把一切都交出去的独角兽创始人

Bryan Johnson wants to break the chains of biology. I want to break the chains of time. We can always raise another round or do another trade, but you can't raise another decade.

You're running a massive company. You also do investing. You're also in a happy marriage with 4 kids. I know the way that you do it all is delegation. What is the secret that you've figured out?

The cardinal sin of delegation is that it will be faster or better to do it myself. The reason it's a blocker is because it's true.

What have you learned in your extensive founder career? Jonathan, welcome to the podcast.

Thanks for having me.

So, Jonathan, you're running a massive company. What, 4,000 employees?

4,000.

4,000 employees. You also do investing. You're also in a happy marriage with 4 kids. How do you do it all? I know the way that you do it all is delegation. You have a chief of staff and a bunch of EAs. Why don't you talk about the secret that you've figured out that other people can learn from?

My origin story here is that, out of school, I worked at the White House, and I sat in the West Wing next to the president's executive assistants. As you might imagine, the EAs in the West Wing were really freaking good. It set my bar super high for what this EA-client partnership could look like.

When I moved to San Francisco to start Thumbtack, I said, “Okay, I'm not president, but what if I had an EA and support team that was as good as the president's? What could I accomplish?” At Thumbtack, as we were scaling the business, I hired my first assistant in the Philippines, who started helping me with basic stuff—my inbox and calendar—and then we got creative and did more and more complex, interesting things, which we can talk about.

What I found was that the more leverage I got, the more ambition I got, and it just compounded. You start with 1 assistant; now I've got a half dozen, and every assistant gives me more leverage to take on more things.

It's fascinating, because some people say, “Bill Gates is so much richer than me, but he has the same phone that I have.” I remember—maybe it was you, maybe someone else—someone saying, “Wait, Obama should have way better information than me, and yet he's reading the same media.”

This is another example of it, which is the democratization of access. The White House has amazing assistants, but thanks to Athena and technology, at some point everyone will have a certain block.

Exactly. A couple of decades ago, you had to be Marc Andreessen or Vinod Khosla to have a half dozen assistants, and that cost you half a million dollars. Now, with a company like Athena, for $3,000 a month you can have your own assistant. With AI, it's going to become ubiquitous.

There are going to be billions of people learning to delegate to a machine assistant. As your budget increases, you might add a human or a chief of staff on top, but it's a ladder of leverage.

How should we think about what humans can do versus what AI can do? My friend started Clara Labs a decade ago, which was trying to be an AI assistant, effectively. Obviously, this was well before a lot of the innovation, but what is the right way of thinking about the human-AI relationship?

It's like self-driving cars: it's not self-driving overnight. You drive the Tesla at first, and then there's assisted steering and braking, and over time it becomes more autonomous. The same thing is going to happen with assistants.

My view for founders is that if you don't have an assistant, you are the assistant, and you don't want to be the assistant. No matter your budget, you should first start by learning to delegate. If you've only got $20 a month, start by delegating to ChatGPT. Prompt engineering is really just delegating.

If you become world-class at leveraging ChatGPT to brainstorm your goals and figure out how you can take the next step in the business, once you have the budget for a human assistant or chief of staff, you're prepared to go to the next level. You start at $20 a month. If you have the budget, you hire someone directly with a service like Athena. If you have enough money, you hire someone in person—that's 6 figures. If you have a legendary budget, you've got a suite of assistants and a chief of staff following you around.

I've had an assistant for many years now, and she's absolutely amazing. I use her for scheduling and typical assistant activities, but you have many assistants. Our friend Sam Corcos has a bunch of assistants as well. Close the gap between what the people who do it well do and what else people get leverage on or delegate that maybe isn't obvious.

First, you start by taking the pain off your life. I never wait on hold. I never put my credit card on the internet. I would never fill out DMV forms or do passport renewal. The world is full of all these annoyances.

First, you take off the things that drain you, that don't give you leverage. Then, once you've done that, you raise your sights and ask, “What's a new business I could start? How can I scale my business faster? How can I spend more time with my family? What are the things I actually care to do a lot more of?”

I start with basic stuff—inbox, calendar, stuff that everyone does—but then I start experimenting. During the scale-up days of Thumbtack, when we were in a house working all the time, I told Marnie, my assistant, “I don't really have any friends outside of work. The only people I know are in this building. I need to make some friends.”

I told her, “Let's plan a founder dinner at my house every other week.” We did lots of this together back in the day, and that's how I made all my best friends. I walked home from work, and Marnie had invited people, set up the chef and the bartender, and we walked in and made new friends.

I met Catherine, my wife, through that. Then Marnie helped plan our wedding, and now Marnie helps us with our kids. It starts small, and then you just compound these things.

But even in that example, say more about how she knew who to reach out to or what kind of people to invite.

The entry-level delegation is that you delegate by task. You say, “Help me plan this dinner party.” If you just say that, you're not going to get what you want.

The more advanced way to delegate is called “delegate by algorithm,” where you actually export your own internal preferences as you delegate. I would say, “When I plan dinner parties, I like to have 6 to 8 people. I like people to have raised similar amounts of capital, or be at similar stages, or have a similar number of employees.”

You literally write an algorithm for how you find the right people, and then you get feedback on whether it worked or not. Once the algorithm is fully exported from your head, it's just rinse and repeat. Engineers tend to be better at creating these kinds of SOPs, or standard operating procedures, but it's something you can learn to do.

