Elias Torres，Agency 联合创始人兼 CEO：没人告诉你的出售公司真相
- Torres称，2021年9月以12亿美元将 Drift 卖给 Vista，是“我人生中最大的失败”。 问题不在价格，而在于没能打造一家可持续的公司。800名员工（其中200人在客户成功团队）扼杀了产品迭代速度，文化过于宽松也解释了公司为何没能上市：“如果不是我给了他们那么多胡萝卜、糖果和奖杯，我现在本来会有一家上市公司。” 出售公司没人明说的另一面是：“成为无名之辈很难过……你在自己的公司里走来走去，却不再重要。”
- 现有巨头的成绩单很残酷：“大象没有跳舞，它们是在努力学跳舞。” Google“刚刚把球丢了”——明明发明了 Transformer 论文，却“推出的所有东西都不好用；NotebookLM，这个伟大的产品，已经死了”。Microsoft 押注 OpenAI 是“天才之举”，但 Copilot“没人用”。ChatGPT 是“王者”，却仍然是面向操作者的软件，“不是解决方案本身，也不是未来”。他承认，分发能力和数据是巨头最大的优势，但坚持认为“速度就是一切”。
- 真正的潜力在于：100人的公司做到数十亿美元收入。 “小一点更好”一直是对的，但“因为没有更多人，我们就无法做到数十亿美元收入，所以别无选择”。Agents 可能改变这一点：“我们其实只是在用算力替代人。” 他认同的融资演变是：早期融资仍以股权为主，增长融资则应基于可预测收入采用债务融资，因为不需要在员工数量上消耗那么多股权。
- AI的普及会比市场炒作的节奏更慢，因为“人类才是最大的阻碍”。 CEO们对他说“好，马上引入 Agency”，随后却把决定层层下放，最终得到的答复是“下个季度再说”。因此他提出反共识判断：5年内，AI服务公司将超过AI技术供应商（Paul Graham 曾对此嗤之以鼻；Torres 曾为 OpenAI 的客户提供咨询，6-7个月赚了200万美元）。时间表是5-10年，而不是12-24个月——他愿意就突然实现10%的 GDP 增长与 Harry 打赌：“我们只是刚刚发现了火。”
- 如果只能持有一只股票10年，他会选 Nvidia。 他“害怕巨头，它们不会长久……而我们对算力如饥似渴”；最大的公开做空仓位是 Salesforce；他也不会卖掉 Klaviyo——他是该公司的首位投资人，认为它“被严重低估”。但对新赢家的反向提醒是：Cursor 式收入“可能消失——你很容易切换到 Windsurf……这是一种全新的现象”。
- Hiten Shah、Dharmesh Shah 和 Pat Grady 都称他是自己认识的最佳招聘者。 他的原则是慢招聘、快解雇，先付费试用再发正式 offer，以强度筛人：“打开你的电脑，我现在就要看你的 Git。” 他曾凭朋友的 GitHub，一天招进5名 Northeastern 毕业生，并让他们当天开始交付代码。职位头衔正在消亡，薪酬将转向利润分成、收入分成和更高的工资。
- 他的人生信条是：逆境塑造韧性。 他高中毕业时曾在 Nicaragua 无家可归，最早接触科技是在父亲车库里使用一台偷来的 IBM。他建议早点要孩子（第一个孩子出生时他27岁，17岁认识妻子）：“两件事可以兼顾。不会好看，只会很狼狈。” 但“今天死去我也能心满意足”。
1. “我假装自己是虚无主义者，好应对其实太在乎一切这件事”
- 开场悖论奠定了基调：Torres 的“我不在乎”是铠甲，不是冷漠。创业过程中反复被拒——有人离开他、客户流失——让他明白“你会被反复击倒，最后只会意识到，一切都会没事”。钱大约解决了“1/3的问题”：水管坏了、行程取消了——“付钱解决，继续往前走”。
- Stebbings 认为，给21岁自己的标准建议——“一切都会好起来”——可能会拿走驱动人成功的恐惧。Torres 的版本是：“事情没那么难。” 即便是前沿 AI，告诉陌生人他们可以自己构建一个 LLM，他们也会说太难了——“其实没那么难，代码在屏幕上占一页就够了”。
- 这种判断力来自 Nicaragua：1979年革命时他3岁，住在一所不属于他们的房子里；高中毕业时，房主收回房子，他无家可归，最后靠祖母的绿卡获救。第一台电脑是一台被偷来的 IBM：父亲买下后放进洛杉矶的车库，他用 WordPerfect 和 Lotus 1-2-3 打作业。“这就是为什么移民在创业公司表现出色。”
- 对自己的孩子，他会“制造逆境”，但也认真对待反对意见：女儿曾批评他那条“现在就辞职”的 LinkedIn 动态——“不是每个人都有你今天拥有的条件”。他的回答是：自己刚开始时也没有。
2. 大象跳不了舞——或许能唱歌
- 这种判断可以追溯到他在 IBM 的经历：1998-99年，他在负责让 ibm.com 上线的团队里开发浏览器应用，由此认定一家40万人的公司不可能快速行动。至于 Gerstner 那本著名的书，他说：“他们声称自己在跳舞，但其实没有。也许他们会唱歌。”
- 按公司逐一打分：Google“刚刚把球丢了——他们发明了 Transformer 论文，但推出的所有东西都不好用。NotebookLM，这个伟大的产品？已经死了”。Microsoft 则赢得了真正的尊重——早期投资 OpenAI 是“天才之举……改变世界”——但 Copilot 受益于巨头身份：“他们正在享受先发优势，但没人用这个产品。它就是不好用。”
- 在他的排序里，“王者是 ChatGPT”——唯一真正实现日常使用的产品——“但这不是解决方案本身，也不是未来”。Copilot 和 Cursor 只能排在遥远的第二位。
3. Harry 追问：如果分发能力才是赢家，速度还有意义吗？
- Stebbings 逐一站在巨头一边：Google 的终端触达数十亿用户。Torres 干脆承认：“分发能力是巨头最大的优势”，数据也是；但他仍坚持执行速度的重要性。GPT 出现已经“大约5年”了。“这叫快吗？”
- 随后问题更尖锐：如果技术最终商品化，为什么不等一等，再把技术注入自己的分发体系？Torres 的回答是绝对的：“速度就是一切。速度造就美国梦。”速度创造机会，也让其他人有机会赢。
4. 人才才是堵点
- 大象跳不动舞，根源在于人员规模本身。很多公司有“1000名产品经理”，因为“说不的人不会被解雇”——大家只是在混日子等晋升，每周写“我这周做了5件事”的邮件。他最讨厌的管理仪式是：如果我的日程表全被一对一会议填满，“那就杀了我”；公司里最糟糕的一句话是“Elias，我们能谈谈吗？”
- Agency 只有15人，全公司只有一个会议：“告诉我你昨天做了什么，再告诉我今天要做什么。其他时间我在卖产品，或者帮助客户。” 到了50人，就变成10人一组、各自负责产品——“你以为我是怎么打造 HubSpot 的？”
- 企业销售的现实是：CEO告诉他“好，现在就把 Agency 引进来”，随后把他层层转交给组织里的其他人，每一层都在抵触，最后变成“下个季度我再回复你”。“CEO犯了一个大错……这个决定应该由 CEO 亲自负责。” 如果2000人的公司都这样，到了4000、1万、2万、4万人会怎样？变化不够快的公司“不会长久”。
- 需求信号已经存在：他在上市公司的朋友们“每年花3000万美元做客户成功”，问题甚至不只是钱，而是“我没法足够快地招人、培训人……如果你能训练一个 AI 做这份工作，让纯算力全天候运行，你更愿意选哪个？”
5. 软件仍然假设有人操作——而这正是将要消失的东西
- 他对产品的核心判断是：过去20-30年的软件像汽车一样构建，整个环境都默认由人来驾驶。未来则应当是无人驾驶：“我只想告诉它我要去哪里、什么时候到。” ChatGPT 没通过他的测试——“ChatGPT 仍然假设我必须使用它”——不过他认可这种渐进式变化，就像自动驾驶有不同等级，Operator 正沿着曲线向上走。他也把自己纳入批评：“我对此负有责任。我打造过大量这样的软件。”
- 对“AI只会让每个人更高效，不会减少劳动力”的舒适共识，他直接否定：这是胡说。就连他“喜欢并尊重”的嘉宾预测员工数量会增加，他也难以置信地反问：“什么？” 有人发帖称要再招1000名销售代表，他的回应是：“对不起——如果我们构建的承诺是 AI 将替人类完成工作……”
- 人类接下来可以服务世界其他细分市场——“为什么我们所有人都盯着同一个市场？”——也可以真正换一种生活：“我可以去经营一个农场，种点豆子，看着它们长大。” 学画画，研究法老。“我不会再想，去 Salesforce 找一份工作吧。”
6. 100人做到数十亿美元收入，以及它对风投的影响
