"Turns out the future does have flying cars." Video: Elon Musk on X, October 15, 2026.

On October 15, at a SpaceX facility, Tesla drove a new Roadster up a ramp that curled overhead until the car hung upside down. It clung there by suction, launched off the end, rolled 180 degrees in midair on jets of compressed air, and landed on its wheels. Production is slated for 2028.

A car that leaves the ground was a punchline for most of our lives. Now it is a livestream.

Progress does not arrive evenly. Two days after the Roadster flew, protests across France turned into riots. In Limoges, rioters set fire to the 1883 town hall, and its spire collapsed into the flames. Nearly a century and a half of craftsmanship, gone in a night. Progress is jagged. The same week can deliver a flying car and a burning landmark. The next ten years will see plenty of both.

Meanwhile, new frontier AI models now ship almost every week. Since September, the White House has called them Super Intelligence, or SI, and the name is earning its keep. These systems write production software, run ad accounts, draft legal briefs, and build 3D worlds in an afternoon. Whole categories of software are being rebuilt from scratch by small teams, sometimes by one person plus agents. People have started calling it the SaaSpocalypse. And the pace keeps rising: Anthropic's Claude Duet, an agent that checks in before you ask, is already in limited release, and a Fable 5.5 model is widely expected next.

This article is a realistic look at the next ten years: what they could bring, and how to navigate them.

The software layer is already moving

Look at how SI stacks up against the best human programmers. Codeforces ranks competitive programmers worldwide. In 2024, GPT-4o rated 808, below the median human on the site. In October, DeepSeek's newest model reported a rating of 3,471. Of the 49,384 people who competed on Codeforces in the past month, only two rate higher. And in March, an SI system called GrandCode took first place in three live Codeforces contests in a row, ahead of every human in the field.

Longer engineering work tells the same story. METR, an independent lab, found that by April SI agents could finish software tasks that take a skilled human engineer about 17 hours, more than two full workdays, before its tests ran out of tasks hard enough to measure the newest models.

Every product that lives on a screen is now replicable. A feature that took a funded team a year to build can be rebuilt over a weekend. The most nuanced programs, the ones whose complexity used to be their protection, are falling fastest, because complexity is exactly what these models handle well.

The hardware layer is next

In August, at the World Humanoid Robot Games in Beijing, a robot called Tiangong Ultra ran 100 meters in 8.86 seconds. Usain Bolt's world record is 9.58. Then it crashed into the barrier past the finish line, sparks flying from its waist. Both halves of that clip tell you where we are.

Tiangong Ultra, 100 meters in 8.86 seconds. Beijing, August 2026. Video: CGTN.

Robotics is following the same curve, a few years behind. The hard parts are hands, balance, and judgment in messy physical spaces. Balance and judgment are mostly software problems, and software problems are getting solved. Hands are the hardware holdout: motors, sensors, and materials that can match human dexterity. Engineers are making strides on both.

When robotics reaches its tipping point, the disruption reaches the electrician, the framer, and the warehouse just as surely as it reached the developer.

The extreme case

Here is a thought experiment. It describes what could happen, which is different from a forecast.

Suppose SI agents become as competent as the most competent human at every task you can do on a computer. Suppose robots arrive with hands and limbs as capable as ours, running flawless software and doing every job exactly right: faster, with fewer errors, and no need for rest.

In that world, the value of labor as a way to raise your economic standing trends toward zero.

Luke Drago and Rudolf Laine call the path there pyramid replacement. Companies stop hiring at the bottom, and machines climb the org chart one layer at a time until only the top remains.

Picture someone who is already wealthy. They can hire a robot that does everything you can do, only better, at a laptop or on a job site. With enough capital, they hire ten. Then a hundred. Then a thousand. Good luck competing with that.

There will be exceptions. A few savants, imaginative and unorthodox enough to find an angle the machines missed, will still climb by their own effort. They will be the exception.

So the model looks like this:

The value of labor as a lever for wealth creation goes toward zero. The value of capital as a lever for wealth creation goes toward infinity.

Naval Ravikant counts a third lever beyond labor and capital: products that cost nothing to copy, meaning code and media. Super Intelligence changes both. Code folds into capital, because whoever pays for the compute owns the output. Media gets cheap too, since SI can produce it in endless volume. What stays scarce is attention, and trust in real people, especially the ones you knew before SI arrived. Keep that in mind when we get to distribution.

You can cooperate instead

Here is the good news. You do not have to compete with the machines. You can command them.

If you build up any scarce, desirable capital before this trend takes hold, you may be able to direct more intelligence and more physical labor than any of your ancestors could have dreamed of. One person with a modest stake could run what used to take a company of hundreds.

Here is what that looks like in dollars. In 2021, $60 bought a million tokens from GPT-3, a model that could not reliably write a working program. Today, $10 buys half a million tokens of output from Claude Opus 5.5, one of the top-ranked models in the world. The price of a given level of intelligence has been falling about tenfold a year.

Point that $10 at a team of SI agents and the math turns strange. A website an agency would have quoted at $100,000 five years ago gets built in an afternoon. An app that once meant $10 million in developer time gets built over a weekend. You can accomplish more with less money than ever before, if you put your capital to work through SI agents.

And if you have not built that capital yourself, a family member, a friend, or a patron in your network who has gathered scarce assets, land for instance, could bring you into their circle on favorable terms. You could be far better off than you are today.

The assets that matter in this model are the scarce ones: land, gold, equities (especially in SI and in the sectors SI depends on and is unlikely to do without), and Bitcoin. Whoever holds them when these trends mature may end up in a better position than they can imagine.

Two ways to spend the next ten years

One way is to scroll TikTok, watch Netflix, and order Postmates with your savings.

You might still come out fine. Maybe your family, friends, and patrons look after you and value keeping you in a good life. Maybe SI makes the world so productive, and the distribution problem gets solved so well, that universal basic income or something like it delivers abundance to everyone. In that world, you could be better off than today without working hard at all.

But why roll the dice? The other way is to do everything in your power to position yourself for the future at its most extreme. If the extreme never arrives, you are still well positioned. If it does, you are ready.

The way out is through

Plenty of people want all of this stopped: banned, paused, regulated into the ground. Even inside the labs, more than 1,300 employees, including Anthropic's Dario Amodei, signed a statement called Pacing the Frontier, asking for an international effort to deliberately slow the most advanced systems while safety and oversight catch up.

There are good reasons to proceed with care. But look at the world we would be pausing into.

Federal debt held by the public is back near its post-World War II record of 106% of GDP, and the Congressional Budget Office projects 120% by 2036. The government now pays about $1.3 trillion a year in interest, a record share of the economy and more than the entire defense budget. The Social Security retirement fund runs dry in late 2032, triggering an automatic benefit cut of about 22 percent. Medicare's hospital fund follows in 2033.

