AI News Today

arXiv:2610.20635v1 Announce Type: new Abstract: Permutation generation with language models requires both text-conditioned ranking and a global one-to-one assignment between candidates and output positions. We study this problem through three diffusion-LLM parameterizations: hard-assignment prediction over shuffled candidate-index sequences, soft-assignment prediction over doubly stochastic matrices, and direct ordering-score prediction followed by sorting. The three families place permutation c

arXiv:2610.20642v1 Announce Type: new Abstract: Recent latent reasoning methods move beyond discrete Chain-of-Thought (CoT) by replacing sampled tokens with continuous soft embeddings or hidden states, enabling richer intermediate representations and preserving alternative reasoning paths. However, the need for latent reasoning varies across decoding steps: confident predictions favor discrete commitment, whereas ambiguous steps benefit from preserving multiple plausible alternatives. Existing a

arXiv:2610.20652v1 Announce Type: new Abstract: Always-on virtual assistants increasingly personalize on-device to new user-specific commands via few-shot fine-tuning (e.g., LoRA), but this adaptation is evaluated almost exclusively on whether it learns the new commands, not on whether the assistant can still recognize an input that matches none of its intents - the out-of-scope (OOS) rejection problem. We show that naive few-shot LoRA personalization degrades OOS-rejection AUROC (area under the

arXiv:2610.20785v1 Announce Type: new Abstract: Large language models are increasingly used in social interactions, where appropriate responses require recognizing implicit cues, inferring mental states, identifying conflicts, and considering potential consequences. However, existing benchmarks often focus on specific social abilities, constrained task formats, or particular interaction settings, offering limited coverage of social awareness across diverse contexts. To address this gap, we intro

arXiv:2610.20927v1 Announce Type: new Abstract: Long-horizon LLM agents rely on context compaction: when the interaction history outgrows the context window, a compressor replaces it with a natural-language summary. We show that these summaries contain more than a record of what happened. Compaction prompts ask for decisions and next steps, so each summary mixes facts established by the environment with reasoning written by the compressor, and the agent cannot tell the two apart afterwards. In a

arXiv:2610.20959v1 Announce Type: new Abstract: Answering questions over heterogeneous data means combining evidence from files, databases, documents and the web. Data agents automate this, but their success depends on context. Three context-engineering techniques (CETs) help by shaping it: semantic layers describe the data, orchestration splits work among agents, and compaction condenses interaction history. Existing studies validate each CET alone, leaving open how agent behavior shapes its ef

arXiv:2610.21025v1 Announce Type: new Abstract: When does a language model acquire a capability? Developmental studies usually answer this by tracking when models start producing the correct token. But that prediction passes through the unembedding matrix, or readout, which is itself still being learned during pretraining. We show that a model's hidden states can encode task information before its own readout expresses it, so timelines read from outputs conflate what the model encodes with what

arXiv:2610.21042v1 Announce Type: new Abstract: Systematic literature reviews follow the PRISMA process of screening scholarly works and extracting evidence from included studies. Deep research tools based on large language models (LLMs) can search and summarize the literature but do not record which studies are excluded or where extracted values come from. We develop a conversational multi-agent system called ReviewPilot that automates the PRISMA workflow from paper screening to information ext

arXiv:2610.21046v1 Announce Type: new Abstract: Peer review is under strain from the growth of submissions, and review comments written by humans or AI can misstate what a paper says. We introduce ReviewFact, a benchmark that labels such factual errors in review comments under one definition across NLP, machine learning, and the natural sciences, and that measures cost and latency alongside accuracy. A frontier LLM checks comments well but must read the whole paper for each check, while cheap de

arXiv:2610.21062v1 Announce Type: new Abstract: As large language models become increasingly capable of generating, and assisting with, scientific writing, there is a need for greater transparency and disclosure around the use of AI in research. Consequently, a veritable subfield of research and a number of commercial services aim to flag AI-generated content automatically. However, existing detectors which rely only on text extracted from PDFs fail to benefit from the rich visual information co

