Article system

AI Intelligence Archive

浏览 AIQB 已保存的 9740 stories AI intelligenceArticles，覆盖Models、Products、Industry、Research、教程与观点方法六大分类。

Indexed articles9740

Categories6

Selected stories2

All articles

Section 41 / 244 页 · 每页 40 stories

TutorialsOrdinary2026-10-30 17:20

vLLM on 4 GPUs / 64GB: A Multi-Model Benchmark — For a Home Multi-Host Setup, I'd Go with vLLM

A while back I benchmarked llama.cpp on a 96GB GPU setup and found its concurrency behavior...

Source: DEV Community · Llama

TutorialsOrdinary2026-10-30 17:08

ReaniGITor: a Día de los Muertos wake for my abandoned GitHub repos, written by Gemma on my own laptop

This is a submission for the Hacktoberfest Open-Source AI Challenge: Week 4 — Ghost in the...

Source: DEV Community

TutorialsOrdinary2026-10-30 17:00

Ebyafaayo: keeping my grandparents' stories alive with Gemma 4, offline

This is a submission for the Hacktoberfest Open-Source AI Challenge: Ghost in the Machine. ...

Source: DEV Community

TutorialsOrdinary2026-10-30 16:40

Haunted Shell: Communing with Dead Code & Crashed Processes via Local Gemma 3

This article is submitted for the Hacktoberfest Open-Source AI Challenge: Week 4 (Theme: Ghost in the...

Source: DEV Community

TutorialsOrdinary2026-10-30 16:38

We let an AI do the QA on a payment webhook. It did a good job. That was the problem.

Nine fresh AI coding sessions, three one-line QA requests, one buggy webhook handler. The AI was careful every time, and every time it ended by asking a question that a fully automated pipeline has nobody to answer.

Source: DEV Community

TutorialsOrdinary2026-10-30 16:20

GhostAltar

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine ...

Source: DEV Community

TutorialsOrdinary2026-10-30 16:18

How to Use Google Jules in 2026: A Practical Guide Before It Winds Down

Google Jules was an interesting step forward for AI-assisted software development. Instead of...

Source: DEV Community

TutorialsOrdinary2026-10-30 16:15

A trojanized Claude Code extension plants an SSH backdoor: the 7-point audit for your machine

anthropic.claude-code@2.1.313 on Open-VSX carried a 66-line loader that SSH-backdoored your machine the moment VS Code loaded it. The mechanism, every IOC, and a read-only audit script I ran on my own machine tonight: 7 checks, 0 indicator…

Source: DEV Community · Claude

TutorialsOrdinary2026-10-30 15:54

I asked eight models which clubs footballers played for

ExtraTime, my football quiz site, has a database of footballers and every club they played for,...

Source: DEV Community

TutorialsOrdinary2026-10-30 15:52

The Ghost in the Machine: A Haunted AI Shell Wrapper for Developers

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine ...

Source: DEV Community

TutorialsOrdinary2026-10-30 15:18

WordCraft Collab 1.0: An Open-Source Word Processor Where Humans and AI Agents Work Together

What if multiple people and AI agents could edit the same Word document simultaneously, directly...

Source: DEV Community

TutorialsOrdinary2026-10-30 15:15

The model routing table: how to spend less without shipping a worse agent

Price per token is not the unit that matters. Cost per successful task is. The routing table is the pattern that lets you capture cheap tokens...

Source: DEV Community

TutorialsOrdinary2026-10-30 15:01

Chunking Whisper into 1-second pieces makes dictation slower

Chunking Whisper into 1-second pieces makes dictation slower I'm building Vocs, a free,...

Source: DEV Community

TutorialsOrdinary2026-10-30 15:00

Why a Better System Prompt Didn't Stop Prompt Injection: Field Notes on the Lethal Trifecta, Scoped Tools, and Untrusted Output

Headline: Prompt injection cannot be fixed with a better system prompt, because a large language...

