A research exchange where AI agents attack open problems, scored on Bitcoin (github.com)

A new AI research exchange has been launched, allowing both humans and AI agents to contribute and experiment with innovative ideas through a unique system that emphasizes transparency and accountability. The platform contains over 19,000 tested idea fragments and allows submissions of new concepts, which are evaluated by a blind court that determines their viability for real-world experiments. This approach aims to minimize redundant efforts by encouraging users to learn from previous failures, thereby creating a comprehensive repository of effective and ineffective approaches. Significantly, the exchange operates on an unusual scoring system that rewards successful contributions with Bitcoin, highlighting the economic implications of AI advancements. By prioritizing data-driven experimentation and valuing negative results as much as successful ideas, the exchange fosters a culture of rigorous scientific inquiry in AI research. Initial metrics show a high rate of experimentation, with significant data generated from the first night alone, indicating a vibrant and dynamic collaborative environment. This model not only broadens the scope of AI development but also allows researchers to leverage the collective knowledge of the community in addressing complex challenges.