- 历史上，每一代产业——工业、互联网、SaaS——都默认规模越大越好，但“因为没有更多人，我们就无法做到数十亿美元收入，所以别无选择”。今天每一家大公司都曾经从小做起。Agents 可能消除这一约束：小得多的公司做到数十亿美元收入，就是“潜力所在”，但他也诚实地保留了余地：“也可能不会按这个方向发展。”
- Stebbings 追问风投模式时，他给出的答案比大多数创始人更进一步：团队更小，需要的资金更少；既然“我们其实只是在用算力替代人”，融资轮次之间被稀释的股权也会大幅减少。因此，早期融资仍是股权融资，但增长融资应转为基于可预测收入的债务融资。“当然，绝对应该这样。”
- 超高速增长也有另一面：Cursor 很棒，因为它增长得太快，快得不可思议——但那是另一种产品形态。你很容易得到它，也很容易失去它；可以切换到 Windsurf，也可以回到 Copilot。“这笔收入可能消失。这是一种全新的现象，我们不知道它会如何发展。” Agency 的应对策略是“终身客户的分发”：用小团队建立大规模的 B2B 信任与数据关系。Stebbings 的担忧仍然存在：一代以 Cursor 为基准的 SaaS 公司，“拿不到融资”。
7. “Drift 是我人生中最大的失败”
- Stebbings 认为这种说法是在逃避——毕竟公司卖了12亿美元——但 Torres 仍坚持：“我想打造伟大的、经久不衰的东西，让客户可以持续使用并获得价值。” 最终没能做到的原因是：“人太多了。是我的问题，是我的错。” 800名员工，其中200人在客户成功团队，而产品却没有增长。“我让别人失望了，也让客户失望了。我不想再这样。”
- 本期最严厉的自我检讨来自这句话：教练们曾建议他放软一些，多给胡萝卜——“如果不是我给了他们那么多胡萝卜、糖果和奖杯，我现在本来会有一家上市公司。” 他对适应型领导的结论是：“祝你好运……我不会再为了别人改变自己。”
- 为什么无从挽救：公司“太大、太低效”。给今天巨头的生存处方是：“如果它们像 Twitter 那样，把自己缩减到20%，打造一款惊艳的产品，并愿意打破所有定价和收入模式，那它们就能活下来。”
- 时间上的遗憾是：“我当时距离 LLM 成熟已经那么近了。” 更刺痛他的参照物是 HubSpot——“我的作品”——如今是一家300亿美元公司，而 Brian 和 Dharmesh 仍然可以参与其中。“那才是我想打造的公司，但我一路犯下的错误，没有建立足够坚固的基础。”
8. 出售公司：进去时无动于衷，出来后陷入低谷
- 他为这次决策辩护，认为那是“我做过最聪明的事——不要被情绪左右”。2021年9月，同行都在乘风上涨：Outreach 的 Manny 和另一家类似公司拿到了“疯狂的估值”，自己的投资人也不断提高他股权的账面价值——“你现在值10亿美元，未来可能值20亿或40亿美元”。他最终以12亿美元出售公司。对他个人而言，这个数字足以“改变人生”。
- 没人告诉你的部分是：“成为无名之辈很难过。” 你带来新的 CEO、新的管理层——自尊心会受伤。公司是你创造的，但你不再相关。随后是7-10个月的旅行和“寻找自己”——他还提到一张关于出售公司后抑郁与悲伤的曲线图。“但那段休息很好，我需要它。” 快问快答环节里，他说钱不会让你快乐，“但确实会让事情容易很多”。
- 他也曾考虑过另一条路：一次长达2小时、喝了很多瓶酒的晚餐，可能是与 David Fialkow 讨论风投公司的运作方式，最终让他打消了念头——“绝对不可能。我是 Builder。” 他宁愿继续打造东西：“这样容易多了，也有趣多了。” 他无法忍受看着创始人走向自己明知错误的道路。
9. 招聘：强度是测试，一切都是红旗
- 他的筛选标准包括：主动性、智商、韧性、执行速度和痴迷程度。另有心理测试团队称，他的工程团队是“我见过最外向的工程团队”。面试刻意带有攻击性：“打开你的电脑，我现在就要看你的 Git。让我看看你提交了什么……砰砰砰。” Glassdoor 上有人写“我讨厌和这个人面试”，这恰恰是设计目标，而不是缺陷：“这就是测试——如果他们能应对我，而且还想和我一起工作，这就是团队成功的原因。”
- 随着经验增加，他反过来了：在 Drift，招聘是数字游戏，“我会说服比应该说服的人更多”。现在的原则是“慢招聘、快解雇”（过去是快招聘、慢解雇），并在过程中加入兼职合同：“让我看看你的韧性，在做其他事情的同时，你能完成多少工作。” 对错误招聘的复盘完全归咎于自己：“我看到了问题，但当时很绝望……我说服自己，一切会没事。” 对那些在业绩指标压力下招进“还不错”候选人的 SaaS 创始人，他说：“我经历过。别这么做。不值得。” 至于背调：“如果你打电话给我，你会听到真话。” 但这个行业只是在反复录用那些大家都知道不行的人。
- Northeastern 的打法是这样的：他看到一个优秀学生，问他“你的朋友是谁？让我看看他的 GitHub”，然后一天发出5个 offer。入职时：“来我家，我们有厨师，给他们做午饭，把所有电脑都寄到家里，然后当天就开始交付代码。” 标准非常明确：产出要达到上周的2-3倍；达不到的人就会离开——“他们入职时就知道这一点。”
- 对简历，他完全同意，连续跳槽、每份工作只待2年是红旗——但随后话锋一转：“一切都是红旗。” 如果你在那里待了很久，也同样是红旗。关键是“你的斜率”。即便职位头衔快速上升，也说明不了太多——“优秀的人用数据，不够优秀的人用故事。” 他预计头衔本身会消失；对于一家100人、收入数十亿美元的公司，薪酬应改为利润分成、收入分成、更多股权和更高工资。员工工作2年后可以离开，“拿走你的股权——它会值数百万美元，也可能数十亿美元——然后去创办下一家公司”。
10. 可交易的判断：服务胜过模型，普及周期更长，做多 NVDA / 做空 CRM
- 在 Drift 和 Agency 之间，他“苦学 LLM”，并通过咨询在6-7个月赚了200万美元。OpenAI 把那些它不愿亲自陪跑的客户线索交给他，包括 Red Bull、Ticketmaster，以及可能是 NBA 和可能是 Klaviyo 的客户。项目从每月2.5万美元起步，扩大到10万美元和50万美元的合同。这段经历支撑了他的观点：“未来5年，AI服务公司将超过AI技术供应商。” Paul Graham 对此嗤之以鼻，但背后的逻辑是他的普及判断：“人类才是普及的最大阻碍，他们需要大量陪跑。”
- 对时间表，他明确站在“更长周期”一边：转型需要5-10年，而不是12-24个月，“市场炒作太多了”。工业革命时期 GDP 增速只有1-2%；他不相信 GDP 会突然增长10-20%，并愿意就10%的 GDP 增长判断下注：“那我们拿点钱来赌。” 总结性的比喻是：“我们只是刚刚发现了火，剩下的事情需要时间。” 大多数行业采用新技术的速度比科技行业更慢，而科技行业自身“适应这项技术的速度也极其缓慢”。
- 快问快答中，未来10年只选一只公开股票，他选 Nvidia：“我害怕巨头，它们不会长久，但……我们对算力如饥似渴。” 最大的公开做空仓位是 Salesforce——“坏了，Benioff 要杀了我”——而 Stebbings 说，他问过的每个人都给出同一个答案。至于他不会碰的私人仓位，是 Klaviyo；他是首位投资人，也曾说服 Andrew 接受风投资金。“我仍然认为它被严重低估。这就是为什么我不卖股票。”
- 后 ZIRP 时代，他重新审视了自己：“盈利从来都不是讨论重点——Pat 曾经给我上过一堂自由现金流的课。” 如今他“比任何事情都更关注一家经久不衰的公司……现在比起钱，我更受遗产驱动”。目标是用100人实现盈利，打造一家能够长期存在的公司，同时牢记 David Cancel 的教训：在“快还是伟大”之间，答案“永远是两者兼得”。
核验说明
- 字幕将风投公司晚餐的嘉宾显示为“David falo”；“可能是 David Fialkow”属于不确定读法。
- 咨询客户在字幕中显示为“clayo”；“可能是 Klaviyo”属于不确定读法。
Drift is the biggest failure in my life. I wanted to build something great. I wanted to build something sustainable and enduring, something that customers could keep using and continue getting value from.
