Benchmarking C versus Rust for SQLite number parsing extensions
Simon Willison compared C and Rust implementations of a SQLite parse_number() extension, finding optimized C achieved 16-19 ns per row.
Data & spreadsheets
CSV and Excel imports, data cleaning and moving data between systems.
Grist introduces window functions to fix spreadsheet data integrity
Grist adds PREVIOUS and NEXT functions to handle running totals and ordered lookups without the fragility of traditional cell references.
Data engineer swaps complex big data stacks for DuckDB and DuckLake
Veteran data engineer Matt Martin details his shift from Spark and Iceberg to DuckDB, citing reduced operational overhead and faster metadata handling with the new DuckLake format.
New tool shrinks spreadsheet data for AI agents by 1000x
Sheetgraph converts complex Excel and Google Sheets files into compact reports, allowing AI agents to analyze large datasets without exceeding token limits.
Grist's hidden footprint reveals secure self-hosting patterns
ZoomEye scans from October 2026 show Grist has a minimal public presence, with only one instance on its default port, indicating strong internal deployment habits.
DuckDB 2.0 adds server mode, async I/O and variant type support
DuckDB 2.0.0 LTS introduces client-server capabilities via the Quack protocol, asynchronous I/O for faster remote reads, and a first-class VARIANT type for semi-structured data.
OpenClear: API-first positive pay system with robust bulk import
GitHub project OpenClear offers a self-hosted positive pay system with multi-bank support, featuring secure bulk imports for checks, ACH, and accounts.
DuckDB 2.0 alpha delivers faster S3 reads and recursive queries
The DuckDB 2.0 alpha introduces async I/O for cloud storage, optimized recursive CTEs, and a new VARIANT data type that shreds JSON for better performance.
Reducing LLM costs by moving spreadsheet aggregation to the server side
A data engineer reduced token usage by 2,300 times for large spreadsheets by performing calculations on a local server instead of sending raw rows to an AI model.
Vindle optimizes massive CSV imports by replacing PHP with Rust
Vindle reduced processing time for 27 GB of product data by moving CSV filtering from PHP to a custom Rust tool, eliminating redundant parsing passes.
