Thoughts on AI & Software
Perspectives from the Kaizen team on building enterprise AI products and modern software systems.
0 guessed
Miles, position and hours from the source
The Three Numbers Your TMS Shouldn't Guess: LoadHive with PC*Miler, Samsara and Motive
How far the load really is, where the truck really is, and how many hours the driver really has left. How LoadHive gets each one from the source, and what that does for the quote, dispatch, detention, settlements and IFTA.
Try the best-fit driver demo
Aug 2026
Lumino Split: Split the Bill Without Typing the Bill
Every other split app starts you at an empty form. Lumino starts at the charge already sitting on your card. Try the live cent-exact allocator and watch six payments collapse into two.
Jul 2026
Knowledge Graph vs. Context Graph: Context Engineering for AI Agents
Your knowledge graph knows everything — that's the problem. The runtime layer between enterprise memory and an agent that can actually decide, and why the permission gate belongs third, not last.
Jul 2026
AI Agent Memory in Production: Architecture, Failure Modes & Fixes
The model remembers nothing between calls — every bit of "memory" is something you engineer. The four memory types, where they break at scale, and how to make them production-grade.
Jun 2026
Google's 5 Free AI Agent Whitepapers, Explained
Five production-grade whitepapers on building AI agents — from vibe coding to security to spec-driven production — with direct links to read each.