For centuries, the carrier pigeon was the undisputed king of global communication. They were the fastest way to move vital information across continents. Empires relied on them. Financial dynasties used them to dominate markets. Breeding and training these birds was a massive, highly respected industry.
Then someone invented the telegraph.
The telegraph did not make the pigeon faster. It bypassed the bird entirely. A mechanical system of copper wires proved superior to a biological system of feathers and muscle. The transition was violent and rapid. In a single generation, the carrier pigeon went from an indispensable economic asset to a park nuisance. There was no retraining program for the birds. They were simply retired.
You might think it is insulting to compare a modern professional to a bird. But we are facing the exact same paradigm shift right now in 2026.
For the last hundred years, humans have been the only viable system for processing complex information. We are biological algorithms. Companies paid us to write software, design marketing campaigns, and map corporate strategy because no other system could handle ambiguous reasoning. We built our entire society around the assumption that this biological processing power would always be valuable.
That assumption is now dead.
Agentic AI systems are the new telegraph. A decade ago, people worried about robots taking over factory floors and self-driving cars replacing truckers. Those were narrow tools. Today, we have synthetic cognition. Swarms of autonomous agents now design, write, and deploy entire software architectures in an afternoon. AI models generate photorealistic films from text prompts. Autonomous reasoning engines navigate complex corporate litigation without human oversight. These new systems do not sleep. They do not demand salaries. They do not make errors born of boredom or ego.
When faced with this reality, people retreat into a predictable defense. They claim their specific job requires a unique human element. Ten years ago, they said algorithms lacked creativity. Now that AI routinely wins design competitions and generates chart-topping music, the goalposts have moved. The modern defense is that AI will always need a human supervisor.
This is a comforting lie.
The supervisor role is a temporary bridge. As these models gain deeper reasoning capabilities, they correct their own errors and manage their own workflows. Believing that your specific brand of human management is magic is a biological conceit. It is not an economic reality.
We like to point to the past and say technology always creates new jobs. But previous industrial revolutions created jobs because human cognition was still required to operate the new machines. The tractor replaced the plow, but it still required a driver. The current generation of AI does not require a driver. These systems are designed to operate autonomously, learn from their environments, and optimize their own processes.
We are approaching a threshold where human labor is no longer the foundational engine of economic growth. The transition is happening rapidly across every white-collar sector. The trajectory is fixed. Economics favors efficiency. Biological workers are expensive, inconsistent, and slow.
Society operates on the assumption that employment is the default state of adult life. The challenge is not stopping the algorithms. The challenge is figuring out what happens to the pigeons when the wires are laid.
Margin Notes
Drop your email to get early access to tools and engineering insights before anyone else.