Session IV | Pirate Wires Keynote with Garry Tan

Yeah, a strong recommendation. Yes, forcing is a bit strong. Like, do we force anyone to do anything in America anymore?

Yeah, I mean, San Francisco right now is sort of brimming with the kind of crazy energy that I can only imagine what that felt like during the personal computer revolution down in the South Bay with the Homebrew Computer Club. What's happening is that the rocks can talk, literally. We have taught silicon language, and large language models were trained by engineers right here in San Francisco at OpenAI and Anthropic, are now sort of empowering a whole new generation of startups that are actually not just making these toy dinky demos that you see on X, but things that are actually creating new jobs, new revenue. So something weird happened even in the past couple years at YC. The typical company increased their revenue in a matter of 10 weeks by 5x. So in aggregate, all the YC... we fund about 270 companies every six months, so going on 600 a year, and in aggregate they went from sort of just under five or six million dollars a year in revenue across all those companies to more than $30 million in a matter of 10 weeks. And that has never happened before. It's the best indicator that people are actually making real software that actually solves people's problems, and those people love those problems being solved so much they're paying a lot of money very quickly.

Yeah, I mean, let's see. I did not consider myself really that much of a political animal at all until really when COVID happened. I started seeing basically elders in my community being stabbed, robbed, pushed down. Vicha Ratanapakdee was a Thai man in his 70s who was killed on the streets of San Francisco. And I saw the reaction from the DA at the time, this man named Chesa Boudin. You know, son of... did someone just make a fart noise? I thought I heard that. I mean, this is someone who literally said, you know what, that was a temper tantrum, I want to let this person out. I mean, in so many words, like he did not take crime against our community seriously. And I looked at the numbers. San Francisco is 30% Asian-American, the voter base is 20 to 25% Asian-American, and yet how could we have an elected official really say openly that, you know what, I don't really... I don't care. Like, it's like that Kanye West moment where it's like, so-and-so didn't care about black people. It was like, how can we have an elected official, the person who's supposed to be the top cop of San Francisco, basically come out and say it's open season on Asian-American elders? And I got angry. I just couldn't believe that this was the deal. And actually, I think we were on a Clubhouse with him, and he was an incredible... I had been drinking, this was like... I was in Miami at the time, I was just vibing. It was the Clubhouse heyday, nothing had been serious at that point, all the conversations were very casual. And I was just doing a primer on local politics, just how it worked, who was your supervisor, which the average person in SF that I talked to did not know. And Chesa Boudin just showed up, and I was like, oh no, I have to take this seriously now. But he did log off the second Balaji came on, which is hilarious, of course. I'm not doing this. So there are levels to this game, I guess.

I think the wild thing is there is sort of a meme out there around AI safety that there are probably some real risks. I mean, everyone sort of walks around and says, you know, what's your P(doom)? And then the tricky thing, I think what Scott Wiener figured out was that this topic in particular is a little bit the Jets versus the Sharks. I think he came out and actually said this, that it's the Jets versus the Sharks, and in this case, he didn't realize that and he sort of ran out and started supporting the Jets real fast. So I don't know if he would have done the same thing, I can't speak for him, but I think he did realize there's a lot more nuance to the debate than not. Personally, I think it's really important for us to maintain the ability of open markets, of open source, of people actually being able to get lots of choice about large language models. You know, maybe there was a moment, maybe two years ago, that we were sort of worried that it was possible that a big tech situation could spring up again with access to the world's best models, and that might have been ensconced by really, really overzealous AI regulation that would basically be like the thing that we see in America over and over again: a new industry pops up and then sooner or later there's regulatory capture. And that was a real thing that we were pretty worried about back then. We're thankfully less worried simply because there is open source. We fund thousands of companies who actually need access and they need choices. And so in the broader debate, being able to maintain access to great LLMs is sort of the same thing as being able to access APIs, being able to... the whole innovation economy is sort of enabled by open standards and open source. And that's really the big danger, I think, when you have regulation, especially at a local level. Like, I actually think that the Biden EO, you know, it was not that bad, and actually maybe that was the correct level at which this type of regulation should happen. And the number one thing that I'm really worried about is, you know, this is a powerful technology. It's like they invented cars. These are bicycles for the mind. These are self-driving cars for the mind. And you know, do we prosecute the people who create cars? That would really sort of hamper society, actually. No, we regulate and we're very careful about what people use cars for, and we hold the people who misuse cars responsible. And that's really what I hope happens with this whole topic.

I hope Newsom doesn't sign it. I hope he vetoes it. I have no inside knowledge beyond that, but you know, I genuinely hope he does because if we want California to continue to be this magnet for the best talent in the world, we don't pass things like this.

