Cast your mind back to the early 2000s. You’re standing in a dimly lit, slightly sticky-floored bar. A song comes on. It’s an absolute banger, but you have no idea who sings it. Suddenly, your friend whips out their phone, taps a button, and holds it up to the speaker like Rafiki holding up Simba in The Lion King.
Ten seconds later, the screen proudly displays the artist and track name.
That was Shazam. And for most of us, it was the closest thing to actual magic we had ever witnessed a piece of plastic and glass perform. It was our gateway drug to machine intelligence. Shazam trained us to accept that software could “hear,” process, and match patterns better than our own brains. It wasn't true AI—it was just incredibly fast pattern recognition—but it perfectly softened us up for the technological tidal wave that was coming.
Fast forward to today, and we are drowning in that tidal wave. But if we peel back the marketing fluff, the reality of what AI currently is—and how people are misusing it—is actually pretty hilarious.
The Great “Glorified Word Processor” Irony
Let’s get one thing straight: people are expecting miracles from the wrong places.
Take generative AI. At its absolute core, beneath the sleek interfaces and the billions of parameters, a generic AI bot is essentially a hyper-caffeinated, insanely well-read version of your phone’s autocomplete. It was built to predict the next logical word in a sentence. It is a glorified word processor that accidentally got so massive it learned how to mimic logic.
Yet, how do people use these generic bots? They ask them to run their entire businesses, diagnose their weird rashes, and write production-level code for tasks they have no business touching. It is the digital equivalent of using a Swiss Army knife to perform open-heart surgery. Sure, it has a blade, but please don't.
The most powerful, world-changing AIs aren't the generic chatbots trying to do a hundred things decently. The real heavy hitters are the specialized AIs—the ones trained on hyper-specific data to fold proteins, optimize supply chains, or manage power grids. But the general public is too busy trying to get a generic AI to write poems about their cats to notice.
The Graveyard of Hype: Dot-Coms and EVs
So, will AI take over the world? It’s no longer a question of if AI will break through, but how. And if history has taught us anything, it’s that the path to a technological revolution is never a straight line up. It looks more like a rollercoaster designed by a lunatic.
Look at the Dot-Com Bubble in the late 90s. Everyone completely lost their minds. If you pitched a company called SellMeSocks investors would throw fifty million dollars at your face before you even finished your PowerPoint. When the bubble burst, investors lost their shirts. It looked like a massive failure. But under the ashes of that hype, the infrastructure was laid. The fiber optic cables were buried. The behavioral shift happened. Fast forward to today, and that reckless hype cycle gave us the foundation for modern e-commerce. We practically live online now.
Or look at Electric Vehicles (EVs). Ten years ago, we were promised a golden, emission-free utopia by next Tuesday. Everyone bought into the hype, but then reality hit: range anxiety, lack of charging infrastructure, and supply chain nightmares. The hype stalled. But look at where we are now. Thanks to the current global energy crisis and skyrocketing oil prices, EVs are getting a massive, pragmatic second wind. The hype faded, but the utility remained.
AI is going through the exact same cycle. We are currently at the peak of the "throw money at anything with 'AI' in the name" phase. Eventually, the hype will cool off, the useless generic startups will die out, and what will be left is the quiet, highly specialized infrastructure that actually runs the world.
The Shovel Sellers: The Only True Winners
But here is the funniest part of this entire technological circus. If you zoom out and look at the Dot-Com boom, the EV revolution, and the current AI gold rush, there is only one group of people who consistently win. Every. Single. Time.
During a gold rush, the people who get rich usually aren't the ones digging in the dirt. It's the guys selling the shovels.
In our digital age, the shovels are made of silicon. The only absolute, undisputed winners of every single tech revolution are the chip manufacturers.
You wanted to build a massive Dot-Com server farm in 1999? You needed chips.
You want to build a smart EV that can drive itself and manage battery load? You need chips.
You want to train a massive AI language model to write a Hive post? You need tens of thousands of insanely expensive GPUs.
While the rest of the world argues over which AI coin is going to the moon, or whether a generic chatbot is going to steal their marketing job, companies like Nvidia, TSMC, and ASML are sitting back, sipping champagne, and watching the money roll in. They don't care who wins the AI war, because they are supplying the ammo to both sides.
And the best part? They are already quietly designing and manufacturing the next generation of chips for whatever the next massive technological hype cycle will be. I have no idea what that next revolution will look like—quantum computing, neural links, holographic pets—but I am absolutely certain about one thing: the chip manufacturers will be the undisputed winners once again.
What do you guys think? Are we expecting too much from generic AI, and when do you think the "AI Bubble" will finally pop and leave us with just the useful stuff? Let’s discuss in the comments!
Have a great weekend,
Peter