Ought
Ought is a non-profit research lab focused on delegating high-quality reasoning to advanced machine learning systems and has spun off Elicit, a public benefit corporation designed to assist with literature reviews.
Ought Company History
Ought was founded as a non-profit research lab dedicated to developing mechanisms for delegating high-quality reasoning to advanced machine learning systems. The organization is focused on scaling up good reasoning with the help of machine learning and aims to make these systems useful for addressing open-ended questions while avoiding alignment risks posed by end-to-end optimization. Ought has spun off Elicit as an independent public benefit corporation and has gained funding from notable organizations such as Open Philanthropy, Jaan Tallin, and the Future of Life Institute.
Elicit's Role and Funding
Elicit, a product developed by Ought, was spun off as an independent public benefit corporation. It has raised $9 million in venture funding to further its development. Elicit assists with literature reviews but does not automate them completely. The tool's effectiveness depends heavily on the quality of the underlying research and includes interactions with multiple machine learning APIs for various tasks. Users are encouraged to double-check Elicit’s results and understand its limitations.
Ought's Research and Development
Ought has developed several key tools such as the Interactive Composition Explorer (ICE) and the Factored Cognition Primer. The organization focuses on process-based supervision for language models to ensure high-quality reasoning capabilities. The development work is influenced by the latest advancements in machine learning technology and aims at balancing clarity and accessibility with comprehensiveness and nuance.
Ought’s Funding and Support
Ought has received funding from multiple grants and contributions from organizations like Open Philanthropy and the Future of Life Institute, along with individual backers identifying with effective altruism and longtermism communities. This funding supports Ought’s mission to scale up good reasoning using advanced machine learning systems and to develop mechanisms that mitigate alignment risks associated with end-to-end optimization.
Distributed Team Locations
Ought operates with a distributed team located in various regions, including the Bay Area, Austin, New York, and Oristà. This diverse geographical presence allows the organization to pool talent and perspectives from different areas to advance its research and development in machine learning and reasoning mechanisms.