The tests passed. The feature had never worked.
With an AI assistant, the test, the example and the docs may all come from the process that wrote the feature. Two ways to check.
147 articles in this category · 468,808 views.
RSS feed
With an AI assistant, the test, the example and the docs may all come from the process that wrote the feature. Two ways to check.

Nicolas Duminil
78 viewsThe same small AI service in Spring AI, LangChain4j on Spring Boot and on Quarkus, held to one shared test. Where they differ, and where …

Anton Epple
71 viewsMany LLM calls in Java apps are multiple choice. dukehoff-decide answers them with a small local model and returns a typed value with …

Viktoria Evdokimova
92 viewsA practical, tool-agnostic workflow for specs, tests, review loops, worktrees, and team-owned configuration when coding agents join …

My job portal asks Gemini why a job fits a candidate. Every claim must quote its source, and a few lines of plain Java drop any quote that …

Luis Majano
87 viewsBoxLang AI 3.6 connects AI agents to browser automation, audio tools, streaming speech, MCP stdio servers, MariaDB vector memory, and …

Anton Epple
199 viewsCoding agents rename methods with grep and edit loops. A headless Apache NetBeans gives them real refactorings over MCP, with fewer turns …

Frank Delporte
113 viewsEvery new Foojay article is announced on Mastodon, Bluesky and LinkedIn by one small JBang script. How it works and which problems we had to …

Oumaima Zerouali
141 viewsSix new JCast episodes from June to October 2026: AI hype and agents, clean code, cloud costs, Quarkus with Kevin Dubois, and meeting-heavy …

Luis Majano
186 viewsBoxLang 1.18.0 adds Azure Functions, cluster-safe scheduled tasks, a more automation-friendly CLI, and refreshed docs for developers and AI …