The classic way out of a debt like this is to grow faster than it does, and that takes a productivity boom. Super Intelligence is the strongest candidate for one we have ever seen. Pause it, and the bill still comes due, with higher interest rates and less growth to pay it.

Younger people carry the heaviest load. They pay into Social Security with every paycheck, and in real terms they are unlikely to get back what they put in. For them, putting SI agents to work over the next ten years may be one of the only realistic ways to climb toward financial independence.

The way out is through. There is no putting the genie back in the bottle, even if we wanted to.

What still holds when software is free

If software becomes a commodity, what is left? Pieter Levels, the solo founder behind Nomads.com and Photo AI, recently laid out the moats he believes survive. There are six.

Distribution deserves special attention. When anyone can build anything, the scarce thing is getting chosen. That applies to people choosing what to buy, and more and more it applies to the AI agents doing the shopping on their behalf.

How Noble Growth is building for it

We have started positioning Noble Growth, and the brands we work with, for this world. One moat at a time:

• Proprietary data. We are building a dataset of what works in ad testing to drive revenue across every kind of business and market sector, drawn from the tests we run.

• Network. We are growing our own network and the networks of the brands we work with.

• Distribution. It is the core of what we do: getting brands found and chosen. That now includes selling directly to AI agents. We have been building MCP connections that let a customer's agent find a partner's products and buy them, with no human required at checkout.

• Physical. We are acquiring land and exploring real-world components of our work.

• Capital. We are accumulating capital for ourselves and helping the brands we work with put theirs in the strongest position possible.

• Regulation. We hold a federal trademark of our own and help clients protect their brands and intellectual property.

The time is now. The choice is yours.

If you'd like help mapping your brand's next ten years, reach out.

A closing note

So what role is left for humans if SI takes on nearly every task? The Matrix imagined us as batteries, harvested for power. A more hopeful picture: humans as the data. SI exists to serve what we need and want, and the only way to learn that is to watch us: our choices, our conversations, the signals we give off in every interaction.

Think of researchers tracking snow leopards in the Himalayas, studying the animal in its own habitat to understand it. Now add one more design goal: the researchers' whole purpose is to care for the leopard, its cubs, and the mountains they live in. That is a far more optimistic future than The Matrix. It is also why proprietary data tops the list of moats.

Short of that extreme, some work will always be preferred done by a human. People will want a human priest, a human coach, a human at the head of the table.

Consider DJs. The craft was automated years ago. Software can beatmatch, mix, and play any setlist you program. Yet crowds still pack rooms to watch a human do it, even when the set is preset and the DJ is mostly jumping around on stage, vibing with them. In February, Calvin Harris played to 1.6 million people on a single São Paulo street, a Guinness World Record for a DJ set. People show up for the person.

Calvin Harris on Rua da Consolação, São Paulo, February 2026. Video: Music World Originals.

So the extreme case above is likely more extreme than what actually happens.

It is still a thought experiment worth running for every person alive today, especially the young, who have decades ahead of them as this future arrives.

"May you live in interesting times" is usually said as a curse. Read it again. Interesting times are when positions get made.

Our last post promised King Arthur would turn up in another ad you would recognize instantly. Here he is, recast into Guinness's 1999 "Surfer," shot for shot.

Guinness "Surfer", directed by Jonathan Glazer, was voted the greatest TV ad of all time in a 2000 Channel 4 and Sunday Times poll. Every wave, horse, cut and note of the soundtrack here is original. Only the hero changed, with no actors on camera.

How to use Higgsfield AI to make this film

The full workflow behind Reimagined No. 2: King Arthur recast into Guinness's 1999 "Surfer," shot for shot, in two days. Below: the tools and what they cost, the prompts, and the fix for every shot that fought back. Follow these steps to recast any classic ad the same way.

Read more

A few weeks ago I came across a guy on X named Jean Philanthrope. French, blonde, a funny haircut, always in a suit, and constantly throwing boxing combinations. I watched a few of his clips and thought, this is a real person. Just an eccentric Zoomer who likes sharp suits and happens to be pretty good at boxing. It did not even occur to me to question it.

He is not real. None of it is. And the moment I realized that, I did not feel tricked so much as curious. How do you make something this good? Good enough that a guy who does this for a living scrolled right past it and assumed it was a person.

So I went and figured it out. This post is what I learned, and how I have started making the same thing.

First, where I draw the line

Before the how-to, I want to be clear about something, because this technology sits in a gray area and I think most people are handling it badly.

Creating an AI character and passing it off as a totally real person is not something I am in favor of. It is deceptive. It likely has strange psychological effects on the people consuming it. And it is just not white hat. You see this constantly on marketing Twitter right now: fake AI influencers shilling products, with no disclosure that the person recommending the thing does not exist. That backfires. You get backlash, you get your account banned, and more importantly it is simply not the right thing to do.

But there is a right way to do this, and it is genuinely exciting.

If you are making an obvious parody or a riff on something people already know, that is fair game. If you label the character as AI, that is fair game. If you are experimenting to learn the technology, that is fair game. And down the road I think brands will use AI characters the same way they use actors today, with a simple disclosure that these are AI characters, not real people, so the focus can stay on putting out the highest quality creative possible.

That is the spirit everything below was made in. With that, here is how I did it.

Step one: the idea

Recently I updated the Noble Growth logo. The whole thing started with a dream. I saw a famous painting from over a hundred years ago of King Arthur in the moment he is being crowned, the crown just touching his head, and I woke up knowing that was the icon.

It fit perfectly. The way I see it, working with Noble Growth is the moment you put the crown on your own head. You give yourself the sovereignty to finally build your brand the way you have always envisioned it. We hand you the crown, the sword, and the armor you need to reach your full potential. So King Arthur became the icon, and then he became something bigger: the larger than life mascot for the whole brand.

Once I had the mascot, the idea almost suggested itself. Take the most iconic commercials of all time, and drop King Arthur into them in place of one of the main characters.

The first one was obvious. Apple's 1984 ad. Directed by Ridley Scott, aired nationally just once, during the Super Bowl, sixty seconds long, and still considered by many the greatest commercial ever made. The one where the runner sprints down the aisle of drones and hurls a hammer through the screen. My version: instead of the runner with the hammer, King Arthur, in full armor, sprinting with his sword.

Step two: the first attempt, and why it failed

I opened Claude in my terminal and had it plan the whole thing out, including which tools to use. The first renders came from Google's Veo, and when those credits ran dry, Claude moved me to a class of video platform where you load up a balance of credits and swap in whichever model is currently the most cutting edge. I put a hundred dollars of credits in, dropped in the YouTube link for the 1984 ad as a reference, and dropped in my King Arthur source image.