Yeah, of course. I'm pretty bullish. It was in a bad state, I think, three or four years ago. San Francisco has a $14 billion budget. Every other day, even right now, we're starting to see that there's sort of an outsource of state capacity. And then unfortunately, if you trace the money, that money goes to nonprofits that are spending money not to actually serve the people but to enrich themselves, and you know, sometimes used for political purposes to ensconce... I mean, it is its own form of capture. If you look at every major issue in San Francisco or most parts of California, you'll see the same thing, actually. Why can't we build? For instance, it's because we prevent people from building, even things that have 30 or 50% affordable housing, because the people who try to build the housing have to pay off the nonprofits like Todco. And if we were looking at this, if this were a New York Times article about a third-world country, you wouldn't be surprised. But we should be surprised. We should expect more. This is a grift. This should be illegal. I do not understand why this is possible that you can have a sitting supervisor in the city say, oh, this is a real project. There was a developer trying to make 30% affordable housing on a Nordstrom parking lot, and a sitting supervisor said, oh, I don't know why we're even here talking about this. You know what they needed to do was go talk to John Elberling at Todco and pay him off or work out a deal, and then this thing would be built. And I'm like, that's corruption. That should not be legal. And so when you wonder, like, why is rent so high, why are there all these problems in the city, if you dig a little bit, it's a lack of state capacity, and it is... we tolerate... it feels like we spend so much... it's incredible how much time we spend thinking about the big national issue, and then we totally neglect what's going on in our very backyard. And these are elections that are won or lost by dozens of votes, maybe a hundred votes. And they go on to... Apology actually has a great line about this where they're talking about billionaires, but there are 11 supervisors, each of whom, if you divide the $14 billion budget out, they're responsible for way more than a billion dollars in government spend. And it goes to their cronies, their nonprofit friends. The nonprofits take their workers and go work on their campaigns. Like, it is an entire loop. And that is, in a nutshell, I think, the Doom Loop. We are literally ensconced in this moment where the government funds are actually being spent on things that do not serve the people, that only ensconce the existing politicians in there. And then if we're not paying attention, then it just continues, it gets worse and worse, the budget keeps going up, we keep passing more and more taxes. I'm the... on the political scale, I'm actually fine with the taxes. What I'm not okay with is for us to spend billions of dollars to not serve the people.

Oh yeah, they're still in power. Sorry, guys.

I mean, the hope at the end of the day, I think, is still technology. It's founders, it's people working in tech. I'm hopeful and thankful for people to continue to come to San Francisco despite all the problems. And the reason why people do it is the money's here, the talent is here. And I mean, okay, let me talk about YC for just a moment. The reason why I push people to be in San Francisco, and we all want it to happen, is because literally the rate of unicorns is twice as much if you base your company in San Francisco or the Bay Area. It literally is the high-order bit. It is the one thing that will double your odds of being a successful billion-dollar company. And this is just basic agglomeration effects. If you have all the people... most people from the rest of the... we have to appreciate this about San Francisco: if you come here and you earnestly want to solve a problem using technology, you're in a cafe, half of the people, three-quarters of the other people in that cafe are going to be like, that's awesome, in fact, I want to help you. And if you go to any other cafe in almost any other part of the world, they're going to be like, oh, sorry, you don't have a real job, do you? And they'll back away slowly. And that is incredibly special. This is something that has been happening in the Bay Area for decades. And then now because of the AI boom, the LLMs were literally birthed right here in San Francisco. All the 20-somethings, all the super cracked engineers, they want to be here because they know the high-order bit is here. And so that is the origin of the Boom Loop. We are going to create dozens, if not a hundred, companies worth billions of dollars, and they're going to fill all the office towers, they're going to fill all the apartments. But we have to build enough of them.

It's definitely there. There are lots of people who absolutely love San Francisco and we're working together to try and fight for this. I mean, Chris Larson, Mike Morris, there's so many people who... they already earned... like, what's funny is the founders actually are a little bit premature. They come to me and they say, like, well, I'm here, what can I do? And I'm like, A, you should vote, B, you should join GrowSF or TogetherSF or Stop Crime SF. Like, go find people, just go spend your time. I mean, in AI and in civic engagement, attention is all you need. And so I actually think that that's a really important part. All we have to do is, at dinner, when something happens, talk about it. And then don't go skin deep. Don't take what the Chronicle has to say. Dig a little bit deeper. And if you work through these local political groups, you're going to meet literally the person who works in the city who will tell you, oh, that's corrupt and rotten out because of this. And then when you find out, put it out there, tweet it, make an X post. I will repost you. Like, you know, earlier we were talking with Francis Fukuyama about this idea of how media is shifting. And what I realized is that's actually what individuals are like. All the institutions are a little bit ruined, and all of the trust has now been atomized down to the level of the individual. That's why it's Substack or it's your individual writers. There's new voices and new... like, Pirate Wires is a great example of there's new atomization. And that's what it takes, actually. Attention is all you need. And if we do not pay attention, if we don't actually talk about what's really going on, there's no one else who's going to do it. And then we're just going to read this dumb voter card, and we're going to read the Pissed Off Voter Guide, run by the people who have run the city into the ground. Exactly. And that's sort of how we got here. It's like low-information voters not caring about the local individual issues. And the good news is it doesn't take a lot. Just pay attention. You don't have to be as mad as I am. You can just, when November comes around, let's read the GrowSF guide. And if you're in these groups, like if you're in the United Democratic Club, it's a fantastic Democratic Club that anyone who lives in San Francisco can join. And if you go to those events, you will understand what is actually going on in the city. And your vote and your work in those organizations, it moves hundreds if not thousands of votes. I think this November we are actually pretty much on deck to retake the Board of Supervisors and have a moderate mayor in the city, and that has not happened in many, many years.