Not smart.

“The fault, dear Brutus,” paraphrases Cassius, “is not in AI, but in ourselves.”

I’m speaking here, of course, about the way the AI industry is not only sticking with a losing playbook but repeatedly losing the plot.

Case in point: after Mark Zuckerberg's recent Today interview, you have to wonder if he's the one who whispered in the president's ear, "Let’s rebrand ‘artificial’ as ‘super-intelligence.’"

I dunno. Maybe the theory is that if you call the vegetables organic, people will swallow them whole.

But it seems a total non-starter when 73% of Americans poll as deeply concerned about a technology they fear can't be controlled.

I call it the persistent Frankenstein problem. You lab- build the thing, give it a prompt, and off it goes, Pac-Manning its way through jobs, energy, water, privacy, and, if the expert doomers have the odds right, all of us.

Then you double down by calling AI inevitable. Translation: You not only have no choice; you have no role.

Funny-not-funny: As telecom and healthcare have taught us, people value having a choice more than making one.  For them, having no choice is the denial of agency masquerading as progress.

Then there's Nvidia's Jensen Huang, telling Ezra Klein that AI development should accelerate to become safer. Intriguingly, he references the lives that would have been saved if the car industry had started out with built-in seatbelts, airbags, and ABS. Fair enough. 

Except for two flies in the insect-covered ointment:

Cars don’t evolve at the speed of Moore's Law, operate across global borders, or rewrite their own programming. No single car can break humanity.

Second, even the safest cars have been deliberately driven into crowds.

Common sense says that bad actors will inevitably turn AI to bad use. So maybe our Frankenstein will be able to beat their Frankenstein. But defending an ever-expanding surface of vulnerabilities is a bet, not a strategy. 

Especially if the bad guys only need to win once.

But instead of nattering on about abstruse-ities like open weights, compute thresholds, regulatory capture, and hallucinations, maybe we cut to the chase and ask how AI gets off the circular hamster wheel.

Being proudly simple-minded, I'd argue you blaze the path forward by building on common sense marketing principles:

- A value proposition is what people are buying, not what you think you’re selling.
- It's  not what you say. It's what people hear and conclude for themselves.
- You can’t persuade people to rethink what they already “know,” but you can give them the reasons to persuade themselves.
- Trust isn’t freely given. It's earned.
- While actions matter more, words are damned important.
- To shift a persistently dark narrative, build on winning points of light.

Obvs, none of this is rocket science. Although, considering the industry we’re talking about, maybe that's the problem.

But following a losing playbook is also a choice, and far from a winning one.

No matter how brilliant you are.