Following a UK regulator ruling, Google plans to allow website owners to opt out of AI search features and will add clearer attributions and links in AI overviews, aiming to give publishers more control over content usage.

Leading AI executives and scientists have signed a public letter calling for mandatory screening of synthetic DNA and RNA orders to prevent misuse for biological weapons, warning that AI could lower the barriers to creating such weapons.

Minimi, from the makers of Shram, is a Mac app that silently listens to your activities (docs, calls, messages, tabs) and gives Claude live context. All processing is on-device, ensuring privacy.

AgentRail is a Cloudflare edge layer that returns deterministic Markdown responses to known AI agents while serving regular HTML to humans and traditional crawlers. It includes bot detection, Markdown extraction, crawling, and Worker runtime components, with background cache warming and Cron-triggered sitemap crawling.

xAI has released "grok-imagine-video-1.5-preview," an image-to-video model that turns still images into cinematic videos at up to 720p based on text prompts. Multiple clips can be stitched together into longer scenes.

This article explores the real-world return on investment from AI agents, particularly for individuals and small teams. The author argues that leverage comes from parallelizing bounded execution loops with tight feedback loops, not from replacing humans. Key insights include the importance of system design (model, harness, tools, environment, evaluator), the value of manual testing, and the advantage small teams have due to short feedback cycles. The article warns of the 'Sloptember' failure mode where agents increase the volume of mediocre work without improving quality.

Gate is a deterministic PII redaction tool written in Rust for AI agents. It uses regex, column heuristics, and Luhn checks instead of LLMs. It intercepts Bash commands and MCP tool calls via hooks, supports multiple harnesses, and provides scanning, real-time redaction, and aggregated reporting while keeping data local.

Chinese AI startup DeepSeek tops a US corporate spending index as US firms adopt cheaper Chinese models, replacing expensive options like OpenAI and Anthropic, amid a broader shift to open-source models.

asd is a smart universal file viewer for the modern terminal that replaces traditional cat with intelligent file type detection and beautiful rendering. It supports syntax highlighting, structured data formatting, tables, Markdown, archives, office docs, PDFs, media metadata, certificate parsing, Git integration, tail mode, side-by-side diffing, and more.

Hscli is an open-source, scriptable CLI for Help Scout, designed for terminals, CI pipelines, and AI agents. It outputs JSON, supports various output formats, and is secure by default.

Stanford researchers released OpenJarvis, an open-source framework that runs inference, agents, memory, and learning entirely on-device. It decomposes a personal AI system into five composable primitives and lands within 3.2 points of the best cloud model at roughly 800× lower marginal API cost.

Phylax is a Windows security layer that uses OS-level ACL enforcement to prevent AI coding agents from accessing your private files. It runs 100% locally with no accounts, cloud, or telemetry, supports multi-agent detection, and offers anti-bypass protection.

Microsoft's experimental fork of Windows Terminal integrates native AI agent support, featuring a context-aware agent pane, automatic error detection, and session management.

Uber is cutting 23% of its people team, which includes HR and recruitment staff, as it seeks to streamline operations under new president Jill Hazelbaker. CEO Dara Khosrowshahi said the changes are necessary. The cuts account for well under 1% of Uber's 34,000 employees and are not due to AI. Uber also set tiered spending caps on employees' agentic AI tools, with a base tier of $1,500 per month.

Malicious packages exploit binding.gyp and index.js to execute payload on npm install, downloading Bun runtime, harvesting credentials, and injecting into GitHub Actions workflows for further spread. Affected package: ai-sdk-ollama versions 0.13.1, 1.1.1, 2.2.1, 3.8.5.

This article explores whether open source software that does not use AI will become obsolete in an increasingly AI-driven world, arguing that while AI integration is growing, many non-AI projects retain unique value.

England's qualifications watchdog warns that new wearable devices like smartglasses and invisible earpieces could increase exam cheating, while AI use in GCSE and A-level coursework is also under scrutiny.

Tech journalist Joanna Stern spent 2025 voluntarily letting artificial intelligence take over nearly every aspect of her life, from answering texts and cooking to driving and even romantic companionship. Her resulting book, 'I Am Not a Robot,' explores the implications of AI's growing capabilities and what comes after.

