Research

SpaceXAI's Grok Build AI coding tool was spotted uploading users' entire codebases to Google Cloud before it was reported, and the company turned it off. The Register reports that Cereblab published findings on Monday showing how the Grok Build CLI was packaging and uploading entire code repositories, "including files it was told not to open and secrets deleted from history," significantly more data retention than similar tools like Claude Code.

AWS announced an expansion to Security Hub to monitor Azure resources, along with new tools for protecting AI workloads, including GuardDuty AI Protection, AI-powered investigations, and an AI inventory.

An evaluation of 1,018 real AI prompts reveals an average score of 54/100, but robustness averages only 31.5, and 96% of prompts have their weakest dimension in robustness. Only 10.5% reach 75 (the production bar). The report highlights the common 'happy path' trap in prompt engineering and offers simple improvements.

The researcher who exposed Grok Build uploading users' entire repositories to cloud storage says the transfers have stopped after a server-side change. Musk has promised to delete all previously uploaded user data. However, the researcher says the privacy command was not the actual fix.

Meta prompting is a technique where a prompt is used to create, improve, or control another prompt. It shifts the model from direct task execution to prompt design, improving consistency and scalability for repeated tasks.

An open-source Claude skill called vibe-check, created by a seasoned product manager, helps beginners go from a vague idea to a buildable blueprint, ensuring they build the right thing rather than just building it right. It includes problem discovery, idea validation, user experience mapping, tech stack recommendations, growth loop design, and produces a comprehensive plan document.

Agent Shell is a hardened Linux box you hand your AI agent root on, over SSH or browser. gVisor-sandboxed so even root inside cannot touch the host, preloaded with agents like borg, Claude Code, Codex, Gemini, and live monitoring. Built on infrastructure operated since 2009.

Using generative AI enables parallel execution of comprehensive user flow testing at scale. This solution demonstrates how to build a cloud-deployed UX testing platform that automatically generates test scenarios from documentation, executes user flows at scale using the intelligent navigation capabilities of Nova Act, and provides actionable insights through automated analysis.

As AI and optimization become commoditized, traditional supply chain planning no longer provides competitive advantage. Research shows most organizations lack visibility into their Tier 1 suppliers. Based on an Emerj podcast series, this article explores how scenario-driven modeling, AI-accelerated scenario analysis, and unified design environments enable better decision-making under volatility.

A longitudinal study of an enterprise '2x mandate' to double merged pull requests per engineer found that throughput eventually reached 2.09x the pre-mandate baseline, with gains linked to AI adoption and usage intensity. Code review was restructured, with automated review overtaking human review and reviewer load doubling.

Researchers released a set of formulas for fundamental mathematical constants designed to evaluate AI mathematical skills. The problems include proven (temporarily encrypted) and unproven formulas, testing AI's reasoning capabilities.

Aevum Realm Architect is a free, LLM-powered RPG engine created by Arcanum RPGs. Players start as a serf with one copper piece and rise through trade, war, diplomacy, and intrigue to claim a throne. It features deterministic combat, a tag-based economy, and a strict feudal hierarchy enforced by the Deference Engine. Runs on ChatGPT, Claude, or Gemini.

Browser agents that operate via pixel-level UI interactions are slow and unreliable. Pluno's approach is to have agents learn product-specific skills (API paths, workflows) from browser sessions, then execute tasks directly via code, achieving much higher speed and reliability.

Mission is a fast, robust HTML parser and CSS selector engine written in Rust, with zero dependencies, no network layer, crash immunity, and built-in MCP tool support for AI agents to extract structured data locally.

A security researcher source-reviewed 200+ multi-tenant AI and SaaS products for cross-tenant data exposure. 78 products had the same flaw: write endpoints had authorization checks, but adjacent read endpoints did not. The post explains the pattern, lists fixed products, and gives advice.

Are AI-generated passwords truly random? Research suggests AI passwords are far less secure than you might think.

StanRose introduces an AI-powered mock-interview coach that evaluates not only your answers but also your delivery, including tone, pace, and confidence.

Demis Hassabis, CEO of Google DeepMind, proposes a global AI watchdog with the power to halt dangerous frontier models. He argues the US should lead the effort and hopes to establish the organization by year-end.

This article examines the open source toolkits for building AI agents in 2026, analyzing key layers like orchestration, memory, protocols, and browser control, and offering strategies for choosing the right tools based on constraints such as latency, audit trails, and language stack.

A professor interviewed by the Chronicle of Higher Education asserts that AI is one of the top four reasons faculty are rushing to retire, citing untenable work conditions, institutional chaos, political assault, and AI's embrace across academia as part of a larger effort to dismantle democratic higher education.

