Everyone's obsessing over which AI model is smarter.
Wrong thing to watch.
The real story isn't about the models. It's about what happens when AI gets so cheap, so fast, and so good at the boring stuff that the only thing left with real value is the thing machines can't fake.
Your taste. Your perspective. Your ability to know what resonates before anyone else does.
That's where we're headed. And it's closer than most people think.
The Price Collapse Is Real — And It Changes Everything

Here's a number that should stop you cold.
In late 2022, running one million tokens through an AI model cost around $20. By August 2025, that same million tokens cost roughly $0.40. That's a 95% price collapse in under three years.
And it's still going.
Right now, some models are processing tokens for as little as $0.08 per million. The median price decline is running at 200x per year. Gartner is already predicting that by 2026, AI service cost will become a bigger competitive factor than raw performance.
Think about what that means.
When something drops 95% in price, it stops being a competitive advantage. It becomes infrastructure — like electricity. Like the internet. You don't win because you have electricity. You win because of what you do with it.
That's where AI is right now. Heading straight for commodity status.
When that happens, the question changes from "who has the best AI?" to "who has the best ideas?"
The Death of the Content Farm

I want to paint you a picture of what's already happening.
Right now, thousands of businesses built on templated content, SEO-optimized fluff, and mass-produced social posts are in trouble. Not because AI is bad. Because AI is too good at making exactly that kind of content.
Stanford's AI researchers are already tracking this at the task level — studying where AI is boosting productivity, displacing workers, and creating new roles. The categories getting hit hardest? The ones that don't require a point of view.
The summary. The first draft. The research brief. The data entry. The follow-up email. The meeting recap.
These are gone, or going. Not because they're not valuable. Because any model at $0.08 per million tokens can do them faster than a human, at scale, without breaks.
What's left?
The stuff that only sounds like you.
The take that only makes sense coming from someone who's actually lived it. The insight that lands because the audience trusts the person saying it. The creative spark that knows what will resonate before it gets tested.
You can't buy that at $0.08 per million tokens.
Free email without sacrificing your privacy
Gmail tracks you. Proton doesn’t. Get private email that puts your data — and your privacy — first.
The Data on Creators Is Wild

Here's what the numbers look like on the other side of this.
The U.S. creator ad market hit $37 billion in 2025 — up 26% year over year. YouTube topped TV viewership every single month throughout 2025. The Interactive Advertising Bureau expects the growth to continue through 2026.
At the same time, expensive celebrity endorsements are fading. Brands are shifting dollars to niche creators with real trust and real engagement — because they're proving more valuable per dollar than anyone expected.
Later's 2026 Creator Economy report identified what's driving this. A few things jumped out:
Creators are licensing their likeness to AI companies. Think about that. The human, the voice, the persona — not the content they produce, but who they are — has become an asset worth licensing. That only happens when identity is the scarce thing.
Written POV content is making a comeback. The flood of AI-generated text is actually making newsletter and long-form writing more valuable, not less. Platforms are rewarding original human perspective. Creators with real opinions and first-person experience are standing out harder than they did two years ago.
And authenticity is now verifiable. EY's 2026 media report put it plainly: "AI slop won't satiate consumers craving authenticity." Audiences have gotten good at sensing the difference between something a person actually thought about and something that was generated. The gap between "real" and "AI-produced" is becoming a brand signal.
The Historical Pattern That Should Make You Optimistic

We've seen this exact movie before. Twice.
Digital cameras. When they arrived, suddenly anyone could take a technically perfect photo. The barrier to entry dropped to zero. So what happened? Did photographers lose their jobs?
The mediocre ones did. But the best photographers — the ones with a genuine eye — became more valuable, not less. Their skill was never about operating a camera. It was about seeing. And when the camera was no longer the hard part, seeing became everything.
Music production software. Same story. When anyone could make a beat with a laptop and GarageBand, music production democratized overnight. And the producers who understood vibe, feeling, and what made a song connect? They became essential. Because the technical barrier disappeared, and all that remained was taste.
AI is doing the same thing to content creation right now.
The technical barrier — the ability to write competently, produce a polished post, generate a decent image — is collapsing. The floor is rising. Anyone can look good now.
Which means looking good no longer gets you anywhere.
What matters is whether you have something to say.
What the Experts Are Getting Wrong

Most AI takes right now fall into one of two camps.
Camp one: AI will take all our jobs and the machines are coming.
Camp two: AI is just a tool, don't worry about it.
Both miss it.
The real story is more specific. AI is taking the boring jobs — the ones that feel like work but don't require real judgment. And it's creating a massive vacuum at the top.
Microsoft's 2026 AI predictions noted that AI won't just summarize papers and write reports anymore — it's actively joining the process of discovery in physics, chemistry, and biology. The boring work is handled so thoroughly that even in scientific research, the frontier has shifted.
For creators, the parallel holds. The commodity stuff — the templated post, the generic email sequence, the average newsletter — is fully automated. Not coming, already there.
The thing that can't be automated is a perspective. A specific human way of seeing things that an audience has decided to trust.
That's your moat.
The New Tools Worth Knowing

Before I wrap this up, a few things I've been watching.
Kimi K2.5 — One of the most cost-efficient models out there right now. The team built OpenClaw on it — an AI assistant with memory and personality that deploys in one click through Kimi's platform. If you want to see what low-cost AI agents actually look like in practice, this is worth exploring.
Manus — They built a Telegram-based AI agent that can run tasks securely in the background. The interesting thing here isn't the interface, it's the framing: an AI that works for you while you sleep. manus.im if you want to check it out.
DoAnything — Lives up to the name. General-purpose AI agent interface worth having in your toolkit. doanything.com
SeedDance 2.0 from ByteDance — New AI video creation tool. Still early, but ByteDance has the resources to build something serious here. seed.bytedance.com
The through-line on all of these: the cost of building and deploying AI-powered experiences keeps dropping. The tools are getting better and cheaper at the same time.
That's the cost collapse story playing out in real time.
The Bottom Line

Here's how I think about where we are.
Six months from now, the basic stuff is fully automated. Email drafts, meeting summaries, first-pass content, scheduling, reporting, customer FAQs — handled. The cost is too low, the quality too high, the speed too fast for humans to compete on tasks like these.
Six months from now, the premium stuff — the creative stuff — commands a higher premium than ever. Because when everyone can make decent content, the only content worth paying attention to is the stuff that has a genuine point of view behind it.
The photographers who kept working after digital cameras weren't the technical experts. They were the ones with something to see.
The producers who kept working after GarageBand weren't the ones who understood compression ratios. They were the ones who understood what made a song feel like something.
The creators who win in the next 12 months won't be the ones with the best prompts.
They'll be the ones with something to say.
That's the bet I'm making. And if you're reading this, I think you probably are too.
One More Thing
I've spent the last two years building AI systems for my own businesses.
Not because I'm a developer. I'm not. Never wrote a line of code before any of this.
But because I was drowning in the boring stuff — follow-ups, scheduling, reporting, the tasks that kept me working nights and weekends on things that should have run themselves.
So I fixed it. Built automations that handle it. Got my time back.
Now I help other people do the same thing.
It starts with a 60-minute call. We dig into your operations, figure out exactly where your hours are going, and I give you a clear roadmap for automating the stuff that's eating your week.
It's $25. No pitch. No pressure.
You leave with a plan you can execute yourself. And if it makes sense to work together on the build, I offer done-for-you automation starting at $3,500.
If you're tired of working in the business instead of on it — this is where you start.
Questions? Just reply to this email. I read everything.

