Zac Johnson

Online since 1997. I've watched every major tech shift happen in real time — AI is the biggest one yet. Follow for takes you can actually use.

[Zac Johnson @zacjohnson]

I started my first online business in 1997.
Watched search change everything.
Then social. Then mobile.
AI makes all of those look small.
Still early. Most people will realize it too late.
Follow if you want to stay ahead of it.

[Zac Johnson @zacjohnson]

@AndySkraga DIY

[Zac Johnson @zacjohnson]

@AndySkraga PDF

[Zac Johnson @zacjohnson]

@adcrookes Backlink

[Zac Johnson @zacjohnson]

@publisherinabox CURATE

[Zac Johnson @zacjohnson]

@recap_david RENOVATE

[Zac Johnson @zacjohnson]

@Charles_SEO Secret

[Zac Johnson @zacjohnson]

@AndySkraga TOBI

[Zac Johnson @zacjohnson]

@ThebilalSEO Brain

[Zac Johnson @zacjohnson]

@hamza_automates 200

[Zac Johnson @zacjohnson]

@publisherinabox DM me

[Zac Johnson @zacjohnson]

@jakezward Replay

[Zac Johnson @zacjohnson]

@recap_david AI ads

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
Anthropic got publicly punished by the DOW for refusing to play ball → instead of it killing them, millions of people heard about Claude for the first time and picked a side.

Claude jumped from #131 to #1 on the App Store in days. Free users up 60%, paid subs doubled, daily signup records breaking every single day this week.

The Pentagon contract they lost? Worth ~$200M. Anthropic makes $14B/year. That’s 1.4% of revenue traded for the kind of brand moment money literally cannot buy.

Then the killer move → they shipped memory to the free plan right as millions of new users flooded in. That’s not luck, that’s a team that had retention plays loaded and ready to fire.

Katy Perry posting her Claude Pro sub, Reddit organizing ChatGPT cancellations, 700+ employees at Google and OpenAI signing letters backing Anthropic → you can’t manufacture that with any ad budget.

Big idea: The best competitive moat in AI isn’t technology, it’s trust — and Anthropic just proved that standing on principle can be the most aggressive growth strategy in the market.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
Oracle’s Larry Ellison told investors his plan: every body cam, doorbell cam, dash cam, and drone feeds into one AI system that watches everything in real time. Cops can’t turn cameras off, and AI flags incidents automatically.

The twist → it’s not just watching police, it’s watching you too. All those feeds get analyzed by AI on Oracle’s servers, not your local government’s.

Instead of police chases → autonomous drones just lock onto your car and follow you. No human decision needed.

The company pitching this built its first database for the CIA and holds $320B in government contracts → so the infrastructure is already halfway there.

No law was passed. No vote happened. A private corporation is just selling total surveillance as a product to the agencies that can deploy it.

Big idea: When one company builds the eye that never blinks and sells it to the people with guns and badges, “keeping you safe” and “keeping you controlled” become the same feature.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
OpenAI cut a deal with the Department of Defense. People freaked out. Now they’re adding explicit language saying the AI can’t be used to spy on Americans — including buying data brokers’ info to track people.

They’re also drawing a line → NSA and other intel agencies don’t get access unless a whole new contract is negotiated separately.

Sam’s basically saying “we want a seat at the table, not to BE the table.” The government makes decisions, OpenAI advises. He even said he’d go to jail before following an unconstitutional order.

The honest part → he admitted dropping this deal on a Friday was a mistake. Complex stuff needs clear communication, not a news dump before the weekend.

He also said Anthropic shouldn’t be cut out and should get the same terms. That’s notable — publicly defending your competitor signals this is about industry norms, not just one company’s deal.

Big idea: The real story isn’t that OpenAI is working with the military — it’s whether the guardrails they’re writing down actually have teeth when it matters.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
Your brain has a “sweet spot” where you do your best work — not too bored, not too stressed. That’s called the flow channel.

Most people protect that zone by staying comfortable → small challenges, small growth, small results.

The hack is counterintuitive: deliberately take on 10-100x more responsibility than feels sane. Your nervous system panics, then adapts — building new neural pathways just to survive the load.