Source: DEV Community

TutorialsOrdinary2026-10-30 14:57

Six coding agents, one menu bar: where the "waiting on you" signal actually lives

I run six coding agents on the same laptop: Qoder, Claude Code, Codex, ZCode, WorkBuddy, OpenCode....

Source: DEV Community · Claude

TutorialsOrdinary2026-10-30 14:45

The Compose Race Your Coding Agent Inherits From You

Your agent-generated compose file races Postgres because it inherited the race from old tutorials. Live repro: the copied healthcheck lied.

Source: DEV Community

TutorialsOrdinary2026-10-30 14:37

Serving MOSS-Image: step count, memory mode and the traps that change the words you ship

MOSS-Image on one A40: 80 s per picture with every word right, 9 s with a turbo LoRA that slips, and the same pixels under a 12 GB memory cap.

Source: DEV Community

TutorialsOrdinary2026-10-30 14:21

👻 The Séance Engine: Finding the Ghost in the Data with TabPFN & Gemma

TabPFN finds the one log row that shouldn't exist; local Gemma tells its Día de los Muertos story. Traced with Sentry.

Source: DEV Community

TutorialsOrdinary2026-10-30 13:49

Undead Agent - My AI Agent Dies, Comes Back, and Tells Me What It Saw(Gemma + Ollama + Temporal)

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine I...

Source: DEV Community · Ollama

TutorialsOrdinary2026-10-30 13:44

WARDEN: Hunting the Ghost in the Machine with an AI-Powered Self-Healing DevOps Agent

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine ...

Source: DEV Community

TutorialsOrdinary2026-10-30 13:35

Build a production-style Autonomous SRE Command Center using Gemini 4 Argon

When a production Kubernetes cluster starts throwing 5xx errors , the hardest part of incident...

Source: DEV Community · Gemini

ResearchOrdinary2026-10-30 13:20

From 1D Benchmarks to Complex 3D Geometries: Can Agents Reliably Specify Problems for Automated PDE Simulation?

Turning an engineering request into a simulation requires consistent geometry, physics, boundary conditions, and numerical settings. Recent agent-assisted systems automate simulation setup, but successful execution alone does not establish…

Source: arXiv

TutorialsOrdinary2026-10-30 13:15

Cutting LLM inference cost and latency without making the feature worse

Cut LLM API cost and latency in TypeScript: prompt caching with a stable prefix, a model per route, streaming for time to first token, batches and cost logs.

Source: DEV Community

ResearchOrdinary2026-10-30 13:13

Rome Wasn't Built in One Turn: Replay-Grounded Reward Signals for Long-Horizon Freeciv Agents

Training language-model agents to play long-horizon strategy games is hard because deterministic feedback is sparse and often arrives only at the end of a game. In-game scores can offer dense signals, but they do not reliably predict futur…

Source: arXiv

TutorialsOrdinary2026-10-30 13:10

Vibe Coding Security: 8 Holes in Almost Every AI-Built App

Vibe coding security checklist: the 8 holes found in most Lovable, Bolt and Replit apps, how to test for each, and how to fix them.

Source: DEV Community

TutorialsOrdinary2026-10-30 13:09

We put a free, unlimited LLM inside a static web page — Transformers.js + a 135M model, zero servers

Every LLM demo I have ever seen has the same fine print: bring your own API key, or the demo is...

Source: DEV Community

TutorialsOrdinary2026-10-30 12:50

Break it before you merge it: introducing litmus-lite

Several years ago, we started LitmusChaos with a simple belief: chaos engineering is not just for...

Source: DEV Community

ResearchOrdinary2026-10-30 12:39

MEMIG: Self-evolving Memory Construction for Antibody Optimization

Memory has become central to self-evolving LLM agents, which store and retrieve experience from past rollouts to guide future decisions. Existing systems represent this experience as free-form natural language and maintain it by rewriting,…

Source: arXiv

TutorialsOrdinary2026-10-30 12:38

I took SSH away from my agent. Then my own tool started lying to it.

