EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 17·5 min read
Amazon, which started off selling books, is destroying rare texts to train AI
TL;DR
Amazon is acquiring rare books, destroying them to scan for AI training data, which raises ethical concerns about the preservation of cultural heritage.
✦ Why It Matters
Consider the ethical implications of using rare texts for AI training in your projects.
Key Takeaways
- Evaluate the sources of training data for your AI models.
- Consider the ethical implications of using rare or unique texts.
- Explore alternative data sources to avoid model collapse.
- Stay informed about industry practices regarding data acquisition.
How It Works
Amazon acquires rare books, removes their spines, and scans the pages to create digital text. This process allows them to gather unique data that is not available online, which is crucial for training LLMs effectively.