AI startup Conduit builds largest brain-to-language dataset in underground lab
Published on: Dec. 13, 2025, 1:32 p.m. | Source: The Indian Express
A small AI startup in San Francisco has gathered 10,000 hours of brain-language data from thousands of volunteers to teach machines to translate thoughts into text. They used custom-built headsets and a conversation-based approach to collect the data, overcoming challenges such as electrical noise.