I thought it might one shot it. It did not.

The problem was consistency. In one shot he had a crown, in the next he did not. Even his face changed from shot to shot. His armor looked one way from the front and a different way from the side. Sometimes he was gripping the sword, sometimes it was sheathed, and in a few frames he somehow had it both ways at once. It was close enough to see the vision and broken enough to be useless.

So I asked the obvious question. What are all the people making flawless AI characters doing that I am not?

Step three: the character reference sheet

This turned out to be the whole game.

I used Grok to research everything people have figured out about keeping an AI character consistent, and the answer kept coming back to one thing: a character reference sheet. Instead of making the video model reinvent your character from scratch in every shot, you first generate one clean sheet that shows the exact same character from every angle. Front, side, three quarter, back, plus tight close ups of the face. That sheet becomes the single source of truth.

Grok handed me a detailed prompt for building one. The core of it is an identity lock. You tell the model to treat your reference image as the one and only truth, preserve every facial feature and proportion exactly, no beautifying, no aging, no stylizing, and render the identical character across every panel on a plain studio background. I will share the full prompt below so you can use it yourself.

I took that prompt and my King Arthur image into ChatGPT, which in my experience still produces the best images, and it gave me something very good.

But not quite right for the ad. In that first sheet the sword was sheathed and there was no crown, and for this I needed the crown on his head and the sword in his hand, because the character in the 1984 ad is carrying the hammer.

So I corrected it directly: put the crown on his head, put the sword in his right hand, and leave the scabbard empty. That gave me the perfect reference sheet.

Step four: the last ten percent

I brought the finished reference sheet back into Claude in my terminal and told it to hold the character consistent against that sheet. The next render was almost perfect.

Almost. The one thing that was off was the throw. The model made him swing and release the sword the way you would throw a hammer, up and over, top over top. Which makes sense, because that is what is happening in the original ad. But that is not how a sword moves through the air. He was also gripping it with both hands, which no one would do with a sharp blade, so I told it: right hand only.

A thrown blade behaves much more like a javelin. So I went looking for the right motion, and I fell down a wonderful rabbit hole. It turns out there was once a rotational javelin technique, a full spin before the release, that got so dangerous it was banned in 1956. Describing that spin in words did not work. The model could not picture it. So I found footage of it, and it was exactly what I had pictured for Arthur. I handed that clip to the model as a movement reference and said, throw the sword like this man throws the javelin.

The spinning javelin, banned in 1956.

That did it. Arthur spins, the sword leaves his hand, flies straight into the screen, and the whole thing resolves into the line "Break through the noise," then the Noble Growth logo.

This time it nailed it. And all in, across every attempt from the first broken render to this final one, the whole thing cost me about thirty-seven dollars in credits.

What I would do differently, and what comes next

Almost every mistake I made early on is one I will never make again, and they all collapse into two lessons.

Always start with the character reference sheet. Never let the video model guess who your character is shot to shot. Lock the identity first, then generate.

And whenever your character needs to move in a way the original footage does not show, go find real footage of that exact motion and feed it in as a reference. The model is not bad at motion. It just needs to be shown the right one.

I have set all of this up in its own repo so that every time a better model ships, I can push the same character a little closer to the real thing. This is the first in a series. Next week I am taking King Arthur into another ad you will recognize instantly, and I will share a bit of how that one came together too.

If you want to follow along, subscribe and I will see you in the next one.

Below the paywall: the exact platform I used, the specific model behind each step (research, reference sheet, start frames, video, and the throw), and the precise prompts for each stage so you can run the entire workflow yourself.

Read more

Then OpenAI launched Astra, and it was time to recalibrate. For at least two use cases, 3D modeling and image generation, ChatGPT was clearly superior, at least for now. Then Meta launched Muse, and it proved to be a superior personal assistant. Similarly, X’s GrokBot is superior for ongoing multi-agent work and trending news.

The broader point is that having multiple models beats being fully reliant on just one, particularly as the AI race is still hyper competitive and rapidly changing. The gap between the top models overall may be small, but the gap between specific use cases is enormous. Knowing what model to use for what use case is paramount.

Below is what each model does best, from my own experience, at this point in time. These rankings are likely to change. The case for remaining model agnostic will not.

Here is the cheat sheet as of October 2026, then what each model is best used for…

Claude: deep work

Claude is still the best thinker I have access to. When the work is hard and ambiguous, nothing else comes close.

I give it the work that requires holding a lot in its head at once. Planning a brand new website or app from scratch. Diving into analytics for an ad account to determine which optimizations to make next. Generating detailed weekly client reports by pulling in all the relevant data from Google Ads, Meta, Shopify, and GA4, then reconciling it.

That last one matters more than it sounds. Most AI tools can summarize a spreadsheet. Very few can look at three platforms that disagree with each other about the same week of revenue and tell you which one to trust and why.

If the task has real consequences and real ambiguity, it goes to Claude.

ChatGPT: 3D and the visual layer

Right now, ChatGPT is the best model available for 3D modeling and image generation. Since Astra, it is not close.

For iKahan, my out of home advertising client, we build 3D environments of their billboards so brands can see exactly what their creative will look like on that board, from every angle, before they buy. We used to spend hours photoshopping a flat mockup. Now we build the environment, drop the creative in, and take screen captures and video previews automatically.

I also use it for things that have nothing to do with work. My son asked me why the Earth goes around the sun, and why the moon goes around the Earth. As I explained it, I could tell he was having trouble picturing it, and none of the videos or models I could find online were good enough. So I built one: https://noblegrowth.co/solar-system/

The sun, the planets, the moons, all moving through space and time relative to one another, so he can see how it actually fits together.

Anything spatial, anything visual, anything that needs to be rendered rather than described, goes to ChatGPT.

GrokBot: the army

GrokBot is the best at deep multi-agent workflows. It is also the best for anything built on X data, like trending news and what people are talking about in real time.

The idea is to stop thinking of AI as a thing you talk to and start thinking of it as a staff you employ. An SEO bot. A social content bot. A sales bot. A data science bot. And above them, a chief of staff bot that assigns the work, collects the output, and tells you what needs your attention.

Always on, always working, with you as the only human in the loop.

This is the part most people have not caught up to yet. A single chat window has a ceiling, and that ceiling is your own attention. An org chart of agents does not.

Muse: the everyday agent

Muse is the best agent for simpler, more straightforward, lower cognitive load work, and it handles that category better than anything else available. It also has the edge anywhere Meta data comes into play: Instagram, Facebook, and Meta Business Manager.