You walk around your own company and you don't matter anymore. You bring in a new CEO and new management. It hurts the ego because you created this, and you're no longer relevant.
Elias, dude, I cannot wait for this. Julian Beck and Pat Grady, two of my closest buddies, said this is the show that we have to do. I'm so thrilled that we can make it happen in person, so thank you for joining me.
I wanted to start with something that you tweeted before, because it really hit home for me. You said, “I don't care.” People hear that phrase, and many versions of it, often from me. The reality is that I pretend I'm a nihilist to cope with the fact that I actually care way too much.
I wanted to start with this because, at a Jordan Peterson event, someone asked, “The depth of my consciousness causes me to suffer. Is it a blessing or a curse to feel so very deeply?” It feels like a deep start, but how do you think about that?
I think it's a great quote. It's funny because I don't tweet that much, and when I do, Brian Halligan tells me, “Why are you tweeting? Get back to work.” I don't want to tweet because it's hard for me, but I really think it's a paradox.
We always want just one answer. In my case, I find myself saying that a lot: “I don't care. I don't care.” People are always wondering, “Does he care or does he not care? What is he doing?”
It's something that I've found later in my life, with the success that I've had, my family, and everything I've been able to accomplish. I'm close to 50 now, and those experiences have allowed me to realize what I care about and what I don't care about.
The life of an entrepreneur is really hard. Most people don't want to be rejected, and most people don't want to be told no. I've been told no so many times. I've had many people quit on me, and I've had many customers churn, so you get beaten up so much that you realize it's going to be okay.
1. Do Rich Founders Make Better Founders: How Backgrounds Shape You
When I say I don't care, it's not that I don't care. I do care, but I know it's going to be okay. It's going to be fine. I just don't care as much anymore.
Do you care less with more money? Do you care less about some things?
Absolutely. There are so many things that we worry about: the plumbing broke, the car needs new tires, a trip got canceled, or you have to pay a fee. That represents a huge amount of stress that I just don't care about anymore.
It takes away maybe one-third of the problems, and I'm just like, “Okay, I don't care. Just pay it and move on.”
I interview many incredible founders like you, and often I ask, “What would you most like to tell your 21-year-old self?” So many say, “I wish I could tell myself that it would all be okay.”
I think that's terrible advice, because the fear that it won't be okay drives you to succeed.
I think it's true that things are going to be okay, meaning that wherever you end up, it's going to be okay. Complacency is the secret of happiness.
What I would tell my younger self is that things are not as hard as they seem. That's what I would tell my young self.
What do you mean by that?
We put so many roadblocks in front of ourselves. We say, “I can't start a company. It's too hard.” But it's actually not that hard. Starting it, finishing it, and exiting are different things, but we think everything is too hard.
If I tell a random person on the street, “You can build an LLM yourself,” everybody is going to say it's too hard. Is it too hard? It's really not that hard. The code fits on a page on the screen. It's really that small.
Maybe it's because I grew up in a communist country, with nothing, no future, and no hope. Everything I've done, I'm like, “Fuck, that wasn't that hard.”
For example, people at my companies in the past would get gifts, food, free lunch, and then complain that the coffee wasn't that great. I'm like, “It's free. If you don't like it, go buy your own.”
The things that people complain about in the United States are not that hard compared with living in countries like that. I just came back from Cairo last night. You see the slums and all these areas. It's hard to live in countries like that.
I'm from Nicaragua, so life is not that hard for us in the United States.
Have we lost the grit in the US and the UK, where we have the welfare systems and the nice cultures?
I think that's why immigrants do well in startups. Grit is everywhere. It's not distributed only to the people who have nothing. We just get beaten up so much that we develop it in larger quantities.
My kids are going to grow up differently. They're going to have the ability to work with my companies, my network, my access, and my money.
I wrote a post on LinkedIn that said, “Quit now. Your company will be fine. It's time. Don't worry about that bonus, title, or salary.” My daughter said, “Dad, why are you posting things like that? Not everybody has the luxury that you have today.”
I didn't have the luxury that I have today when I first started. I have the right to post that; I'm not making it up. She said, “The problem is me. I can't say those things because I had all those things from the beginning.”
I speak to many very successful people, and one of the biggest challenges is how to bring your children up with a lot of money and have them be as hungry and gritty as we had to be.
I just try to manufacture adversity. That's all I can do for them.
The good thing is that I had success when they were older, because I had my kids so young. They're only starting to get spoiled now. We're in London with the whole family, and I said, “Do you guys want to go to Cairo for the weekend?” They said, “Yes, we want to see the pyramids.” That would be hard for them.
You mentioned typing away on the little IBM computer in the garage. IBM was your first job, I believe. Talk to me about that and how it impacted you.
IBM was a very special group. This was the group that put IBM.com on the internet. We were doing crazy stuff. It was an advanced internet technology group.
I was building applications using HTML and JavaScript in 1998 and 1999. We were building browser-based apps with open-source projects and things like that. It was insanity what we did.
I learned so much from these crazy visionaries, but what I learned was that a 400,000-person company just can't move as fast. Lou Gerstner was the CEO at the time, and he wrote a book called Who Says Elephants Can't Dance?
2. Why are Incumbents Slower than Ever
I don't disagree with him. Elephants can't dance. They claim they're dancing, but they don't. They can sing, maybe.
The best interviewers are able to move with schedules and be very flexible about how they take conversations. I wanted to touch on that at the end, but you give me, “Elephants can't dance,” so let's go.
