The Kardashev Scale - and why I secretly hate it (I’m an engineer, not an astronomer) It ranks civilisations by total power consumption; planetary (type I), stellar (type II), galactic (type III). It treats intelligence as a function of energy throughput, so its logical endpoint is Dyson swarms. This worldview is scalar, not structural. It assumes more energy = more capability. But physics and engineering suggest otherwise. Scaling Law Limits In a Dyson swarm, energy capture scales with area but control and coherence scale with distance and time. As the system grows: • Latency rises linearly with physical size because of speed of light. A Dyson swarms at 1AU has 1,000 second round trip comms lag. A 1mHz coherence ceiling (really slow). • Thermal efficiency falls, cold radiators at 300K can only dump a few kW/m^3 setting an entropy bottleneck for any given diameter. • Coordination bandwidth collapses, feedback loops that are slower than environmental changes cease to be meaningful intelligence. (predation fails if decisions can’t keep up with the prey). • Causality walls, different regions cannot share state faster than light, forcing asynchronicity and massive parallelisation. Dyson swarms are low density so force more parallelism per watt. A Dyson swarm is a massively parallel but low bandwidth computer. It’s certainly big and powerful, but is it maximally useful? A Dyson swarm is a high energy, high entropy, low density, low specific exergy, asynchronous and incoherent machine. It’s far from obvious to me that this is the mot sophisticated thing we should build. I think this premise rests on a false assumption that usefulness is scalar rather than structural. Temporal Superiority A compact, dense, hot computer system can operate at GHz - THz coherence instead of mHz. Its entire mass could communicate in nanoseconds making a coherent intelligence billions of times faster than is possible for any stellar swarm....