Are you a horse?

Not a day passes without a new headline of how AI will replace workers. We all read them, and we all wonder. What does this mean for me?

Decades ago, Nobel laureate Wassily Leontief famously observed that human labor could eventually be replaced by technology just like horses were replaced by tractors. In the early 20th century, horses were not replaced because they were lazy; they were replaced because the internal combustion engine made them economically irrelevant. The "subsistence wage" of a horse (oats, water and shelter) simply became higher than the "operating wage" of a tractor.

Today, we face our own "tractor moment." We are entering an era where the cost of human labor is colliding with an AI "compute wage" that is plummeting. A CEO recently replaced a six-person marketing team—costing over $300,000 annually—with AI software costing just $2,400 a year. Work that took humans 38 hours was completed in 18 minutes.

The Speed of Obsolescence The timeline for this shift is shrinking rapidly. Geoffrey Hinton, the "Godfather of AI," warns that AI capabilities are now doubling roughly every seven months. This exponential growth means that by the time you master a routine workflow, an algorithm may already be doing it faster and cheaper. Goldman Sachs estimates that 25% of all work tasks in the US could be automated, with administrative and legal sectors facing the highest exposure.

The Counter Narrative: You Are Not the Horse, You Are the Driver While the statistics sound grim, Shyam Sankar, the CTO of Palantir, offers a critical antidote to the panic: Doomerism is a luxury we cannot afford. Sankar argues that AI is not a replacement for the worker; it is a 50x force multiplier.

He describes AI as the "printing press of our age". Just as the printing press broke the monopoly on knowledge, AI breaks the monopoly on technical skill. It allows an ICU nurse to process patient data instantly to spend more time at the bedside, or a manufacturing technician to fix complex machinery without waiting for an engineer. In this view, the worker is not a cost center to be cut, but a "co-creator of value" who wields the machine.

The Ultimatum: Wield It or Be Replaced To avoid the fate of the horse, you must stop competing against the machine and start "orchestrating" it. If your job is generating first drafts, summarizing emails, or writing basic code, you are in the "subsistence zone".

1. Don't just ask AI for answers. Learn to frame the problem. Your value is no longer in memorizing facts, but in asking the non-obvious, high-judgment questions that steer the model toward breakthrough insights.
2. Master the Cultural & Political Context: AI can write a technically flawless contract or operational plan, but it cannot navigate office politics, read room dynamics, or anticipate why a rational proposal will fail inside a specific corporate culture. Context is your moat. This human ability to navigate irrationality and nuance is currently untouchable by algorithms.
3. Become an Agent Orchestrator: Stop viewing yourself as a solo worker. Start training to be an "architect" or manager who directs fleets of AI agents to complete complex workflows. The future belongs to those who can manage the machine, not those who compete against it.
4. Reclaim Your Agency: As Sankar argues, "The American worker is not deficient; he is under-leveraged. AI is the lever". Aggressively automate your own drudgery so you can spend your day on high-leverage judgment.

The horse had no choice when the tractor arrived. You do. The era of the "information worker" is dying, but the era of the "insight worker"—who commands fleets of AI agents to do the heavy lifting—has just begun.