Our thesis and mission

Autodidact Labs is an applied ML and agentic minilab built out of Harvard University that seeks to push the needle forward on continual learning and adaptation through context management, composition, and foundational research.

The foundational policy of any agent is to become an extension of its user.

We believe that any agent can be reduced to how well it interprets and performs its user's intentions. In order to do that, the system must first understand its user, continually learning and adapting to their preferences.