

Big data is dead, analytics is alive (Practical AI #292)
Oct 24, 2024
Till Döhmen, a member of MotherDuck who helped develop DuckDB, joins Kurt Mackey, co-founder of Fly.io, to discuss the future of analytics. They highlight the shift from big data to innovative solutions like DuckDB, capable of lightning-fast queries right on your laptop. The conversation includes AI's role in analytics with features such as text-to-SQL and vector search. They also explore how DuckDB enhances workflows for developers and integrates seamlessly with various data sources, making it a game-changer in the data landscape.
AI Snips
Chapters
Transcript
Episode notes
Initial DuckDB encounter
- Adithya Krishnan initially dismissed DuckDB's speed on a 100MB dataset.
- His experience with Spark made him accustomed to slower processing times.
DuckDB Wasm
- Till Döhmen's interest in DuckDB stemmed from its Wasm version and in-browser analytics capabilities.
- He was particularly impressed by its geospatial analytics potential in the browser.
DuckDB's in-process nature
- DuckDB is an in-process database, maximizing local compute resources.
- This contrasts with traditional client-server databases that introduce transfer protocol bottlenecks.