Discovered Materials, a young San Francisco-based startup specializing in finding new substances for advanced chips, has announced a major challenge featuring a prize pool totaling one million dollars. The company is calling on artificial intelligence models to solve five critical and previously unsolved problems in the field of materials science. Although modern frontier AI models have achieved remarkable success in solving complex mathematics problems and advancing life sciences research, they have not yet successfully navigated the intricate physics involved in advanced materials. According to co-founder Advaith Sridhar, specifying problems clearly in this scientific domain is exceptionally difficult because it lacks the definitive yes-or-no clarity found in mathematics. To establish these goals, the startup gathered an advisory group of academic experts from leading global universities to outline five distinct open challenges. Felipe Jornada, a Stanford professor who assisted in developing these hurdles, noted that explaining currently observed but unexplainable phenomena would greatly improve physics understanding and directly shape future electronics, energy storage, and manufacturing. Before launching the initiative, the startup tested existing lab models on the tasks, allowing autonomous AI agents with internet access to work for days without making significant progress. However, the creators emphasize that the primary objective is establishing clear goalposts to measure AI capability and deepen fundamental scientific understanding, rather than retaining exclusive rights to the solutions. The startup's core business focuses on developing innovative materials for semiconductors through partnerships with industry giants, and leadership expresses increasing confidence in discovering industrially useful materials rapidly. Source: TechCrunch #AI #MaterialsScience #Startups #TechNews #Innovation
Ver más en News Today