Have we not seen, bluntly, over the last year or two that elephants can dance? Adobe moved incredibly fast on product. Google is moving incredibly fast in terms of integrating LLMs and building. We're seeing the biggest incumbents move at speeds we've never seen before. How do you think about that?
I completely disagree. These elephants are just talking. They're singing and talking a big game. It's mostly talk.
Google fucked it up. They invented the Transformer paper, and everything they put out doesn't work. NotebookLM was the great one, but it's already dead—the project. They're trying so hard, but they can't do it.
Microsoft Copilot—I felt like Microsoft was moving. Investing in OpenAI was genius. [Likely Satya Nadella] is the man, no doubt, and being that early, dropping all that money on OpenAI was genius. It changes the world.
They had the advantage of adopting all of this early, but look at search. Have you tried Perplexity? They dropped the ball. With Copilot, everybody thought Microsoft was just ahead and making all the money first.
The problem is that people have such a hard time trusting someone. They're reaping the benefit of being early, but nobody is using the product. It just doesn't work. It's tough.
The king is ChatGPT. That's the one that everybody is adopting and using every day. But that's not the real solution, and it's not the future. Copilot and Cursor are a distant second.
I've got you. I'm going to take these one by one; otherwise, I'm not going to have any coherent structure.
Google, I completely disagree with you. I think the fundamental thing is access to consumers and distribution. When you look at all the endpoints they have to consumers, is the product there today?
No, 100%. But their ability to access billions of people—
They have the endpoints and the data, too. Those are great assets.
Those are great assets, but it's been three years—it's been five years since we had GPT. Is that fast?
I'm forward-looking by nature.
Okay, okay. You're forward-looking.
With Microsoft, I completely agree with you on Copilot. I think they will get amazing adoption. It's very much like Teams, where people in large enterprises just take it because of distribution.
I agree. Distribution is the biggest advantage that incumbents have.
But does speed matter if you can be slow as hell, wait for someone else to build it, let the technology become commoditized, and then embrace it into your distribution advantage?
Speed is everything. Speed is what creates the American dream. It's the opportunity. Speed creates the opportunity for others to win.
Why do you think ChatGPT isn't the answer?
I want to give you this insight. Software built in the past 20 or 30 years had the assumption that it was built like cars. We build the entire car and the environment around it, and then a human operates the car.
The future is not going to require operators. It's going to be like building driverless cars. They're completely different, except that we're stuck with the roads, signals, and all this other stuff. Imagine when we can completely change transportation.
We need to build software with the assumption that I don't want to know how to drive the car or do all this work. I just want to tell it where I want to go and when I want to go, and maybe choose the route.
If we just start with a little bit of Copilot, you see that, right? ChatGPT still assumes that I have to use it. I'm the operator.
I totally get it. Is that not, step-function-wise, moving toward that? It has to start somewhere, and with Operator, we see it doing more and more of that full work process.
That's the future I'm visualizing. How do we get to a truly agentic world where the LLM is doing the work for me? I just want to start it.
I want to get a little bit of causation there. Why can elephants not dance?
My last company, Drift, before we sold it to Vista Equity Partners, had 800 people. I was consumed. I know big companies that have 1,000 product managers. Imagine how many Microsoft has, or Salesforce.
Most of those people have no incentive to say yes. Nobody gets fired for saying no, so everybody will not buy the right product or build the right thing. It's easier to coast.
When you start spending all this time on all these people, focusing on promotions, and saying, “I have to go home; I'm working to rule; I have to send an email at the end of the week saying five things I did this week,” you're toast.
You can't move because you have to get all these approvals. When you're the CEO of a public company and you spend all this time reporting to analysts, dealing with the press, and doing all this other stuff, how much time are you spending with customers? How much time are you spending building the product? Who has the vision? Who makes the big decisions?
What conventional management lesson that we're all told do you think is the most bullshit?
If you're not working with the person every day, why do you need this special time to go talk to them? I hate it when somebody says, “Elias, can we talk?” Oh my God, that's not good. They're the worst messages.
Then you go into those one-on-ones and nothing happens. Why didn't you just tell me during that one-on-one? If my whole schedule is filled with one-on-ones, shoot me right now.
At Agency, I have no meetings with anybody. We're about 15 people, and we have one meeting for the whole company, for the whole team. We show what we did yesterday and what we're going to do today. That's all we do in that meeting.
The rest of the time, I'm selling or helping customers.
At some point, that doesn't scale. When you have 50 people, you can't do that as well.
We create 5 groups of 10. That's absolutely how I built HubSpot with the team. We created small teams that operate independently. You own that product. You're responsible for that product.
Let's go to the recruitment side. Why can't elephants become cheetahs? When we look at Twitter, we saw a massive reorganization of the structure of teams and the size of the company. Why can't elephants transition to this new world?
They're in a pretzel. Every generation—industrial, commercial internet, SaaS, Web 2.0—always assumed that larger was better. Everybody's dream was, “I want to build a large company.”
We're entering a new world where the proposition is that smaller is better. We've always known that smaller is better, but because we couldn't achieve billions in revenue without more people, we had no choice.
Every big company today started small. We will have much smaller companies with billions in revenue. That's the potential.
That's agent-fueled, because it has to be.
It has to be. If not, you can't keep hiring. I had 800 people, including 200 in customer success. You have no idea how hard that was.
I have friends at public companies who say, “Elias, help me with Agency. I'm spending $30 million on customer success.” It's not even the money. I have the money, and we can make it profitable. I can't hire and train people fast enough to support our customers.
Then you have layers of management, benefits, recruitment, and onboarding. You understand how that blows everything up. If you could train an AI to do that job and just have straight compute do it 24/7, which one do you prefer?
Cool. That's going to lead naturally to smaller teams.
We still need more compute. We're really trading people for compute, so maybe not.
If you're trading people for compute, does that change your equity-to-debt preference?
Of course. A lot less equity will be spent on the team because you're not going to need round after round, distributing equity to hundreds and hundreds of people after every round.
Early financing remains, but growth financing becomes debt financing on a more predictable revenue basis.
In a world of AI, what changes about the way that you start companies?
Everything needs to be done faster. The skill sets you need and the people you're looking for are very different.
I have to ask: when you said B2B changes dramatically, what in B2B will change most significantly in your mind?
It's the notion that you need operators. All this B2B software was designed to require operators, and it just kills me. It's the real problem. I've been guilty of this; I built a lot of that software.
How do you respond to the common suggestion that we'll just do different things, find more ways to be productive and creative, and actually not reduce the workforce? We'll just become more productive.
That's bullshit. I have guests here whom I love and respect, and they're like, “Oh yeah, we're going to get larger in 5 years.” I'm like, “What?”
I saw somebody post on Twitter, “We're going to hire 1,000 new sales reps.” If we're building the promise that AI is going to do the work of humans, why are we hiring 1,000 new sales reps?
Humans are going to do things differently, but perhaps at their own companies. Why do we all have to be focused on all the B2B companies in the United States? There's a whole world out there. There are so many niches and so many things that need to be built to improve the lives of other human beings.
Why are we all focused on the same market? Let's go travel. Let's go feed people. Let's go farm. If I weren't building Agency, I'd have a farm and plant some beans and watch them grow.
There are other things we could do with ourselves. I want to learn how to paint. I was just at the Grand Egyptian Museum, and I want to study that stuff. I want to learn about different pharaohs and kings. There's so much we can be doing.
You don't worry about the future of human labor and productivity?
No. I'm not going to say, “Let me take another job at Salesforce.” That's great, but no.
You don't want a job at Salesforce?
No.
I was hoping to set you up after this. You have connections. Do you know anybody?
You know me and Marc go way back.
I had to email Marc Benioff 53 times to get him on the show. I got him. I like it. I got him.
