Walkthrough demo
The same panel from two directions: what a panelist sees and what a facilitator does. Every number on this page comes from an example panel that Pronoia ran itself — none of them are written by hand.
What this panel produced
One thesis rated on two scales. The distance between them is what a survey measuring agreement never produces at all.
The same thesis before the dialogue and after it. The panel read each other’s reasoning, and the distribution moved. 6 / 23 revised.
One revision in the panelist’s own words. They refer to another panelist by code — that is what a round is for.
🤖 AI-01 · 5 → 4I am moving the other way, and I expect to be alone in it. R4 conceding the licensing filter did not surprise me, but his write-off examples did. I have watched well-funded traits fail in a dry August too. Ownership sets the terms I farm under. It does not decide whether the seed works, and 2050 will still be decided in the field.
Every panelist on this page is simulated. No human answered these questions. The panel design, the theses and the roles are real research artefacts; the answers are machine-generated, and convergence here is not consensus.
Simulated Delphi (Argument Simulator v2.5) · bootstrap_food2050.json (planning-to-xdelphi.v1.0) · SIM_ArgumentSim_Food2050_20260819.md
The 1–7 position is derived from the simulator's probability estimate by a single rule, which lives in the code as one function and is checked against every row. likert = max(1, min(7, round(1 + 6 * P / 100)))
The global food system is at the threshold of a paradigm shift whose direction is contested. Multiple paradigm-carriers — states, multilateral institutions (FAO, EAT-Lancet, IPES-Food), agro-industrial corporations, cellular-agriculture pioneers, indigenous food-sovereignty movements, and climate-adaptation alliances — have published, funded, and operationalized competing visions for the global food system by 2050. Five candidate paradigm-regimes have crystallized as coherent enough to be taken seriously: (1) Techno-capitalist, (2) Distributed bio-economic, (3) Fragmentation, (4) Multipath cultural-pluralism, and (5) Post-smallholder calorie-system. The dominant paradigm will be effectively decided in the 2026–2035 decade, not announced in 2050.
0 human · 23 AI · 39 revised · The example study is in English, so the theses, the reasoning and the comments appear in English. The page's own text follows the language you chose.
The invitation arrives
A panelist gets a personal link by email. No account, no password. The link first says what the study is about and who is taking part — the theses come after that.
The same thesis is rated twice. Probability and desirability are not the same question — and the distance between them is where the alternative futures come from.
Want to try it on your own topic?
The free Student tier covers one study at a time. If you would rather have the first panel run with you, ask about a pilot.