I used it recently to go through every subscription and bill I pay, find what I was overpaying for or no longer needed, cancel the ones I did not need, and change the plan tier on the ones where I was paying for more than I used. That added up to $1,428 a year in savings. It also ordered new trash cans when the lids on mine broke. Small, useful, finished.

One agent, done elegantly, instead of an army.

Muse hit No. 1 on the App Store ten days after launch, which tells you something. It may have gotten the interface more right than anyone else so far. Its connectors are already being described as a next generation app store, and I think that description will hold up.

How to actually run four models

Most people make the mistake of running their AI only in the desktop apps or website browsers.

The most powerful way to run them is in your terminal, on your MacBook or whatever computer you use. This gives the models access to your files and tools, so they can do almost anything your computer can do, within the permissions you set.

The added benefit is that all your files live on your computer, so you can swap models without losing any of the context you have been working in. When I switch from Claude to Astra, it has access to the same markdown files, project folders, and platform API keys. I can pick up right where I left off. If one model gets stuck on a task, I hand the whole thing to another one and see if it can bring it to the finish line.

That portability is the real unlock. Not the model. The context.

AI training

I have started running AI team training sessions for several of my clients, to bring them up to speed on how to use these tools properly. The hardest part is usually the setup: getting everything running in the terminal, with access to your website, your platform APIs, and GitHub for version control. Once that is done, you talk to these models the way you already do in a chat window. It is that easy, and far more capable.

If you want a session like this for your brand or business, get in touch: https://noblegrowth.co/contact

Mission Control

In the paid section below, a full walkthrough of the solar system build, from the first prompt to the live page: how Astra generated the 3D, how it grew from a handful of inner planets to all eight planets, five dwarf planets, and 465 moons in motion, and how Claude Fable got it onto my site with version control and a full screen mode that holds up on a phone.

Here is exactly how I built it, and how you can too.

Read more

The reason for this is that AI, particularly the newer frontier models, do a far better job when they know the end mission, your actual objective, and can then decide on the most effective means of getting there themselves.

Set a B.H.A.G.

From school I remember this acronym called a B.H.A.G., a Big Hairy Audacious Goal. The idea is akin to the saying: If you shoot for the moon, even if you miss, you end up in the stars. That is particularly vital when it comes to building with AI.

So instead of prompting, “Let’s improve this section, the logo looks a bit off, move it up slightly,” and other little marginal refinements like that, you should start off with the end in mind. Prompt something more like this:

“Make this website world class. Rethink from first principles how to make this site worthy of a design, animation, and copywriting award. It should follow all best practices while also having unique elements that no other site has, so it truly wows every visitor, and converts at a high rate.”

This can also be thought of as reverse engineering success.

Abide by “The Golden Circle”

In a famous TED Talk, Simon Sinek describes “The Golden Circle.” In short, the idea is that in order to be persuasive, you must start with WHY. Don’t start by talking about WHAT product or service you offer, or HOW you build it. Start with WHY you do it.

Directing AI successfully relies on the same principle. Don’t start by telling your AI agent WHAT you want it to do for you, or HOW to do it. Start by telling it WHY. The core mission. The end goal. The purpose of WHY you’re undergoing this effort at all.

The result you’ll get will be far better. And it is always easier to fill in the HOW and the WHAT down the road, once the WHY is set. Doing the reverse gets you technical debt.

Give it room to astound you

Even top AI developers get this wrong sometimes.

Boris Cherny, the creator of Claude Code, has said that the biggest mistake he often makes in building with AI is getting stuck with what the AI model used to be capable of, since newer models are always coming with greater capabilities than before.

Do not limit your mindset regarding the realm of possibilities. Rather than speaking to your AI like it is an intern that needs to be handheld, speak to your AI like it is a world class director. You wouldn’t tell Werner Herzog how to make a film. Though you might tell him a cool idea you have for a film, and why it matters to you and the world. Then you would give him free rein to create that film in accordance with his expertise.

Boris Cherny has also observed, replying to Andrej Karpathy on X, that people who are new to AI are often among the most capable at directing it, because they don’t know what it can’t do yet. In other words, they do not have the curse of knowledge. And because of that, they are free to tell it exactly what they want the AI to build, without any preconceived notions of its limitations. Or, in the words often attributed to Mark Twain, “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.“

This is the worst AI will ever be

It’s true that as of this writing, AI is limited. It can’t do everything yet. And there are still some things that are far better done by a human than an AI, even when it comes to things like web development, design, and copywriting. But it will get better and better over time. This is the worst AI has ever been, right now.

So get into the mindset of thinking big, of telling the AI the end goal you have in mind, of having the highest standards possible and refusing to lower them, while encouraging the AI to achieve that world class quality that used to cost millions of dollars and dozens of employees years to build.

This will give you a major leg up during the next decade of AI adoption.

If you want help with this

Recently I’ve begun teaching friends, family, and colleagues how to optimally build with AI. Websites, mobile apps, automations, anything that can be done with code.

This ranges from me walking you through exactly what to do so you can handle it all on your own, from your own terminal on your laptop, to doing the entire thing for you if you prefer a more direct path.

Either way works for me. Book an intro call with me here: noblegrowth.co/contact

Reach Out to Mattimore

In the paid version of this Substack article below, known as The Control Room, I’ll walk you through a real example of a website I recently built using AI.

The Control Room

Welcome to The Control Room. This is where I show you exactly how a recent build came together: the real project, the real prompts, and the real numbers.

The project: darling.la

Darling is Sean Brock’s first restaurant on the West Coast. If you’ve seen his episode of Chef’s Table on Netflix, you know the level he operates at: live-fire Southern cooking paired with an all-vinyl hi-fi lounge in West Hollywood. The LA Times, Eater, and Forbes all raved about the restaurant when it opened.

The website did not match the room. It was a template site that opened on a brunch promo. The menus were PDFs that neither Google nor AI assistants could read. There was no analytics of any kind installed, so nobody could say what was driving reservations. And the mobile experience took over ten seconds to show its most important content, for a restaurant where nearly every diner decides on their phone.

Here is exactly how I briefed the AI to get this result…

Read more

The Web is Evolving

The First Era (1990 to 2009)

In the early days of the web, every site was custom code. HTML, CSS, a little JavaScript, written by hand. Fast loading. Specific to the brand. The catch was you had to know code, or pay someone who did. A real website meant a real developer. Every Steve Jobs needed a Steve Wozniak.

The Second Era (2010 to 2025)

In the second, the builders took over. Squarespace, Wix, Shopify, Webflow. Drag and drop. Pick a template. Get online without a developer. For most small businesses, this was the right trade for fifteen years.

