26 Oct AI is Coming
Don’t be Scared! | by Mark Cooke
Anthropologists have a word for the uncomfortable in-between, the neither here nor there, the not like it was but also not what it’s going to be: this is the liminal period, or liminality. Victor Turner, who made the concept famous, called it being “betwixt and between.” You’re welcome – that might show up in the crossword some day. But that’s where AI has taken us, and that’s what is causing all the anxiety. It’s here and can’t be ignored, but most of us are not quite sure what to do with it.
The good news: we have been here before, and there is a pattern to it. In 1972, two paleontologists, Niles Eldredge and Stephen Jay Gould, looked hard at the fossil record and noticed what the textbooks glossed over. Species don’t change gradually the way we like to imagine. They sit in stasis – sometimes millions of years of nothing much happening – and then, in a geological blink, everything changes at once. They called it “punctuated equilibrium”: long stretches of stability, punctuated by bursts of rapid transformation (again, you’re welcome, it’s good for trivia night).
Sound familiar? It should, because economies and cultures behave similarly. The printing press, the steam engine, electricity, the internet. Each was a punctuation: arriving after a long plateau, rearranging how we work and live, then settling into an equilibrium looking nothing like the old one. Artificial intelligence is the next one, and I would argue we are standing in the liminal burst right now. Economics and cultural behaviors are shifting, and pretending otherwise is how institutions end up in the fossil record. But – and here is the subtitle of this article doing its job – don’t be scared. What comes through these bursts are the ones prepared to adapt, and I want to share some advice.
The history of AI is older than most people think, and knowing how we arrived in this liminal moment takes away a lot of the mystique. In 1950 Alan Turing asked, quite plainly, “Can machines think?” The decades that followed tried to answer him using logical rules. Engineers hand-wrote “if this, then that” thousands of times over. It was brittle work. The machines could play checkers but couldn’t tell a cat from a dog, and the field famously froze through a couple of “AI winters” when the funding and the faith both ran dry.
What thawed it out wasn’t a clever new rule. It was data. While the rule-writers struggled, the world quietly digitized everything, and the statisticians moved in. Statistics matured into data science (my beloved era – as some of you have endured my past articles), and data science asked a better question: instead of telling the machine the rules, what if we showed it enough examples to derive them itself?
That question led to neural networks – software loosely inspired by the connected neurons in our heads. Feed a network enough examples and it learns the pattern, no hand-written logic required. Add enormous data and computing power and those networks got deep, then got good at the hardest pattern of all: human language. Natural language processing taught them to read and write it, and the culmination is the natural language interface – NLI, one more acronym for your collection.
In plain terms: for seventy years, humans learned the computer’s language. Punch cards, then code, then menus and clicks and dropdowns. Now, for the first time, the computer has learned ours. You describe what you want in ordinary English, and the software does it. That reversal is the punctuation, as significant as the moment we stopped needing a specialist to run the printing press.
So what changes? Economically, the same thing as every prior burst: the work gets restructured. Tasks that were expensive become cheap, roles reorganize around judgment rather than repetition, and new occupations appear that we don’t have names for yet. Culturally, how we write, learn, research, and decide is transforming in real time – ask anyone with a teenager. None of this is a reason for fear. Liminality is a call for intention. Quality by Design, you might say.
Which brings me to the advice portion, and it is probably not what you expect from a technologist. The best practice for mastering these tools is not another software certification. It is to go back to art, literature, poetry, history, and science. I am completely serious (and thrilled).
Here is why. When the interface is human language, the quality of what you get out is bounded by what you can imagine and express. The machine supplies the horsepower; you supply the destination and its description. Literature stocks the mind with ideas and metaphors worth reaching for. Poetry teaches precision – saying exactly what you mean in the fewest, truest words, which is now a marketable skill. History gives you pattern recognition across centuries, so you can tell a genuine punctuation from a fad. Science teaches you to frame a hypothesis, test it, and mistrust an answer that merely flatters you – a critical habit when the machine will happily generate confident nonsense.
The liberal arts, in other words, have quietly become the most practical technical training available. The people who will do amazing things with AI-enabled software are not the ones who can build the tool, but the ones who trained their minds to imagine clearly and express precisely, then let the tool close the gap between idea and artifact.
Every liminal period ends in a new equilibrium, and every prior punctuation rewarded the curious and the adaptable. This one will too. So no, don’t be scared.
Pick up a book. Read a poem. Study some history. Sharpen the oldest tools we have – imagination and language – because they just became the newest ones. And if you want to talk about any of it, track me down. I am easy to find.