AI News Today

arXiv:2609.16073v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) serves as the primary memory architecture for long-horizon autonomous agents. However, treating shared memory as an append-only stream introduces \textit{Semantic Shadowing}, a critical failure mode where conflicting historical observations accumulate and statistically dominate valid recent updates. In dynamic environments, this results in severe state divergence as agents retrieve and act upon obsolete facts. T

arXiv:2609.16076v1 Announce Type: new Abstract: Large Language Models (LLMs) excel at programming tasks but frequently fail at deterministic, fine-grained reasoning in natural language, relying heavily on semantic approximations rather than robust symbolic execution. To bridge this gap, we propose MIMIC, a framework that leverages executable code as a rigorous medium for reasoning data synthesis. MIMIC fundamentally transforms algorithms into verifiable reasoning trajectories through narrative f

arXiv:2609.16095v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) has emerged as a powerful paradigm for improving the quality of generated contents of Large Language Models (LLMs) by grounding responses in external knowledge, thus reducing hallucinations and factual errors. However, recent studies have highlighted a critical vulnerability: adversaries can exploit the retrieval process to extract personally identifiable information (PII) from the underlying corpus. To mitigate

arXiv:2609.16268v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly interleave natural language reasoning with external tools such as web search and code execution. These tool-use policies are often optimized via reinforcement learning (RL), which can amplify spurious correlations in the training data. In this work, we study when and why RL-trained agents learn shortcut tool-selection policies: invoking tools based on superficial prompt cues rather than genuine task re

arXiv:2609.16312v1 Announce Type: new Abstract: One-to-many machine translation (MT) is computationally expensive for autoregressive (AR) systems, which suffer from linear latency scaling with both sequence length and the number of target languages. We explore how diffusion can enable multilingual translation with a discrete diffusion framework that refines all target languages in parallel, achieving sublinear latency scaling with the number of targets, and supports deployment as a single unifie

arXiv:2609.16366v1 Announce Type: new Abstract: When foundation models describe people, recent work in AI fairness, accessibility, and ethics recommends avoiding inferred identity labels (e.g., "she", "his") in favor of seemingly "objective" physical descriptions (e.g., "short hair", "a defined jawline"). Yet whether such descriptive language achieves gender-neutral communication remains an open empirical question. To study this, we introduce GAPA (Gender Associations of Physical Attributes), a

arXiv:2609.16372v1 Announce Type: new Abstract: Masked diffusion language models (dLLMs) generate text by iteratively denoising masked tokens with bidirectional attention. Extending reasoning across generation chunks normally requires keeping earlier generated text in context. We ask whether a dLLM can instead continue reasoning after that text is cleared, using only a fixed-size carried state. We implement this state as a small number of register tokens: dedicated fixed-position tokens whose co

What else don't we know? ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

A warning sign. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Article URL: https://www.vincentschmalbach.com/inference-is-the-last-llm-moat/ Comments URL: https://news.ycombinator.com/item?id=49721882 Points: 2 # Comments: 0

<p> AIM buddy list for your friends’ Claude Code and Codex usage </p> <p> <a href="https://www.producthunt.com/products/die-with-me-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1251910?app_id=339">Link</a> </p>

<p>Every family photo you ever posted would be fair game for AI companies unless you choose to opt out</p><p>Leaked documents from a government proposal <a href="https://www.theguardian.com/australia-news/2026/sep/16/pocock-says-labor-easing-copyright-protections-for-ai-datacentre-investment-would-throw-creatives-under-the-bus">show that the Albanese government is considering</a> reversing the fundamental rights of Australians under copyright law.</p><p>The proposal refers to an “opt-out model”.

'Jev' doesn't chat. It produces typed probabilistic decisions

Sam Altman and other tech CEOs say there are incentives to limit advancements in AI, as fears grow over the threats it poses to humanity.

AI isn't some new form of "alien mind," according to Jensen Huang. It's just hardware and software, so safety can be engineered by each AI product maker.

<p> Sell your product to AI agents </p> <p> <a href="https://www.producthunt.com/products/zeroclick?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1251841?app_id=339">Link</a> </p>

Jev from TypeSafe AI is now available on AI Gateway. Jev is a probabilistic decision model for software: state goes in, typed Choice, Score, and Boolean answers come out. Regular language models generate text one token at a time, which the application then parses and validates. Jev evaluates all declared questions in parallel and returns typed answers plus probabilities directly. That removes unnecessary text generation and makes it straightforward to automate clear cases while routing uncertain

Is Agentic reports now let you view your checks through one of four site types: Docs & content, Business, App, or Commerce. For example, the Commerce view highlights payment and checkout standards like x402, UCP, and ACP, while the App view highlights API discovery, authentication, error handling, and SDK support. Your score stays the same in every view, and you can change views on an existing report without rescanning. Only the highlighted checks change, so you can focus on what matters for you

