When Teams Become Optional

In the AI Story

Putnam's framework clarifies why this matters beyond productivity. Teams are not merely production units; they are the primary site of social capital formation in professional life. The repeated interaction under shared stakes, the demonstrated reliability under uncertainty, the mutual accountability that develops when five people must coordinate to ship a product — these produce the trust that makes the team resilient, the norms that make the organization functional, and the professional identity that makes the work meaningful. When the team becomes optional, these byproducts of teamwork disappear even if the team formally persists.

The Trivandrum training that Edo Segal recounts exemplifies the transformation. Twenty engineers, each suddenly capable of work that previously required a full team. The exhilaration was genuine — new capability, new reach, new possibilities. But the social infrastructure question went unasked: What happens to the relationships among those twenty people when their work no longer structurally requires them to depend on each other? Segal's decision to keep the team was protective of headcount. It did not automatically preserve the social function of the team, which depends on genuine interdependence rather than mere co-location.

The phenomenon is visible across the industry. Startups that previously required technical co-founders can now be launched by solo founders using AI assistants. Cross-functional teams whose existence was justified by the necessity of integrating design, engineering, and product can now be replaced by one person prompting across domains. The code review that required human attention because no tool could catch design flaws can now be handled by AI that reviews more comprehensively than most humans. Each replacement is individually rational. The collective effect is the quiet dissolution of the structural interdependence that made professional communities cohere.

The counterargument — that AI collaboration is itself a form of teamwork — fails Putnam's criteria. The human-AI dialogue may be intellectually stimulating, may produce good output, may even feel like partnership. But it does not build interpersonal trust, because trust requires the possibility of betrayal by a being with independent interests. The AI does not have interests. It does not choose to be reliable. It is reliable by engineering or unreliable by error, and neither constitutes the demonstrated choice under uncertainty that trust requires. The developer who trusts Claude may be confident in its outputs. That confidence does not translate into generalized trust in other humans, which is the form of trust that enables cooperation beyond the individual and the machine.

Origin

The concept crystallizes the confluence of AI capability expansion and Putnam's social capital framework. It names the transition that technologists have documented empirically — solo founders building products, cross-functional boundaries collapsing, traditional team structures dissolving — and asks the question the technologists have not: What is the social cost of making teams structurally unnecessary? The Putnam simulation's contribution is recognizing that this is not a new question but the application of a sixty-year research program to a novel technological condition.

Key Ideas

Optionality kills social capital. When collaboration is optional, it competes with individual productivity — and loses, because productivity is measured and social capital is not.

Co-location is not collaboration. The team that continues to exist in formal terms but no longer depends on each other for output is a team in name only. The trust-building mechanism has been eliminated.

The team was always a school. Its productive function was necessary but not sufficient. The social function — teaching cooperation, building trust, transmitting professional norms — was the invisible infrastructure of professional life.

Rebuilding requires redesign. Protecting headcount is insufficient. Organizations must redesign work to create genuine interdependence around judgment, deliberation, and the decisions that AI cannot make.