Suno hack exposes scraped training data: millions of hours from YouTube and podcasts
Justgototheeffinmoon · reddit · 2026-07-27
Generative music startup Suno was recently hacked, leading to the leak of its source code and some user data. According to 404 Media, the leaked files explicitly reveal the inventory of Suno's training data sources.
The leaked data shows Suno scraped 113,879 hours of YouTube Music audio (over 2 million clips), 12,287 hours from Deezer, 17,615 hours from Genius, and around 1 million hours of podcasts. The code also included routines specifically searching for acapella versions of songs.
This breach is highly significant for ongoing copyright lawsuits. The RIAA previously sued Suno for "stream ripping" from YouTube, and this leaked inventory turns those allegations into documented internal evidence.
More from Multimodal
- Third-party test: Claude Opus 5.5 renders finer 3D scenes but costs 13x more than GPT-6 Sol — testingcatalog · 2026-09-23
- ComfyUI trick: aux preprocessor + Qwen transfers poses across characters with one prompt — Acceptable-Work8202 · 2026-09-23
- Same portrait prompt across Midjourney V6.1, V7 and V8.2: do older models look better? — tisch_eins · 2026-09-23
- Testing AI character consistency across a 20-image travel sequence — SiennaVaire · 2026-09-23
- Midjourney v8.2 Faces: New Portrait Generation Samples Shared — azed_ai · 2026-09-23
- One-sentence prompt generates lifelike dog video, shown side-by-side with the real one — wgrathwohl · 2026-09-23