A developer’s AI-assisted analysis found a NoShake variant in Baidu Netdisk iOS that uses Core Location’s CLHeading to obtain device orientation instead of relying on conventional CoreMotion motion detection. cnBeta reported that the app could still open shake-triggered ads after motion and fitness access was disabled, although the analysis did not capture a live ad response or establish the parameters used online.
Anthropic announced upgrades for Claude across iPhone, iPad, Android, Chromebook, web and desktop. The iPhone app’s voice mode now supports Apple CarPlay, allowing drivers to ask Claude questions by voice while driving. Anthropic also announced upcoming pull-down search for iPhone and iPad, side-by-side and full-screen views for generated content, multi-account switching, and desktop-style controls on large-screen Android devices. The company said Claude now supports 33 languages and that voice recognition and response speed have improved. Pull-down search remains forthcoming.
Apple launched the 24-inch M6 iMac with a 10-core CPU and GPU, or a 12-core configuration for $1,699, while the base model starts at $1,499. The all-in-one retains its Liquid Retina LCD display, 12-megapixel Center Stage camera, and two Thunderbolt 4 ports, adding Wi-Fi 7, Bluetooth 6, 2.5-gigabit Ethernet, and twice the storage read and write speeds. Gizmodo said M6's neural accelerators make it capable of running some mid-range on-device AI models, but its performance comparison with the M4 iMac remains an estimate rather than a confirmed benchmark.
Wccftech, citing DigiTimes, reports that TSMC's A14 process is planned to enter mass production in the first half of 2028, following risk production in the second half of 2027. Fab 20 and Fab 25 are named as the primary production sites. ETNews, as cited by Wccftech, reports that A14 is intended to scale SRAM bit cells from 0.021 µm² to 0.017 µm². DigiTimes also reportedly estimates that TSMC's 2028 A14 capacity could be 7.5 times Intel's 14A capacity; this figure is attributed to the report.
The Wall Street Journal, citing TechInsights, reported that AI hyperscalers’ demand helped raise NAND flash prices to roughly five times their level a year earlier, while Meta, Microsoft and Amazon signed higher-priced long-term supply agreements with Kioxia. TechInsights estimates that Apple may spend about $25 billion this year on roughly the same flash volume it bought for $6 billion in 2025, as cloud operators’ purchasing now exceeds Apple’s. The Wall Street Journal also estimated that upgrading an iPhone 18 Pro from 256GB to 1TB costs consumers an extra $600 against an estimated $192 incremental chip cost, though the reported premium varies by calculation method.
TMTPost's Market Value Observer examines how Apple supply-chain companies including Lens Technology, Luxshare Precision, Goertek and Foxconn Industrial Internet are expanding into AI servers, liquid cooling, AI glasses and robotics. It cites Lens Technology's proposed $25 million cash acquisition of a 59.78% stake in PMG, which would indirectly control YuanShi Technology and strengthen its server-cabinet and liquid-cooling business. The analysis says these firms are reusing precision-manufacturing, thermal-management, optics and large-scale delivery capabilities built in consumer electronics.
According to 9to5Mac and cnBeta, U.S. District Judge Edward Davila set July 19, 2027, at 9 a.m. for the trial of Apple’s trade-secrets lawsuit against OpenAI, two former employees and io Products. The case-management order sets deadlines for document production, fact and expert discovery, motions and pretrial proceedings. The court has also granted Apple additional expedited discovery. Apple alleges that confidential information about its products and technologies was improperly accessed or used to advance OpenAI’s consumer-hardware development. OpenAI denies Apple’s allegations and says the former employees may have retained or been able to access Apple files without knowledge. Lead counsel are scheduled to meet on June 1 to discuss a potential settlement and other pretrial matters, but the report says this is a routine procedural step and does not establish that settlement talks are underway.
A federal district court in Illinois rejected Apple’s request to dismiss an Apple Photos class action, sending the dispute to a jury. Plaintiffs allege that the preinstalled Photos feature collected voiceprints and faceprints without adequate notice or explicit written consent. Apple disputes the plaintiffs’ claims, arguing that its numerical-string comparison does not constitute biometric scanning.
Amp introduced macOS Satellites, letting its coding agent automatically start a cloud Mac with Xcode and simulators to build, test, and run iOS or Mac apps while the main thread remains on a Linux orb. The beta includes a Mac screen stream, temporary instances reused for 20 minutes after a task, and pricing of $2 per hour for mac.medium or $4 per hour for mac.large. Amp demonstrated the Penny app on iPhone 17 Pro and iPhone Duo and an iPod app for Mac, but these capabilities are vendor claims.
Mattt announced AnySpeechRecognizer and AnyDecisionModel, two Swift packages for Apple platforms. AnySpeechRecognizer mirrors Apple’s speech-to-text API, supports selectable engines such as NVIDIA Parakeet and Alibaba Qwen3-ASR running locally through Core ML, and adds speaker diarization. AnyDecisionModel returns typed answers and probabilities for yes-or-no, multiple-choice, and rating-scale questions, using on-device models through MLX, Core AI, or llama.cpp, or decision models from Jev or Ollama. The packages can download open models from Hugging Face. Reported speed, memory, and word-error-rate results came from one machine and a few datasets; testing was limited to macOS 27, with no iPhone results.