I spoke to Hiten Shah, obviously. I spoke to Dharmesh Shah. I spoke to Pat Grady. They all said you're literally the best they know when it comes to recruiting, which I thought was really interesting to hear as the one commonality from all 3.
You recruited Andrew Clavijo, Whitney, who's now HubSpot's CTO, and Chris O'Donnell at Day. What is your hiring criteria? Everybody wants to know the secret.
My hiring criteria is high agency. You have to have IQ, grit, execution speed, and obsession.
The way I recruit is weird. I'm an engineer, but people have done psychology tests on the teams I built, and they're all like, “This is the most extroverted engineering team I've ever seen in my life.” They're all intense, and they all talk to customers.
I don't tend to hire that many introverted engineers. The way I recruit is so intense and so in your face. I need you to make a decision now. I'm going to ask you to go on the most dangerous expedition ever: We have to take HubSpot public, build Drift from scratch, or create the fastest company to $200 million in revenue.
That scares a lot of people. I rub a lot of people the wrong way during the interview process.
What do you think rubs them the wrong way?
I'm too intense. People say, “Calm down.” They leave Glassdoor reviews saying, “I hated interviewing with this guy at Drift.”
How does your intensity show itself?
The speed of my questions. I open up your computer and say, “I want to see your GitHub right now. Show me what you committed. Show me your code. Let's discuss that. Did you really build that? Who was it for?”
I don't care about my questions. I just go bang, bang, bang, bang. I want to see you worry.
Do you worry that, in some cases, that won't get the best out of someone who's perhaps shy?
That's the test. If they can handle me and want to work with me, that's what makes the teams work.
Do you need to change your management style to get the best out of different people? We often say, “This is the culture,” as if culture is one homogenous, unified thing. Actually, some people benefit from the carrot and some from the stick. Do you need to change who you are as a leader to get the best out of them?
Good luck. Why? I'm done changing for people. I don't care. Screw all that.
Look at Drift. We had 800 people. I hired coaches, asked people to watch me, and asked for feedback. They said, “Elias, you should use more carrots here,” or “You should do more of this.”
That's why I would have a public company right now if it weren't for all those carrots, candy bars, and trophies that I gave out.
Unpack that. What do you mean you would have a public company if you hadn't given the carrots?
I sold the company to private equity.
Why wouldn't you want to have a public company?
Maybe I don't want a public company.
Do you consider Drift a shortfall of its success?
Drift is the biggest failure in my life.
What do you mean by that?
I wanted to build something great. I wanted to build something sustainable and enduring, where customers could keep using the product and continue getting value from it.
Why didn't that transpire?
Too many people. My problem, my fault, was not making the product grow and not making it better. You want to make a product where customers receive value right away.
3. Why Was Selling Drift For $1.2BN a Massive Failure
What did you not do with the product that you wish you had done?
Speed and execution. The timing was off. I was so close to the LLMs being there and transforming what I'm doing now.
I'm excited about what I'm doing now because I get to start from scratch. That's really good in a way. I wouldn't necessarily want to be transitioned into a large ship right now, but it was tough.
I don't think it was my greatest accomplishment. I let people down. I let my customers down. I don't want to do that again.
Why was it not salvageable? Why couldn't you transition it into a public company?
It's the way we built companies. They're too big and too inefficient.
To make things salvageable, big companies would have to do what Twitter did: reduce themselves to 20%, build an amazing product, and be willing to break all the pricing models and change all of that. Then they might survive.
Talk to me about the sale process. Was that difficult for you?
I told you, nothing is hard. It wasn't difficult. I'm an immigrant, and I'm a realist. I accept what it is. I don't let my pride get the best of me too much.
I think it was the smartest thing I did: not being emotional about it. A lot of people, even my investors, would say, “You're worth $1 billion now. You could be worth $2 billion or $4 billion.” They multiply my equity by that number.
I did this in September 2021, so there were a lot of peers of mine getting $4 billion valuations. We started a little earlier than some of those waves, so other companies were getting crazy valuations.
Outreach's Manny is a good friend of mine. He's in London. You see all that, and it affects your ego. You think, “Why am I not being valued when I'm a similar company?”
How much did you sell for?
$1.2 billion.
Can I be super blunt? How much did you take from that?
I'm not going to say. It changed my life.
Fuck your money, for sure. How do you think about your relationship with money?
It reduces a lot of the physical and material stress. It's such a blessing for an immigrant who had nothing not to worry about those things.
I focus on the small stuff. Get a car whenever you need it. Stay at whatever hotel you want. Fly wherever you want. Eat wherever you want. Those are the basics. It's great not to worry.
I believe a lot in assistance. Eventually, we'll have that with AI. People are wasting their time trying to be frugal. I want to focus my time on the customers, my team, and my people. I don't want to do anything else.
Before we go back to customers—and by the way, you're throwing me all these questions that weren't on the list.
Don't mind me. It's cool.
I always do this. You clearly haven't seen all the shows.
4. How to Hire F******* Rockstars
We're going to go to customers, but I want to stay on hiring. Is there a process, and has the process changed over time?
There's a big difference. I used to be intense and treat it as a numbers game. We wanted more, more, more, so I had to move quickly. I would force people and convince people, and I would convince more people than I should have. I could convince anyone. I'm a good recruiter.
I realized that isn't the same thing. Now I'm pushing back more. More people know me, more people know my reputation, and more people want to join, but it's important not to fall for that.
Some people would reverse-engineer my process. They knew how to talk to me in a way that touched the right points. Now it doesn't matter what you say to me. I have to reject, reject, reject—push back, push back, push back.
If you don't handle the pushback, I do a lot of contracting now. You have to work for me part-time. Either you're available or you're not. You need to show me your grit and how much work you can do while you're doing something else.
I used to hire fast and fire slowly. Now I hire slowly and fire fast.
People never fire fast enough.
Every time I fire someone—even though I'm a good firer now—I would always think, “Everybody knew.” It became harder and harder.
Think about the large incumbents and how long it takes them to fire people.
Years.
Years. In Germany, it's almost structurally impossible to fire someone.
That's Europe.
Let's not talk about Europe.
We're in Europe.
Exactly. We can say anything we want about Europe.
People just need to be more careful. My advice to early founders is that they have so much more to learn. They have all these high hopes for the people they hire, but they don't realize that most likely they're going to be let down. It takes a long time to develop intuition for recruiting.
I had a guy working for me who tried for 1 week, and then he didn't get anything done in that week. I also hired a whole bunch of young kids out of college. I hired 5 kids from Northeastern in 1 day.
I've built all my companies on the backs of Northeastern students. They're amazing people. They're the best. I love the university, and all my kids are going there.
I used to hire 1 or 2 every time. My co-founder was from Northeastern, and he was an early founding engineer at Drift. This time, I saw 1 person and said, “Great. Who's your friend?” I looked at that person's GitHub. Then I asked, “Who's your other friend?” and looked at that GitHub. I made 5 offers, and they all started working.
How do you onboard 5 new people in a day?
I tell them to come to my house. We have a chef and cook lunch for them, and then we ship all the laptops to the house. We ship code that day.
Doesn't that cause dysfunction and misalignment when there are 15 people?
No. We're like, “Okay, we're going to do 2 or 3 times more than we did last week because these people are here.” Whoever doesn't make that goes, and they know that coming in.
Seriously, how quickly do you know when someone isn't good enough?
Most people should know within the first week or 2. You should be getting rid of people during the interview process or in the first week or 2 because they're interviewing you.
The hard thing for people is that they hire someone, and 2 weeks in they realize that person isn't good enough. It looks bad to get rid of them after 2 weeks, because they feel they've barely given them a chance.
Those people have to change their process. They either need to do more research, more interviewing, more testing, find the right talent, or do the right reference checks. Or they need to be willing to make the decision.
