U.S. autonomous ground vehicles are operating in Ukraine
Forterra has deployed more than 100 self-driving ATVs in Ukrainian conflict zones.
News
Short posts prepared from approved sources with clear links back to the original articles.
RSS feedForterra has deployed more than 100 self-driving ATVs in Ukrainian conflict zones.
A new Hugging Face post describes running AI workloads on any cloud while storing data on Hugging Face with zero-egress storage through SkyPilot.
An InfoQ presentation by Aaron Erickson explains how NVIDIA designs AI agent hierarchies and balances deterministic tools with agentic discovery.
It is working with Hugging Face to bring new models and frameworks to LeRobot, targeting shared tools for physical AI and robotics research.
Ars Technica examines how progress in AI could make autonomous robot workers more practical in workplaces and potentially homes.
Agentic workflows plan, loop, and burst differently from single model calls, changing the GPU and infrastructure requirements underneath.
MIT News reports that a USAF cadet and Lincoln Laboratory researcher found AI chatbots can help nontechnical service members build useful software applications.
GitHub’s Q1 2026 Innovation Graph update says global developer communities are growing faster and collaboration is reaching new highs across many economies.
A British space startup launched a longevity lab into orbit to generate data for AI models of proteins linked to age-related disease.
Microsoft is replacing some OpenAI and Anthropic model usage in Copilot products with its own MAI models to reduce external model costs.
AWS detailed best practices for multi-dataset Topics in Amazon Quick Chat, aimed at natural-language exploration across related datasets.
Hugging Face and Microsoft are connecting Hugging Face models with Foundry managed compute, giving developers another managed path for running open models.
Savi raised $7 million and is launching an iPhone and Android app aimed at protecting consumers from realistic AI-enabled scams, including fake ransom calls.
Anthropic is rolling out Claude Cowork on mobile and web, starting with Max subscribers before broader availability in coming weeks.
DeepSeek is exploring its own chips as U.S. export controls pressure Chinese AI companies to reduce outside dependencies.
AWS added release-management capabilities to its DevOps Agent so it can assess code changes and autonomously test software before production.
IT leaders need durable AI architecture foundations as organizations expand toward agentic systems and fast-changing use cases.
Its Vera CPUs are aimed at agentic AI workloads where tool calling, code execution, and response-time-critical tasks put CPUs back on the critical path.
RunPod’s State of AI report draws on production data from more than 500,000 developers to show what AI workloads are actually running.
Researchers are starting new work on how AI is changing jobs, productivity, and economic outcomes.
Solos has introduced smart glasses with a clip-on cover for the cameras, highlighting the privacy tension around AI wearables.
Microsoft’s June Foundry update makes Claude generally available and adds new agent features including autopilot agents, expanded Toolboxes and Routines, and Agent Optimizer preview work.
OpenAI published a customer story on Australian Payments Plus using ChatGPT Enterprise and Codex to speed work across complex payments workflows.
Google expanded managed agents in the Gemini API with background tasks, remote MCP support, and additional features for developer workflows.
OpenRouter ran 1,730 visual reasoning questions across five models and found that lowering image detail can reduce accuracy without reliably cutting costs.
NVIDIA’s Kyber NVL144 AI server rack has been delayed by more than a year because of circuit-board manufacturing problems.
ByteDance and Alibaba are shutting down custom AI companion features in response to new Chinese regulations, The Decoder reports.
Some wealthy families are paying high tuition for schools that put AI tutors and experimental education models at the center.
TechCrunch is tracking large 2026 tech layoffs where employers cited AI, including Microsoft’s latest cuts and broader industry automation claims.
The Paris startup hub is preparing a new F/ai accelerator batch aimed at helping early AI startups move from product to revenue.
A new Hugging Face post explains how Photoroom built PRX training data with mixed sources, VLM re-captioning, streamable shards, filtering, and deduplication.
AWS detailed an integration that streams SageMaker AI benchmark and inference recommendation results into a managed MLflow app in real time.
GPT-4 led a capability index for about a year, while recent leading models have averaged only about seven weeks at the top.
Tencent released Hy3, a 295B-parameter open model that activates 21B parameters at a time and is pitched as matching much larger systems.
AWS published a SageMaker HyperPod pipeline for multi-turn reinforcement learning using Amazon Nova Forge and event-driven training data workflows.
AWS presented a multi-step image-redaction pipeline that combines Amazon Nova reasoning, Segment Anything, SageMaker, and Textract.
Cloudflare and AWS have both implemented x402 stablecoin micropayments at edge networks, reviving HTTP 402 for agent-to-service payments.
Zhipu AI is adding GLM-5.2 to its ZCode development environment and pitching low-cost coding-agent workflows with long context.
Countries are investing in AI capabilities to advance economic priorities, protect data, and build domestic technology capacity.
Hugging Face updated its Kernels project with trusted publishers, kernel signing, revamped CLIs, and broader framework and backend coverage.