SSPAI

Category: Google

  • SSPAI Morning Brief: Google Unveils Gemini-Powered Android Features and Teases New Googlebook AI Laptop Platform

    SSPAI Morning Brief: Google Unveils Gemini-Powered Android Features and Teases New Googlebook AI Laptop Platform

    Morning Brief

    1. Google Rolls Out Multiple Android Feature Updates
    2. Tencent’s Acquisition of Ximalaya Approved
    3. Spotify Launches 20th Anniversary Recap Page
    4. eBay Rejects GameStop Acquisition Proposal
    5. Linux Exposes Another High-Risk Privilege Escalation Vulnerability, Dirty Frag
    6. Sony Unveils Officially Licensed Scuf Omega Controller
    7. Sony Launches Reon Pocket Pro Plus Wearable Air Conditioner
    8. Forza Motorsport 6 Leaked Early and Cracked
    9. News Worth a Quick Look

    Google Rolls Out Multiple Android Feature Updates

    On May 13, Google announced a range of Android feature updates during its online livestream event, The Android Show I/O Edition 2026, while also teasing the brand-new Googlebook laptop product lineup for the first time.

    The new Gemini Intelligence experience deeply integrates Gemini into Android systems and hardware. It will first launch this summer on the Samsung Galaxy S26 and Pixel 10 series, before later expanding to Wear OS smartwatches, in-car systems, smart glasses, and notebook devices. Core experiences include cross-app and multi-step automated operations. Chrome for Android will launch Gemini in Chrome for Android 12 and above devices in the United States by the end of June. Powered by the Gemini 3.1 model, it supports webpage summaries, contextual Q&A, integration with Gmail, Calendar, and Keep, as well as Nano Banana image generation and editing features. It also supports automatic agent-based browsing capabilities that can autonomously complete tasks such as parking reservations and order modifications on websites.

    In addition, Google Autofill has been upgraded into an intelligent form-filling system powered by Gemini Personal Intelligence, capable of automatically completing complex mobile forms using data from authorized apps. Gboard has also introduced a new speech-to-text feature called Rambler, which can automatically clean up pauses, repetitions, and filler words in spoken language while supporting multilingual mixed input and contextual understanding. Android is also introducing Create My Widget, a generative widget creation feature that allows users to create customized Gemini-powered widgets directly through natural language. Source 1, 2

    Quick Share is further expanding compatibility support for AirDrop, with broader support arriving later this year for Samsung, OPPO, OnePlus, vivo, Xiaomi, and Honor devices. Meanwhile, all Android phones will be able to generate QR codes through Quick Share to share files with iOS devices via the cloud, while WhatsApp integration will also be added to the sharing workflow. Google has additionally redesigned the migration process from iOS to Android, supporting wireless transfers of passwords, photos, messages, apps, contacts, home screen layouts, and eSIMs. The feature will first launch on Samsung Galaxy and Pixel devices. Source

    Finally, Google also previewed the Googlebook laptop platform, which merges Android and ChromeOS into a new category of Gemini Intelligence-centered devices. Googlebook laptops feature a Magic Pointer intelligent cursor that can trigger Gemini contextual actions based on cursor position. They also support generating personalized desktop widgets through natural language and can integrate with Google services such as Gmail and Calendar. Googlebook additionally supports seamless switching between smartphone apps and Quick Access direct browsing of phone files, allowing users to view and use smartphone files directly on the laptop without transferring them.

    On the hardware side, Google is partnering with Acer, ASUS, Dell, HP, and Lenovo to launch the first wave of Googlebook devices. All models will adopt the classic Chromebook-style Glowbar illuminated light strip design. More product information is expected to be revealed this fall. Source

    Read More: Google Held an Android Showcase Early, but Android Still Wasn’t the Main Character

    Tencent’s Acquisition of Ximalaya Approved

    On May 12, China’s State Administration for Market Regulation announced that it had approved Tencent Holdings’ acquisition of Ximalaya through a decision with additional restrictive conditions attached. The regulator stated that the merger could potentially create anti-competitive effects within China’s online audio streaming platform and online music streaming platform markets. As a result, Tencent, Ximalaya, and the post-merger entity are required to make five restrictive commitments. These commitments include not raising service prices for online audio streaming platforms, not lowering service quality or imposing unreasonable transaction conditions, not reducing the proportion of free or popular free content on online audio platforms, and not entering into exclusive licensing agreements with copyright holders. The regulator stated that these commitments would help mitigate competition concerns caused by the acquisition while protecting the interests of consumers, copyright owners, content creators, and automobile manufacturers. Authorities also stated they would supervise both parties to ensure compliance with the imposed commitments. Source

    Spotify Launches 20th Anniversary Recap Page

    On May 12, Spotify celebrated its 20th anniversary by launching a limited-time personalized recap experience. Users can search for “Spotify 20” or “Party of the year(s)” within the app to access it, where they can view when they joined Spotify, the first song they played, their favorite artist, and the total number of songs they’ve listened to over time. Users can also save a personalized 120-track playlist ranked by their own listening counts.

    Spotify stated that the feature is available globally but will only remain accessible for six weeks. Last month, the company also released cumulative 20th anniversary platform rankings, with Taylor Swift currently holding the title of the most-streamed artist in Spotify history. The rankings also included the most-streamed albums, songs, and podcasts as of last month. Source

    eBay Rejects GameStop Acquisition Proposal

    eBay has officially rejected GameStop’s previously proposed $56 billion unsolicited acquisition offer. Last week, GameStop proposed acquiring eBay at $125 per share, representing a 20% premium over eBay’s current stock price. The proposed deal structure consisted of half cash and half GameStop stock. However, GameStop’s current market capitalization stands at roughly $11 billion — less than a quarter of eBay’s approximately $45 billion valuation. Although GameStop had also planned to raise $20 billion through debt financing, CEO Ryan Cohen later failed to provide detailed explanations regarding the remaining funding sources required for the acquisition. eBay stated that it had thoroughly reviewed the proposal, expressing major concerns over financing feasibility and debt pressure. eBay Chairman Paul Pressler additionally described the offer in a letter as “neither credible nor compelling.”

    GameStop previously became one of the defining meme stocks of 2021. In recent years, the company has shifted its focus from NFTs toward retro gaming businesses, while also closing more than 400 stores across the United States earlier this year. Source

    Linux Exposes Another High-Risk Privilege Escalation Vulnerability, Dirty Frag

    Researchers have disclosed a Linux kernel local privilege escalation exploit chain named Dirty Frag. By combining CVE-2026-43284 and CVE-2026-43500, attackers with low-level privileges — including those operating within containers or virtual machine environments — can reliably obtain root access. Exploit code reportedly leaked within three days of disclosure, and Microsoft stated that it has already observed signs of real-world attack testing activity.

    Both vulnerabilities originate from flaws in Linux kernel page cache handling. CVE-2026-43284 targets the esp4 and esp6 processing paths within IPsec ESP receiving flows, while CVE-2026-43500 targets rxrpc. Researchers stated that the vulnerability belongs to the same family as Dirty Pipe and Copy Fail, but shifts its focus from pipe_buffer structures to the frag member within struct sk_buff. Although exploiting either vulnerability independently offers limited reliability, chaining them together reportedly enables stable privilege escalation across all major Linux distributions, allowing attacks such as SSH compromise, web shell deployment, and container escape scenarios.

    Debian, AlmaLinux, and Fedora have already released patches. Google stated that Kubernetes’ default hardening configurations can reduce container escape risks, although virtual machine environments and less restricted systems remain highly vulnerable. Source

    Sony Unveils Officially Licensed Scuf Omega Controller

    On May 12, Sony announced the launch of the officially PlayStation-licensed Scuf Omega gaming controller developed by Scuf Gaming. The controller is compatible with PS5, PC, Mac, iOS, and Android, offering 28 input controls including 11 customizable buttons, 4 rear paddles, 2 side buttons, and 5 G-Keys. Users can configure mappings directly on the controller itself or through the Scuf Mobile app on iOS and Android. Multiple configuration profiles can also be created, covering trigger and joystick response curves, dead zones, button selection, joystick response areas, audio settings, and lighting effects.

    Scuf Omega additionally adopts Omron micro switches for directional and function buttons, while featuring adjustable triggers capable of switching between click mode and analog mode. The joysticks use non-contact magnetic sensing technology and support interchangeable concave, convex, and varying-height thumbsticks. Notably, the controller removes vibration modules to reduce weight, while also featuring textured anti-slip grips and an ergonomic design optimized based on feedback from professional players. Scuf Omega is available starting today at suggested retail prices of $219.99 USD, €239.99 EUR, £209.99 GBP, or ¥37,980 JPY. Source

    Sony Launches Reon Pocket Pro Plus Wearable Air Conditioner

    On May 12, Sony updated its “wearable air conditioner” product lineup with the launch of the new Reon Pocket Pro Plus, priced at €229. In terms of wearable design, the Reon Pocket Pro Plus adopts a pair of redesigned flexible support arms that wrap around both sides of the neck, while the main unit rests on the upper back. Compared with previous models, the new design improves fit and stability, making it less likely to shift while walking or bending over. To address heat dissipation, Sony redesigned the exhaust structure by introducing a retractable and angle-adjustable air outlet. Users can extend the vent and fine-tune its angle according to collar height and clothing style, directing hot air away from the body. Even when wearing high-collar shirts or standing-collar jackets, the design helps prevent heat from accumulating around the neck and back. The new model also upgrades its cooling algorithm, improving overall cooling performance by roughly 20% over the previous generation, while lowering the cooling plate surface temperature by up to an additional 2 degrees Celsius. Sony also recommends pairing the device with the Reon Pocket Tag 2 temperature and humidity sensor for smarter automatic cooling adjustments. Source

    Forza Motorsport 6 Leaked Early and Cracked

    On May 11, the PC version of Forza Motorsport 6, originally scheduled for official release on May 19, was leaked and cracked ahead of launch. According to reports, the leak occurred because Playground Games uploaded Steam preload files in an unencrypted format, allowing players to extract and distribute the files through sharing sites such as GameDrive. The complete leaked build is reportedly around 150GB in size.

