Researchers find people increasingly juggle multiple multimodal LLMs at once, building shifting hierarchies between primary and secondary systems based on task, effort, and perceived credibility. arXiv preprint arXiv:2603.26107

#AIResearch #HCI #LLMs #MultiAgent
https://arxiv.org/abs/2603.26107

Would be hilarious if it was #Microsoft, above all, who would be responsible to finally pop the #AIBubble because they decide to withdraw all their funding from #OpenAI to prevent #LLMs from vibe-reconstructing all their proprietary software […]

A new benchmark called KULTURE Bench targets Korean cultural understanding in language models, using cultural news, idioms, and poetry to test comprehension at word, sentence, and paragraph levels.

#AI #LLMs #KULTUREBench
https://arxiv.org/abs/2412.07251

Linux kernel contributors are debating whether to keep the label for AI-assisted patches.

Torvalds opposes restricting tools; the challenge is prese…

https://en.hacks.gr/o-linus-torvalds-yperaspizetai-ti-chrisi-technitis-noimosynis-apo-programmatistes-toy-linux/

#LinuxKernel #LLMs #AITransparency

Researchers propose a framework for evaluating LLMs as scientific inquiry agents, extending the Countdown task to test hypothesis generation, reasoning, and exploration in an interactive setting. It aims to assess how models integrate…

#AI #LLMs #MachineLearning
https://arxiv.org/abs/2610.21604

www.theguardian.com/technology/2...

The true effect of ‘AI’. LLMs affect.

#llms

The true effect of ‘AI’. LLMs affect.

#llms
https://www.theguardian.com/technology/2026/oct/13/ai-companies-rushed-workers-fallout?CMP=Share_iOSApp_Other

AI agents get more useful when they can work with the tools you already use.

#hermesagent #aiagents #llms

LLMs are machines that turn software engineers into gambling addicts.

#ai #LLMs #gambling #SoftwareEngineering

A new framework called Multi-Agent Transactive Memory proposes population-level storage and retrieval of LLM agent trajectories, so consumer agents can reuse procedural knowledge from producer agents instead of rediscovering…

#AIagents #RAG #LLMs #MultiAgentSystems
https://arxiv.org/abs/2606.19911

New arXiv paper proposes disagreement-guided weak-to-strong routing for small models on edge devices, letting stronger models step in without needing ground-truth supervision for open-ended and agentic tasks.

#EdgeAI #ModelRouting #AIResearch #LLMs
https://arxiv.org/abs/2610.22245

New arXiv paper introduces an AI tutor for non-native English students in computing courses, using a specialized retrieval-augmented generation pipeline that handles foreign-language questions and produces bilingual…

#AIinEducation #EdTech #LLMs #ComputingEducation
https://arxiv.org/abs/2610.22163

#PwnEscape (CVE-2026-99999): a hardcore Guest-to-Host escape based only off IPv4 packets, that impacts most #cloud providers. The timeframe for discovery does show how fundamentally usage of #LLMs changes #cybersecurity research (and how effective #BugBounty programs are).
pwn.ai/blog/pwnescape

Well, well, well. @bsky.app is using #LLMs to steal work. Tragically predictable.

A new paper, Prism-1, examines how System One decision models that read choices from language models often violate rational choice axioms like order independence and Luce's choice axiom, even with techniques like shuffled-order training or…

#AI #Prompting #LLMs
https://arxiv.org/abs/2610.22066

Tool-augmented LLMs can improve medical calculation reliability by structuring calculators into reusable executable tools rather than relying on direct prompting for the entire workflow. This approach separates variable…

#Prompting #LLMs #HealthcareAI #ToolUse
https://arxiv.org/abs/2610.21477

AlphaDiana is a new arXiv paper introducing a modular measurement system for evaluating AI agents on verifiable reasoning tasks, separating harness control loops from file reuse and memory retention. It benchmarks three…

#OpenSource #AIagents #Benchmarking #LLMs
https://arxiv.org/abs/2610.20675

arXiv paper CoCA reframes rubric-based RL for LLMs as criterion-aware credit assignment, using rubric info as a relevance cue rather than self-distillation targets to avoid reward degradation. A useful step…

#OpenSourceAI #ReinforcementLearning #LLMs #AIResearch
https://arxiv.org/abs/2610.17733

A new framework called Agentic Architect uses LLMs for agentic CPU microarchitectural design-space exploration, combining code evolution with cycle-accurate simulation to optimize performance, area, and power…

#AI #ComputerArchitecture #LLMs #HardwareDesign
https://arxiv.org/abs/2604.25083