The argument in marketing circles for the last two years has been: AI makes content production cheaper, so produce more of it. That logic is half-right and mostly dangerous.
What actually changed
AI did not just make content cheaper — it made average content worthless. When anyone with a prompt can generate a 1,500-word blog post on any topic in three minutes, the supply curve shifts infinitely to the right. Price falls to marginal cost, which is close to zero. In economic terms, undifferentiated content is now a free good.
The businesses responding by pumping out more AI-generated content are optimising for an economy that no longer exists. They are manufacturing sand in the desert.
What has value now
When production is free, value shifts to the things AI cannot easily replicate:
Proprietary data and experience. An article that analyses 500 real campaigns you ran has permanent value. An article that summarises publicly available knowledge has none. The moat is your data, not your writing.
Distinctive voice and taste. AI can write competently in any style, but it cannot be Derek Thompson or Rex Sorgatz or your most idiosyncratic operator. Personal voice — specific, opinionated, inconvenient — commands attention because it is an honest signal of a human paying specific attention.
Timely, direct response to what just happened. AI systems have training cutoffs and slow retrieval loops. A human writing about something that happened this week, with informed context, is publishing information the AI ecosystem literally cannot produce yet. Speed of thought is a content advantage.
Embodied reporting. Went to the conference, interviewed the person, watched the thing happen. None of this scales with AI. All of it is worth more now than it was five years ago.
The practical shift
Reduce your content volume. Increase the time you spend per piece. Pay your best operators to write, and stop trying to get your marketing team to "sound like" them. Shift budget from production to distribution — because the bottleneck is not making it, it is getting the right people to see it.
And stop measuring content success by output. Measure it by citations, signals, pipeline, and brand search lift. A single piece that gets referenced in AI responses for a year is worth more than 50 posts that get zero engagement.
The cheap-content era ended about 18 months ago. Most teams have not caught up.
