GoKawiil
Tech News
The company behind a cloud data warehouse rebuilt its backend around DuckDB five months ago to give customers dedicated instances and a cost-based optimizer. After three months of re-architecting, it says it has removed DuckDB from its stack because customer workloads grew larger and more complex than its single-node, pre-sized instances could handle efficiently.
The DuckDB team released version 2.0.0, a long-term support build nicknamed 'Cinnamon Teal,' incorporating nearly 20,000 commits since version 1.5. The update adds a polished Quack protocol, asynchronous I/O, a new VARIANT data type, an agent mode, triggers, and an extensible SQL parser. DuckDB says this LTS release will be supported until at least October 2027, alongside continued support for the 1.4 and extended 1.5 lines.
Will Lachance describes his experience using Implementation Decision Records (IDRs), a practice he devised to get better output from LLM coding agents and keep a durable log of implementation decisions. He found that when given a vague prompt, an agent will often just fill in the IDR with a generic restatement of what it already intended to build, producing a document that is lossy and arguably worse than a plain code diff.
DuckDB 2.0 introduces VARIANT as a native data type, replacing text-based JSON storage. The engine analyzes JSON fields during checkpointing and extracts consistent fields—like a field that is always a number or always text—into real typed columns, leaving only the inconsistent or rare fields in a binary remainder.
DuckDB released a DuckLake extension, enabling DuckDB to read and write an open lakehouse table format that stores metadata in a catalog database and data as Parquet files. Users can install it via INSTALL ducklake, attach a DuckLake database with standard ATTACH syntax, and then create, query and modify tables using regular SQL, including updates, time travel queries, and schema evolution.
Polars has released version 2.0, introducing initial out-of-core processing, significant performance improvements, and comprehensive SQL support. Benchmark tests show Polars surpassing DuckDB and DataFusion in TPC-H and TPC-DS workloads, highlighting its growing capabilities for data analytics tasks.
Planet Labs, the San Francisco-based Earth-imaging company, maintains a public ephemerides feed that lists tracking data for its satellite fleet. A query of this data against catalog records shows 97 Planet Labs satellites currently active, spanning several constellations including the newly launched Pelican-11, which was deployed into low Earth orbit via SpaceX's Transporter 17 mission on July 7th.
Keenable has released SELECT, an MCP server that lets an agent query the web using DuckDB SELECT statements instead of clicking through search results. The tool combines web search and fetch operators with semantic extraction functions inside SQL, filtering pages with exact WHERE clauses before sending only surviving rows to an LLM for field extraction. Each generated report links to the full trajectory of queries, tool calls, and result sets used to build it.
DuckLabs, the Amsterdam-based company founded by DuckDB creators Mark Raasveldt and Hannes Mühleisen, announced it will join Amazon Web Services effective early September. The team plans to stay together in Amsterdam and continue developing DuckDB, DuckLake, Quack and related community projects. DuckLabs said the core DuckDB stack will remain free and open source under the MIT license, with the nonprofit DuckDB Foundation continuing to oversee governance.