I’ve spent more than 20 years in marketing and technology. I’ve lived through the dot-com bubble, the social media gold rush, big data, and blockchain’s promise to reinvent everything from banking to lettuce. Every one of those cycles followed the same arc: breathless promises, frantic investment, disappointment, and then, quietly, the actually useful stuff survived and changed how we work. Gartner calls this a Hype Cycle. And yeah, they are right.
AI is different in one important way. The useful stuff isn’t waiting at the end of the cycle. It’s already here, demonstrably, in my daily work and probably yours. And that’s exactly what makes this hype cycle so hard to navigate. The technology is REAL, and the hype is REAL, at the same time, and most of the loudest voices have a financial interest in you not being able to tell the difference.
So here’s my honest accounting. What I love, what I hate, and what I think we owe the people/generation coming after us.
The love is earned
I use AI every day. Not because a vendor told me to, but because it works. It compresses research that used to take a day into an hour. It gives me a thinking partner at 6 a.m. when no one else is awake. It is a brainstorm partner when I’m on a dog walk or in the car and I want to talk through an idea, presentation, or process. It handles the first 70 percent of a draft so I can spend my energy on the 30 percent that actually requires judgment.
Notice the framing, though. First 70 percent. Thinking partner. Compresses research. In every case where AI delivers real value for me, there’s a human in the loop who owns the outcome. That’s not a limitation of the technology. That’s the correct operating model for it. It’s the system I’ve purposely built to keep me in control.
AI is a tool. A pretty fabulous and remarkable one, maybe the most significant one of my career. But a tool doesn’t carry accountability. It doesn’t validate its own output, it doesn’t take responsibility when it’s wrong, and it doesn’t get fired when the campaign tanks or the numbers don’t hold up. People do. The organizations getting real value from AI right now are the ones that understand this: ethics, validation, and ownership of outcomes are not compliance overhead. They’re the whole system that makes the tool safe to use at speed.
And this is genuinely an inflection point for how work gets done. I truly believe this. I’ve been saying it for years, too. This is not about a feature release. This change is one that will take years to fully absorb and normalize. It will and has begun to reshape what entry-level jobs look like, what skills matter, and how prepared the next generation is for the careers they’re walking into. That question matters enough to me that I’m building something around it. Get AI Literate exists because I believe AI literacy, the judgment to use these tools well, is about to be as foundational as writing.
The hate is also earned
Here’s what I can’t stand: the hype machine wrapped around all of this.
We are being sold a story where every product is AI-powered, every job is about to vanish (though jobs are coming back…we are on a roller coaster for now), and every company without an AI strategy is already dead. Meanwhile, the measured reality is sobering. MIT researchers found that 95 percent of enterprise generative AI pilots are failing to deliver measurable returns. McKinsey’s own State of AI research shows that while 88 percent of organizations now use AI, only 39 percent can attribute any bottom-line impact to it at all. Gartner has generative AI sliding through its trough of disillusionment. None of that means the technology doesn’t work. It means the gap between the pitch and the practice is enormous, and the pitch is what’s setting budgets, headcount plans, and boardroom expectations. To me, this is not surprising.
I’ve watched this movie before. The hype cycle doesn’t just waste money. It wastes trust. It pushes leaders to deploy AI where it doesn’t belong, skip the human oversight that makes it safe, and then blame the technology when the predictable failure arrives. This is a typical change management crisis: when new technology appears so transformative, everyone puts the cart before the horse in an attempt to get an edge, only to forget how to implement it.
And beneath the hype, there are risks I think about that have nothing to do with quarterly disappointment.
Control of what comes next. The labs are openly racing toward AGI, artificial general intelligence, and beyond. Ask ten researchers to define AGI and you’ll get ten answers, which is itself the problem. We’re sprinting toward something we haven’t defined, with no agreed mechanism for controlling it in a meaningful way. Hundreds of the field’s own leaders signed a statement putting AI risk in the same sentence as pandemics and nuclear war. I don’t know how much weight to give the most extreme scenarios. I do know that “we’ll figure out control later” is not a plan.
Who we’re trusting. The honest question isn’t whether the AI labs are staffed by good people. Many of them are. It’s whether we should expect any company to consistently choose the right thing over growth when tens of billions of dollars and Silicon Valley’s entire status hierarchy are pulling the other way. History says incentives win. Self-regulation in a capital race is a bet I wouldn’t make with my own money, and we’re currently making it with everyone’s.
The environmental bill. This buildout has a physical footprint we’re only starting to price in. A UN University report projects AI could consume 945 terawatt-hours of electricity annually by 2030, roughly triple the combined usage of Pakistan, Bangladesh, and Nigeria, along with trillions of litres of water. Data centers are landing in American communities that are already seeing the effects in their utility bills and water systems. That cost is real whether or not the pilots deliver ROI.
The concentration risk. The US economy is leaning on this one sector to a degree that should make everyone nervous. Harvard economist Jason Furman calculated that investment in information processing equipment and software, about 4 percent of GDP, accounted for 92 percent of US GDP growth in the first half of 2025. Strip out the AI buildout and the economy grew 0.1 percent. When one bet is carrying the whole economy, the question isn’t just whether the bet pays off. It’s what happens to everything else if it doesn’t.
Holding both
So do I love AI or hate it? Yes.
I love what it does in the hands of a skilled person who stays accountable for the output. I hate the hype cycle that sells it as a replacement for judgment, and I take seriously the risks that the hype conveniently talks past.
Both things are true, and I’ve stopped trying to resolve the tension in my head and my heart. The tension is the point. The people who will navigate the next decade well are not the true believers or the doomers. They’re the ones who can hold “this tool is changing my work for the better” and “this industry needs guardrails, scrutiny, and accountability” in their heads at the same time. These are literal conversations I have with my 18yo son who is entering college who genuinely despises AI.
We’re at the start of a transition that will take years to play out, and the generation entering the workforce right now will live with the choices being made today, in labs and boardrooms they have no say in. The least we can do is teach them to use these tools with judgment, verify what the tools produce, and never confuse output with ownership.
That’s not anti-AI. That’s what taking AI seriously actually looks like.
There’s a related thread I’m pulling on next: what happens to quality when everyone uses the same tools to produce “good enough” at scale. More on that soon.
