Brass Tacks: What You Need to Get Started with Auto-GPT
Updated: May 3
This post is the second in the series about trying to get Auto-GPT to generate BigFix fixlet content for new vulnerabilities.
This post is about how to get the resources you need to get yourself set up with Auto-GPT. If you're not interested to do that: life is short, skip this article!
Resources
Here's a list of resources I have found helpful:
You'll need an OpenAI API Key. This key will be used for token charges ($.002 per 1000 tokens). Once you generate a key: copy it and save it somewhere safe because OpenAI will never show you that key again. (Don't worry, you can rotate keys if you lose it.)
You may want a Pinecone API Key. Pinecone is a vector database
If you want to use Auto-GPT to generate images, you'll need an API key for that as well. DALL-E is supported.
If you want Auto-GPT's AI to limit or filter its search: you can create a custom search engine on Google.
Auto-GPT on GitHub. You'll either clone the repo or you'll download the latest stable build and extract the .zip to get the Auto-GPT dependencies and files.
Auto-GPT on Discord this is an active community.
Auto-GPT on Reddit because sometimes you just need a meme.
How I Got Started
A former colleague and longtime friend Dale Hattaja heard me raving about how learning (and eventually teaching) AI prompting sparked a passion I hadn't felt for technology since BigFix. Since Dale and I share a BigFix connection, he mentioned Auto-GPT and suggested I check it out. I saw a demo on youtube - it's embedded in the first linked resource above - and that inspired me to get started. I probably watched that demo 5x at 50% speed to grasp the concept of to how Auto-GPT works. From there I searched for a good set of instructions (first link in the resources above) and got going.
Running on macOS Ventura
I'm using a 2018 Macbook Air thats been upgraded to Ventura. For the requirements, Python v3 was already installed with the OS. I added the GitHub Desktop App, and created a new OpenAI API just for Auto-GPT so I could track the spend.
I started by cloning the Auto-GPT repo but issues with it followed by a tour through the GitHub issues make me realize quickly that I'd do better running the stable version. So I...
went to GitHub and navigated to and downloaded the latest stable release of Auto-GPT
extracted it
opened Terminal from the path where Auto-GPT was extracted
installed the requirements
pip3 install -r requirements.txt