Share more about how, for people who have multiple systems, you differentiate. How do you delegate across the team?

Start with 1, and then, like any team, you specialize. I just had 1 assistant who did everything as a generalist, and now I have a half dozen, which I know sounds crazy, but each one specializes in different things. One is on work, one is on home, one is on kids and family, one is on travel, and one is on finances.

That specialization is helpful because the finance person is just thinking about the balance sheet, sending wires, and all of that. Then I have a chief of staff who sits on top and distributes work across all of them.

What are other principles? I love “delegate by algorithm,” not by task. What other principles or frameworks are really important or non-obvious to get the most out of an assistant?

The cardinal sin of delegation is that it will be faster or better to do it myself. That's the number 1 blocker for most people, and the reason it's a blocker is because it's true. It will actually be faster or better if you do it yourself that first time.

It takes more effort to delegate, to teach someone how to do it. It might not be as fast or as good as you'd like it, but the only way you get leverage is by going through that work. You have to overcome that activation energy.

The second mistake people make is that they don't compound for the long term. They hire someone and then switch assistants every 6 or 12 months. But it's the compounding that creates incredible leverage. I've been working with Marnie for a decade—she's like a little sister now—and she knows everything about me.

Those are just some common principles. In terms of the way to delegate, I think this is just tactical—the methods. The most common way people delegate is with their thumbs on their phone.

That’s the worst way. Very slow. It’s better to use all 10 fingers at a computer, but that’s actually pretty slow as well. The best way to really delegate is by using your voice. With voice, you can talk 2 to 3 times faster. You can do it on the go: on a date, between lunch, or in the Uber.

At Thumbtack during hyperscale times, I would walk between meetings and voice-note an assistant: “Here are the takeaways. Pre-draft these 5 emails. Follow up with this person.” All my work from that meeting was actually delegated by the time I started the next meeting, versus everything accumulating until you get to the end of the day and you’re like, “I have 100 things to do.” Learning to delegate by voice is important. If you look at all the super-delegators at Athena, they’re all just delegating by voice all day long.

Right? Yeah, it’s fascinating. At some point, it’s going to be that I don’t even have to say it; I just think it.

Yeah. This is part of our vision at Athena longer term: you shouldn’t have to actually vocalize it.

No.

We’ve built an internal demo of this. It’s not live for customers, but the way it works is that we built something that watches your screen as you work and takes screenshots. When it identifies things that you should be delegating to your assistant, it automatically adds them to your assistant’s task list. Your assistant then says, “Oh, Erik would want my help with this, or maybe he’ll do this on his own.” That creates the reinforcement learning, effectively, to tune the model to pull things off your screen. You don’t have to vocalize it.

The person who built this internally has been using it, and the majority of his delegations are now machine-generated.

Yeah.

He’s not talking. He’s just working, and the machine is magically pulling tasks and putting them on his assistant’s plate. It’ll take a while to bring that to market and fruition, but that’s the future.

There should also be a version where you see in Slack that it’s someone’s birthday, or someone has a baby, or something, and it suggests, “Hey, do you want to get this person this specific gift?”

I think this is a place where we’re thinking a lot about the human-machine merger and how we get the best of both. That’s really the vision at Athena: the best human assistant and the best AI wrapped into one combined product that’s seamless. Humans obviously have human touch and good UX, but what machines can do is remember everything and be super proactive. They can look through every email and every calendar and remind you to do things that you would not have thought to do otherwise.

One thing you’ve always been fascinated by is international labor. Some call it labor arbitrage. What’s the right way of thinking about what people in, say, the Philippines can do or can’t do versus assistants in the US? How do people think about it? Obviously, there’s a huge cost savings. What are the things that you keep in that cost savings, and what are the things that you don’t? What’s the right way of thinking about it?

America’s strengths are entrepreneurship, capital, innovation, and technology. The more you can leverage someone to do back-office tasks, the more you can spend on technology, product, and so on. I think the thing Athena focuses on is an executive assistant—someone who can do all of this admin work. But people are using global talent for all sorts of things: legal, medical, and everything else.

Yeah. So let’s get a little bit into the finance side. I’m curious because I don’t leverage mine there. What are the ways that people could get leverage from this? Are you still interfacing with the accountants, or are they? What’s the right way of thinking about it?

The first thing I’d recommend, if you’re getting an assistant, is to have them help you save money. If you have a big budget, it’s not a big deal, but the first project can be: “Hey, help pay for your salary by looking through all my subscriptions, finding things to cancel, saving me money, and finding me refunds.” That’s a cost-saving bucket.

Then there’s admin—paying bills. Then there’s coordination, which is what you’re talking about: basically being a project manager. So, coordinating when the tax attorney asks for something from the accountant, and there are a bunch of documents, and it’s just moving paperwork between systems. At enough scale and complexity, that becomes effectively a full-time job.

Obviously, you need someone you truly trust to go deep with. We recommend that you build this trust over time: give them email and calendar access at first, and a bank account later. But as you earn that trust, the more access you give, effectively, the more leverage you get.

Yeah. It’s really interesting. In terms of scaling this company, you’ve dealt with this at Thumbtack as well, in terms of workforces in different areas. How do you think about culture between them? What are the challenges, and what’s really important to get right?

