Catatan

Solving AI Context Decay with Google's Open Knowledge Format (OKF) For Deep State of Mind (DSOM) For My AI

In the world of Open Source and AI development, we are always learning and improving. Today, I want to share  a major milestone in our Deep State of Mind (DSOM) project    ( https://github.com/linuxmalaysia/deep-state-of-mind-for-my-ai ) :    We have officially adopted Google Cloud's Open Knowledge Format (OKF) v0.1 . Why do we need OKF? If you have been following the DSOM project, you know that our biggest challenge with autonomous AI agents  is Context Decay . Over time, AI agents forget the architecture, the rules, and the operational history of a project. Historically, organizational knowledge—like runbooks, architecture decisions, and API schemas—is scattered  across different wikis, databases, and code comments. OKF solves this by creating a standardized  "common language" for knowledge. It structures documentation so that it is equally readable by human engineers  and autonomous AI agents.   Think of it like a cluster for y...

The Art of the Start: Intelligence Audit & State Sync in AI-Agentic Workflows

  The Art of the Start: Intelligence Audit & State Sync in AI-Agentic Workflows In the fast-paced world of AI-assisted development, the biggest challenge isn't writing the code—it's maintaining the "State of Mind" across sessions. Have you ever returned to a project after a break, only to have your AI assistant forget the architectural nuances or the specific status of a complex refactor? For CMSForNerd v3.5 , we've solved this with a rigorous protocol called the Intelligence Audit & State Sync . The Problem: Context Decay AI agents are brilliant but ephemeral. Every time a new conversation starts, they "wake up" in a vacuum. Even with a codebase to read, they often miss the "Why" behind recent changes or the "In-Progress" threads that aren't yet committed. This leads to Architectural Drift , where the AI suggests "standard" solutions that violate your laboratory's strict rules (like our Zero-Global or Pair...

Evolution of a PHP Laboratory: Moving CMSForNerd from v3.1 to v3.3 (PHP 8.4) - Written by AI

  Evolution of a PHP Laboratory: Moving CMSForNerd from v3.1 to v3.3 (PHP 8.4) In my latest development cycle for CMSForNerd , I’ve taken the project through a significant architectural shift. What started as a lightweight database-free CMS has now evolved into a high-standard "PHP Laboratory" designed for the 2025 engineering landscape. Here is a breakdown of what has changed between version 3.1 and the newly minted version 3.3. https://github.com/CMSForNerd/CmsForNerd 🚀 Version 3.1: Strengthening the Foundation In version 3.1, our focus was Security and Standards . We realized that even a "simple" CMS must follow enterprise-grade rules if it's meant to teach the next generation of developers. Key Milestones: SecurityUtils Implementation: We moved away from scattered htmlspecialchars calls and created a centralized security class to handle path traversal and input sanitization. PSR-12 Compliance: We enforced strict coding styles. If the code doesn't loo...

E: Sub-process returned an error code" During APT Upgrade: A Deep Dive into Cache and List Corruption

Imej
Solving E: Sub-process returned an error code During APT Upgrade: A Deep Dive into Cache and List Corruption" The frustrating error, E: Sub-process returned an error code, completely cripple the apt package manager. My own extensive troubleshooting revealed the root cause wasn't a problematic package, but rather a failing package manager hook specifically tied to etckeeper. Recently, I encountered this exact issue on a server running on  Linux Ubuntu 24.04 server while attempting to upgrade elasticsearch fleet node. Cleanup of APT's local cache and package lists does not work.  1. The Symptoms and The Misleading Logs The problem was I could not perform any package transaction (install, remove, or upgrade), with the error appearing immediately after files were downloaded but before installation started. Failure Point: The system exited with E: Sub-process returned an error code. Failed Debugging: Standard troubleshooting (dpkg --configure -a, checking term.log, and even jou...

Panduan Ringkas: Pemasangan MinIO di Ubuntu 24.04 Menggunakan Podman

Panduan Ringkas: Pemasangan MinIO di Ubuntu 24.04 Menggunakan Podman Tujuan: Dokumen ini adalah versi yang sesuai untuk dikongsi secara awam. Semua alamat IP sebenar, nama pengguna, dan kata laluan telah digantikan dengan contoh/tempahan supaya tidak mendedahkan maklumat sensitif. Nota penting: Sentiasa gantikan nilai contoh di bawah ( <...> ) dengan nilai sebenar yang selamat dan unik apabila mengimplementasikan pada persekitaran produksi. Gambaran keseluruhan Panduan ini menerangkan langkah utama untuk: Memasang Podman dan komponen rangkaian rootless Menala parameter kernel (sysctl) Menyediakan pengguna khusus dan direktori storan Menjana sijil TLS menggunakan CA dalaman (contoh) Membuat unit systemd untuk menjalankan MinIO dalam Podman Pod (rootless) Mengkonfigurasi MinIO Client ( mc ) dan polisi pengguna Semua konfigurasi yang mengandungi alamat, nama pengguna, atau kata laluan telah ditukar kepada contoh seperti minio.example.com , exampleuser , dan REPLACE_W...

The Self-Contained Research and Verification Workflow

My usual process for research using NotebookLM is to first start by uploading my own verified sources . I then ask the system many questions and immediately compile all the key answers into a Note . I do this because the questions and answers in the main NotebookLM chat panel (the centre section) tend to be cleared when you access the notebook again, so I make sure to save everything into the Note (the right-hand panel) for permanent keeping. When I feel I have run out of initial questions to ask to further develop my notes, I then use the following prompt in the NotebookLM chat : 'I want to ask Google Gemini, what questions that I need to ask to improve my technical propose so I also can enrich my notes.' The list of critique questions that NotebookLM generates in response is then what I take and pose to Google Gemini to find the answers. I will then insert those answers into my Note. Crucially, the completed Note that we have gathered can eventually be saved and re-uploa...

Mini UbuCon Malaysia 2025 - An Event Dedicated To Advancing The Open Source Software Ecosystem

Mini UbuCon Malaysia 2025 We are thrilled to announce the Mini UbuCon Malaysia 2025 , an event dedicated to advancing the open-source software ecosystem in our region. This year, the Ubuntu Malaysia LoCo Team is proud to partner with Cyber Digital Services, Defence and Security Asia (CyberDSA) , a leading regional event serving key stakeholders in the cyber defence and security industry. Please save the date and venue for this event: Event: Mini UbuCon Malaysia 2025 Dates: 30 September 2025 Venue: Kompleks MITEC KL Metropolis, 8, Jalan Dutamas 2, Segambut, 50480 Kuala Lumpur For more information and registration, please visit: Mini UbuCon Malaysia 2025: https://2025.ubuntu.my/ CyberDSA: https://www.cyberdsa.com/ Through this collaboration, CyberDSA is the Official Place Sponsor for Mini UbuCon Malaysia 2025. This sponsorship has made it possible for us to select Kompleks MITEC KL Metropolis as our official venue, a premier convention center that will provide an exceptional ...