Tony Stark didn’t invent the idea of time travel. But in partnership with a machine, he figured out how to do it.
In one scene in Avengers: Endgame, Tony uses his living room as a lab, engaging in a back-and-forth with his AI assistant, Friday, as they work through scenarios projected in a floating hologram. He tells her, “Run a simulation with a Möbius strip… invert it.” The model stabilizes on the screen. Tony leans back in shock, whispering, “Sh—.” Behind him, his daughter Morgan—who has snuck out of bed and is sitting on the stairs—echoes Tony’s spicy language, then grins and blackmails him into promising ice cream in exchange for not repeating it in front of her mother.
It’s a lighthearted comic-book version of a scientific breakthrough, but it’s also instructive as a posture to adopt as we head back for another, AI-influenced school year. –Projected over his dining room table, Tony shapes the premise. He imagines the structure, defines the parameters, and sets the problem in motion. Friday does the computation—much, much faster than even a prodigious genius like Tony can. But while the machine can run scenarios, it can’t decide what’s worth modeling in the first place. No one thinks Tony’s cheating—we know who’s in creative control. But unlike Tony’s eccentric billionaire approach, our agency must come from something more grounded: protocols, routines, and a positive mindset we intentionally cultivate in ourselves and our students.
Learning to See Like Carpet Merchants
We don’t question Tony’s authorship because we can see the signature of his genius in the work. The Möbius strip wasn’t Friday’s idea—it was his. The machine accelerated the process, but the insight and the structure came from him. He handled the qualitative work; the machine handled the quantitative effort.
Learning to work with AI requires the same discernment—not just in how we use it, but in how we judge what it gives back. And for that, we might take a cue from Persian carpet merchants.
To the untrained eye, a rug may look beautiful, symmetrical, flawless. But master appraisers know that artistic value isn’t found in mechanical symmetry. Machines can replicate patterns with mathematical precision but they can’t replicate an artisan’s judgment—those tiny, intentional irregularities that signal a human hand making aesthetic decisions in real time. Craftsmanship lives in the imperfections a machine would never allow.
Here are three routines practiced by Persian carpet experts when they examine a rug—and what each teaches us about holding onto the thread of creative control while enlisting a machine to run the calculus:
- Flip it over
Expert merchants turn the carpet over to inspect the back. In hand-knotted rugs, slight irregularities from knot to knot mark the choices of a human hand more than they signal error. Machine-made rugs, by contrast, show uniform gridwork and glued mesh—symmetry without style, precision without decision. Students and teachers should flip their AI responses on their backs and examine what holds them together. Read past the surface polish and feel for gridwork. If the language is too clean, the logic too pat, question the model’s assumptions, trace its sources. See if meaning and structure still hold up from beneath. - Inspect knot irregularities
Not every irregularity is a mistake. In hand-knotted carpets, slight inconsistencies from knot to knot don’t diminish the work—they reveal a signature. These are the lightly brushed fingerprints of the weaver: a shift in tension, a reinterpretation of shape, a choice to sacrifice precision for flair. In machine-made rugs, the pattern is perfect—but flat. Students and teachers alike need to learn to see that difference. It’s a skill—recognizing the signs of intention, including their own. Human thought leaves traces: hesitation, inflection, voice, friction. When AI-generated language flows too smoothly, it often means no one wrestled with the threads of thought. The ideas weren’t written—they were printed. - Trace the weave
Carpet merchants don’t just look—they touch. Their palms move across the backing, like reading poetry in Braille. A natural rug breathes: knotted wool, cotton, silk. Synthetic blends feel colder. Flatter.
Think of Saito in Inception. Pressed against the floor of an incepted apartment, being interrogated, he chuckles under his breath:
“I’ve always hated this carpet. It’s stained and frayed in such distinctive ways, but very definitely made of wool. Right now I’m lying on polyester. Which means I’m not lying on my carpet in my apartment… I’m still dreaming.”
He didn’t spot the simulation on visual inspection—he felt the dream collapse beneath his hands.
Students and teachers need to learn to approach AI writing the same way. On the surface, it may look fine. The sentences are clean, the logic neat. But you can’t just scan it—you have to read it out loud. That’s how we run our hands across the weave. That’s how we feel where something is off. Is there warmth, friction, the tug of an idea being threaded? Or is it cold and smooth, a polyester understanding stitched on an LLM assembly line?
Weaving Ourselves Into The Loop
This kind of discernment isn’t automatic. It takes training—intentional practice, repeated exposure, and reflection. With practice, like Saito, we’ll learn to distinguish between the artificial and the real.
Flipping the rug, tracing the weave, feeling where the material breathes—these metaphorical steps aren’t just about critiquing what AI produces. They’re about understanding our human output better.
It’s about authorship. Every time students and teachers pause to examine what makes an idea feel real—where the structure reveals cohesion, where the pattern carries style, where density supplies the kind of illustrative detail that gives ideas credibility—they’re not evaluating output. They’re noticing the qualities that make a piece of thinking recognizably human. Theirs.
Tony didn’t ask Friday to invent time travel, to solve the problem for him. He reimagined the structure—an inverted Möbius strip—and Friday ran the calculations. Tony set the loop, and together they tied the knot. It’s like when someone holds the loop of a bow steady while you pull the ends tight—the AI supports, but it’s by human hands that the bow comes together.
As a result, we learn to more deeply appreciate what we sound like in our own intellectual voice. Which turns of phrase we use naturally in conversation and writing. What they reveal about our beliefs and assumptions.
When teachers and students learn to divide the labor like Stark used Friday—using AI not to replace their thinking, but to enhance it—they’ll do more than work efficiently. By examining its outputs the way a merchant flips a rug, they’ll become more thoughtful, more deliberate. And in mastering both, they’ll learn to recognize quality work from above and below.
Statement of AI Use
AI tools were used during the development of this piece only to support brainstorming and explore options for language refinement. All associated research was conducted by the author. All conceptual development, structure, and phrasing are the author’s own.
Also by the author:
Ethical AI: Why the Monkey In the Middle Matters, TIE Online
Modern Learning Lab 