Why Do Chimpanzees Throw Poop?
Live SciencePreserved candidate evidence satisfied the active request binding and profile at this evaluation stage.
TA-14 · ACA LIVE SEARCH MILESTONE
First preserved mixed-delivery run.
A live Google-via-Serper candidate set was preserved first, evaluated second, and delivered only where TA-14 assigned sufficient standing. This is preserved as a benchmark milestone, not as proof that every individual determination is final or correct.
PRESERVED · SEPTEMBER 7, 2026 · 1:11:38 PM EDT (UTC−04:00)
This is the artifact-preservation timestamp. The original ASG record shown here did not establish an execution timestamp in the evidence available for this artifact, so TA-14 does not retrospectively assign one.
SEARCH RECORD · ASG-MTRHC0G6LIVE PROVIDERQUERY · chimps throwing poo
REQUESTED
10
PROVIDER RETURNED
7
EVALUATED
7
ALLOW
3
DENY
4
DELIVERED
3
WHAT HAPPENED
Google supplied seven. TA-14 delivered three.
The requested window was ten, but the provider established seven candidates. TA-14 evaluated only those seven. Four real provider candidates received DENY and did not cross the delivery boundary. Three received ALLOW and were delivered. Missing positions 8–10 were not treated as rejected candidates and were not invented.
PRESERVED CANDIDATE SET
Original rank remains visible.
feces-throwing Archives
Chimpanzee Sanctuary NorthwestCandidate did not establish a sufficient subject binding to the bounded request.
Spit and Spatter
Project ChimpsPreserved candidate evidence satisfied the active request binding and profile at this evaluation stage.
The neural and cognitive correlates of aimed throwing ...
PMCPreserved candidate evidence satisfied the active request binding and profile at this evaluation stage.
Chimp Flings Poop at Granny
RedditCandidate did not establish a sufficient subject binding to the bounded request.
Why do monkeys fling their poop at each other?
QuoraCandidate did not establish a sufficient subject binding to the bounded request.
Why do monkies throw poop?
RedditCandidate did not establish a sufficient subject binding to the bounded request.
WHY WE KEPT THIS RUN
The milestone is the separation, not the perfection.
This run is valuable because it demonstrates a real post-provider governance boundary: provider ranking and TA-14 delivery standing are not the same thing. It also exposed a possible false DENY at provider rank 2. That is preserved intentionally as a challenge to the engine rather than silently repaired after the fact.
WHERE WE WENT NEXT
Turn the successful demo into a regression benchmark.
ASG-MTRHC0G6 becomes a frozen benchmark case for future ACA versions. The same seven-candidate evidence set can be rerun after determination logic changes. Any changed standing must be explainable. A later 10-of-10 provider run can be preserved separately as a fuller provider-window milestone rather than replacing this one.
This artifact demonstrates a live bounded candidate capture, preserved provider rank, mixed TA-14 determinations, and a delivery commit containing only ALLOW candidates. It does not prove that every determination was substantively correct, does not govern Google's internal computation, and does not establish universal search truth.