My experiment with ChatGPT began in early 2023, with humble tasks like summarizing meetings, drafting emails, and researching topics. It was a useful tool, but I didn't feel like I was unlocking its full potential - it was more of an enhanced assistant than a game-changer. That all changed when I started feeding it more context: mission statements, goals, strategies, competitive research, and product information. The outputs became less generic and more insightful, allowing me to use AI to think through strategic decisions in a way that would have taken a human days to compile.
As I delved deeper, my needs evolved. I wanted more than just a chat window; I needed a system that could monitor information, perform research, and work on projects over extended periods. I also wanted it to operate independently of my primary computer, where I could give it more freedom without worrying about potential disruptions. This was the catalyst for building my home AI lab.
I had a decent GPU (RTX 3090) leftover from my crypto-mining days, along with enough spare parts to assemble another computer without breaking the bank. I installed Linux, as the agent framework I wanted to try, Hermes, didn't support Windows at the time. Then, I used Claude Chat, Claude Code, and Hermes to get everything up and running. The process was slow and not exactly elegant - I spent a lot of time moving instructions between tools - but I learned how the pieces worked together and what I actually needed from the system.
The first project that made the effort worthwhile was a competitive-intelligence monitor for my day job at a Pacific Northwest credit union. Every Sunday night, it reviewed competitors' websites, social channels, press releases, and hiring activity, analyzing their products, positioning, and changes since the previous week. Before this, competitive intelligence was a manual or outsourced task that quickly became outdated. The new system kept running, providing me and other leaders with up-to-date information without the need for periodic research projects. It still does today.
This early success led me to explore turning these ideas into products. I worked on productizing the competitive-intelligence system and built a job finder for friends and family. However, I soon realized that the major AI platforms were rapidly adding similar capabilities, making it challenging to differentiate my solutions. I've written more about what happened to those product ideas, but the main lesson was that building software is no longer the difficult part - the challenge now lies in finding a problem that the platforms won't absorb, differentiating your solution, and out-marketing the competition.
My setup has continued to evolve. I eventually returned to ChatGPT and gave Codex the necessary permissions to work directly on the Linux box, streamlining the process. Now, I can describe what I want in the morning, let it work, and come home to a prototype - the original goal I had for Hermes. ChatGPT and Codex are my primary tools today, although I still use Perplexity for research and haven't completely given up on Claude. I jump between these platforms, depending on the task, and I'm less interested in committing to one tool than in maintaining the ability to experiment with all of them.
One thing I've come to realize is that these tools don't eliminate the need for human experience. With roughly 25 years in marketing and business leadership under my belt, I can usually recognize when an AI-generated answer sounds plausible but is incomplete or wrong. I know which assumptions to question and where business reality is more complicated than the output suggests. An earlier version of me wouldn't have been able to use these systems in the same way.
The biggest change in how I think about the lab is that I'm less focused on building software products and more focused on using AI to solve business problems. Sometimes that means creating an internal tool, and other times it means helping a business leader understand what AI could do for their team or process. Over time, it may lead me into consulting, helping multiple organizations work through these questions.
I'm aware of the risks AI poses for jobs, the economy, and society, but I don't think it's going away. Putting my head in the sand doesn't seem like a good plan. The lab gives me a place to keep learning while the technology, hype, and business models sort themselves out. For now, I can learn about something in the morning and find out whether it's actually useful before dinner. That's enough reason to keep the machine running.