AI makes everything easy — and that's why it won't give you an edge

Executive overview

When a capability becomes cheap and abundant, it loses its value. AI is doing this to content, marketing, and software simultaneously — and most businesses haven't updated their strategy to match.

The only durable competitive advantage has always been whatever is hard and rare. AI doesn't change that principle; it changes which things qualify.

The value was never in the thing — it was always in the difficulty of getting the thing.

The abundance trap

• When something becomes abundant, it stops being scarce, and scarcity is what creates value.
• Seeing AI save you five grand on content doesn't mean you saved five grand — that content is now worth close to nothing.
• The "gold from thin air" narrative ignores that universal access to gold makes gold worthless.
• Knowing this intellectually and behaving accordingly are different things — most people are still acting as if the old scarcity holds.

Content: scarcity was the value, not the quality

• Good content in 2023 worked because most people couldn't produce it — standing out was the mechanism.
• AI lets everyone produce good (or great) content, which means nobody stands out.
• Photography is the precedent: smartphones took annual photos from ~80 billion to 1.4 trillion, and made photography a commodity for everyone except those whose vision can't be replicated.
• Action step: stop asking how to produce more; ask what you know from specific experience that an AI or competitor cannot replicate.
• Scarcity is the content strategy now, not volume.

Marketing automation: we've watched this film before

• Cold calling worked when it required genuine effort; auto-dialers made it easy and people stopped answering.
• Early email got 40–50% open rates when it was novel; automation flooded inboxes and average rates dropped to ~20% even for opted-in lists.
• The first banner ad (HotWired, 1994) got a 44% click-through rate; today's average is around 0.1%.
• The pattern is consistent: the moment a channel becomes easy to scale, it stops working — not because it was bad, but because it was no longer rare.
• AI agents automating outreach will produce the same result at speed.
• Action step: if your marketing can be fully automated, it's already dying. Find activities that are hard, slow, and human — friction is the feature.

Software: zero economic profit incoming

• Anyone can now build internal tools with AI — that's a genuine productivity gain.
• Building software to sell is different: if you can build it, so can everyone else, and so can the major AI providers inside their own platforms.
• This is zero economic profit — barriers to entry drop to near zero, competition floods in, margins disappear (Porter's threat of new entrants).
• Same reason you can't make money selling white t-shirts: demand exists, but anyone can do it.
• Action step: ask whether someone with a ChatGPT or Claude subscription could rebuild your product in a weekend. If yes, it's a feature, not a business.

What still qualifies as hard and rare

• Relationships — being in a room with someone, building trust over time, reading people. Can't be faked or automated. Rising restaurant prices, booming live events, and office returns all reflect this.
• Judgment — sitting with genuinely ambiguous situations where data tells you nothing useful, then making a call and owning it. AI gives analysis and options; it can't look your board in the eye and say "we're doing this."
• Original thinking — seeing what the data doesn't show, connecting dots no one else has connected, holding a contrarian position when everyone disagrees. The scarcest resource in business; always has been.

The right question to ask yourself

• Not "what are you good at?" — loads of things people are good at can now be done by AI or by each other.
• The question is: what is the thing that only I can do?
• What would be genuinely hard to replace even in a world full of AI?
• If you can't answer clearly, that's the strategic problem — not your AI stack or automation setup.
• AI tools can serve as nodes within higher-value activities, handling support functions efficiently. But the value lives in the hard, rare thing they support — not in the tools themselves.