The rubric is the new prompt
The most important change in the AI tooling landscape in the first half of 2026 is not a model. It is a shift in what the tools accept as input: not just a task, but a definition of done — and a harness that iterates against it without you.
Three examples, all shipped between April and June 2026:
- Anthropic Managed Agents — Outcomes. You send a
define_outcomeevent with a task description and a rubric. A separate grader model (its own context, not the agent’s) scores each iteration against the rubric and feeds per-criterion gaps back. The loop runs iterate → grade → revise until satisfied, capped at a max iteration count. - Claude Code —
/goal. A fresh evaluator model checks a completion condition after every turn and keeps the agent working — potentially for days — until the condition holds. - Claude Fable 5 (June 9, 2026 — the first generally available Mythos-class model, a tier above Opus). Anthropic’s own migration guidance is explicit: give the full task specification up front in a single well-specified turn and run at high effort. The model’s headline capability is long-horizon autonomy against a stated goal.
Grok Build runs the same play from the other side: its plan mode produces a plan you approve, comment on, or rewrite before execution, and post-approval deviations surface as reviewable diffs. Different mechanism, same contract — the human’s contribution moves from steering turns to specifying success.
Why this favors the meta-thinker
This site has argued that the meta-thinker’s skill — describe the problem well enough that someone good could solve it — becomes the entire job. That was a working style. As of mid-2026 it is a literal interface: the spec is no longer something you hold in your head while you steer; it is the artifact the harness consumes.
Which means rubric quality is now load-bearing. The graders are competent but literal. They verify what you stated, nothing else:
- Checkable beats aspirational. “The output CSV has a numeric
pricecolumn for every SKU” is gradeable. “The data looks good” is not — a vague rubric produces a confident loop converging on the wrong thing, with the grader cheerfully reportingsatisfied. - Independent criteria beat bundled ones. Graders score per criterion. Five separate checkable statements give the agent five separate gaps to close; one paragraph gives it a vibe.
- The rubric can’t see what you didn’t mount. It grades the artifact in the session, not your live schema or last quarter’s column semantics. The checks that need the real world stay outside the loop.
What stays human
Two things, and they are the same two things as before.
First, the rubric itself. Writing one forces the question every verification habit on this site comes back to: what must be true for this to be safe? If you can’t write the rubric, you weren’t ready to delegate the task — to a model or to anyone.
Second, the acceptance. A grader’s satisfied is a claim like any other model output. Before the change lands in the permanent record — the DAG, the registry, the paper — you run the actual command, read the actual table, check the actual decision log. The loop got cheaper and longer; the last step didn’t move.
Put the rubric in the registry next to the decision it produced. Six months from now, “what did we ask the agent to achieve, exactly?” will be the most useful provenance question you can answer.