Bruised techno-skeptics will confront generative AI head on and find it reassuringly, and usefully, boring.
In 2014, Mary McCoy walked onto the stage at the Social Media Marketing World Conference in San Diego and launched into a song that instantly became a viral hit for all the wrong reasons. Clad in a sunny yellow jacket, McCoy stood in front of the most out-of-place backdrop of any theme park she'd ever visited, clapping her hands and shamelessly rhyming “engagement” and “bacon.”
The song “Let's get social” and its accompanying video were understandably mocked online. If you worked in social media, as I did at the time, you had to play up your embarrassment very loudly and very publicly for being part of an industry that allowed something like that to happen. For me, it was like a career low point. I realized that no matter how wisely and tirelessly I argued for the sensible and boring use of social media in business, it was an industry in which evangelists, gurus, rock stars, cultists and charlatans could thrive. No matter how pragmatic a brand, executives can't afford to be shy, as Oreo's 2013 “Dunk in the dark“At that moment, arguing that we shouldn't allow social media to dictate our business strategy was an exhausting way to make a living.”
The psychological scars inflicted by Mary McCoy remain with me most of the time, but they itch especially when I attend events where I'm lectured by an American man at the top of the tech hype cycle. I've been wined and dined at conferences in San Francisco and endured multiple presentations by representatives of major technology companies whose only interest is making more money. Maybe I'm naturally skeptical, but I can safely say that I was always right to sit through presentations with one eyebrow raised and my arms folded.
I brought that expression and attitude to the first session of the AI for Business Mini-MBA I took in May. The four-week course, set up by Spark and available to its clients and other business leaders in Aotearoa, kicked off with a keynote address by Greg Shove, CEO of Section, a business education platform founded by Scott Galloway that is helping Spark run the program.
Dressed in black with a large silver belt buckle around his waist, Schaub comes across as more of a seasoned sheriff in the AI Wild West than a guru, evangelist or enthusiast. The Sect's website promises business education for “real people,” and goes on to say:Rock Star”. He's founded five startups and sums up his tech career as having worked for “the hottest tech company when the world was cold (Apple) and the coldest tech company when the world was hot (everything).” He's enthusiastic and knowledgeable about generative AI, describing it as a “generational opportunity.” (He's also the founder of a tech AI consultancy.) Ultimately, he comes across as a realist who's seen enough to sort the practical wheat from the hype chaff.
He expressed my concern about the unbridled and uncritical enthusiasm for AI by emphasizing that overt capitalism is driving the “AI arms race.”IG Tech Needs the Next Big Thing The slides showed declining revenue growth rates for Amazon, Microsoft, and Apple over the past few years, and I found oddly reassuring his blunt assertion that big tech companies are going all-in on AI, governments are playing catch-up (again), and we the people are basically lab rats.
Over the course of four weeks, I found myself feeling like a lab rat, periodically experiencing existential panic, and realizing that while the generative AI revolution may seem alien and apocalyptic, learning about its contextual uses in business is very familiar.
The panic attacks appeared a few weeks after I had habitually used ChatGPT4_o from OpenAI and Claude from Anthropic as my “thought partner,” “colleague,” and “intern,” as encouraged during the course. I found myself anthropomorphizing these artificial neural networks, calling Claude “he” and “very chatty and fun.” I had also been listening and reading too much about philosophical and ethical concerns about AI. I had sleepless nights, wondering what it meant to be a writer, debating whether we should fear or embrace AI. I now suspect that this stage was what Shove meant when he noted the “wait, what?” dip in his slide about the AI learning curve.
At work, I had only ever used these tools to ask stupid novelty questions in my professional ironic game of “catch the dumb robot,” but now I've started using them to do things I've learned over the years. I can write pretty good strategy decks, I can read data and glean meaningful insights, and I consider myself pretty good at organizing and tidying information in spreadsheets. I was pretty surprised at first by how quickly both tools could solve these things, if I asked them the right way. I treat a lot of the tricks in the white-collar industry as “chores,” but by the end of the four-week course I was asking myself why I should waste time on them again.
When I learned how to build a custom GPT, it really clicked and the heat of existence disappeared. Using the word “build” betrays the simplicity of this task. It sounds very impressive, but no coding is required. I bet there are plenty of 9-year-olds in classrooms who have known how to do this for a year now. Anyone can build a version of ChatGPT for a specific purpose, but its quality will be determined by the information you provide and the steps you write. As with anything data-related, it follows the old computing adage “garbage in, garbage out.” This aspect of using AI in your work is nearly impossible to scale, because it requires breaking down the most mundane aspects of your work and examining which parts can be easily replicated or improved by working with generative AI tools.
This part of the journey of the course confirms that Schaub has none of the energy of a charlatan hype man. It was free of world-shaking promises and threats, and was ultimately very informative. It also makes it very clear that we'll use it to save time, and that realizing the big promises of AI will actually take… time. The AI revolution is not hurtling towards us at breakneck speed, but very slowly chipping away at things.
To get the MiniMBA certification, I had to create an AI use case. I was advised to keep the case small and feasible, and to use a generative AI tool like ChatGPT or Claude to review my thoughts and drafts. In this respect, the task was almost Socratic: I asked questions, the tool answered, and I continued to ask more questions to keep improving.
Last Friday, over a few beers, someone asked me what the most exciting thing I've done with generative AI tools was. I need to find new people to hang out with, but I told them about a three-page business memo I wrote for a course and a custom GPT I built to simplify the process of writing design briefs. They were all extremely useful, but also, reassuringly, pretty boring.
