Focus of Letter 9

The focus of the week will be devoted to Handy, of course for its ease of use, its installation and local operation and above all its more than satisfactory results with the models offered.

Handy’s choice is also for its history: its author, developer, after breaking his finger, is looking for the speech recognition tool that will suit him. Between high costs, stored data or lower performance, he decides to develop it himself from shared open-source “speech to text” models. He succeeds with a simple and effective tool on all systems that he decides to make 100% open-source, fully shareable and modifiable for other evolutions.

Behind this beautiful story, there are the sharing of open-source technologies – totally open as here or semi-open for many language models – which are often at the base or simply used in many generative AI tools, often erased by the big announcements of new “revolutionary” models (at least for a week or two…) and major publishers.

Above all, there is the demonstration that the willingness to share and effective sharing is a possible model, thanks to previous sharing and future appropriations that will allow other applications.