RAG vs Fine‑Tuning: Which One Should You Pay For?
A practical decision guide: when retrieval (RAG) is enough, when fine‑tuning helps, and how to avoid spending money on the wrong lever.
Writing
Articles on reliability, shipping, and production problem-solving.
How to Hire an AI Engineer: Roles, Interview Loops, and a 1‑Week Paid Trial
A practical hiring guide for startups: what “AI engineer” should mean, what to test for, and how to de-risk with a paid trial.
Prototype Sprint vs MVP: What to Build First (and why it matters)
Prototype or MVP? A simple decision guide for founders, plus example scopes that prevent the common “half-build” trap.
Inside Upright’s Rails Engine Architecture for Self-Hosted Synthetic Monitoring
Upright is 37signals’ self-hosted synthetic monitoring app packaged as a Rails engine. Here’s why the engine + subdomain design makes operations calmer.
Action Push Native: Rails Push Notifications in Food & Fit Without the Glue Code
How Food & Fit uses 37signals’ Action Push Native to send push notifications with retries, safer token ownership, and a Rails-first delivery layer.
MVP Development Cost in 2026: Realistic Budgets + Timelines
A practical way to budget an MVP: what drives cost, what to cut, and how to get a real quote without wasting weeks.
Tests Are an API for Agents: The Fastest Path to Reliable Diffs
Treat tests as the contract between you and coding agents so you ship smaller diffs, cut retries, and avoid silent regressions when speed increases.
Compound Engineering for Codex and Rails Without the Rework
Learn how to make Codex and Rails work compound over time with durable repo memory, contract tests, and clear decision rules that prevent repeat mistakes.
Hosting Is a Trust Problem: Choose a Host for Weird Days
Choose hosting by pricing the weird days: lockouts, restores, support, logs, and security defaults—so recovery stays predictable at 2 a.m.
Ready to make your MVP production-ready?
Free 15-min call → paid diagnostic → 1-week sprint with real fixes in production.