The Moat Was Made of YAML

AIOpenSourceSaaSConferences
The Moat Was Made of YAML

I gave this talk at Carolina Code Conference 2026 in Greenville. The short version: the moat around every SaaS product you pay for was never the software... it was the operational friction of running the alternative yourself. TLS certs, SSO config, upgrades that don't brick the instance, backups somebody's actually tested, the 2am page. AI turns out to be very good at exactly that kind of YAML, and Jevons' paradox says that when a resource gets radically cheaper you don't use less of it... you expand what's worth doing with it. So self-hosting expands, capture shifts toward mature open source and the people who run it, and the friction that remains is organizational, not technical: before anything reaches production, someone still has to say why it should exist, why it's safe enough, and what they'll do when it's wrong. AI makes the YAML free. It doesn't make the answerability free. The moat was made of YAML... the new moat is made of judgment.

The session recording will be embedded here once Carolina Code Conference publishes it.

The slides

SaaS stocks fell ~20% in Q1... the SaaSpocalypse

Ed Zitron's bear case

Built Ford Inevitable

Anatomy of the moat

AI is great at YAML. That's it. That's the talk.

Jevons, 1865: efficiency went up... so did consumption

Imagination is a function of what is currently cheap

Bifurcation: what loses, what survives, what gains

The homelab

Anthropic's revenue curve

The constraint is probably made of meat

Three questions: why it should exist, why it's safe enough, what you'll do when it's wrong

The accountability contract

42% of committed code is AI-generated or assisted

The honest caveats

The moat was made of YAML. The new moat is made of judgment.

Q&A / where to find me