Yet translation was the main application for applied language machine learning.
When would've thought that to solve natural language translation, one would first need to solve... natural language.
All those arguments about agents and hallucinations kind of distracted people from noticing we've accodentally built a universal translator.
When would've thought that to solve natural language translation, one would first need to solve... natural language.
All those arguments about agents and hallucinations kind of distracted people from noticing we've accodentally built a universal translator.