When last we spoke (CurryOn Prague, 2015), Pony had just been open sourced. Since then, we have developed a small but thriving community that mostly uses Pony for low-latency, high-throughput, concurrent server-side computing. Let’s talk about how Pony’s type system and runtime put actors to work to deliver a fast, predictable, GC’d, native-code, actor-model environment, including: data-race freedom, optimised binaries, the workstealing scheduler, fast zero-copy messaging, and efficient no-stop-the-world GC. And let’s talk about how industry users and academic researchers have been working together to improve Pony for "real-world" projects.