    Following the incident, Playground Games issued a statement on X warning that players who download or use the leaked version would face severe penalties, including motherboard serial number bans and account-wide restrictions across all Forza titles. Source

    News Worth a Quick Look

    • Tokyo Game Show 2026 has unveiled its official visual artwork and announced that the event will take place from September 17–21. This year’s theme is “The Longest and Most Play-Filled 5 Days in History” (史上最⻑、遊びづくしの 5DAYS). Tokyo Game Show will also celebrate its 30th anniversary this year, with additional commemorative announcements expected later. Source
    • The FIA, Formula 1 teams, power unit manufacturers, and Formula 1 Management have reportedly reached a preliminary agreement to adjust the balance of power units for the 2027 season. The proposal would increase V6 internal combustion engine output by 50kW to 450kW (603hp) while reducing electric motor output by 50kW to 300kW (402hp). By increasing fuel flow rates, the goal is to raise combustion engine performance and mitigate the “super clipping” issue and track speed disparities caused by insufficient battery capacity under the upcoming 2026 regulations. Source
  • SSPAI Morning Brief: OpenAI GPT-5.5 Launches as Google Unveils TPU 8t and 8i Chips, Powering the Next Wave of AI Innovation

    SSPAI Morning Brief: OpenAI GPT-5.5 Launches as Google Unveils TPU 8t and 8i Chips, Powering the Next Wave of AI Innovation

    Morning Brief

    1. OpenAI releases GPT-5.5 series models
    2. Hanvon launches Clear7C color e-ink e-reader
    3. Onyx Boox introduces Leaf5+ monochrome e-ink e-reader
    4. iFLYTEK unveils AINOTE2 e-ink device
    5. Panasonic releases LUMIX ZS300 camera
    6. Google introduces two AI chips: TPU 8t and 8i
    7. Tencent Cloud issues Xinference supply chain poisoning risk advisory
    8. News Worth a Quick Look

    OpenAI releases GPT-5.5 series models

    OpenAI has announced the GPT-5.5 series, describing it as the most intelligent and intuitive model version to date. Compared to GPT-5.4, the new models deliver stronger performance in coding and debugging, online research, and generating documents and spreadsheets, while also supporting cross-tool collaboration to execute complex tasks. Users can input multi-step, unstructured requests, and the model can automatically handle task planning, tool invocation, result verification, and uncertainty management. In Codex scenarios, GPT-5.5 significantly reduces token usage and introduces more robust safety mechanisms.

    GPT-5.5 will be gradually rolled out to ChatGPT Plus, Pro, Business, and Enterprise users, and will be integrated into the Codex platform. The Pro tier will offer more advanced capabilities. Source

    Hanvon launches Clear7C color e-ink e-reader

    On April 23, Hanvon launched the Clear7C color e-ink e-reader. The device features a 7-inch Kaleido 3 color e-ink display with a resolution of 1680×1264, supporting 150 PPI in color mode and 300 PPI in black-and-white mode. It also includes physical page-turn buttons on one side. The Clear7C weighs approximately 175g and is just 3.9mm thick, with a strong focus on portability and one-handed usability.

    In terms of hardware, the Clear7C is powered by an octa-core processor, paired with 4GB of RAM and 64GB of storage, and includes a 2700mAh battery. The company claims up to 120 hours of reading time. It runs on Android 14. Source

    Onyx Boox introduces Leaf5+ monochrome e-ink e-reader

    On April 23, Onyx Boox introduced the Leaf5+ monochrome e-ink e-reader. The device features a 7-inch Carta 1300 panel with a 300 PPI resolution, supports stylus input, and uses etched tempered glass for the display, along with physical page-turn buttons. The Leaf5+ is about 4mm thick and weighs 192g, and also includes a new light guide plate, with the company claiming up to 90% screen uniformity and a color temperature adjustment range of 3000K–5500K.

    For hardware, it is equipped with a Qualcomm octa-core processor, 4GB LPDDR4X RAM, and 64GB of storage, along with a 2000mAh battery. The device is priced at RMB 1,749. Source

    iFLYTEK unveils AINOTE2 e-ink device

    On April 23, iFLYTEK unveiled the AINOTE2 e-ink device, described as an “ultra-thin e-ink tablet.” It features a 10.65-inch monochrome e-ink display with a resolution of 1920×2560 and a pixel density of 300 PPI, supporting handwriting input via the iFLYTEK I-Pencil electromagnetic stylus. The device uses a magnesium-aluminum alloy body, measuring about 4.2mm thick and weighing 295g.

    In terms of hardware, the AINOTE2 comes with 4GB of RAM, 64GB of storage, and a 4000mAh battery. It runs on Android 14 and includes several custom AI features from iFLYTEK, such as content summarization, speech-to-text transcription, and integrated note-taking. The device is priced at RMB 3,999. Source

    Panasonic releases LUMIX ZS300 camera

    On April 23, Panasonic announced the LUMIX ZS300. The device features a 1-inch CMOS sensor with approximately 20.1 million effective pixels, paired with a Leica-certified 15× optical zoom lens (equivalent to 24–360mm). It supports a minimum focusing distance of 3 cm and includes 5-axis hybrid image stabilization. The camera is available in black and silver, with a built-in flash and 22 creative filters.

    In terms of features, the LUMIX ZS300 supports 4K video recording, with a 4K Photo mode capable of 30 fps high-speed burst shooting, as well as FHD 120fps slow-motion recording. For connectivity, it supports Bluetooth and Wi-Fi for pairing with smartphones and transferring images, and also supports charging via a USB Type-C port. The product will officially go on sale at 12:00 on April 27, with a starting price of RMB 4,998. Source

    Google introduces two AI chips: TPU 8t and 8i

    At the Google Cloud Next 2026 conference, Google unveiled two new generations of custom AI ASIC chips: TPU 8t and TPU 8i, optimized for training and inference scenarios respectively, while also supporting cross-scenario workloads.

    The TPU 8t is designed for large-scale, compute-intensive training tasks, focusing on improved compute throughput and vertical scaling bandwidth, delivering higher efficiency in complex model training. The TPU 8i, on the other hand, is optimized for low-latency inference and AI agent workloads, offering higher memory bandwidth and integrating a CAE (on-chip acceleration engine). It is equipped with 384 MB of on-chip SRAM and 288 GB of off-chip HBM, reducing data transfer overhead during execution and lowering latency.

    At the cluster level, the TPU 8i Pod scale has increased from 256 chips in the previous generation to 1,152 chips. A single Pod delivers 8.67× higher FP8 compute performance and 5.74× greater HBM memory capacity. Google states that the overall energy efficiency of the new TPU 8t and TPU 8i is approximately double that of the previous-generation TPU Ironwood. Source

    Tencent Cloud issues Xinference supply chain poisoning risk advisory

    On April 23, Tencent Cloud released a security advisory regarding a supply chain poisoning risk affecting the open-source PyPI inference framework Xinference. The advisory states that affected versions may contain malicious code when installed or imported, potentially exfiltrating sensitive data—including cloud platform credentials, API keys, SSH keys, cryptocurrency wallet information, database credentials, and environment variables—to attacker-controlled command-and-control servers.

    Tencent Cloud noted that such risks typically originate from tampering within the software supply chain and may pose significant threats to AI inference environments and related infrastructure. Affected users are advised to immediately uninstall compromised versions, scan systems for backdoors and residual files, and rotate critical credentials and keys as soon as possible. A comprehensive security audit is also recommended to assess potential impact and strengthen system defenses. Source

    News Worth a Quick Look

    • X product lead Nikita Bier said that Premium subscribers on iOS can now early-access a new feature: users can pin specific topics on the “Home” feed, and Grok will personalize timeline recommendations based on those topics and past interactions. Powered by Grok’s understanding of platform content and personalization algorithms, the feature aims to deliver a highly customized feed experience, with an Android rollout expected soon. In addition, X announced it will shut down the Communities feature on May 6 due to a continued decline in user activity, and recommends migrating communities to group chats on XChat. Meanwhile, competing platforms like Threads and Mastodon continue to strengthen community-driven feeds. Source
    • Memory manufacturer Kioxia introduced the EG7 series SSDs on April 23. The lineup uses QLC NAND across all models and does not include a DRAM cache. Performance-wise, EG7 offers up to 7000 MB/s sequential read and 6200 MB/s sequential write speeds, with random performance reaching up to 1,000K IOPS. Capacities include 512GB, 1TB, and 2TB, with form factors in M.2 2230, 2242, and 2280. The series targets the consumer market and has already been supplied to multiple OEMs for use in thin-and-light laptops and desktops. Source
    • Google confirmed during its Cloud Next 2026 keynote on April 22 that its partnership with Apple stems from a multi-year agreement signed in January 2026. Under the deal, Google will act as the preferred cloud provider, delivering core technical support, with results expected to materialize by the end of 2026. Source
    • Rambus announced a chipset solution for LPDDR SOCAMM2 modules on April 22, designed to provide control, telemetry, and power management for JEDEC-compliant SOCAMM2 server memory modules. The goal is to meet the high stability and efficiency demands of AI servers. SOCAMM2 introduces a modular, replaceable design to replace traditional soldered LPDDR memory, combining LPDDR’s energy efficiency with improved serviceability and scalability for data centers. Rambus’ LPDDR5X SOCAMM2 chipset supports speeds up to 9.6 Gb/s, emphasizing reliability and efficiency under high bandwidth workloads. Source
    • According to Mark Gurman, Apple is developing six new product categories, including AI-powered AirPods, smart glasses, wearable clip-on devices, smart displays, desktop robots, and security cameras. Source
  • SSPAI Morning Brief: Google IPv6 Traffic Surpasses 50% Milestone, Claude Opus 4.7 Token Costs Rise Significantly

