What AI actually is
The AI shelf, fully stocked. What a model is, which companies train them and which only rent them, the parts an AI product is assembled from, and an honest account of what the technology does today.
Lessons
- What do the AI words on an engineering job description mean? The one guide to the AI vocabulary: what a model is, who trains them, what companies build on top, and what the technology reliably does today.
- What is a large language model, and who makes GPT, Claude, and Gemini? What a model is, what training means, how a large language model turns text into more text, and the labs behind the three names everyone knows.
- Who are the AI companies, and what does each one sell? The industry map: labs that train models, clouds that rent the computers, the chipmaker that sells the GPUs, open-weight models, and the application companies that build on rented models.
- What is the difference between building AI and using AI? The highest-leverage distinction in modern engineering hiring. It re-prices most AI requirements on the spot.
- What are RAG, AI agents, evals, and MCP? The technique map on the using side, each one ranked by how rare the skill is and what hiring it implies.
- When a company says it is an "AI company", what did it actually build? The product shapes AI ships in: a chatbot, a copilot feature inside an existing product, an agent that does a task end to end, a workflow with a model in the middle, and a wrapper around someone else’s model.
- Is AI overhyped? The honest state of the argument: what models reliably do, what hallucination is, what slop means, and why serious engineers land in both the enthusiast and the skeptic camp.