October 9, 2025
It's wild listening to bossa nova me, beethoven me, and bollywood me talk about voice AI latency.
Communicating with people across language barriers is now a solved problem (from a low-level technology perspective, at least).
Three years ago, if you'd told me we'd be able to do very high-quality machine translation between a wide range of language pairs, it would have blown my mind.
And this is only one of several similarly sized new abilities: we blew by the Turing test in both text and audio; we're there for non-realtime video, too, and pretty close for realtime video; LLMs write lots of quite complex code for me; almost every day I ask an LLM to compile a custom research report about some specific topic.
In the early days of the Internet, the global communication possibilities were what excited me most. We got a lot of other things, too. A vast and new kind of library. Location-aware computing. New media formats and creative possibilities.
LLM translation feels like coming full circle back to the moment when I first learned I could STOR a file on an FTP server and anybody with a network connection could instantly RETR it. And bonus points if you had started using a new-fangled thing called a "web browser."
Thanks to @craigsdennis, @thorwebdev, and @elevenlabsio for the great video clips.
AI IS FOR EVERYONE!
We worked with @elevenlabs Productions to get dubs in German, Hindi, and Portuguese.
They do an incredible job of making sure that context is retained and help you scale your content to more people.
Here watch @kwindla speak in multiple languages!
