Tooltician / Engineering Journal

Build. Measure. Verify.

A place for engineering experiments, tools I actually use, and technical decisions explained with evidence. No showcase benchmarks or conclusions the numbers cannot support.

Latest articles / 01

Python · Performance · Experiment

Python 3.13 vs. 3.14 vs. 3.15: newer doesn't always win

14 workloads, three versions, three separate sessions and 1,512 benchmark executions on an Intel Core i5-12600K. Which improvements replicated, where an older version won and what the results still cannot tell us.

Carlos Ortega González
Read the experiment →