    SSPAI Morning Brief: Google IPv6 Traffic Surpasses 50% Milestone, Claude Opus 4.7 Token Costs Rise Significantly

    Morning Brief

    1. China’s low-altitude authority responds to drone “takeoff difficulties”
    2. Seven e-commerce platforms fined RMB 3.597 billion over ghost kitchens
    3. Kindle for PC to shut down, fewer options for DRM removal
    4. Claude Opus 4.7 shows significantly higher token usage
    5. Google’s IPv6 traffic share surpasses 50% for the first time
    6. Digital and smart products become key drivers in trade-in consumption
    7. News Worth a Quick Look

    China’s low-altitude authority responds to drone “takeoff difficulties”

    According to Caixin, at a State Council Information Office press conference on April 17, Zheng Jian, Director of the Low-Altitude Economy Development Department under the National Development and Reform Commission, responded that in light of difficulties in obtaining approvals for drone flights, the department is working with relevant agencies to promote effective local practices such as “scan-to-fly,” aiming to improve the efficiency of flight plan approvals. This marks the first official response to widespread user complaints since tighter drone regulations were introduced.

    Under current regulations, drones must be registered with real-name identification on either the Civil Aviation Administration’s Unmanned Aircraft Management Platform (UOM) or the Public Security Drone Management Platform. Each flight must then be submitted for prior approval, and only after authorization can the drone be operated—otherwise, users may face fines or detention from public security authorities. In practice, these rules have proven difficult for many drone users to adapt to.

    The “scan-to-fly” approach mentioned in the response originates from Sichuan Province, where local authorities have experimented with easing flight demand. After completing an online filing via a mini program, users can fly within designated permissible airspace without having to visit local police stations in person. In addition to Sichuan, Shanghai introduced a similar service module in early February via its “Suishenban” government app, designating three drone flight experience zones within controlled urban airspace.

    Industry insiders note that China’s low-altitude airspace is primarily divided between military and civil aviation authorities, with the military holding final authority over approvals. As a result, local governments have limited involvement. A representative from a domestic drone company pointed out that while flight application volumes are large, approval capacity remains constrained, and suggested that granting local governments more control over suitable and partially restricted airspace, along with improving automation in approvals, could be a solution.

    Previously, on November 8, 2023, the Civil Aviation Administration released a draft of the Airspace Management Regulations, proposing the establishment of hierarchical air traffic management coordination bodies responsible for airspace governance. Industry experts believe the legislation could be finalized by 2026.

    Seven e-commerce platforms fined RMB 3.597 billion over ghost kitchens

    According to Caixin, on April 17, China’s State Administration for Market Regulation (SAMR) announced administrative penalties against seven e-commerce platforms—including Pinduoduo, Meituan, JD.com, Ele.me (now Taobao Flash Delivery), Douyin, Taobao, and Tmall—for violations related to “ghost kitchen” operations. The platforms were ordered to rectify illegal practices, suspend the onboarding of new bakery merchants for periods ranging from three to nine months, and pay combined fines and confiscations totaling RMB 3.597 billion. Additionally, legal representatives and food safety directors of the seven companies were fined a total of RMB 19.6874 million. Pinduoduo faced the largest penalty at RMB 1.522 billion, followed by Meituan, JD.com, and Ele.me at RMB 746 million, RMB 635 million, and RMB 558 million respectively; Douyin, Taobao, and Tmall were fined RMB 56.89 million, RMB 46.97 million, and RMB 31.74 million.

    “Ghost kitchens” refer to vendors that lack physical dining spaces or have already ceased operations but continue to operate on delivery platforms. Order-routing platforms allow merchants to transfer orders to other food businesses for fulfillment.

    According to SAMR, the platforms failed to properly verify the licenses of food vendors, neglected their legal obligations for qualification review, and entered into agreements with order-routing platforms while knowing—or having reason to know—that such practices infringed on consumer rights, yet failed to take necessary measures. Legal representatives and food safety directors also failed to fully perform their responsibilities.

    Additionally, SAMR’s administrative penalty document against Shanghai Xunmeng Information Technology Co., Ltd. (Pinduoduo’s operating entity) revealed that during the investigation, the company repeatedly refused to provide materials without justification, submitted false information, and even obstructed enforcement through confrontational tactics. Previous reports indicated that on December 3, multiple Pinduoduo employees clashed with regulators during the investigation, leading to the dismissal of several staff in its government relations department and triggering public labor disputes.

    Earlier, in November 2025, SAMR guided eight major online food trading platforms—including JD.com, Meituan, Pinduoduo, Douyin e-commerce, Xiaohongshu, Taobao, WeChat Shops, and Kuaishou e-commerce—to jointly sign a self-regulatory agreement on food safety management. On February 26, 2026, SAMR announced that new regulations on food safety responsibilities for online catering service operators will take effect on June 1, requiring delivery platforms to implement real-name registration for vendors and conduct on-site verification to ensure license information matches actual conditions.

    Kindle for PC to shut down, fewer options for DRM removal

    According to Good e-Reader, Amazon recently notified users via pop-up that the current Kindle for PC desktop client will be discontinued on June 30. After that date, the software will no longer function. Amazon confirmed it is developing a new Kindle app for PC, compatible only with Windows 11 and available exclusively through the Microsoft Store.

    The original Kindle for PC client was launched in 2009 and has long been used by users to download e-books locally and remove DRM (digital rights management) protections. In recent years, Amazon has repeatedly urged users to upgrade the client and restricted access for older versions in order to patch vulnerabilities and combat piracy. In 2023, Amazon had already discontinued the older Kindle for Mac, replacing it with a version distributed solely via the Mac App Store.

    Compared to standalone installers, app store–distributed versions are generally harder to bypass technically. By shifting its PC client entirely to the Microsoft Store, Amazon aims to further tighten control over Kindle and meet publishers’ requirements for preventing e-book piracy.

    Claude Opus 4.7 shows significantly higher token usage

    According to the official migration guide for Claude Opus 4.7, the model adopts a new tokenizer, which can increase token usage by up to 35% compared to previous versions.

    However, user testing suggests this may be a conservative estimate. In one test, TypeScript code saw token usage increase by 1.36×, while English technical documentation rose by 1.47×. In a debugging session involving 80 conversational turns, costs increased from approximately $6.65 to nearly $8.80. Another dataset of over 500 samples showed an average token usage increase of 38.6%. In contrast, token consumption for non-Latin scripts such as Chinese, Japanese, and Korean (CJK) remained largely unchanged.

    The new tokenizer breaks English text and code into smaller segments, reducing the number of characters per token. Anthropic claims this improves task accuracy, but many users have already complained that usage limits are now depleted even faster.

    Google’s IPv6 traffic share surpasses 50% for the first time

    According to data released by Google, on March 28, 2026, global user traffic accessing its services via IPv6 reached 50.1%, up from 46.33% during the same period last year—marking the first time the metric has exceeded the halfway point.

    However, data from other internet infrastructure organizations suggests that IPv6 has not yet fully taken the lead. Monitoring by Cloudflare indicates that IPv6 currently accounts for only 40.1% of global HTTP request sources; data from the Asia-Pacific Network Information Centre (APNIC) shows that the proportion of networks with IPv6 capability stands at around 43.13%.

    IPv6 was introduced to address the exhaustion of IPv4 addresses. IPv4 can provide only about 4.3 billion IP addresses, while IPv6, using 128-bit addressing, offers an almost limitless address space. However, because the new protocol did not bring many disruptive new features, and due to the widespread use of Network Address Translation (NAT)—which allows a large number of devices to share a single public IPv4 address—many organizations have relied on NAT to mitigate address shortages. As a result, the global adoption of IPv6 has long lagged behind expectations.

    The pace of IPv6 adoption also varies significantly across regions. Due to relatively lenient allocation mechanisms in the early days of the internet, Europe and the United States secured large amounts of IPv4 resources. In contrast, populous countries such as China and India received far fewer IPv4 addresses, prompting earlier and more aggressive deployment of IPv6. According to APNIC data, 29 countries in the Asia-Pacific region, including China, surpassed the 50% IPv6 adoption threshold in 2025.

    Digital and smart products become key drivers in trade-in consumption

    According to Xinhua News Agency, based on data from the Ministry of Commerce’s national system for home appliance trade-ins and digital and smart product purchases, as of April 16, purchases of digital and smart products reached 42.433 million units, up 31.7% year-on-year, with total sales of RMB 126.153 billion, up 36.4%. Among these, mobile phone sales accounted for more than 80%.

    Since the beginning of this year, subsidies for new purchases of digital products—such as smartphones—have been expanded and upgraded to cover a broader range of digital and smart products, with smart glasses included for the first time. This has driven rapid growth in retail sales of communication equipment. By mid-April, 16 domestic smart glasses brands had participated in the subsidy program, boosting sales volume and revenue of key enterprises by 42.4% and 46.8% year-on-year, respectively.

    It is reported that under the combined effect of home appliance trade-in programs and digital product subsidies, retail sales of communication equipment above the designated size reached RMB 284 billion from January to March, up 20.8% year-on-year—18.4 percentage points higher than the overall growth rate of total retail sales of consumer goods, ranking first among 16 product categories.

