"GTM in git" (John Kutay, Rippling), assessed 2026-10-04
Why this is in the vault
Founder shared it (iMessage 2026-10-05 00:10 ET). It independently states the Denominator post #1 thesis (shared definitions across agents) and gives a scale example for Fifth Job post 3 (agents can find customers).
Summary (paraphrase)
Talk turned article. Claim: coding agents are the best agents; put GTM operations in a form they can reason over (a git monorepo holding business context, skills that query read replicas). Marketers start working like AI engineers: define evals, validate data, inspect failures. Rippling's GrowthOS hosts agents, data and workflows for sellers and marketers: 2,000+ internal users, 1M+ AI prompts per month, pipelines at 1B+ rows/day (self-reported). Key line on definitions: the same account definition should be interpreted the same way by a coding agent and a seller's agent; corrections should improve the shared system. Cites Ian Macomber, "The Shape and Feel of the Post-AI Data Stack" (iandmacomber.com/blog/post-ai-data-stack).
Mapping against Ray Data Co
- Denominator post #1 ("agents need your definitions"): outside, at-scale corroboration.
- Fifth Job post 3 (agents find customers; a human vouches): example of agents doing research, audience building, signal testing.
- RDCO itself: our vault + skills pattern is the same "context in git" shape.
- Ship 30 topic candidate: "Your definitions belong in git."
Sponsorship
Self-promotional: promotes Rippling's GrowthOS (author works there). Scale numbers are self-reported; "coding agents are the best agents" is opinion.