Dave Gambrill

You've seen my AI videos. 👀

This is how I make them in minutes on my phone...

https://gambrill.com/aivids

Grab that now before I start charging for it.

👊🏻

Create Amazing AI Videos With Your Phone in Minutes Learn how to create AI videos with Google Gemini using your phone, AI avatar, prompts, and images in this free beginner-friendly mini-course.

I think a lot of people still fundamentally misunderstand what’s happening with AI right now.

They think this is mostly about:
- writing emails faster
- making funny images
- generating social media posts
- automating customer service

And yes, AI can do all of those things.

But that’s the surface layer.

The real story is discovery.

The real story is exploration.

The real story is what happens when you point massive compute, reinforcement learning, simulation, probabilistic reasoning, and relentless iteration at problems humans either:
- couldn’t solve,
- stopped trying to solve,
- or never realized had another path forward.

And honestly, we’ve been watching this unfold for years already.

A little over a decade ago, one of the biggest AI milestones in history happened when AlphaGo defeated world champion Lee Sedol in the game Go.

Now if you’re unfamiliar with Go, it’s an ancient strategy game that’s dramatically more complex than chess in terms of possible board states.

The number of possible game positions is so absurdly large that people often compare it to the number of atoms in the observable universe.

In other words:
you cannot brute force Go the same way old-school chess computers brute forced chess.

The search space is too large.

There are too many possibilities.
Too many branches.
Too many paths.
Too many combinations.

For years, people thought Go would remain one of those deeply human games that computers couldn’t truly master because intuition mattered so much.

Then AlphaGo showed up.

And what shocked experts wasn’t just that it won.

It was HOW it won.

There were moments where the AI made moves professional players thought were mistakes.

Some commentators literally thought the machine had malfunctioned.

Until 20 moves later when they realized:
“Oh no… it saw something we didn’t.”

That’s the important part.

The system wasn’t merely copying human behavior.

It was exploring solution spaces differently than humans do.

It found moves humans rarely considered.
Strategies humans undervalued.
Paths humans overlooked.

And after enough games, human players themselves started learning from the AI.

Think about that for a second.

Humans taught the machine.
Then the machine started teaching humans.

That was one of the first big public glimpses into what this era might become.

And now we’re seeing the same pattern emerge in science and mathematics.

A few years ago, there was a massive challenge in biology called protein folding.

Scientists knew proteins were critically important because proteins basically drive life itself. But understanding how proteins fold into their final 3D structures was incredibly difficult.

And that structure matters.

It determines behavior.
It determines function.
It determines interaction.
It determines whether something heals, harms, activates, blocks, repairs, or breaks.

For decades, protein folding was one of the grand scientific challenges of biology.

Then DeepMind created AlphaFold.

And AlphaFold essentially solved a problem that had bottlenecked scientific progress for decades.

But here’s the part that really matters:

They didn’t stop at solving a few proteins.

They essentially said:
“What if we solve ALL the known proteins we can?”

And they did.

Millions of protein structures mapped and released publicly for scientists around the world.

Think about the scale of that shift.

What was once a years-long bottleneck became a searchable resource.

That’s not just automation.

That’s acceleration of human discovery itself.

And now we’re seeing yet another example.

I just read about an AI model helping make a breakthrough on a geometry problem first posed by Paul Erdős back in 1946.

Nearly 80 years ago.

Generations of brilliant mathematicians worked on this thing.

Tiny incremental progress.
Lots of dead ends.
Long-standing assumptions nobody could fully crack.

Then AI helps uncover a path researchers hadn’t seriously considered before.

Again:
this is not really about geometry.

It’s about search space.

Humans are extraordinary at reasoning.

But we are limited by:
- time
- energy
- memory
- assumptions
- cognitive bias
- experience
- pattern recognition capacity

We tend to stay relatively close to known territory.

AI can explore differently.

Not because it’s conscious.
Not because it’s magical.
Not because it “thinks” exactly like humans.

But because it can:
- iterate endlessly
- simulate millions of outcomes
- optimize continuously
- identify hidden patterns
- test weird possibilities humans dismiss
- connect ideas across domains at enormous scale

Sometimes it brute forces.

Sometimes it stumbles into elegance.

Sometimes it discovers paths humans never even noticed.

And when the possibility space becomes massive enough, this becomes transformational.

That’s why this matters far beyond games and math.

Because now apply this to:
- manufacturing
- chemistry
- materials science
- medicine
- energy systems
- logistics
- aerospace
- genetics
- robotics
- agriculture
- climate science
- battery technology

There are solutions hidden inside these giant possibility spaces that humans alone cannot realistically explore fast enough.

New medicines.
New materials.
New engineering designs.
New manufacturing methods.
New energy systems.
New optimization models.
New treatments.
New supply chain efficiencies.

Not because AI replaces scientists or engineers.

But because it becomes an exploration engine for them.

A tireless collaborator.
A simulation partner.
An idea generator.
A pattern recognition system operating at superhuman scale.

And I still think most people are underestimating this.

A lot of the public conversation is still focused on:
“Can AI write my emails?”
“Can it make a logo?”
“Can it generate content?”

Meanwhile, these systems are beginning to help humanity navigate possibility spaces we literally could not traverse before.

That changes everything.

Because once discovery itself accelerates, civilization accelerates.

Some of the most important breakthroughs of the next decade may not come purely from human intuition anymore.

They may come from humans working alongside systems capable of exploring more possibilities in a day than entire teams could explore in a lifetime.

That’s why I keep saying this isn’t just another technology trend.

This is a foundational shift.

👊🏻

The AI space is moving fast right now.

In just the last couple of days, OpenAI rolled out major Codex upgrades, Google pushed a stack of new Gemini features, Anthropic released a new model, and Perplexity made another strong move toward AI that can actually operate across apps, files, and workflows, not just answer prompts.

That’s the part that matters.

We’re not just watching smarter chatbots show up.

We’re watching the interface to work itself start changing.

This is where it usually breaks down for people:

They keep treating AI like a novelty instead of a skill.

They watch from the sidelines.

They wait until the tools feel “settled.”

That’s a mistake.

The people who win with this stuff are not the ones who waited for perfect clarity.
They’re the ones who got in the game early, built reps, and learned by doing.

You do not need to master everything overnight. But you do need to start.

Pick a tool.
Pick a workflow.
Pick one real use case in your business.
Then start building familiarity now, while most people are still just spectating.

Because once these tools move from “interesting” to “expected,” the gap between users and non-users gets very real, very fast.

👊🏻