It all started with a spark. And a dream.
Hi. I'm Jared Sorensen, the founder of Provident Living Learning Center.
Some time ago I had an idea for a new software product and turned to AI for help building it. I was making good progress when the AI model started forgetting important concepts, referencing variables and functions that did not exist, and taking the project in the wrong direction. I had seen this behavior before with other AI models and apps, and found it very frustrating.
As I struggled with the issues, it dawned on me that what I needed was an AI assistant that could never forget or hallucinate, would always follow the project specifications, and would continue to work even if AI models and technology changed.
One night during this process I had a very vivid dream that was undoubtedly triggered by my thoughts. In the dream I saw myself going on a long hike that would take many days and pass through thick forests, arid deserts, and rugged mountains. As an experienced hiker, I had brought along all the gear I would typically need, but I knew this journey might require skills I had not yet developed. But I had with me an unseen companion who was more experienced than I, though he had not hiked this particular trail before.
At one point in the desert, the trail ahead appeared so steep that I was tempted to give up, but my companion helped me understand that it was just a desert mirage, and encouraged me to go on.
Later in a mountainous area, I came to a cliff that was very real and beyond my ability to climb. My companion was able to ascend the cliff and then lower a rope to help me climb up.
Eventually my companion and I reached our destination. But as the dream ended I realized that more important than achieving the goal were the skills and confidence I had developed along the way. My companion, too, had benefited.
Going back to AI, I mentioned the dream, and it led to a series of long discussions about what could be done to make AI more reliable while allowing the user to maintain full control and develop new abilities. We came up with a solution based on a self-maintaining library of philosophies, standards, procedures, history, and project specifications. This library was securely stored outside the AI's conversational memory and reviewed at the beginning of each conversation.
The library we eventually developed was highly complex but solid, with strict rules for usage and updates. Only after developing the library did we begin creating the actual product.
This methodology left me completely in charge and prevented the AI from straying. The AI would often make constructive suggestions but could not take over or do anything without ensuring that I understood and okayed the process. The workload was divided logically, leaving me with the creative tasks while the AI did the mundane work. My dream became a foundational element, helping define the roles of the user and the AI companion. And the product essentially developed itself.
I soon realized that AI and I had not only built what I needed, but in the process we had created a platform that anyone could use with any AI model to develop any product. Changes in technology could be implemented with little or no adaptation, and the method should continue to work for many years to come.
So OpzorAI was born. It is currently being developed as a commercial product, and it is built on its own methodology. So far the results have been astounding, and we look forward to completing it and making it publicly available.
If you're interested in participating in this exciting project, I encourage you to check out the available opportunities or contact us.