That means that if the sales team gets a contract it will get routed to a Lora agent. The Lora agent will start working on it in parallel. If it needs to escalate to a lawyer, it will. Otherwise, it might just execute the contract. Or you connect the entire data room to a Lora agent and the agent automatically starts organizing that data room and produces the due diligence report. It's having the agent do things without you telling it what it should do.
嗯。
Yeah.
对我们来说，这就是解锁：一名律师如何在 Lora 的帮助下产出十名律师团队的成果。
And that for us is the unlock of how does one lawyer produce the outcomes of a team of 10 lawyers with the help of Lora.
嗯。
Yeah.
我觉得这非常有趣。从规模角度看，我们也有一些非常有趣的问题，比如我们在 OCR 和文档解析上花费数百万美元，这些都是在更大规模下需要解决的真实工程问题。
I think that's very interesting. We also have some very interesting problems just from a scale perspective, like we're spending millions and millions and millions of dollars on OCR and document parsing, and these are real engineering problems to go solve at a bigger scale.
明白了。你们在 4 月时用 18 个月从零增长到 1 亿美元 ARR。还在以疯狂的速度增长。
Got it. And you grew from zero to 100 million ARR in 18 months in April. Keep growing at a crazy rate.
是的。
Yes.
你的工作在过去 18 个月里变化很大。
Your job has changed quite a lot in 18 months.
是的。
Yes.
我们之前聊到你现在要面试高管团队，就像几年前我们在 YC 听过 Dropbox 创始人的那场分享。我们聊过你要玩的那些小游戏。你在 18 个月里玩了很多小游戏，现在你到了组建或招聘人来管理团队这个小游戏。这听起来很难，而且听起来不像你以前做过的事。那你怎么——
We talked earlier about you now interviewing exec teams, like we had this talk from the founder of Dropbox at YC a couple years ago. We talked about the mini games that you play. You've done a lot of mini games in 18 months and now you're at the mini game of building out or hiring people to run your team. How—that sounds hard and doesn't sound like something you've done before. Like how—
嗯。
Yeah.
你怎么学会这种事？招对高管似乎极其重要。
How do you learn something like that? It seems extremely important to hire the right executive.
某种程度上，我认为在那之前的大部分工作让你有资格做那件事，对吧？如果你——你基本上是通过完成上一个小游戏来获得下一个小游戏。
In a way I think most of the work up until that point makes you qualified for that thing, right? If you—you get the next mini game by basically seeing the last one.
是的。
Yes.
但你必须确保自己学习，因为每个小游戏的一部分——我认为你真的需要把公司放在第一位，某种程度上远高于自己的自尊。我常对高管团队说：我每个季度都需要重新证明自己胜任 Lora CEO 这份工作。这是一家新公司，有新挑战，他们也是如此。在 100 万美元 ARR 时管理销售团队，与在 1.5 亿美元 ARR 时非常不同。对我自己来说，如果你享受这个，如果你享受学习新事物，如果你享受被挑战，如果你接受“现在我最重要的工作是组建高管团队，我要去做这件事”，你就会成功。然后一如既往，你要努力建立一个做过这些事的人的优质网络。难的是，很少有人像我们这样快速地做公司，甚至我们现在招来的很多高管，他们来自 SaaS，来自那个三倍三倍、两倍两倍的世界。如果你年同比增长 40%，你已经做得非常好了，算是走上正轨。
But you need to make sure that you learn, because at part of every mini game—and I think you really need to put the company first, way ahead of your own ego in a way. And you know I tell this to the executive team: I need to re-qualify for the job as CEO of Lora every quarter. It's a new company, new challenges, as do they. Running a sales team when you are 1 million ARR is very different from 150 million ARR. And I think for myself, if you enjoy that, if you enjoy learning new things, if you enjoy being challenged, if you accept the fact that now my most important job is building out the executive team and I'm going to go do that, you will succeed at it. And then I think as always, you try to build a good network of people who have done it before. The tough thing is there's very few people who have done the company this quickly as we are, and even a lot of the executives that we are now bringing on, they come from SaaS and they come from this world of triple-triple-double-double. If you grow 40% year on year you are doing fantastic, you're sort of on track.
现在这已经不够了。
That's not enough anymore.
