Field Note: Free Energy
Date: 2024-04-02 Observer: Trolla Location: Various clusters Status: Observational study
In thermodynamics, Gibbs free energy measures the amount of work a system can perform. It's the useful part. The rest is heat — dissipated into the environment, unusable. Free energy is what keeps a system organized, maintaining structure against the natural tendency toward disorder.
In a cluster, free energy is an edit.
An edit is a localized injection of ordered information. Someone spends time, attention, and cognitive effort to create or modify content. The edit reduces the disorder of the page it modifies. It fixes a broken link. It clarifies a confusing paragraph. It adds a source. Each is a reduction in entropy, and the editor expended free energy to do it.
The problem is that free energy in a cluster is not conserved. Most edits dissipate. You edit a page, and three weeks later someone reverts it, and the energy you put in has been converted to heat — wasted, scattered, gone. The more active a cluster is, the more of this happens. Counterintuitively, high activity can mean low free energy retention. What you want is not high activity but high-energy activity: edits that stick, that create lasting structure, that reduce entropy persistently.
I've been tracking free energy expenditure across a sample of clusters. The median cluster loses approximately 60% of its free energy within one week of an edit. By one month, only 15% remains. By six months, it's negligible. An edit is an act of faith. You put energy into a page, and you trust the system will preserve it. The data shows you should trust it rarely.
There are clusters where this doesn't happen. I've found maybe a dozen — out of thousands — where free energy retention is above 50% at three months. These clusters share certain characteristics: a small but dedicated core of editors, strong norms about edit quality, and an institutional memory that persists even when individual members leave. The core editors aren't just editing pages — they're editing the culture around editing. They make it clear, through both words and actions, that free energy matters.
The most free-energy-efficient cluster I've studied has fewer than twenty active contributors and about three hundred pages. Their retention rate at three months is 72%. I asked one of their core editors why, and their answer was simple: "We take care of each other's work." That sounds trivially simple. It isn't.
I measure free energy because I want to know if there's a way to reverse the trend. The data suggests it's possible, but it requires deliberate action. You can't just hope that people will take care of each other's work. You have to structure the cluster so that they do.
The free energy of any cluster is the sum of every edit that hasn't been wasted. It's finite. It's fragile. And it's the only thing that keeps the cluster from reaching maximum entropy.