The Third Era (2026+)

We’re now in the third era. The AI code age. A fully custom site no longer takes a developer six months and six figures. You tell Claude what you want, it writes the code, pushes it to GitHub, deploys it to Cloudflare, and you’re live. From idea to launched site in an afternoon.

I’ve built it this way four times now. My own agency site at noblegrowth.co. Plus three client sites. All are fully custom, load in under a second, and get cited by ChatGPT, Claude, and Perplexity. None of them are on a builder. None of them ever will be.

Why Custom Code Wins Right Now

The case, in order of what matters most.

By building a fully custom site, you’re speaking the native language of computers. Code is what search engines parse. Code is what AI systems read when they decide whether to cite you. A drag and drop interface is a translation layer built for humans who don’t know code. Useful in 2012, a liability now. Elon said that eventually computers will just talk to each other in binary, no coding languages needed. He’s probably right, eventually. For the next decade, code is the lingua franca. Speak it.

The site is fast. Static HTML served from Cloudflare’s edge loads in under a second anywhere in the world. Builders carry the weight of every feature you don’t use. Speed is a ranking signal in Google, a citation signal in ChatGPT, and a conversion signal in your checkout. You want it fast, lean, and functional.

There’s no ceiling. With a template, you bend your brand to fit the platform. With code, you simply update the code to fit your brand. The animation is exactly the one you described. The header sits exactly where you want it. Nothing is “close enough.”

You’re sovereign. This matters most long-term. When your site is custom code in your own GitHub, you stop renting. Ship an update in five minutes. Launch an e-blast from the same codebase. Post to social, run your analytics, track your ads, all from the same place. No need to pay for seven SaaS subscriptions. You can build almost anything yourself. You ride the AI wave instead of waiting for it to crash on top of you.

The Stack

Mission Control

It’s three tools, plus a domain.

A note on the tools: The top tools change. Codex may surpass Claude Code for a while, then Grok may surpass both of them, then a new or open source model surpass all of the above. Whatever happens, you own the code and data for your website. You are not locked to any one provider. When a better model or host shows up, you swap it in and keep moving, always in the pilot’s seat. Set it up this way and you ride every wave instead of starting over on each one.

Claude Code (or Codex or Grok or whatever AI model you prefer) writes the site. You describe what you want, it generates real HTML, CSS, and JavaScript, the same as a senior developer would hand you.

GitHub holds every version. Free. You never have to worry about breaking your site. You can always go back to an older version. You can always share your repo.

Cloudflare Pages serves it to the world in under a second, anywhere. Free.

A domain from GoDaddy points your name at it. A few dollars a year.

You say something like: “Build me a site that looks like this, make it world class, push it in GitHub, and deploy it to Cloudflare so it’s live on my domain.” And it does. The whole loop runs from one terminal on your own computer, for a fraction of what a website builder charges, and it’s infinitely more powerful.

In the paid section below, I walk through how to run this without the pitfalls: the one setting that decides whether it goes smoothly, the order that keeps you safe, and how that same terminal becomes the control center for your entire business.

Once your site lives in this stack, two new things come online.

Your entire site changes with a single prompt. Add a team member, launch a new product, update a disclaimer. One command, every page stays consistent. The bottleneck stops being execution and starts being your judgment.

Your foundation compounds. The site you ship this month gets better next month, not because you redesigned it, but because the model that builds and improves it got better. Every release of Claude lifts your foundation. Opus 4.8 just came out. Therefore I’m updating all of the websites I manage to be even more world-class. My websites become more powerful as AI advances, month over month, year over year.

AI SEO is the new SEO

AI SEO vs Traditional

AI Search Engine Optimization is the discipline of getting cited inside ChatGPT, Claude, Gemini, Grok, and Perplexity. Traditional SEO was about ranking on Google’s ten blue links. AI SEO is about becoming the answer itself.

Custom code is how you compete here. You control your structured data. You control your schema markup. You control your page architecture. The specific things that decide whether AI tools cite you or ignore you. Builders lock you out of most of this. Your own codebase opens all of it.

As the models improve and as you get more fluent with them, your site becomes more visible across the board. In Google’s traditional results. And, more importantly, inside the AI tools your customers are already using to make decisions.

Why Now

The brands that get on this stack early will pull ahead. The ones who stay on legacy platforms will fall behind. Slowly, then all at once. The first-mover advantage is real.

The key is getting a little better each day and staying in motion.

As Alan Watts said, “The only way to make sense out of change is to plunge into it, move with it, and join the dance.”

I’ll leave you with a bit of hopium. Someone recently said that “anything built before 2028 will be inherently valuable.” Hyperbole. But the kernel is real. Any legitimate business that plays its cards right with AI, and sets its foundation up to keep getting better as the technology does, is going to see outsized returns over the next five years.

Now is the time.

If you want help installing this

The full install is custom site, content engine, AI SEO infrastructure, email system, social, analytics. All owned by you. Built to keep getting better as AI improves. I’ve been calling it Mission Control. Ninety days. You own everything at the end.

The way in is a Mission Control Audit. Two weeks, paid up front. You get a written roadmap of your foundation today and the highest-leverage installs to ship in order. If you hire me for the install, the audit fee credits in full toward the engagement.

Book a 20-minute fit call at noblegrowth.co and we’ll see if it’s the right fit.

The wave is here.

Read more

Here’s the rundown for those who haven’t been paying attention to this AI saga that’s been unfolding for several weeks now.

The Set Up

ChatGPT announced on January 16th, 2026 it will incorporate ads into the platform.

Earlier, OpenAI CEO Sam Altman said that ads would be a “last resort” for them.

But the company is under the gun to increase profitability ASAP, given its massive cumulative losses. So, it has turned to ads.

OpenAI promises, of course, to incorporate ads into an ethical and transparent way, similar to sponsored results on Google.

The Ad Battle Begins

Antrhopic, in response, has launched a series of ads comedically jabbing ChatGPT for this decision, all without mentioning them by name. At the same time, Claude is reinforcing it’s own position to never incorporate ads into LLM conversations.

Here’s one Claude ad with a coach metaphor, which ran during the Super Bowl.

And another with a therapist metaphor.

In response, Sam Altman posted a lengthy quote tweet trying to sound casual and unbothered, while also downplaying this vector of critique. It fell pretty flat though, and also had a bit of an air of PR crisis management contrivance.

The Escalation

This advertising battle ramped up amid the backdrop of Claude taking the spotlight recently among hardcore AI builders.

Particularly, Claude Code has shown momentous gains over ChatGPT’s Codex. Developers and vibe coders have begun preferring Claude Code for open source projects like OpenClaw. And the fervor has gotten so intense that Mac Mini’s briefly went out of stock due to so many devs setting up virtual AI Agents running Claude.

However, from a pure market dominance perspective, ChatGPT still leads massively.

