Shortcuts

If you're doing a lot of work with LLMs, then you already know that ... well ... it's a lot of work!

It's also a lot of typing and words (hey, it's in the name!).

So thoughts invariably turn to "how can this be more efficient?"

I am convinced that a macropad built specifically for prompting workflows would be a huge boon. But failing mustering the entrepreneurial energy required to make this happen, I'll probably resolve, instead, to just programming a regular one.

Some things you buy off the shelf or download:

Text macro utilities

If you're copying text from a web UI rather than working with APIs and scripts, you can use utilities such as: Autokey-GTK, xdotool, Espanso, Text Blaze and (on Windows) AutoHotKey, Breevy (etc, etc).

Obsidian has great support for hotkeys and macros.

One major advantage of running prompts via scripts is that you can just run the output straight into a folder. But this can come at a major downside: you miss out on the very useful contextual element of LLMs that happens after the first output when you can begin asking follow up questions.

There are also:

• Browser extensions for exporting text from web UIs like ChatGPT
• Voice control utilities for prompting with your voice whether the particular device you're using has native support for that or not