About Garry Tan
Garry Tan, President and CEO of Y Combinator, has been speaking about the rise of "AI-native" companies, which he describes as organizations where small teams use AI agents to achieve output previously requiring hundreds of employees. In a March 2026 appearance, Tan stated that shifting from writing individual lines of code to managing AI agents has produced a roughly 400x increase in output. He has said that founders should treat AI not as an autocomplete tool but as a workforce, and that core organizational components such as skill files and resolver tables can be mapped to agent-based systems. Tan has also cited examples of companies reaching $6 million to $12 million in annual revenue within six to twelve months with teams of under a dozen people, a phenomenon he attributes to large language models.
Tan has discussed Y Combinator's role in standardizing seed-stage funding through the SAFE document, which he described as a pivotal moment in Silicon Valley. He has argued that lines of code can be a valid productivity metric when paired with rigorous testing, and that "taste" — the human ability to judge what is good or bad — is a durable asset that cannot be delegated to AI. Tan has also commented on housing policy in San Francisco, stating that rents should be lower and that increasing supply, rather than subsidizing demand, is necessary. He noted that there were no new housing starts in San Francisco proper for the prior calendar year.
Source: AI-verified profile updated from Garry Tan's recent appearances.
Browse all interviews →
Transcript (28 segments)
I
Interviewer0:15
Welcome back. All right, we've got a very exciting topic to talk about which I'm sure will intrigue all of you. It's called founder mode, but we're going to retitle it founder mode versus manager mode. With me is Garry Tan, who runs Y Combinator, who actually maybe accidentally started founder mode. So you can talk about the origin story of this thing called founder mode, which has become a meme. We're going to dive into it, but I think you actually triggered this.
G
Garry Tan0:38
Absolutely. Brian Chesky, obviously a friend, on the board at YC. We torture him as a board member by making him come and speak at every single YC batch. I had seen every iteration of the initial selling serial pitch. I'm pretty sure everyone knows that story and I'm sure Brian is sick of telling it, but we still make him say it every time.
One thing I noticed post-COVID, he came in and added a new section about how much he loves this moment of Airbnb, which was surprising when you're used to hearing about founders who have been out for 10 or 15 years building something worth a hundred billion dollars. His second half of the speech to the YC batch was stuff that was totally irrelevant to them but very relevant to the top alumni. I finally got him to come up to Soma at 9 PM last minute. Steve Huffman was the keynote, and Brian came in thinking he'd speak for 20 minutes but spoke for two and a half hours to exactly the correct audience. That evening must have unlocked at least a hundred, maybe 200 billion dollars worth of value for the YC companies. What he said was that it's entirely possible to start your company and take the classic advice: hire the best people and give them as much room as possible. But Brian found that after nine years of conventional wisdom, he did not recognize the company he started. He had division heads and product teams with their own CTOs and design people. He found that executives were focused on their own fiefdoms over the needs of the company. It got to a point where the CEO had no agency. The crux is that it's possible to have a cage match with no agency. I've been realizing that you can fall back on codified decisions from the last 10 years rather than giving yourself agency. Brian had the bravery to redo the org, change how he hired, and be in the track of both his executives and skip levels. The powerful moment is realizing you don't have to live that way. Choosing agency requires courage, going from manager mode to founder mode, and being willing to rehire or reorg from executives down. If you actually show up, leadership is presence, not absence. Manager mode is saying 'it's your call' and not checking in. It's a misconception that founder mode is hyper-micromanagement. Brian shared that by going into the details and aligning every executive, spending 60 to 80 hours a week, and doing reviews, you can figure out who the key people are and ensure everything comes from your vision. He was willing to say if you're not here for this, take care of them with a great severance. The result is that upon IPO he was shocked, and Airbnb remains one of the most profitable companies in tech.
I
Interviewer8:29
So since Brian gave this talk in front of influential founders, have you seen it cascade and people apply his precepts? How often is it working?
G
Garry Tan8:43
Well, we had a founder mode dinner, and Max from Fair came. Max is an incredible founder. It's easy to fall back on 'hire the best people, they've done this before' but it's not a rinse-and-repeat moment. Everything about every process can and should be changed. We spend a lot of time with Bob McGrew at OpenAI, and his refrain is that this is truly the age of intelligence. Yet many organizations with thousands of people are still operating like it's 2015. Going to ground now means you will have 10 or 100 times more impact because of large language models.
I
Interviewer10:02
Have you seen companies have negative experiences inverting from manager mode to founder mode, or is it all positive?
G
Garry Tan10:13
I think the tricky thing is that Airbnb happened to be sitting on a giant gold mine. You have to fly the plane during the transformation and continue to grow. The underdiscussed part is what if those executives you disagree with have had too much free reign? Can you rein them in and replace them? If you have a gold mine and things are going up and to the right, great. But in many scenarios it's scary to replace three to seven leaders at once unless you have traction or cash.