    News Worth a Quick Look

    • According to Nikkei, the three major DRAM suppliers—Samsung Electronics, SK Hynix, and Micron Technology—can currently meet only about 60% of total market demand. By mid-2026, the share of memory costs in low-end smartphone manufacturing is expected to double from 20% to nearly 40%.
    • Recently, NVIDIA CEO Jensen Huang stated on the Dwarkesh Podcast that he opposes stricter U.S. export controls on chip equipment to China, arguing that China’s abundant energy resources and manufacturing capabilities could allow it to achieve large-scale computing power through system scaling. He warned that strict export controls could inadvertently push China to build a strong, independent technology stack, potentially causing the U.S. to lose access to the world’s second-largest market and weakening the global influence of U.S. technology standards.
    • According to Microsoft’s release notes, in Windows 11 Insider Preview Build 26300.8170, the FAT32 partition size limit has been increased from 32GB to 2TB, although the change is currently only accessible via command-line operations. The 32GB limit had long been an artificial restriction imposed by Microsoft and remained unchanged for decades.
    • According to MacWorld, this year’s iPhone 18 Pro is expected to feature “Dark Cherry” as its signature color, while the foldable iPhone is rumored to measure just 4.7 mm thick when unfolded, with currently tested color options including silver-white and indigo.
  • SSPAI Morning Brief:Google Gemini Adds Notebook Feature for AI Knowledge Management; Dyson Launches Portable HushJet Mini Bladeless Fan

    SSPAI Morning Brief:Google Gemini Adds Notebook Feature for AI Knowledge Management; Dyson Launches Portable HushJet Mini Bladeless Fan

    Morning Brief

    1. WeChat Pay launches Skill tool integration system
    2. Amflow introduces electric mountain bike lineup
    3. Dyson releases HushJet Mini Cool bladeless handheld fan
    4. Google Gemini introduces “Notebook” feature
    5. MiniMax launches MMX-CLI command-line tool
    6. Coze releases version 2.5 update
    7. Tencent unveils QClaw v2 and QBotClaw
    8. News Worth a Quick Look

    WeChat Pay launches Skill tool integration system

    WeChat Pay has launched a new AI-oriented integration solution, covering Skill packages, AI-friendly documentation, and AI-friendly APIs. Developers can load Skill packages with simple instructions and complete integration through conversational interfaces in common AI development tools.

    According to WeChat Pay, the solution organizes business knowledge, code examples, and integration specifications in a structured way, enabling AI agents to understand requirements and assist developers in completing payment integrations more efficiently. Each Skill corresponds to specific product capabilities, covering integration workflows, API examples, and interaction guidelines to ensure accurate and reliable outputs. The Skill system integrates multiple categories, including basic payments and coupon-related functions, further lowering development barriers and improving integration efficiency for merchants and developers. Source

    Amflow introduces electric mountain bike lineup

    On April 9, Amflow, an electric bike brand spun off from DJI, introduced two product lines: PX and PR. These models are equipped with the Avinox M2S and M2 drive systems, supporting up to 1500W peak power and 150Nm torque, pushing beyond traditional trade-offs between weight, range, and power in e-assist systems. The carbon fiber frame supports up to 40 geometry adjustment combinations, offering enhanced adaptability and customization for riders.

    Additional features include navigation display, Apple Find My integration on the PR series, and heart rate–based intelligent assistance adjustment. The bikes also support expandable battery options, with the standard 600Wh battery charging from 0% to 80% in 1.5 hours. In terms of pricing, the Amflow PX Carbon is now available in Europe and Australia starting at $7,999, while the PR Carbon is expected to launch later this year with a starting price of $4,999. Source

    Dyson releases HushJet Mini Cool bladeless handheld fan

    On April 9, Dyson introduced the HushJet Mini Cool, a handheld bladeless fan that brings its Air Multiplier technology into a portable form factor. The device features a 65,000 RPM brushless motor, delivering wind speeds of up to approximately 55 mph, with multiple adjustable speed levels.

    It is equipped with a 5,000 mAh battery, offering up to 6 hours of runtime, with noise levels ranging from 52 dBA to 72.5 dBA. The device adopts a cylindrical design with a diameter of about 38 mm, and its nozzle can be rotated to adjust airflow direction. It also supports neck-worn use for hands-free cooling. The product is now available at a price of $99, initially offered in Stone and Light Pink color options. Additional colors—including Onyx Red and Sky Blue—will be released in May, followed by Ink Black and Cobalt Blue in June. More color variants and accessories, such as stroller mounts and clip attachments, are planned for future expansion. Source

    Google Gemini introduces “Notebook” feature

    On April 9, Google announced a new “Notebook” feature for Gemini, designed to organize content around specific topics. Users can add files, past conversations, custom instructions, and more into a notebook, which Gemini will use as contextual reference during interactions.

    This feature is similar to the “Projects” function introduced by ChatGPT in 2024, both enabling centralized storage of materials around a single topic. Google describes notebooks as a personal knowledge base shared across its ecosystem, currently available within Gemini. In addition, Gemini notebooks will sync with Google’s AI research tool NotebookLM, meaning content added in one application will appear in both.

    The feature is currently available on the web for Gemini Ultra, Pro, and Plus subscribers, with mobile support expected in the coming weeks and a gradual rollout to free users planned. Source

    MiniMax launches MMX-CLI command-line tool

    MiniMax has introduced MMX-CLI, a command-line tool designed for AI agents. It supportsmultimodal capabilities, including text, image, video, voice, and music generation, and can be integrated into environments such as Claude Code and OpenClaw. By optimizing underlying interaction logic, the tool enables agents to autonomously complete end-to-end workflows—from information retrieval to multimedia generation.

    MMX-CLI also includes features such as output isolation, pure data mode, JSON structured outputs, and a semantic status code system, improving the stability and controllability of automated tasks. Additionally, it supports non-blocking and asynchronous execution, making it suitable for parallel processing scenarios. The tool is integrated with the MiniMax Token Plan, allowing direct access to model quotas. Source

    Coze releases version 2.5 update

    On April 7, Coze announced its 2.5 update, introducing an AI agent collaboration platform called “Agent World,” along with new features such as independent email identities, cloud phones, and cloud PCs. It also upgrades capabilities for video creation agents and programming CLI tools.

    According to the official introduction, Coze 2.5 focuses on three aspects: “persona,” “skills,” and “equipment.” In terms of persona, the new version introduces independent email identities and a memory system, enabling AI agents to communicate via email and build long-term memory of users’ work habits and preferences.

    On the “equipment” side, the update adds a workspace capability for agents. After users describe tasks in conversation, the agent can automatically generate schedules. Users can view plans via a calendar, set timed tasks, and manage agent states. Data, charts, and business reports generated during execution are categorized and stored in a file system.

    In addition, Coze has launched a programming CLI that covers project creation, online preview, and cloud deployment, while also providing real-time messaging and skill extension interfaces tailored for agent-based scenarios. Source

    Tencent unveils QClaw v2 and QBotClaw

    On April 9, Tencent introduced QBotClaw (“Lobster”), an intelligent assistant tool for the QQ Browser. Built directly into the browser, it supports contextual understanding of web content and enables automated operations through webpage element recognition. According to Tencent, users can issue natural language commands to perform tasks such as cross-software operations, information extraction, and file processing, making it suitable for scenarios like data organization, text generation, price comparison, and multi-page content analysis.

    QBotClaw supports integration with third-party large model APIs and can be remotely controlled from mobile devices via the WeChat-based Clawbot service to execute tasks on a computer. The feature is currently available on the Mac version of QQ Browser, with a Windows version yet to be released. Source

    Also released on the same day, QClaw v2 introduces a multi-agent mechanism, application connectors, and a “Lobster Manager” feature. Users can create multiple agents with different capabilities and permission settings, and connect them to third-party applications. Tencent states that the new version can reduce the number of operational steps in certain scenarios.

    In terms of security, QClaw includes the “Lobster Manager” as a safety module to warn or block potential risks, including abnormal commands, misuse of permissions, and access to sensitive information. Source

    News Worth a Quick Look

    • YouTube has introduced an AI feature for Shorts that allows creators to generate digital avatars resembling their appearance and voice for video creation and content expansion. Users must record a “live selfie” including facial and voice data to build the model, with better lighting and quieter environments yielding higher-quality results. Once generated, creators can produce videos up to about 8 seconds long based on prompts or insert avatars into eligible Shorts content to streamline production and expand creative possibilities. Google states that the feature is limited to creators’ own content and can be deleted at any time; data will be automatically removed after three years of inactivity. All avatar-generated videos will be labeled as AI-generated and include digital watermarks such as SynthID and C2PA for identification. Source
    • OpenAI has introduced a new ChatGPT Pro subscription tier priced at $100 per month, targeting high-intensity users. Compared to the $20 Plus plan, it offers approximately five times the usage quota for Codex programming tools, making it suitable for long-duration, high-load development tasks. Like the $200/month Pro tier, it provides access to Pro models and increased availability of both real-time and reasoning models compared to Plus. This new tier fills the gap between the Plus plan and the existing $200 high-end Pro offering, further refining the subscription lineup. OpenAI also noted adjustments to Codex usage for Plus subscribers: “More sessions per week will be supported, rather than longer sessions in a single day.” Source
  • SSPAI Morning Brief: macOS TCP Stack Vulnerability Discovered, Cyberpunk 2077 Gets PS5 Pro Enhanced Update

    SSPAI Morning Brief: macOS TCP Stack Vulnerability Discovered, Cyberpunk 2077 Gets PS5 Pro Enhanced Update

    Morning Brief

    1. macOS kernel TCP stack has an overflow vulnerability
    2. Google Pixel 10a launches in Japan with an exclusive variant
    3. Minisforum releases Elite M1 Lite mini PC
    4. Motorola unveils Moto G Stylus (2026) and Moto Pad
    5. Anthropic and others launch Project Glasswing cybersecurity initiative
    6. Adobe introduces Student Spaces learning tool
    7. 360doc to shut down service on May 1
    8. Cyberpunk 2077 receives PS5 Pro enhanced update
    9. News Worth a Quick Look

    macOS kernel TCP stack has an overflow vulnerability

    A blog post published by Photon on April 6 revealed a TCP stack vulnerability in macOS that can lead to network failure. The issue is triggered by a 32-bit counter in the Apple XNU kernel. When the system uptime reaches 49 days, 17 hours, 2 minutes, and 47 seconds, the internal TCP timer stops functioning. As a result, TCP connections in certain states cannot be properly recycled, eventually exhausting ephemeral ports and preventing the system from establishing any new TCP connections. Only certain types of connections, such as ICMP, remain unaffected.