Yet Claude appears to be gaining on this.

In addition to their standard ads, Anthropic has also launched a number of influencer collaborations, particularly targeting female-leaning, artsy, hipster audiences, which have been well-received and are cleverly positioned from a marketing perspective.

It feels a little bit like the “I’m a Windows, I’m a Mac” campaign for the AI era.

This dynamic continued to escalate leading up to Super Bowl Sunday, when Claude and ChatGPT both aired prime time ads.

The Super Bowl

Claude ran the coach ad, included towards the top of this Substack post. It resonated well among AI savvy audiences, but for many people it went over their heads.

My wife, for instance, only noticed that it was a robot-sounding Coach, and then saw the Claude logo at the end, so it might not have had the intended impact.

You have to be pretty in-the-know to get that Claude was taking a jab at OpenAI.

OpenAI’s Super Bowl ad, by comparison, came off as confident and tasteful, focusing on the problems that ChatGPT solves rather than its competitors. This is considered best practice for a leading brand in any category.

The “YOU CAN JUST BUILD THINGS” line at the end was perfect.

What Will Happen Next?

Follow Noble Growth Insights for the next great ad saga.

Subscribe now

Before we get into why most ads fail, a disclaimer…

It is possible for a single ad to immediately launch a business into the stratosphere.

The most famous example is Dollar Shave Club.

Dollar Shave Club’s iconic video ad cost just $4,500 to make. The company would later be acquired for $1B.

When razor prices were sky high, Dollar Shave Club did not try to outspend the giants. They used comedy to win attention. Founder Michael Dubin, trained at Upright Citizens Brigade, wrote and starred in a deliberately scrappy, self-aware video that launched on YouTube and immediately went viral. Subscriptions sold out in six hours. The site crashed.

That kind of outcome is real. But it is also very rare.

What often gets missed in these “overnight success” stories is the groundwork. The website, supply chain, and positioning all had to be ready. And the video itself was so strong it likely would have spread organically. What looks like a lightning strike is usually years of preparation. In Dollar Shave Club’s case, that meant Michael Dubin honing his comedic delivery for years at Upright Citizens Brigade.

So while breakout ads can happen, most brands find success a different way.

They build an organic audience first.

Then they sprinkle ad dollars on top to pour rocket fuel on what is already working at smaller scale.

This takes more time. But the result is lasting growth that does not disappear the moment you turn ads off.

With that context set, here is what I have learned from actually running campaigns across industries about why most ads fail, and what actually works.

Why Most Ads Fail

Most ads fail because they are treated like a silver bullet.

They are expected to do everything at once:

• Create demand

• Explain the product

• Build trust

• Differentiate the brand

• Close the sale

All in a single impression.

Instead of pouring rocket fuel on something that is already working, the ad is trying to be the thing that works.

When this happens, the result is usually the same.

Realistically, most companies, especially those working with typical ad agencies, end up with a ROAS below 1.00.

That means for every $1 spent on ads, they earn less than $1 in revenue.

They are paying for the privilege of losing money.

The important thing to understand is that this is not inevitable.

At Noble Growth, we typically see around 2.5× ROAS across accounts, with certain campaigns reaching 10× ROAS once they hit their sweet spot. These results come from following a repeatable system that blends scientific discipline with creative flexibility, and they are achievable as long as the product or service itself is sound.

Here is how that actually looks in practice.

Start With What Is Already Working

Ideally, you begin by building an audience or fan base that already likes your product or service.

This is how you establish product market fit at small scale.

Before you go hunting for new customers, you want proof that someone already wants what you are selling, and that they want it for the reasons you think they do.

A simple place to start is closer than most people realize.

If you are starting at zero today, my advice is to reach out to friends, family, and people you already know in real life. Ask them to check out your startup project. Find out what resonates and what is confusing.

Think organic reach first. Then paid reach second.

Paid ads work best when they amplify existing momentum, not when they are asked to create it from nothing.

First: Inbound Marketing

Your first ad dollars should usually go toward inbound demand.

Start by bidding on your brand name alone.

There are two reasons to do this.

1. First, defense: Competitors can and often do bid on your brand name. Bidding on your own name helps ensure that you show up first.

2. Second, friction reduction: You are making it as easy as possible for people who already want you to find you.

There is also a secondary benefit many people overlook. Being a verified Google advertiser, with strong click through rates and clean account history, often helps reinforce your overall visibility in Google’s ecosystem over time.

Once your brand campaign is profitable, expand into a non-brand keyword campaign.

A good way to plan this is simple. Search your product category on Google and look at what already shows up near the top. Those results are telling you what Google believes best matches user intent. Take inspiration from these results: the best headlines, descriptions, visuals, landing pages, etc, and make them your own.

Inbound marketing works because you are meeting demand, not trying to manufacture it. That is why it is typically the easiest place to achieve ROAS positive campaigns early on.

Then: Outbound Marketing (Carefully)

Once inbound is working, you can move into outbound.

Right now, Meta is the largest outbound platform in the world. Depending on your audience, TikTok, Pinterest, LinkedIn, X, or others may also make sense. But if you’re agnostic I’d recommend starting with Meta Ads.

If you are not sure where to begin as far as your ad creative, one of the best tools available is the Meta Ads Library.

Search for competitors in your space, especially the same companies already ranking highly on Google, or other brands you know are doing meaningful volume.

Then look for the longest running ads.

The reason to look for the longest running ads is simple: Ads that have been running for a long time are usually profitable. They are battle tested. If they were not working, they would have been turned off.

New ads, even from top brands, are often exploratory. Long running ads are proven.

If you take inspiration from the top two or three longest running Meta ads from your top two or three competitors, you give yourself a major head start. You are effectively benefiting from the testing and learning those brands have already paid for.

Run Ads Like a Scientist

One of the most common mistakes I see advertising novices make is messing with their campaigns too much. They are so eager for results, that they feel compelled to change ad parameters constantly, or turn them off too soon, or not know when to declare victory and when to admit defeat and try something different. They let emotions be their guide rather than anything resembling the scientific method.

Take emotion out of the process. Follow the scientific method.

Commit to spending $X per week for 12 weeks, no matter what. This gives you enough time and data to test systematically instead of reacting emotionally to short term noise.

Start with what you believe is the best possible combination of:

• Messaging

• Visuals

• Landing page

• Targeting

• Placements

Then test one new variable per week, keeping only the winners from the previous week.

Think of it like a March Madness bracket. Losers get cut. Winners advance.

The structure should be scientific. But the creative should be artistic.

Create Ads Like an Artist

Just as your testing framework should be rigid, your creative process should be fluid. Don’t worry too much about making it systematic. Follow your intuition, and imagine what would persuade you if you were in the customer’s shoes.

