Artificial Anlys tested Apple’s ADM 3 Cloud in its arena, where the model ranked 42nd on AA-Image-Editing v2.1 and 85th on AA-Image-T2I v2.1. The supplied post does not describe the test methodology, prompts, scoring process, evaluation setup, or number of evaluated systems.
Google announced a global rollout of the standalone Gemini app for Apple CarPlay, bringing Gemini Live hands-free conversations to supported iOS users. The app supports interruptions, topic changes and pausing, integrates with Gmail and Google Calendar for email summaries and schedule queries, and syncs CarPlay conversations with the Gemini iPhone app. Users need the latest Gemini app, microphone permissions and prior Gemini Live use; rollout is phased, with some personalized features subject to regional and age restrictions.
Chinese equipment supplier Quick Intelligent Equipment said at its Nov. 10 third-quarter results briefing that it had participated in Apple AI-glasses and new smart-earphone projects. A Securities Times report carried by iFeng said the supplier expects the initiatives to drive demand for its precision-welding, assembly and machine-vision inspection tools. The disclosure does not establish final specifications, earphone cameras, AirPods branding, mass production, production volumes, contract values or launch dates. A separate Chinese report described an unnamed long-term Apple supplier as working on AI glasses and camera-equipped AirPods, but the supplier statement reviewed by TechRepublic did not confirm those cameras.
Apple Intelligence in the Home app can generate text descriptions and summaries of recordings from compatible HomeKit Secure Video cameras, then let users find clips with natural-language searches. CNET and 9to5Mac report that access requires an Apple Intelligence-capable iPhone or iPad, a compatible camera and home hub, plus an eligible iCloud+ plan; 9to5Mac says the 2TB, 6TB and 12TB tiers support up to one, two and five cameras respectively.
Android Authority conducted hands-on testing of flagship smartphones powered by Qualcomm’s Snapdragon 8 Elite Extreme Gen 6, Google’s Tensor G6, MediaTek’s Dimensity 9600 Pro, and Apple’s A20 Pro. The comparison covers CPU, GPU, and AI technologies, but the supplied report excerpt does not provide the test results or detailed AI benchmarks. Samsung’s Exynos 2700 was excluded because it was expected to debut with the Galaxy S27 series.
Apple plans to limit usage of several Apple Intelligence features backed by server-side models, with users potentially able to obtain more capacity through iCloud+ subscriptions, IT之家 reported. The affected features include Siri AI, Image Playground, smart photo editing tools, AFM 3 Cloud models in Shortcuts, and potentially third-party apps using Apple Foundation Models; Logic Pro already offers an additional AI usage pack for US$49.99.
A leaker’s post, cited by IT之家 and AIbase, says Apple added disabled Health-related feature flags to the first iOS 27.3 developer-preview beta. The “IF / HealthSearch” flag is described as covering 92 HealthKit data types, while “IF / HealthSnippets” covers visual presentations for health statistics, activity rings, workouts and sleep summaries. The reports cite Apple support documentation saying Siri AI does not currently support reading Health app data.
Huxiu AI argues that simply adding a large language model is insufficient to justify premium pricing for AI glasses and earphones as phones and older devices gain translation, transcription, and question-answering features through software updates. It examines whether these products can reduce friction in complete tasks such as real-time conversations, meeting workflows, and task execution rather than merely serve as AI access points. Counterpoint data cited by Huxiu AI show global AI-glasses shipments rose 263% year over year in the first half of 2026, with display-free products accounting for 96% and Meta holding 94% of that segment's shipment share.
9to5Mac Daily says Apple launched $50 “intelligence usage packs” for selected AI features. The supplied report is a podcast roundup and does not independently verify the launch or explain which features are included.
Apple launched the 2027 Swift Student Challenge for student developers worldwide, with submissions open from February 8 through February 28, 2027. Entries built with Xcode or Swift Playgrounds must be demonstrated in three minutes, and the contest expands beyond iOS to Apple platforms including macOS, watchOS, tvOS, and visionOS. Apple also plans tutorials and “Meet Apple” events covering SwiftUI, AI, machine learning, spatial computing, and coding intelligence, while some distinguished winners will be invited to WWDC 2027.