A report by Dongdu Finance says Meta launched Muse in the United States and Canada on September 8. The personal agent can handle tasks such as comparing prices, placing and paying for orders, booking travel and restaurants, filling out forms, sending emails, arranging schedules and dealing with customer service. The report argues that by bypassing product interfaces, personal agents could weaken platforms’ control over traffic distribution and potentially reduce advertising, commission and bidding-based revenue. It contrasts the relatively open and fragmented US internet ecosystem, where transactions often occur on independent sites, with China’s more closed platform ecosystems that handle traffic and transactions internally.
Anthropic reported that an unpublished internal Claude research model autonomously identified vulnerabilities and wrote exploit code across Chrome, Safari, Android, Windows, and Linux, producing five runnable attack chains. The Chrome report reportedly yielded 62 critical or high-severity CVEs, while the Linux chain CVE-2026-98432 was validated in 53 minutes and reached a root shell on a server running a multipath TCP listener. Anthropic said the vulnerabilities were disclosed through coordinated disclosure and fixed by the relevant vendors.
Apple unveiled its first touchscreen MacBook Pro in 14-inch and 16-inch models, pairing the new PROLED XDR tandem-OLED displays with M5 Pro or M5 Max chips. The machines offer up to 128GB of unified memory, and Apple says the M5 Max configuration can run large language models locally; U.S. pricing starts at $2,899 for the 14-inch model and $3,399 for the 16-inch model.
Apple opened preorders for the 14-inch MacBook Pro with the M6 chip at RMB 15,999, according to IT之家 AI, with delivery expected as early as November 6. The machine has a 12-core CPU, 12-core GPU, and dual 16-core Neural Engines, while Apple says its on-device AI performance is up to 13 times that of the 13-inch M1 model. According to cnBeta, citing Apple’s technical specifications, the 16GB configuration provides 153GB/s of memory bandwidth, while the 24GB and 32GB configurations provide 170GB/s.
Apple’s 24-inch iMac powered by the M6 chip has opened for preorder on JD.com in China, starting at 12,499 yuan and 10,999 yuan after the national subsidy. The two-Thunderbolt configuration has 10-core CPU and GPU components, while the four-Thunderbolt model has 12-core CPU and GPU components; both reportedly include dual 16-core NPUs. IT之家 also reports a 24-inch 4.5K display, Wi-Fi 7, Bluetooth 6, and optional 2.5Gb Ethernet.
IT之家 reports that Apple’s iPad mini (A20 Pro), commonly called the “iPad mini 8,” has opened for preorder and will go on sale November 6. Pricing starts at RMB 5,499, or RMB 4,999 after the national subsidy, with 128GB, 256GB, and 512GB Wi-Fi and cellular configurations offered. The tablet has an OLED display, an A20 Pro chip, neural accelerators for Apple Intelligence and other AI tasks, and an IP67 dust-and-water resistance rating.
IT之家, citing CCTV Finance, reports that products with an average price above 4,000 yuan account for more than 68% of the second-hand phone market in Shenzhen’s Huaqiangbei. An industry representative said rising memory prices have kept second-hand phone prices firm this year, with some popular models commanding premiums. The iPhone 15 Pro Max, for example, rebounded from the low-3,000-yuan range last year to 4,300–4,400 yuan. The report also says nearly new or unopened second-hand phones have become sought after by merchants.
Apple plans to bring Siri Recap and Live Rewind to Apple Watch Series 12 and Ultra 4 later this year, according to 9to5Mac. Apple says Siri Recap creates Apple Intelligence-generated summaries of conversations that auto-delete after seven days unless saved, while Live Rewind displays a text snippet from the previous 15 seconds after a Digital Crown double-press. Apple says raw recordings are not retained and stored text and summaries are end-to-end encrypted, although the final implementation and broader privacy treatment remain subject to scrutiny.
TechRepublic compares Siri AI and Google Gemini across everyday questions, personal context, screen awareness, app automation, privacy, and device compatibility. It reports that Siri AI is built into compatible Apple devices, while Gemini is available on supported Android phones and through a standalone iPhone app, with both offering conversations, writing help, planning, and supported cross-app actions. The comparison uses official documentation and published reporting but does not independently benchmark speed, accuracy, or task completion on equivalent devices.
Safari Notify Me, part of Apple Intelligence, can monitor selected webpages for changes and send alerts when conditions such as a price drop or product restock occur. Users can describe what to track in plain language and adjust the checking interval. The feature arrived with Apple’s September 2026 software releases, including iOS 27, iPadOS 27, and macOS 27 Golden Gate. Apple’s documentation does not establish detection accuracy for login-protected pages, CAPTCHA challenges, or dynamically generated inventory data. Notifications appear on the device running the monitor, rather than necessarily syncing to other devices.
