Static

Engram is a sampler that turns broken AI hallucinations into music

First reported by The Verge ·

The signal ●○○○ Compiled by AI from The Verge, the single source so far
Why you might care

Experimental music producers can now use AI to create unique, broken sounds without needing extensive coding knowledge.

What happened

Music startup Thoughtful Things has launched a Kickstarter for Engram, a new AI-powered sampler and groovebox. Unlike AI music generators focused on commercial hits, Engram is designed for experimental sound design, transforming incoming audio and generating novel, often uncanny, sounds. It runs a custom-trained, proprietary "tiny AI" model locally, ensuring it does not use non-commercially licensed data. The device is inspired by circuit bending, allowing users to manipulate and even break its AI models. Engram is positioned as a tool for exploring AI hallucinations in audio, with plans to release its firmware for customization and user-created models. The Kickstarter campaign offers an early bird price of $675, with an anticipated retail price between $850 and $900.

What it means

Engram represents a significant shift in the AI audio landscape, moving beyond commercially viable music generation to embrace the avant-garde and experimental. By focusing on "hallucinations" and "broken" sounds, the device taps into the unpredictable nature of AI, offering artists a new palette for sonic exploration. The local processing and commitment to using only commercially licensed training data also address growing concerns about copyright and data provenance in AI development.

This approach democratizes access to experimental AI sound design, making it accessible to musicians and sound artists who may not have deep technical expertise. The potential for open-source firmware further encourages community-driven innovation, allowing users to customize and extend Engram's capabilities. This focus on user modification and exploration positions Engram not just as an instrument, but as a platform for future AI audio research and artistic practice.

AI-written summary. May contain errors.