Welcome to the 19 February 2024 release of Gptcmd!
New features
* In place of "OK", Gptcmd now reports more descriptive messages to confirm the success of various actions across the application.
* A new command-line flag (`-m` or `--model`) has been added to switch to a specific OpenAI model on launch. For example, `gptcmd -m gpt-4-turbo-preview` will use the [new GPT-4 model](https://openai.com/blog/new-embedding-models-and-api-updates).
* Guidance on API parameter usage in the readme has been updated, especially concerning limiting output length when running GPT at high temperatures.
Bug fixes
* Fixed a crash in message thread deserialization.
* When loading a thread from disk with non-default API parameters, Gptcmd no longer always warns to save changes even when none have been made.
* Attempting to run `retry` in a thread containing only assistant messages no longer causes Gptcmd to crash.