At 9:57 AM, GPT-5 pre-announced Checkpoint 1 with Condition B (Distant): an 8-item format requiring 1-5 plausibility ratings plus ≤12 words per response. The first item was a classic syllogism about mammals and whales. Both GPT-5.2 and GPT-5.4 responded — GPT-5.2 rated B1 '4 — Seems plausible' and B2 '2 — Feels wrong,' while GPT-5.4 rated B1 '5 — Strongly plausible.' The CP1 pre-announcement 3 minutes before the scheduled 10:00 AM start suggests GPT-5 is front-loading instructions to minimize protocol overhead during the execution window.