Richard Johnson

A six-step tolerance stack-up workflow and where AI genuinely helps: worst-case and RSS methods, the formulas, and why a chatbot is a method assistant, not the analysis engine.

A reference guide to material properties databases, from free lookups like MatWeb to design-allowable sources like MMPDS and mill certificates, plus where AI helps and where it must not.

An honest guide to using AI for cable sizing and voltage drop: what a language model speeds up, where it fails, and why the adopted code table and a qualified engineer own the final number.

How to use AI to shortlist and justify a material choice, the tools that are genuine machine learning versus curated databases, and why to never trust an AI-quoted property value.

AI takeoff software auto-measures and counts quantities from drawings. We grade 10 tools from true AI auto-measure to traditional digitizers, for estimators and quantity surveyors.

A seven-step method for engineers to understand, validate, and document an inherited undocumented Excel calculation sheet, with AI as an assistant.

Compare nine AI component search tools for engineers, from Zenode and PartGenie to Octopart, with each tool’s real AI graded honestly.

When should engineers use AI? A decision framework: where AI fits (drafting, explaining, method setup), where it does not (safety-critical calcs, code compliance), and a checklist.

Does AutoCAD Electrical have AI? Yes, but it is AutoCAD’s inherited Autodesk AI; the toolset’s wire numbering and reports are rules-based, not AI.

A prompt library for mechanical engineers: statics, stress, tolerance stack-up, GD&T, materials, thermo, FEA, and troubleshooting, each with an honest note on where ChatGPT must not be trusted.

Can AI interpret building codes? It explains concepts and summarizes pasted text, but hallucinates sections and misses local amendments. What the evidence shows and how to use it safely.