AI Research

Envirotech Vehicles closed its merger with Azio AI on July 2, 2026, ahead of schedule, accelerating its pivot to AI data centers and enterprise GPU compute. The combined company targets the $487 billion AI infrastructure…

Former Tesla scientist Rémi Cadene launched UMA, a Paris-based startup building a lightweight, AI-powered humanoid robot called Northstar for manufacturing, logistics, and homes. The company has secured backing from top …

DeepSeek is developing its own AI inference chip to reduce reliance on Nvidia and Huawei hardware, according to a Reuters report citing three sources. The move aims to control serving costs and latency amid U.S. export c…

DeepSeek is developing its own AI inference chip to reduce dependence on Nvidia and Huawei, according to a Reuters report citing three sources. The Chinese lab has been in talks with chip-design and foundry partners for …

Cryptography expert Filippo Valsorda argues that hybrid signatures combining classical and post-quantum schemes are often not worthwhile due to hidden costs, despite their apparent simplicity. He states that pure ML-DSA …

Mulot, an open-source AI agent framework, uses open-weights LLMs like GLM-5.2 to drive a headless Chromium browser with a Burp-style toolkit, achieving 87% on OverTheWire Natas and 73% on Root-Me Web-Server challenges. T…

UMA, a Physical AI company, unveiled its first humanoid robot design and Real-Time Learning architecture at Machina Summit, enabling robots to learn skills through demonstration rather than manual programming. The compan…

AI search now favors brands with strong reputations, consistent messaging, and trusted third-party validation, requiring marketers to shift focus from visibility to being chosen. A single AI prompt compresses the user jo…

Anthropic published research on July 6, 2026, revealing that its Claude AI model contains a small set of internal neural patterns called J-space, identified using a Jacobian-based technique named J-lens. Editing these pa…

A new interactive tool called AI Power Atlas has been released to help policymakers analyze countries' AI strengths, weaknesses, and dependencies across six levers of power including compute, model capability, talent, da…

Ant Group's embodied AI unit, Robbyant, launched LingBot-Depth 2.0 and LingBot-Vision, claiming the models enable robots to accurately perceive transparent objects like glass and mirrors. The company says LingBot-Vision …

CNCF Ambassador Lin Sun of Solo.io wrote on July 7, 2026 that sandboxing alone is insufficient for running large-scale AI agent fleets, contrasting the Kubernetes SIG Apps project agent-sandbox with the newer agent-subst…

Scientists at National Taiwan University measured the transfer length of metal contacts in 2D-material transistors to be just 2-3 nanometers, far smaller than previous estimates. This finding suggests that contacts can b…

Meta researchers released SWE-Together, a benchmark built from 11,260 real developer-agent sessions that tests AI coding agents on collaborative performance under user corrections, not just one-shot task completion. Clau…

Researchers at Delft University of Technology have created Mycelium, a self-replicating seedbox for the Tribler BitTorrent client, as part of a broader effort to build a decentralized digital democracy. The project, fund…

Boston Dynamics vice president Matt Malchano says AI is enabling autonomous robots to perform complex tasks in workplaces and potentially homes, a shift from 15 years ago when robots could only navigate from point A to B…

A new Clovion AI study reveals that 62% of brand recommendations from AI assistants vanish after a single buyer question, based on 69,120 multi-turn conversations across Claude, ChatGPT, and Gemini. The research also fou…

China is considering new restrictions limiting foreign access to its most advanced AI models, treating them as strategic national assets. The move escalates the tech cold war with the US, which has already imposed export…

Bespoke Labs, a 40-person AI startup, raised $40M from insiders at Anthropic, OpenAI, and Meta to explore whether improved training environments can outperform larger models.

Correctover Research Group published the first formal conformance standard for AI agents, the Correctover Conformance Standard (CCS) v1.0, after auditing over 8,000 real API calls across multiple providers. The standard …