---
title: 'Outcome maxing'
description: 'An emerging label for optimizing an AI workflow around accepted, durable results instead of easy-to-count activity.'
canonical_url: 'https://darkfactory.dev/glossary/outcome-maxing'
markdown_url: 'https://darkfactory.dev/glossary/outcome-maxing.md'
collection: glossary
date_published: '2026-08-05T00:00:00-04:00'
date_modified: '2026-08-26T00:00:00-04:00'
---

# Outcome maxing


## Definition

**Outcome maxing** is an emerging label for optimizing an AI workflow around accepted, durable results instead of activity proxies such as token volume, prompts, model spend, pull requests, or generated lines. Dark Factory Dev uses the term as a direct counterweight to token-maxing: more inference is justified only when it improves the result that matters.

The word **outcome** must be operationalized before it can guide a system. Name the unit, acceptance evidence, durability window, risk constraints, and full cost boundary. Otherwise a workflow can win by producing more low-risk tasks, weakening the acceptance test, moving failures downstream, or spending scarce review attention that the dashboard does not count.

## Where the phrase sits

The term follows the playful **maxing** or **maxxing** construction used in token-maxing and value-maxxing discussions. The available record does not establish a single coiner.

## Why it is contested

Outcome maxing has no settled technical definition, and the word **outcome** can hide incompatible targets. One system may count completed tasks, another accepted durable work, and another business value after cost and risk. Each choice creates different incentives and can be gamed by weakening acceptance or moving damage outside the measurement window. This glossary uses outcome maxing as a policy direction and requires the outcome measure to be stated. Cost per accepted durable outcome is one possible implementation.

## Distinguish it from nearby terms

Value-maxxing asks about broader economic value. Token efficiency asks how much useful work is obtained from model consumption. Outcome maxing says what the workflow should optimize. Specification gaming and reward hacking describe ways the chosen measure can be satisfied without delivering the intended result.

## Check your understanding

An agent team doubles merged pull requests by selecting smaller tickets and leaving integration defects for another team. Did it improve outcomes? Not under a measure that includes task mix, acceptance, downstream incidents, durability, and human correction cost. It improved a count.

## Also called

outcome maxxing, outcome-maxxing

## Related terms

- [Token maxing](https://darkfactory.dev/glossary/token-maxing)
- [Token minning](https://darkfactory.dev/glossary/token-minning)
- [Token efficiency](https://darkfactory.dev/glossary/token-efficiency)
- [Cost per accepted durable outcome](https://darkfactory.dev/glossary/cost-per-accepted-durable-outcome)
- [Acceptance criteria](https://darkfactory.dev/glossary/acceptance-criteria)
- [Production truth](https://darkfactory.dev/glossary/production-truth)
- [Verification gate](https://darkfactory.dev/glossary/verification-gate)

## Related factory areas

- [Verification, evaluation & quality truth](https://darkfactory.dev/factory/verification)
- [Economics, capacity & factory FinOps](https://darkfactory.dev/factory/economics-finops)

## Evidence and further reading

- [Workplaces look for cheaper AI as tokenmaxxing fades as a corporate fad](https://apnews.com/article/31bb80ac1cd7862d05f6397177d826b1)
- [Value-Maxxing and the New Economics of AI Labor](https://economy.ac/research/2026/05/202605289132)
- [Stop 'tokenmaxxing' and deploy AI sensibly instead](https://doi.org/10.1038/s42256-026-01253-5)
- [The Harness Effect: How Orchestration Design Sets the Token Economics of Enterprise Agentic AI](https://arxiv.org/abs/2607.06906)