Thorsten Meyer AI has published Outcome-First Decisions, an open-source AGPL-3.0 framework that asks operators to judge each initiative by current outcomes and ongoing cost, then place it in one of three categories: keep, change, or kill. The release matters because it formalizes a stopping process for portfolios where projects can continue after they have stopped earning attention or funding.

The supplied Thorsten Meyer AI material describes the project as part of Built in Public Day 8 of 19 and says it is available on GitHub under AGPL-3.0. The framework centers on a Worth Filter, which weighs a forward-looking question: whether the outcome now expected from an initiative is worth its ongoing cost.

The framework returns three verdicts. Keep applies when the outcome still justifies the cost. Change applies when the underlying opportunity may be sound but the current form is not working. Kill applies when continued upkeep no longer matches the outcome, with capacity moved elsewhere.

The source material frames the tool as decision support, not an automated decision-maker. It says the verdicts are based on the inputs provided, may be wrong, and should be checked independently before action is taken.

Stopping Becomes A Portfolio Process

The release addresses a recurring management problem: initiatives can survive because teams remember the money, time, identity, and effort already invested. Thorsten Meyer AI argues that those backward-looking factors should not decide whether the next period of work is worth funding.

For readers who manage products, content operations, experiments, or AI-assisted portfolios, the practical impact is a repeatable review method. A tool that makes endings speakable can free attention, maintenance budget, and planning capacity for work that is still producing results.

Day Eight In The Build Series

The material identifies Outcome-First Decisions as Day 8 of a 19-part Built in Public series on ThorstenMeyerAI.com. It places the project inside an operator portfolio described as 18 products sharing a local-first, provider-agnostic foundation.

Within that series, the source says the decision layer now runs from validate to plan to review. Outcome-First Decisions is the review piece: it closes the loop by asking whether existing work still earns its place instead of only screening new ideas.

The source also discloses that the commentary was produced with AI assistance under human editorial oversight and that the views are the author’s own and may change.

Adoption And Scope Still Unclear

The source material confirms the license, GitHub availability, verdict structure, and the author’s intended use inside the operator portfolio. It does not provide a repository URL, release tag, number of users, example input schema, or evidence that outside teams have adopted the framework.

It is also not yet clear how much of the process is manual, how verdicts are generated in practice, or how conflicts are handled when data quality is weak. The source says the framework’s outputs can be wrong and should be verified before acting.

Repository Details And User Tests

The next items to watch are the public GitHub repository, setup instructions, worked examples, and later Built in Public entries that show the framework operating across real portfolio decisions. Those materials would clarify how the Worth Filter is applied, what data it needs, and how a kill verdict is handled after review.

Key Questions

What is Outcome-First Decisions?

It is a decision framework from Thorsten Meyer AI for reviewing active initiatives by outcome and ongoing cost. It produces one of three verdicts: keep, change, or kill.

What is the Worth Filter?

The Worth Filter is the framework’s central test. It asks whether the outcome an initiative is producing now, or is expected to produce next, is worth the cost of continuing it.

Does the framework make final decisions?

No. The source says the verdicts are reasoning aids based on inputs and may be wrong. It tells users to verify independently before acting.

Why does the kill verdict matter?

The source argues that dead or weak projects can consume attention, upkeep, and capital even when the cost is hard to see. A kill verdict gives teams a named way to end work that no longer pays for its place.

Is Outcome-First Decisions open source?

Yes. The supplied material says Outcome-First Decisions is open source under AGPL-3.0 and available on GitHub, though it does not include the repository URL.

Source: Thorsten Meyer AI