All Articles

143663 articles · Page 451 of 7184

Apr 23, 2026 · Artificial Intelligence

Architectural Fixes for LLM Hallucinations: Inference Parameters, RAG, Constrained Decoding, and Post‑Generation Validation

The article breaks down LLM hallucination mitigation into five layers—runtime inference parameters, retrieval‑augmented generation and prompting tricks, constrained decoding with confidence calibration, post‑generation verification checks, and domain‑specific fine‑tuning plus continuous evaluation—showing how each layer reduces false, confident outputs.

LLMRAGconstrained decoding

0 likes · 11 min read

Architectural Fixes for LLM Hallucinations: Inference Parameters, RAG, Constrained Decoding, and Post‑Generation Validation

Smart Workplace Lab

Apr 23, 2026 · Artificial Intelligence

Think Standard Scripts Solve It? Uncover the Real Issue with High‑EQ AI Prompt Tuning

The article explains why using formal, standard language makes AI‑generated workplace messages sound robotic and presents a three‑step protocol—high‑quality phrase extraction, persona‑mapping prompts, and forbidden‑word rules—to feed the model with emotionally intelligent corpora for more natural communication.

AI Prompt EngineeringLanguage ModelPrompt design

0 likes · 5 min read

Think Standard Scripts Solve It? Uncover the Real Issue with High‑EQ AI Prompt Tuning

Raymond Ops

Apr 23, 2026 · Operations

Advanced Nginx Load Balancing: How to Choose and Tune Layer 4 vs Layer 7

This guide walks through the differences between 4‑layer (TCP) and 7‑layer (HTTP) load balancing in Nginx, explains when to use each, and provides step‑by‑step configuration examples, health‑check setups, performance tuning, SSL handling, WebSocket support, and common pitfalls.

ConfigurationLayer 4Layer 7

0 likes · 25 min read

Advanced Nginx Load Balancing: How to Choose and Tune Layer 4 vs Layer 7

Meituan Technology Team

Apr 23, 2026 · Artificial Intelligence

LARYBench Introduces an ImageNet‑Style Benchmark for Embodied Action Representations Learned from Human Video

LARYBench (Latent Action Representation Yielding Benchmark) provides the first systematic, ImageNet‑scale evaluation for implicit action representations derived from large‑scale human video, decoupling representation quality from downstream control, and shows that general‑purpose vision models outperform specialized embodied models in both action generalization and control precision across diverse robot morphologies and environments.

Vision-Language-Actionaction representationbenchmark

0 likes · 13 min read

LARYBench Introduces an ImageNet‑Style Benchmark for Embodied Action Representations Learned from Human Video

mikechen

Apr 23, 2026 · Backend Development

How Nginx Handles One Million Concurrent Connections: 4 Key Techniques

The article explains how Nginx sustains one million simultaneous connections by using asynchronous non‑blocking I/O, a robust multi‑process architecture, zero‑copy file transmission, and optimized caching strategies, with concrete configuration examples and performance reasoning.

Zero-Copyasynchronous I/Ocaching

0 likes · 4 min read

How Nginx Handles One Million Concurrent Connections: 4 Key Techniques

Top Architect

Apr 23, 2026 · Backend Development

Replacing MyBatis with MyBatis‑Plus Can Trigger LocalDateTime Conversion Errors

A newcomer swapped MyBatis for MyBatis‑Plus in an old Java project, which caused a runtime exception about unsupported java.time.LocalDateTime conversion; the root cause was a MyBatis version change and an outdated MySQL JDBC driver, fixed by upgrading the driver.

JDBCMyBatisMySQL

0 likes · 9 min read

Replacing MyBatis with MyBatis‑Plus Can Trigger LocalDateTime Conversion Errors

Top Architect

Apr 23, 2026 · Industry Insights

Inside the OpenAI Codex Leak: GPT‑5.5, Glacier, Heisenberg and What They Reveal

A recent OpenAI Codex leak exposed internal models—including GPT‑5.5, Glacier, Heisenberg and Arcanine—triggering analysis of the accidental staging‑to‑production push, the shift toward agentic AI, speculative new architectures, and the community debate over whether the incident was a genuine engineering mishap or a calculated marketing move.