Yeah, I mean, we look for cultures that have a strong affinity for American culture. The Philippines has a long history with America. They know American pop culture and sports better than I do. They have a very strong work ethic, and it’s a caretaking culture. There are lots of nurses, and one of the reasons we picked it for assistants is that being an executive assistant is really taking care of someone. Their full-time job is to make your life better, take things off your plate, and help you be successful.

I think one of the barriers for some people to get an assistant is that it feels indulgent. It’s like, “Should I really have someone who’s just taking care of me?” But the way I flip that around is that not delegating and not giving yourself leverage is holding you back. It’s also preventing you from creating a job for someone who’s excited to help you.

Yeah.

For the people on our team, they get to work with a CEO and see behind the scenes of an exciting startup that’s a totally different life from the one they live. For them, that’s a cool, exciting, life-changing thing.

Yeah. One thing I think people should be using assistants more for is not only delegation but also accountability. For example, you want to be exercising a certain amount, spending your time doing X, Y, and Z, or making sure you go to your appointments or something.

Oh, we have some clients who actually exercise live with their assistants for this reason. They both have a fitness goal, so they pull up videos at a set time and work out together. That’s a place where human accountability is a little more powerful than machine accountability.

ChatGPT actually has a feature now where it will automatically ping you about things. If you can’t afford an assistant, you can just tell ChatGPT, “Here are my top 5 goals. Ask me how I’m doing every day, log my results, and give me feedback.” You can get that for $20 a month.

I’ve done similar versions of this with a human assistant. I think we may have done this in the past, where I’m like, “Hey, I want to work out, meditate, and eat clean. Send me a ping on WhatsApp every day, and I’ll just reply yes or no, yes or no, whether I did it.” At the end of each week, send me a scorecard and compare me to Erik or Katherine, and we can have a little competition with it.

Helping you hit your goals is exactly what assistants should be doing. The other thing I think is an interesting opportunity to get more leverage is that these people are not just operations-savvy but also pretty intuitive. They go into caring fields, so there are opportunities for coaching, too. They’re the people who have the most context on you, and they can give feedback on how you’re engaging. They can tell you, “Hey, you’re tired today.”

That was my first observation when I was in the West Wing. The assistants who sat next to the president were incredible. They were so good. They went on to be elected to Congress and became amazing people.

At the end of the day, the president and his advisers would have been meeting with senators and all these people. They’d finish the meeting, lean back, and have the assistant come over, and they’d just be like, “What’s going on?” It was more than just accomplishing things. This was their closest confidant who saw behind the scenes of everything. They knew the good and the bad. They knew the president had gotten kicked in the nuts a couple of times that day by different things.

They’re there for them. I think that human element is going to remain important for a long time, no matter how good AI gets. It’s something you should lean into. We have clients who tell us, “I’m going through a divorce,” or, “I’m depressed,” and their assistant is there for them in a way that others aren’t. They can’t tell their company or their team about some of these things, but this is the person who’s on the inside with them.

Yeah, totally. Relatedly, one of the things you’ve had really interesting insights on over the years is how to spend time—thinking about time management not only in terms of ways to get more of it through delegating, but also how to prioritize. I’ve learned this from you, but you think about goals on an annual, quarterly, and other time horizons. Why don’t we start with LBD and how you think about goal-setting in general, and then go a bit deeper?

Yeah, I mean, my wife, Katherine, and I met as friends. We developed this friendship over one of these entrepreneurship dinners.

We read a book called How Will You Measure Your Life? by Clayton Christensen. He basically told the story of how Harvard Business School professors’ kids come back at reunions and they’re Fortune 500 CEOs, super rich, and their lives are in shambles. They’re divorced, have all these problems, their kids don’t talk to them, and they’re going to jail. The rigor they apply to their businesses—the quarterly reviews and the metrics—was completely absent from their lives.

His point was, how do you want to measure your life? Katherine and I read this book when we were just friends, but we were both like, “I want to be on your life board of directors. Let’s freaking do that.” So we started a tradition a decade ago. Every quarter, we get together and fill out a survey on our relationship—strengths and weaknesses, kind of a SWOT analysis, geeky stuff—and then we talk about what we can improve and what we could do better. Anything you invest in, the more you invest, the better it gets.

Totally. I think that’s a key point because some people are turned off by quantification. They want their relationship to be relaxed or turned off or something. I think it’s less that you have to quantify every element of it and more that you don’t want the lack of quantification to mean that you’re not prioritizing it. This is just an exercise to make sure that you’re holding yourselves accountable to how you want to be.

Exactly. Be purposeful and get the thing you want. If you’re a geek like us, then measure it and track everything. If you just want to go have a date, drink wine, and talk about your relationship, that works too. I think it’s just that people prioritize fitness or their work, but they never think about, “How do I prioritize my relationship? What’s the fitness or workout or yoga equivalent for my relationship?”

Having been in many goal-setting sessions with you, one of the things I think you intuitively understand or have learned, and are often instilling in others, is this idea of prioritization. When people set goals, they set a lot of goals, and they don’t think about, “What is the one that makes everything else much easier?” What is the one thing that, if we did it, would make it a great year, a great quarter, or whatever it is?

Who sets too many goals?

You know, me a little bit. But, yeah.

