A guitarist with five million YouTube subscribers just became one of the most visible mainstream voices arguing that local AI models will replace cloud-based ones. Rick Beato - best known for music theory breakdowns and industry commentary - published a video drawing a direct line between how the music industry failed and how commercial AI companies are headed for the same cliff.
The Music Industry Parallel
Beato's argument is straightforward: the music industry ignored consumer behavior, locked content behind paywalls and DRM, and lost to piracy and free alternatives. He sees commercial AI following the same script - subscription fatigue, aggressive data collection, and increasingly restrictive terms of service will push users toward free, locally-run alternatives that already exist and keep improving.
It is not a new argument in AI circles, but hearing it from someone whose audience is musicians, producers, and creative professionals - not developers - signals how far the local AI conversation has spread beyond tech communities.
LM Studio and Qwen on Camera
The most interesting part of the video is practical. Beato walks through setting up LM Studio (a desktop app for running AI models on your own computer with no cloud connection) and runs Qwen 3.5-35B, a 35-billion-parameter open model from Alibaba. He shows it handling creative tasks, answering questions, and working entirely offline.
For his audience, many of whom have powerful machines for music production and video editing, the hardware barrier is lower than they might expect. A decent workstation with 24GB of VRAM - common in creative setups with an RTX 4090 - can run a 35B model reasonably well.
Privacy as the Selling Point
Beato hammers privacy throughout. Musicians and producers routinely work with unreleased material, client projects, and contractual obligations around confidentiality. Sending lyrics, melodies, or business documents to a cloud AI service creates risks that running a model on your own hardware simply does not.
This is the argument that tends to land hardest outside tech circles. Developers debate benchmarks and token speeds. Creative professionals and small business owners care about whether their unpublished work is being used to train someone else's product.
The video is a useful marker of where local AI adoption stands: the tools are mature enough that a non-technical creator can demo them to millions of viewers without it feeling like a stretch. LM Studio, Ollama, and similar apps have made "run your own AI" a realistic option for people who have never opened a terminal. The models are not as capable as GPT-4o or Claude for every task, but for a growing number of use cases, they are good enough - and the privacy tradeoff tips the scale.