Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-29 · source-backed
Built from 1.3 billion collected video URLs of which 80 million downloaded successfully, shipping 55 million clips with synthetic captions and 300 million extracted frames for image-text training, documented in arXiv 2608.24845. The license is research only and explicitly not commercial, which makes this a reproducibility and safety-analysis resource rather than a training set for a product. Read the license before anyone on your team gets ideas about the frame set. (LAION)
Each link below shares sources, entities, or timing with this story.
Shared entity: URLs / Earlier coverage / Tension
Both cover URLs; earlier URLs coverage from 2026-08-28; pushes against this story (against).
Shared entity: Built / Earlier coverage / Tension
Both cover Built; earlier Built coverage from 2026-08-28; pushes against this story (against).
Shared entity: URLs / Earlier coverage / Tension
Both cover URLs; earlier URLs coverage from 2026-08-06; pushes against this story (against).
Both cover URLs; earlier URLs coverage from 2026-07-27; pushes against this story (but).
Both cover URLs; earlier URLs coverage from 2026-07-25; pushes against this story (against).
Shared entity: Built / Earlier coverage / Tension
Both cover Built; earlier Built coverage from 2026-07-15; pushes against this story (but).
Both cover Built; earlier Built coverage from 2026-06-18; pushes against this story (but).
Both cover Built; earlier Built coverage from 2026-04-16; pushes against this story (vs).