I think my view is that there’s a power law in everything: a power law in business, a power law in venture, and a power law in your life. There’s typically 1 thing every month or quarter that, if you accomplish it, is worth more than everything else combined. Figuring out what that 1 thing is is what you should do. Figure that thing out, and then go all in on it.

It’s not to say that having other goals isn’t worthwhile, but if you haven’t identified the 1 thing and gone all in to make that happen, then I think it’s not optimized. I think it’s also easier to procrastinate on the most important thing in your life by having a long list.

And you do a bunch of different things, but there’s that 1 thing. Find your partner, start the business, or fix your health. That really is more foundational than everything else.

So, to that end, I’m curious if you could share more frameworks or principles you live by in terms of time. For example, I believe you try to stack meetings into 1 day or as few days as possible. What are certain frameworks that are important for you in terms of time and energy?

At a high level, Bryan Johnson wants to break the chains of biology or longevity. I want to break the chains of time. The question I ask myself is, “What’s the most valuable asset in the world?”

It’s not gold, Bitcoin, or NVIDIA clusters. It’s time. We can always raise another round or do another trade, but you can’t raise another decade. If time is the primary asset, and it’s more foundational than everything else, we should focus on owning and controlling it.

This means different things for different people. If you’re in hardcore scale mode, that may mean half-hour meetings for 14 hours a day, where you’re grinding out interviews and doing things that only you can do as a founder. Or it may mean clearing your schedule and having time to dream about the future of the product or big deals. You just have to know what stage of the business you’re in and then design your calendar purposefully around that. If you look back on the last month and your calendar doesn’t reflect your highest goals, then you’re not doing it right.

I remember Keith Rabois sharing this idea with me: At the end of every week, do a calendar audit of the past week. What are the meetings you had that you wish you hadn’t had? Then just look at the next week.

There’s a place where Athena assistants can do this proactively for you, but this is something AI is obviously going to automate, and it will be very powerful. At the end of each week, it just tells you, “Here’s what you prioritized. Are those actually your goals?”

Yeah.

If not, then next week we should adjust those things.

Yeah. I’m curious about your interpretation of meetings. Some people are like, “I want to have as few meetings as possible. Meetings aren’t the most efficient; I’d rather do everything asynchronously.” What’s your thinking between asynchronous and synchronous communication?

Naval has a good line: A phone call is better than a meeting. A voice note is better than a call. A text is better than a voice note. I think, in general, that’s true.

There’s almost a generational thing where I feel like you’re in this a little bit: There’s a reluctance to call people on the phone. We do everything by text and email, but calling people is super effective. It’s really high bandwidth, and you can figure things out in 5 minutes. I personally try to minimize meetings and instead pick up the phone and talk to someone for 5 minutes.

When I’m in more of a chairman mode, I prioritize as few meetings as possible and use voice notes. What I request is voice notes out and text messages back, because listening to voice notes is slow.

It’s assault as marketing. Yeah, it is interesting. One of my favorite self-help books is The 7 Habits of Highly Effective People, and one of its principles is this idea of efficiency when it comes to achieving tasks or accomplishing tasks, but effectiveness when it comes to people.

The efficient path might not be the effective path because you might save some time on the front end, but you might accrue a debt on the back end. It might be efficient to just send a text or something, but you’re creating a dynamic. Phone calls are obviously underrated.

Yeah, that’s true for remote cultures, too. Remote cultures build up debt from a lack of bandwidth. If you’re in a remote culture, you should get together every couple of months for a few days in person.

How do you think about when founders are asking for advice, like, “How should I spend my time? What’s most important?” What frameworks are helpful for founders? Obviously, there are a million things they could be doing, and there’s not enough time in the day for even a fraction of them. What are the most important things to get right, whether it’s your time at—

Thumbtack, look, it’s: Build your product, talk to customers, raise capital, and build your team. When founders ask me how to leverage an assistant, I turn it around: What are your top 2 goals for the month?

If it’s recruiting, you have to be closing candidates, but you don’t need to be sourcing all the emails. You could send a voice note and say, “Here’s a template I want you to use to reach out. Let me go through my contacts—here are 20 people. Send an outreach email asking them for suggestions for this role.”

If your goal is to raise capital, your assistant could help project-manage the process. Step 1 is to set your goals, and then you break them down into subtasks. Your assistant won’t be able to run the fundraise, but they can run components of it. Your job is to decompose things into the smallest parts that you can hand off.

Maybe you should have some sort of forward-deployed role. It’s not just the assistant; it’s also someone who helps you really get the most leverage out of it and watches your activity.

One of our learnings from this business, when we first started it, was that we thought we were just going to hire the best assistants, match them with clients, and boom—amazing. But we matched the first 5 assistants, and every single client said the same thing: “How do I delegate my calendar? How do I delegate my inbox?”

So it became clear that we needed to invest more in teaching the clients how to delegate, because that’s as much of a constraint as the assistants. We’ve got a few thousand assistants and clients. If you look at the best-performing assistants at Athena, they work for the best delegators. It’s not a coincidence.

The reason they’re the best is that someone has figured out how to export their ideas, decompose projects, and create SOPs in a way that helps the assistant fly. The more mediocre delegators struggle with that and need more training.

One thing that you've been phenomenal at, whether it's Thumbtack or hiring executives, is hiring executives. What do you think you do differently? What are the principles or frameworks that guide your executive hiring process?

