Information Thermodynamics
The writing on this site builds on a specific framework: treating connected human systems as thermodynamic networks where capability-energy diffuses through graph edges, gradients represent inequality, and the system’s trajectory depends on the topology of connections.
Nodes as Negentropy Engines
Schrödinger defined life as the ability to locally reverse entropy by exporting disorder to the environment. Every human is a negentropy engine — taking in energy, maintaining internal order, and producing surplus structure that propagates through the network. A farmer grows more food than they eat. A teacher transmits more knowledge than they consumed learning it. An engineer builds tools that amplify the negentropic capacity of everyone who uses them.
The species runs roughly 8 billion of these engines. Most of them are throttled — consuming their entire energy budget on biological self-maintenance, producing no surplus order beyond their own bodies. The configuration problem is not resources. It is activation.
Edges as Diffusion Paths
The connections between nodes — communication channels, trade routes, shared languages, tool access — determine how capability-energy flows through the system. Increase the conductivity of the network and diffusion accelerates. Sever edges and gradients persist. Every communication technology in history — writing, printing, telegraph, internet — was an edge-conductivity upgrade that produced a spike in collective negentropic output.
Superlinear Scaling
Geoffrey West and the Santa Fe Institute showed that when humans aggregate into denser networks, creative and economic output scales superlinearly — double the connected population and you get more than double the innovation. The graph doesn’t just sum its nodes. It multiplies them.
But only when the nodes are active. A network of a million where most are in survival mode doesn’t get the scaling bonus. You need the nodes online.
Attractors and Absorbing States
Complex systems have basins of attraction — configurations they tend to fall into and stay in. Some basins are productive: self-sustaining diffusion, expanding capability, growing the container. Some are traps: absorbing states where the dynamics that could escape the basin have been suppressed below the activation threshold needed to generate escape velocity.
The shape of the graph determines which basins exist and how deep they are. Change the topology and you change the landscape of possible futures.
Layer 0
Information wants to diffuse, but atoms don’t. You can’t download a chip fab. You can’t torrent a lithography machine. You can’t open-source a gigawatt of electricity.
Control at the physical substrate layer — hardware, energy, manufacturing capacity — has fundamentally different diffusion properties than control at the information layer. This is where the optimistic thermodynamic argument (diffusion is downhill, insulation is self-undermining) meets its critical caveat: what if the insulation operates at a layer where the leaky-insulation argument doesn’t apply?
The answer to that question may determine more about the next century than any policy debate or technological breakthrough.