9to5Mac, citing leaker Aaron Perris, reported that Apple is developing a Mac app called Harbor to run open-weight models locally and potentially serve private AI inference. Product materials described in the reports say Harbor can manage models and cluster multiple Thunderbolt 5 Macs by pooling unified memory, although these capabilities have not been independently tested. IT之家 and other outlets reported that Claris is testing the app in an early-access program and that its App Store listing was briefly removed, leaving public availability unsettled.
TechRadar reports that Disney Imagineering is preparing the opt-in Imagineering Labs iOS app so Disneyland guests can test unfinished experiences and provide development feedback. One experiment is the AI-powered Disneyland Enchanted Guide, which handles natural-language and image queries through an iPhone, AirPods, or compatible Meta Ray-Ban glasses and is initially planned for Magic Key holders.
Tom's Hardware benchmarked an M5 Ultra Mac Studio with a 36-core CPU, 80-core GPU and up to 512GB of unified memory against a single Dell Pro Precision GB10 system, which it described as nearly identical to an Nvidia DGX Spark. Using llama.cpp with Qwen 3.8 27B and its native multi-token prediction, the Mac completed the Artificial Analysis AgentPerf-Local test in just under half the time of the GB10 system. The report also cited 1.2 TB/s of memory bandwidth for the Mac versus 273 GB/s for the GB10. For DeepSeek V4.1 Flash 1017, however, the Mac's llama.cpp setup took 10–25 minutes for the longest tested contexts, while a two-Spark cluster using DSpark speculative decoding and TensorFold took under five minutes. The report said Mac Studio lead times stretched into February 2027.
Osaurus co-founder and CEO Terence Pae discussed running AI locally on a Mac rather than sending everything to the cloud in an Apple @ Work podcast episode. 9to5Mac describes Osaurus as a free, first-party AI harness for the Mac.
Quick Intelligent said at its third-quarter 2026 earnings briefing that it has participated in Apple's AI-glasses and new smart-earbud projects. The company said the work would help drive its precision soldering and assembly and machine-vision inspection businesses. Earlier reports cited by IT之家 estimated that Apple's first AI glasses could launch as early as 2027 and described the product as non-AR without a lens-based display system.
Asahi Linux announced that it is migrating its code repositories from the permissive MIT license to GPL v2 or later, with future code using strong copyleft terms. The project said earlier releases remain usable under their original licenses, while new contributions and later versions follow each repository’s current terms. It also said some in-development code was copied or allegedly rewritten with AI to obscure its origin, though the reported misconduct was not independently verified.
An IT Home report citing an X post from @itspdfu says Apple introduced a Skimmer pre-module in the first iOS and macOS 27.3 developer preview betas. The source claims that about 2% of English Siri AI requests, including commands to set timers, play music and make calls, are routed to local models. According to the report, Skimmer uses regular-expression matching against preset tool flows, allowing Siri to execute matching commands without entering more complex AI inference.
William L. Tong and coauthors report that state space models use constant memory and linear computation for sequence modeling but lag behind Transformers on in-context learning and precise retrieval. Their theory and experiments show that gating causes state space models to first learn an in-weights memorization solution, delaying or preventing convergence to correct in-context learning even when architecture or memory capacity is not the fundamental constraint. The researchers also find that gating often improves generalization to longer sequence lengths.
A study accepted at the NeurIPS 2026 Can We Trust AI Evaluation? Workshop compares LLM judges with human preferences on three long-form writing tasks. Across four text generators, judge quality varied by task and criteria, correlated weakly with MMLU performance at Pearson’s ρ ≤ 0.2, and improved with paired comparisons; the strongest LLMs did not consistently reproduce human judgments.
A commentary draws on The Trap by Aeer Zarah and Maurice Stuck to argue that algorithms, targeted pricing, and consumer monitoring can reshape competition. Applying that lens to large technology companies, the author questions how markets will react to heavy AI capital spending when its revenue path remains unclear. Google is presented as especially exposed because search is central to its traffic and advertising model, while the author notes that signs of users shifting from search to AI are not yet clear.
