Darwin AI · · 565 words · 2 min
You're not competing against AI. You're competing against the person who uses it well.
Everyone asks me about the future of work. There are two horizons.
Short term, today. Santi Bilinkis (Argentine entrepreneur) said it best — you’re not competing against AI, you’re competing against the person who uses AI well. If you use AI fluently, you produce more, create more, make better decisions, and your seat on any team becomes more valuable, not less. So step one is just: use this thing. Aggressively. Daily. The same way the computer transformed work in 1995, AI is transforming work now, except faster.
Medium term. AI will take real decisions. It will make better decisions than we do in many domains. It will work 24/7, faster. There’s an obvious rule of thumb — the more your job is “open a computer and use software,” the more replaceable it is. If your job already happens through screens, the screen can be automated. The work that requires atoms — physical presence, physical action — is harder to automate, until robotics catches up. The work that requires being a human — looking another human in the eye and saying “I promise you this” — is hardest of all. That’s where value will go.
YOUR JOB IS MAINLY… ┌──────────────────┐ ┌──────────────────┐ ┌──────────────────┐ │ open a laptop │ │ touch atoms │ │ be human │ │ use software │ │ build, fix, │ │ trust, eye │ │ │ │ deliver │ │ contact, voice │ │ │ │ │ │ │ │ [REPLACED │ │ [REPLACED │ │ [VALUE KEEPS │ │ FIRST] │ │ LATER, BY │ │ GROWING] │ │ │ │ ROBOTICS] │ │ │ └──────────────────┘ └──────────────────┘ └──────────────────┘ AI eats screens robotics eats atoms this part is yours
So what happens? Two patterns:
More people become bosses of AI agents. Inside companies, fewer humans, each one managing larger fleets of AI work. Outside companies, more solo operators serving niche customers very well, with AI doing 90% of the operational lift.
More people open small businesses with deep human service. Everything that requires trust, presence, eye contact, accountability. These were less prestigious before. They become more valuable now. Skilled trades, in-person services, relationship-driven sales — these aren’t going away.
A friend in San Francisco shared an education-cycle observation: after the 90s software revolution, parents told their kids “study so you can have a white-collar job.” For 20 years that meant: study something that connects you to a computer. We’re entering a new cycle. The next 20 years of “study something useful” probably means studying people and physical reality — relationships, presence, atoms. The white-collar conveyor belt that started with the PC is ending with AI.
In the very long run — robots build robots, AI runs every screen — most repetitive work gets automated. Then humanity moves to the next problems: leaving the planet, or finding new ways to organize ourselves where work and worth aren’t synonymous. We can argue about which.
For today: be the best in the room at using AI, and prepare for a career that involves humans or atoms, at least for a few years. Both directions look right.