Atomically Precise Manufacturing aka Nanotechnology
If you are an expert on A.P.M., you will read these pages through a very different lens than most of my audience. Like superconductors, nanotechnology is treated as a magic buzzword without accurate definition by many people. Sadly, the word "nanotechnology" has been coopted and reused to describe thin sheets of bulk material, small grained randomly made particles, and all kinds of bulk chemistry/material researches that have little to do with Eric Drexler's original definition involving nanoscale structures made from precisely and intentionally positioned atoms. In some people's brains, nanotechnology has oxidized and gone rancid, and I refuse to treat Eric's great ideas that way, even inadvertently through my own misunderstandings of Atomically Precise Manufacturing ( his newer and hopefully less corruptable neologism ).
A.P.M. is making great progress, and will certainly show widely deployed results sooner than 2061, the 75th anniversary of Eric Drexler's "Engines of Creation". However, in 2012, we cannot accurately predict the course of progress of A.P.M., or which server-sky-relevant manufacturing processes it will impact first. Will it be computation? Solar cells? Radiation resistance? The problem is not the relevance of molecular scale manufacturing, but in which directions it will expand the "possibility space" first, or more accurately, which directions will be hindered by stumbling blocks and lack of application-specific robustness. While Eric has a much better idea about near-term directions, and may write about it here someday, I am not competent to make such predictions.
In the long term, Server Sky will be entirely transformed by A.P.M., and we should leave "hooks in the design" for rapid insertion of these technologies (or any others). But we can't wait for A.P.M. to appear - Server Sky can be deployed now, and modified later. We must keep the possibility of "hostile goo" in mind - older thinsat designs still in orbit may later be vulnerable to molecular scale modification or destruction. We should budget for on-orbit security upgrades that protect legacy thinsats from such modification, and develop and deploy these upgrades well in advance of need.
Take the musings on this site as an approximation. If you are an expert on A.P.M., re-vector my musings in the directions your expertise may take them. If you know little about nano-scale bulk materials or A.P.M., assume that the possibility space around Server Sky is huge and undefined. The spot I describe here is in that possibility space, but the paths actually taken will be more feasible, cost effective, and environmentally benign, driven by the specifics we discover while moving forwards. Take the optimism and enormous benefits described here and multiply them, perhaps beyond my understanding or yours.
Atomically Precise Manufacturing is by no means the only potential technology that will change the eventual implementation of Server Sky and space energy. Advances in laser efficiency and steerability could replace radio transmission with light. Economic improvements in rural China, India, or Africa could radically alter applications and provide millions of developers. And war could change everything. If you have expertise in any of these areas, and can describe their effects in quantifiable design-applicable ways, please write about them.