It’s time for another data/AI roundup and here are the highlights from February👇

𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐜𝐞 & 𝐀𝐈

Inside OpenAI’s in-house data agent

A practical guide to which AI to use in the agentic era

Why judgment may not be uniquely human after all

How Codex is being used for serious research automation

Why semantic linking matters for giving data meaning

𝐃𝐚𝐭𝐚 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠

A portable analytics stack built on DuckDB, DuckLake, dlt and SQLMesh

Why healing tables beat slow-motion backfill disasters

The case for MetadataOps engineers

How to use AI tools without losing data engineering fundamentals

Why 5-second BigQuery queries can still be expensive

𝐃𝐚𝐭𝐚 𝐀𝐧𝐚𝐥𝐲𝐬𝐢𝐬 & 𝐁𝐈

The state of machine learning competitions in 2025

Plus: why AI is eating software’s TAM, what world models could unlock in robotics and why AI may intensify work instead of reducing it.