Migration has run continuously for three hundred thousand years and it is not stopping now
Every person alive is descended from someone who left.
People left East Africa and kept going until there was no more land to cross. Bantu farmers carried iron across a continent. Merchants and monks moved along the Silk Road. Guest workers rebuilt postwar Europe. Gulf cities were built by imported labor. A generation of Bangalore engineers went to Silicon Valley and helped turn software into an industry.
The pattern is simple. People move when the place on the other side makes their work, their money, or their children’s future worth more.
It is still true. Take two people with the same skills and put one in San Francisco and one in Paris. Their talent did not change, but the room around them did: one person is closer to capital, colleagues, customers, and the institutions that make work move faster.
Paul Graham explained why in Cities and Ambition: cities send a message about what is normal. In some places the message is “you could do more.” That message travels badly over Zoom.
Being in the same room is a luxury good
The return-to-office fight got framed as a culture war, but most companies wanted people back in some form. The harder question was who could afford to make that happen.
The lease is only the visible cost. The real bill is relocation packages, immigration counsel, temporary housing, school placement, tax questions, spouses, pets, furniture, and months of someone’s life spent getting stable instead of building.
The top AI labs and fastest-growing companies globally paid anyway. Not because offices are fashionable again, but because proximity still changes the work. Judgment, trust, taste, and momentum all travel faster in person.
What slows them down is not conviction. It is the infrastructure and support required to move one person from where they are to where the work is happening: visas, housing, payroll, compliance, and a dozen small dependencies that each introduce delay.
The smart money is already moving people
Tom Hammer wrote that “30 of the 70 companies in our last @speedrun batch had founders born outside the US”. He also said that if Speedrun keeps doing its job, that number only goes up.
That is the point. The next generation of companies is already international before it is mature. Founders are building products and teams across borders from day one, while others are stuck in jobs that keep them close to a visa but far from the slope they could be on.
Entrepreneur First pulled operations out of cities like Singapore and opened in San Francisco. That is not a small operational choice. It is a bet that being in the same place is worth the cost.
A six-week visa delay used to cost one person. Now it costs a team
Every technology that raises output per person also raises the stakes of where that person sits. AI made that jump much larger, but most companies are still treating migration like an HR workflow.
One strong engineer with a good agent setup can now do work that used to require a team. The bottleneck moved. It is less about typing the code and more about deciding what should exist, what is worth shipping, and which tradeoffs matter.
That changes what a stuck visa costs. Six weeks of waiting used to mean six weeks of one person’s output. Now it can mean six weeks of decisions not made, agents not directed, customers not reached, and work that never got sequenced.
Companies can feel the cost even if they do not have language for it yet. Some are already paying meaningful bonuses just to get a new hire in the door a week or two earlier.
Remote is the cheaper option. Relocation is the better investment
The obvious reply is remote work. If the work is digital, why should anyone need to move at all?
Because only half of the frontier stayed digital. Money is flowing into data centers, robotics labs, battery plants, manufacturing capacity, defense, hardware, and physical infrastructure. The teams building that future often need to be near the machines, the labs, the customers, and each other.
The people who can do this work are rare and randomly distributed. The places where the work is happening are not random at all. That mismatch is the problem.
The people building physical AI often need to be physically present. The world is reorganizing around that fact faster than the immigration and relocation systems can keep up.
Everyone is building infrastructure for AGI. Nobody is building it for humans
Most AI companies produce things that live on a screen: text, code, images, analysis, workflows. Gullie is different because our output has to land in the physical world.
Our output is a person standing in a different country.
That means the product cannot stop at a recommendation. It has to carry the move through vendors, lawyers, payments, housing, timing, documents, and all the edge cases that only appear once someone’s life is in motion.
We are rebuilding relocation around speed and transparency because the old tempo no longer matches the rest of the world. People who get answers from AI in seconds are not going to accept “we’ll get back to you next week.” Companies that move at AI speed cannot have their hiring plans blocked by an opaque relocation process.
Migration is one of the oldest technologies humans have. The strange thing is how little infrastructure we have built for it. AGI may change what work looks like, but someone still has to get the right person to the right room.