Insights

Hot take: Why perfectly optimized brands are leaving us perfectly bored

AI can build your assets, but it can’t build you a hit.

hot take blog cover

There’s a very specific feeling that occurs roughly four seconds before you decide to buy something you didn’t know you needed. It isn’t a data point. It isn’t an engagement metric or a conversion flow. It’s instead an ethereal swell; a fleeting, internal nudge; a feeling. And this flutter of the spirit is precisely what AI gets wrong.

Strauss Zelnick, CEO of Take-Two and overseer of the sprawling, billion-dollar hit-factory that is Grand Theft Auto, recently laid down some hard truths about AI. He argues that AI can build your assets, but it can’t build you a hit.

“Speed isn’t the issue,” Zelnick noted, with the taught patience of a man who has seen a thousand disruptive technologies come and go. And he’s right. We’re obsessed with (and distracted by) raw speed and output. We want our logos in seconds, our copy in milliseconds, our campaigns generated by a prompt before we’ve even finished concepting the concept. But speed is just a measurement of friction. It doesn't tell you if the thing you’ve produced is actually worth the time it took to render.

At TinyWins, we’ve always believed that the most powerful brands aren't built on the back of sheer velocity, but on the back of feelings. We see AI for what it is, a spectacular tool. It’s a hyper-efficient hammer, a paintbrush that never dries out. But a paintbrush doesn't decide to paint the Mona Lisa. It doesn't feel those ethereal swells that move a warm-blooded artist to take a risk, to zig when the data says zag, to find the soul in say, a pair of sneakers.*

Zelnick’s most profound point was about the nature of surprise. “All hits, by their very nature, are unexpected,” he said. This is the central paradox that the LLMs can’t solve. An AI, by its very definition, is a backward-looking machine. It looks at everything that has ever existed and spits out the most statistically probable version of whatever it is you’re doing. But as Zelnick points out, “Things that are data-driven in their entirety can’t be unexpected.”

If you follow the data to its logical conclusion, you end up with a perfectly optimized, entirely forgettable average. You end up with content rather than culture. You end up with a product that looks like a brand, sounds like a brand, but leaves the heart entirely unmoved.

To create a hit — whether it’s a world-conquering video game or a brand identity that sticks to your ribs — you have to be willing to get weird. You have to be willing to use your brain. You have to lead with those complex, messy feelings that a machine would classify as noise but we recognize as truth.

At the end of the day, AI might give us the flat-pack furniture and the pre-painted walls, but it won’t give us the feeling of a home that’s been lived in. It can build the world, but it can't compel us to exist in it. For that, you still need humans. You still need unpredictable, spontaneous, day-dreamy, voyeuristic, uncanny, flawed humans. They’re the only technology guaranteed to make you feel something you didn't expect to feel. They are the hit-makers.

*See “REEBOKS LET U.B.U.” campaign by Chiat\Day, 1988