What AI Can’t Kill

Hyperfocal isn’t dead and the products do different things for now, but I can see the trajectory and so can anyone paying attention.

It’s a matter of time before Adobe absorbs the functionality I spent months building, and when that happens, the feature becomes a checkbox in someone else’s product rather than a standalone business.

I got outcompeted by a model update at a company with thousands of engineers, and no amount of indie hustle was going to change that. And that experience has forced me to rethink what’s actually worth building right now.

The Kill Zone

I’m not the only one watching this happen. In early 2026, over a trillion dollars in market capitalization was wiped from software stocks in a single week after investors realized that AI agents are beginning to replace entire product categories.

Analysts started calling it the “SaaSpocalypse” and for once the dramatic label actually fit.

We’ve watched copywriting tools die when ChatGPT got good enough, image generation startups crater when Midjourney improved, and code assistant companies sweat every Claude  Code release. Gartner is predicting that 35% of point-product SaaS tools will be replaced by AI agents by 2030.

If your product is a thin wrapper on top of an AI capability, you’re building in the kill zone.

The moment the underlying model gets better, your value proposition evaporates. And the improvement cycle is accelerating, which means the window between “we have a viable product” and “a platform just absorbed our functionality” is getting shorter every quarter.

This applies to indie builders as much as it does to funded startups. Maybe more, because we don’t have the resources to pivot fast or the runway to survive a sudden irrelevance event.

So the question becomes what you can actually build that survives this.

The Layers That Survive

When I strip away everything AI can replicate or absorb, four layers remain.

• Infrastructure
• Distribution
• Personal brand
• Community.

Each one requires time and human trust to build, which is exactly why a model update can’t shortcut them into existence. At least not yet.

These layers share a common trait. They’re all about relationships and positioning rather than features and functionality, and while a product can theoretically be cloned, the years of accumulated trust behind these layers are extraordinarily hard to replicate.

1. Infrastructure

The plumbing of the internet isn’t going anywhere.

DNS, hosting, email delivery, payment processing, domains, servers. These are the roads and pipes that everything runs on, including AI itself. The models need infrastructure to function.

Companies like Cloudflare, Stripe, AWS, Ready, and Vercel don’t get disrupted by a new model release because the models depend on them.

I’m not suggesting every indie builder go start a hosting company. But the principle scales down.

If you own a piece of the infrastructure that other products or workflows depend on, you’re in a fundamentally different position than someone building a feature that a platform can absorb. The closer you are to the pipes, the safer you are from the waves above.

2. Distribution

This is the layer I’ve been building for over 15 years, and it’s the one I understand best.

Brett and I run DailyPhotoTips with 27,000 subscribers and TheDailyPreset with 9,000. Marko and I run DailyTips.dev for a growing developer audience. All 50/50 partnerships, all built on email, all representing a direct line to real humans who chose to hear from us.

I like how that sounds.

No algorithm sits between us and our audience, and essentially no platform can throttle our reach. No AI update can replicate the trust that 15 years of showing up daily has built.

I wrote about how distribution is the whole game a while back, and the thesis has only gotten stronger since. When anyone can build a product in a weekend using AI, the person who can put it in front of the right audience wins.

Distribution has always mattered, but what’s changed is that it’s now the scarce resource while building has become abundant.

The builders who spent years growing newsletters, cultivating email lists, building social followings, earning karma on Reddit, establishing credibility in niche forums, and developing real relationships with their audience are sitting on the most durable asset in the game right now, even if it doesn’t feel as exciting as shipping a new SaaS.

The big companies are starting to figure this out too. HubSpot acquired Starter Story, a bootstrapped media brand with 800,000 YouTube subscribers and a 275,000-person newsletter, adding it to a portfolio that already includes The Hustle and My First Million. A $12 billion software company looked at the landscape and decided that owning an audience of founders was worth more than building another feature.

Greg Isenberg put it well when he said that media is becoming the base layer of software companies, not as marketing but as infrastructure. When AI keeps making products easier to replicate, the company that controls where attention flows controls where demand flows.

