AI News Today

arXiv:2609.03340v1 Announce Type: new Abstract: Distributed LLM-agent teams can read the latest shared facts and still act on an obsolete plan. A planner may derive an action from requirement $r_3$, another agent may commit $r_4$, and an executor may receive $r_4$ without replacing the plan derived from $r_3$. We call this \emph{stale-plan execution}: state freshness does not establish that the plan authorizing an action remains valid. We introduce PlanFence, a dependency-scoped action-validatio

arXiv:2609.03402v1 Announce Type: new Abstract: Artificial intelligence (AI) teaching assistants powered by large language models (LLMs) offer scalable educational support but often provide limited personalization. This study presents a prompt-engineering-based framework for personalizing general-purpose LLM/RAG-based AI teaching assistants such as Jill Watson across academic disciplines and courses. The framework adapts responses using six learner-specific dimensions: self-assessment, abstracti

arXiv:2609.03407v1 Announce Type: new Abstract: People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory context. Most prior work has studied this context through single-turn judgments or pressure-laden rebuttals, assumptions that poorly match how guidance is sought in real-world contexts. These assumptions leave unclear whether narration alone, without an explicit opposing position, can shift model j

arXiv:2609.03416v1 Announce Type: new Abstract: LLM-empowered paper-code discrepancy detection has received growing concern since the scaling of research submissions exceeds the manual review capability. However, the limited context capacity and one-sided discrepancy detection of existing single-agent LLM paradigms lead to an inferior recall performance in detecting discrepancies. In this paper, we propose Dude, the first Dual-Detection Multi-Agent System for paper-code discrepancy detection. We

arXiv:2609.03423v1 Announce Type: new Abstract: Full-duplex voice agents must continuously decide when to listen, backchannel, interrupt, handle speech overlaps, take the floor, and yield. Existing benchmarks largely test these behaviors through explicit turn-management instructions, while deployed agents are often configured through roles or personas from which the appropriate conversational behavior must be inferred. We introduce DuplexSpeechBench-IFEval (DSB-IFEval) for evaluating implicit in

arXiv:2609.03438v1 Announce Type: new Abstract: Graphical user interface (GUI) agents are increasingly used to execute natural-language instructions on user interfaces, yet real users may issue infeasible instructions due to benign mistakes. A reliable agent should not only know how to act, but also when not to act. In this work, we introduce CONFLICTGUI, a benchmark covering instruction-internal conflicts and instruction-GUI context conflicts to study conflict-aware termination. Our evaluation

arXiv:2609.03460v1 Announce Type: new Abstract: As generative AI makes polished prose cheap to produce, users can no longer rely on fluency as a proxy for truth. We call this failure mode the Fluency Trap: users trust fluent hallucinations while also discounting accurate content once it is disclosed as AI-generated. Binary ``Made with AI'' labels respond with authorship disclosure, but they do not show what supports a claim. We propose Provenance Density, an evidence-visualization interface that

arXiv:2609.03493v1 Announce Type: new Abstract: Modern vision-language models (VLMs) can directly answer many image-grounded questions, yet they often struggle with complex queries requiring fine-grained visual details or external knowledge. To acquire this missing evidence, agentic VLMs invoke tools such as image cropping, image search, and text search. However, existing training paradigms primarily evaluate tool-use based on final answer correctness, leaving evidence acquisition and utilizatio

arXiv:2609.03494v1 Announce Type: new Abstract: Long-output reasoning has made the key--value (KV) cache a critical memory bottleneck for efficient LLM serving. Existing KV compression methods usually rely on a predefined per-request budget and adjust only which KV states are retained, leaving the total capacity fixed throughout decoding. However, reasoning workloads exhibit substantial demand variation: different requests require different KV capacities, and the attention demand of an individua

arXiv:2609.03503v1 Announce Type: new Abstract: With the rapid development of the Internet of Things, computation intensive directed acyclic graph (DAG) tasks have become increasingly common in cloud-edge-end collaborative environments. However, cloud, edge, and end nodes are highly heterogeneous in computing capacity, network bandwidth, and energy consumption, which makes the efficient scheduling of tasks with complex dependencies an NP-hard problem. Traditional heuristic algorithms and convent