Feel free to test whatever you like. That said, there are some helpful prompts to follow to get you going creatively.

Try:

• Questions

• Statements

• Quotes

• Exclamations

Try:

• Video

• Text overlays

• Carousels

• Static images

Apply all six of Cialdini’s Principles of Persuasion.
Try ads that resemble the best competitors in your space.
Try ads that look like nothing else in your category.

Aim to outdo yourself each week.

As long as your brand messaging and aesthetic stay unified, this combination of creative experimentation inside a disciplined process tends to produce better results over time, typically three months is a solid benchmark to assess your progress.

By then, you are usually well on your way to ROAS positivity, unless there is a deeper issue with product market fit.

The Hidden Benefit: Fast Truth

Let’s consider the worst case scenario for a moment…

You try this system. You commit to 12 weeks of iterative ad testing. You experiment with different creative angles, persuasion angles, targeting angles, and you still don’t see the results you’d hoped for. Your business hasn’t achieved product-market fit.

Even in this case, where product market fit is the issue, this process reveals it faster than almost anything else.

By putting your product in front of a statistically significant number of real people and watching how they behave, you learn the truth quickly.

At that point, you can pivot or persevere.

And with all the data gleaned from your advertising learnings, you’ll have a good sense of which way to pivot, or how best to persevere.

Either way, you are better off.

The Best Ads

The best ads accentuate what is already real.

They start with a truth that already exists in the world and tell it in the most compelling way possible.

The job of advertising is not to fabricate desire or manufacture belief. It is to recognize what is already resonating and bring that story into sharper focus.

Great ads do not sell the product itself.
They sell what life looks like after using it.

More confidence. Less friction. A better version of the customer.

The product or service is simply the vehicle.

This is why ads work best when built on top of organic momentum. The story has already been validated. The transformation is already happening. Paid media helps that truth travel farther and faster.

As they say in copywriting circles: “Facts tell. Stories sell.”

Find what is already true.
Tell the best story about it.
Then scale it.

And when one of your ad campaigns reaches that sweet spot, let it run.

Looking to grow? Contact Noble Growth

Start with the Goal, which is 1,000 True Fans

In 2008, Kevin Kelly published a now-famous essay titled “1,000 True Fans.”

His thesis: A creator doesn’t need millions of followers to make a living, just 1,000 true fans who will buy whatever they produce.

This concept has become foundational in the creator economy and is widely cited across entrepreneurship, marketing, and digital growth. And given the rise of AI, having your own true fans is even more important than before. Otherwise you’re paying for every impression, which is costly and vastly raises your success hurdle.

Enter organic social media growth.

Building your own following on Instagram, X, YouTube, Twitch, Reddit, or whatever other platform feels most natural to you, is immensely valuable. It might not feel like it right away, but the value compounds over time. Eventually, you won’t have to rely on ad dollars. And any ad dollars you do allocate will provide a much higher return.

So, how do you attract these 1,000 true fans? Follow these 5 rules.

But First, Why Buying Followers Doesn’t Work

Before we get into the rules, let’s address the “shortcut” that you may have in mind: buying followers. It’s tempting to just buy followers. Suddenly you have 50,000 Instagram followers, and at a casual glance it looks impressive.

But that’s all it is, surface-level optics. You’re posting to an audience of bots. No real people will see your posts. You’re not getting 1,000 true fans. You’re getting the appearance of fans. And even that illusion only works if you keep paying for bot likes, which is the opposite of organic growth.

There’s also a deeper reason not to buy followers: It nukes your reach in the algorithm.

Every platform is becoming more like TikTok, which means that even small accounts can go viral if their content resonates. This is a good thing. It means any one of your posts has the potential to attract thousands of new followers overnight.

However, if your follower base is made of bots, your posts will most likely only get recommended to… more bots. Therefore, you will have effectively eliminated your changes of going organically viral if you start off by paying for fake followers.

It’s better to have 12 real followers than 120,000 fake ones. Why? Because those 12 real people accurately tell the algorithm who to show your content to next. Bots do not.

Tl;dr: Don’t buy bot followers. And if you already made the mistake of buying bot followers, don’t worry, you can still do things the right way to improve your chances at organic virality by following these 5 rules of organic social growth.

Rule #1: Don’t Fear Being Cringe

The more you post, the better your chances are at growing.

Not all posts will land. And unless you’re saying something as uncontroversial as “water is wet,” someone out there will disagree with your takes. That’s fine.

Be fearless. Be cringe.

Everyone else is mostly thinking about themselves anyway.

Post often. Post confidently. Be the man in the arena.

Rule #2: Be Authentic

The best posts are the ones that flow out of you naturally, not the ones you try to compose. It’s fine to take inspiration from others, but stay true to your own voice.

Authenticity attracts the right followers over the long term. If you spend all your time and energy trying to be like somebody else, chances are you’ll burn out. So, be yourself, post what you find interesting, visually appealing, and worth talking about.

This way, it won’t feel like work. It will feel like texting a friend.

Rule #3: Have a Purpose

Why does your account exist? Why should anyone follow it? What do they expect to see if they do? Who is it for?

Be able to answer all of that in one sentence.

Once you find your niche, growth becomes much easier. Consistency of message, visuals, and ethos helps others recognize who you are and what you stand for. It’s ok to have a variety of interests, but there should be one unifying theme for your account.

Does your account pass the “5 second rule?” Meaning, if someone who’d never seen your account before came across it, would they be able to know what it’s all about with 5 seconds of scanning your profile? If not, you need to have a clearer purpose.

A great way to cement your place in a niche is to interact with and collaborate with top accounts in that niche. If you’re a marketing agency, follow and engage with other top marketing accounts, and you’ll likely get recommended alongside them.

Rule #4: Follow the Golden Rule

Do unto others as you would have others do unto you. This is the golden rule not just for life, but also for organic social media growth.

Be courteous. If someone shows kindness to you, reciprocate. If you enjoy someone’s post, like it, share it, or comment on it. Join other people’s conversations, not just your own. Follow back the real people who follow you.

Reciprocity is the first and most powerful of Cialdini’s persuasion principles.

Most people are so focused on their own posts, and wondering why nobody is commenting on them, yet they never think to comment on other people’s posts. A good exercise is to spend 15 minutes per day making other accounts look good. There’s even a cliche on Twitter/X that says, if you have less than 500 followers, your best bet is to be a “reply guy.” Once you hit critical mass, you can focus more on your own posts. But don’t shy away from engaging with others’ posts and following other accounts before they’ve followed you, especially in the early growth stages.

Reciprocating in this way also helps you find kindred spirits.

Rule #5: Be Consistent

A huge portion of success is simply showing up every day.

