cartero Tuesday, August 11, 2026 · No. 25887
PostgreSQL

Radim Marek: Same rows, different SUM

Everyone knows not to store money as a double precision. One can hope. The rule is so well drilled that it has stopped being interesting, and it is also not where the trouble usually starts. The float is already in the schema before anyone weighs in on it: a measurement column someone later sums for a report, telemetry that drifts into a finance dashboard, a third-party feed ingested as double precision because that is how it arrived. Here is the part the rule does not warn you about. Take a ...

Database Administration and Tooling

I built a CLI tool that makes AI schema-aware before it writes SQL

I wrote a longer post about a small internal tool I built after running into the same AI failure mode again and again: AI could write decent SQL syntax, but it kept guessing the wrong tables, joins, and business meaning in a large Oracle database. The real problem was not prompting. It was context. So I built `db-cli`: a simple read-only CLI that lets an AI assistant explore schema metadata before writing SQL. The workflow is basically: ```text search -> inspect table context -> run sma...

AI Hardware

The Consensus Weekly, June 27, 2026

From The Consensus The feature in OxCaml that more languages should steal In most languages, you hunt down allocations with a profiler and they creep back the moment you touch the hot path. Jane Street's superset of OCaml lets you flip that around: annotate a function with [@zero_alloc] and the comp