Record

Things I pick up.

Notes on building, learning, growing, listening, and becoming—kept in one place as life changes.

Index / topics

Browse by topic

Topics are pointers into the record, not a separate section.

All notes

5 entries

  1. AI

    Pi — The Coding Agent That Chose to Do Less

    Pi is an open-source terminal coding agent that deliberately ships without features most agents consider essential. This article examines why radical minimalism, a TypeScript extension system, and a tree-structured session model make it one of the most versatile coding agent frameworks available.

    • #pi
    • #agent
    • #architecture
    • #developer-tools
    • #open-source
    • #minimalism
  2. Growth

    A User Manual for My Younger Self

    Four things I wish I had known at twenty — about starting before you're ready, choosing environments over prestige, staying connected without performing, and turning curiosity into completion.

    • #personal
    • #mindset
    • #INTP
    • #life-advice
    • #self-reflection
  3. AI

    Dissecting Claude Code — An Architecture Deep Dive

    A deep dive into Claude Code's open-source codebase, covering the three-level paradigm shift, the agent loop with 4-stage context compression, streaming parallel tool execution, 66+ built-in tools across 6 categories, the dual-call skill system, four-type memory taxonomy, and three multi-agent collaboration modes.

    • #claude-code
    • #agent
    • #architecture
    • #developer-tools
    • #open-source
  4. AI

    Post-Training in LLMs — From SFT to Reasoning RL

    A detailed walkthrough of every major post-training paradigm for large language models, from basic SFT through RLHF, RLAIF, DPO, ORPO/IPO/KTO/SimPO, reasoning RL, Constitutional AI, and tool-use training — with formulas, trade-offs, and the modern multi-stage pipeline that leading models use in 2025-2026.

    • #LLM
    • #post-training
    • #RLHF
    • #alignment
    • #fine-tuning
    • #DPO
    • #reasoning
  5. AI

    Three AI Trends Reshaping 2026

    A deep look at three converging AI trends in 2026 — from half-duplex chat to full-duplex real-time interaction, from multi-modal input to world understanding, and from semantic reasoning to predicting and acting in the physical world.

    • #AI-trends
    • #multimodal
    • #agent
    • #full-duplex
    • #world-model
    • #embodied-AI