A large language model is a model trained to produce plausible, well-formed text in response to a prompt. It is well suited to language tasks such as summarizing, explaining, and drafting or checking code, and it has no special mechanism for seeing future market outcomes.
How it works
The model generates text by predicting what is likely to follow a prompt, which makes its output fluent and confident.
Its reliability depends on the task: text-in, text-out work can be spot-checked against a source, while predictions about future prices cannot.
Why it matters
Fluency is not accuracy. A confident, well-written answer can be entirely invented, which is why outputs must be verified.
Used for language tasks and verified against data, a language model speeds up research without being a source of truth.
A simple market example
A trader asks a language model to summarize an earnings call. The summary condenses the key points and can be checked against the transcript. Asking the same model for tomorrow's price is a different category, one it has no mechanism to support.
Common mistakes
Treating a fluent answer as evidence the claim is true.
Using the model for price prediction, which is outside what it is built to do.
Frequently asked questions
Is a language model good at predicting prices?
No. It is trained to produce text, not to foresee market outcomes, regardless of how confident the output sounds.
How do I use it safely?
Use it for language tasks and verify every output against the underlying data before relying on it.
Does a confident answer mean it is right?
No. Confidence is a property of the writing, not evidence of correctness.
Educational content only. Definitions describe common market usage and may vary by jurisdiction, instrument, or institution.