Announcement·
Phonetico Speech: 14.7 Hours of Tigrinya
Phonetico's first public speech dataset. CC-BY-4.0. 14.7 hours of read Tigrinya from 161 speakers, collected and verified end to end.
Solen Feyissa
Essays & Dispatches
Blog
Notes from the work: Ethiopian languages, the machinery that hears them, and what we’re learning as we build.
Research Note·
Tokenizer Extension for Amharic Whisper
A controlled four-condition study isolates a tokenizer fix that recovers the unused portion of Whisper's decoder, with no change to data, model, or compute.
Phonetico
Perspective·
Low-Resource Is Not a Property of the Language
Many Ethiopian languages are categorized as 'low-resource', a term that describes an attention and investment gap, not a linguistic reality. Phonetico is building the infrastructure to close it.
Solen Feyissa