The builder behind NextGen.ing. (Click the photo - it changes.)
RJ has been writing software for two decades - long enough to have shipped in Java, JavaScript, Node, .NET, and C++, and long enough to be skeptical of anything pitched as a silver bullet. NextGen.ing is where he works out, in public, what agentic development actually looks like when you take it seriously.
The short version of his view: agents are extraordinary at producing code and genuinely bad at owning judgment. So he runs a multi-model setup - sending each task to whichever model does it best - and keeps the human firmly on intent, architecture, and review. The highest-leverage thing an experienced engineer does in 2026 is not type faster; it is know where AI helps and where it quietly makes things worse.
He still spins up local environments and prototypes alongside the work, mostly to hunt down the repetitive tasks worth automating - a quick script here, a custom MCP server there. He cares as much about quality as about velocity, on the theory that a release which ships on time but does not work is just a cleanup project wearing a deadline. Tests, AI-assisted review, and real feedback loops are how he keeps both.
NextGen.ing is one of a family of sites where he writes about and builds for the AI-native era. None of this is a pitch - it is a working notebook, kept in the open.
Engineering and the craft of building software in the agentic era.
Notes on AI-native software development and where the field is heading.
Free, privacy-first developer utilities that run entirely in your browser.
The short bio and a map to everything else RJ is building.
Code and open-source experiments at github.com/rjlsoftware.
Professional background and the longer story at linkedin.com/in/rjlindelof.