# Assay Bench

Source: https://dinand.com/experiments/assay-bench/
Author: Dinand Tinholt

> A workbench for measuring the quality and checking cost of AI outputs when there is no simple answer key.

The question behind the system

> The cost that constrains an AI process may not be generation. It may be the cost of becoming confident enough to trust the output.

What it is designed to do

- Report quality with uncertainty and explicit blind spots
- Plan samples against both confidence and checking cost
- Keep judged cases, methods and human effort inspectable

Study domain: Synthetic contract exceptions

Product: Next.js + SQLite

Companion: Separate experimental rig

The public boundary

## What this page does—and does not—open.

The application sits beside a process rather than controlling it. This public profile includes no seed database, credentials, model endpoints, client material or live review workflow.

Separate application

No public deployment is linked yet. The introduction remains available without exposing a private service or local data.

[Read the monitoring study ](https://dinand.com/research/when-monitoring-stops-scaling/)

Continue the thread

## A next room to enter.

[Choose another path ](https://dinand.com/)

[System profile · Paper-only research prototype

### Personal Quant Agent

A research system that scans public market data, tests a small strategy library, applies explicit risk controls and records simulated decisions in a paper portfolio.

](https://dinand.com/experiments/personal-quant-agent/)[System profile · Working research product

### Pakhuis

A seventeenth-century Amsterdam canal warehouse, built in simulation, where two robots count stock and a reviewer decides what reaches the record.

](https://dinand.com/experiments/pakhuis/)[Experiments · Interactive tool

### Monitoring cost explorer

Explore how portfolio size changes modeled oversight cost.

](https://dinand.com/experiments/monitoring-cost-explorer/)

[All experiments ](https://dinand.com/experiments/)[Back home ](https://dinand.com/)

---
Site guide for AI agents: https://dinand.com/llms.txt