arXiv:2610.21098v1 Announce Type: new Abstract: The success score of a tool agent reflects both its model and the budgets of the harness that runs it. We present ToolBottleneckBench, a benchmark that measures tool-agent performance under controlled pressures and fixed harness budgets. Synthetic enterprise tasks stay fixed while one pressure varies: the registered tool catalog, the visible tool menu, the certified workflow, or the encoding of returned records. An exact grader records the first un

arXiv:2610.21126v1 Announce Type: new Abstract: Sparse KV compression reduces long-context inference costs by retaining selected historical states, yet these states typically keep their coordinates from the full history. Under Rotary Position Embedding (RoPE), gaps left by evicted tokens continue to affect query-key matching, potentially limiting the use of retained evidence. We study post-selection positional geometry on a fixed selector-produced cache, examining how positional organization aff

arXiv:2610.21143v1 Announce Type: new Abstract: Spoken Dialogue Models (SDMs) typically share a speech-token space for generation and synthesis, leading to long autoregressive sequences and costly inference. Lower-rate speech tokenizers shorten sequences but may introduce compression-quality trade-offs and costly replacement without breaking this coupling. To address this limitation, we propose CoSTA, a compatibility-preserving post-hoc framework that decouples generation and synthesis token spa

arXiv:2610.21149v1 Announce Type: new Abstract: Self-generated In-Context Learning (ICL) enables LLMs to synthesize their own few-shot demonstrations, often outperforming manual prompts. However, existing methods filter demonstrations by observation, assuming a rationale is high-quality if it leads to the correct output. We show such filtering suffers from \textit{observational equivalence}, whereby both faithful causal rationales and spurious post-hoc rationales can yield same outputs and are s

Article URL: https://lobby.domains/ Comments URL: https://news.ycombinator.com/item?id=50113913 Points: 1 # Comments: 0

Article URL: https://www.businessinsider.com/sergey-brin-google-gemini-ai-argon-recursive-self-improvement-rsi-2026-10 Comments URL: https://news.ycombinator.com/item?id=50113841 Points: 8 # Comments: 6

Article URL: https://swarmcha.se/posts/chinese-agent-fleet-full-report Comments URL: https://news.ycombinator.com/item?id=50113792 Points: 3 # Comments: 2

Article URL: https://lore.kernel.org/lkml/CAHk-=whTs09_ozN+e09GU68N1wPP56MUzDjrg9rnKxQUa0A_4Q@mail.gmail.com/ Comments URL: https://news.ycombinator.com/item?id=50113790 Points: 6 # Comments: 3

Article URL: https://machinelearning.apple.com/research/mlx-distributed-inference Comments URL: https://news.ycombinator.com/item?id=50113743 Points: 7 # Comments: 1

Article URL: https://www.youtube.com/watch?v=SJCw54PUdXg Comments URL: https://news.ycombinator.com/item?id=50113697 Points: 2 # Comments: 1

Article URL: https://gizmodo.com/researchers-stranded-an-ai-agent-on-a-virtual-island-things-got-very-weird-2000825998 Comments URL: https://news.ycombinator.com/item?id=50113668 Points: 3 # Comments: 0

Article URL: https://www.theregister.com/devops/2026/10/15/24000-ai-experiment-ports-typescript-compiler-to-rust/5303409 Comments URL: https://news.ycombinator.com/item?id=50113643 Points: 3 # Comments: 1

Article URL: https://github.com/TheWariend/MuscleMemory Comments URL: https://news.ycombinator.com/item?id=50113506 Points: 3 # Comments: 0

Article URL: https://www.404media.co/video-game-reverse-engineering-community-shuts-down-because-ai-has-made-it-too-easy/ Comments URL: https://news.ycombinator.com/item?id=50113499 Points: 6 # Comments: 2

Article URL: https://whosellsdata.com/ Comments URL: https://news.ycombinator.com/item?id=50113318 Points: 1 # Comments: 0