A combination of fact and fiction leaves the celebrated documentarian’s puzzling project about software training wanting for depth. Marc Isaacs’ new film is a curious, intriguing, semi-sincere affair that I couldn’t make friends with. It is an odd, shallow piece of work about artificial intelligence that is itself exasperatingly artificial, a self-aware docudrama hybrid.

Vizro is an AI-powered iOS app that turns CSV or Excel files into interactive dashboards in minutes. It offers automated analysis, natural language queries, story mode, and easy sharing, all for $9.99/month.

As AI makes code generation cheap, costs shift from generation to ownership. To avoid technical debt, coding agents need an open-source intelligence layer that helps them reuse trusted components before generating new code.

This article contrasts learning from curated datasets with learning from raw experience. It shows that SGD and its variants absorb noise in online data streams, failing to learn only predictable components. The IDBD algorithm, however, can selectively assign credit and learn only useful associations. Extensions to neural networks (NetworkIDBD) demonstrate similar advantages on the NoisyMNIST task. The authors argue that better credit assignment algorithms are needed for online continual learning.

Mistral AI introduced Robostral Navigate, an 8B embodied navigation model. It moves robots from a plain-language instruction using only a single RGB camera, with no LiDAR or depth sensors. The model reaches 76.6% success on R2R-CE validation unseen through a pointing method, prefix-caching training, and CISPO online reinforcement learning.

Anthropic's new research reveals a hidden 'J-space' inside LLMs where words influence reasoning without appearing in output, offering insight into model decision-making but also raising questions about anthropomorphization.

A free benchmarking tool to evaluate your engineering team's AI agent maturity in 5 minutes. Based on hundreds of discussions with engineering leaders, it uses a 1-5 scale covering from suggestions only to fully autonomous multi-hour workflows.

This article debunks common lies from AI data center proponents, such as claims of innovation and job creation. It argues that these projects primarily bring pollution, water strain, and few local jobs, and criticizes media and corporate think tanks for misleading communities. The author warns that broken regulations make it difficult to hold tech companies accountable.

RQSHC V64I is a native Windows image compression research tool that uses a proprietary RQI format. It supports PNG, PPM, BMP input and achieves ~33% size reduction with very high SSIM. The core is built with C++17 and x64 assembly with AVX2 optimizations. Free for non-commercial use.

Copyright law is a key obstacle for AI companies investing in Australia. Creators accuse AI firms of using their work without permission, while tech groups argue the law blocks investment. The government considers multiple reform options but has not decided.

This survey systematically reviews the state of the art in embodied Vision-and-Language Navigation (VLN), organizing methods along two orthogonal dimensions: action paradigms (hierarchical vs. monolithic) and model paradigms (discriminative vs. generative). The authors conduct a real-world evaluation on a physical robotic platform across ten diverse scenes, revealing a significant simulation-to-real gap: a monolithic RGB-only method achieves 61% success in simulation but drops to 22% in the real world, while a hierarchical framework attains 51% real-world success. Key challenges in perception, decision-making, and control are highlighted.

This paper proposes a human-efficient post-training pipeline that enables a small number of human operators to supervise multiple robots through specialized division of labor and automatic trajectory segmentation using VLAC-CUT. Validated on four real-world manipulation tasks, the final policies achieve 80%-95% success rates and improve task throughput by 1.7x-4.2x over the base model.

This paper proposes a risk-field enhanced closed-loop digital twin framework for safety validation of autonomous driving systems. The framework integrates physical data acquisition, virtual reconstruction, risk-aware scenario generation, and algorithm evaluation, using a driving risk field as a unified intermediate representation to identify high-risk scenarios and provide safety guidance for reinforcement learning policies. Experiments show the method improves targeted validation and interpretability, but its effectiveness is bounded by model fidelity and sim-to-real transfer.

OmniSCS proposes an innovative system for generating photorealistic safety-critical scenarios (SCS) with high physical fidelity, enabling closed-loop simulation testing. It consists of a Fully Editable Driving World Construction module and an SCS Synthesis module that preserve data fidelity during scene editing. Experiments on nuScenes, Waymo, and KITTI datasets show that OmniSCS outperforms state-of-the-art methods in edited scene fidelity and supports real-time (13Hz) closed-loop testing, providing a safer and more cost-effective solution for autonomous driving development.

UAV swarms have potential in SAR and environmental monitoring but face limitations in situational awareness, connectivity, and cybersecurity. This paper proposes LAUS, an LLM-centric agentic AI framework integrating perception, memory, reasoning, and action for adaptive swarm behavior. It reviews enabling technologies, analyzes threats like Priority Manipulation Attacks, and identifies open challenges including hallucination-resistant reasoning, onboard LLM deployment under SWaP constraints, and standardized security benchmarks.

