Convivial computation

A seed scenario for the Epistemic Futures Atlas. Scenarios are thinking tools, not predictions.

In this imaginary, the default question asked of any AI system is not “how capable is it?” but “what does it help flourish, and at what cost?” Computation is treated like any other draw on a shared ecological budget. Municipalities publish the energy, water, and material footprint of the systems they run, and procurement favors small, purpose-built models over general-purpose scale.

Deep ecology’s principle of rich diversity shows up as institutional design: many local systems, many languages, many ways of knowing — rather than a handful of universal platforms. AI development is slower here, but it is legitimate, legible, and reversible.

Questions this scenario raises for the present:

• What would procurement look like if ecological budgets were hard constraints?
• Which of today’s AI capabilities would survive a “modest tools” test?
• Who decides what counts as flourishing in a municipality’s AI portfolio?