arXiv:2609.03527v1 Announce Type: new Abstract: Neonatal respiratory diseases are a major cause of neonatal morbidity and mortality, posing substantial challenges in clinical practice. Despite recent advances, existing Multimodal Large Language Models (MLLMs) face two key limitations in neonatal diagnosis: (1) domain gap arising from predominantly adult training data; (2) insufficient integration of multidimensional clinical context for accurate diagnosis. To address these challenges, we collect

arXiv:2609.03546v1 Announce Type: new Abstract: We present Dalek, a closed machine designed for agents that realizes self-maintenance, self-evolution, self-reproduction, and self-organization on any substrate satisfying a general host contract. The machine is built from three primitives---actors, messages, and channels. Four obligations---a host boundary, a construction language, admissible transitions, and rule heredity---give its boundary, identity, and closure a structural basis. Von Neumann'

arXiv:2609.03553v1 Announce Type: new Abstract: Policy analysis requires more than predicting whether a proposal will pass: it requires identifying who will be affected, how those actors respond, and what follows. LLM-based policy simulations model these processes at scale, but their validity is hard to establish when plausible behaviour is never compared with observed outcomes. We introduce GPS-Bench, an evidence-grounded benchmark for governance policy simulation that links policies to relevan

arXiv:2609.03580v1 Announce Type: new Abstract: The growing scale of academic peer review has motivated the use of Large Language Models (LLMs) as review assistants, yet LLMs can generate fluent but unsupported claims that undermine review reliability. Existing hallucination benchmarks are not designed for peer review, where verification requires grounding claims in long, technical papers. We introduce HalluPeer, a benchmark for detecting hallucinations in scientific peer reviews, providing alig

arXiv:2609.03586v1 Announce Type: new Abstract: Audio-video diffusion models rely on cross-modal attention to coordinate text, sound, and visual content, yet this same mechanism can introduce subtle and systematic semantic leakage. We study these models by probing and analyzing the ``attention triangle,'' comprising the three cross-attention edges connecting the text, audio, and video streams, and examine how semantic information is routed across modalities during generation. Our analysis reveal

arXiv:2609.03588v1 Announce Type: new Abstract: As LLMs increasingly act through tools, they must reconcile user instructions, parametric knowledge, and dynamic environmental observations before taking actions. We introduce KC-Bench, a controlled multi-turn benchmark for measuring this capability across world-knowledge conflicts, input inconsistencies, and multi-source temporal conflicts. Its 238 tasks are manually screened from more than 1,000 generated candidates and combine a user simulator,

arXiv:2609.03635v1 Announce Type: new Abstract: Clinical trials in the UK can cost up to {\pounds}1.3 million, with approximately 90% drug failure rate. Toxicity is a major contributing factor in drug failure. Testing is time and cost intensive. In recent years, the use of artificial intelligence has been increasingly explored to aid in the prediction of drug toxicity, with extensive use of large language models (LLMs). However, LLMs can show considerable variation when minor changes are made to

arXiv:2609.03702v1 Announce Type: new Abstract: General-purpose code embeddings power tools for code search, classification, and retrieval. Compact transformer encoders for code typically rely on either human-written docstrings (labor-intensive and inconsistent) or mined structural signals such as execution traces (setting-specific and costly to collect). We empirically study an alternative: contrastive pretraining of small encoders with synthetically generated natural-language descriptions emph

<p>I got access to GPT-6 Astra this afternoon, so naturally I used it to generate <a href="https://simonwillison.net/tags/pelican-riding-a-bicycle/">SVGs of pelicans riding bicycles</a> - at low, medium, high, xhigh and max reasoning levels (Astra doesn't support reasoning=none). Then I rendered those pelicans in <a href="https://static.simonwillison.net/static/2026/gpt-6-and-5.6-pelicans.html">a comparison grid</a> with GPT-5.6 Sol, Terra, and Luna, and beyond being fun the result was surprisin

Article URL: https://engineering.atspotify.com/2026/9/portal-by-spotify-cut-my-claude-code-token-usage-by-90 Comments URL: https://news.ycombinator.com/item?id=49571465 Points: 218 # Comments: 120

The round is being raised just months after the robot data startup exited from stealth.