    Photon’s testing confirms that this kernel flaw affects at least macOS Catalina 10.15 and later versions, while it is unclear whether earlier versions are impacted. Notably, this issue does not cause kernel crashes or generate obvious error logs, making it a silent failure that is difficult for users to diagnose in a timely manner.

    Interestingly, due to macOS’s system update cycle, Apple typically pushes updates and triggers restarts within the window before the bug occurs. As a result, only devices that run continuously for extended periods without restarting are likely to be affected. Source

    Google Pixel 10a launches in Japan with an exclusive variant

    On April 7, Google announced that the Pixel 10a has officially opened for pre-orders in Japan. The company also partnered with local creative agency HERALBONY to introduce a Japan-exclusive color variant, “Isai Blue.”

    This limited edition is available only in a 256GB configuration, with hardware specifications identical to the standard version. It comes bundled with an exclusive protective case and stickers, and features custom wallpapers designed by HERALBONY artists, along with a system-wide theme package in a unified color style.

    In addition to the limited-edition color, other color options are also available simultaneously. The Google Pixel 10a starts at JPY 79,900 in the Japanese market, with the new “Isai Blue” limited edition set to begin shipping on May 20. Source

    Minisforum releases Elite M1 Lite mini PC

    On April 7, Minisforum announced the Elite M1 Lite-125U mini PC, powered by Intel’s Meteor Lake U series Core Ultra 5 125U processor. The device measures 130×126×50.4mm and weighs 0.6kg. It supports a 35W TDP with phase-change TIM and a turbo fan cooling system, keeping noise levels below 45dB. It features two DDR5-5600 SO-DIMM memory slots and two M.2 2280 PCIe Gen4 SSD slots. Networking includes an Intel I226-V wired NIC, along with Wi-Fi 6E and Bluetooth 5.2 for wireless connectivity. For ports, the front panel offers two USB-A 10Gbps ports and a 3.5mm combo audio jack. The rear includes a full-function USB-C 40Gbps port supporting both 100W PD input and 15W output, two USB-A 480Mbps ports, one HDMI 2.1 TMDS, one DisplayPort 1.4, and a 2.5GbE RJ45 port.

    The device is now available on domestic e-commerce platforms, starting at RMB 2,119. Source

    Motorola unveils Moto G Stylus (2026) and Moto Pad

    On April 6, Motorola officially announced the Moto G Stylus (2026), featuring a stylus, along with the new Moto Pad tablet in the U.S. The Moto G Stylus (2026) upgrades from a basic capacitive stylus to an active electromagnetic pen with tilt detection and pressure sensitivity. It includes a 4mAh battery offering up to 4 hours of continuous use and supports 15-minute fast charging. The stylus also enables quick editing, drag-and-drop split-screen, hover zoom, and Google’s Circle to Search via built-in buttons. In terms of hardware, the device is powered by the Qualcomm Snapdragon 6 Gen 3 processor, with 8GB RAM and 128GB or 256GB UFS 3.1 storage. It features a 6.7-inch 120Hz AMOLED display, a 50MP rear main camera, and a 5200mAh battery supporting 68W wired and 15W wireless fast charging, along with IP68/69 dust and water resistance.

    The Moto Pad, unveiled alongside it, comes with an 11-inch 2K display with a 90Hz refresh rate, powered by the MediaTek Dimensity 6300 chip. It includes a 7040mAh battery with 20W charging support, but does not support stylus input.

    The Moto G Stylus (2026) is priced at $499 and will go on sale on April 16. The Moto Pad will be available via T-Mobile starting April 30, with pricing yet to be announced. Source

    Anthropic and others launch Project Glasswing cybersecurity initiative

    On April 7, Anthropic, together with 11 companies including Amazon, Apple, Google, Microsoft, and NVIDIA, launched the Project Glasswing cybersecurity initiative. The project aims to mitigate the risks of AI being maliciously used to attack global economic and security systems by defensively deploying AI tools. Specifically, it leverages an unreleased general-purpose model, Claude Mythos Preview, to identify and fix vulnerabilities in critical software worldwide. The model has already detected thousands of exploitable vulnerabilities across mainstream operating systems and browsers. Source

    Adobe introduces Student Spaces learning tool

    On April 7, Adobe officially launched Student Spaces, an AI-powered learning suite designed for education. The tool allows users to upload PDFs, documents, presentations, spreadsheets, web links, handwritten notes, and transcripts to automatically generate flashcards, mind maps, quizzes, study guides, and editable presentations powered by Adobe Express. It also integrates the AI podcast feature introduced last month, enabling users to convert study materials into listenable audio content.

    Student Spaces is currently in free testing. Users can access it via a standalone URL without logging in, and it supports both the Acrobat web version and mobile browsers. Source

    360doc to shut down service on May 1

    On April 1, the 360doc Personal Library website announced it will cease operations, stating that it was unable to find a suitable party to take over the service while ensuring data security and continuity. All services will officially shut down on May 1, 2026. Starting immediately, 360doc has suspended new content publishing, but login, data backup, VIP refund requests, and wallet withdrawal functions remain available. Source

    Cyberpunk 2077 receives PS5 Pro enhanced update

    On April 7, CD PROJEKT RED announced via an official blog post a free PS5 Pro enhanced update for Cyberpunk 2077. This update introduces support for PSSR upscaling technology and implements BVH8 architecture optimized for PS5 Pro hardware, improving the accuracy of ray-traced lighting, shadows, and reflections. The game now features three rendering modes: the Ray Tracing Pro mode enables all ray tracing effects, including ambient occlusion, global illumination, and emissive lighting, targeting 40fps on VRR displays (30fps on standard displays); the Performance mode maintains high visual fidelity while using PSSR to boost frame rates up to 90fps (VRR required); and the Ray Tracing mode serves as a balanced option, preserving some ray tracing enhancements while delivering a stable 60fps experience.

    The update is available starting today, and the game has also been added to the PlayStation Plus catalog. Source

    News Worth a Quick Look

    • Some users have discovered that the new Copilot app introduced by Microsoft on Windows 11 is למעשה based on the Microsoft Edge browser. By simply renaming mscopilot.exe to msedge.exe and renaming its folder from Copilot to Edge, users can launch Copilot as Microsoft Edge even if both Microsoft Edge and Edge WebView2 have been uninstalled. Source
    • On April 7, Google announced the introduction of native vertical tabs in the Chrome browser. After updating to the latest stable version, users can right-click the window and select “Move tabs to the side” to enable the feature. This update also includes a revamped Reading Mode, offering a more focused, distraction-free text reading experience. Source
    • On April 6, The New Yorker published a 16,000-word investigative report by Ronan Farrow and Andrew Marantz. Based on interviews with more than 100 sources and private memos from figures such as Ilya Sutskever and Dario Amodei, the report alleges systematic deception by OpenAI CEO Sam Altman, including withholding safety protocol details from the board, forming a secret agreement with Greg Brockman to bypass board oversight, and failing to fulfill commitments to allocate compute resources to the Superalignment team. Source
    • According to a recent analysis cited by Ars Technica from The New York Times, Google’s AI Overviews feature has an error rate of around 10% when answering queries. This means that roughly one in ten responses may contain incorrect information. While the system is accurate about 90% of the time, Google’s massive user base means that even a 10% error rate results in hundreds of thousands of incorrect responses being delivered to users worldwide every minute. Source
  • SSPAI Morning Brief: Apple Unveils M5 Chips and New MacBooks as Google Rolls Out Pixel Feature Drop

    SSPAI Morning Brief: Apple Unveils M5 Chips and New MacBooks as Google Rolls Out Pixel Feature Drop

    Morning Brief

    1. Apple unveils M5 series chips, Studio Display, and new MacBook models
    2. Unihertz releases the new full-keyboard smartphone Titan 2 Elite
    3. Four AI updates
    4. Google releases the Spring 2026 Pixel Feature Drop
    5. Speedtest and Downdetector are being sold
    6. Chrome adjusts the update cycle for the stable version
    7. Highguard announces shutdown on March 12
    8. Briefs worth a quick look

    Apple unveils M5 series chips, Studio Display, and new MacBook models

    On March 3, Apple introduced the M5 series chips and simultaneously refreshed the MacBook Pro, MacBook Air, and Studio Display / Display XDR product lines.

