Welcome to the Machine
In the spring of 2024 I created an obituary for my father. He took ill rather suddenly, his final weeks were a whirlwind, so after his passing we decided to hold a gathering for friends and loved ones who had been shocked by his rapid demise. We planned a celebration of life to take place in his hometown, at his favorite venue, on what would have been his 76th birthday.
A little-known aspect of my father that I found particularly pleasing was his penchant for writing poetry. He deeply felt the wonder and beauty in life and chose to express his appreciation with poems. He wasn’t prolific or published, but when I was honored with the task of writing his obituary, I felt it appropriate to do so poetically. Well… not right away. Let me back up a bit.
I began by going to Legacy.com. I was vaguely aware that they had tools to help with writing obituaries, certainly forms for entering vital statistics and whatnot. This was the case and, once I completed the data entry, the site presented style options for generating the content. This piqued my interest. Being a curious tinkerer, I ordered up an AI-composed draft in each available style. This proved enlightening but, looking back, it was purely academic, the poetic style would be my obvious choice regardless of the other offerings.
Since each draft was different—I clicked up a dozen in the poetic style. Apparently this approach is acceptable; I mean, CTRL+A, CTRL+C worked—without triggering any FBI prosecution threats. Confident in the legitimacy of my endeavor, and having plopped my dozen drafts into OneDrive, I began stitching together the summary of my father’s life, intent on making it worthy of the life he loved.
This is where having a dozen different drafts proved quite useful. I was able to pick and choose bits and pieces. Keeping language I liked such as: “In the quiet repose of twilight, the world softly whispers farewell…” while ensuring the poetic prose authentically reflected my father’s personality. I was able to weave his interests together in phrases like: “His heart beat in rhythm with the crack of the bat and the gentle arc of a ball in flight” and ”With a sportsman’s spirit and a craftsman’s grace, he walked the fields and fairways of life”.
The resulting obituary garnered a good bit of attention and, in the face of all the compliments, I fessed up to using AI. Surprisingly, that didn’t diminish the praise from well-wishers at my dad’s final birthday party. If anything, many seemed more impressed.
In the year that followed I often thought of this experience. I marveled at how, given a specific goal, AI was able to help me achieve something more meaningful than I could have alone. It seems odd, being a technology specialist and all, that creating the obituary was my first straight-up use of AI and that I had stumbled—accidentally—into a glimpse of the powerful potential of its collaboration.
One year later
By “collaboration,” I don’t count Alexa’s maintaining of our grocery list or her ability to serve up my ‘Princess Lullaby’ playlist on demand. I wouldn’t be surprised if most people don’t even consider that an interaction with AI. Moreover, Legacy.com’s training-wheel interface, designed for grieving loved ones, was collaborative-esque, but I was curious: what could happen with the training wheels removed?
However, there is some fear-mongering surrounding the use of AI. Certainly not, as far as I’ve heard, related to Google Maps recalculating routes in response to wrong turns, but what about all those stories of “cheating” on homework? Or, something closer to home—AI taking over all the development work. I smirked. But should we worry about SkyNet deciding it’d be better off without humans? The impressiveness of my earlier AI-collaboration overshadowed these vague concerns and I decided to forge ahead.
I didn’t have a specific task this time, so I tried to imagine something that would benefit from an intelligent, directed interaction with AI—specifically, ChatGPT. Clicking the link provided by Google, I landed at a clean, pristine, and refreshingly minimalist interface. The label above the lone input control read “What’s on your mind today?”. Looking at that now, it reminds me of Facebook’s post prompt. Hold on, I’ve got to check… Wow, it’s nearly identical. I wonder if that phrase has some statistical significance for drawing people in. (Spoiler: it does).
I offered my thoughts as:
Just want to get familiar with you and begin a productive working relationship
The response from AI was immediate and uninspiringly generic. Was I expecting something poetic instead of something along the lines of: I’d like to help. What do you want to do? No, in hindsight, the response was clearly an enticement to draw me in—tempting me to provide more information. It was as if the AI thought I was beating around the bush and wanted to avoid having to pull teeth in order to get started. Game to venture forth, I replied:
I love the idea of journaling but can’t seem to stay consistent. Do you have any ideas on how I can maintain an active journal?
Thus, it began. I thought I was being clever—putting the onus on ChatGPT by asking if it had any ideas. [Snicker, snicker]. The breadth of suggestions stunned me. It was suddenly obvious that any productivity hindrances would be on me. I stopped laughing. The illusion of producing via productivity discussions vanished, replaced by the realization that I’d have to take action to make progress.
This productive working relationship I was after would require focused attention on my part. I’d need to craft carefully constructed questions since each prompt produced such a breadth of information from the AI. I relished the thought of discovering gateways in those responses—deep rabbit holes of related, even if slightly tangential, inquiry. For example, the AI’s reply included this suggestion:
- Try Different Formats
- Examples: Bullet points, mind maps, or sketches.
- Why: Variety keeps it engaging and creative.
Recovering slowly from the stark realization of my role in this relationship, I acknowledged ignorance while embracing ambition and posed the follow-up question:
What are mind maps?
I really should have expected this new can of worms. I was reeling since that was only the third prompt I’d submitted. My side of the conversation comprised fewer than 50 words—half certainly superfluous—yet ChatGPT provided multiple pages in response—information well-organized and easy to consume. It included graphs illustrating the inherent complexities of mind map connections, explanations of how each node and branch could inspire journal entries, an outline of branches and sub-branches, and—most importantly—a question:
Would you like help creating a mind map for your journaling topics?
Insidious! Exactly the enticement I should have expected. By now, I was completely intrigued. ChatGPT’s initial response alone was enough to level up my journaling (assuming I followed the suggestions), but I wanted to see how deep I could go. I focused on a single datum—a tiny drop in the ocean of information—and pursued the creation of a mind map.
I was enthralled with the idea of harnessing the torrent of data. It was spread out before me, literally at my fingertips, awaiting a few choice prompts to produce more. But, recognizing my inexperience, I decided to seek the wisdom of a guide who might know how to best craft them—I asked the AI.
A bit meta, sure, but I was doubtlessly employing the services of a Terrisman Keeper with infinite metal minds—there’s a rabbit hole for you to follow… The last time we’d worked together, we were stitching stanzas to farewell my father. Now, I was mapping an adventure—one that was bound to be more practical (but would get stranger) from here.