Short one today. I am in Madrid for the weekend visiting my son Somers, who is finishing his dissertation here. Bad golf later. Long lunch with him and Kirsten. Indonesia plans for the summer. Somewhere in the middle of all that is the question I keep landing on as a father, and the same question I keep landing on as someone running a small business: what AI leverage for small teams actually looks like five years from now, and what kind of work is worth picking right now.
Somers is one of my bros, and my best friend. He is hysterical when he wants to be, smart, and a gifted athlete. When he was up in Verbier for spring break he landed his first 1080 on skis. I take that to mean he spun three times in the air, he was also likely inverted and had to find a lot of air to carry that kind of momentum. Impressive, not something I ever was able to do in my younger days, or even want to try. He and I have done a lot of bonding since his brother Finn passed in 2023, of course we all have. Learning to live together as a 3 person family was awkward at first, but now we're becoming pretty familiar with it. Finn was the lynchpin of the family, we all knew it. It's even more plain now that he is gone. One of the things we share is surfing the eastern island chains of Indonesia. Three summers running, and we are going again this July and August. I am not thinking about that this weekend, though. This weekend it is focus time with Somers and Kirsten. Laughing, eating great food, and probably some bad golf today. That is another thing Somers and I bond over. Neither of us are good at it. In fact, we are both pretty awful, and we love it anyway.
One of his keen interests is trading. It started with crypto, like a lot of his generation, enamoured with the stories of quick gains and the easy logistics of moving between currencies. After getting burned, probably more than once, he switched to a specific trade he runs at the open of the New York Stock Exchange. I do not really know what the trade is, but he has been getting good at it. He has a 60-plus per cent win rate, and a rule that he always risks 2 to 1. The upside on every trade is a hundred per cent greater than the downside. The maths is not hard. Win more than half the time, with twice the upside on every trade, and you are in the money.
He is finishing his dissertation now, and is interested in commodities trading. There is a clear career path for that in Geneva, a two-hour train ride from where we live, and a strong masters programme in shipping and finance built around it. One of the lanes that interests him most is putting deals together. Human to human, matching commodities to the markets that need them. I think that is a clever career path to pick in the age of AI, when most uni grads are sweating about getting any job at all. Let the other traders and finance guys fight over equities positions in London or New York. Somers can use his Swiss residency to find a nice niche market that favours his high EQ skills. Being good with people, of course highly important in the age of AI.
Commodities are essential for human survival. People have been trading them for as long as people have been able to barter for them. That is not going away even if robots end up doing most of what we currently do. Robots do not eat wheat. They do not need coal to keep warm. Humans do, and my suspicion is that AI is not going to be the thing brokering the deals for the things humans actually need to live. AI will find the deals. It will help along the way. But the human with a warehouse full of grain in Africa is not going to be wiring money to an AI in Europe to close a contract. There is a human on the other end of that line, managing a relationship with another human. The AI sits in the middle and helps. The relationship still belongs to the people.
That is the same lens I am running on my own use of AI. Right now I might look like a wizard because I know how to wield the tools. Three years from now, AI will be good enough that anyone can speak into their phone and the model will just do the thing. By then, people may not need a guy like me for the wielding. But my hunch is that the people who stay out in front of AI, who keep getting better at it as it gets better, are still going to be needed. They will just be needed differently.
I also think the markets and the services will fragment. Less need for mammoth software companies servicing tens of thousands of users with one product. My belief is the age of AI gets run by small teams of uber proficient domain experts, each one magnifying their own productivity to service smaller, more specific needs. That is the basic vision behind TestVentures, and I have written about it in a few different shapes already, most recently in Fat Skills, Thin Harness. A small team of specialists. A 10x dev. A 10x designer. A 10x sales person. Maybe an AI agent builder and an AI admin to keep the whole thing running. Done well, with no overlap, all of us rowing the same boat, staying nimble, iterating as the markets and the models change, we could do very well. We do not need more than 50 to 100 clients that come and go as the work calls for it.
That is the vision at least, and the way I am thinking about workflows in the upcoming age. It is a fun, slightly terrifying time to be alive. Terrifying because for Somers and his cohort, there are going to be changes in the next 10 to 15 years that none of us can really picture yet. Let us see how the theories hold up.
What I am sure of is this: being on the front of this AI wave is where we want to be. Both of us.
Monthly Revenues $11,800 | Clients 2 | Prospects 0, this weekend is for the family.