navigation
30 articles about navigation in AI news
BeliefDiffusion Uses Diffusion Models for Robot Navigation in Partially
BeliefDiffusion combines diffusion models with MPC for robot navigation in partially observable environments, outperforming model-free RL and generative baselines in synthetic maps.
Stanford's EgoNav Trains Robot Navigation on 5 Hours of Human Video, Enables Zero-Shot Control of Unitree G1
Stanford's EgoNav system uses a 5-hour egocentric video walk of campus to train a diffusion model that enables zero-shot navigation for a Unitree G1 humanoid robot, eliminating the need for robot-specific training data.
QAsk-Nav Benchmark Enables Separate Scoring of Navigation and Dialogue for Collaborative AI Agents
A new benchmark called QAsk-Nav enables separate evaluation of navigation and question-asking for collaborative embodied AI agents. The accompanying Light-CoNav model outperforms state-of-the-art methods while being significantly more efficient.
How to Cut Claude Code's Token Costs 32% by Fixing Its Navigation Problem
Claude Code agents waste tokens on grep-style navigation. A new open-source tool gives them IDE-like navigation, cutting costs 32% and doubling efficiency.
Google Maps Gets an AI Brain: How Gemini Transforms Navigation from Directions to Dialogue
Google is fundamentally reshaping Maps by integrating its Gemini AI, launching 'Ask Maps' for conversational discovery and 'Immersive Navigation' for a complete visual and data-driven route overhaul. This represents a shift from static maps to intelligent, proactive travel assistants.
Doby Cuts Claude Code Navigation Tokens by 95% with Spec-First Workflow
A spec-first fix workflow that slashes navigation tokens 95% and enforces plan docs as source of truth before code changes.
Alibaba's Qwen-RobotNav Unifies Robot Navigation in One 2B-8B Model
Alibaba's Qwen-RobotNav unifies VLN, ObjectNav, tracking, and autonomous driving in a 2B-8B model, deploying zero-shot to quadruped robots via a configurable observation protocol.
Wikipedia Navigation Challenge Exposes Critical Gaps in AI Planning Abilities
Researchers introduce LLM-WikiRace, a benchmark testing how well AI models navigate Wikipedia links between concepts. While top models like Gemini-3 show superhuman performance on easy tasks, success rates plummet to just 23% on hard challenges, revealing fundamental limitations in long-term planning.
Mistral AI Ships Robostral Navigate for Physical AI Push
Mistral AI released Robostral Navigate, a robotics navigation model for physical AI, targeting European industrial customers. The company is negotiating a €3B funding round at a €20B valuation.
Dimos OS Launches as Open-Source Robot OS with AI Agent MCP Access
Dimos OS is a new open-source operating system for robots that lets developers write Python modules and gives AI agents direct control via MCP. It includes a full navigation stack and supports hardware like Unitree G1 and DJI drones.
New Framework Reveals LLM GUI Agents Don't Navigate Like Humans
Researchers introduced a trace-level framework to compare human and GUI-agent behavior in a production search system. While the agent matched human success rates and query alignment, its navigation was systematically more search-centric and less exploratory. This reveals a critical gap in using agents as user proxies.
Google DeepMind Unveils Gemini-Powered Browser That Generates Websites in Real-Time
Google DeepMind has demonstrated a browser prototype powered by Gemini 3.1 Flash-Lite that generates complete HTML/CSS websites dynamically based on user prompts and navigation context, shifting from static page retrieval to on-demand interface generation.
Cursor Launches Instant Grep: Millisecond Local Search Across Millions of Files
Cursor has launched Instant Grep, a local search tool that performs millisecond-level regex searches across millions of files. The feature is integrated into the Cursor IDE, targeting developers needing fast, offline code navigation.
AgentComm-Bench Exposes Catastrophic Failure Modes in Cooperative Embodied AI Under Real-World Network Conditions
Researchers introduce AgentComm-Bench, a benchmark that stress-tests multi-agent embodied AI systems under six real-world network impairments. It reveals performance drops of over 96% in navigation and 85% in perception F1, highlighting a critical gap between lab evaluations and deployable systems.
