If you're one of the many CarPlay users who prefers Google Maps for navigation, Google has begun rolling out a small but meaningful update for the iOS version of Maps. Now you can use the company's conversational driving experience, Gemini in navigation, through CarPlay. With Maps open, you'll see a new Gemini icon toward the bottom right of the interface. Just tap the bottom to start asking questions of the company's chatbot. "It's a conversational way to find stops on-the-go, report traffic disruptions, get information about places, and more," Google explains, adding Gemini is particularly adept at answering multi-part questions like "what park is famous for gorgeous views overlooking the city?"

OpenAI has'work to do to rebuild trust' in Australia, executive tells AI inquiry Mon 5 Oct 2026 20.53 EDTLast modified on Tue 6 Oct 2026 03.33 EDT Accept'bad things' in return for benefits of AI, says Sam Altman "We've just shared information about how the training works, and then discussed what its implications would be with respect to our capacity to operate here," he said. Anthropic suggests current Australian laws mean AI companies need to obtain a licence for the use of every piece of content on the internet as training material. Bleich said this was "technically impossible". Anthropic was keen to explore what the Australian government wants to do to incentivise training in Australia, and Bleich said Australia could shape frontier models if they were trained here, otherwise it would ultimately rely on models trained elsewhere. The multibillion-dollar company notified the government via an email to a public-facing address three months after the breach.

Alongside a new soundbar and headphones, Sonos is announcing an updated app and some major changes to the underlying OS that powers all of its products. Forgive me for immediately delving into the recent past, but this is likely the biggest software update the company has announced since its absolutely disastrous app update in 2024. Without rehashing too much, Sonos rushed the app update out to enable functionality for its new Ace headphones, but it was missing a ton of key features and was extremely buggy. The cascading effects of this resulted in the dismissal of CEO Patrick Spence and a 2025 where the company didn't release any new products. We have to imagine that Sonos learned enough from the 2024 debacle to make sure things are locked up for the launch of Sonos 27 -- yup, the company is giving its OS annual version numbers going forward.

Tabitha Britt is an award-winning freelance journalist, editor, and SEO/AEO strategist. Aside from reviewing dating apps and sex toys for Mashable, Tabitha is also the founding editor-in-chief of DO YOU ENDO -- a digital magazine by individuals with endometriosis, for individuals with endometriosis. She has a Master's degree in Creative Publishing and Critical Journalism from The New School for Social Research and is a grad of Sextech School. You can find more of her work in various online publications, including,, and . All products featured here are independently selected by our editors and writers.

The real world is messy and unstructured. Uncovering critical information often requires active, goal-driven exploration. It remains to be seen whether VisionLanguage Models (VLMs), which recently emerged as a popular zero-shot tool in many difficult tasks, can operate effectively in such conditions. In this paper, we answer this question by introducing FlySearch, a 3D, outdoor, photorealistic environment for searching and navigating to objects in complex scenes. We define three sets of scenarios with varying difficulty and observe that state-of-the-art VLMs cannot reliably solve even the simplest exploration tasks, with the gap to human performance increasing as the tasks get harder. We identify a set of central causes, ranging from vision hallucination, through context misunderstanding, to task planning failures, and we show that some of them can be addressed by finetuning. We publicly release the benchmark, scenarios, and the underlying codebase.

Vision-and-Language Navigation (VLN) is a core task where embodied agents leverage their spatial mobility to navigate in 3D environments toward designated destinations based on natural language instructions. Recently, video-language large models (Video-VLMs) with strong generalization capabilities and rich commonsense knowledge have shown remarkable performance when applied to VLN tasks. However, these models still encounter the following challenges when applied to real-world 3D navigation: 1) Insufficient understanding of 3D geometry and spatial semantics; 2) Limited capacity for large-scale exploration and long-term environmental memory; 3) Poor adaptability to dynamic and changing environments.To address these limitations, we propose Dynam3D, a dynamic layered 3D representation model that leverages language-aligned, generalizable, and hierarchical 3D representations as visual input to train 3D-VLM in navigation action prediction. Given posed RGB-D images, our Dynam3D projects 2D CLIP features into 3D space and constructs multi-level 3D patch-instance-zone representations for 3D geometric and semantic understanding with a dynamic and layer-wise update strategy. Our Dynam3D is capable of online encoding and localization of 3D instances, and dynamically updates them in changing environments to provide large-scale exploration and long-term memory capabilities for navigation.

