Testing an AI Agent Harness Over a Few Weekends
A short write-up on using Cursor, Kimi K2.5, Laravel, React, and shadcn/ui to test whether a tightly planned agent loop can build a useful app cheaply.
Explore a topic
Thoughts, experiments, and lessons on Coding.
3 articles Newest first
Browse all topicsA short write-up on using Cursor, Kimi K2.5, Laravel, React, and shadcn/ui to test whether a tightly planned agent loop can build a useful app cheaply.
AI coding models and agents are powerful, but they still need tight supervision. Here is the process I use to keep quality high and avoid costly drift.
I just built an iOS app solo. New models like Opus 4.6 and GPT-5.4 make it possible — but only if you bring real developer skills to the table.