Apple released iOS 27.2 on October 26, expanding Siri with historical conversation memory and access to personal data such as voice mail and message attachments. The update also adds an exclusion option so users can control which apps Siri can access, while TMTPost examines the speculative possibility that AI-mediated app actions could shift traffic away from the App Store.
Judge Edward Davila granted several of Apple’s previously deferred discovery requests in its lawsuit alleging trade-secret misappropriation by former employees Chang Liu and Tang Tan, OpenAI, and io Products. The approved discovery covers Liu’s use of ChatGPT, Claude, Codex, OpenClaw and other AI tools; devices, accounts and messaging apps the former employees allegedly used to access, store or share Apple information; and testimony concerning OpenAI’s recruitment and hardware work involving batteries, power management and metal finishing. The defendants have up to 14 days to produce records and answer Apple’s questions. The judge denied requests concerning OpenAI’s own devices and accounts and internal investigations, while postponing two corporate-testimony requests for regular merits discovery.
Apple released the M6 MacBook Pro and iMac, with China starting prices of 15,999 yuan and 12,499 yuan, respectively, according to ifanr. The M6 has a 12-core CPU and 12-core GPU, while the MacBook Pro ships with 16GB of unified memory and offers up to 32GB at 170GB/s bandwidth. Apple claims up to 30% better graphics performance and 25% better on-device AI workload performance than M5, but the report provides no independent measurements.
SK Hynix expects RAM and flash memory shortages to persist through at least 2030 as AI-driven demand continues to outpace supply, according to 36Kr and 9to5Mac. The company attributes the imbalance to the difficulty of leading-edge production processes and the long lead times for new facilities. The Register reports that 14 major buyers have long-term priority-order agreements covering more than half of SK Hynix’s output in 2027 and 2028. These agreements include a mechanism allowing prices to reflect market conditions. Apple is believed to be among the buyers, but its participation is not confirmed.
Apple released iOS 27.2 to all supported iPhones on October 26 after six weeks of beta testing. The update expands Siri AI into five additional languages, adds conversation recall and broader contextual information, and brings Multiview for up to four live sports streams, expanded FaceTime Dual Capture, new Apple Podcasts sections, and iPhone Duo fold-state automation. It also includes a large security patch set; CNET reported 141 bugs and security issues, while iThome Taiwan counted 163 vulnerabilities in the iOS and iPadOS advisories, and one fix addressed unauthorized camera access.
Gizmodo, citing MacRumors, reports that Apple’s iPhone Duo does not support the FaceTime Eye Contact feature, which uses on-device AI to make users appear to look into the camera during a call. The feature has been available across numerous iPhone models since iOS 14, but Apple has not explained the omission or said when it might be added.
Apple unveiled a new iPad mini with a larger 8.5-inch OLED display, an A20 Pro chip, Wi-Fi 7 and a starting price of $699. The tablet is 5.4 millimeters thick and weighs 255 grams. Apple says the A20 Pro provides up to 60% faster multitasking and 4.5 times faster AI performance and is built to run Siri AI. Other additions include “Acoustic Display” speakers, a landscape front camera and an optional $100 nano-texture coating designed to reduce glare.
TechNews reports that Apple launched a new 24-inch iMac powered by the M6 chip, with up to 12 CPU cores and 12 GPU cores. It features on-device AI capabilities, memory bandwidth of up to 170GB/s, macOS 27 Golden Gate and Apple Intelligence, and can run AI models locally with applications such as LM Studio Bionic and Topaz Studio. Taiwan pricing starts at NT$49,900, with preorders opening October 29 and sales beginning November 6. Apple says prompt-processing performance in LM Studio is up to five times that of M4 iMacs and up to 12 times that of M1 models.
Apple announced beta AI assistants for Pages, Keynote, Numbers, and Logic Pro, built with a combination of Apple and OpenAI models. Apple says Pages can draft and restructure text, Keynote can design and refine slides, Numbers can calculate, organize, and visualize data, and Logic Pro can guide music-production workflows. MCP support will let external AI agents and tools read content, make edits, and automate work in Pages, Keynote, and Numbers on macOS 27 or later. Other updates include Generate Fill in Pixelmator Pro and built-in MotionVFX in Final Cut Pro. Apple plans to release the updates in November at no extra charge for existing Creator Studio subscribers, with no exact date announced.
Apple launched a 24-inch iMac powered by the M6 chip while retaining the M4 model's display and overall design. The M6 offers up to 12 CPU cores, 12 GPU cores and dual 16-core Neural Engines, with memory bandwidth of up to 170GB/s. Apple says the system provides up to 40% faster CPU performance, twice the graphics and storage performance and five times the on-device AI performance of the M4 model. It adds Apple's N1 chip for Wi-Fi 7 and Bluetooth 6, plus an optional 2.5Gb Ethernet adapter. Preorders begin October 28, with availability scheduled for November 6. The US starting price is $1,499.
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…