Abadugu_blog

My default is local agents. When a task needs more compute, I use OpenRouter to reach models that cost less than the frontier APIs while still covering most of my coding and text work.

A quick look at Typst and how it compares to markdown and LaTex.

Thoughts about the state of being ok when you cannot make two things exactly the same.

There is a lot of synthetic data that is being generated by LLMs, These include false metrics.

Creating a practical, no nonsense guide for using ChatGPT-5 efficiently by manual model switching and prompt style to each task while avoiding wasted tokens and bad outputs.

OpenAI’s removal of GPT-4o, o3, and other models after GPT-5’s launch breaks fundamental MLOps principles. Without model versioning and control, data science workflows become unreliable. Local LLMs offer a better alternative for maintaining consistency.

Ongoing guide to replace default bloated and worse apps on Android with lightweight and FOSS versions.

How I rebuilt my search strategy in 2025 after declining quality in search results

I sent a picture to ChatGPT for OCR and it gave me a response that leaked from somewhere.

Exploring the advantages of hosting your own blog versus relying on platforms like Substack and Medium