One thing that's obvious as you hire executives is that the more senior they are, the better they are at interviewing, almost by definition. When you get to the C-suite, they're all really good at interviewing, so you actually have to discount the interview more as they become more senior. I use references much more the higher they go.

The other thing that I think people should do more of is ask people for their 360 reviews. At their last company, 10 people wrote reviews of them that are honest about their strengths and weaknesses. I just ask to see them, and I say, “I'll share mine; you share yours.” I actually think this would be a cool kind of reference-process startup where you could ping the LLM about someone versus pinging their references. I try to get to more ground truth on the person versus—

The UX of an interview is just going to be pretty good when you get to the senior level. What's most important when you're doing reference checks? What's the way to get the most signal out of them?

What I often tell people is, “Hey, I'm probably going to hire this person. Now tell me all the things you don't like about them, or things that they could be better at.” People are so biased toward telling you all the good things. You have to get them to see it as a safe place to share strengths and weaknesses. Do enough references until they start to sound the same. Once they start to sound the same, then you've got a signal.

One question I like to ask is, “What skill sets would be really important to complement this person with?” That's another way of saying, “What are they bad at?” One of my favorite questions is, “If this didn't work out in 6 months or a year, why would that be?”

The best way to source talent—I just hired an executive recently—is to ask the people you respect the most, who have the highest standards, “Who's the best person?” Here are the 3 people who have the highest standards in this role. I ask them for the 2 or 3 best people, and I go try to hire one of them.

That is the rapid way to hire an executive. Using an executive recruiter, you spend 3 months and get all these cold leads, but going through a network of high-bar people is super efficient.

There's some maxim that almost half of executives don't work out within a 12- to 24-month period. That checks out. Why is that?

I think it's difficult to judge. The interview process is not representative of the work they're doing, and that's ultimately what's broken about interviewing. Interviewing should be someone doing the work they're going to do.

I think Weebly, back in the day, would actually have executives come work for 2 weeks on a trial before they ever hired them. That's difficult to pull off because people have other jobs, but I do think something like that makes a lot of sense. In a lot of cases, your interview process is just kind of irrelevant compared to the work they're doing.

Keith has this maxim that you're not going to get 100% hiring, in the same way you're not going to get 100% investing. You also want to avoid the trap of not taking enough risk or moving too slowly.

As long as you cut your losses quickly, it's okay. If you don't cut losses, then you compound a hole, and that's where it's painful. The other thing I'd say is more useful for basically all interview processes is project-based work.

You know, hiring a C-suite executive for one part of the team, I just tell them what the pain points in the organization are and what I want to fix. I'm like, “Come up with a plan. How would you fix this? If we hire you, you're going to do it.” That sort of practical work is more useful.

That's really interesting. Going back to other founder principles, what is your stance on transparency with the company? What information is helpful for the company and employees to know? Going back to the Thumbtack days, what principles guide how much you share?

Default transparency is obviously the best, but there are obviously some lines. At Thumbtack, there was a moment—you may remember this, a decade ago—when I woke up and had a message from someone on our team in the Philippines. It said, “Thumbtack does not exist on the internet.”

I was like, “What?” I went to Google, typed in “Thumbtack,” and nothing showed up. We basically received the death penalty from Google. Our traffic went to 0, and our revenue went to 0. There was a misunderstanding. They thought we'd done something wrong, but they eliminated us from the internet.

I remember waking up and going to the office, and there was a class of 25 new Thumbtackers whose first day it was. I was supposed to do the onboarding with them. There was also a journalist walking around outside trying to get a quote. That's one of those times when you can't share everything. If I just walked into a new class and said, “Hey, the business might be dead,” that's not going to work.

You have to get a handle on the situation. The people who need to know know, and then once there's a plan and a solution in place, the whole company is told.

That makes sense. At Thumbtack, you were a non-CEO co-founder, and you lasted a decade—or almost a decade—and you're still on the board, of course. I feel like it's pretty rare when I think about co-founding teams that the non-CEO, especially if they're not a CTO, lasts over 5 years or something. I'm curious what's really important to get right in co-founder relationships, or how you think about advising companies you invest in.

I heard Sir Michael at Sequoia say something that was, “There's only one founder.”

Yeah.

He said it at a founder event, and I was like, “That's rude. There are lots of co-founders here.” Now, a decade later, I'm like, of course there is only one co-founder who goes the distance, because ultimately, as you scale the organization, it only makes sense to have 1 person on top.

I think that's just the reality, and it's good to know that going into starting a business. In terms of who to pick, my advice is to pick someone you want to get married to.

You're going to have all the same relationship issues you have with a partner. There's no makeup sex, but there are still highs and lows, and you want someone that you ultimately deeply trust. There's going to be stress and challenges, but as long as you trust the foundation, then you've got a good thing to build on.

I would say one of the biggest—if I call it a mistake—at Thumbtack, because I love my co-founders, is something I wouldn't recommend others do: We had 4 co-founders, and 1 was technical. He left after about 2 years, and then the other 3 were non-technical. This is the most idiotic way to start a technology business, but we had no idea what we were doing, and it worked out.

How did you divide and conquer between those 3 non-technical co-founders?

We just jumped on things—whoever was good at something—but there was no clear executive.

