Wow, 12 bars of 26 notes constitute copyright violation, and I presume monetary damages.
Given that all the hyperscalers, Meta, Google OpenAI, anthropic, etc have all read the entire web for weights, why are the hyperscalers not committing copyright violation on a previously unimaginable scale?