Distilly

Developers have created an open-source tool named Distilly, hosted on GitHub by project creator titanwings, that analyzes a person's text records to build reusable AI profiles. The software examines past writings, transcripts, and workplace communications to capture how someone solves problems and expresses ideas. It then packages those patterns into modular files that compatible AI assistants can load to emulate that specific individual.

When a key colleague leaves an organization, their institutional memory and unique decision-making style often leave with them. Distilly offers a practical way to preserve that operational knowledge so an automated assistant can continue offering advice or writing documents in a familiar voice. Outside the workplace, users could theoretically preserve the communication styles of relatives or learn from the published thinking patterns of public figures.

It remains unclear how accurately these generated profiles handle totally novel scenarios where no past text exists to guide them. The repository documentation clarifies that the system maps observed behaviors rather than cloning actual human consciousness. This introduces an open question regarding workplace accountability, as teams will need to decide who takes responsibility when an automated system relying on a former worker's profile gives flawed advice.