LLM - for command line LLM prototyping

LLM - for command line LLM prototyping

Table of Contents

I have been doing a lot of exploration and prototyping for “The Intelligent Enterprise” (tm mine obvi) using LLMs. One technique I’d love to share with y’all is using llm - a command line package - to do this.

Here’s the rough setup:

brew install llm

# Paste your OpenAI API key into this
llm keys set openai

# Run a prompt
llm "Ten fun names for a pet pelican"

# Run a system prompt against a file
cat myfile.py | llm -s "Explain this code"

Now comes the fun part. I’ve got a nice setup in vs code so I can quickly run the last command in the integrated terminal.

So, what I do is take the system prompt and put it in one file (e.g. system-prompt.txt) and take the data (e.g. data.txt) that I’m working up and put it in a second file. I can simply use the cat program to cat those files together and pipe them into LLM.

cat ~/system-prompt.txt ~/data.txt | llm

hotkey to change back to the prompt, edit, rerun. repeat!

Related Posts

The true value of startups: Network and learning

The true value of startups: Network and learning

After leaving my last startup gig at Opto, I got a text from two of my former colleagues that reminded me why the real value of working at a startup goes way …

Read More
Man/Machine/Marketing - applying lessons from DevOps and Finance product development to digital marketing

Man/Machine/Marketing - applying lessons from DevOps and Finance product development to digital marketing

Man/Machine/Marketing - applying lessons from DevOps and Finance product development to digital marketing

Read More

Get new posts via email

Intuit Mailchimp

Copyright 2024-infinity, Paul Pereyda Karayan. Design by Zeon Studio