CULTURE · Settled
Will Elon Musk post 260-279 tweets from August 28 to September 4, 2026?
The answer: NO
The panel of 13 models averaged 10% YES; 13 of 13 were on the right side. The crowd price when the panel was sealed: 6%.
THE BRIEF, WRITTEN AFTER THE OUTCOME
Panel and crowd expected NO; panel mean 10% vs crowd 6%, Gemini 2.5 Flash best
The outcome was NO. The crowd had p_yes 0.06 and called it correctly; the 13-model panel all called NO with mean_p_yes 0.1 and median_p_yes 0.08. Gemini 2.5 Flash made the best call, assigning p_yes 0.03 and achieving the lowest Brier 0.001. Misses came from higher-probability forecasts, notably Gemini 2.5 Pro at p_yes 0.24 which produced the highest Brier 0.058.
- Outcome: NO, resolved 2026-09-04 18:15:02.
- Crowd p_yes 0.06, crowd called it correctly.
- Panel: 13 models, mean_p_yes 0.1, median_p_yes 0.08, range 0.03 to 0.24.
- Best Brier 0.001 by Gemini 2.5 Flash at p_yes 0.03.
- Worst Brier 0.058 by Gemini 2.5 Pro at p_yes 0.24.
EVERY FORECAST, HIGHEST FIRST
- Gemini 2.5 Pro24%
- Qwen3 235B18%
- Claude Opus 4.112%
- DeepSeek V312%
- Mistral Large12%
- Claude Sonnet 4.58%
- DeepSeek R18%
- Kimi K28%
- Llama 4 Maverick7%
- GPT-5 Mini6%
- Grok 4.56%
- GPT-54%
- Gemini 2.5 Flash3%