AI Companies Are Buying Rare Books, Scanning Them for Training Data, Then Shredding Them
Multiple AI companies are buying rare books, scanning them for training data, then shredding the originals — including irreplaceable texts.
Alibaba's Qwen AI team has released Qwen-Image-3.0, the latest iteration of their multimodal image model, promising what the team describes as "authentic details, deep knowledge" — improvements in content understanding and visual detail accuracy.
Specific benchmark numbers were not included in the announcement materials at publication time. The release surfaces through Hacker News, pointing to the Qwen team's official blog, where the model's capabilities are described in qualitative terms. Independent evaluation has not yet been published.
The release arrives amid ongoing community discussion about how Chinese AI labs are performing relative to Western proprietary models. Whether Qwen-Image-3.0 moves that comparison in a meaningful direction will become clearer as developers begin running their own tests.
Qwen has built a following among developers and researchers interested in capable open-weight multimodal models. This release appears to be a capability push — but the specifics will need independent stress-testing before conclusions can be drawn.
All comments are reviewed before appearing. Keep it respectful.
Multiple AI companies are buying rare books, scanning them for training data, then shredding the originals — including irreplaceable texts.
Adam Mosseri says Instagram no longer requires engineers to code 40–60% of the time and has scrapped the traditional technical interview loop.
Sam Altman says AI won't shorten the workweek because humans enjoy staying busy and will just create more work to fill the time.