Platform as a Product, 9 years later

As part of the updating our platform as a product paper (almost ten years old, and now updated for 2026), I interviewed one of the people involved in it from State Farm and Pivotal, and now MomentumAI, going over how the concept has changed since then from platform engineering, to internal developer portals (remember when the ‘P’ in “IDP” meant “portal”? what a whacky time to have been alive!), and so on. Here’s the interviews:

Some fun quotes

“Your build pipeline and your platform are, like, always in a cold war together.” — Coté, 12:07

“‘Give us a Kubernetes cluster and get out of our way.’ It’s not that that’s what they wanted, it’s what they knew you would give them — so that you’re not the barrier to them building their own platform for their own set of applications.” — Caleb Washburn, 21:17

“The portal became the human API to getting all these things done, which was good, ’cause it helped with productivity. But now the human is maybe not involved in this initial project setup. Now I’m orchestrating with an AI agent that I wanna go build a new product.” — Caleb Washburn, 26:08

“Take my application, give me a URL that terminates TLS and adds it to DNS. If you solve that problem pretty elegantly, you’re solving like 80% of what people are deploying applications for.” — Caleb Washburn, 28:55

“It’s like, ‘Do you know who your developers are?’ They’re like, ‘We’ve never talked to those humans ever.’” — Caleb Washburn, 31:52

“If you had a menu and no one ever bought anything off the menu, would you view that as a successful restaurant?” — Caleb Washburn, 34:32

Check out the paper, too. Like many great ideas, not much has changed over the years, just gotten better, so you’ll find the overview of platform as a product helpful.