AIGPT-5.5Glacier

0 likes · 11 min read

Inside the OpenAI Codex Leak: GPT‑5.5, Glacier, Heisenberg and What They Reveal

Top Architect

Apr 23, 2026 · Fundamentals

12 Surefire Ways to Write Unmaintainable Code

The article lists twelve concrete anti‑patterns—such as over‑splitting microservices, writing megamethods, deep nesting, misleading comments, copy‑pasting code, ignoring logging, and over‑engineering with heavyweight frameworks—that dramatically reduce code readability and make maintenance a nightmare, illustrated through the fictional programmer Er Gou and his bewildered teammates.

Software Architecturebad practicescode quality

0 likes · 13 min read

12 Surefire Ways to Write Unmaintainable Code

Machine Heart

Apr 23, 2026 · Artificial Intelligence

First Survey of Attention Sink: From Utilization and Understanding to Elimination in Transformers

This survey reviews over 180 papers on the Attention Sink phenomenon in Transformers, outlining its three-stage evolution—from early exploitation to mechanistic interpretation and finally strategic mitigation—while detailing utilization tactics, theoretical explanations, removal techniques, and promising future research directions.

Attention SinkTransformersmitigation

0 likes · 9 min read

First Survey of Attention Sink: From Utilization and Understanding to Elimination in Transformers

Machine Heart

Apr 23, 2026 · Artificial Intelligence

DeepSeek Unveils Tile Kernels and DeepEP V2 – Is V4 on the Horizon?

DeepSeek recently opened the Tile Kernels repository and released DeepEP V2, detailing new GPU kernel features, a fully JIT-enabled expert parallelism redesign that boosts peak performance by up to 1.3× while cutting SM usage fourfold, and hinting at an upcoming V4 release.

DeepEP V2DeepSeekExpert Parallelism

0 likes · 6 min read

DeepSeek Unveils Tile Kernels and DeepEP V2 – Is V4 on the Horizon?

AI Open-Source Efficiency Guide

Apr 23, 2026 · Artificial Intelligence

LLM Wiki: A Karpathy‑Inspired Personal Knowledge Base Now Available as a Desktop App

LLM Wiki is an open‑source, cross‑platform desktop application that transforms documents into an organized, interlinked knowledge base; unlike traditional RAG it incrementally builds a persistent wiki, offers a three‑layer architecture, Obsidian compatibility, and provides step‑by‑step installation and quick‑start guidance.

Knowledge BaseLLM WikiObsidian

0 likes · 6 min read

LLM Wiki: A Karpathy‑Inspired Personal Knowledge Base Now Available as a Desktop App

Shi's AI Notebook

Apr 23, 2026 · Artificial Intelligence

Decoding Anthropic’s Agent Evaluation Methodology: Challenges, Graders, and Best Practices

Anthropic’s engineering blog outlines a systematic approach to evaluating AI agents, highlighting why agents are harder to test than traditional software, defining key concepts like tasks, trials, transcripts, and outcomes, and detailing the three grader types, evaluation timing, and practical decisions for building robust eval pipelines.

AI agentsLLM-as-judgecapability eval

0 likes · 23 min read

Decoding Anthropic’s Agent Evaluation Methodology: Challenges, Graders, and Best Practices

MaGe Linux Operations

Apr 23, 2026 · Databases

How to Diagnose and Optimize MySQL Slow Queries Beyond Adding Indexes

This guide walks through a systematic approach to identify, analyze, and fix MySQL slow queries by enabling the slow‑query log, interpreting its format, using tools like mysqldumpslow and pt‑query‑digest, examining execution plans with EXPLAIN, designing proper indexes, rewriting SQL, tuning server parameters, and establishing continuous monitoring to prevent regressions.

EXPLAINIndex OptimizationMySQL

0 likes · 34 min read

How to Diagnose and Optimize MySQL Slow Queries Beyond Adding Indexes

Tencent Cloud Developer

Apr 23, 2026 · Artificial Intelligence

Hy3 Preview: First Post‑Rebuild Model with Dramatically Boosted Agent Capabilities

Tencent releases and open‑sources Hy3 preview, a 295‑billion‑parameter mixed‑expert LLM supporting 256K context, built on rebuilt pre‑training and RL infrastructure and guided by three principles—systematic capability, authentic evaluation, and cost efficiency—delivering strong gains in complex reasoning, context learning, code and agent tasks, and is already deployed across multiple Tencent products.

