Writing

Data Materialization is a Convergence Problem
Data materialization isn’t an orchestration problem, it’s a convergence problem, and we need a new system to handle it. Part of the Is the orchestrator dead or alive? symposium on Data People, Etc.
It's Time to Retire the CSV
Despite its ubiquity and ease of access, CSV is a wretched way to exchange data. The time has long passed to retire CSV and replace it with something better.
S3 Intelligent-Tiering: What It Takes To Actually Break Even
When does it make sense for an object to be in Amazon S3’s Intelligent-Tiering (“S3-IT”) storage class? The answer, unfortunately, is “it depends”. (Published on the Duckbill Group blog.)