This track features:

> Custom Node.js based MIDI.js Electron app with human-authored randomized chords sequencing

> Klatt synthesis with DECTalk which you can get open source from Github. This track features a curated randomized list of Klatt-based speech synthesis utterances with a custom voicing

This track is about artificial intelligence. Today's large language models use a concept called 'temperature' which introduces randomized choices of next best tokens given the previous tokens and a corpus of text. When we lean on language models, what influence does randomization have on the process?

Tokens are sequences of characters used in natural language processing. Where does a word start & where does it end?