The Object Goal Navigation (ObjectNav) task challenges agents to locate a specified object in an unseen environment by imagining unobserved regions of the scene. Prior approaches rely on deterministic and discriminative models to complete semantic maps, overlooking the inherent uncertainty in indoor layouts and limiting their ability to generalize to unseen environments. In this work, we propose GOAL, a generative flow-based framework that models the semantic distribution of indoor environments by bridging observed regions with LLM-enriched fullscene semantic maps. During training, spatial priors inferred from large language models (LLMs) are encoded as two-dimensional Gaussian fields and injected into target maps, distilling rich contextual knowledge into the flow model and enabling more generalizable completions. Extensive experiments demonstrate that GOAL achieves state-of-the-art performance on MP3D and Gibson, and shows strong generalization in transfer settings to HM3D. Codes and pretrained models are available at https://github.com/Badi-Li/GOAL.

Embodied agents are expected to perform object navigation in dynamic, open-world environments. However, existing approaches typically rely on static trajectories and a fixed set of object categories during training, overlooking the real-world requirement for continual adaptation to evolving scenarios. To facilitate related studies, we introduce the continual object navigation benchmark, which requires agents to acquire navigation skills for new object categories while avoiding catastrophic forgetting of previously learned knowledge. To tackle this challenge, we propose C-Nav, a continual visual navigation framework that integrates two key innovations: (1) A dual-path anti-forgetting mechanism, which comprises feature distillation that aligns multi-modal inputs into a consistent representation space to ensure representation consistency, and feature replay that retains temporal features within the action decoder to ensure policy consistency.

Unmanned Aerial Vehicles (UAVs) are evolving into language-interactive platforms, enabling more intuitive forms of human-drone interaction. While prior works have primarily focused on high-level planning and long-horizon navigation, we shift attention to language-guided fine-grained trajectory control, where UAVs execute short-range, reactive flight behaviors in response to language instructions. We formalize this problem as the Flying-on-a-Word (Flow) task and introduce UAV imitation learning as an effective approach. In this framework, UAVs learn fine-grained control policies by mimicking eUAxpert pilotVtrajectoriesFlopaired withwatomic Fly around the tree ahead Land on the left side of carlanguage instructions. To support this paradigm, we present UAV-Flow, the firstreal-world benchmark for language-conditioned, fine-grained UAV control.

Multimodal Large Language Models (MLLMs) have demonstrated strong generalization in vision-language tasks, yet their ability to understand and act within embodied environments remains underexplored. We present NavBench, a benchmark to evaluate the embodied navigation capabilities of MLLMs under zero-shot settings. NavBench consists of two components: (1) navigation comprehension, assessed through three cognitively grounded tasks including global instruction alignment, temporal progress estimation, and local observation-action reasoning, covering 3,200 question-answer pairs; and (2) step-by-step execution in 432 episodes across 72 indoor scenes, stratified by spatial, cognitive, and execution complexity. To support real-world deployment, we introduce a pipeline that converts MLLMs' outputs into robotic actions. We evaluate both proprietary and open-source models, finding that GPT-4o performs well across tasks, while lighter open-source models succeed in simpler cases. Results also show that models with higher comprehension scores tend to achieve better execution performance. Providing map-based context improves decision accuracy, especially in medium-difficulty scenarios. However, most models struggle with temporal understanding, particularly in estimating progress during navigation, which may pose a key challenge.