Explores co-ownership of musician AI models for creative collaboration.
Connects machine learning to traditional music appreciation and context.
Discusses the need for platforms to document and release AI-generated music.
Deep dives
Introduction of the Artist and Musical Background
The artist, Moises Sorta Valenzuela, introduces himself as a sound artist and electronic musician with roots in Tijuana, Mexico. He discusses his background in traditional and electronic music, as well as his transition into solo work focused on generative sound and neural audio synthesis.
Influence of Traditional and Diasporic Music on Machine Learning Interest
Moises reflects on his band's sampling of Latin American traditional music, highlighting the connection between machine learning interest and music rooted in traditions. Sampling from diverse internet sources led to a deep appreciation for the social and historical context of music, aligning with machine learning's data-driven nature.
Connection Between Speculative Sound Art and Machine Learning
Moises describes his entry into AI through speculative sound art discussions, particularly influenced by controversial artist Goody-Pel's futuristic AI concepts. The exploration of deep learning papers like WaveNet intrigued Moises, leading him to combine interest in chance composition and neural audio synthesis.
Project 'Muto Alismo' and Collaborative Neural Music Synthesis
Moises introduces 'Muto Alismo,' a project involving 13 artists contributing individual data sets for neural music synthesis models. Each artist trained their model, fostering a co-op approach with shared ownership of generated music. The project blends features and patterns to create a new sonic universe, emphasizing collective model ownership in music composition.
Innovative Digital Music Creation Project
The podcast discusses an innovative digital music creation project where the host plans to release musical instruments loaded with various models, allowing artists to use their own models with ownership rights. The collaborative project aims for consensus-based ownership, enabling artists to continue using their models even after the project ends.
Challenges and Opportunities in AI Music Industry
The episode explores the challenges and opportunities in the AI music industry, highlighting the lack of dedicated labels capturing advancements in neural ML audio. The discussion delves into the need for platforms to document and release AI-generated music, emphasizing the potential for a new economy and online music label to address the evolving technologies in music production.
We catch up with Moisés Horta Valenzuela (Hexorcismos) to discuss his remarkable experiments with music and machine learning, and his upcoming software project Semilla that experiments with co-ownership of musician AI models shared for anyone to use. Just...the best.
