Words, when we speak, fly away, only the written ones remain. We have been saying it to each other for two thousand years: verba volant, scripta manent. Then came Musixmatch that thanks to an artificial intelligence, and to the work of its community, has transcribed all or almost all the lyrics of the songs in the world and continues to do so. And now he wants to do the same with podcasts.
The podcast with the voices of young people that we cannot ignore
by Riccardo Luna
Stop the words that fly, in fact, and write them so that everyone can read them, understand them, look for them, share them. A huge challenge: there are four million podcasts and sixty million episodes in circulation; it has been calculated that two more are added every minute. How to transcribe them all? And above all why? I asked the founder of Musixmatch, Max Ciociola, who is not an American, but was born in Puglia, he founded the company in Bologna, where he was studying, about ten years ago, aiming to create the largest archive of texts of songs; even there, an impossible yet successful undertaking, today Musixmatch grinds profits and a few months ago he sold the majority to a Californian group but he remained at the helm.
Number one reason for transcribing podcasts? He told me: for all those who have hearing problems and so they can at least read the content, this would be enough; and then for those who do not speak the language of the podcast well enough and instead are facilitated by reading them; and finally, to make podcast contents accessible through a search engine that can recognize what has been said at any time even if that content is not in the title. In short, to find out what the world is talking about. Not what we write, but what we say. And create new connections and, ultimately, value in this.
Elections
Podcast – Next Gen: new voices to vote
Behind this encyclopedic challenge is the artificial intelligence developed by Musixmatch, that of song lyrics; and there is a neural network that goes in search of connections and meanings and that was called Umberto, in homage to Umberto Eco, not only because he was from Bologna, but because he was a world reference point for what concerns the study of the meaning of the words. Musixmatch is thus a candidate to become the Google of audio in the digital world, a world in which words continue to fly, but then remain.