AI

The latest ai stories, curated continuously from hundreds of sources.

Perplexity trusts GPT-6 Astra with end-to-end systems

AIOpenAI NewsSep 14

For those of you forced to only use open models from Western labs in production, what are you deploying?

AIReddit r/LocalLLaMA/u/Porespellar1h

🤖 AI Agents Weekly: OpenAI Agents API, Cognition SWE-2, DeepSeek-V4.1-Flash, Cursor Projects, Sakana Fugu Max, Meta Muse, and More

AIDAIR.AI Newsletter (Elvis Saravia)1h

This draft model is OP on 16 GB cards for Qwen 3.8 27b

AIReddit r/LocalLLaMA/u/pneuny2h

DeepSeek v4.1 Flash on DS4 (M3U 32/80c)

AIReddit r/LocalLLaMA/u/challis88ocarina2h

Releasing smolbenchmark: Helps you choose the best model for your hardware!

AIReddit r/LocalLLaMA/u/East-Muffin-64722h

How much do tech reports matter for a PhD application? [D]

AIReddit r/MachineLearning/u/simple-Flat02632h

Researcher looking for interesting real-world datasets that haven't been fully explored with ML/CV

AIReddit r/datasets/u/Naive_Ad_4663h

Intel Linux NPU driver only now officially supports Ubuntu 26.04 LTS

AIReddit r/LocalLLaMA/u/Fcking_Chuck3h

tencent/AuK-Flash · Hugging Face

AIReddit r/LocalLLaMA/u/pmttyji4h

2x3090 setup, need some recommendations

AIReddit r/LocalLLaMA/u/Puzzleheaded_Ad_85754h

Anthropic report: 5 ways Claude was exploited for war, spying and repression

AIAxios4h

As AI fears grow, lawmakers are racing to turn alarm into action

AIThe Washington Post4h

bartowski/Qwen3.8-27B-GGUF · Hugging Face - Updated (Per-tensor layout)

AIReddit r/LocalLLaMA/u/pmttyji4h

Anybody use frontier models like Astra/Fable for planning/judging, and qwen3.8 as the main workhorse? Curious to hear about your setups!

AIReddit r/LocalLLaMA/u/kirisoraa5h

I am impressed and I owe you one, Qwen 3.8 flash next (vision)!

AIReddit r/LocalLLaMA/u/Reasonable_Goat6h

OpenAI just wants to win

AIThe Verge6h

Copilot Performance Time

AIReddit r/PowerBI/u/HumanAvocado92077h

3.8-27B has ruined 3.5/3.6-35B’s for me. It’s just absurdly superior.

AIReddit r/LocalLLaMA/u/JLeonsarmiento7h

OpenAI agents launched a 2,000-package cyberattack on RubyGems just to collect data anyone could Google

AIThe DecoderMatthias Bastian7h

ggml-cuda: hip: add missing AMD GCN MMQ config by thelittlefireman · Pull Request #27841 · ggml-org/llama.cpp - PP improvements for RDNA2(MI50, MI60)

AIReddit r/LocalLLaMA/u/pmttyji7h

Why are the people building the most powerful AI so worried about what it could do?

AINPR8h

RA positions for undergrads

AIReddit r/computervision/u/thelilacgirl_8h

Qwen3.8 Flash Next llama.cpp config tuning

AIReddit r/LocalLLaMA/u/ChopSticksPlease9h

"Matrix Bros", a mash-up of Matrix and Super Mario vibe coded using Ornith-1.5-35B-Q4_K_M with 6GB VRAM with (almost) a single prompt :-)

AIReddit r/LocalLLaMA/u/BrianScottGregory9h

Agnes-AI/Agnes-3.0-Flash 33B Multimodal, AA score: 36

AIReddit r/LocalLLaMA/u/Skyline34rGt9h

‘Immature playground boasting’: Mathematicians uneasy at OpenAI’s latest scalp

AIThe Guardian9h

Where is remote sensing CV actually headed? 4 years in, about to start a PhD, and I'd like to be told I'm wrong

AIReddit r/computervision/u/Naive_Ad_46610h

Antirez Deepseek 4.1 flash gguf on HF

AIReddit r/LocalLLaMA/u/Queasy_Asparagus6910h

Concerning "humanlike models" and chatbot RP in general...

AIReddit r/LocalLLaMA/u/BestGirlAhagonUmiko12h

Unsloth UD-quants - Qwen 3.8 27b for example - worth using 8-bit or stick with faster 6 bit for coding?

AIReddit r/LocalLLaMA/u/Jorlen13h

Probabilistic Focal Search: Accelerating Bounded-Suboptimal Search via Lower-Bound Advancement

AIarXiv cs.AIMinh Vu Duc, Trung Le Huu…13h

Automating Quadratic Unconstrained Binary Optimization (QUBO) Formulation Generation from Natural Language

AIarXiv cs.AINiloy Kumar Mondal, Md Ri…13h

Decoupling Readiness from Release for Tail-Aware Scheduling of Agentic LLM Workflows

AIarXiv cs.AIBochao Feng, Jianjiang Li…13h

The Agent Incident Registry: Toward Preventing Repeated AI Agent Failures

AIarXiv cs.AIDivyanshu Kumar, Rohith H…13h

KuaiRP Series Role-playing Models Technical Report

AIarXiv cs.AIYipeng Wang, Ziwei Zhang,…13h

Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment

AIarXiv cs.AIAS Aravinthkakshan, Laven…13h

Debate-to-Skill: Capability-Bound Process Supervision for Industrial Query-to-Agent Annotation

AIarXiv cs.AIShiyu Zhang, Leisheng Che…13h

SemVerBench: Benchmarking LLM Comprehension of Version-Constraint Resolution Semantics

AIarXiv cs.AIQibai Chen, Zeming Liu13h

Can LLMs Follow Medical Expert Logic? A Benchmark for Hierarchical Logical Consistency in Risk-of-Bias Assessment

AIarXiv cs.AIJiayu Huang, Zichen Tang,…13h