A practical question about using Claude for price forecasting

A price project is using several kinds of past data to predict one target price. The main question is which is best for this kind of work.

There is also a practical choice to make: use Claude to produce the forecasts directly, or use Claude to write code for a proper model. The deeper issue is where an LLM is actually useful and where a normal model is the better tool.

Key points

  • The project uses historical input data to predict a target price.
  • The choice is about which to use for work.
  • One option is to ask Claude for forecasts directly.
  • Another option is to use Claude to write code for a separate model.
  • The core question is where an LLM helps versus where a normal model is needed.
Read original