Yeah. I think that's not recommended, because it's obvious how that could create tension. You worked through it, but that's why it's nonstandard and not best practice.

Right. Right. But going back to the original story, I remember you got into GSB and were going to go to grad school, but you had your buddies from college and were thinking of starting this business, which wasn't even a personal pain point. You just thought it would be a good business.

You emailed the GSB people to let them know you weren't going to accept, and they sent you back this very demoralizing email about how you were making the biggest mistake of your life. It turned out pretty well. Pretty good decision. How did you, at the time, think about getting enough conviction to take the leap and do it?

It was scary. I got into this good school, and it seemed like a safe bet. I knew I wanted to start a business, and if I went to business school, I knew I'd have all this debt. I probably would need to get a nice-paying job afterward. If I was going to go for it, I should probably just go for it.

When the dean sent me that email, I was like, “This is the biggest mistake of your life.” He said something like, “Startups will come and go, but you can only come to Stanford once,” or something. “This is a once-in-a-lifetime opportunity.”

I never responded, and I have it saved as an email that I'm just going to respond to whenever Thumbtack goes public and be like, “I made the right decision.” I have nothing against this guy; he's just doing his job. But it's good to have a little chip on your shoulder.

Yeah, totally. Okay, so there are EAs, and then there are chiefs of staff. How should we think about the difference there? What’s really important to get right in a chief of staff? It’s a role that everyone, in addition to an EA, wishes they had, but it seems hard to find the right person.

There’s no formal definition. An EA is typically more administrative, taking things off your plate, and a chief of staff is more offensive. It’s the type of person who can follow you into meetings. You could deploy them: “Hey, actually take this meeting for me,” or, “Hey, this problem came up today. Go figure out how to solve it.” A chief of staff typically has the capabilities to become a founder themselves, and the only reason they take the job as chief of staff is so they can see what it’s like to be a founder, so they can go off and do it. You get people who do more tours of duty, and you get super-high capability, but you don’t get the same longevity as an assistant.

What have you looked for in a chief of staff? What’s the right balance? There’s this tension between getting someone younger who’s initially going to do it for a couple of years and is more ambitious, versus someone who’s more seasoned but maybe not as high-slope. How do you think about it?

The way I break it down is: for an assistant, you find someone you want to go a decade with. The intimacy and the compounding of that relationship are super important, so you go deep. That means you probably don’t find a founder, because they’re not going to be excited to do this. It’s someone who’s much more of a caretaker and is excited to be number 2.

For a chief of staff, I do pure slope, where you know the person is going to move on in a couple of years, but they could start a business. Instead, you get them by your side to help solve problems. There are people like Lonsdale who have a chief of staff for a year or 2, then they start a company, you invest in them, and they build this reputation for doing. Then they have these alumni, almost, chief of staff.

Yeah, Bezos has someone who follows him around and goes to every meeting. It’s cool. Let’s talk about some people who get an enormous amount done. I’m curious what lessons we can draw from them—whether it’s someone like Lonsdale, Elon, or Sam Altman. Who are folks that you’ve particularly learned something from, even just watching from afar, about how they build their world and operate?

I’ll just say one for someone like Lonsdale: he’s done this really well. They compound these certain talent networks. Whether it’s Stanford Review or just Stanford in general, they’re often hiring people from there and getting involved in those talent pipelines. They’ve created a proprietary, compounding talent base. They’re deep in the hackathon scene, but they just build this talent base.

I mean, talent access is part of it. I think the other thing is that there are some people who have limitless ambition, and you’re very much this way: “What if I could do 10 times more things?”

I think lots of people just think that’s impossible.

But it’s not. It is possible, and with delegation and empowerment, you can keep compounding bigger and bigger scale. The other thing that’s interesting, from working with these power delegators, is that people assume people of great power, wealth, or resources can afford this team, and that’s why they have this ambition. But we’ve actually seen the opposite is the case: as people get more leverage with assistants, their ambition increases.

It’s difficult to have big ambition if you’re weighed down by DMV forms and passports. But once that’s off your plate, you raise your sights and can be more ambitious. Then you go from assistant to chief of staff and keep leveling up until you’re Lonsdale or the president.

I think another thing I’ve admired about some of these people is that they remarkably hold very few grudges. Apparently, Peter Thiel kicked Elon out of PayPal, or a lot of these people have had beefs decades ago, and yet they’re quick to reconcile and just do business again.

Yeah. I mean, look at what Vice President Vance said about President Trump before they were partners.

Trump’s actually quite forgiving. A lot of people say a lot of bad things about him, and he’ll continue to work with them as long as they change their mind.

Yeah. I think there’s just a desire to win at all costs, and grudges are not helpful.

Yeah. Even Trump and Elon, right? They got into it and reconciled. TBD where it stands by the time this comes out. There are also some people who seem to have a pretty big operation. Bill Gates has a pretty big team, pretty big operation, but Bezos and Elon maybe have pretty lean operations. I’m curious how you think about that.

Elon’s the one that makes no sense. He claims not to have an assistant and to be scheduling things himself. There’s a story of Bill Gates trying to schedule a meeting with Elon, saying, “Connect me with your assistant,” and Elon says, “No, I am my assistant.” I hardly believe it, but Elon’s just insane, so maybe he is that way.

