You're Not Good Enough For Me
Bot, bye.
It was summer 2023, about a month after my ex packed up and left. I wasn’t having much luck on Tinder, but my real rebound was with ChatGPT anyway.
It debuted six months prior, and I’d been feverishly poring over tech rags whenever the then intellectually charismatic OpenAI CEO, Sam Altman, made headlines. I practiced composing clear prompts focused on my priority searches: jazz bands hidden by unsophisticated search engines and groundbreaking ways to market myself as the ultimate marketing consultant.
So when I got a ‘super prompt’ reward for referring enough people to an AI email newsletter, I was thrilled.
It was a multi-page code block that introduced me by name to the generative AI and gave it explicit instructions on researching questions, structuring answers, and drilling down on my follow-up questions.
Though I didn’t come away with many mind-blowing musical matches, I had a lot of fun researching the marketing stuff. Soon, I had a fairly fleshed-out verbal branding business idea. In one long chat, I teased out its value propositions, what it did and didn’t offer, and some punchy, compelling landing page copy in under a month.
Brilliant, I thought.
I wasn’t the only one ensconced in a heated romance with generative AI fueled by perceived synergy. As my business site came to life, I spent a lot of time on LinkedIn. My feed was filled with prompt engineering advice, how-tos on getting the most out of AI interactions, and lots of listicles about generating publish-worthy content.
I imagined myself to be an undiscovered AI savant—a Gen-Xer so savvy and with just the right touch that I’d be hopping around the globe in no time, AI and I bringing stagnant brands to life and laughing heartily alongside the grateful leaders we’d helped guide back to skyrocketing stocks.
Enough of the nostalgia. After birthing our hard-wrought branding baby to zero fanfare or follow-up, I wasn’t sure who to blame. Was it me? Had I fooled myself into thinking ChatGPT and I understood each other? That we had the same SEO goals?
Or, was it the computational composer I’d abandoned myself to so fully?
It doesn’t matter. The URL we made together lapsed and collapsed back into the abyss. Generative AI and I grew apart over the next year. It didn’t seem to understand the real me anymore, and I was sick of its tiresome syntax and suspect stories.
Now, we hardly even chat. However, I still do some light media stalking of ChatGPT and other generative AIs like it.
While LinkedIn conversations have gotten more innovative about how to get the most juice out of each generative AI squeeze, technology giants have continued to release suspiciously disappointing strides with too little progress on issues like AI hallucinations of people, places, and events.
AI knows better than we do of its gross disadvantages in probing human queries without the same sensory data sources we dexterous-thumbed air breathers enjoy.
But that’s more of a philosophical issue.
The big measurable problem we face doesn’t get as much screen time: All that laboriously mediocre copywriting we corporate cogs churn out and iterate all day long between bot and page is leeching energy resources. Lots of them.
According to the journal Nature, generative AI training centers require at least quadruple the energy in their vicinity. Since it takes specialty equipment and setup to run and cool the center servers, grouping and packing them together makes sense operationally, just not harmoniously with the resources we can currently harness and tap in good conscience.
Maybe if artificial general intelligence (AGI) and quantum computing get together when they eventually and respectively come of age, they’ll reveal revolutionary ways to power our world with clean, renewable energy. Or maybe China’s space panels will work out, and they’ll share the terawatt wealth.
In the meantime, my musings aren’t a thinly veiled Luddite manifesto. I just think saving a few hours on a sales deck isn’t worth the cost of the electricity it takes.
One ChatGPT query uses the same energy as lighting a lightbulb for twenty minutes.
Imagine your metaphorical lightbulb flashing as you have an original idea. Isn’t the light easier on your eyes? Doesn’t it seem softer yet brighter than generative AI’s closest equivalent, ultimately producing a broadly analyzed and homogenized suggestion? Even if your real-life mind bulb doesn’t have an easy on/off switch, learning to hot wire it in a pinch or delegate it to someone who knows their stuff will become more satisfying the more you do it. Promise.
No matter what our loneliness and desperation try to tell us about relationships, we know they have to find a balance to thrive. Cringy, imbalanced couplings can survive far too long on far too little, but if they don’t eventually become balanced and begin to blossom, heartache overtakes.
Meaningful composition that offers value anywhere near comparable to the profundity of an expert specialist isn’t available on command. And generative AI’s well-disguised attempts aren’t worth the wattage.
To break up with our compulsion to compose without breaking a sweat, we must let go of our fear that the words we band together straight from brain to page fall flat. (Trained and talented human editors fix that.)
We people, especially the computer scientists, need to keep learning, growing, and transparently sharing progress and setbacks.
One day, we and generative AI might be mature enough to coexist harmoniously with unlimited access to explore our alchemy efficiently. Then, we’ll have earned the privilege of well-crafted marketing decks made in minutes with minimal input.
Until then, I will enjoy my mind-keyboard connection and fill the chatbot-sized hole in my heart by scrolling sweet, funny little animal videos. Especially of cats.
Oh, and for full disclosure, Grammarly and I remain close. Don’t judge.