AI policy · October 2026

The Pill and the Dish

Governments are writing rules for artificial intelligence without knowing which way it will develop. This project looks for the policies that still work whichever future arrives.

The analysis sets AI aside. It follows two imaginary technologies with similar economics. One is an augmentation pill that people gradually allow to act on their behalf. The other is a lab-grown DishBrain that may someday have interests of its own. Each is traced stage by stage. At every stage the question is which policies would help, and whether those policies would still work at later stages. The results are then mapped back to AI.

Key findings

  1. Both futures cross the same threshold. Each reaches a point past which the causes of the technology's behavior can no longer be established from outside. The project calls this the visibility threshold.
  2. Two familiar tools fail at that threshold. One is labeling machine-made work. The other is assigning blame to whoever caused a harm. Both are useful today, and neither should anchor a long-term strategy.
  3. A core set of policies works in both futures. These policies attach to things that stay observable: physical inputs, named responsible parties, measurable effects, and the system as a whole.
  4. Timing is decisive. Nearly every policy that matters later is cheap to adopt early and very difficult to adopt late.
  5. Durability is not enough. A rule built to outlast the threshold will also outlast the officials who wrote it. Every proposal is also tested for what it costs, who bears that cost, and what it could do in the wrong hands.

Read

New to the argument? Start with the essay. The report holds the full analysis and every policy design.

Essay

The Visibility Threshold

About 4,600 words · 11 pages · PDF, 0.1 MB

The argument in one sitting. It explains why the visibility threshold should organize AI policy, why labeling and fault-finding break there, and which seven commitments survive.

Policy report

The Pill and the Dish: Robust AI Policy for Two Possible Futures

52 pages · PDF, 0.8 MB

The full analysis. It covers both scenarios stage by stage and proposes eighteen policies. Each one is tested for durability past the threshold, for cost, for who bears that cost, and for its potential for abuse.

Summary deck

The Pill and the Dish, in 19 slides

19 slides · PDF, 4.8 MB

A short visual introduction to the project, suited to a briefing or a first look.

How to cite

Fischer, V. (2026). The Pill and the Dish: Robust AI Policy for Two Possible Futures. Zenodo. https://doi.org/10.5281/zenodo.23247500

Fischer, V. (2026). The Visibility Threshold. Zenodo. https://doi.org/10.5281/zenodo.23247334

BibTeX
@misc{fischer2026pill,
  author    = {Fischer, Verlyn},
  title     = {The Pill and the Dish: Robust AI Policy
               for Two Possible Futures},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23247500},
  url       = {https://doi.org/10.5281/zenodo.23247500}
}

@misc{fischer2026visibility,
  author    = {Fischer, Verlyn},
  title     = {The Visibility Threshold},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23247334},
  url       = {https://doi.org/10.5281/zenodo.23247334}
}

All three documents are licensed under CC BY 4.0. You may share and adapt them with credit.