A solo AI operator tested llama.cpp, OpenRouter, Gemini batch, and OpenAI Batch under one strict rule for the 2asy.ai extraction pipeline. What broke and what fit.
I Built a Local LLM Rig to Escape API Bills. Then I Paid OpenAI Again.
Related
I built a "what do I need on my final" calculator with AI auto-fill (Next.js 16 + Gemini)
Every finals season I'd end up doing the same algebra on a notes app: "I have a 91% in homework...
Loopcraft: Stop Prompting, Start Designing Loops
The shift from prompt engineering to loop engineering. How Karpathy, Steipete, and Boris Cherny design systems that prompt their agents.
練習跟同事的AI Agent協作
> 2026 年 3 月 | 整理自社群公開發表的實戰經驗與開源專案