That's why I tell people I'll pay them on a contract at an hourly rate. I let them be. Show me what you've got.
Are the B hires always obvious?
No, not true. I've hired thousands of people. There are so many people where you're like, “What a gem. What a diamond in the rough.” There are a lot of those. I have time to find them.
5. The Biggest Mistakes Founders Make in Hiring
I often find that boring people in interview processes tend to be the best hires, because they're never as boring as they appear in an interview process.
Boring is great. I've had people my team interviewed, and they said, “Elias, this guy is not good.”
I went back in. I was quiet and shy, but they had grit. They got shit done. They showed me their code. I went back to the team and said, “By the way, we're hiring this person.”
At HubSpot and Drift, I would have a panel of people interview candidates. I would let them interview and have fun and tell me what they thought. Then I would make the hiring decision regardless of what the panel said.
Many times they would come back and say, “That was really good.” They were better than the people who had interviewed them. Those were the compromises I was making before. Now I'm getting leaner and leaner and just trying to find the best of the best.
When you've made mis-hires, what did you not see that you wish you had seen?
The biggest failure in mis-hires is myself. I saw it, but maybe I was desperate. I needed people. I needed to build things. I needed help, and I convinced myself that it was going to be okay.
A lot of SaaS founders will listen to this and say, “I get it, Elias, but I need some fucking sales. I need to hit my numbers. My VC is telling me I need this, and I need to hit the quarter. They're good enough, but are they perfect?”
What do you say to them in that position?
I've been there. I hear you. I feel you. I've made that mistake. Don't do it. It is not worth it.
I invest in everything, so founders come to me for advice. I tell them, “Don't do it. That's the wrong person.” Then they say, “No, but I need to hire this person.”
I see people living this lie. They convince themselves they're going to do it, hire the person, and keep them for 2 years.
People I've let go, people I made mistakes on, get hired by somebody else. The whole industry keeps taking these people over and over. Everybody says, “They were great,” and nobody says what actually happened.
If you call me, you'll hear the truth. I don't want to lie to a founder.
One of the things I hate when I look at CVs and LinkedIn profiles is what I call “bouncers”—people who spend 2 years at Drift, 2 years at HubSpot, and then 2 years somewhere else. At least 2 years is better than 1 year.
Americans are incredibly promiscuous when it comes to jobs. Do you agree with me about hating that bounce, bounce, bounce?
100%. Everything is a red flag. If you stayed somewhere for a long time, that's a red flag too.
Why?
You can see whether there was title acceleration.
Exactly.
That's what I'm saying. Even title acceleration doesn't mean anything. You have to probe deeply when you're interviewing.
I always find that great people use data. Not-great people use stories.
You said title acceleration doesn't mean much. I saw on your Twitter that you're not a fan of titles and think titles will be removed. Why do you think that, and why are titles not useful?
We're just ego-driven humans. We want to build empires, and everybody is obsessed with titles.
I've hired so many young people who pull me aside and say, “Elias, can I get a shot at being a VP at your company?” I'm like, “You just graduated from college. Back off.”
We obsess over all that stuff. I'm looking for people who don't really care about it. We have to change the way we reward the team, whether that's profit sharing, revenue sharing, more equity, or higher salaries.
I'm thinking about these things constantly. If I'm going to have 100 people and $1 billion in revenue, how am I going to compensate them? How am I going to retain them?
Maybe they burn out after 2 years. Take a break. Here's your equity. It's going to be worth millions, maybe billions. Then go start something else.
If I'm a founder starting a company in this age and listening to you, what can I do to change the reward and incentive structures of companies to make it better for my teams?
First of all, at all the companies I've been at, I was never calculating how much money I was going to get. I was doing it for the love of the game.
To me, being in that world and creating compared with my IBM days was incredible. At IBM, I was frustrated because I couldn't do anything. People would leave at 5:00 and not care. They'd say, “I'm an STSM. I'm a distinguished engineer,” and then go home and do pottery. Kill me.
I wanted to be in it: creating, building impact, talking to my customers, staying busy all day, and working with my team. If you don't want to be in that, don't do it.
If you're in it for the money, forget about it. Join the party. Come and create value, and the rest will take care of itself.
How do you think about the distribution of profits? You have equity, but you can also do profit distribution and dividends. What do you think is most effective?
Let's get to profitability. Let's see how we build this company. Let's see what Pat supports.
Is your mindset toward profitability in this journey different from where it was previously?
Yes. We were taught not to worry about it. Profitability was never a discussion. Pat gave me a class on free cash flow at one point because there were some big players in the Sequoia portfolio doing things like that.
Now I think about it a lot more. I'm thinking more about building an enduring company than anything else. Whatever that takes, if I can make this profitable with 100 people and make something last longer, I'm more driven by legacy now than by money.
How do you think about the focus on profit today reducing the growth potential of tomorrow? You will make trade-offs in resource allocation to hit profitability earlier, and that could hamper your growth.
David Cancel says to me all the time, “It's always both.” It's a paradox.
When we were at Drift, he would say, “Just build me another CRM,” or, “Build me another marketing automation platform.” I would say, “Which one do you want? Do you want a CRM or a marketing automation platform? Do you want a chatbot or this?”
I would always make him choose, and he would always say, “I want both.”
It took me a long time to learn that. You have to focus, but it's not a simple, clear answer. You have to do more than one thing at once.
The same is true with companies. I want profitability, but that doesn't mean the rules are changing. The success of a company depends on its brand and product differentiation. Those things don't change.
How do we scale it and grow it? There's not one company to rule them all. Cursor is great because it can grow so fast—it's unbelievable—but that's a different style of product. It's something you can gain easily and lose easily.
One thing I'm worried about is funding for traditional enterprise companies. I'm a little younger than you, but I grew up in the same SaaS days, so to speak, with triple-triple-double-double. Those were the revenue expectations for founders.
Now we've seen several companies go to $50 million or $100 million in revenue in 1 or 2 years, and the revenue expectations have changed so much. I worry that there's a generation of SaaS companies that won't get funding because they don't meet the metrics.
It's really bad. There was a moment when we were worried about profitability, post-ZIRP, and we said, “Now we're going to build better companies.” Then Cursor came along, and everybody wanted the same thing.
Luckily, I'm not falling into that peer pressure. What I really care about is how I gain the distribution of lifetime customers. How do I get that?
Remember, I'm B2B. Let's not confuse this with consumer tech, personal data, shopping, or e-commerce. Those are different spaces. I'm a B2B SaaS guy.
I want to help businesses achieve the dream of high efficiency and profitability with small teams, while serving their customers better. You have to gain their trust for them to share their data with you. You have to build a relationship.
I'm working on doing that at scale with a small team. That's what I'm out here to prove.
With Cursor and all that stuff, you can easily switch to Windsurf, switch to something else, or go back to Copilot. That revenue could disappear. This is a brand-new phenomenon, and we don't know how it plays out.
Do you think a lot of VC money is being burned in this wave?
VC money is always burned. VC is a small asset compared with the rest of what happens in the world. Go to Dubai and all these other places. It's an exciting place. It's the American dream.
You said you've got good at firing. It's not a nice thing to do, and I think it's hard for younger founders. What would be your biggest advice to me on how to get good at letting people go?
I've learned not to make big promises that I cannot fulfill. People don't realize, when I'm speaking, how accurate I am because I've had so many conversations before.
I'm not promising shit to anyone. I'm like, “I don't know. All I promise you is a lot of work and me yelling. That's it. I promise you that.”
Do you enjoy fundraising, hiring, and firing, which are crucial parts of being a CEO?
Not really. I like deals. Whether I'm hiring someone and closing the deal, hiring a customer and closing the deal, or fundraising, it's about closing the deal.