I think most normal humans, if you want to get leverage, have to delegate. One of the cool personal teams I saw was this guy who’s a billionaire and runs a hedge fund. He has a team of 40, including 8 personal executive assistants. All of the executive assistants graduated from Princeton, so it’s an insanely high bar.

I was asking the woman who managed the EA team, “How do you decide what to prioritize your team on?” She said, “If my team can spend 1 day saving my principal 1 minute, it’s well worth our time.” I thought that was so next-level and definitely something I would love to aspire to, but it’s the billionaire budget.

Yeah, it’s pretty funny.

Also, just in terms of prioritization, we’re hiring assistants here, and someone was saying, “Do you want them to be in all meetings with you?” I was like, “Well, if they’re in meetings with me, then they’re not working.” I want them to be finding ways to give me leverage and take things off my plate.

I mean, one of our visions at Athena is that you can have digital assistants that listen to you and work with you, connected to your Zoom, inbox, and calendar, and mine all of that digital exhaust for things your assistant should work on. Your assistant can’t possibly sit in every meeting if they’re going to do anything, but if the digital machine assistants are surveying it and pulling things out, that’s a pretty cool combination.

Yeah. I think it’s interesting to look at how people build out their universe. For example, Marc and Ben funnel everything through a16z and are all in on a16z. Of course, they have other intellectual or philanthropic interests, but mostly it’s about consolidation.

Whereas someone like Peter Thiel fragments it a bit more. There’s Founders Fund, but then there are several other funds, companies, and all sorts of initiatives that are somewhat collaborative but independent. He sits on top, but he also seems to let them operate; he’s not the CEO of them. He has this talent network that he’s worked with for a long time that goes in almost boomerang-like fashion across the universe.

Lonsdale feels like he’s more hands-on. He’s got the HBCU universe, but then, with his other initiatives, he’s quick to institutionalize something, whether it’s a think tank or something else. It’s no coincidence that Peter Thiel would do the Thiel Fellowship and Lonsdale would create a university or something. Thiel is almost early to things: he creates a format or a template for something. He does the hard work of field-building intellectually, whether it’s doing something in a controversial space like longevity, charter cities, or higher education. Then people like Lonsdale institutionalize it to some degree.

Elon has also been described as having this SWAT team of people who help him from company to company. He’s able to recruit world-class people to lead these efforts.

Yeah, I think it comes down to what’s the right org structure. There’s actually no right org structure; it’s what’s the right org structure for you as a founder. Marc Andreessen and Ben Horowitz are company builders, so they’re building their leverage like a company. a16z is a company.

Not that Peter Thiel is not an entrepreneur, but he’s more of a philosopher today, and so his system is more like a philosopher.

It's all these wild and crazy ideas. Jensen Huang has 46 direct reports or something, which is absolutely insane, but it works for him. You just have to know what your style is, and then build it for you.

Totally. Someone like Sam Altman is a dealmaker, a fundraiser, and a talent identifier. I remember he was at YC and talking to one of my friends about a universal basic income company. My friend didn't end up doing it, but I was like, “Huh, that's interesting. Why is he even doing that?” He's focused on waste, and he's like—

He has a knack for good ideas.

He has a knack for good ideas. This person was 22. He has a knack for high-slope talent, and he has a superpower around raising money.

Yeah. One of the things that I think new founders—or at least me as a new founder—was, “Okay, what are my weaknesses, and how do I get better at them?”

But then a second-time founder is like, “Forget all of my weaknesses. I'm not even going to try. I'm just going to be really good at my strengths, and I'm going to hire other people to do all the things I'm not good at.” That is way more effective.

Yeah, totally. There used to be this sort of skepticism—and there still is, actually—that you can't really incubate a company like this. The CEO has to be the co-founder; they have to drive it. Where do you stand on that debate?

I think the biggest companies aren't incubated because the biggest companies have a founder who is monomaniacal from the beginning and compounds it for 50 years.

Yeah, totally.

That's not incubation. Incubation is, “Let me try a bunch of things.” It's not this monomaniacal, power-law commitment. So if you're just trying to build a trillion-dollar business, I don't think that's the way to do it.

But look, we have a mutual friend who likes to remain anonymous, who incubates lots of companies and launches 4 companies a year.

He loves doing it, he's going to make a ton of money doing it, and that's just his style. He's probably not going to have trillion-dollar businesses, but he's going to have 50 $100-million businesses, which is still incredible.

Right. Yeah. I think one use case is more like breeding. Sam Altman, in some sense, incubated OpenAI in that it wasn't the main thing he was working on. But once he identified that it could be this massive opportunity, he went all in with it.

Yeah. It's more like a venture portfolio. There's probably a bunch of things Peter Thiel has done that we don't even remember, because of seasteading and things like that. They just disappear into the ether, and then the things that work compound. If you have the resources to launch 10 new things and then just double down on the 1 thing that's big every year, that's a cool approach.

Right. A decade ago, Marc Andreessen wrote this blog post on productivity. They were asking him how he thinks about meetings, scheduling, and whatever, and he said, “You know what? I try not to schedule things in advance because I don't know how I'll feel next week, and I want to do things in flow.”

When people are trying to get my time, what I typically say is, “Call me right now, or feel free to check in next week and we'll see how I feel.”

Yeah. It is better aligned with the flow state.