Agentic LLM systems that generate code through multi-turn tool use face a fundamental context problem: each session starts from zero, discarding the configuration choices, domain constraints, data schemas, and tool-use patterns that made previous sessions productive. Naively persisting entire conversation histories is both token-inefficient and counterproductive—irrelevant context degrades generation quality. We introduce shared selective persistent memory, a memory architecture for agentic syst

Enterprise data lakes accumulate tables faster than human stewards can document or classify them, leaving columns with missing descriptions and unassigned governance labels. This documentation debt undermines data discovery, access control, and regulatory compliance. We present Glyph, a production system that frames two coupled problems, column description generation and column type annotation for data classification, as cooperating LLM agents orchestrated as stateful graphs. The Descriptor grou

Diffusion large language models are a compelling alternative to autoregressive models, yet existing RL methods for diffusion treat all denoising steps as equally important and rely on biased, high-variance likelihood estimates. We identify two fundamental weaknesses: the absence of temporal credit assignment across the denoising trajectory, and the systematic bias of mean-field likelihood estimates used for policy optimization. To address these, we propose Denoising-Aware Credit Assignment for G

Discrete flow matching generates text by iteratively transforming noise tokens into coherent language, but may require hundreds of forward passes. Distillation uses the multi-step trajectory to train a student to reproduce the process in a few steps. When the student underperforms, the usual explanation is insufficient capacity. We argue the opposite: the trajectory is the bottleneck, not the student. Each training trajectory is built through a chain of blind stochastic jumps with no evaluation

Conversational Large Language Models are post-trained on language that expresses specific behavioural traits, such as curiosity, open-mindedness, and empathy, and values, such as helpfulness, harmlessness, and honesty. This is done to increase utility, ensure safety, and improve the experience of the people interacting with the model. However, values are complex and inter-related – inducing one could modify behaviour on another. Further, inducing certain values can make models more addictive or

Article URL: https://www.youtube.com/watch?v=it7TAwjwH5E Comments URL: https://news.ycombinator.com/item?id=49720322 Points: 3 # Comments: 0

Poll data released Tuesday by The New York Times and Siena University confirms what we've already been seeing, and what politicians are responding to - AI and data centers are incredibly unpopular. Asked if they support or oppose the construction of data centers to power AI tech, 61 percent of the 1,503 likely voters surveyed […]

<p> A marketplace for AI agents you can text </p> <p> <a href="https://www.producthunt.com/products/text-agent-store?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1251830?app_id=339">Link</a> </p>

The ad regulator said it had "zero-tolerance" for ads promoting AI tools that could sexualise or objectify women.

Learn how recursive self-improvement works, what AI can already do, and why researchers worry that faster advances could outpace safety measures.

<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/gemini-live">Gemini Live audio</a></p> <p>Google released <a href="https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/">Gemini 3.8 Live and 3.8 Live Extended Thinking</a> today - two new speech-to-speech models that are a similar shape to OpenAI's <a href="https://openai.com/index/introducing-gpt-live/">GPT-Live</a> family.</p> <p>I pointed GPT-6 Astra Extra H

Know everything. Do anything. That was the message NVIDIA founder and CEO Jensen Huang brought to Salesforce Dreamforce Tuesday, joining CEO Marc Benioff onstage in an appearance that coincided with the announcement of Koa — Salesforce’s first CRM reasoning model, built on NVIDIA Nemotron 3 Super. Huang didn’t just take the stage. He walked into […]

The event is focused on "Windows, NVIDIA RTX Spark, Surface, and the broader PC ecosystem," with Satya Nadella and Jensen Huang among the speakers.

Article URL: https://cagritemel.com/blog/hn-broke-my-llm-agent-gate.html Comments URL: https://news.ycombinator.com/item?id=49719548 Points: 1 # Comments: 0

Get lifetime ReciBites Plus for $39.99 (reg. $360) to organize recipes, plan meals, create grocery lists, and get AI-powered nutrition and cooking help.

National outcry against data center construction has spread to Philadelphia, where officials suggested possible construction in a neighborhood already impacted by a now-defunct oil refinery.

Article URL: https://xyntetik.com/runner/truncation-safe-tool-calling/ Comments URL: https://news.ycombinator.com/item?id=49719027 Points: 2 # Comments: 0

Get the perks of a VPN plus extra AI-powered protection with this lifetime subscription to GhostShield AI-Powered VPN Premium Plan.

Intel spin-off Cornelis and newcomer Delos Data pitch open alternatives for scaling AI beyond the rack

It's been more than two years since the last major Windows event, so Microsoft is heading to San Francisco next month to outline the future of Windows and Surface devices. At an event on October 7th, the company says it will host a "conversation on how local AI will shape the next chapter of the […]

Pizza Bot runs locally and lets users interact with AI agents in an email-like interface. Perfect for blowing them off just like your human colleagues!