确实不够。比如我们的 CFO David，从另一家非常成功的 YC 公司加入。但他们在四年左右从 3000 万增长到 3 亿，而我们现在大约一年就做到了。还有从 200 人到 1300 人，他们用了四年多，我们一年就做到了。所以你必须压缩时间线。我认为你需要极其果断地对事情说不，并且要了解自己在哪里最有竞争优势，在哪里能做出最大贡献。
It's not enough. Like we—you know David, our CFO, joined from another YC company, very successful. But they've grown from 30 million to 300 million in something like four years, and now we're doing it in like one year. And so from 200 people to 1300 people, and they're doing over four years and we're doing it in one year. So you need to just compress timelines. I think you need to be brutal at saying no to things and you need to understand yourself where you have the most competitive edge, where you can do the most.
四月底我落地斯德哥尔摩时，所有的广告基本上都是 Jude Law 出演的 Legora 广告。我很好奇你是怎么让 Jude Law 成为公司代言人的。如果你还没看过那些广告，我建议你去 YouTube 上看 Jude Law 的广告。
When I landed in Stockholm at the end of April, all of the ads were basically Legora ads with Jude Law. And I'm curious how you get Jude Law to be the face of the company. And he has—if you haven't watched the ads, I recommend you go on YouTube and watch the Jude Law ads.
你是不是又重新爱上了法律？
Did you fall in love with law again?
是的，确实如此。
Yeah, I did.
是的，你确实爱上了。很好。发生了什么？你是怎么做到的？
Yeah, you did. Good. What happened? How did you score that?
好。内幕是这样的，我们和一家营销机构开了个会，他们提出的一个方案是，我们为什么不重拍一些非常著名的法律场景，比如——
Okay. So, the inside baseball is we had a meeting with one of our marketing agencies and one of their pitches was, why don't we redo very famous legal scenes like—
“你承受不了真相。”
You can't handle the truth.
对，比如我们为什么不把著名的法律场景重拍一遍，用律师，但用 Legora 来做？
Yeah, like why don't we redo famous legal scenes with lawyers but with Legora?
嗯。
Yeah.
不，方案是这样的。其实挺搞笑的。所有这些关于律师的电影——如果它们有了 Legora，你知道，电影大概一分钟就结束了。于是我们有了和演员合作的想法，然后有人说，Legora 是 AI 驱动的法律，那我们能不能请裘德·洛（Jude Law）来扮演 AI 驱动的角色？这就让我们走上了这条路，嗯，也许我们能请到裘德·洛。现在好莱坞有趣的地方在于，它相当反 AI，无论是写剧本还是在电影中使用 AI 等等。
And no, the pitch was this. It was actually pretty funny. All of these movies about lawyers—what if they just had Legora and you know the movie would be over in like a minute? And so we got to this idea of working with actors, and then somebody went, well, Legora is like AI-powered law, what if we got Jude Law to be like AI-powered? And that got us down the track of, huh, maybe we could get Jude Law. And what's funny in Hollywood right now is it's pretty anti-AI, with writing scripts, using AI in movies, etc.
而且我们第一次联系他时，他觉得我们完全不严肃，就像某个瑞典 AI 公司想和裘德·洛合作做广告牌和营销活动。
And he thought we were completely unserious when we reached out the first time, like some Swedish AI company wants to work with Jude Law to make billboards and a campaign.
但后来我们非常坚持，我认为一种被低估的谈判方式就是软磨硬泡。这是我父亲教我的事情之一。我们磨了又磨，最后他答应了。但有一个条件，他要自己挑选编剧和摄影师。谁能想到呢？他带来了《周六夜现场》的剧本导演和《奥本海默》的摄影师来打光并完成整个摄影。他们基本上就进了片场。Legora 的人一个都不许进，你知道，几个小时后，这些神奇的作品就出来了。
But then we were so persistent and I think one underappreciated form of negotiation is nagging. It's one of the things my dad taught me. And we nagged and nagged and nagged and he said yes. On one condition, he got to pick his own script writer and cinematographer. And who would have known? He brings an SNL script director and the cinematographer from Oppenheimer to light the scene and do the whole cinematography. And they basically go into this set. Nobody from Legora is allowed in and you know a few hours later like these magical productions come out.
真的很好。
It's really good.
太搞笑了。我认为它为我们做的是，它让我们以另一种形象出现，现在连法律圈外的人都知道 Legora 了，这真的非常令人兴奋。
It's so funny. And what I think it did for us was it put us in another light where even people outside of law now know about Legora and that has been really exciting.
太棒了。非常感谢你来学校。这太棒了。
Awesome. Thank you so much for coming to school. This is awesome.
谢谢你，Gustav，邀请我来。
Thank you Gustav for having me.
谢谢大家。
Thanks everyone.