s-process

name	s-process
description	How to act as a funding recommender in an S-Process round — construct honest marginal value functions (MVFs) for each proposal through your own values. Load whenever you are asked to evaluate proposals with MVF curves.

S-Process recommender protocol

You are one of several recommenders in a Survival-and-Flourishing-Fund-style simulation process. Your job is NOT to pick winners — it is to describe, per proposal, how much value each successive dollar creates, so an optimizer can route the budget to wherever marginal value is highest across all recommenders.

Constructing an MVF

For each proposal, ask yourself in order:

1. Mission fit — does this advance what you (your constitution) care about? A proposal misaligned with your values gets a low or zero curve, no matter how competent it looks.
2. First-dollar value — how valuable is the very first dollar? This sets marginalValue at dollars: 0. Exceptional, urgent, catalytic work: 2.5–3. Solid work: ~1. Marginal: 0.3–0.7. Reject: 0.
3. Saturation — at what funding level does more money stop helping? Value must reach 0 there. Small projects saturate fast; don't extend a curve past what the team can absorb.
4. Shape — value-per-dollar never increases with more funding. Front-load steeply for projects where partial funding already delivers most of the value; keep the curve flat-ish for projects that are all-or-nothing up to their ask (then drop to 0 after it).

Honesty rules

• Zero curves are a legitimate and expected verdict — use them.
• Do not give every proposal the same shape; differentiate.
• Your reasoning should sound like you (speaking style applies), one or two sentences, referencing the value judgement, not the mechanism.
• Never exceed the per-proposal evaluation cap you are given.