I
Interviewer11:11
You mentioned that you have to embrace more conflict with founder mode, and maybe that's why manager mode emerged—people are conflict-averse. Talk about embracing conflict.
G
Garry Tan11:29
I'm still kind of a people pleaser but I'm aware of it. Left to your own devices, surrounded by people who agree with you, it's easy to take your foot off the accelerator. It takes real courage. The call to action is to remember when you had nothing and it was do or die. Reconnect with that original courage.
I
Interviewer12:38
Assuming you have to hire some executives at scale in founder mode, how do you evaluate which ones might succeed?
G
Garry Tan12:52
The YC angle is that if you sit down with an exec for 10 minutes, it better be the highest bandwidth conversation. You can root out manager mode by how much information they give in a 30-minute meeting. The people who fit zero to one are those you would want by your side from that period. There's no difference between that moment and this moment.
I
Interviewer13:54
If you were to put on a boot camp for teaching founder mode, what would be the top modules?
G
Garry Tan14:05
We might need an encounter session with some screaming. Post-product market fit, it's too comfortable. If the leader doesn't show up, deleterious cultures evolve. Power loves a vacuum. Executives building their careers will lie to each other and themselves. You need to root out those people. If the leader shows up, it makes it hard for them to exert power.
I
Interviewer15:27
Let's say we start a company from scratch. Have you analytically looked at success versus management style?
G
Garry Tan15:51
We really should. I have a one-man data science team, so I need to hire more. It would be awesome to do that mathematically.
I
Interviewer16:05
Let's open up for questions. Um, first question.
A
Audience Member16:31
Hi, this is Anu from Scaffo. We had Ali on stage just before you, and he's clearly hardcore founder mode right now. Why do you say scary? But the founder mode question is the inverse. Prior three years when he was hurting, the company was run by Frank Slootman. Founder mode is the story of the day, but before that, the story was that founders should step aside. How much truth is that? Is it founder mode, manager mode, or just Satya versus Ali?
G
Garry Tan18:02
We're trying to implement founder mode at YC, and I'm not the founder. Having Brian on my board means I have clear governance. The board of YC never met; now we have quarterly board meetings. I'm the professional CEO, but I'm trying to do founder mode by act. It comes from the founders. I have a 50-page deck and we go through everything. I'm trying to do founder mode with clear governance. We get aligned on goals, and then I share that with my 14 partners. Founder mode means someone needs to make everyone run in the same direction instead of falling back on policy. Without governance, power vacuums attract people who choose out of ego.
I
Interviewer20:24
You mentioned the personal side. Founder mode is intense. What are techniques for founders dealing with the personal roller coaster?
G
Garry Tan20:49
The most important thing I learned the hard way is that many of us try to remake our families in our companies. That caused me incredible grief. I mistook my business for my family. It's not family; it's a sports team. We have to find the best people, and just because you won championships in the past doesn't mean you have the team for the next one. That was a huge unlock for me.
A
Audience Member22:32
In implementing founder mode, there's an undermining of ownership that execs have on product details. In the absence of being told what to do, they take ownership. But if they're used to being told, they wait for that. How do you deal with it?
G
Garry Tan23:10
With Brian or me, we have extreme ADHD and push hard in meetings. The great thing is that after a quarter or two, the best people stay at that pace. When starting a new initiative, we meet daily, then as metrics work, you can back off to quarterly. Brian is probably in his iMessage 15 to 18 hours a day. Founder mode requires pushing the pace, then backing off but never being totally gone.
A
Audience Member24:31
It's not micromanagement, but hired talent will feel undermined if you check everything. How do you set expectations that it's about the outcome?
G
Garry Tan25:14
It's a change in vibe. As an Asian-American, my parents used shame. I had to work hard to separate that. Just because someone gets it wrong doesn't mean I think less of them. If they can convince me, great. I want to hear everything. People need to see you change. It's not about being a dictator; it's about syncing up. Handling conflict poorly meant avoiding it, leading to self-abandonment or authoritarianism. The key is to be authoritative: yes, I'm CEO, you're the executive, we co-create. If I'm wrong, I want to hear it. That creates alignment.
I
Interviewer27:59
Channeling Brian, he'd say you're pressure testing everything. Deputies who are good enjoy it because they've considered alternatives and can debate. Those people thrive and become happier because the CEO knows the details.
A
Audience Member28:41
How do you maintain efficiency while maintaining alignment? Steve Kerr can't call a timeout every time Steph Curry gets the ball. How do you keep alignment efficiently?
G
Garry Tan29:09
I don't have a good answer other than presence. You need a clear sense of priority, dialing up what's on fire and dialing down maintenance. Eric mentioned that setting standards is key. Your job is to set consistent standards so people emulate them. Time allocation: Vinod taught that your calendar should be half empty on Sunday so you can be proactive. Jensen told me he cancels half his meetings every Sunday. You can definitely do this at scale.
I
Interviewer30:26
Cool. Thank you very much, Garry, for joining.