art: Tim Parish, The Journeybook anthology
The Digital McKenna, Terence (DMTi) Initiative
I, Terence is an experiment to bring alive the spirit of psychedelic philosopher Terence McKenna (November 16, 1946 – April 3, 2000) by harnessing existing content in the public domain, his talks, ideas, videos, photos, geo-mapping, etc, with the Lifenaut system of AI intelligence into a functioning Avatar interface.
The platform is designed for multi-users to submit content that will eventually aggregate into a personality matrix ("mind file"). It's all experimental but the foundations are there: Lifenaut is a not-for-profit company with good intentions to explore this new AI modality, and all content is retained for ownership by the original makers. The platform will hold the content and resynthesize with an AI interface, it will learn and grow in useability to enable an interface with the public and the spirit of Terence's ideas, perhaps eventually synthesizing new contours.
Tuesday April 3, 2012 is the 12th anniversary of Terence's passing and the launch of the beta-version of the iTerence project, aka the Digital McKenna, Terence Initiative (DMTi)... We ask the global psychedelic community to crowdsource any bootleg talks, podcasts, images, text, etc. from Terence. This is a non-commercial, open-community project that in essence aggregates existing content Terence has left us into a new, more useable form.
The Lifenaut system is still in early development in that they are not pulling data or metadata from video, audio or text yet – but this is something they are working on, trying to develop algorithms to extract ideas and voice synthesis from video and audio. We reinterate that they are not capable of doing this presently, although by aggregating this content now we will be positioned to work with it when their system is able in the near future. People with relevant skillbases wishing to donate time and expertise or money to Lifenaut may help speed this process to fruition.
At the moment we believe the power in the project is in training the virtual Avatar. After 1000 lines of conversation that digital Terence Avatar starts to become coherent. We have been spending time training the iTerence Avatar by feeding it interview question and answers and quotes from existing talks. The Avatar is still relatively incoherent, but it is getting there. Every day we work with it, it is increasingly more responsive. At the moment only a generic male voice is possible, but we hope that a closer simulation of Terence's voice will eventually be possible from learning from the audio samples.