French startup Genesis AI unveils Eno, a robot designed around human capability rather than appearance. Eno may lack a head and legs, uses a wheeled base, and folds like a deck chair, but features hands that match human anatomy. Production is planned by end of 2026, targeting manufacturing, logistics, and more.

High cost of intensive AI use sees companies reining in AI spending

Spaturzu is an open-source SDK that wraps existing LLM clients (OpenAI, Anthropic, etc.) with a single import change to enable per-agent cost attribution, budget enforcement, and cross-provider fallback, without altering call sites.

A Pentagon official revealed that Elon Musk's Grok AI was used to launch over 2,000 munitions at Iran. The admission came in a court filing defending xAI against pollution claims. The use of AI in military strikes has raised concerns about civilian casualties and accountability.

Two related Oracle-backed projects published opposing policies on open-source contributions created with generative AI: the OpenJDK Governing Board approved an interim policy prohibiting such contributions, while the Coding Assistants policy from GraalVM permits them. Both projects require contributors to sign the same Oracle Contributor Agreement (OCA) for intellectual property.

Z.AI introduces GLM-5.2, a flagship model for long-horizon tasks with a solid 1M-token context, advanced coding capabilities with flexible effort levels, and an open-source MIT license. It achieves top-tier performance on long-horizon coding benchmarks, rivaling closed-source models.

For the second time in a week, the AUR was found to contain malicious applications. What can Arch Linux users do about this?

Databricks announces new AI platform capabilities at Data+AI Summit 2026: Genie Code for ML (coding agent integrated with ML stack), AI Runtime (public preview of serverless GPU training), and enhanced real-time ML support (low-latency, high-QPS Feature Store and Model Serving). These features accelerate the path from experimentation to production.

Claude Code out of the box only unlocks half its potential. By installing 27 specialized skills, developers can transform it into an AI engineering team comprising a React architect, frontend designer, QA engineer, and more. This article details the skill categories, installation methods, and how to create custom skills to maximize AI-assisted development.

Lovable has updated its Terms of Service to unify credits across Platform, Cloud, and AI Gateway, effective August 15, 2026. The new policy also clarifies subdomain rights and introduces auto-reload for credit purchases.

Mistral AI announces a new model coming this summer, the start of a larger family of fat but sparse models. Early access program in July for key partners.

ax-audit is an open-source CLI tool that audits websites for AI Agent Experience (AX) readiness, offering 18 checks including LLMs.txt, Robots.txt, content negotiation, and more, with scoring and detailed reports.

MiniMax released MSA, a sparse attention built on Grouped Query Attention. A lightweight Index Branch selects Top-k key-value blocks per query and GQA group; the Main Branch attends only to those blocks. It matches GQA on downstream benchmarks while reducing per-token attention compute 28.4× at 1M context.

Fig0 is an AI-powered tool that transforms text, sketches, references, and PDFs into publication-ready, editable vector scientific figures. It supports multiple export formats and is used by over 150,000 researchers.

Computational language built for humans and AI agents.

OpenAI introduced Deployment Simulation on June 16, 2026. The method replays past conversations through a new candidate model before release, then grades the completions to estimate deployment-time rates of undesired behavior. It reported a 1.5x median multiplicative error, reduces evaluation awareness, and extends to agentic coding via simulated tool calls.

Z.ai released GLM-5.2, an MIT-licensed open-weight frontier model focused on coding and long-horizon agentic tasks. It achieves top scores in frontend coding benchmarks, trailing only Fable 5, and leads in Design Arena. The model features a 1M-token context window, IndexShare sparse attention optimization, and improved multi-token prediction for speculative decoding. Community reactions are mixed: some hail it as a viable open-source alternative to proprietary models, while others call for more rigorous evaluations.

A solo founder launched a data annotation job aggregator that evolved into a hybrid platform with reviews, educational content, an AI Academy, and a global talent pool. With 10k monthly visitors and over 100 talent pool signups, the platform is on track to hit its first $1k month from candidate referrals, all within six months.

Cursor launches Origin, a code storage and Git hosting service for teams and AI agents to host, review, and collaborate on code. Available this fall.

AI Brand Kits is a brand system generator designed for AI coding agents. It instantly generates font pairs and color palettes, offers live previews, and exports DESIGN.md files to ensure brand consistency when working with AI tools.

Henji is an AI writing assistant that learns your personal style to generate replies, letting you keep the thinking while it handles the writing, avoiding generic AI tones.

This article analyzes four structural failure modes of AI memory systems at scale: cosine similarity's inability to discriminate within a domain, the decoupling of extraction quality from retrieval precision, session drift compounding noise across turns, and latency figures hiding session-level degradation. It proposes alias-weighted BM25 leveraging lexical priming as a replacement for semantic similarity.

Chatr is an AI tool that handles customer support, collects feedback, books appointments, and sells products.

This paper presents a modality-adaptive contactless respiratory rate monitoring framework for heterogeneous mobile robots with onboard edge computing. It combines brightness-adaptive sensor selection across RGB, thermal, NIR, and low-light cameras, keypoint-guided chest ROI extraction, and SQI-based filtering. Experiments on quadruped and wheeled platforms demonstrate generalization without per-platform retuning: RGB covers up to 8m, NIR up to 6m, thermal short-range only, low-light effective in complete darkness up to 8m. The framework supports autonomous triage and victim assessment in hazardous environments.

A new study proposes a transformer-based warm-start method for sequential convex programming (SCP) in the terminal approach of a space manipulator to a tumbling target, reducing iterations by 28% and runtime by 23% while preserving optimal control cost.

This paper introduces APOLLO, a hybrid framework that combines a lightweight personalized embedding model with selective large language model (LLM) assistance for abstention-aware object rearrangement in cluttered, partially erroneous environments. The framework uses uncertainty estimates to decide when to invoke the LLM, balancing efficiency, privacy, and reasoning. A new synthetic dataset, APOR, is also introduced. Experiments show APOLLO reduces LLM usage while maintaining or improving performance over prior methods.

To overcome the scarcity of paired vision-action data in robotics, researchers propose CAIP, a vision encoder that uses hand poses from egocentric human video as proxy for robot actions. Using only 88 hours of robot data and 32,041 hours of human video, CAIP outperforms leading encoders by over 30% on dexterous manipulation tasks.

ACE-Ego-0 is a unified Vision-Language-Action (VLA) pretraining framework that integrates egocentric human videos with robot data by converting human videos into pseudo-action trajectories and using a reliability-aware training objective, boosting model performance on benchmarks and real-world tasks.

This paper presents VL-MemKnG, a hybrid memory framework that combines a spatio-temporal knowledge graph with persistent segment-level contextual memory for question answering over long egocentric navigation videos. It improves Top-1 retrieval accuracy from 58% to 67% and Recall@1 from 34.50% to 40.55% on the WalkieKnowledgeT+ benchmark, outperforming methods including Gemini 2.5 Pro and Qwen 3.5+.

This paper proposes ParkingTransformer, a novel framework that leverages multi-view perception and scene understanding capability of Large Language Models (LLMs) for end-to-end autonomous parking. By combining trajectory queries with LLMs implicit state features, it outputs planning trajectories directly, eliminating dense BEV representations. It introduces 3D positional encoding, a fixed-window streaming mechanism, and a coarse-to-fine decoding strategy. Experiments on CARLA simulator achieve a driving score of 61.32, and real-world experiments show an average success rate of 88.70%.