A term earns an entry here when a lesson defines it, and the entry links back to that definition. Six of the twelve lessons are written, so this page currently holds their terms and nothing else. It grows as each remaining lesson is published.
A to C
Agent — a model wrapped in a program that can act: open files, edit them, run commands, fetch pages. The model proposes, the program executes. Chat versus agent
AI coding tool — a program that can read your project files, write code into them and, in some forms, run commands. Different from a chat window you paste code into. What this can and cannot do yet
Allowlist — a standing approval for a named action so it stops asking. A decision about every future case, not this one. Chat versus agent
Cached prefix — the opening stretch of a request a vendor recognises from a recent one and re-reads at a fraction of the input price. Change anything inside it and it is gone. Tokens and false economies
Compaction — a product summarizing older turns of a conversation into a shorter form to stay under the context limit. Selective compression, not lossless memory. The context window
Context window — the text a model can see while producing its next answer, cleared when the session ends. The context window
D to S
Distractor — material that looks relevant but is not, competing with the fact you needed. Harder to ignore the closer it sits to the truth. The context window
Effective context — the part of what you sent that the model actually uses well. Always smaller than the part it accepted. The context window
Effort — how much work a model spends on one request: thinking time, tool calls, how much it explains. Set separately from which model you chose. Models and effort
False economy — a saving that is real on the line you measured and larger somewhere you did not look. Tokens and false economies
Input and output tokens — text sent versus text produced, billed at different rates. Output costs several times more. What this costs
Knowledge cutoff — the date past which a model has no reliable knowledge of the world. Two models sold side by side can be more than a year apart. Models and effort
Model family — several models from one vendor at different points on capability, speed and price, sharing one interface, so switching is a setting not a migration. Models and effort
Permission prompt — the product asking before an action it does not auto-approve. Answering once is not the same as allowlisting. Chat versus agent
Prompt caching — a discount on an opening the model has seen recently, re-read at a fraction of the normal input price for a short window. What this costs
Prompt injection — instructions hidden in content the model reads, written to make it act against you. The reason fetched text is treated as untrusted. Chat versus agent
Round trip — one full exchange: everything so far read in, one answer generated. The unit a retry costs. Tokens and false economies
Scope — the boundary of what you are asking for right now: what is included, and what is deliberately left out. What this can and cannot do yet
Session — one continuous conversation with the tool, from opening it to closing it. What it knows resets when you start a new one. What this can and cannot do yet
Solo-operator ceiling — the limit set by how much of a project one person can hold, decide about and verify, rather than by how fast code can be produced. What this can and cannot do yet
Spend limit — a ceiling you set on an account so a runaway loop stops rather than bills. What this costs
T to Z
Thinking tokens — the reasoning a model does before answering, billed as output whether or not any of it is shown to you. Models and effort
Token — the unit text is split into before a model reads it. One vendor puts it at about four characters of English, so 100 tokens run to 60 or 80 words (checked 10 August 2026). Context window sizes are quoted in tokens. The context window
Tokenizer — the step that splits text into tokens before a model reads it. The same text costs a different number of tokens on a newer model, and long words break into several. Tokens and false economies
Tool call — one request from a model to act, and the unit that permission is granted or refused on. Chat versus agent
Usage limit — the cap a subscription enforces inside a rolling window. It stops you and then resets; it does not bill you. What this costs
Usage pricing — paying by the volume of text instead of a flat fee, quoted per million tokens. There is no ceiling unless you set one. What this costs