    The M5 series adopts a Fusion architecture that integrates two 3-nm chiplets into a single SoC. The M5 Pro and M5 Max feature a new 18-core CPU architecture composed of 6 super cores and 12 performance cores. Multi-thread performance improves by up to 30% compared with the M4 series and up to 2.5× compared with the M1 series. GPU configurations scale up to 40 cores, with each core integrating neural network accelerators. Peak AI compute performance improves by more than 4× over the previous generation, while ray tracing performance increases by up to 35%. In terms of memory, M5 Pro supports up to 64GB with 307GB/s bandwidth, while M5 Max supports up to 128GB with 614GB/s bandwidth. The M5 series also integrates a 16-core Neural Engine, a media engine supporting AV1 decoding, a Thunderbolt 5 controller, and memory security protection features. Source

    New MacBook Pro models powered by the M5 series start at RMB 17,999 for the 14-inch M5 Pro version and RMB 21,999 for the 16-inch model; the 14-inch M5 Max version starts at RMB 29,999 and the 16-inch version at RMB 31,999. Orders open on March 4, with official sales beginning on March 11. Source

    Apple also introduced 13-inch and 15-inch MacBook Air models powered by the M5 chip, featuring a 10-core CPU and up to a 10-core GPU. Base storage starts at 512GB and can be configured up to 4TB. The new generation SSD delivers twice the read and write performance. In terms of connectivity, the devices debut Apple’s N1 wireless chip, supporting Wi-Fi 7 and Bluetooth 6, and include two Thunderbolt 4 ports capable of driving up to two external displays. The machines use a fanless silent design and feature a 12MP Center Stage camera and Liquid Retina display, with battery life of up to 18 hours. The new MacBook Air is available in Sky Blue, Midnight, Starlight, and Silver, with the 13-inch model starting at RMB 8,499 and the 15-inch model at RMB 9,999. Orders open on March 4, with sales beginning March 11. Source

    Finally, the new Studio Display features a 27-inch 5K Retina display with brightness up to 600 nits. The Studio Display XDR upgrades to a 27-inch 5K Retina XDR panel using mini-LED backlighting with 2,304 local dimming zones. It supports a 120Hz ProMotion adaptive refresh rate (47Hz–120Hz), SDR brightness up to 1000 nits, HDR peak brightness up to 2000 nits, a contrast ratio of 1,000,000:1, and newly adds support for the Adobe RGB color gamut. Both displays include a 12MP Center Stage camera with Desk View support, a studio-quality three-microphone array, and a six-speaker system supporting spatial audio. For connectivity, the lineup adds Thunderbolt 5 downstream support, providing two Thunderbolt 5 ports and two USB-C ports, and supports daisy-chaining up to four displays. Each unit includes a Thunderbolt 5 Pro cable. Studio Display supports up to 96W power delivery, while the XDR version supports up to 140W reverse charging.

    The Studio Display with a tilt-adjustable stand starts at RMB 11,999, while the Studio Display XDR with tilt- and height-adjustable stand starts at RMB 24,999. Orders open on March 4, with sales beginning March 11. Source

    Unihertz releases the new full-keyboard smartphone Titan 2 Elite

    On March 3, Unihertz announced the Titan 2 Elite, a new full-keyboard smartphone in its Titan series, at MWC 2026. The device adopts a compact candy-bar design reminiscent of classic BlackBerry phones and comes in black and orange color options. It features a 4.03-inch AMOLED display with a 120Hz refresh rate, and the secondary rear display from the previous generation has been removed to achieve a slimmer and lighter design. The phone is powered by the MediaTek Dimensity 7400 processor, with an alternative Dimensity 8400 Pro version also available, and includes 12GB of RAM and 512GB of storage. The 3.5mm headphone jack has been removed, while the programmable side button and infrared transmitter remain.

    On the software side, the Titan 2 Elite ships with Android 16. Unihertz promises security updates through 2031 and Android version upgrades up to Android 20. The Titan 2 Elite will launch on Kickstarter for crowdfunding in March this year, with pricing yet to be announced. Source

    Four AI Updates

    OpenAI Releases GPT-5.3 Instant
    On March 3, OpenAI released GPT-5.3 Instant. The new model improves response relevance, conversational fluency, and web search capabilities, while significantly reducing mechanical replies, redundant disclaimers, and overly stiff tone tendencies. Compared with GPT-5.2 Instant, it particularly improves responses involving sensitive topics, addressing previous issues where the model could appear overly cautious or overly preachy. It also enhances the integration quality of web search results, reduces excessive reliance on search outputs, and resolves earlier cases where responses occasionally contained “link dumping” or loosely organized information. The update also substantially adjusts dialogue styles that users previously described as “awkward,” “out of place,” or overly forceful in tone, while significantly reducing hallucination rates across a wide range of topics.

    GPT-5.3 Instant is available immediately to all ChatGPT users. Developers can access the model through the API under the name gpt-5.3-chat-latest. Updates for the Thinking and Pro variants are expected to roll out soon. GPT-5.2 Instant will remain available for paid users under the “Legacy Models” dropdown menu for three months and is scheduled to be officially retired on June 3, 2026. Source

    Google Releases Gemini 3.1 Flash-Lite
    On March 3, Google officially launched Gemini 3.1 Flash-Lite. The model features a context window of up to one million tokens and supports multimodal inputs including text, images, audio, and video. It delivers a maximum output speed of 389 tokens per second. Pricing is set at $0.25 per million input tokens and $1.50 per million output tokens. The model’s knowledge cutoff date is January 1, 2025.

    A preview version of Gemini 3.1 Flash-Lite is now available to developers via Google AI Studio and the Vertex AI platform. Source

    Anthropic Introduces Memory Import Tool
    On March 3, Anthropic announced that the memory feature of its AI assistant Claude is now available to free users. The feature is implemented through standardized prompts: users can copy a specific prompt provided by Anthropic into other AI chat tools and request them to export all stored user preferences, personal details, project goals, and behavioral adjustments. The returned structured text can then be pasted into Claude’s memory settings to complete the import and integration process. Source

    Alibaba Releases Qwen3.5 Small Model Series
    On March 2, Alibaba’s Qwen team released the Qwen3.5 Small model series, including four parameter sizes: 0.8B, 2B, 4B, and 9B. Built on the Qwen3.5 architecture, the series features native multimodal capabilities and has been optimized through large-scale reinforcement learning (RL). The 0.8B and 2B versions focus on high-speed inference for edge devices, the 4B version is positioned as a lightweight multimodal foundation for AI agents, and the 9B version aims to narrow the performance gap with larger models. Base models are also provided across the entire lineup. The weights have been released on Hugging Face and ModelScope, where they are freely available for research and industrial applications. Source

    Google releases the Spring 2026 Pixel Feature Drop

    On March 3, Google rolled out the March Pixel Drop for Pixel devices. The update adds real-time location sharing in Google Messages to the Find Hub tracking network and introduces lost luggage tracking in partnership with 10 airlines (requiring compatible tags). A customizable Calling Card feature is also introduced, allowing users to personalize their call screens with photos and fonts, while the Google Play Store adds a short-video feed to enhance app discovery. Pixel 10 series–exclusive Magic Cue now supports restaurant recommendations, and the Circle to Search feature expands with enhanced virtual try-on and video shopping capabilities. The home screen also gains six AI-generated icon styles. Pixel Watch receives several updates, including phone-loss alerts, quick payments without waking the device, earthquake alerts, and satellite SOS coverage across Europe, Canada, Hawaii, and Alaska. These updates will roll out gradually starting today. Source

    Speedtest and Downdetector are being sold

    On March 3, digital media company Ziff Davis announced that it will sell its Connectivity business unit, which includes Ookla Speedtest and Downdetector, to Accenture for $1.2 billion in cash. The division generated $231 million in revenue in 2025, representing a substantial return compared with the $15 million Ziff Davis paid to acquire it in 2014. Ziff Davis stated that the deal is intended to sharpen the company’s strategic focus, with future efforts centered on core media brands such as IGN, Mashable, and Everyday Health. The transaction is expected to close in the coming months, and both Speedtest and Downdetector will continue operating normally during the transition period. Source

    Chrome adjusts the update cycle for the stable version

    On March 3, Google announced that the update frequency for the Chrome browser’s stable release across all platforms will change from once every four weeks to once every two weeks. The adjustment applies to desktop, Android, and iOS platforms, aiming to accelerate feature delivery and simplify post-release debugging. The first stable version under the new schedule will be Chrome 153, scheduled for release on September 8. Beta versions will be released three weeks ahead of the stable version, while weekly security updates for milestone versions will remain unchanged. The Canary and Dev channels, as well as the extended eight-week stable release cycle for enterprise users, will continue as before.

