Our great sponsors
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
The reason for this is that the training data needed for this simply does not exist. To make such an app, you would need tons of phonetically transcribed recordings from tons of places in tons of languages to act as training data. We just don't have that, and even if we did, phonetic transcriptions miss a lot, like precise vowel height. As a case in point, even major voice assistants like Alexa are worse at understanding non-standard speakers than (relatively) standard speakers. Hell, I use Mycroft AI on a Raspberry Pi as a voice assistant and it can't even understand women because there isn't enough data available from recordings of women.