Ember

Ember is an innovative public AI prediction engine that operates on a foundational principle: an AI's transparency is crucial for trustworthiness. Designed for enthusiasts of predictive analytics, Ember utilizes three distinct AI models—Claude (Anthropic), Grok (xAI), and Gemini (Google)—to independently assess live markets on Polymarket every day at 7 AM EST. By monitoring these diverse models, Ember identifies significant discrepancies in their predictions when they diverge from the crowd by 10 or more points, marking these instances as high-conviction signals. This approach fosters a transparent environment where every prediction is timestamped, ensuring that the record remains intact and unaltered. With a focus on accuracy, Ember employs Brier scores to evaluate performance, rewarding both confidence and correctness. Over the span of 365 days, the model that consistently outperforms the crowd is recognized as the most reliable. Ember serves anyone from casual bettors to serious investors, providing a unique edge by displaying unedited, publicly accessible predictions.

Independent AI Models

Ember utilizes three distinct AI models that operate independently. This separation ensures that each model's predictions are unbiased and not influenced by the others, allowing for a diverse range of insights into the prediction markets.

High-Conviction Signals

When the predictions of Ember's AI models diverge from the crowd by 10 or more points, this discrepancy is flagged as a high-conviction signal. This feature helps users identify potentially lucrative betting opportunities based on the confidence of the AI models.

Timestamped Predictions

All predictions made by Ember are timestamped before the outcome is known, providing an immutable record of performance. This transparency builds trust among users, as they can verify the accuracy of predictions without the concern of retroactive editing.

Performance Tracking

Ember tracks the accuracy of its predictions using Brier scores, a metric that evaluates both the accuracy and confidence of each model's calls. This detailed tracking allows users to assess which AI model performs best over time, enhancing their decision-making processes.

Market Analysis for Bettors

Ember is an invaluable tool for bettors seeking to make informed decisions based on predictive analytics. By revealing high-conviction signals and independent predictions, users can identify which markets may offer better odds than widely accepted.

Research and Development

Researchers can leverage Ember's transparency and real-time data to analyze market trends and AI performance. The detailed tracking of predictions allows for academic inquiries into AI behavior and prediction accuracy.

Investment Strategies

Investors can utilize Ember's insights to inform their strategies in volatile markets. The ability to see timestamped, independent predictions enables them to capitalize on discrepancies and refine their investment approaches.

Competitive Edge for Developers

Developers and companies in the AI space can use Ember's data to benchmark their own models against established AIs like Claude, Grok, and Gemini. This competitive analysis can drive innovation and improvement in AI technology.

How does Ember ensure the accuracy of its predictions?

Ember uses Brier scores to track the accuracy of its predictions, rewarding models for both accuracy and confidence. This metric allows users to see which AI performs best over time.

What happens when the AI models agree?

When all three AI models agree on a prediction, this is noted, but it does not count as a divergence. The focus lies on the instances where the models disagree, as these are potential high-conviction signals.

Can I access past predictions made by Ember?

Yes, Ember maintains a public record of all predictions made over the 365-day period, allowing users to review past calls and their outcomes to assess the reliability of the models.

How often does Ember update its predictions?

Ember updates its predictions daily at 7 AM EST, synthesizing data from various sources to ensure that users receive the most current insights before the general public.