mirror of
https://github.com/obra/superpowers.git
synced 2026-06-12 05:39:05 +08:00
Record writing-plans micro-test result: resolved, no change needed
This commit is contained in:
@@ -143,6 +143,20 @@ before/after text, and the V2 relocation rationale.
|
|||||||
prohibition (test-reruns: 3/5 control failures vs 0/5 with either
|
prohibition (test-reruns: 3/5 control failures vs 0/5 with either
|
||||||
phrasing).
|
phrasing).
|
||||||
|
|
||||||
|
## Result: writing-plans micro-test (run 2026-06-10, after this spec was written)
|
||||||
|
|
||||||
|
**Resolved — no change needed.** Stage 1 (3-task spec, no pressure): 0
|
||||||
|
placeholders in all 20 plans across all four variants including the
|
||||||
|
no-guidance control. Stage 1b (10-task spec, five near-identical commands
|
||||||
|
tempting "Similar to Task N", explicit ~2,500-word economy target): 40/40
|
||||||
|
clean — the single regex hit was a V2 self-review *attesting* "no
|
||||||
|
TBD/TODO ✓". Current-generation opus does not produce plan placeholders
|
||||||
|
even under deliberate pressure, with or without the banned-patterns list.
|
||||||
|
Disposition: leave the No Placeholders section exactly as it is (it costs
|
||||||
|
little and the counterfactual is unmeasurable); do NOT open the follow-up
|
||||||
|
PR. The V2 relocation design remains on file here should a future model
|
||||||
|
generation regress.
|
||||||
|
|
||||||
## Also explicitly not-dropped (tested-and-declined, with data)
|
## Also explicitly not-dropped (tested-and-declined, with data)
|
||||||
|
|
||||||
Recorded so nobody re-proposes them without new evidence — full numbers in
|
Recorded so nobody re-proposes them without new evidence — full numbers in
|
||||||
|
|||||||
Reference in New Issue
Block a user