https://greglum.com/learnLearn AI ยท Greg LumUnderstand AI through everyday questions, examples, and sources. Start with Making Sense of AI, or follow a path for using, building with, or reviewing AI.2026-10-28T00:00:00ZGreg Lumhttps://greglum.com/https://greglum.com/learn/how-ai-answers-workHow does an AI answer get made?2026-10-27T00:00:00Z2026-10-28T00:00:00ZFollow a question through context, text generation, and optional search. Learn what each step can miss.https://greglum.com/learn/getting-help-from-aiGetting Useful Help from AI2026-10-27T00:00:00Z2026-10-28T00:00:00ZGive an assistant enough information, improve a request, and decide whether the result helps.https://greglum.com/learn/ai-privacy-and-actionsWhat should I share, and what can it do?2026-10-27T00:00:00Z2026-10-28T00:00:00ZUnderstand saved conversations, training choices, connected tools, and the difference between drafting and sending.https://greglum.com/learn/understanding-ai-newsMaking Sense of What You Hear About AI2026-10-27T00:00:00Z2026-10-28T00:00:00ZAsk useful questions about product announcements, comparisons, and predictions about the future.https://greglum.com/learn/writing-versus-choosingWriting an Answer or Choosing One2026-10-28T00:00:00Z2026-10-28T00:00:00ZTell writing, finding, ranking, and choosing apart, and see which jobs belong to code or a workflow instead.https://greglum.com/learn/finding-before-answeringFinding Before Answering2026-10-28T00:00:00Z2026-10-28T00:00:00ZFollow one question from search to reply, see why a missing passage cannot be recovered by a better choice, and measure each stage separately.https://greglum.com/learn/reading-model-claimsHow to Read a Model Claim2026-10-28T00:00:00Z2026-10-28T00:00:00ZAsk seven questions of a vendor or benchmark comparison, from baseline and version to sample size, total cost, and who ran it.https://greglum.com/learn/cited-answersHow can I check an AI answer?2026-10-23T00:00:00Z2026-10-27T00:00:00ZCheck the facts, open the sources, and notice what is missing. A confident answer or a citation can still be wrong.https://greglum.com/learn/what-is-aiWhat do people mean by AI?2026-10-27T00:00:00Z2026-10-27T00:00:00ZLook beyond chatbots to understand prediction, recognition, recommendations, and generative AI.https://greglum.com/learn/why-ai-nowWhy is everyone talking about AI now?2026-10-27T00:00:00Z2026-10-27T00:00:00ZA short history of what changed, from learning patterns to conversational tools.https://greglum.com/learn/ai-names-explainedChatGPT, Claude, and all those names2026-10-27T00:00:00Z2026-10-27T00:00:00ZTell a company, a model, an assistant, and a tool apart without needing a technical background.https://greglum.com/learn/designing-ai-systemsDesigning the System Around a Model2026-10-26T00:00:00Z2026-10-26T00:00:00ZChoose the simplest AI design that does the job, then decide its context, limits, recovery, and evidence, with Ask the Archive as the example.https://greglum.com/learn/bounded-decisionsShould a Model Make This Decision?2026-10-25T00:00:00Z2026-10-25T00:00:00ZCompare code, classifiers, and model-based decisions. Set an error limit and test it on fresh cases.https://greglum.com/learn/one-rulebook-two-agentsOne Rulebook for Two Agents2026-10-23T00:00:00Z2026-10-23T00:00:00ZClaude Code and Codex read the same user-level rules file. One import line and one symlink do it.https://greglum.com/learn/public-repo-to-better-skillFrom a Public Repo to a Better Skill2026-10-23T00:00:00Z2026-10-23T00:00:00ZCompare public skills at fixed revisions, check their licenses, and test each accepted change against a known failure.https://greglum.com/learn/catching-what-agents-missCatching What Agents Miss2026-10-23T00:00:00Z2026-10-23T00:00:00ZFour defects in one Draftroom change show how local checks, independent review, browser measurements, and accessibility tests work together.tag:greglum.com,2026-10-23:learn-log/2Prepared my skills repository for public release2026-10-23T00:00:00Z2026-10-23T00:00:00ZA cleaned export of the catalog, with an installer dry run, a validator, and license notices.tag:greglum.com,2026-10-23:learn-log/1Studied public motion skills for the film skill2026-10-23T00:00:00Z2026-10-23T00:00:00ZAdopted timing defaults, script checks, and a critique loop into motion-film, with each source and license recorded.tag:greglum.com,2026-10-22:learn-log/2Cleaned up my development environment2026-10-22T00:00:00Z2026-10-22T00:00:00ZRetired Zed, Neovim, Warp, Podman, Caveman, an offline semantic search tool, and a local docs index. Context7 replaced the docs index.tag:greglum.com,2026-10-22:learn-log/1Wrote down how I split work between agents2026-10-22T00:00:00Z2026-10-22T00:00:00ZClaude Code mainly implements. Codex mainly refines writing, researches, and configures the machine. It is a habit, not a rule.tag:greglum.com,2026-10-20:learn-log/1Four problems caught before one release2026-10-20T00:00:00Z2026-10-20T00:00:00ZAn agent-built Draftroom change passed the local gate. Reviewers, a phone screenshot, and CI still caught four problems before merge.tag:greglum.com,2026-10-18:learn-log/1Added How I Build to this site2026-10-18T00:00:00Z2026-10-18T00:00:00ZMy Claude Code and Codex setup, shared rules, skills, and checks, on one page.tag:greglum.com,2026-10-16:learn-log/1Gave the site a written point of view2026-10-16T00:00:00Z2026-10-16T00:00:00ZAfter Katie Dill's talk on "zombie UI", the design spec gained a point of view and principles that agents follow.