Hy3-previewTencent AIagent capabilities

0 likes · 12 min read

Hy3 Preview: First Post‑Rebuild Model with Dramatically Boosted Agent Capabilities

Tech Musings

Apr 23, 2026 · Industry Insights

Goodbye Roo Code and VS Code: How AI Teammates Evolve from Plugins to Full‑Team Members

The Roo Code team announced the shutdown of its VS Code extension and a shift to the cloud‑based AI teammate Roomote, arguing that AI can now close the “error → fix → PR” loop autonomously, making traditional IDE plugins obsolete and moving code collaboration into shared Slack channels.

AI assistantIDERoomote

0 likes · 9 min read

Goodbye Roo Code and VS Code: How AI Teammates Evolve from Plugins to Full‑Team Members

DataFunTalk

Apr 23, 2026 · Artificial Intelligence

Why Palantir’s Valuation Soars: Large Models as the Brain, Ontology as the Skeleton and Memory

In a 90‑minute round‑table hosted by DataFun, experts from banking risk control and cloud observability dissect how Palantir’s ontology—structured as a graph that links entities, metrics and logs—complements large‑model AI, solves data chaos, and becomes the practical backbone for trustworthy enterprise AI.

Data ModelingKnowledge GraphOntology

0 likes · 16 min read

Why Palantir’s Valuation Soars: Large Models as the Brain, Ontology as the Skeleton and Memory

ITPUB

Apr 23, 2026 · Industry Insights

Musk Claims Grok 5 Is AGI as xAI Unveils Two Trillion‑Parameter Models in One Month

Elon Musk announced that Grok 5 is AGI while xAI races through a month‑long rollout of Grok 4.3 (0.5 T), Grok 4.4 (1 T), Grok 4.5 (1.5 T) and a 6‑trillion‑parameter Grok 5, sparking intense debate over whether sheer scale can bridge the AGI gap.

AGIAI competitionGrok

0 likes · 10 min read

Musk Claims Grok 5 Is AGI as xAI Unveils Two Trillion‑Parameter Models in One Month

DataFunSummit

Apr 23, 2026 · Databases

How Hologres Dynamic Table Redefines Data Processing with Incremental Computing

The article analyzes the limitations of traditional batch and stream processing, introduces Hologres Dynamic Table as a declarative, incremental‑compute framework that bridges the gap between low‑cost batch jobs and low‑latency streaming, and validates its performance with benchmarks and real‑world case studies.

Cloud Data WarehouseData ProcessingDynamic Table

0 likes · 13 min read

How Hologres Dynamic Table Redefines Data Processing with Incremental Computing

DataFunSummit

Apr 23, 2026 · Artificial Intelligence

Ontology + Large Model: How Knora Solves Hallucination and Execution Gaps in Enterprise AI

The article details how Knora 4.0 integrates ontology with large‑model AI to create a reusable, extensible enterprise AI platform that mitigates hallucination, stabilises output, and enables autonomous end‑to‑end execution, illustrated with LED production line case studies, architectural breakdowns, and a roadmap for future intelligent agents.

Autonomous AgentsKnowledge GraphOntology

0 likes · 17 min read

Ontology + Large Model: How Knora Solves Hallucination and Execution Gaps in Enterprise AI

ByteDance Data Platform

Apr 23, 2026 · Artificial Intelligence

How LanceDB Powers Enterprise‑Scale Memory in OpenClaw Agents

This article details the technical evaluation and deep integration of LanceDB as a memory plugin for the OpenClaw‑based ArkClaw agent platform, covering plugin selection, core enhancements such as mixed retrieval, hierarchical memory, Autodream processing, Context Engine optimizations, Git‑style version control, and the vision of a unified edge‑cloud memory lake.

AI agentsArkClawLLM memory

0 likes · 12 min read

How LanceDB Powers Enterprise‑Scale Memory in OpenClaw Agents