Data◆Article
The experimental prototype works extremely well: 1,000 simulated revisions of a document compressed from 20.4 MB of raw revision text to 80.3 KB as a Zstandard-compressed JSON array.
GPT-5.6 Sol Pro built the prototype (38 minutes of work), and the results confirm the approach: a roughly 250x compression ratio. ✦ AI generated
The author (on GPT-5.6 Sol Pro output) · Simon Willison's Weblog · 2026-08-09 · original ↗
The approach works really well! 1,000 simulated revisions to a document resulted in 20.4 MB of raw revision text that compressed to 80.3 KB as Zstandard-compressed JSON array.
Read full article ↗excerpt · fair-use quotation
- ·1,000 simulated document revisions processed in prototype
- ·Raw revision text: 20.4 MB
- ·Zstandard-compressed JSON array: 80.3 KB
- ·Roughly 250x compression ratio confirmed
- ·Approach works extremely well, per author
- ·Built by GPT-5.6 Sol Pro in 38 minutes
- ·Results confirm the compression approach
Around this claim
This moment responds to
gives example → Bundling every version of a document into a single blob and applying a good compression algorithm should wipe out most of the redundant text.The author (as spoken to GPT-Live) · Simon Willison's Webloggives example → The scheme is a BLOB history column holding a compressed JSON array of all previous documents, plus an uncompressed parallel JSON array of timestamps.The author (as spoken to GPT-Live) · Simon Willison's Webloggives example → Storing revision histories in relational databases is perennially difficult, and column-per-version approaches waste enormous space on long documents.The author (as spoken to GPT-Live) · Simon Willison's Weblog