Sam Altman on YC, Universal Basic Income, Trump, New Cities
On April 13th, 2017, Sam Altman was interviewed by Nellie Bowles (what an interesting background) at an event in San Francisco. The talk was recorded and made into a podcast.
Here’s a summary of the podcast, with my brief commentary in {cursive}.
Enjoy, and/or discuss on Hacker News.
Q (Question from Nellie): What do you/YC do?
We give people money and advice, help them start startups.
We built a platform to attract founders. We get 125 companies in a batch, twice a year. A startup is usually 2–3 people, some software or hardware. We have funded enough companies that we now have a network.
Q: What’s your day like?
Mostly email. {like Naval Ravikant}.
Sometimes I meet new problems, e.g. Boom (supersonic airplanes), or how can I find one user for my product.
At YC we don’t focus on a particular industry. I like meeting new entrepreneurs. I spend lots of time learning new tech.
We mostly “pick new founders and new ideas”.
Q: How did you get interested in tech?
I was a nerd. Still am. Always interested in tech, when it was deeply uncool. E.g. learning Qbasic in elementary school. It felt like magic.
As humans we have evolved very fast, from being “primitive” not long ago to today’s cool technology.
Q: How did you become Sam Altman? Your startup Loopt?
Loopt wasn’t very successful. I turned out to be better at investing. I started early on, before valuations went crazy. My first investment was Stripe,in which the strike price for the investment was less than $1M post-money.
I then got close to YC.
I started believing that startups were going to change the world. Not just software, but also biotech, nuclear fusion. They start with low amounts of capital, iterate quickly.
I then Raised a fund. Scaled it up, then talked to Paul Graham for almost two years about taking over YC. I had a plan for YC.
Q: What did you learn from Loopt?
You can’t create a market that doesn’t want to exist. We thought location to find friends and stuff was going to be big, but we were wrong.
Q: You invest through YC, and yourself. What do you invest? E.g. Cruise?
In general, I do few non-YC investing. I use my capital for things I think are really important for a good future, that if I don’t invest they won’t happen.
I thought driverless cars could help with the Housing crisis, reduce traffic deaths.
Kyle Vogt (founder of Cruise) came to YC with nothing, and said “I am going to build a self-driving car in 10 weeks”. I said “No you aren’t”. But he did. After 10 weeks, his car drove itself to the YC office.
Good engineer. Vision for the future.
Helion (nuclear fusion company): a potentially huge impact for humanity and quality of life.
I want anyone to do fusion, any company. It would be transformative.
With super cheap energy we can get clean water, food, better transportation.
Q: What are the trends? Is VR still hot?
You should be skeptical about trends. It’s important to think as indipendently as you can, unless it’s an absolute megawave that is going to spawn a lot of companies.
To give you an answer: Machine Learning applied to any vertical — e.g. farming, radiology, self-driving cars.
There are hundred-billion dollar companies waiting to be started.
Q: Tell me about the Universal Basic Income project.
Universal Basic Income: It’s now 100 families in Oakland (they will be 1,000 soon), we give them $1,500 a month each.
I don’t think UBI is the only answer.
I think we are about to see a change in the economy in the order of the industrial revolution. A huge amount of human jobs will change. We’ll find new jobs on the other side. Maybe we’ll all be artists? I don’t know.
In the meantime, we’re going to have a real disruption. Not many talk about what we, as the tech industry, are doing to solve the problems we are creating.
At a minimum, we should create a floor, so that no one is worrying about a place to live, what to eat, healthcare.
What most people get wrong: if most robots take the jobs, our cost of living will go down a lot.
Mass unemployment (driven by automation and robots) means a lot of money to redistribute.
Important question: What makes people happy?
And what if we don’t need work to survive?
We’ll find answers to this.
People want to contribute.
They want to have a say in the future.
I’ve talked to low income people, most people don’t want a hand out, most want a meaningful job.
Q: How are you measuring success for this program?
I hope it turns out to be a good program, success means getting good data to decide whether it’s a good policy or not. Are people more or less happy? What about their work?
YC Research, which runs the program, is a non-profit part of YC, separate office, hands-off from the rest of YC.
Q: Do people like the program?
I’ve been asked not to comment until when we’re done, and then we’ll share a lot.
Q: Lot of folks think it’s strange to have startup investors doing an experiment to families in Oakland.
Totally understand the sentiment. We’re not the only fan of UBI. There’s a broad support. We are in a place where we can actually study it. We can contribute to the policy discussion with data.
We talked to community groups and politicians, and we picked Oakland because they were very supportive of the experiment.
The Government will eventually do it if it turns out to be a good policy.
This is a mega change in society. We have evolved to survive. If you take away that need or drive, what’s going to happen? I am optimistic, and I think we need data on how it impacts people.
Q: What jobs do you think are AI-proof?
Hard question. Which won’t? In the short term, all repetitive work that doesn’t require human connection is going to be vastly replaced by AI.
Some repetitive jobs, however, still needs a human connection.
Q: Is this reason for the rise of populism?
I do. When I went to meet 100 Trump supporters, it was striking to me how much San Francisco is an incredibly optimistic place, in a way that most of the rest of the country isn’t.
People think Globalization = Technology, but it’s not true.