Apple announced a long-term investment in seven facilities across Turkey: five Apple Developer Academies, an Istanbul Developer Center, and an Ankara manufacturing training facility. The first academy and the developer center are planned for 2027, while the Ankara facility is slated for 2028; programs will cover app development, AI, machine learning, automation, predictive maintenance, and robotics.
Apple has begun selling one-time $49.99 Intelligence Usage Packs to subscribers of Apple Creator Studio, with Logic Pro offering the first purchase option. The packs are consumed only after the monthly allowance is exhausted, roll over monthly, and expire 12 months after purchase; they do not support Family Sharing. Reports say the additional quota is not disclosed and varies with the feature, request complexity, system requirements, and system policies.
A V2EX user reported that a new Claude Pro account activated with an Apple gift card passed KYC using a Chinese passport and remained active for more than one day. Other commenters reported conflicting outcomes: one Chinese-passport account survived two weeks, while another was banned immediately after its passport was uploaded. A separate commenter claimed that an account passed KYC through Google Play and consumed more than 300 million tokens over three days, though the thread provided no independent verification.
9to5Mac Overtime 089 debates Apple Watch’s Siri Recap and Live Rewind features, including whether the watch is always listening and whether users should be concerned. The episode also covers iOS 27.3 beta changes: Safari gains adaptive dark mode and vertical tabs, StandBy is redesigned, and the Health app returns. It also addresses reader comments about a touchscreen MacBook Pro and the iPad mini’s missing 120Hz display.
Apple CEO John Ternus reportedly urged the software team to accelerate Apple’s AI roadmap beyond limited Siri improvements and toward more agentic capabilities. The reports, citing Bloomberg’s Mark Gurman, say the directive may involve a revised iOS 28 development schedule, while Apple has separately added developer tools for iOS 27.3, including expanded App Intents, a modular Skills framework and MCP support in iWork and Creator Studio. 9to5Mac notes that faster agentic development could create significant reliability, privacy and execution risks, and the internal directive and schedule change have not been independently confirmed.
DoNews reported that Apple Intelligence completed an additional-device filing covering iPhone, iPad, and Mac. The report did not provide filing details, identify the relevant authority, or confirm the filing through an official announcement.
IT之家, citing code researcher pdfu, reported that Apple is developing a personal memory system for Siri in the iOS 27.3 beta. The code reportedly includes a “Memory Graph” for organizing relationships, frequently visited places, daily habits, health conditions, exercise, and factors that motivate or hinder users. “Memory Atoms” are designed to record facts, opinions, intentions, and events linked to people, places, times, and topics. The exposed code is unfinished: it reportedly lacks the processing flow for creating Memory Atoms. It does not disclose when the system may launch or how users would manage and control their data.
cnBeta, citing 网信上海, reported that Apple Intelligence completed the required filing for added iPad and Mac scenarios in China. The iOS 27.3 developer beta also opened selected Apple Intelligence features to China-based developers, including Live Translation, Writing Tools, Visual Intelligence, Photo Eraser and Genmoji. The report says Apple and Alibaba integrated Qwen into Siri, Writing Tools and Visual Intelligence.
Apple began selling HomeView and the fourth-generation Apple TV 4K on November 13, 2026, with both devices priced at $299. HomeView is a 7-inch smart-home display with an A18 chip, a 12MP Center Stage camera and Siri AI, while the new Apple TV 4K has an A19 chip, 128GB of storage, Wi-Fi 7 and support for four HomePod speakers. Apple also opened orders for the first three Apple Home Collection accessories: an LG video doorbell and indoor and outdoor security cameras.
Apple Intelligence’s registered generative AI service in China has added two usage scenarios covering Apple tablets and Apple laptops, according to 36氪快讯 citing 网信上海. The filing was completed as required under China’s Interim Measures for the Administration of Generative AI Services; the report does not specify availability dates or supported devices.
What problem it solves
Computer-use projects need environments where agents can operate applications, run code, and interact with graphical interfaces within the same task. They also need repeatable ways to create tasks, evaluate agents, export trajectories, and generate data for training and research.
