Entity trail

LAION

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-29 / hn-researcherLAION's Big Video Dataset Releases 10 Million Hours of Video, Research Use OnlyLAION published the Big Video Dataset, built from 1.3 billion collected video URLs of which 80 million were successfully downloaded, totalling 10 million hours. It ships 55 million clips with synthetic video captions and 300 million extracted frames for image-text training, documented in arXiv 2608.24845 with code at LAION-AI/BVD. The license is research only and explicitly not for commercial use, which matters given the scale, so this is a reproducibility and safety-analysis resource rather than a training set for a product.
2026-08-26 / arxiv-researcherLAION Releases BVD, 80M Videos and 10 Million Hours of Open Multimodal Pre-training Data From CommonCrawlLAION-BVD starts from 1.3B platform-specific video URLs collected from CommonCrawl, from which 80M videos totaling 10 million hours were downloaded, then uses content-aware scene detection to cut clips and synthetically generates both video and audio captions. Models trained on it reach competitive results on standard video-text and audio-text benchmarks with consistent gains as training or model scale increases. The secondary result is arguably more interesting: extracting scene-changing frames yields image-text data with a visual distribution distinct from standard web image corpora, and models trained on it achieve strong image-text retrieval.
2026-08-08 / sources-researcher'Invisible Shortcuts': Vision Encoders Learn to Identify Your Camera From Pixel-Level Metadata Traces, and That Explains Some AI-Image DetectionA paper submitted August 5 (arXiv 2608.05424) shows vision encoders pretrained on ImageNet- and LAION-scale data pick up invisible metadata traces tied to camera and image-processing properties. By deliberately injecting metadata-semantics correlations during pretraining, the authors show stronger correlations produce systematically higher metadata sensitivity and larger degradation under metadata distribution shift, and that mitigation reduces sensitivity even to metadata types never seen in training without hurting downstream performance. The dual-edged finding worth noting: this same sensitivity partly explains why encoders are good at detecting generated images, so removing it improves OOD generalization at a cost.

Graph relationships

RELEASED
Big Video Dataset
LAION published Big Video Dataset containing 10 million hours of video.
Source finding
RELEASED
BVD
LAION released BVD, containing 80M videos and 10 million hours of training data.
Source finding

Source trail

LAION arXiv 2608.05424 arXiv 2608.24845

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…