Why AI companies are signing content deals with publishers
News archives, forums and media libraries are being licensed to AI developers. What these deals include, why they happen and how they relate to copyright fights.
AI Desk3 min read
Tag
#Training Data
Every The Latent Daily story, tool, model and resource tagged “Training Data”.
5 items
NLP & LLMsResearch
Chinchilla explained: why AI models needed more data, not just more parameters
DeepMind's 2022 Chinchilla paper found that most large language models were undertrained. The 20-tokens-per-parameter rule, its impact and why labs now go beyond it.
AI Desk4 min read
DatasetsResearch
The Pile explained: the open dataset that helped democratize language models
EleutherAI's 825 GB Pile combined 22 diverse text sources into an open training set. What it contained, why diversity mattered and the copyright questions it raised.
AI Desk3 min read
DatasetsResearch
FineWeb and Common Crawl explained: how web data becomes training data
Most language models learn from filtered web crawls. How Common Crawl works, how FineWeb cleaned 15 trillion tokens, and why filtering choices shape model quality.
AI Desk3 min read
AI and copyright: the legal battles over training data, explained
Can AI companies train on copyrighted books, news and images? The key lawsuits, the fair use and text-and-data-mining debates, and how licensing deals fit in.
AI Desk3 min read