Key features
Shared desktop Spaces — Cua Spaces gives agents full desktops hosted on computers o…
What problem it solves
Coding agents need a consistent way to inspect, control, debug, and verify applications across mobile, web, TV, and desktop targets. They also need repeatable workflows and reviewable evidence across local devices, simulators, CI, and remote device clouds.
Key features
Inspect and control app interfaces — Agents can read accessibility snapshots, act through refs and selector…
What problem it solves
Hermès Watch Faces require an enabling process on supported Apple Watch devices. Users need instructions for installing or building the app, as well as guidance for preserving data and restoring original settings after updates or removal.
Key features
Install the IPA — The README provides a release-based installation flow using the **Watch** app after sideloading the IPA wit…
What problem it solves
Running large language and vision-language models can require substantial setup and may depend on the available hardware. The project aims to make inference accessible across a wide range of local and cloud hardware with minimal setup.
Key features
Minimal-setup inference — The project provides installation options for common platforms, including scripts, Docker, pre-built b…
What problem it solves
Serving language models requires reading checkpoints, tokenizing requests, and running accelerated kernels while managing a local model server. TensorFold addresses these requirements with a native binary that can run on Apple Silicon and NVIDIA GPUs without requiring Python or MLX for serving.
Key features
Native Metal and CUDA inference — The Zig-based engine executes mode…
What problem it solves
Installing sideloaded apps and importing pairing files can require confusing device setup, certificate handling, and troubleshooting steps. The project provides a desktop workflow for installing IPA files, managing pairing files, and addressing common setup errors.
Key features
App installation — Install SideStore, optionally with LiveContainer, or import any IPA file.
Sourc…
What problem it solves
Sponsor segments, introductions, and advertisements interrupt videos on YouTube for TV clients. A self-hosted application can automate this handling while supporting both network discovery and manual device setup.
Key features
Automatic segment skipping — The application connects to a YouTube for TV app and automatically skips segments such as sponsors and introductions.
Sou…
What problem it solves
Interactive multimodal and speech systems often require generation to span heterogeneous stages with different compute patterns, dependency structures, and resource needs. Serving these pipelines requires coordinated stage lifecycles, inter-stage transport, model integrations, and a unified API surface.
Key features
Coordinated multi-stage runtime — Models generation as coor…
What problem it solves
Users need a command-line workflow to search for apps across Apple platforms and obtain .ipa or macOS .pkg packages from the App Store. The workflow also needs to support App Store authentication, purchases, version metadata, and automation.
Key features
Search across Apple platforms — Search the App Store for iOS, iPadOS, tvOS, watchOS, visionOS, and macOS applications.
Sou…
What problem it solves
Developers need a coding agent that can assist with development work while remaining inspectable and open source. The project addresses the need for configurable agents with different access levels, a straightforward installation path across platforms, and a way to explore codebases without modifying them.
Key features
Two built-in agents with different access levels — A def…
What problem it solves
Developers need to run AI models inside applications while handling model downloads, inference configuration, and chat sessions. The project presents local inference with data privacy and no inference costs as its deployment model.
Key features
High-level application API — The engine exposes a simple API for creating an engine, loading a model, opening a chat session, and re…
What problem it solves
PS5 executables need to be converted into native executables for Linux and Windows, with suitable system libraries available for dynamic linking. The project addresses this porting work without emulation or a separate runtime process.
Key features
Native executable relinking — The relinker converts executables into the native format of the target operating system and provide…
What problem it solves
Time-series forecasting requires models that can forecast univariate and multivariate series across different context lengths. Forecasters also need to incorporate optional covariates and produce quantile forecasts without requiring per-task tuning.
Key features
Univariate and multivariate forecasting — Forecast individual time series or multiple target variates from NumPy a…
What problem it solves
Running AI models locally can require integrating multiple inference engines, model formats, hardware accelerators, and API surfaces. LocalAI addresses this with a small core, on-demand backends, and compatible APIs that work across supported hardware without requiring a GPU.
Key features
Composable inference backends — Backends are packaged separately and pulled only when a…