The AI Learning Hub Journal

Trained Once, Used Millions of Times

Two completely different things people both call "AI"TRAINING — happens once, before you ever see itBillions of examples go inbooks, code, web pages, conversationsPredict, compare, adjustthe model guesses the next token, checks thereal one, and nudges its weights a fractionFinished weights come outa fixed file of numbers — that file is the modelTakes weeks or months of computeCosts millions and needs a datacentreDone by the lab, long before your chatUSING IT — happens every single time you askOne prompt goes inyour question, plus the chat so farThe frozen weights runone pass through the same unchanged numbersfor every token it writes back to youAn answer comes outthe weights are exactly as they were beforeTakes a second or twoCosts a fraction of a pennyHappens millions of times a dayCorrecting the model in a chat does not update it — it only adds text to that one conversationYour messages may still be stored, and may be used to help train a future version — that is a separate decision
Training builds the model, using it only reads the model — your chat does not teach it anything on the spot

Two Completely Different Phases

There is a moment when a model is built and a moment when it is used, and they have almost nothing in common. Training happens once, takes months of work, runs on enormous clusters of specialised chips, and costs an amount of money that only large organisations can spend. It produces a fixed set of numbers — the model's weights. After that, using the model is a comparatively small computation that happens every time anyone sends a message. When you chat, you are not training anything. You are running a finished artefact, the same frozen set of numbers everyone else is running.

  • Training: one-time, enormous, produces the weights that define the model
  • Inference: what happens on every single message, fast and comparatively cheap
  • Your conversation does not update the weights — the model does not "learn from you" mid-chat
  • Newer versions come from new training runs, not from the model quietly improving on its own

What "Learning From Data" Actually Means

It does not mean the model stored the internet somewhere and looks things up. Training adjusts billions of numerical weights so that the model gets better at one task: predicting the next token in real text. Statistical regularities in language get compressed into those weights — grammar, facts that appear consistently, the shape of a good argument, how code is structured, and also every bias and error that recurs in the source material. What survives compression is what was common and consistent. Rare details get blurred or lost, which is precisely where confident invention creeps in later.

  • The output of training is weights — numbers — not a stored copy of the training text
  • Frequently repeated, consistent information survives compression well; one-off details often do not
  • Patterns in the data become patterns in the output, including the unfair ones
  • A later stage tunes the model with human feedback so it answers helpfully rather than just continuing text

Knowledge Cutoffs and Why Memory Is a Product Feature

Because the weights were fixed at the end of training, a model's built-in knowledge stops at a certain point. Anything after that is invisible unless the product goes and fetches it — which is what happens when a tool searches the web, reads a file you upload, or pulls from a company's documents. Similarly, when an assistant "remembers" your name across sessions, that is the app storing text and quietly re-inserting it into the context window. Both memory and up-to-date knowledge are things built around the model, not properties of the model itself. Knowing the difference tells you what to trust.

  • Knowledge cutoff: the model has no built-in awareness of events after training ended
  • Search, file upload and retrieval add fresh information by putting it into the context window
  • Persistent "memory" is stored text replayed to the model, not the model recalling you
  • When accuracy on recent events matters, use a tool that cites sources and check them

Prefer slides, quizzes, and saved progress? Read this lesson in the library — free, no sign-up.