With the current wave of populism / Trumpism, people really blame globalization (easier enemy for Trump to talk about), and automation, which is an orthogonal vector.
Q: What inspired you to go meet these Trump supporters?
If I get data that my model of the world is wrong, I go get new data. The Trump phenomenon — clearly something about my model of the world was wrong. There was something I wasn’t understanding.
I had to get new data then.
Q: You should talk to my Dad.
A lot of my friends told me I should talk to their Dad.
I try not to fall into this insular traps. I didn’t know how to find 100 Trump voters.
Q: Your work in politics. Organizing tech power players start some movement.
I want to do something. We don’t have a name for it.
Are there a set of statement that we can ask a tech company to commit to? E.g. Data protection. Immigrants. Committing to creating jobs throughout the country.
We had some informal meetings. I’ve talked to a few hundred people about it.
We haven’t figured out what’s the right thing to do — I hope we do soon. It’s more of a workers console. It should come more from employees.
Q: Trump tech advisory — Elon Musk, etc.
I’m happy Elon Musk is there. I couldn’t stomach myself.
I wasn’t asked to be part of it. I don’t think Trump likes people that have been vocally opposed to him.
Q: One of the most controversial figures in Silicon Valley is Peter Thiel. One of your best friends.
There was a lot of anti-Peter, protests outside his house. He was one of the few vocal Trump supporters, I disagree with his endorsement but nobody deserves what happened to Peter (constant barrage of negativity, protests, etc.). Peter should be able to support a political candidate, especially if it’s unpopular.
It’s very hard for me to defend Trump, but Peter’s support of him doesn’t mean he hates immigrants, for example.
Q: What are the things that Trump is right about?
He’s right that a lot of Americans have been left behind. The tech industry has some responsibility for this.
But we need an economy that has more winners.
There’s a graph that shows productivity gains over the last 60–70 years, and then average wage growth adjusted for inflation, which has been flat ever since the late 1970s. This is the problem.
I’ve heard this a lot when I spoke to Trump voters.
I don’t think he’s right on how to solve it, but he has identified this thing in a way that spoke to people in a special way.
Something has gone wrong with the American dream.
Q: Does Peter likes the direction where this is going?
I don’t know. You can ask him.
Q: How much of a treat is Trump to Silicon Valley?
I don’t really know. It’s hard to say what Trump is really going to do. Most campaign promises have not happened. It’s hard to separate rhetoric from actions.
Q: Local government. I’ve heard you’ve tried to find someone to run for Governor of California?
The California Governor is an interesting position in terms of opposing Trump. Not a lot of people wanna get into politics right now.
This was mostly a few dinner conversations, not much more than that.
I agree with Trump’s diagnosis of the problem: the system is rigged, we have economic injustice.
Q: Who did you approach?
I want to be respectful of people’s privacy.
Q: Would you ever run?
I don’t think charisma is my strength. It’s a politician skill, it would be a challenge for me.
Q: What about finding a Koch brother in Silicon Valley?
I’ve been trying. Unsuccessful. Maybe a progressive version of a Koch brother.
Q: It could be you.
Not quite enough money. Maybe some day. The amount of resources that that organization deploys every year, for decades, is unbelievable.
You need someone to commit several Billion dollars in cash over several years.
Q: What do you say about Jared Kushner and the innovation group?
I don’t know anything about it. We’ve never met in person.
I hope for their success. They are our executives. I hate the spirit in San Francisco of constant negativity and hoping they fail. Let’s be proactive instead.
Q: You are building a city.
Not yet.
One of the biggest problems facing the country, and certainly the Bay Area, is the cost of housing. It compounds into everything else.
There has been a complete failure of state and local governments to build enough housing to fill demand. People are priced out of their home.
What we are researching now: if we find land to build, can we build houses for $100,000, so that a family could afford a house?
Can we use technology and some policy work to make housing affordable again? Otherwise, most of the wealth that will be created will flow to land owners.
If we can figure this out, we’ll then look for a place to do this.
In Missouri, where I’m from, there are homes that cost $100,000.
Q: You want to build a suburb, then?
No. We don’t want to do that.
Can we get to the same price point ($100,000), in a place where people are near their job, close to a city, etc?
On a side note, the number of people who are pro-environment, and anti-housing in the city, is nuts to me.
Q: Tech industry is not known for being great civic engagement folks (Google buses, etc). How would you make a city that has a civic soul?
How can we avoid building a “smarter divide”?
We would build this in a community where we could engage. Tech have been bad at engaging with cities, and vice versa.
It will only work if we’re able to do it with really good engagement. The Oakland UBI was a good thing, for example.
Another issue: the current political climate, everything gets tribal. 50% of people would now object to their kids marrying someone of the opposing party.
Super-tribalistic thing that supports fringe behavior is bad.
Q: You are a survivalist. A prepper.
Is this really what you rather talk about, and not Open.AI? There are things that matter in the world, and then my weird hobbies, a boyhood survivalist fantasy, that’s it. The world is unstable, but of all the things we can talk about…
[The conversation then goes into Q&A]
(49:40) Question from the audience: What do you think of Fully automated luxurious gay space communism?
I’m sorry, I don’t know what that means. (here’s the link)
[The Q&A continues…]
