AI

AI shows up on this blog as four separate arguments, not one topic. The first is the explainer thread, written for someone who needs the vocabulary before the next vendor call: what a large language model actually is, what a context window buys you, and a direct comparison of Claude, GPT-4 and Gemini covering context size, coding, writing, analysis and where each one is weak. The question it answers is which model you standardize on.

The second thread is agents. There is a real difference between a chatbot that answers a question and an agent that reads a request, checks a system of record, acts, and reports back. The post here walks that loop step by step, perceive, reason, plan, execute, evaluate, and contrasts it with the if-then automation most teams already have.

The third is AI inside one industry at a time. Two long pieces, one on construction contractors and one on independent medical practices, look at where AI is actually being adopted in those sectors, lead scoring, follow-up, scheduling, review monitoring, and what stops smaller firms from starting.

The fourth is the skeptical one. AI React Bias argues that coding assistants recommend React for nearly everything because React dominates their training data, and that for a marketing site or a blog you inherit JavaScript you never needed. Read it before you let an assistant pick your stack.

No AI strategy abstractions here. Keep scrolling if you want to know what these tools do and where they already work.