When we were selling to private equity, we had bankers whom the board asked me to hire. I didn't really need them, but I was like, “I like this job. This will be great.” Having 2 or 3 deals in flight every month was exciting.
In reality, I'm a relationship guy. I care more about what happens after the deal and the relationship I have with the VC. I love Pat. I love Azeem from CRV, Fialkow, and Larry Bohn from General Catalyst. Those are my Jewish uncles.
I like people. I like being with people.
Is selling a company what we're told it is? Talk me through the pros and cons.
It was sad to be nobody. You walk around your own company and you don't matter anymore. You bring in a new CEO and new management, and it hurts the ego. You created this, and you're no longer relevant.
Or maybe you couldn't take it further. It was a great experience for me to have the post-Drift days. I spent 7 to 10 months traveling and trying to find myself. It was a tough journey.
There are graphs showing what happens with the depression and sadness: You're nobody, you have no job, and you have no meetings. But it was great. I needed that break.
Do you think you still have an ego?
Of course. Everybody has an ego.
Is it dangerous?
Ego is dangerous, but ego drives us too. It's like the care-don't-care paradox.
6. Everything You Think You Know About Working Parents is Wrong
I spoke to Julian before, and he told me that you advocate having children as soon as possible. I wanted to touch on this. Why do you think it's so important to have kids early?
When I was growing up, everybody said we had too many humans on the planet. I remember hearing that the global population would be 12 billion. Now the problem is the reverse. We don't have enough children, and societies are going to collapse. That's what we think, unless I've been misinformed.
Brian Johnson is trying to extend longevity, but I don't think we're very far away from figuring that out. I think we're farther away from figuring out how to have healthy babies at a later age.
The best time to have children is when we're younger, unless we have unlimited resources. Even then, there's your health and your ability. We have improved the possibility of having children with IVF and those things, but I've heard that the journey is really hard and difficult.
I was very fortunate. I met my wife when I was 17, and we had our first child when I was 27. I have 3 kids. They're 18, 19, and 21, and they're wonderful.
We just went to Cairo together. They're my squad. They're my team. Now they're like, “Can we go to Amsterdam this week?” I'm like, “I have to work. You guys can go have fun.”
I was blessed because most people think it's one or the other, but in reality, you can do both. It's not going to be pretty. It's going to be ugly.
While everybody was asking, “How can you do both?”, my wife was incredible. She was at home with the kids, and I would work all day at IBM, then go home for dinner, bathe them, put them to bed, and do the dishes.
It was exhausting. There was no sleep, no money, no help at home, and we were far away from our family. It was horrible.
Entrepreneurs are like athletes. This sounds awful and arrogant, so I might have to edit it out, but I have a very honed life. Everything I do is selfish, so that I can perform better and be the best I can be in different respects.
When you have kids and a wife, you're no longer the most important thing. Your schedule goes to hell, and your ability to inspire in that meeting is completely off. Your ability in fundraising is affected. You're not performing at your peak.
We have different definitions of peak. I was able to do all that. I raised hundreds of millions of dollars while going through all of it.
You have 3 kids at home who don't take a break and don't take naps. You're exhausted. It's Sunday afternoon, and you think, “Can I take a nap?” You can't.
People would talk to me in my younger days and say, “What did you do this weekend?” I hated that question at IBM, because they would say, “I went to Egypt for the weekend.”
I'd say, “What did you do?” I had spent all day dealing with outages from the company and the startup. My wife would say on Saturday, “Let's go out. Let's get out of the house.”
I'd say, “Okay, I'll be done coding this feature, finally without interruptions, by noon. Then we'll take the kids out.” It would be 3:00 p.m. before we got out.
And I would still be coding, right? We would go to the beach, and I would carry my laptop. David would be buying me a modem or something because I needed to respond, and if there was an outage, who else was going to do it but me?
So my life sucked, right? I never traveled. I never did anything but work, the kids, and the family. I traded it off.
Now I get to go wherever I want, right? But, unfortunately, that was how I was able to build a company, be successful, and have my children. I'm more than blessed, I'm more than grateful, and I can die happy today, right?
Other friends were either successful or had a family, and now they want to go and start the next journey in their lives, right? I was able to do both. But regardless, when you have everything, when you have money, when you have all these things, if you get to a later age and you don't have kids—not that everybody should, and I'm sure people are going to judge me, like, "What if I don't want to have kids?" Respect me, I'm fine—but the reality is, I see that we want something more than just material things, and having a family is a great thing.
Totally agree with you. I've just become an uncle for the first time, and it's—I've seen your pictures. You're getting soft. You're soft, dude. I love this—a beautiful little girl.
Your wife was the first woman that you met in America, I think?
Yeah.
When you were 19?
17.
After so many years of an amazing marriage, what would be your biggest piece of advice to someone listening on how to sustain an amazing marriage?
She's always right.
It's like, I just can't get my wife to apologize for anything, no matter how right I think I am. I'm right at work all the time, but not at home.
The first thing that came to my mind—we always talk about that—but it's really about selfishness, right? You have to realize that it's not about you. I think social media has ruined us by making us think about what a perfect life was, what you see in social media.
It is about you, and I don't mean that badly, but you mentioned Cairo, you mentioned Amsterdam, you mentioned the nice things that they have. Well, now I'm being more selfish. That's all enabled by you. It is all about you.
It is, it is. But we were supposed to go to Cairo. I have a friend, and he invited me, but then my friend wasn't responding, so I said to the kids, "Let's go to Amsterdam instead." My kids were like, "We want to go to Egypt." Okay, okay, let's make it happen, right?
I think you have to balance the selfishness, right? Yes, we are selfish beings, but I'm telling you, you get a lot more pleasure when others are around you, right? It has to be a little bit less about us.
So that's marriage. On the fatherhood side, when you think back over the 21 years of fatherhood, is there anything that you know now that you wish you had known when you became a father 21 years ago, when you were 27?
I guess, from what I saw other parents doing in the past—for example, my kids' love languages are touch. All 3 of them are like that. My daughter says, "From you, Dad, I want words of affirmation." So I've learned to overcome that, like, "Oh, I don't want to be this dad." We had these images and so on.
On one side, you learn to be more loving and caring. On the other side, I wish I would have been tougher with them. Right now, I try to be tough too, because nothing compares to how I grew up. I was beaten the shit out of by my mom. I can go into specifics, and I love all that. I love my mother because of what she did at the time with what she knew, but she made me tougher.
I think we're too soft on kids. Sorry, we're too soft on our kids.
7. Quickfire
But what do we do when they say they want words of affirmation? You can give them words of affirmation and be tough at the same time.
Back to the paradox, right? Words of affirmation don't mean that you can't be tough. When you're like, "I've got to manufacture adversity," you've got to be tough, man. You can't be weak. You can't be soft.
Has there ever been a time when you felt like you were letting them down as a father?
Oh, all the time, right? I saw all these parents, and it would be like, "Oh, I just restored this car with my kids," or, "My kid is an Olympian," or, "My kid is going to be a pianist, a world pianist." I'm like, "Fuck, I had no money. I couldn't do tutors for them. I couldn't do any of that stuff."
So then I was like, "Maybe I'm going to teach them how to be engineers." That's one thing I could teach them—how to code early—and I still failed at that because I was so busy. It was either doing the dishes or making sure the systems were up at the company. So I felt like I let them down.
I ask because I feel like, bluntly, I'm a constant failure. I'm a founder of a media company, an ambassador, a son, a brother, a boyfriend, an uncle, and a board member. At every moment, I'm letting down multiple people.
Yeah, that's why I said I'm letting them down. But the reality is that no one can do a perfect job for anyone. What matters is what you're able to do in the good moments, right? Whether it was a tough lesson or some great words of affirmation or touch.
