I like New York more than San Francisco. It isn't close.
I would still tell almost anyone building in AI to move to San Francisco first.
Both of those are true, and this year I got to test them back to back. I moved to San Francisco in January and stayed until July. I stopped in Pittsburgh, where I was born. In September I moved to New York.
San Francisco makes it easy to do nothing but work
People say this like it's a complaint. For a builder it is the whole point.
The city is quiet at night. Very little pulls at you. The person at the next table is probably debugging something, and the first question at any dinner is what you're building. After a few months you forget there was ever another question.
That sameness is a tool. You don't have to manufacture focus in San Francisco, because the city supplies it. Every hour you spend not working feels slightly out of place, so you mostly don't spend them.
Why the builders are there
The labs are there. OpenAI and Anthropic are headquartered in the city, and a large share of the people building the models live within a short drive of each other.
That density helps in a few ways. You hear what people are trying long before it shows up in a launch post. When everyone around you is shipping, a slow week is obvious. And the engineers you want to hire are already in the same rooms you are.
If the hard part of your company is still making the thing work, there is no better place to be.
A stop in Pittsburgh
Between the two, I went back to Pittsburgh. It is a good place to look at seven months of work from 2,260 miles away, and to decide what the next seven should be for.
New York is a better place to live
New York has more of everything, all the time. You walk everywhere. Every kind of person and every industry is a few subway stops away, at any hour. A free evening in San Francisco tends to turn into more work. In New York it can turn into anything.
I love living here. That part was easy to decide.
It is also where the customers are
Lunon rewires how companies operate for the age of AI. The companies that need that most mostly aren't tech companies. They are in finance, law, healthcare, media and the headquarters of large companies, and a lot of them are here.
In San Francisco, the people you meet are building what you're building. In New York, the people you meet might buy it.
The order matters
Go to New York too early and you are selling something that isn't finished, surrounded by things that are much more fun than the work. Stay in San Francisco too long and you build something great for an audience made mostly of other builders.
So my advice is the route I took. Build in San Francisco while the hard part is making it work. Move to New York when the hard part is getting it to the people who need it.
I'm in New York now. That tells you which half of the job I'm on.
