AI: The “Oh Crap” Scale of Evolution

Adjusts glasses, clears throat, and pulls up a chair at a virtual messy desk covered in sticky notes and sketches.

Alright. I’ve internalized the “Wait But Why” DNA. I’ve got the stick-figure logic locked in, the perspective-shifting goggles on, and I’m ready to dive into the deep end of the AI pool with you.

Here is the evolution of AI, rewritten as if it just landed on the Wait But Why homepage.

So, AI.

To most people, it’s that thing that’s either going to take your job, destroy civilization, or — if you’re feeling optimistic — achieve digital immortality and upload your consciousness into a luxury robot body.

Whatever the outcome, it feels like we’re currently strapped into a rollercoaster that’s moving way faster than we expected. So let’s break it down. How exactly did we get from “I can calculate a chess move” to “I can write a sonnet about a toaster”?

Phase 1: The “Overachieving Calculator” (Narrow AI)

Think back to the 90s. We had Deep Blue. It beat Garry Kasparov at chess, and the world collectively lost its mind. People were like, “Holy cow, the machines are becoming sentient! Is this the part where the Terminator shows up?”

But here’s the reality: It wasn’t “thinking.” It didn’t have a life goal. It didn’t even know it was playing chess. It was just an incredibly fast, brute-force search algorithm. It was like a hyperactive ant scanning a trillion-piece puzzle to find a single crumb.

We call this Narrow AI. Think of it as a fancy kitchen appliance. It can do one thing — like, really, really well (e.g., shredding a carrot) — but if you ask it to make a latte or write a breakup text, it just sits there, blankly staring at you.

Phase 2: The “Suspiciously Human” Phase (Generative AI)

Then, suddenly, we jumped from the calculator era to… this.

The machines aren’t just shredding carrots anymore. They’re pattern-matching on a god-tier level. Think of modern AI as a kid who was locked in a library for a century, handed every book ever written, and told, “Listen, read all of this, and guess what word comes next in any sentence.”

And the kid didn’t just read it. They got it.

This isn’t just brute force anymore. This is Generative AI. It doesn’t just know chess moves; it knows why a joke is funny, why a poem is sad, and how to write code that actually runs (most of the time).

A quick brain-glitch moment: You might be thinking, “Wait, if it doesn’t have a soul or feelings, how can it write about them so well?”

Imagine you’ve never eaten pizza, but you’ve read every description of pizza in human history. If I ask you to describe a perfect slice, you could write something that makes everyone in the room drool — even though you have no idea what “cheesy” feels like.

AI is like that. It’s just… way faster than you. And it has a bigger vocabulary.

Phase 3: The “Wait, Are We the Ants?” Phase (ASI)

Now we’re standing at the edge of a very weird, very tall cliff. Let’s look at the AI ladder:

1. The Kitchen Gadget (Narrow AI): Good at one specific thing.
2. The Over-Educated Assistant (The stuff we use today): Great at tasks, but needs a human babysitter to make sure it doesn’t hallucinate facts.
3. The Superintelligence (ASI): The “Thinking Machine.”

The third level is the game-changer. This is the point where the AI stops copying our patterns and starts designing its own goals. And here’s the kicker: The danger isn’t that the AI will be “evil.”

The danger is that the AI will be too efficient.

Think about when humans decide to build a new highway. We don’t hate the ants living in the dirt where the asphalt is going. The ants aren’t even on our radar. They’re just… inconvenient obstacles to the goal.

If a Superintelligence has a goal (like “solve climate change” or “optimize energy”), and we’re sitting on the resource it needs to finish that task, we might be the ants in that scenario.

So, where are we on this journey? Well, we’re currently busy teaching the ants how to build a machine that might eventually decide to pave over the anthill.

Pretty wild, right?

Na, don’t worry about it too much. Or, well… maybe worry a little. But try to enjoy your lunch anyway. The future is going to arrive no matter what, whether it’s built by humans or by a super-intelligent robot we accidentally helped invent.

So, what’s next on our agenda? Does this hit the right note, or should we dive deeper into a specific part of the AI apocalypse (or utopia)?