Over the past decade, we’ve dramatically improved how software gets built. AI can now generate code, help design systems, and accelerate development workflows in ways that were unimaginable just a few years ago.
But one part of the process still feels stuck: deploying and operating infrastructure.
Even today, going from “working code” to “live application” often involves navigating cloud dashboards, configuring environments, managing permissions, and wiring together tools. That operational overhead slows down experimentation and adds unnecessary friction.
Recently, I wanted to explore what deployment could look like if we treated infrastructure as something AI could reason about — not just something we configure manually.
So I tried deploying an application using Claude together with NexusAI.
Top comments (0)