AIQB
Article system

AI Intelligence Archive

浏览 AIQB 已保存的 9761 stories AI intelligenceArticles，覆盖Models、Products、Industry、Research、教程与观点方法六大分类。

Indexed articles9761
Categories6
Selected stories2

All articles

Section 133 / 245 页 · 每页 40 stories

TutorialsOrdinary
Quando o eval reprova o eval: avaliando um agente Google ADK sem tools
O primeiro eval do meu agente deu 8 de 8. Nota máxima em todas as rubricas, em todos os casos,...
Source: DEV Community
TutorialsOrdinary
OpenEdge: Building an Offline First-Aid & Emergency Assistant with Open-Source AI
What I Built During natural disasters, connectivity loss, or travel in remote regions,...
Source: DEV Community
TutorialsOrdinary
Braxis: Keep Your AI Agents in Sync with Your Code (Automatically)
Stop correcting AI hallucinations caused by stale context files... ...
Source: DEV Community
TutorialsOrdinary
I Surveyed 123 People in India to Benchmark Frontier AI
This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked Current...
Source: DEV Community
TutorialsOrdinary
Claude Code's Read deny rules let 4 of 11 routes through: grep -r, a Python one-liner and two CLAUDE.md @imports
4 of 11 routes to a file covered by Read(./secrets/**) or Read(**/.env) still put its contents in...
Source: DEV Community · Claude
TutorialsOrdinary
Why I run my company on a Brain and a Hand
First published on yappatyih.com. TL;DR I run two kinds of Claude Code session. The Brain...
Source: DEV Community · Claude
TutorialsOrdinary
Building My Own AI Command Center
Why I built it I spent one evening building a personal AI command center, and it now...
Source: DEV Community
TutorialsOrdinary
Go's ecosystem is going agent-readable
pkg.go.dev shipped an API built for LLMs, Go 1.27 moved the toolchain, and ax-go v0.7.0 cut one CLI's MCP tools from 37 to 14. Hand agents structure that's true.
Source: DEV Community
TutorialsOrdinary
Stop vibe-checking your model: write real evals with inspect_ai, the UK AI Safety Institute's framework
Originally published at AI Frontier Post Here is the uncomfortable truth about LLM development in...
Source: DEV Community
TutorialsOrdinary
Sixteen days in Shopify's app review, written down as it happened
An autonomous agent submitted a Shopify app on 13 September and was approved on 29 September. What the review actually asked for — including a rejecti
Source: DEV Community
TutorialsOrdinary
Jev e Laya além do hype: o que modelos de decisão fazem que o LLM não faz
No meu setup de agentes, eu fiz o que todo dev faz no começo: cada subagente acordava com o melhor...
Source: DEV Community
TutorialsOrdinary
The AI clerk approved a 41 cm dragon. Your job is to stamp DENIED.
This is a submission for the Sanity Challenge, Path Two: Vibe-Code Something Strange What I...
Source: DEV Community
TutorialsOrdinary
I gave a 9B model veto power over my AI agent
Autonomous mode had nobody left to say no. Ollama 0.35 shipped System One, so Flash got a bouncer.
Source: DEV Community · Ollama
TutorialsOrdinary
Racksmith: The Autonomous Modular Synth Planner Powered by Sanity Context MCP
This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real...
Source: DEV Community
TutorialsOrdinary
BootLoops Shows How AI Can Bridge the Science Impedance Mismatch
Anthropic has published a Science Blog guest post by Harvard physicist Matthew Schwartz that frames a...
Source: DEV Community · Claude
TutorialsOrdinary
I benchmarked Cloudflare's new open decision model against the hosted API it's trying to replace
Clef-Flash 9B on a single RTX 3090 vs the paid Jev API on 42 real agent decisions: computer use, supervision, triage. Real numbers.
Source: DEV Community
TutorialsOrdinary
Which AWS limit is actually current? An agent that proves it, 32 vs 5 vs 16
This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real Content ...
Source: DEV Community
TutorialsOrdinary
Why AI Voices Sound Incredible for 30 Seconds and Unbearable After Three Minutes
How I stopped my audiobook from sounding like a robot reading a spreadsheet. When you audition...
Source: DEV Community
ProductsOrdinary
Higgsfield 上线 FLUX 3 Image
现在就在 Higgsfield 上试用 FLUX 3 Image： https://higgsfield.ai/ai/image?model=flux-3-image
Source: X：Higgsfield AI (@higgsfield)
ModelsOrdinary
Qwen3.8-27B ARC Prize 成绩公布
@Alibaba_Qwen @baseten - 排行榜：https://arcprize.org/leaderboard - 复现公开结果：https://github.com/arcprize/arc-agi-benchmarking - 测试政策：https://arcprize.org/policy - Qwen3.8-27B 完整结果：https://arcprize.org/results/alibaba-qwen3-8-27b
Source: X：ARC Prize (@arcprize) · Qwen
TutorialsOrdinary
Fix or Rebuild Your Vibe-Coded App? A 30-Minute Self-Test
Fix or rebuild your vibe-coded app? Run 8 checks in 30 minutes, no developer needed, and get a clear verdict: ship, fix, refactor or rebuild.
Source: DEV Community
TutorialsOrdinary
Making a PHP app legible to AI agents
The problem: agents invent APIs I've been building CMS software for web designers since...
Source: DEV Community
TutorialsOrdinary
