Cultivate

Next experiments

Make the next step something we can learn from.

The immediate aim is one complete product experiment: a need, competing explanations, a fair comparison and a decision supported by the result.

That gives us a foundation for testing organisations themselves. The programme below is proposed work; it does not report completed experiment results.

What we have to build on

Partial

A working delivery foundation

The recorded Slop Filter work includes issues, isolated worktrees, pull requests and automated checks. The 11 October inspection also recorded a running Gas City setup. Some installed-browser and live-provider checks remained incomplete.

This is a dated evidence record, rather than a live status feed.

Proposed

A product and organisation learning loop

The next task is to connect that delivery machinery to a product decision. Controlled comparisons of organisational arrangements and transferable improvements still need to be demonstrated.

How we would investigate this

The next proposed experiment

Can we make the feed more useful without losing worthwhile posts?

Compare the current filter with a small number of alternatives, such as personal rules or a different classification approach. Agree the acceptable trade-offs, evaluation sample and total budget before running it.

The output is a decision and its evidence. It may support a change, another round of research or keeping the current approach. No improvement threshold or result is claimed yet.

Inspect the proposed product milestone

A practical programme

Make the product comparison credible

Choose a feed problem, record the current baseline and agree the measures, sample and budget. Check that the evaluator can recognise known successes and failures.

The decision it informs: Is the comparison trustworthy enough to guide a product decision?

Complete one learning loop

Compare a small set of interventions, including the current approach. Link each hypothesis to its implementation, evidence, cost and final decision.

The decision it informs: What should we continue, change or stop—and why?

Compare organisational arrangements

Once evaluation works, compare solo, cooperative and independent approaches under equal total resource budgets. Repeat across tasks and runs.

The decision it informs: Does the arrangement improve outcomes enough to justify its total cost?

Preserve and reuse the learning

Share accepted findings from all approaches, including failed attempts. Record which later decisions use them.

The decision it informs: Can the next team avoid rediscovering what we already learned?

Improve the recipe and test its reach

Change one organisational feature, compare it with the earlier recipe, then try it on another product or problem.

The decision it informs: Does the improvement transfer beyond the case it was designed for?

For readers planning the work

The capability map shows the building blocks and their dependencies. An epic groups related work; a milestone describes a testable result. These details help plan implementation alongside the wider journey.

Engineering: capability details

The ability to plan, implement, review and ship software through issues, branches, worktrees, tests and recoverable orchestration.

E1

Coordinated coding agents

Working for Slop Filter and evidenced in Delegate instructions. The coordination record is strong; product decision quality is still mostly human-led.

  • Working Issue, branch and PR workflow is visible
  • Tested Worktrees are mandatory for concurrent work

E2

Persistent software factory

Working locally. This website repository is registered as its own rig. The full recovery and elasticity envelope is not yet audited.

  • Working Mayor can coordinate registered rigs
  • Partial Remote supervision path is established

E3

Verified continuous delivery

Partial. Slop Filter has strong automated checks, but PR #121 explicitly left installed signed-in Brave verification undone.

  • Partial Automated release verification is recorded
  • Next Browser verification is required before delivery claims

E4

Economic and self-improving factory

Proposed. The brief makes cost and resource allocation central, but this website does not claim the factory optimises itself today.

  • Proposed Cost and capacity become part of engineering decisions
Product: capability details

The ability to define outcomes, gather evidence, compare hypotheses, run experiments and decide when not to build.

P1

Outcome measurement and evaluation

Next. Slop Filter has correction records and synthetic cases, but the first protected product baseline still needs approval.

  • Next Outcome and baseline are written down
  • Proposed Evaluation boundaries are explicit

P2

Competing hypotheses and experiments

Proposed. The assignment defines the experiment but says it must not be run automatically as part of this website build.

  • Proposed Competing Slop Filter hypotheses are ready to test

P3

Autonomous product discovery

Research. The concept is defined, but no completed discovery loop has been verified.

  • Research Agents propose product work from evidence
Organisation: capability details

The ability to hold roles, delegated authority, budgets, review duties and governance boundaries across teams and institutions.

O1

Delegated roles and authority

Next. The boundaries are clear in the brief and repository instructions; the product charter has not yet been created.

  • Next Authority boundaries are explicit

O2

Autonomous product organisation

Proposed. The website documents the target organisation; it does not claim one is already operating.

  • Proposed Product organisation charter exists

O3

Competing organisations

Research. Slop Filter is useful for a small comparison but too small for a standing ecosystem.

  • Research Competing organisations can be compared