Experiments Building Auranite: How I Vibe Coded a Map of the Unexplained
A behind-the-scenes look at Auranite.com, from the original idea to the PHP backend, moderation workflow, media pipeline, and interactive world map.
7/13/2026
Blog
- Experiments
-
News GPT-5.6 Sol and the Shift from Code Generation to Code Completion
Why OpenAI's GPT-5.6 Sol matters less as a faster autocomplete model and more as an agent that can carry engineering work to done.
7/12/2026
-
Tips & Tricks Terra, Luna, and the End of Using One AI Model for Everything
A practical model-routing guide for choosing OpenAI's Sol, Terra, or Luna by task risk, complexity, speed, and cost.
7/11/2026
-
News Claude Fable 5 and the Rise of the Multi-Day Coding Agent
What Anthropic's Claude Fable 5 says about long-horizon coding, asynchronous work, visual checks, and the changing role of the developer.
7/10/2026
-
Thoughts Sol vs. Fable: Stop Asking Which Model Is Best
A grounded way to compare GPT-5.6 Sol and Claude Fable 5 using your own tasks, acceptance criteria, review effort, and cost per accepted change.
7/9/2026
-
Thoughts Vibe Coding After Autocomplete: The Agent Is Now the Interface
AI coding has moved from suggesting the next line to operating editors, terminals, browsers, and cloud environments. Here's what that changes.
7/8/2026
-
How-Tos The Production-Readiness Cliff in Vibe-Coded Apps
AI app builders make polished prototypes quickly, but recent research shows why backend integrity, security, and concurrency still need deliberate engineering.
7/7/2026
-
Tips & Tricks Context Engineering Is the New Prompt Engineering
Why files, repository rules, examples, tools, and feedback loops now matter more than finding one magical prompt.
7/6/2026
- Tips & Tricks
Multi-Agent Vibe Coding Without the Chaos
How to use parallel coding agents, worktrees, clear ownership, and an integration pass without multiplying conflicts and confusion.
7/5/2026
-
Thoughts The Metric That Matters: Cost per Solved Coding Task
Token prices are only the beginning. Measure retries, review, cleanup, latency, and accepted changes to understand the real economics of AI coding.
7/4/2026