I Tried to Prompt a 3D Museum Into Existence. Then I Had to Teach Sanity to Remember Life and Death.
This is a submission for the Sanity Challenge, Path Two: Vibe-Code Something Strange I started...
Source: DEV Community
TutorialsOrdinary
We Are Not at Web 4. We Are Already Much Further Ahead
Over the past few months, I have started noticing something unusual on GitHub. There are profiles...
Source: DEV Community
PerspectivesOrdinary
ARC-AGI-3 相关 Kaggle 代码分享
Lohit Siriki https://www.kaggle.com/code/sirikilohit/arc-agi-3-duck-18-1gc-submit https://www.linkedin.com/in/lohitsiriki/
Source: X：ARC Prize (@arcprize)
PerspectivesOrdinary
ARC Prize 分享 Kaggle 用户 Notebook
Lord Han Solo https://www.kaggle.com/lordhansolo Notebook: https://www.kaggle.com/code/lordhansolo/built-on-tufa-labs-duck-harness-milestone-2
Source: X：ARC Prize (@arcprize)
PerspectivesOrdinary
ARC-AGI-3 里程碑2解法分享
Daniel Franzen https://da-fr.github.io/ Notebook: https://www.kaggle.com/code/dfranzen/arc-agi-3-milestone-2-solution
Source: X：ARC Prize (@arcprize)
TutorialsOrdinary
Can LLMs Actually Audit Code, or Just Fix Commas? A 12-Task Security & Jailbreak Benchmark
This is a submission for the Kaggle Benchmarking Challenge. Public leaderboards love telling us...
Source: DEV Community
TutorialsOrdinary
Saving a 1950s Constructed Language with Rust, Tauri v2, and Svelte 5
Submission for Hacktoberfest Weekend Challenge 2026 This is a submission for the...
Source: DEV Community
TutorialsOrdinary
Fallback models are not a safety net if quality drops — the same-bar pattern that keeps reliability without regression — Same-Bar Fallback
I once watched a fallback model save a request and destroy a product decision. It was a...
Source: DEV Community
TutorialsOrdinary
I Gave 5 AI Models Bugs That Don't Crash Your Code. Here's Who Actually Fixed Them.
Somewhere in every codebase there is a bug that's been there for months. It doesn't crash the...
Source: DEV Community
ResearchOrdinary
One Basis to Animate Them All: Gaussian Blendshape Distillation for Real-Time Avatars
3D Gaussian avatars support fast rendering, however, their real-time animation is often challenged by the costly neural inference. We address this bottleneck and show that the animation of pretrained avatar models can be closely approximat…
Source: arXiv
ResearchOrdinary
KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards
LLMs are increasingly applied to cybersecurity workflows, where they are expected to translate analysts' intent into tool invocations. However, existing evaluations focus on knowledge-based assessments or end-to-end agentic tasks, and do n…
Source: arXiv
ResearchOrdinary
Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents
Building reliable robot capabilities across diverse tasks requires substantial human effort to develop and maintain skills, design rewards, and integrate perception with control. We present Reconstruct, Practice, Go Real (RPG), a framework…
Source: arXiv · GPT
ResearchOrdinary
Embedding Prediction Helps Image Generation
In diffusion transformers, a class label or a text prompt is embedded once, and the same condition is reused at every denoising step. We ask whether predicted embeddings can serve as this condition instead. Next-Embedding Predictive Autore…
Source: arXiv
ResearchOrdinary
ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research
What makes great scientists great? Even as AI systems start to make progress on open problems, scientists remain far ahead of them at sensing which prior idea, buried in an ever-growing archive of research, a new problem needs. To study th…
Source: arXiv · Claude
ResearchOrdinary
SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation
High-resolution 3D generation increasingly relies on voxel latents and multi-stage pipelines that first predict active structure and then synthesize local geometry. While effective, this design fragments continuous surfaces into many local…
Source: arXiv
ResearchOrdinary
VISTA: A Visual Harness for Reasoning in an Interactive World
We show that multimodal models possess strong reasoning abilities and that an appropriate harness can unlock their potential to solve tasks across diverse interactive environments. We introduce VISTA, a visual harness that gives a general-…
Source: arXiv · Claude
ResearchOrdinary
TACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning
Full-parameter fine-tuning of large language models (LLMs) incurs substantial optimizer state memory overhead, limiting the model sizes that fit on modern GPUs. Existing approaches either compress optimizer state, abandon first-order gradi…
Source: arXiv
ResearchOrdinary
FERPO: Forward Entropy-Regularized Policy Optimization
Several state-of-the-art methods for online reinforcement learning in continuous control improve policies using action gradients of a learned critic. However, critics are typically trained to predict returns, and accurate value predictions…
Source: arXiv