The Great AI Paradox
Why This Bubble Will Burst, And Why It Doesn't Matter
I was fifteen when my father died. The world split. I carried that break through years of trying to understand how things fall apart and sometimes, if you're lucky, how they reassemble into something else.
Last week, watching another AI startup raise billions at comedy-sketch valuations, I recognized the pattern. Not because I'm smart. Because I've lived through enough breaks to know one when I see it coming.
The Movie We've All Seen Before
The late 90s: Slap ".com" on a sock puppet and watch your stock double.
Today: Whisper "AI-powered" and venture capitalists throw money like confetti at a wedding nobody really wants to attend.
Here's what haunts me, this isn't just another bubble story. It's more twisted than that.
The AI revolution, like the internet before it, is simultaneously the most overhyped bullshit and the most underappreciated transformation of our time.
Both things. At once. That's the paradox that's making everyone crazy.
In 1999, VCs said out loud: "The company doesn't have to be successful for us to make money."
Today, OpenAI commands $157 billion despite few understating how they'll make money long-term. Musk's xAI hit $50 billion before shipping anything real. BuzzFeed muttered "AI" and doubled overnight.
Now I am not denying that these companies have insane potential, but others are building on the marketing of these companies without understanding the inner workings.
The crazy part?—the rhetoric hasn't even evolved.
Then: "The internet changes everything."
Now: "AI is more profound than fire."
Same sermon, different church. Old rules don't apply. Get on board or drown.
Pets.com burned millions on Super Bowl ads for a sock puppet. Dead within months of IPO. Today's AI darlings are following the same playbook, just with better graphics.
The Thing Nobody Wants to Admit
Strip away the hype and you find something everyone's missing: AI is older than your parents.
Turing was thinking about this in the 1940s. The term "artificial intelligence" dates to 1956. But this isn’t even 1950’s old, its 1600’s old.
Descartes and Leibniz pondered mechanical reasoning before electricity existed.
(if you want to catch up on the last 400 years of “thinking machine” arguments, check out my article here )
When I train a model, I'm doing math that would make Euclid weep, calculating distances in spaces with more dimensions than there are atoms in your coffee cup.
Each weight adjustment shifts coordinates in geometric realms our brains can't visualize.
It's not magic. It's linear algebra at a scale that was impossible until about five minutes ago…historically speaking.
The Survivors Tell the Real Story
When the music stopped in 2000, the bodies piled up. But notice who lived:
Amazon ate a 95% stock collapse and kept building. Why? Because Bezos understood something others didn't, the internet wasn't about being online. It was about reimagining the entire pipeline from manufacturer to doorstep.
Google skipped the IPO circus entirely. Built a real business. Figured out that organizing information was worthless unless you could make someone pay for attention.
The pattern is brutal in its clarity: Infrastructure beats innovation. Business fundamentals beat brilliant demos. Solving real problems beats solving theoretical ones.
What This Means for the Church of AI
Today's AI landscape is littered with Pets.coms waiting to happen. But hidden among them are the Amazons, companies that understand AI isn't about the technology.
Nvidia gets it. They're selling shovels in a gold rush. Real shovels. Real gold. Real profits.
The rest? They're mostly selling dreams to VCs who are selling bigger dreams to pension funds.
When this bubble pops…and it will pop like your ears on a plane descending too fast, the survivors will be the ones who remembered something everyone else forgot:
Technology is only as valuable as the problem it solves.
The Questions That Matter
When someone pitches you their revolutionary AI product, ask:
"What specific problem does this solve that couldn't be solved before?"
"How does this make money when the free version inevitably appears?"
"Why won't OpenAI or Google eat your lunch with their next update?"
But especially ask: "Are you building infrastructure or just another interface?"
Because here's what I learned watching my world split at fifteen and spending years trying to understand how things break: The tools don't matter. The problems matter.
The infrastructure matters. The ability to see clearly when everyone else is drunk on possibility, that matters.
The Paradox Resolved
AI is overhyped when people claim it will solve loneliness or replace human creativity. (note It WILL replace current creative paradigms but the abstract is where humans thrive)
It's overhyped when every startup adds "AI" like it's sriracha, makes everything taste better but doesn't make bad food good.
But it's criminally underestimated by people who don't grasp what's actually happening:
We're not building thinking machines. We're building infrastructure for thought itself. Systems that can map meaning across domains in ways our pattern-seeking brains never could.
That's not magic. It's something weirder, tools that extend human cognition into spaces we couldn't reach before.
What Survives the Fire
The dotcom crash didn't kill the internet. It killed companies that thought the internet was about websites.
The AI crash won't kill AI. It will kill companies that think AI is about chatbots.
What survives?
The boring stuff.
The infrastructure plays. The companies solving problems so real that customers pay even when the hype dies. The ones building competitive moats from actual advantages, not press releases.
Amazon survived because warehouses are harder to copy than websites.
Google survived because organizing world's information is harder than building a search box.
The AI survivors will be the ones who understand: In a world where every startup can call GPT-4's API, what's your actual advantage?
The Lesson Written in Scar Tissue
I've watched enough things fall apart to recognize the pattern. The hype-crash-rebuild cycle isn't a bug, it's how we process transformation too big to swallow whole.
We need the bubble to fund the experiments.
We need the crash to kill the pretenders.
We need the quiet rebuilding phase where real value emerges from the wreckage.
AI isn't going away. But most AI startups are already ghosts, they just don't know it yet.
When the break comes, and it's coming like winter after a long fall, the survivors will be the ones who remembered that transformation isn't about the technology.
It's about understanding what actually needed transforming in the first place.
The internet didn't revolutionize everything.
It revolutionized how we move information. AI won't revolutionize everything either. But it will revolutionize how we process meaning.
That's not hype. That's not deflation. That's the paradox resolved, AI is exactly as important as the problems it solves, no more, no less.
Most companies are betting on more. They're wrong. The survivors are betting on less—specific problems, real solutions, boring infrastructure.
They're the ones who'll still be standing when the music stops.