Amal Jolly

The problem was that the output was bound to atoms. Skill was locked in a person’s head. Knowledge was trapped on paper in a cabinet. To get anything done, you had to physically gather all these atoms in one room. Humans got better and better at arranging them, and we called it management. Work was a problem of logistics before it was a problem of ideas.

Then came the digital wave. We diligently copied the paper office onto our screens, swapping atoms for bits. The filing cabinet became SharePoint and CRMs. The memo became email. The chatter around a desk became Slack. The project plan on the wall became Asana. We got better and better at arranging these bits, and we called it collaboration.

We perfected a system for coordinating human experts. We reached the logical peak of productivity.

To see what happens next, you have to break work down to its physics. Any collective effort, from building a cathedral to shipping software, is composed of four fundamental elements: the Intent we define, the Knowledge it requires, the Orchestration that connects and guides it, and finally, the Skill to execute it.

For all of history, that fourth element, Skill, was the most expensive and constrained. It was sourced from years of human training and experience, and the entire system of the office was built around optimising its performance.

Then, in late 2022, Skill underwent a phase transition with LLM-based AI. For the first time, high-end cognitive Skill was no longer a scarce resource. It became a utility. The solid, expensive element of human expertise became as fluid and abundant as water. You could summon it on demand.

When a core element of a system becomes free and infinite, the physics of value changes completely. A world built to manage scarce Skill is the wrong framework for a world of abundant intelligence. If the "how" becomes a commodity, human value is no longer in the doing of the task. It collapses into the realms of judgment and direction.

This is the domain of the Conductor.

The Conductor's value is not in playing an instrument. It is in the decisions that separate a templated performance from a transcendent one. Anyone can direct the orchestra to play loudly; the Conductor knows when to demand silence. It is the judgment to know why a piece should be played, the taste to decide how it should feel, and the vision to guide dozens of individual players toward a single, coherent output.

The AI agents are the orchestra, a collection of world-class musicians who can play any note perfectly. But they cannot decide which music to play, or how it should feel. They await direction.

This focus on judgment, direction, and vision isn't just a change in roles; it signifies the emergence of a new, now fundamental, fifth element of work.

If the first four elements define the mechanics of getting things done, this final element defines the quality and purpose of the outcome. It was always present, but it was a luxury, overshadowed by the brute-force requirement of Skill. In a world of abundant Skill, this element is no longer a luxury; it becomes the new centre of gravity for all valuable work.

This fifth element is Taste.

Taste, uniquely human, represents the subtle yet profound capacity for discernment — the ability to distinguish good from great, the meaningful from the mundane. It’s the human sensitivity to nuance, purpose, and emotion. As Kant put it in Critique of Judgement, taste is how we bridge reason and perception. It’s how we turn feeling into something shareable. It lets us move between the abstract and the concrete — to sense what’s right before we can explain why. It’s not just preference. It’s judgment shaped by experience, made legible to others.

This new reality demands a new workspace. The old model, a task manager that is merely a list of labels, is a relic designed to orchestrate scarce Skill. The new model requires the Task itself to become the conductor's podium, an environment where the elements combine in service of a singular vision. It must be a container where Intent is defined, Knowledge is supplied, and Orchestration is the primary human activity. From this podium, the Conductor directs the fleet of AI specialists, shaping their flawless execution into a coherent and valuable output.

This is the final abstraction.

The old roles of arranging atoms, bits, and even skills are being rapidly automated and made available in abundance. The gap between inspiration and creation is narrowing; what we create matters more deeply than ever. Human value condenses into a single essential role.

The last human job is to provide the Taste.

Leonard Bernstein rehearsing Beethoven's Fidelio, 1971