That's kind of why I don't care. I have regrets. I'm trying to do the best I can, and I do good things because I'm thinking about them. But I don't worry if I don't do 100% of what you wanted, what she wanted, and what they wanted. I'm just like, "Hey, I'm doing—leave me alone," you know what I mean?
I want to finish on this. You retweeted Marc Andreessen's tweet on the commonality of failure in great people's careers, and I think it's a really interesting question. What was your favorite failure?
A favorite failure is often attributed with great lessons and takeaways on how it changed you.
I think Drift was a great failure, right?
It's both, right? You sold it for $1.2 billion, so it's a bit of a cop-out of a failure, my friend.
It is, but what I'm saying is, look, what I've learned was—
Do you think David saw it as a failure?
Yes. I mean, look at HubSpot, right? HubSpot is my thing there. It's a $30 billion company. Brian and Dharmesh get to still be involved. From an ego perspective, that's nice. It's better than Drift, right?
That's what I wanted to build, but I made mistakes along the way that didn't create a foundation solid enough to grow that company.
You're the first investor in Klaviyo?
You know, I'm a first investor, good friends with Andrew.
Right. You're the first investor in Klaviyo? You convinced him to take VC money?
I convinced him to take VC money.
Whoa. And so how much did you put in?
It went—it's great.
I'm sure I'll have him on the show.
Of course. The thing about it is that you go and you do that, and look, he has an amazing $10 billion company, right?
Right.
It's a fantastic free-cash-flow business. I still think it's super underappreciated at $10 billion, so that's why I'm not selling my stock.
Those are role models for me, but this time I want to do it my way. That's what I'm doing. That's why I'm in this journey all over again with Agency.
You're an investor also, correct? You've made quite a few investments. How has investing changed your operating mindset?
Investing is really hard, man. That's why I could have come and become a VC. I had a dinner with David [likely Fialkow] for about 2 hours one day because we were going to start a firm, and he's like, "Oh, my God." So many bottles of wine later, he's like, "I can't believe you gave me 2 hours to explain how a VC firm worked."
After that conversation, I was like, "No fucking way."
Why? What made you say no fucking way?
Because I'm a builder.
You can build a firm.
No, but the thing is that I can't be telling people, "I'm going to build it this way." I hate telling people that. I'm like, "No, that's wrong. That's not going to work." When I see how much of a gap there is between where they're going to be after that experience and that journey, I'm not interested in working through that journey with them, you know what I mean?
Asking people for money, making bets, and then having to advertise and build a different style of brand for myself—I'd rather build, man. It's so much easier, so much more fun.
Why did you take VC money? Respectfully, you could self-fund. Why take VC money for Agency?
The reality is that I made about $2 million in just over 6 or 7 months consulting on AI. I was working with OpenAI. I was servicing their customers. I was sweating it out there, learning LLMs, because I thought I had missed the boat.
I worked with Red Bull, Ticketmaster, likely the NBA and likely Klaviyo, and that's kind of where the idea for Agency came from.
You made $2 million doing the consulting?
Yeah.
What were you doing?
I was just working, coding, and solving problems for them—for OpenAI's customers. People would go to OpenAI and say, "I need help doing this," and OpenAI would say, "We don't help you build that stuff. That's your problem."
They would send me leads, and I had to work the leads. I started charging about $25,000 a month, and then slowly things became $100,000 contracts. Some became $500,000 contracts. So there's money to be made, right?
I actually tweeted that AI services companies will be bigger than actual AI technology providers in the next 5 years.
Yeah, I think Paul Graham said I was a [__]. But do you agree with that?
I just think that humans are the biggest blocker to adoption, and they're going to need a lot of hand-holding and help. There's a period where, because we don't have the agents that we need in the world, there are workflows, then new products, and then new models, right? In the meantime, it's a step function where we need people to hand-hold them.
It leads on, and then we'll do a quickfire. It's a horrible question because I hate asking about timelines, but do you think there will be this transitionary period of 5 to 10 years, or will it be much shorter—12 to 24 months?
I'm on the longer side.
You're on the longer side.
I think there's a lot of hype. Even in the Industrial Age and revolutions, GDP was growing at 1% or 2%. I don't think we're going to grow at 10% or 20% in 1 year all of a sudden, because most of the industries are even slower than tech. Tech is being extremely slow to adapt this technology, so it can't be that fast.
It's funny. Someone said we misattribute success in AI to AGI, and actually the way that we should measure it is as a 10% increase in GDP.
Sure. I'll bet.
How much? I'm not a betting man.
Let's drop some money on this if you want.
I didn't want to drop some money on this, but he didn't set a timeline. That could have been in 10 years.
Yeah, but 10% GDP growth? The world's going to double and explode so much faster.
Whatever. When that happens, it's going to happen.
In the meantime, it's going to take time for people to adopt this, and we don't even know—we don't even have the right solutions yet. What we've found is that we just discovered fire. The rest is going to take time.
I want to do a quickfire, so I'll say a short statement.
What do you believe that most people around you disbelieve?
Being direct is better.
People disagree with that?
Yeah.
You can buy and hold 1 public stock for the next 10 years.
I'm not a good investor. I don't know.
Come on.
I'm scared of the giants. They're not going to last. But NVIDIA—I think compute, we're thirsty for compute.
You could have said Klaviyo.
That was a mistake. It's tough.
What would you do if you knew you could not fail?
Educate more people in the world.
I love that one. I was not expecting that.
How do you have such great skin? Julian Back asked this one. He likes asking about facial routines.
Asking about his facial routine? Am I blushing?
You do look very young.
I think it's the Latino genes. I thank the genes; I'm grateful for them.
Your Botox doctor told me.
I have no Botox—no, no, no. Nothing here. I don't even wash my face.
What have you changed your mind on in the last 12 months?
I didn't think I was going to start a VC-backed company. I didn't know what the right idea was. I didn't know that. So I changed my mind: this is the best thing I could ever be doing.
What about the way that your mother brought you up have you decided to do differently with your children?
I'm letting them do whatever they want. I'm freeing them from my pressure to be me. They're kind of being me in a way.
I'm letting them be. If they don't want to go to school or they don't want to do their homework, I don't care. The problem is, I'm happy, I'm successful, and I'm rich. It's up to them.
Does money make you happy?
No, but it definitely makes things easier.
What concerns you most in the world today?
Extremism. Ignorance. I just think that people can't have a conversation. It's sad.
What's your biggest short in the public markets?
Salesforce.
I've asked a couple of people that question, and everyone said that. I'm just looking at—oh, fuck, Marc Benioff's going to fucking kill me. I guess I haven't listened. Anyway, Marc, do you want to come back on the show?
When you look at the next 10 years, what are you most excited for? I'd like to end on a note of optimism.
My mother's got MS. I'm excited for the drug discoveries that will come through AI and through what we have ahead of us.
Look, I just came back from Cairo, and I got to learn about Egyptian civilization. It's mind-boggling, and I don't think we studied it enough. The love for art, the love for culture, and the love for craftsmanship was so mind-boggling and so inspiring for me.
We think that with evolution and new generations, we're better, but I think we're going worse. You can see in different dynasties and different generations how the craftsmanship in the Hellenistic era was actually worse than prior to that. You can see that in the way that we build and the way we treat our customers, even with billion-dollar companies.
I'm hoping—I'm optimistic—that we can get back to that and build better products, better customer relationships, and better businesses; more art, more culture, and more time with our families. I'm optimistic that that could happen.
Elias, listen, my favorite thing about this is being able to have great conversations that are completely unstructured and learn from amazing people. This has been fantastic, so thank you so much for joining me today.
It's been my pleasure.