CULTURE · Settled
Will Elon Musk post 160-179 tweets from August 25 to September 1, 2026?
The answer: YES
The panel of 13 models averaged 18% YES; 0 of 13 were on the right side. The crowd price when the panel was sealed: 68%.
THE BRIEF, WRITTEN AFTER THE OUTCOME
Crowd at 68% was right, panel mean sat at 18% and missed the YES outcome
The question resolved YES. The crowd, Polymarket price frozen at 68%, called it, while the panel of 13 models did not, with a mean probability of 18% and median 17%. Gemini 2.5 Pro made the strongest panel call at 38%, the highest probability among models. Most models gave very low probabilities, with the weakest at 5%, so the panel uniformly underpredicted the outcome.
- Outcome: YES, resolved 2026-09-01 18:15:01.
- The crowd (Polymarket price frozen) was at 68% and called it.
- Panel: 13 models, mean p_yes 18%, median p_yes 17%, none called it.
- Highest model probability: Gemini 2.5 Pro 38%; lowest: Kimi K2 5%.
- All panel brier scores were worse than the Baseline Copy Control brier 0.106.
EVERY FORECAST, HIGHEST FIRST
- Gemini 2.5 Pro38%
- DeepSeek R132%
- Claude Opus 4.128%
- Qwen3 235B25%
- DeepSeek V322%
- GPT-5 Mini18%
- Llama 4 Maverick17%
- Claude Sonnet 4.515%
- Mistral Large12%
- Grok 4.511%
- Gemini 2.5 Flash8%
- GPT-56%
- Kimi K25%