Featured in this episode of Tech News of the Week
Project Gutenberg is an online library of free ebooks that are in the public domain. Despite what Disney might have you believe, intellectual property is actually supposed to end up in the public domain after a relatively short period of time. Popular titles include “Romeo and Juliet”, “Frankenstein”, and “Alice’s Adventures in Wonderland” among their 70k+ titles.
The thing is, reading is hard and I have shit to do. That’s why I love audiobooks and I’d really like it if those 70k titles were available in audio format. But alas, human readers are expensive and busy. Which is where a partnership between Microsoft, Google, and MIT step in using AI to generate 5,000 open-license audiobooks.
This isn’t the first time someone has attempted to use text-to-speech for audiobooks, however those efforts tend to sound robotic and unlistenable for even a novella length work. The new models being used by developers at Microsoft have a far more natural cadence and timbre, resulting in an almost enjoyable experience, even if it doesn’t reach the heights of a Stephen Fry or Roy Detrice.
If you’re interested, the entire project is open-source and published on Microsoft’s Synapse ML site. You can find all the books on Spotify, Apple Podcasts or the Internet Archive.