Claude AI Gains Computer Control Feature: Opens Apps, Navigates Browser, Fills Spreadsheets
Anthropic's Claude AI can now be enabled to directly control a user's computer to perform tasks like opening applications, browser navigation, and spreadsheet work. This represents a significant shift from chat-based interaction to direct system automation.
MiRA Framework Boosts Gemma3-12B to 43% Success Rate on WebArena-Lite, Surpassing GPT-4 and WebRL
Researchers propose MiRA, a milestone-based RL framework that improves long-horizon planning in LLM agents. It boosts Gemma3-12B's web navigation success from 6.4% to 43%, outperforming GPT-4-Turbo (17.6%) and the previous SOTA WebRL (38.4%).
InterDeepResearch: A New Framework for Human-Agent Collaborative Information Seeking
Researchers propose InterDeepResearch, an interactive system that enables human collaboration with LLM-powered research agents. It addresses limitations of autonomous systems by improving observability, steerability, and context navigation for complex information tasks.
Bridging the StarCraft Gap: New AI Benchmark Makes Strategy Research Accessible
Researchers introduce Two-Bridge Map Suite, a lightweight StarCraft II benchmark that isolates tactical skills without full-game complexity. This open-source tool enables reinforcement learning experiments on realistic budgets by focusing on navigation and combat mechanics.
FDM-1: The AI That Learned to Use Computers by Watching 11 Million Hours of Screen Recordings
Standard Intelligence has unveiled FDM-1, an AI system trained on 11 million hours of screen recordings that can perform complex computer tasks like CAD design, web navigation, and even simulated driving with minimal fine-tuning.
Switchboard's Grid View Gives You Bird's-Eye Control of Claude Code Sessions
Switchboard v0.0.16 adds a grid view that shows all your Claude Code sessions at once with live terminal previews, status indicators, and quick navigation.
Dyna-2 World-Action Model Trained on 1M Hours Video
Dyna-2, trained on 1M+ hours of egocentric video, jointly predicts future video and actions. Claims new scaling laws but no benchmarks or technical details released.
Fix 80% of A11y Bugs Automatically
Install axe-core MCP in Claude Code to auto-detect and patch 80% of WCAG violations, replicating Evinced's agentic a11y workflow.
NVIDIA's Molt: 9.2K-Line RL Framework Scales to 1T-Parameter MoE Models
NVIDIA released Molt, a 9.2K-line PyTorch RL framework scaling to 1T-parameter MoE models via vLLM, targeting agentic tasks with fully-async rollout.
Alibaba Releases RynnBrain 1.1 Embodied AI Models at 2B-122B Scales
Alibaba released RynnBrain 1.1 on Hugging Face with 2B, 9B, and 122B-A10B MoE models for robot manipulation, but disclosed no benchmarks.
ActiveVision Benchmark: Humans 96.1%, Best AI 10.6%
ActiveVision benchmark: humans 96.1%, best AI 10.6%. The 85.5-point gap reveals fundamental limits in iterative visual reasoning for current models.
Microsoft Fara1.5-27B Open-Source Agent Scores 72.3% on Web Tasks
Microsoft released Fara1.5-27B, a vision-only web browsing agent scoring 72.3% on Online-Mind2Web, open-source on Hugging Face.
Nvidia, Hugging Face Open-Source Robot Models to Democratize Physical AI
Nvidia and Hugging Face open-sourced robot models to democratize physical AI, providing pre-trained models and simulation tools on the Hugging Face hub.
LLM agents fail nonlinearly as tasks lengthen, 27-paper synthesis finds
27-paper synthesis finds LLM agent failures compound nonlinearly with task length. Six failure clusters identified across 19 benchmarks.
AWS launches MCP server for its open-data registry
AWS launched an MCP server for its Registry of Open Data, giving AI agents natural-language access to 170+ public datasets.
Tencent Hunyuan Hy3: 295B MoE Hits 90% Agent Task Resolution
Tencent launched Hunyuan Hy3, a 295B MoE model with 21B active parameters, claiming 90% agent task resolution, surpassing DeepSeek V4 Pro and Qwen 3.7 Max.