Social Capital and Innovation: Evidence from Facebook Friendship Networks -- by Brad Cannon, David Hirshleifer, Joshua Thornton

TEXT START: Using Facebook friendship data, we study how three aspects of social capital shape innovative activity.

The Dissection

The paper shows that innovation is heavily gated by access to affluent social networks. Economic Connectedness—having a larger share of high-income friends—is associated with substantially more patents and breakthrough patents, with evidence pointing to financing as a mechanism.

What the text is really doing is converting unequal access to money and elite networks into an innovation story. Its strongest finding is not that “social capital” broadly creates progress. It is that wealth-linked networks route capital toward already advantaged places and people.

The Core Fallacy

The paper treats increased patent output as evidence of systemic productive health. Under the Discontinuity Thesis, that inference is defective. AI-driven innovation can increase while the mass employment–wage–consumption circuit is being severed. More patents may accelerate labor displacement, capital concentration, and the collapse of ordinary productive participation.

The study measures where innovation is produced. It does not establish that innovation preserves broad employment, wages, or human economic necessity. Its impressive percentages are therefore compatible with terminal capitalism: the machine is becoming more productive as the population becomes less necessary to it.

Hidden Assumptions

• Patents are treated as a sufficient proxy for social and economic progress.
• Innovation is implicitly assumed to diffuse into broad employment and wage gains.
• Financing is treated as a growth mechanism rather than as a channel for concentrating control of AI capital.
• Elite-connected networks are treated as productive infrastructure, without confronting their exclusionary function.
• Existing institutions are assumed to remain capable of converting innovation into mass participation.
• Facebook friendship ties are treated as a meaningful measure of social capital, even though the strongest result may simply be the persistence of wealth-based selection.
• The paper does not test whether innovation remains human-labor intensive once AI dominates cognitive production.

Social Function

Classification: partial truth, elite self-exoneration, and transition management.

The finding is probably useful and may be causally meaningful. It is not copium in the narrow empirical sense. But its framing sanitizes hierarchy: privileged access to wealthy networks becomes “social capital,” while the underlying mechanism is preferential financing for people already near capital.

As transition management, the paper maps a real survival channel. Those embedded in high-income financing and innovation networks are more likely to remain useful to Sovereigns. It says almost nothing about the majority outside those networks, whose productive role can still be eliminated by the innovation being financed.

The Verdict

A valid local result with no exculpatory systemic meaning. Economic Connectedness is not a rescue for mass capitalism; it is an access pipe into the shrinking zone where capital, AI, and human indispensability still overlap. The paper identifies where innovation gets funded while ignoring what that innovation does to everyone else.

It records a functioning organ in a dying organism—and risks mistaking the organ’s increased output for the organism’s survival.