Posts

Written down, so it isn't forgotten.

• No Longer Writing for Utility: Upgrading Both Content and Framework

  14 Jan, 2026

  When the requirement shifts from a "public resume" to "long-term writing," the tech stack should also switch to something lighter and handier.

• Hands-On with Gemini 3 Pro: Rapidly Developing an Auto-Battler Game from Scratch

  19 Nov, 2025

  Gemini 3 Pro SOTA Model Capability Test

• Immich GPU Acceleration: Empowering the Home AI Photo Library with Integrated Graphics

  14 Nov, 2025

  Engraving memories with computing power.

• Live Photo and HDR Media Library Compression Scheme

  8 Sep, 2025

  Existence precedes perfection.

• Travelogue of Japan

  30 Jul, 2025

  Scenery has no fixed bounds; it is revealed by the mind’s illumination.

• From Megabytes to Megawatts: A Comprehensive Guide to High-Performance LLM and Diffusion Kernels with CUDA and Triton

  17 Jul, 2025

  All content is generated by LLM, please exercise discretion.

• From Zero to LLM: A Complete Guide to the Large Language Model Tech Tree

  2 Jul, 2025

  All content herein was generated by an LLM and compiled into this document to facilitate sharing with family and friends.

• Tinkering with the Cudy TR3000 and ImmortalWrt

  24 Jun, 2025

  How to set up ImmortalWrt on the Cudy TR3000 router.

• Solution for "Could not restore this iPhone because an error occurred" When Restoring iOS from Mac/iTunes Backup

  26 May, 2025

  Data is invaluable.

• The HomeLab Power Diet: How to Encode AV1 Without Burning Your Wallet

  3 Apr, 2025

  Pro tip: This guide has been seasoned with DeepSeek V3's special humor sauce.