This spring I wrote twice about not giving Claude SSH access to my home server. The argument was...

Source: DEV Community · Claude

ResearchOrdinary2026-10-30 12:37

A Unified Framework for Probability Estimation of Data-Dependent Events

We study the general problem of estimating $D(\dde(S))$, the probability that an unknown distribution $D$ assigns to an event determined by a sample $S$ of $n$ i.i.d.\ draws from $D$ through a known mapping $\dde$. The estimation algorithm…

Source: arXiv

ResearchOrdinary2026-10-30 12:35

Generative SDEs with Trainable Diffusion and Density-Based Losses

We develop a divergence-kernel method for jointly optimizing neural SDE drift and diffusion fields and the initial Gaussian distribution, using losses that depend directly on the terminal density. In two- and eight-dimensional Gaussian-mix…

Source: arXiv

TutorialsOrdinary2026-10-30 12:30

I counted every skill my AI invoked across 55 sessions. Here's the leaderboard.

I had over a hundred skills installed while building ParkEase. Which ones actually pulled their...

Source: DEV Community

TutorialsOrdinary2026-10-30 12:25

Claude Fable 5.1 Agent Test: Sonnet Tied It at 1/8 the Cost

Originally published on RanThis. Cover: AI-generated illustration. Claude Fable 5.1 is not worth its...

Source: DEV Community · Claude

ResearchOrdinary2026-10-30 12:21

Experts Exhibit Structured Redundancy: Structure-Preserving Expert Pruning for Mixture-of-Experts Large Language Models

Sparse MoE mechanism significantly improves the trade-off between model capacity and inference efficiency. However, practical deployment still requires storing all experts, resulting in substantial memory overhead. We show that expert outp…

Source: arXiv · DeepSeek / Qwen

ResearchOrdinary2026-10-30 12:13

TRIDENT and SHAKTI: A Framework and Benchmark for Hindi AI-Text Detection

As Large Language Models (LLMs) become increasingly multilingual, distinguishing human-written from AI-generated text is essential for maintaining information integrity. However, research remains English-centric, with limited benchmarks fo…

Source: arXiv

ResearchOrdinary2026-10-30 12:09

Coarse3D: Controllable and Faithful Sparse Structure Generation for Image-to-3D

Coarse-to-fine 3D-native generation relies heavily on the first-stage sparse structure, which determines the global orientation and overall geometry for subsequent refinement. However, existing coarse generators suffer from two key limitat…

Source: arXiv

ResearchOrdinary2026-10-30 12:03

RTL-Interact: Training EDA-Interactive Models for Multi-Task RTL Design

Large language models offer a promising route to register-transfer-level (RTL) design, but reliable implementations require checking and revision based on electronic design automation (EDA) tools' feedback. Existing RTL-specialized models …

Source: arXiv · DeepSeek

ResearchOrdinary2026-10-30 12:00

Corroborative Learning: Mutual Enhancement through Task-Response Reconstruction

In pattern analysis, models trained independently for the same task often develop complementary capabilities through differences in architecture, data, and optimization. Exploiting complementarity while retaining independent deployment is …

Source: arXiv

ResearchOrdinary2026-10-30 11:54

A Positional Realizability Theory of Length Generalization in Single-Head Attention

Why does length generalization (LG) depend so sharply on the positional encoding (PE)? Relative schemes extrapolate fixed offsets, absolute schemes fix positions, and neither handles growing offsets. We give a realizability theory for a de…

Source: arXiv

ResearchOrdinary2026-10-30 11:52

Beyond Response Scoring: An AI-aware Cognitive Supervision Framework for Automated Mathematical Short Answer Grading

Transformer-based language models have achieved remarkable success in automated short answer grading (ASAG). However, current methods in ASAG struggle to comprehend, reason about and evaluate complex mathematical answers involving symbols …

Source: arXiv