OpenAI’s latest agent swarm incident adds urgency to calls for independent investigations as researchers and lawmakers question whether AI labs should control the scope of their own safety reviews.

The gift that keeps on giving. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Learn how to deploy a multimodal WhatsApp ordering assistant that takes customer orders through text, voice notes, and real-time voice calls on a single business number, built on Amazon Bedrock AgentCore with Amazon Nova 2. The channel and ordering layers stay separate, and one shared memory recognizes each customer across all three channels.

Article URL: https://openrouter.ai/openai/gpt-6-astra Comments URL: https://news.ycombinator.com/item?id=49570545 Points: 279 # Comments: 201

Nscale, which recently struck a $45 billion deal with Anthropic, is in talks to raise additional funds in anticipation of an upcoming IPO.

OpenAI on September 3, 2026 introduced Daybreak for Frontline Defenders, a global initiative committing $1 billion in subsidized access to its Daybreak cybersecurity models, training, technical support, and partnerships, and announced a new pilot with the Multi-State Information Sharing and Analysis Center (MS-ISAC) aimed at state, local, tribal, and territorial cyber defenders. The OpenAI announcement said the $1 billion commitment will expand subsidized access to Daybreak cyber models and…

Now comes the harder sell: Convincing customers and merchants to trust AI with purchases

AI workstation promises to put up to 576 GB of HBM3e and 16 TB/s of memory bandwidth on your desk

It's all in the choosing. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Article URL: https://eebench.org/blog/can-ai-design-circuit-boards-yet/ Comments URL: https://news.ycombinator.com/item?id=49569366 Points: 314 # Comments: 189

<p>I like vibe coding. I'm not here to shame anyone out of it. Some of the best things I've shipped started as a half-formed idea typed into a chat box at 11pm.</p> <p>But if you've done this for more than a weekend, you know the shape of the problem. The first prompt is magic. The fifth one is negotiation. By the fifteenth you're re-explaining a decision you already made twice, the agent has quietly renamed your data model, and you're one "fix the header" away from deleting the folder and start

Article URL: https://github.com/gotchat-ai/foundry Comments URL: https://news.ycombinator.com/item?id=49569265 Points: 1 # Comments: 1

<h2> How to connect LLM agents to verified African examination content, deliver progressive Socratic hints, and diagnose student misconceptions without hallucinating. </h2> <p>Large Language Models (LLMs) like Claude 4.6 Opus and GPT-4 possess formidable general reasoning capabilities. However, when deployed in African educational applications, they frequently suffer from two critical pitfalls: syllabus hallucinations and premature answer spoiling.</p> <p>A student preparing for JAMB Chemistry w

<blockquote> <p><strong>"Building builders, not just learners."</strong></p> </blockquote> <p>Technology is changing faster than ever, and cloud computing has become the backbone of today's digital world. From global companies to ambitious startups, organisations everywhere depend on the cloud to build applications, manage data, and deliver smart, scalable solutions. Recognising how important these technologies have become, the <strong>AWS Student Builder Group (SBG)</strong> at <strong>Graphic

<pre style='white-space:pre-wrap;width:81ex'>fix: avoid loading saved prompts for usage metrics (#138428)</pre>

<p>There are two kinds of provider in this field. One explains what you could do with AI. The other operates something with it and therefore knows where it breaks.</p> <p>The difference does not show in the sales conversation, it shows eight weeks later. When a system runs for four months straight, problems appear that are in no manual: storage grows and nobody clears it. A model suddenly answers differently than before. An access token expires, and the failure does not arrive as a failure but a

A lot of my non-developer friends still think that AI is basically just Google or something you ask to write something and then copy paste it around... if they use it at all. When I ask them why they don't use the good stuff like Claude Cowork or its competitors, they tell me a few things: 1. Well, I don't wanna pay a bunch of money for it, and the $20 plans run out real fast. 2. I'm dealing with sensitive info (student grades, therapy notes, legal work, etc) and I don't wanna give it to anyone.

Article URL: https://airuncode.com/ Comments URL: https://news.ycombinator.com/item?id=49569121 Points: 1 # Comments: 0