Tech notes
The article presents a layered roadmap of the 2026 Java AI ecosystem, compares major AI frameworks, LLMs, embedding models, vector databases, and agent toolchains, and offers three concrete stack configurations with cost estimates and practical configuration snippets for architects and technical leaders.
This article presents a comprehensive six‑layer AI Agent architecture, explains the underlying principles of reasoning, tool use, memory, and planning, compares ReAct, Function Calling, and MCP, walks through a real‑world request flow, and offers practical technology‑selection guidance.
This article enumerates essential operational metrics—such as CPU, memory, disk and network I/O, response time, throughput, error rates, availability, MTBF/MTTR, security logs, and capacity‑planning indicators—explaining their meanings and recommended target values to help engineers comprehensively monitor system performance, stability, and efficiency.
The article outlines why monitoring key IT operations metrics is vital for performance, reliability, and cost control, then details 17 common metrics—including availability, failure rate, MTTR, MTBF, response time, throughput, error rate, capacity utilization, latency, data integrity, success rates, waiting time, backup success, recovery time, security patch time, server and network bandwidth utilization—providing definitions, calculation formulas, typical reference values, and applicable scenarios.
The article argues that AI agent development is moving beyond improving model performance toward engineering reliable software systems, introducing concepts such as Harness, Loop, and Graph engineering, and explains how organizing agents, feedback loops, and dependency graphs can turn AI agents into robust, verifiable applications.
The article explains that AI agent failures are rarely due to model intelligence and instead stem from three engineering layers—Harness, Loop, and Graph—detailing how tool access, verification loops, and workflow graphs affect reliability, and provides a checklist for diagnosing which layer is broken.
The article provides a detailed technical comparison of SRAM, DRAM and HBM, explains their roles in modern AI GPUs such as the H100, discusses training and inference memory placement, and evaluates emerging ideas like High‑Bandwidth SRAM and High‑Bandwidth Flash, concluding that all three memory types will coexist in future AI systems.
The article breaks down five durability tiers for database writes—from in‑memory acknowledgment to cross‑availability‑zone replication—explaining how the point at which success is reported determines latency, fault tolerance, and the hidden coordination work a system must perform.
The article explains how moving from single‑loop (Loop Engineering) to interconnected loops forming a graph (Graph Engineering) can prevent agents from optimizing narrow metrics at the expense of real business value, outlining risks like Goodhart’s law and offering a four‑step framework for supervision, conflict resolution, and independent auditing.
The article explains how Linux’s single‑root filesystem, governed by the Filesystem Hierarchy Standard (FHS), resolves early Unix directory chaos through hierarchical organization, functional separation, predictability, and economic design, and details the responsibilities of each standard directory.
