Here’s a thought exercise that’s been stuck in my head all week.
Let’s say AI does what a lot of very smart people think it’s going to do. It reshapes labor markets so thoroughly that some form of universal basic income becomes not just a policy idea but a necessity. Everyone ends up earning something close to an average income — enough to live, not enough to be rich.
Now — what does that person want to buy? And more importantly, what does that person want to subscribe to?
Because if you’re thinking about where to put capital right now, that question matters more than almost anything else in the AI discourse.
The Subscription Question Is the Real Question
I keep coming back to subscriptions because that’s where recurring revenue lives, and recurring revenue is what builds durable businesses. In a world where AI compresses the cost of producing nearly everything — content, code, design, analysis — the things people are still willing to pay for monthly tell you what’s actually scarce.
And scarcity is the whole game here.
Think about what’s already happened to journalism. AI can summarize, rewrite, and generate news-style content at essentially zero marginal cost. So why would anyone pay for a news subscription? The answer is trust, curation, and identity — people subscribe to writers and outlets they feel connected to, not to “information” in the abstract. That’s a pretty important distinction.
Now extend that across every category. If AI can generate a decent workout plan, why pay a trainer? If AI can compose music, why pay for Spotify? If AI can tutor your kids, why pay for online courses?
The answer in every case is the same: people pay for the human layer on top. The taste. The accountability. The relationship. The status signal.
Maslow’s Hierarchy, But Make It Post-AI
The useful framework here is Maslow’s hierarchy of needs — not because it’s a perfect model, but because it forces you to think from the bottom up.
Physiological needs — food, shelter, utilities. In a UBI world, these are covered. But people will still pay more for better versions. Meal kits. Nice apartments. Premium basics. This is where most of an average income goes, and AI doesn’t change that much. You still need a roof and calories.
Safety and security — health insurance, financial planning, home security. AI actually makes some of these cheaper to deliver, which means margins compress. But people will pay for the peace of mind of having a HUMAN confirm that their coverage is right, that their money is allocated properly. Trust-as-a-service is a real category.
Belonging — and here’s where it gets interesting. Social platforms, dating apps, community memberships, fan clubs, group fitness. People are lonely, and that’s not getting better. If anything, AI makes it worse by removing more reasons to interact with other humans during the day. The businesses that create genuine belonging — not parasocial content consumption but actual connection — those are going to be disproportionately valuable.
Esteem — status goods, personal branding, credentials, recognition. This is where the celebrity AI question comes in, and it’s a WEIRD one. We’re already seeing AI versions of public figures being built and deployed. The legal framework is way behind. But the consumer demand is real — people want to interact with personalities they admire, even synthetic versions. The esteem need cuts both ways: consumers want proximity to status, and creators want to scale their presence.
Self-actualization — creative tools, learning platforms, coaching, experiences. This might be the most resilient subscription category of all. Because even if AI handles your job, you still want to feel like you’re growing, building, becoming. Peloton isn’t selling exercise — it’s selling the feeling of being someone who works out. Masterclass isn’t selling skills — it’s selling aspiration.
What Average Can Afford
Here’s the constraint that makes this more than just philosophy. If we’re talking about average income — let’s call it $3,000-4,000/month in today’s dollars — you’ve got maybe $200-400 in discretionary subscription spending after housing, food, transport, and basics.
That’s 3-5 subscriptions. Maybe 6-7 if some are cheap.
So the competition isn’t really between your product and your competitor’s product. It’s between your product and everything else that person could subscribe to with their last $50/month. Netflix vs. the gym. Spotify vs. a news site. A coaching app vs. a dating app.
The winners in that fight are the ones closest to the bottom of Maslow’s hierarchy — or the ones that are so good at delivering belonging or esteem that they feel essential.
Where I Think This Lands
I’ve written before about how knowledge ingestion is the killer AI app nobody’s talking about. I still believe that. But the flip side of “AI can know everything” is “what do humans still value enough to pay for?”
And the answer, I think, is pretty clear: humans pay for other humans. For their attention, their judgment, their presence, their identity. AI makes everything else cheaper, which paradoxically makes the human premium MORE valuable, not less.
If you’re building a business right now, or thinking about where to invest, I’d be looking hard at anything that sits at the intersection of “subscription model” and “delivers something AI fundamentally can’t fake.” Connection. Trust. Status. Growth.
That’s what average wants. And average is about to be a VERY large market.
