Wix spent a decade as one of the most prominent Scala shops in the industry - ~1,500 engineers, ~3,000 services, no staging environment, deploying straight from laptop to production hundreds of times a day. In 2025, we shipped Java. Not as a rewrite, and not as a retreat from functional programming - but as the natural consequence of a bet we'd made years earlier on platform engineering.

This talk is an honest, outsider-friendly tour of that bet. We'll start with what platform engineering actually is (most teams have never seen it done at this scale), how Wix turned "a new microservice needs ~25 concerns" into a one-line declarative macro, and why "the platform is the product, engineers are the clients." Then the twist: because the platform - not the language - owns those concerns, adopting Java became, on the surface, a feature toggle. We'll go under that surface too: the messy JVM interop, what we borrow from modern Java (21+), and the strategic reasoning that made a Scala stronghold change course.

Finally: in an AI era where anyone can generate anything, is platform engineering obsolete - or more essential than ever? I'll argue, self-critically, for the latter.