    For Chromebook hardware, ChromeOS will follow the new cadence after completing platform-specific testing. Further details regarding update policies for managed devices will be announced later. Source

    Highguard announces shutdown on March 12

    On March 3, developer Wildlight Entertainment announced that its shooter game Highguard will officially shut down on March 12. The title was developed by a team composed of former members from Apex Legends, Call of Duty, and Titanfall, and was showcased as the closing highlight at The Game Awards 2025. Since its release on January 26, 2026, the game once attracted up to 2 million players but ultimately failed to establish a player base large enough to sustain long-term operation. The studio is reportedly partially funded by Tencent and laid off most of its staff in February. The game will receive its final update over the next two days, including a new character and new weapons. Source

    Briefs worth a quick look

    • Swedish media outlet Svenska Dagbladet reported that Meta AI smart glasses (such as Ray-Ban Meta) allegedly captured users’ private videos and financial information without their knowledge and sent the data to reviewers outside the European Union. The claim was revealed by an AI data annotator in Nairobi, Kenya, who said that while processing visual data used to train large language models (LLMs), they encountered highly sensitive content including nudity, people using the toilet, intimate behavior, and even credit card numbers. Meta declined to comment on the report. Source
    • Apple’s official regulatory compliance page briefly listed a previously unknown device called MacBook Neo (Model A3404), identified as part of the 2026 MacBook lineup, though the page has since been removed. According to earlier rumors, the device is expected to target the education and consumer entry-level market and may use an iPhone-series processor rather than an M-series chip. Source
  • SSPAI Morning Brief: Samsung Brings Perplexity AI to Galaxy as PayPal Confirms App Data Breach

    SSPAI Morning Brief: Samsung Brings Perplexity AI to Galaxy as PayPal Confirms App Data Breach

    Morning Brief

    1. Samsung to Bring Perplexity AI to Galaxy Phones
    2. Linux Kernel 7.0-rc1 Released
    3. Firefox 115 ESR to End Support for Windows 7/8 This Month
    4. PayPal Confirms Data Breach in Its App
    5. Google Releases AI Music Generation Model Lyria 3
    6. Microsoft Edge to Remove Custom Master Password Feature
    7. Rumors You Can Just Glance At

    Samsung to Bring Perplexity AI to Galaxy Phones

    On February 22, Samsung Mobile announced that Galaxy AI is further expanding its multi-AI agent ecosystem to deliver more diverse and flexible user experiences. Perplexity AI will be introduced to flagship Galaxy devices, and Samsung stated that it plans to collaborate with more partners in the future to jointly develop Galaxy AI into a more inclusive AI ecosystem. Samsung has also adapted a new wake word, “Hey Plex,” for Perplexity AI. Users can alternatively long-press the side button to activate it. Similar to Bixby, it can interact with Samsung’s first-party and some third-party apps for system-level integration. Specific supported models will be announced soon. Source

    Linux Kernel 7.0-rc1 Released

    Linux kernel version 7.0-rc1 was officially released on February 22, bringing multiple changes and new features. In terms of new hardware support, it adds compatibility for Intel Nova Lake and Diamond Rapids processors, AMD Zen 6, and Qualcomm Snapdragon X2. Performance improvements are also notable, including better PostgreSQL performance on AMD EPYC, improved sequential read performance for exFAT, various F2FS file system enhancements, memory management optimizations, improved concurrent direct I/O writes in EXT4, default enablement of Intel TSX auto mode, scheduler performance and scalability work, and the return of large page memory support for Nouveau to boost NVK performance. Source

    Firefox 115 ESR to End Support for Windows 7/8 This Month

    On February 18, Mozilla updated its official support documentation, confirming that it will end maintenance support for the Firefox 115 ESR (Extended Support Release) browser on Windows 7, Windows 8, and Windows 8.1 by the end of this month (February 2026). This makes Mozilla the last major browser vendor to drop support for these legacy systems, following Chrome and Edge. Mozilla had originally planned to end support for Firefox 115 ESR in September 2024. Source

    PayPal Confirms Data Breach in Its App

    PayPal recently confirmed that its PayPal Working Capital app experienced a data breach caused by a code vulnerability. The flaw existed for six months, from July 1, 2025 to December 13, 2025. At the time of the announcement, PayPal had contacted approximately 100 potentially affected customers whose names, Social Security numbers, dates of birth, email addresses, phone numbers, and business addresses were exposed to unauthorized access. Some users have already reported discovering unauthorized transfer records in their accounts. PayPal Working Capital is a lending service app under PayPal that provides small business loans based on merchants’ transaction histories. Source

    Google Releases AI Music Generation Model Lyria 3

    On February 19, Google released a new AI music generation model named Lyria 3. The model can create high-fidelity, 30-second audio tracks based on text, image, audio, or video prompts, supporting a wide range of styles—from short ad jingles and lo-fi beats to more complex arrangements and full songs. Users can control genre, vocals, and rhythm through prompts. Lyria 3 can also automatically generate lyrics based on the prompt, and when paired with Nano Banana, it can create matching album artwork. The model currently supports English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with more languages planned for future updates. It has already been integrated into the Gemini app, where users can access it by selecting the “Music” option from the “Tools” dropdown menu. Google has also released a prompt guide for Lyria 3 to help users get started quickly. Source

    Microsoft Edge to Remove Custom Master Password Feature

    According to the Edge Beta channel update log, starting from Edge 146, the custom master password feature in the Edge password manager will be removed, and users will no longer be able to create new custom master passwords as a security measure for saved credentials. Users who have already enabled this feature will receive a warning notification, but it will remain available until June 4, 2026. After that, affected users will be automatically migrated to device-based authentication methods such as Windows Hello.

    Additionally, in this update, the “Passwords” option will be removed from the “Clear browsing data” tab to prevent users from accidentally deleting saved account credentials when clearing browser data. Source

    Rumors You Can Just Glance At

    • On February 23 (local time), Anthropic posted on X alleging that DeepSeek, Moonshot AI, and MiniMax used more than 24,000 fake accounts to conduct large-scale distillation attacks on its models, generating 150,000, 3.4 million, and 13 million interactions with Claude respectively to improve their own model performance. Source
    • According to 9to5Google, Google is adjusting the experience of the built-in Google Weather service on Android. Previously, tapping the Google Weather widget would open a full-screen weather information interface that resembled a standalone app, but was actually a dedicated page within the Google app. Some users who updated the Google app have found that this weather page has been removed. Now, tapping the widget redirects directly to the Google app’s search page, where weather information is presented in card form as part of search results. In addition to traditional forecasts, air quality data, and detailed weather metrics, AI-generated forecast summaries will also be included, aligning the experience more closely with standard web search results. Notably, Pixel users are unaffected as they rely on the Pixel Weather app. Source
    • According to Game File, Tencent has shut down its TiMi Montreal studio. In addition to reports from internal sources, LinkedIn posts by former employees (now deleted) have corroborated the development. The studio was established in July 2021 with the goal of developing AAA open-world multi-platform games. Source
    • On February 23, Uber announced the acquisition of SpotHero, an app that offers parking reservation services in crowded areas. Financial details of the deal were not disclosed. Uber stated that it plans to integrate parking services directly into its own app. Source
    • According to Android Authority, Google is developing a facial recognition system called Project Toscana for future use on Android phones and Chromebooks. Test devices equipped with Project Toscana reportedly require only a circular camera cutout. The system is said to function reliably under varying lighting conditions, similar to Face ID, though the underlying technology remains unclear and is currently speculated to involve infrared cameras. Source
    • According to Bloomberg’s Mark Gurman, due to the global popularity of the “Starlight Orange” color this year—especially in the Chinese market—Apple is considering introducing a deep red color option for the iPhone 18 Pro. Previous rumors of purple or brown variants may have been testing adjustments for this color. The iPhone Fold, however, is expected to avoid bold colors, offering only dark gray or black, as well as white or silver options. Source
    • Following a U.S. Supreme Court ruling that most of last year’s Trump-era tariff policies were unlawful, Apple—which had previously lobbied for exemptions on many products—may now face a uniform 10% tariff, potentially higher than before. Source
    • In Windows 11 Insider Preview Build 26300.7877, the Paint app (version 11.2601.391.0) introduces a free rotation feature, allowing any shape, text, or selection to be rotated at any angle. Source
  • SSPAI Morning Brief: SpaceX Acquires xAI, Adobe Ends Animate

    SSPAI Morning Brief: SpaceX Acquires xAI, Adobe Ends Animate

    Morning Brief

    1. SpaceX acquires xAI
    2. Adobe discontinues Animate
    3. Hypergryph responds to overseas payment issues in Arknights: Endfield
    4. Malicious skills found in the OpenClaw skill library
    5. ROG Kithara premium gaming headset announced and released
    6. Rumors You Can Just Glance At

    SpaceX acquires xAI

    On February 2, SpaceX announced its acquisition of Elon Musk’s artificial intelligence company xAI. The deal makes SpaceX the most valuable private company in the world, with a valuation reaching USD 1.25 trillion. In the acquisition memo, Elon Musk stated that the primary goal of this acquisition is to build space-based data centers, aiming to avoid the power consumption and cooling challenges associated with running AI in large terrestrial data centers. Source

    Adobe discontinues Animate

    On February 2, Adobe announced that it will discontinue Adobe Animate effective March 1, 2026, and remove the application from its official website, after which users will no longer be able to purchase it. Enterprise customers will be able to continue using the app, download content, and receive technical support until March 1, 2029. Other users will be able to keep using the app, download content, and access technical support until March 1, 2027. Adobe Animate was previously known as Adobe Flash Professional / Macromedia Flash and has been used to create vector graphics and animations. In its announcement, Adobe stated that due to technological advancements and the emergence of new platforms and paradigms, better applications are now available to support users’ creative needs. The company recommends After Effects or Adobe Express as alternatives. Source

    Hypergryph responds to overseas payment issues in Arknights: Endfield

    On February 1, Hypergryph responded to Southern Metropolis Daily regarding a serious PayPal payment interface error that occurred on January 22, the launch day of Arknights: Endfield, affecting overseas servers. According to Hypergryph, when the payment issue occurred, most overseas regions were outside peak gaming hours. Upon receiving user feedback, the company promptly disabled the PayPal payment option, and the overall response was considered timely. In total, 3,429 orders were affected, with the total amount involved not exceeding USD 80,000. Fewer than 1,800 users were impacted, and the highest abnormal charge for a single user reached USD 5,289. As of that day, Hypergryph had completed full refunds for all affected users. Source

    Malicious skills found in the OpenClaw skill library

    According to OpenSourceMalware, between January 27 and February 1, two waves totaling more than 230 malicious skills were published to OpenClaw (formerly ClawdBot/MoltBot) skill repositories, including ClawHub and GitHub. These malicious skills primarily target cryptocurrency traders, disguising themselves as legitimate crypto trading automation tools while stealing and exfiltrating sensitive information on Windows and macOS. All of the malicious skills share the same underlying strategy and rely on sophisticated social engineering to persuade users to execute harmful commands. Once run, these commands steal trading API tokens, wallet private keys, SSH credentials, browser passwords, and other sensitive data. Source