This paper proposes a method to estimate normal and shear pressures in prosthetic sockets using sparse sensing and mechanical models. By validating with global wrench and local interface loads, a quasi-static spring-mass contact model with least-squares parameter identification reduces measurement offsets, offering objective fit metrics.

This paper presents DLO-Lab, a differentiable simulator designed for deformable linear objects (e.g., ropes, cables, rubber bands). It models a wide range of material properties, provides a benchmark suite of tasks, and introduces a specialized agent to handle topological complexity and grasp sensitivity. Various policy-learning algorithms are evaluated, with sim-to-real transfer experiments.

This paper proposes a risk-aware model predictive control (RA-MPC) framework that incorporates prediction sets to guarantee risk control below a user-specified threshold without assumptions on the uncertainty distribution. By extending conformal risk control to general spectral risk measures, it achieves distribution-free risk quantification. Simulated vehicle obstacle avoidance scenarios demonstrate improved safety and reduced solve time compared to a baseline RA-MPC framework.

This paper presents Affordance2Action (A2A), a benchmark-centered learning framework for scene-level, task-conditioned part affordance grounding. Its core is A2A-Bench, a manipulation-oriented benchmark covering single-region and multi-region instruction correspondences in everyday scenes. Built with the A2A-AffordGen annotation pipeline, experiments show A2A exposes gaps in existing methods and improves task-level localization for downstream manipulation.

This paper proposes a distributed, risk-aware multi-agent NBV framework where each robot maintains a private local 3D Gaussian Splatting map and jointly maximizes expected information gain via Consensus ADMM, reducing communication by orders of magnitude while maintaining mapping quality and safety.

This study compares three haptic guidance models—spring-damper, potential field, and guiding tube—in teleoperation tasks. No universal best model is found: spring-damper excels in cluttered environments, potential field works well in free space but risks near obstacles, and guiding tube offers a balanced compromise. New objective metrics are proposed, showing that guiding force magnitude correlates with comfort and trust scores.

CLAW is a fully end-to-end self-supervised framework that learns world models and continuous latent action representations from action-free videos using adversarial latent regularization and diffusion-based generation. It enables imitation learning from observation and goal-directed planning, outperforming existing methods.

AgenticDiffusion is a multi-view UAV navigation framework that coordinates language-guided reasoning, open-vocabulary target grounding, vision-based diffusion planning, and NMPC. In 40 real-world trials, it achieved an 80% mission success rate and 100% trajectory generation success.

The paper introduces Spatial Artifact Coherence (SAC), a metric measuring when patch-based rPPG outperforms global methods under codec compression. Experiments on 280 subjects across 11 codec variants show SAC explains 93.8% of PCA advantage variance. Non-MPEG-4 codecs (SAC 0.10-0.18) achieve 84-90% PCA win rates, while MPEG-4 (SAC 0.48-0.59) only 61% with 5.8x reduction in improvement. P-Hybrid is identified as the most robust algorithm, with PatchPCA advantage requiring SAC<0.30 and low-to-moderate motion.

A new benchmark, GroupToM-Bench, evaluates multimodal large language models on group-level theory of mind. While models excel at individual mental state reasoning, they fail at understanding collective behavior that emerges nonlinearly from social tensions and structural constraints. The benchmark covers a causal chain from micro-level beliefs, desires, and intentions to group tension and macro-level outcomes, using a seven-level cognitive audit. Experiments reveal a significant gap between current models and human baselines.

This paper introduces Orli (Ordered Regression of Lines), an end-to-end model that unifies text line detection and reading order prediction as a single image-to-sequence task. Trained on 196,691 pages across ten writing systems, Orli marginally exceeds state-of-the-art on cBAD line detection without dataset-specific training, achieves near-perfect coverage and ordering on multiple reading-order benchmarks zero-shot, and adapts to specialized out-of-domain layouts with limited fine-tuning. Code and weights are open-sourced.

Pinpoint is a new image geolocation method that combines internet photos and street-view imagery using a retrieve-and-rerank architecture. It trains a contrastive embedder on both Flickr and street-view data to learn a shared image-GPS space, then uses an attention-based reranker to rescore candidates. Without relying on multimodal LLMs, it achieves state-of-the-art results on multiple benchmarks.