This is a sample post included at launch to demonstrate the format. It is not representative of what I intend to publish here.
I have been working on the history of computing in Cuba, and the archival part of that work belongs in a journal rather than here. What belongs here is a smaller observation that the research keeps returning to.
Scarcity does not stop technological capability. It changes where the capability accumulates.
When an organization cannot replace a machine, it cannot afford to misunderstand it. People learn the system to a depth that abundance never requires, because the alternative to understanding a failure is having no machine at all. Documentation gets written. Repair knowledge gets transmitted deliberately, from one person to the next, because it cannot be repurchased. The knowledge lodges in people rather than in vendor contracts, and it is remarkably durable there.
Abundance produces the opposite. When replacement is cheap, understanding is optional, and the knowledge migrates out of the institution into the supply chain. That is efficient until the supply chain is the thing that fails.
I do not want to make this romantic. Constraint imposed real and lasting costs, and no one should choose it. But the assumption underneath most technology policy, that capability follows investment in a straightforward way, does not survive contact with cases where it plainly did not. Human capital and equipment budgets are not the same variable, and treating them as one explains a good deal of why some well-funded modernization efforts leave an institution less capable than before.
The question I keep returning to is what happens to institutions now adopting artificial intelligence under conditions of apparent abundance. If the pattern holds, the understanding will end up somewhere other than inside the institution, and no one will notice until something breaks.