3. Personal Brand

Every founder I know who invested in building their personal brand over the last few years is in a fundamentally different position right now than those who didn’t.

When an AI update kills your product, the founders with a personal brand still have an audience that trusts them as a person. They can pivot to the next thing and bring people with them. The ones who built exclusively behind a faceless company name have to start from zero every time the landscape shifts.

Personal brand accumulates across everything you do.

My photography audience trusts Jon, not just DailyPhotoTips as a URL. Our dev audience trusts me and Marko, not just DailyTips.dev as a domain name. If either of those products disappeared tomorrow, the trust would survive because it’s attached to real people with real opinions and real track records rather than to a logo.

I know this firsthand. When I sold Contrastly, the brand changed hands, but subscribers on DailyPhotoTips and TheDailyPreset still recognized and trusted me because the relationship was with a person, not a product.

AI can generate content that sounds like anyone, but it can’t be anyone.

A real human with real experience, real skin in the game, and real opinions that occasionally piss people off is something people connect with in a way that no model output can replicate. Your name on the emails, your face in the community, your voice on Twitter, all of that is equity that builds over time and that no platform shift can erase.

And personal brand feeds directly into both distribution and community, because people subscribe to newsletters written by people they trust, and they join communities led by people they respect. The brand creates the gravity that holds everything else together.

3. Community

This is the bet I’m making right now, and I think it might be the most durable layer of all.

Email is distribution, and I love email, I’ve been sending daily emails to photography audiences since 2012 and I’ll keep doing it.

But email is fundamentally passive consumption. Someone opens, reads, maybe clicks a link, and goes about their day. The relationship flows in one direction, mostly.

Community is active participation. People talking to each other, helping each other, sharing what they’re working on, forming relationships around shared interests.

The real value of a community extends far beyond what you provide as the creator, because the members start providing value to each other in ways you couldn’t have predicted or manufactured.

AI cannot manufacture that.

You could use AI to generate a thousand forum posts, but you can’t use it to generate the feeling of belonging to a group of people who care about the same things you do. That feeling requires real humans spending real time together, and it requires trust that builds slowly over repeated interactions.

Marko and I are building a developer community on Skool (more on that soon!) alongside DailyTips.dev because we believe this is the layer that’s most resistant to disruption.

Developers and founders will always want to gather, learn from each other, share what they’re building, and talk about the problems they’re solving. The tools and platforms will change, but the fundamental human need to connect with people who share your interests and challenges will not.

I’m not seeing a scenario where an AI model updates and communities suddenly become irrelevant. Humans will always want to gather and talk and learn from each other. The format might evolve, but the need is permanent.

Moving Up the Stack

If you think of these four layers as a stack, the pattern becomes clear.

The product layer sits at the bottom because it’s the most vulnerable.

If your entire business lives there, every model improvement is an existential threat. The move for indie builders right now is to make sure whatever you build sits on top of layers that survive.

I’m not saying stop building products, obviously. Products are how you create value and generate revenue.

But the products should sit on top of distribution you own, a brand people trust, and a community that gives people a reason to stick around even when the product landscape shifts beneath them.

The builders who will thrive in the next few years are the ones who own the layers underneath their products. The ones who will struggle are the ones who built a feature, wrapped it in a landing page, and hoped the product itself would be enough.

The Bet

Hyperfocal might survive, or it might not. Adobe might build exactly what I built, directly into Lightroom, and make my product redundant overnight.

I’ve made peace with that possibility because I’ve been in the game long enough to know that individual products come and go.

But Adobe can’t build my newsletter list. They can’t replicate the trust that Brett and I have built with our photography audience over a decade.

They can’t replicate what Marko and I are building with our developer community, and they can’t replicate the personal brands that hold all of it together.

The products will change, but the layers underneath won’t.

That’s the bet I’m making, and a decade+ of building newsletters have taught me it’s a bet worth taking.