Under the hood (just enough)
What an LLM is (and isn't)
A large language model, given all the text so far, predicts the next token. Repeated thousands of times, that produces fluent output. It is not a database, not a search engine, and not a colleague who remembers yesterday.
Three consequences you will feel on day one
| It is… | So… | What you do |
|---|---|---|
| Memoryless between calls. Every reply is generated from scratch from whatever text is sent this time. | "Memory" features are the harness re-sending saved text. | Put durable material in a project or a file. |
| Optimized for plausibility, not truth. It produces a confident answer whether or not one exists. | This is the root of hallucination. | Give it the source. Tell it "I don't know" is fine. Verify. |
| A language engine, not a calculator or a database. | Arithmetic, dates, counts, and citations are exactly where it slips. | Route those to a tool or check them by hand. |