TL;DR
Thorsten Meyer AI has published Outcome-First Decisions, a decision-support framework for reviewing initiatives by current outcomes and ongoing cost. The source says the AGPL-3.0 project is open source on GitHub and is part of a Built in Public series.
Thorsten Meyer AI has published Outcome-First Decisions: Keep, Change, or Kill, an open-source decision-support framework meant to help operators decide whether ongoing initiatives still justify their cost, a problem the project frames as central to managing a multi-product portfolio.
The framework centers on what it calls the Worth Filter: whether an initiative’s current or expected outcome is worth the cost of continuing it. According to the source material, the filter deliberately excludes sunk cost, past effort and identity-based attachment from the decision.
Outcome-First Decisions returns one of three verdicts. Keep means the outcome justifies continued investment. Change means the underlying idea may still be worth pursuing, but its current form is not working. Kill means the outcome does not justify the ongoing cost and the work should be ended cleanly.
The source says the project is open source under the AGPL-3.0 license and available on GitHub. It is described as local-first and provider-agnostic, with reviews intended to run on owned compute and without being tied to a single AI model provider.
Outcome-First Decisions — keep, change, or kill
The hardest decision isn’t what to start — it’s what to stop. Judge every initiative by the outcome it produces now, not the effort already spent.
Independent commentary, produced with AI assistance under human editorial oversight. The views are the author’s own and may change. Outcome-First Decisions is open source under AGPL-3.0, provided “as is” without warranty; see the repository LICENSE. The framework’s verdicts are reasoning aids based on the inputs given and may be wrong — decision support, not decisions; verify independently before acting. Product and company names are trademarks of their respective owners; mention does not imply endorsement.
Stopping Work Becomes Explicit
The release matters because many operators, teams and founders keep marginal projects alive long after their value has faded. The framework gives that review a defined vocabulary and makes ending work a normal verdict rather than an implied failure.
For readers managing several products, experiments or internal projects, the main issue is capacity. The source argues that dead or low-return work consumes attention, upkeep and capital even when it does not appear as a clear budget line. By forcing a keep, change or kill decision, the framework aims to make that hidden cost visible.
Corporate Divestiture Management: Organizational Techniques for Proactive Divestiture Decision-Making
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Part Of A Decision Layer
Outcome-First Decisions is presented as Day 8 in Thorsten Meyer AI’s 19-part Built in Public series. The source places it inside a broader operator portfolio that includes 18 products and a shared decision layer.
The dispatch says the decision layer now follows a sequence of validate, plan and review. In that structure, Outcome-First Decisions is the review step: it checks whether an initiative still earns its place after earlier validation and planning work.
The project also fits the author’s stated approach of local-first, provider-agnostic tooling and non-developer builds. The source describes the product as a small, opinionated framework rather than a full enterprise portfolio system.
“The hardest decision isn’t what to start. It’s what to stop.”
— Thorsten Meyer AI dispatch
outcome-based initiative review tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Adoption Details Are Limited
Several details are not confirmed in the supplied material. The source does not provide a calendar launch date, repository URL, release version, installation instructions, user numbers or examples of teams using the framework outside the author’s own portfolio.
It is also unclear how the framework scores costs and outcomes in practice, whether it uses structured metrics, free-form reasoning or a mix of both. The source says its verdicts are reasoning aids based on the inputs provided and may be wrong, meaning users still need to verify decisions independently before acting.
Practical Guide to Applied Conformal Prediction in Python: Learn and apply the best uncertainty frameworks to your industry applications
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Repository And Use Cases Await
The next step for readers is to inspect the GitHub repository referenced by the source, review the AGPL-3.0 license and test the framework against real initiatives. Operators considering it should define inputs carefully, including ongoing cost, current traction, expected future return and the cost of keeping the project alive.
Further updates in the Built in Public series may show how Outcome-First Decisions connects with the author’s other portfolio tools and whether the keep, change or kill review becomes part of a larger workflow.
AI-powered project evaluation tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Outcome-First Decisions?
It is a decision-support framework from Thorsten Meyer AI for reviewing initiatives and assigning one of three verdicts: keep, change or kill.
What is the Worth Filter?
The Worth Filter asks whether the outcome an initiative is producing, or is likely to produce next, is worth the ongoing cost of continuing it.
Is the framework open source?
According to the source material, Outcome-First Decisions is open source under the AGPL-3.0 license and is on GitHub.
Does the framework make final business decisions?
No. The source describes its verdicts as reasoning aids and says users should verify independently before acting.
Who is the intended audience?
The source frames the tool for operators managing multiple products, projects or commitments, especially where old work is consuming capacity without clear returns.
Source: Thorsten Meyer AI
