)
I learn most efficiently in a group setting or with a partner. I also enjoy learning through pain and suffering, which the Marine Corps taught me is the most effective way to make a new type of thinking stick. Since COVID, I have worked remotely and have encountered challenges when collaborating efficiently to learn new topics or tackle new problems.
When ChatGPT and other AI products emerged, I took advantage of the co-pilot assistance to learn new topics and methodologies in security. One of my most recent projects was building my home security lab, a task I procrastinated on because I prefer having a mentor to help me. I enjoy building PCs, so I always have spare parts lying around. Finally, I decided to build a home server, using ChatGPT to guide me through installing Proxmox and setting up multiple VMs for various tasks, including Kali Linux, Ubuntu, and Windows.
The ChatGPT Experience
This is where I knocked my head against the keyboard more than I would have liked. ChatGPT is excellent for providing methods and troubleshooting, but it doesn’t always offer the most efficient approach. For example, I wanted to install MISP, an open-source CTI platform, on my Ubuntu server. ChatGPT walked me through installing Apache, MariaDB, and then MISP separately. This led to hours of troubleshooting each service as I copy-pasted issues into ChatGPT, which sent me on a wild goose chase. Eventually, I turned to trusty ole Reddit for others' experiences and solutions.
Finally, I referenced the official documentation for MISP. I discovered that the Ubuntu (24) version wasn’t fully compatible with MISP yet, so I rolled back to version 22, which MISP offers an express install for on their GitHub. After reinstalling Ubuntu, I had MISP running correctly in about 15 minutes. While I appreciated the IT experience of setting up the different services, it was a waste of time and delayed my goal of creating a CTI feed guide for MISP users by hours.
Learning from the Experience
Like others, I fell victim to not checking references and simply believing what ChatGPT told me. One of the gaps in ChatGPT is that it doesn’t provide references for the documentation it recommends unless you ask for them. (Figure 1.) 
Figure 1: No References in ChatGPT
This is where our generative AI, IQ, stands out. At Cybersixgill, our users always check the original source and content of every piece of intel we alert them to. With IQ, we’ve shifted into overdrive to help gain intelligence faster than ever, constantly referencing the source so users can validate the intel for themselves. Cybersixgill’s IQ will reference all the intel it gathered and provide links to the original source for users to validate the findings. (Figure 2)
The Advantage of IQ
There is little to no room for security mistakes—you must get it right the first time. Accurate and timely intel is crucial, and we can’t afford to hallucinate AI providing inaccurate responses without references. Our users have learned they can trust IQ because it will tell you so if it doesn’t have the answer.
)
)
)
)