From Big Data to DuckDB: Ryan Boyd (MotherDuck) on Hyper-Tenancy, Duck Lake, and Making Data Fun.

MotherDuck co-founder Ryan Boyd (https://www.linkedin.com/in/ryguyrg/) joins AI by the Bay to rethink analytics in the LLM era. We dig into why DuckDB’s ruthless efficiency and developer experience often beat “big data,” and how MotherDuck’s hyper-tenancy (many small DuckDBs) and dual execution (local + cloud) enable sub-second queries and instant SQL previews. Ryan demystifies Duck Lake, storing lakehouse metadata in Postgres—to cut the complexity and latency of Iceberg/Delta-style stacks. We also talk practical AI: LLMs need trustworthy, aggregatable facts; analytics engines + clean metadata make that possible. Finally, a peek at community strategy, DuckDB Snippets, the DuckDB UI, and why making data fun isn’t fluff but a product decision.