I mean, that's the ultimate flex. If you can pull that off, that is definitely the ultimate way to live. You just work on whatever you want, and the world opens up to you. Obviously, it's hard for most mere mortals to pull off, but if you can get to that level, that is the dream, right?

Right. For certain meetings that people were cold-reaching out about or whatever, I was supposed to try it. I always anticipate that in a few more days I'll be busier than I am, whereas right now, I've got time. Give me a call. So I tried that a little bit, and then Marc did a follow-up interview where someone asked him about that. He was like, “Oh, you know, that didn't work. I tried, but that didn't work. Now I'm just scheduling everything—”

Just back-to-back all day long.

I was like, “Oh, goddamn it. You've been doing that.” He's also mastered multitasking.

I mean, I think the thing you can learn from Marc and other people is to challenge the defaults occasionally. If you're in back-to-back meetings all the time, try this other system and see if it works. I think most people just go with the flow. The other thing is, most people aren't vicious enough at saying no.

Yeah. Just being super hardcore: Does that conference you're going to matter? Does that random get-to-know-you chat mean anything? I think you just say no to most things. I've—

Back in the Uber-versus-Lyft scaling-war days, I remember emailing Logan a question, probably asking him to come to one of our dinners, and he just responded, “War mode.”

I love that. I'm like, most people should actually be in war mode 90% of the time.

Yeah. That same hardcoreness is just the right way to operate.

What have you learned in your extensive founder career about managing your own psychology, dealing with stress, and dealing with the cognitive dissonance of having to present externally like things are amazing even when you're concerned about Google taking out your business? What advice might you have for others in terms of cultivating this?

Yeah, man. I think it's like a cold plunge or something. It's just shock therapy, and eventually you're bathed in existential fear all the time and you're like, “Whatever, let's bring it on.”

Totally.

I don't think there's much other than knowing that it's going to happen. There are going to be tons of ups and downs. One of the clear lessons from Thumbtack, which has now scaled and done so well, was that there were lots of near misses where we almost died and times when we thought it wasn't going to work. You just have to stay in the game.

Yeah. If you can stay in the game, you can eventually win.

Yeah. Thumbtack was very good at staying in the game so that we could eventually win.

Yeah. It's funny. There are degrees of panic in the beginning. You're getting your feet wet, and it's like, “Oh my God, someone said something mean about me on the internet.” Then it's like, “Oh my God, someone wants to leave my company.”

Then it's like, “Oh my God, we might die. We only have 6 months of payroll.” Then it's like, “Oh my God, we're going to get sued.”

Like, you know—

Well, it's like the first time you get sued, or you're worried about getting sued. Eventually you have a board meeting and there's just a list of lawsuits, and you review them quarterly. That's actually what success is: you're going to have that stuff constantly, and it becomes just part of business.

Yeah, yeah. You just keep building up more capacity to handle these things, and they just—

Yeah. That's where I think co-founders are the best. They're the people you can be totally open with and trust when you say, “Well, that situation's fucked up.”

Totally.

They're someone you can trust and be on the ups and downs with.

Let's go to Athena, and let's go where we started, which is the intersection between humans and AI. How have you thought about building Athena given developments in AI, and what does that mean for Athena?

It started fully bootstrapped and human-only, scaled to 1,000 people with no outside capital. Then we had this AI moment a few years ago, and it was clear we could make a big AI play. This could become a generational business, or we could get railroaded by AI, and so the only option there is to go big.

Our vision here is that you take the best of a human and the best of a machine and wrap it into 1 unified product. What the machine does will increase over time, just like a self-driving car. The human is the UX on top. Over time, the human assistant will focus more on project management and high-leverage work and really be that caretaker, that empathy, that partner. The machine will increasingly do the more rote, mechanical, administrative stuff.

Just as when you get into a Tesla, Elon made them with steering wheels so you could drive them because they didn't drive themselves when he started, every time you get in the car, you're actually teaching the machine how to drive.

Yeah.

These models don't generalize well. You have to bring edge cases under distribution. Every time you drive, the model gets more powerful. The same is true in executive assistance. Our humans are driving. They're doing the tasks, but we're building machine models that are being improved by the way they act.

The more your assistant does tasks for you, the more our models will improve, and that'll allow the human to upgrade what they work on to more high-leverage stuff.

Yeah, that’s exciting. So what else can you share about the company, or things to watch out for, for people who want to get assistance?

Yeah, look, I just say: If you don’t have an assistant, you are the assistant. Step 1: If you’re starting a business, you should hire an assistant. If you’ve got $20 a month, use ChatGPT. If you’ve got $10 an hour, go on Upwork and hire someone yourself.

If you hire someone yourself, my main recommendation is that you need to interview a lot of people. At Athena, we have 50,000 assistants apply per month, and we put them through a huge battery of tests. We hire about 1 in 300. So if you’re going to hire someone yourself, make sure you interview not 10, but probably 50 people.

If you have the resources for $3,000 a month, then you should work with a company like Athena, which can recruit, train, manage, and teach you how to delegate. When you have the resources for an in-person assistant—$100,000 to $150,000 a year—you should do that as well. Ultimately, when you get to a later stage, you have a chief of staff, an in-person assistant, and a fleet of Athena assistants in the cloud.

Yeah, that’s a great overview. Well, let’s wrap the podcast on that.