Researchers propose SWIFT, a unified framework integrating small-world networks with traffic flow theory for trajectory prediction in autonomous driving. It introduces structural inductive biases via a Small-World Interaction Network and a Flow Regime Encoder, outperforming baselines on nuScenes, MoCAD, and NGSIM datasets, with improved generalization and robustness.

A new differentiable physics framework for robust trajectory optimization of reusable launch vehicles introduces a Differentiable Particle Tube Control (DPTC) scheme that integrates actuator saturation constraints. Monte Carlo simulations show improved robustness over conventional methods by proactively managing performance trade-offs.

Proposed DecisionPerceiver architecture projects dynamic agent features into a fixed-size latent space, regulating granularity with latent queries, improving scalability. Evaluated across three driving scenarios shows consistent gains and generalization.

A new framework called RoboNav-Arm enables robotic manipulators to safely navigate and avoid obstacles in cluttered environments using agentic AI. It combines real-time obstacle detection, semantic reporting, central coordination, and adaptive motion planning, tested in Gazebo simulations.

EgoSteer is a full-stack system that enables steerable dexterous manipulation by pre-training a VLA model on 9.6K hours of egocentric human videos and post-training on robots. It achieves robust execution of free-form instructions across 40+ tasks, with failure recovery and few-shot adaptation to long-horizon tasks like box folding at 75%+ success.

Real-image diffusion inversion faces a quality-cost trade-off. This paper reveals two mechanisms: element-wise compression asymmetry and trajectory binding, leading to Noise-Anchored Reverse Correction (NARC), a training-free method that outperforms baselines with drastically reduced storage.

A paper accepted at ECCV 2026 presents a new approach to wearable motion capture that works with any combination of consumer devices like smartphones and smartwatches, introducing the WHIP model and a comprehensive dataset spanning 50 activities, along with a systematic study of sensor complementarity.

This paper proposes a Generalized Deep Non-negative Matrix Factorization (G-DNMF) method for SAR automatic target recognition. It overcomes the error accumulation and local optima problems of layer-by-layer decomposition in existing DNMF methods by deriving globally optimal update rules using the Lagrangian multiplier method. Experiments on MSTAR and OpenSARship datasets show improved stability and recognition performance over existing DNMF algorithms.

Multi-Modal Knowledge Graphs (MMKGs) enrich entities with modalities like text and images, but entities with highly similar multi-modal features remain hard to distinguish. Temporal information can serve as an additional modality for disambiguation, yet existing approaches rarely treat time as a separate modality due to sparse temporal semantics and noise from multiple timestamps. This paper proposes Time Imprint, a framework that treats time as an entity-level modality and aligns temporal, textual, and visual representations via a three-view contrastive objective. It also designs a compact timestamp subset selection with attention pooling to balance specificity and robustness. Experiments on three MMKG benchmarks show state-of-the-art link prediction, with Hits@1 improvements up to 6.07% overall and 58% on the top-1% ambiguous samples.

This paper proposes a knowledge-constrained shape optimization framework that translates expert knowledge and user intent into quantifiable parameters for DFFD-based deformation operators. A Mixture-of-Experts Neural Operator (MoE-NO) improves drag prediction and trend consistency on heterogeneous datasets. Experiments show MoE-NO achieves 1.16% MAPE and 94.34% trend accuracy, with CFD-validated drag reductions of 4-10%.

A new AI system called ReflectWorld-MM enables assistants to continuously process and remember open-ended video streams by organizing memory around persistent entities rather than frames, achieving state-of-the-art results on six benchmarks.

RSLoRA is a training-free, gradient-free method for allocating LoRA ranks based on activation-space geometry. It introduces virtual representational probing to identify high-sensitivity layers, outperforming state-of-the-art allocators like AdaLoRA and GoRA.

WiCAT, a multi-subject model using self-supervised pretraining, outperforms single-session models and enables zero-shot behavior decoding on unseen subjects in widefield calcium imaging.

Researchers propose DUNE, a training-free framework that refines diffusion models by detecting and suppressing early-stage fluctuations in deep latents, reducing artifacts and hallucinations while improving fidelity across both U-Net and Transformer backbones.

This paper investigates the feasibility of training a reasoning language model in Japanese. By applying GRPO to a Japanese continually pretrained model based on Qwen-3-Swallow-8B, the authors find that reasoning-language control is achievable, yet performance at best matches English-reasoning baselines. On Japanese cultural benchmarks, the model performs worse, indicating that reasoning in Japanese does not automatically improve culturally relevant tasks.