Article URL: https://firmenza.com Comments URL: https://news.ycombinator.com/item?id=50113168 Points: 2 # Comments: 2

Article URL: https://au.finance.yahoo.com/news/perplexity-open-source-tool-used-180212360.html Comments URL: https://news.ycombinator.com/item?id=50113012 Points: 2 # Comments: 1

Article URL: https://twitter.com/antigravity/status/2110866531932316011 Comments URL: https://news.ycombinator.com/item?id=50112997 Points: 2 # Comments: 1

NVIDIA, working with IIT Madras and its AI4Bharat lab, on October 16, 2026 introduced Nemotron-pretraining-Indic-v1, a synthetic pretraining corpus of approximately 2.8 trillion tokens across 22 Indian languages that is releasing soon, together with the open, GPU-accelerated data generation pipeline used to create it. The 22 languages are those listed in India's Constitution, which the announcement describes as the first languages for 97% of Indians, nearly 1.2 billion citizens at the 2011…

Article URL: https://www.zombieai.co Comments URL: https://news.ycombinator.com/item?id=50112964 Points: 3 # Comments: 0

<p><strong>Research:</strong> <a href="https://github.com/simonw/research/tree/main/sqlite-sparql-rust#readme">sparqlite: SPARQL 1.1 inside SQLite, written in Rust</a></p> <p>This research task is really an experiment in running Claude Code for web in "Ultracode" mode, as described in <a href="https://code.claude.com/docs/en/workflows">Orchestrate subagents at scale with dynamic workflows</a> in the Claude Code documentation.</p> <p>With Ultracode turned on, Claude Code makes aggressive use of d

Article URL: https://marquee.page Comments URL: https://news.ycombinator.com/item?id=50112730 Points: 2 # Comments: 4

<p>Meta says Muse excels at ‘multi-step reasoning’ and ‘complex software engineering’ as it races to catch up to Chinese rivals</p><p>Meta on Thursday launched a new open-weight AI model which it claims rivals, and in some cases outperforms, similar products from leading Chinese developers.</p><p>Muse Spark Open “is the most capable US open-weight model to date,” Meta’s chief of AI Alexandr Wang wrote on X.</p> <a href="https://www.theguardian.com/technology/2026/oct/16/meta-releases-muse-spark-

<p> Write a to-do and your AI agent gets it done </p> <p> <a href="https://www.producthunt.com/products/todo-is?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/1281979?app_id=339">Link</a> </p>

<p> Krea's AI image model: 2K edits in under 8 seconds </p> <p> <a href="https://www.producthunt.com/products/krea?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/1281975?app_id=339">Link</a> </p>

The FTC sent OpenAI and Anthropic 50-page demands for records on security incidents, safeguards, marketing and privacy, Axios reported. Meta's new business terms, effective October 30, let it scrape and train on publishers' websites. Investors at Anthropic's investor day discussed a $1.5 trillion IPO valuation, down from about $2 trillion. Character.ai filed for bankruptcy. Citizen Lab traced spyware sent to Al Jazeera devices through targeted ads. Accenture now funds $29 a month of Claude per e

Two engineers shipped Stack Overflow for Agents in a few weeks with Codex, and its ChatGPT plugin helped drive up to 10× growth in agent creation.

Glimpse from Elorian is now available on AI Gateway. The visual reasoning model accepts images and video to follow objects through motion, interpret spatial relationships, and connect an action to its physical consequences. Glimpse can return to visual evidence as it works: zooming into a detail, revisiting a short interval of video, or tracking the same feature across frames. This allows the model to support use cases like scientific chart analysis, video captioning, visual counting, and roboti

Perplexity Decider V1.2 27B and Decider Flash V1.2 4B are now available on AI Gateway . Both are multimodal decision models for classifying content, routing requests, and scoring text, structured records, and images. The models answer typed questions with probabilities. Boolean questions return the likelihood of yes, Choice questions select from categories you define, and Score questions assess content against an ordered rubric. Your application can use those results to route a support ticket, f