Jack's Blog

AI Dream Analysis

Reflecting upon 2024, one of the greatest challenges I faced was poor sleep quality. Facing burnout at work, I would often fall into bed as soon as I got home, and my lack of a wind-down routine was causing me to wake up repeatedly throughout the night. While the cycle that formed was mostly negative, a silver lining surfaced: Since I would often wake up mid-REM cycle, I started having almost total recall of my dreams and formed a habit to write them down.

While toying around with Custom GPTs, I came across "Dream Decoder", a GPT that draws upon psychology and symbolism to help analyze dreams. At first, I thought the GPT was fun to toy around with, but I never thought that dream analysis would become one of my favorite generative AI use cases. I could not have been more wrong.

While the Dream Decoder GPT does a decent job analyzing the dreams without context, I estimate that using that approach, I only gleaned new insights ~25% of the time. Here, I define a "new insight" as any realization I made about my emotions, inner state, or patterns of behavior as a result of AI dream analysis that I did not first make on my own.

After a few weeks, I realized that providing personal context improves performance significantly. Currently, when using AI for dream analysis, I glean new insights about my subconscious state closer to ~60% of the time.

My approach is something like:

Without getting too personal, in the past two weeks alone, I have gained insight into:

I also discovered:

If you liked this, please share it! In the future, I expect that technologies will help us better harvest insights from our dreams so that we can live happier, more fulfilling lives. For the moment, the approach detailed here has served me quite well, and I think it will serve you, too.

Happy Dreaming!