Code Is Cheap. Distribution Is the Moat.

In 2025, Collins Dictionary named "vibe coding" its word of the year. A phrase that started as a joke among engineers became the default way a new generation builds software: describe the app in plain English, and a machine writes it.

I am one of those builders. I spent seven years as a Reuters correspondent and five more running brand and content for AI, analytics, and fintech companies. Right now I am building two products by talking to an AI coding assistant. A few years ago that sentence would have sounded absurd. Today it sounds like a Tuesday.

Here is the uncomfortable math for every AI founder reading this: if I can build it, so can ten thousand other people.

The flood

When the cost of making something falls toward zero, supply explodes. We watched it happen with blogs, then podcasts, then mobile apps. Each time, the winners were the people who owned attention when the flood arrived.

Software is having its podcast moment. A week of prompting produces what used to take a funded team a quarter. Every niche now has a dozen AI copilots. Every copilot promises to "transform workflows." Every launch post on LinkedIn uses the same three adjectives.

When everyone can build, the scarce skill is getting chosen.

My read of this market is simple: the value has moved to distribution and to proprietary data. Plenty of AI startups still pour most of their energy into the part of the problem that just got easy.

Three assets that stay scarce

1. A phrase

Categories belong to whoever defines them first and repeats them most. Earlier this year, the leadership team at the company I work with chose a two-word term they wanted to own. I built the blog, the search architecture, and the answer engine structure around it. It reached No. 1 in search in both the US and India.

Once a phrase ranks, every buyer who searches it meets your definition first, and your definition becomes the reference point for the whole market.

2. An audience

At my previous company I grew a LinkedIn newsletter from zero to 73,000 subscribers, with engagement above 20% on every issue. That list became the most valuable marketing asset the company had. Every new report, event, and product announcement started warm.

An audience is the closest thing marketing has to an interest-bearing account. Each week you make a deposit. Each launch you make a withdrawal. The companies that skipped the deposits are now trying to borrow attention at paid-media rates.

3. A place in the answer

Your buyer's first search is increasingly a conversation with a machine. A CFO asks ChatGPT or Perplexity for the best tools in a category and gets a shortlist of three names with a paragraph on each. That paragraph is your new homepage, and you did not write it. When Google shows an AI summary, Pew Research Center found that users click a regular search result in 8% of visits, compared with 15% when no summary appears.

Every month I check how Google's AI answers, ChatGPT, and Perplexity describe the brands I work on. The answers shift. They get facts wrong. They cite sources nobody on the marketing team has read. The brands that win this layer treat it like press coverage: something you earn, monitor, and correct.

The three assets at a glance

Asset	How you build it	How you know it is working
A category phrase	A definition page, a content cluster, and executives who repeat the term	You rank for the term, and AI answers use your definition
An owned audience	A weekly newsletter and a founder voice on LinkedIn	Subscribers grow every week and engagement holds
A place in the AI answer	Clear positioning, consistent descriptions, and earned mentions	Your brand appears in AI answers to buyer questions, described accurately

The two-phase founder

The sharpest founders I follow in 2026 work in two phases. Phase one: vibe code a minimum viable product in a few weeks, on a tool subscription, and put it in front of users. Phase two: once the market responds, bring in engineers and rebuild it properly, using the prototype as a living spec.

Apply the same discipline to distribution. Phase one starts the same week as the prototype:

• Name the problem in public. Write the definition page for the category you want to own before the product has a logo.
• Start the newsletter. One email a week. Ten readers is a start. Ten readers who stay become a hundred.
• Build in public. Share the decisions, the numbers, and the mistakes. Buyers trust the founder they watched make hard calls.
• Ask the machines. Run 20 buyer questions through ChatGPT, Perplexity, and Gemini. Save the answers. Run them again next month.

Phase two arrives when the product is ready. Distribution built this way is ready too.

What this means for you

If you are a founder, the launch plan is now as important as the product roadmap. Put them on the same page, owned by the same people, reviewed in the same meeting.

If you are a marketer, this is the best moment in a decade to be good at your job. The craft skills that looked old fashioned two years ago, writing a sharp definition, building an audience one issue at a time, earning third-party credibility, are the exact skills that decide which AI products get chosen.

Build the product in a weekend. Build the distribution every week after.

Stay on deadline,

Derek