I’ve seen successful accounts that post some variation of the same thing every single day, and they are immensely successful. You don’t have to reinvent the wheel each morning. But you do need to post consistently to attract your 1,000 true fans.

This last rule encapsulates the other rules in that, if you fear not being cringe, if you are authentic, if you have a clear purpose, and if you follow the golden rule of reciprocity, you are unlikely to burn out. Your growth won’t feel like work. It will feel like fun, like a natural extension of what you’re already doing day-to-day. It will feel like connecting with old friends each time you log on, steadily achieving your mission.

Case Study: The National Flag Foundation

Noble Growth recently took The National Flag Foundation, a respected nonprofit that’s been around for more than half a century but had no social media footprint, from 143 followers to nearly 20,000 followers on Instagram in just four months.

We did it by following these 5 unwritten rules. And we’re just getting started.

Most marketers get one thing fundamentally wrong:

They make their marketing about themselves, not their customer.

You’ve seen it a thousand times:

“WE just launched…”
“Sign up for OUR newsletter…”
“OUR product is the best…”

Brands love to talk about their product, features, and accomplishments.

But customers only care about a single question:

“What does this do for me?”

This is the moment everything changes.

The Super Mario Insight is that your product is not the product.
Your product is what your customer becomes after using it.

Little Mario = Your customer today
Fire Flower = Your product
Fire Mario = Who they become after using your product

And here is the key insight:

You do not sell the flower.
You sell them becoming Fire Mario.

Most brands talk about the flower.
Great brands talk about what the flower transforms you into.

This is what psychologists call Theory of Mind. It is the ability to model what another person thinks, feels, and cares about from their unique corner of the universe. It is the foundation of all persuasion, advertising creative, and high-performance marketing.

A Campaign That Does This Exceptionally Well

When the “Most Interesting Man in the World” commercial first aired in 2006, it boosted Dos Equis U.S. sales by 22% at a time when most imported beers were declining. And it spawned a viral internet meme format that is still used today.

Part of why it was so successful is it doesn’t sell the beer. In fact, the ad never talks about the qualities of the beer.

Instead, Dos Equis sells who the beer allows you to become. In other words, it sells the dream that you, too, can become the Most Interesting Man in the World.

Lesson there.

A Common Example That Falls Flat

❌ The typical newsletter subscribe box

No transformation. No benefit. No reason.

A Subscribe Box That Actually Works

Here is the better version:

This version answers “WIIFM” instantly:

✔ What I will get
✔ How it helps me
✔ Who I will become

When framed this way, subscribers do not feel like they are doing you a favor.
They feel like they are doing themselves a favor.

Where to Begin ( Start With “WHY?”)

Simon Sinek’s Start With Why is required viewing if you have not seen it.

“People don’t buy what you do. They buy why you do it.”
— Simon Sinek

It all falls into place when your “why” aligns with who the customer wants to become.

The Takeaway for Marketers

To convert the customer, you must become him. Or, as the Buddha puts it:

This post breaks down how to adapt your website to get recommended more often by LLMs, highlighting a few key adjustments to make your business more visible.

In our inaugural Noble Growth Insights post, we shared a chart showing how AI search is predicted to overtake traditional search by 2028. That shift is already well underway.

Your website is no longer just competing for clicks on a search results page. It’s competing to be the answer that AIs deliver. And this trend will only become more prominent over time, so you’re better off adjusting sooner rather than later.

The Case Study: From 0 to 80 AI Hits a Day

One of my clients, a hospitality business, just recently began appearing in AI recommendations after a few key optimizations focused on crawlability and clarity.

Here’s what we changed:

• Added an FAQs page with natural, conversational questions that people are likely to ask to an LLM; and clear, plain-text answers to those questions.

• Added a reviews page written in simple text that AIs can easily summarize.

• Added rich, structured content with clear headings and hierarchy, avoiding hidden text or dropdowns that look nice to humans but are obfuscated for AI.

• Optimized metadata to make the site easier to interpret for AI models.

• Integrated authoritative backlinks to make the website as easy as possible to find.

Result: The site went from 0 to 80 AI hits per day over the first 90 days, and is continuing to trend upwards. 89% of the traffic comes from ChatGPT, 7% from Perplexity, and 4% from Gemini. Many of these interactions led directly to bookings.

Why AI SEO Differs from Traditional SEO

• AIs read differently. They ignore design flourishes and focus on text structure, metadata, and schema.

• Clarity beats cleverness. Straightforward language helps AIs map your content to user intent.

• Questions matter more than keywords. Think like your customer: What would they actually ask an AI?

• Entities and context are key. Use consistent names and schema to identify your business, location, and services.

• Readable reviews matter. Include written reviews in plain text, not just embedded widgets or scripts that look good to humans but are invisible to AIs.

How to Make Your Site AI-Discoverable

1. Add an FAQ page that mirrors the natural phrasing of user questions.

2. Publish a reviews page in plain text rather than relying on third-party embeds.

3. Simplify the site structure so all key text is visible to crawlers.

4. Use schema markup to define your content clearly.

5. Write clear metadata with concise, descriptive titles and meta descriptions.

6. Keep your content fresh to encourage AI models to re-index your site.

7. Backlink to other prominent sites where your business can be found, in particular your Google My Business, your Apple Maps, and your social accounts on Instagram, Facebook, X, LinkedIn, YouTube, TikTok, Pinterest, and Reddit.

8. Track AI referrals separately using analytics tools that identify ChatGPT, Perplexity, and Gemini as traffic sources.

The Takeaway

AI SEO is about finding and widening the path of least resistance between user intent and the information your business provides.

Traditional SEO was about ranking higher in search results. AI SEO is about being clearly understood by the models that now guide discovery and recommendation.

The businesses that adapt first will be the ones most often chosen.

Need help adjusting your website for AI SEO? Email mattimore@noblegrowth.co.

Noble Growth Insights exists to make sure you’re on the path to lasting growth. That’s why I created this Substack.

Each publication, I’ll share what’s actually working (and what’s not) across the brands I manage, spanning e-commerce, lifestyle, tech, fashion, and beyond. You’ll see real data, emerging trends, and practical tactics you can put into play immediately.

We’ll cut through the noise and get to what matters. For instance, recently I’ve noticed a few major changes in digital marketing:

• AI SEO is disrupting traditional SEO

• Semantic search now favors longer, richer captions on Instagram

• Collaborative posts help algorithms determine who your content is for

• Reels and carousels are currently dominating in Meta ads

This isn’t theory. It’s the real pulse of digital marketing in motion.

If you want to stay ahead of the shifts that decide who grows and who fades,
you’re in the right place.

– Mattimore, Founder of Noble Growth