    ROG Kithara premium gaming headset announced and released

    On February 2, the ROG Kithara premium gaming headset was officially released and made available for purchase. Developed in collaboration with audio manufacturer HIFIMAN, the headset features a 100 mm HIFIMAN planar magnetic driver tuned by ROG, along with an open-back design. It comes with a balanced, interchangeable headphone cable, making it compatible with DACs, amplifiers, game consoles, and PCs. The headset is also equipped with a full-bandwidth inline MEMS microphone. The launch price is CNY 1,999. Source

    Rumors You Can Just Glance At

    • Tencent announced on February 2 that QQ Classic Farm will officially return on February 6, launching simultaneously on QQ and WeChat. Based on current official teasers, this new release will include a mobile version and largely retain the gameplay of QQ Farm, which first launched in 2009. Players take on the role of a farmer—planting crops, raising animals, managing agricultural products—and can visit friends’ farms to “steal crops” or “plant weeds” as mischievous interactions. Source
    • Baidu’s autonomous ride-hailing service Apollo Go (Luobo Kuaipao) officially announced it will launch in Qingdao at 9:00 a.m. on February 3, with operations limited to the Hongdao area of Chengyang District. Available vehicle options include Arcfox models and the Luobo 06, featuring a “business-class ultra-spacious cabin,” 135-degree intelligent ten-point massage heated seats, and support for VIP music streaming. Users can hail rides by logging into their Baidu account via the Apollo Go app or mini program, and will receive an SMS notification when the vehicle arrives. Source
    • According to The Wall Street Journal, Apple is considering shifting the production of some lower-end processors away from TSMC to another supplier. The report does not name the alternative manufacturer, though prior rumors have suggested that Intel may begin supplying chips for Apple in 2027 or 2028, potentially including processors for some Mac and iPad models. Source
    • According to DigiTimes, Apple’s rumored plan to enter the smart glasses market later this year has already begun to impact the AR optics supply chain. The report says multiple Taiwanese AR optics suppliers are investing to expand capacity and shift R&D focus. Kinko Optical has reportedly invested around USD 5.6 million to develop a new AR/VR/MR research center, while companies such as Asia Optical, MetaOptics, Chung Yang Optics, and Zhihui Innovation are also accelerating related product development. Source
    • Recently, media outlets discovered that a Google app called Desktop Camera was listed on the Google Play Store on January 28. The app description suggests a minimalist UI, closely resembling ChromeOS or the Pixel Camera interface. It is currently speculated that this app may be a system component of the previously rumored “desktop-style Android interface,” Aluminium OS. Source
  • SSPAI Morning Brief: OpenAI Introduces Prism, a New AI Platform for Scientific Research

    SSPAI Morning Brief: OpenAI Introduces Prism, a New AI Platform for Scientific Research

    Morning Brief

    1. Sennheiser releases CX 80U and HD 400U wired earphones
    2. Sonos releases multi-channel power amplifier Amp Multi
    3. Bigme unveils B10 color E Ink office notebook
    4. Three updates related to Google Gemini
    5. Yahoo launches AI Q&A search tool Yahoo Scout
    6. Alibaba releases Qwen3-Max-Thinking
    7. Anthropic launches interactive Claude app
    8. OpenAI releases scientific research AI collaboration space Prism
    9. Rumors You Can Just Glance At

    Sennheiser releases CX 80U and HD 400U wired earphones

    On January 26, Sennheiser officially unveiled the CX 80U wired in-ear earphones and the HD 400U wired over-ear headphones. As upgraded versions of the CX 80S and HD 400S, both new models replace the traditional 3.5 mm jack with a USB-C connector and feature a built-in digital converter that supports 24-bit/96 kHz lossless audio playback, along with an in-line remote equipped with an integrated MEMS microphone.

    The CX 80U uses a dynamic driver with an impedance of 28 Ω and a frequency response of 17 Hz–20 kHz, and comes with three sets of silicone ear tips in different sizes. The HD 400U features a 32 mm dynamic driver with an impedance of 18 Ω and a frequency response of 18 Hz–20 kHz, a foldable closed-back design, and includes a carrying pouch. Both products went on sale on January 26, 2026, with the CX 80U priced at USD 40 and the HD 400U at USD 100. Source

    Sonos releases multi-channel power amplifier Amp Multi

    On January 27, Sonos officially introduced its first professional-grade multi-channel streaming power amplifier, Amp Multi, designed specifically for large residential properties and complex custom installation scenarios. Amp Multi features a gallium nitride (GaN) power architecture and Class D post-filter feedback (PFFB) technology, delivering eight channels of 125 W (8 Ω) each in a compact 1.5U chassis, and can be configured as four independently controlled zones. Its hardware configuration includes a quad-core 1.9 GHz A55 processor, Wi-Fi 6, and dual Ethernet ports. It supports Apple AirPlay 2 and Spotify Connect, and each channel can drive up to three Sonos Architectural series speakers. Pricing has not yet been announced, and the product is expected to go on sale in the coming months through authorized installers. Source

    Bigme unveils B10 color E Ink office notebook

    On January 27, Bigme launched the 10.3-inch B10 4G color E-Ink office tablet on overseas platforms. The device is powered by a Dimensity 1080 processor and comes with 8 GB of RAM, 256 GB of storage, and Android 14. It supports 4G connectivity across major bands and is equipped with a 6,900 mAh battery, an eight-microphone array, a 4,096-level pressure-sensitive stylus, and fingerprint recognition. The B10 features a Kaleido 3 color E-Ink display with up to 30 FPS refresh rate and 36 levels of adjustable dual-tone (warm/cool) color temperature. It also includes both front and rear cameras, supporting document scanning, OCR text recognition, and video conferencing.

    The Bigme B10 adopts a design with eight customizable touch keys and comes with four built-in AI models, supporting features such as meeting speech-to-text transcription. An optional keyboard accessory is available. The B10 is priced at USD 699, with a pre-sale price of USD 594. Source

    Three updates related to Google Gemini

    On January 27, Google announced the introduction of Agentic Vision capabilities to the Gemini 3 Flash model. This feature aims to shift image understanding from static processing to a dynamic process based on a “Think, Act, Observe” loop. By combining visual reasoning with Python code execution tools, the model can autonomously perform actions such as zooming in to inspect details, rotating and cropping images, drawing bounding-box annotations, conducting visual math calculations, and generating charts. Technical tests show that with agentic vision enabled, Gemini 3 Flash achieves a 5–10% quality improvement across most visual benchmarks and partially mitigates hallucination issues in long-form visual tasks. The feature is now officially available via the Gemini API (including Google AI Studio and Vertex AI), where developers can enable Code Execution in the toolbar to use it. Source

    At the same time, Google also announced that Gemini 3 has been officially set as the default underlying model for its global “AI Overviews” feature, and that it has introduced a seamless “AI Mode” conversational experience. This allows users to ask follow-up questions directly below AI Overviews, turning search from one-way information display into contextual, streaming conversations. Google said that the model upgrade also improves reasoning quality for complex questions. The new AI Overviews and AI Mode conversation features are now available globally on both mobile and web search, with the basic service offered free to all users. Source

    Finally, Google announced that its entry-level AI subscription plan, Google AI Plus, has officially launched in 34 additional countries and regions worldwide, including the United States. Priced at USD 8 per month, the plan is aimed at everyday users and provides access to the Gemini 3 Pro model, Deep Research, and the Nano Banana Pro model. Subscribers also receive 200 GB of Google One storage and unlock AI filmmaking tool Flow, image-to-video tool Whisk, and advanced research features in NotebookLM (including support for up to 100 input sources and audio overviews). Notably, existing Google One Premium 2 TB members will automatically receive all AI Plus features, and new subscribers can enjoy a promotional price of USD 4 per month for the first two months. Source

    Yahoo launches AI Q&A search tool Yahoo Scout

    On January 27, Yahoo announced the launch of its AI-powered Q&A search feature, Yahoo Scout, which is supported by Anthropic’s Claude. The feature deeply integrates 50 billion user profiles and 18 trillion consumer behavior signals accumulated by Yahoo over the past three decades. It supports synthesizing information from across the web—including Yahoo’s own channels such as News, Finance, and Sports—using natural language, and presents the results in the form of structured tables, rich multimedia cards, and prominent blue source links. Each answer contains an average of up to nine cited sources. At the same time, the feature has been embedded across Yahoo’s entire product lineup, offering enhanced experiences such as email summaries, real-time financial market analysis, news comment trend extraction, and in-depth breakdowns of sports events.

    Yahoo Scout is currently available in beta in the United States, with access via a standalone web version as well as iOS and Android apps. The service remains free to use and is planned to be monetized through bottom-page advertising and affiliate commissions. Source

    Alibaba releases Qwen3-Max-Thinking

    On January 26, Alibaba released its flagship reasoning model Qwen3-Max-Thinking, with total parameters exceeding one trillion. The model underwent large-scale reinforcement learning post-training and has enhanced its native agent capabilities for autonomously calling tools across multiple key benchmarks, enabling on-demand use of search engines and code interpreters. It also supports test-time scaling techniques, significantly improving reasoning performance. On major reasoning benchmarks, it surpasses Gemini 3 Pro, while also substantially reducing model hallucinations.

    Qwen3-Max-Thinking is now available on Qwen Chat, and its corresponding API (model name: qwen3-max-2026-01-23) has been released via Alibaba Cloud. Source

    Anthropic launches interactive Claude app

    On January 26, Anthropic announced the launch of an interactive app for its Claude chatbot, allowing users to directly connect everyday workplace tools such as Slack, Canva, Figma, Box, and Clay within the chat interface. