Norm examines how degrees of agency over the algorithm can be handed back to the user.
Norm is an API built with modular deep neural nets, comprised of a corpus of historical texts, and hip hop lyrics. All of this running directly in the browser with no communication of what you write back to a server. Norm offers opportunities of authorship through control over curatorial algorithms and data embodiment. Played out in familiar instances of “autocomplete” across an arrangement of interfaces - it politicizes the curatorial algorithm as a cultural agent of knowledge production rather than simply a tool for information consumption.
The latest version of Norm is very rough and very crude (literally) and will be updated over the next few weeks.
You can see where it's heading at norm.salted.nu.
It will have pre-trained models (deep neural nets) that can be loaded and run in the browser (currently it only has West Coast hip-hop 😉 ). The autocomplete system will be trained with everything from the collective works of Drake and Arthur Conan Doyle to topics like neuro-biology and law. And will give you control of the voices that influence your writing.
In the meantime you can read my thesis-paper...