Then you drop back to normal. Suddenly your old workload feels like lifting a feather after carrying a boulder. Your capacity is permanently stretched — like a rubber band that doesn’t snap back to its original size.

The catch most people miss → the overload has to be focused in one domain with real consequences, not busy work spread across random projects. One impossibly heavy backpack beats ten medium ones.

Big idea: You don’t grow by gradually adding weight — you grow by taking on something so absurd it forces your biology to unlock reserves you didn’t know existed, then you keep those upgrades forever.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
Gold and Bitcoin both claim to be “safe havens.” But since April 2025, gold is up 60% while Bitcoin is down 30%. That tells you everything.

When wars escalate, gold surges → Bitcoin drops. When stocks sell off, gold holds → Bitcoin sells with them. It trades like a tech stock wearing a gold costume.

The narrative always follows the price. At $126K everyone called Bitcoin “digital gold.” At $66K with 5 red months, nobody mentions that thesis anymore.

Central banks bought 863 tonnes of gold in 2025 — fastest pace in decades. Nobody is buying Bitcoin for sovereign reserves. That’s the institutional tell.

Bitcoin acts like protection only when everything else is already going up. In a world with real geopolitical risk, a weakening dollar, and central banks hedging — you want the asset that’s done the job for 5,000 years.

Big idea: Gold survived empires, Bitcoin survived Twitter — and when real crisis hits, the market shows you which one is actually insurance and which one is just cosplaying.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
A Trump ally just publicly broke ranks over Iran, and the argument is simple → Americans are drowning in debt, can’t afford insurance, and live paycheck to paycheck, so why are we spending blood and treasure on another Middle East war?
The core tension: “America First” was the brand promise.

War in Iran is the opposite of that promise. When your own base starts saying it out loud on camera, that’s not media spin — that’s a real fracture.

Here’s the pattern that matters → every modern Middle East conflict started with “it’ll be quick” and ended with years of escalation, trillions spent, and zero benefit to average Americans.

72% can’t afford health insurance but we can afford another war? That contradiction is what breaks political coalitions. People tolerate a lot, but not hypocrisy they can feel in their wallet.

Big idea: Wars don’t stay contained and neither does the political fallout — when the base feels betrayed, the realignment happens fast.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
Jack Dorsey just fired 4,000 people from Block — not because business is bad, but because AI made them unnecessary. His exact words: business is strong, profit is growing. He just doesn’t need the humans anymore.

Here’s the chain → AI tools get good enough to replace whole teams → one person with AI does what three did in 2023 → companies realize they’re overstaffed → mass cuts begin. Dorsey said he could have done it slowly so nobody noticed. He chose to rip the bandaid. Other CEOs are doing the slow version right now.

This splits the workforce into two groups. Group 1 uses AI like a fancy Google search and gets a marginal boost. Group 2 builds systems and automations that run without them. Dorsey kept Group 2 and cut Group 1. Every company will do this — the only variable is timing.

The survival move isn’t “learn to prompt better.” It’s → audit your daily tasks → automate the repetitive ones → build something new that didn’t exist before → then TELL your boss what you built. The person who deploys AI as infrastructure is the last one cut.

What AI still can’t touch → real relationships, strategic judgment about what matters in YOUR specific context, deciding what should exist vs. just building what’s asked for, and navigating the messy human politics of getting orgs aligned.

Big idea: The window to become the person who deploys AI — not just uses it — is open right now, and every month you wait, the bar gets higher.

[Zac Johnson @zacjohnson]

No BS breakdown (ELI5): 👶🧠
The best-performing ads right now look like they're not ads. One person, one camera, talking like they're texting a friend — no crew, no set, no polish.

It works because your brain dismisses ads before they finish loading. → If it looks native, your guard stays down. → By the time the product appears, you already trust the person.

The formula is 4 steps: hook that stops the scroll → specific problem they actually feel → product as the natural solution → soft CTA that doesn't beg.

The kill shots: leading with the product, weak openers, or vague problems. "I felt tired" = nobody. "These were never allowed in my house growing up" = everyone stops.

Brief the creator on the skeleton, then get out of the way. Over-scripting kills the thing that makes it work.
Big idea: In a world drowning in polished ads, looking real is the most expensive thing you can fake.