你刚刷到这条消息，本来准备顺手划走，但又怕自己错过了真正会影响下一步判断的那一点。
最容易做错的，是Iroh 1.0；代价往往是如果只盯表面热闹，你很容易在错误方向上花掉时间、预算和注意力。；我先给一个保守判断：写 Iroh 应用，第一步该删掉 host:port。。
My conservative take is simple: if you are writing an Iroh app, the first thing to delete is host:port. A release is worth your time not by how many features it lists, but by whether it changes your next decision.
Most coverage will talk about peer-to-peer networking, QUIC, or the fact that 1.0 shipped. The more useful signal is narrower. In the June 15, 2026 launch post, the headline is "Dial Keys, not IPs." The current README repeats the same idea: point at a device's public key, meaning its stable identity, and let Iroh find and maintain the best path.
The docs make the contract even clearer. Iroh's Endpoint API gives each node a stable EndpointID, basically a durable ID, and handles address discovery, NAT traversal, and relay fallback. In plain English, the hard question stops being "Where is the server?
" and becomes "Which identity am I trying to reach?"
That is the part I think people should actually discuss when they talk about Iroh 1.0. Not "new transport," but identity before address. My boundary is narrow: this read is based on three official v1.0 sources, the 2026-06-15 launch post, the current README, and the "What is iroh" docs, not a production rollout. If host:port is still leaking into your app design this week, start there. Share this with the person who is still treating Iroh as just another transport upgrade.
真正该讨论的是：Iroh 1.0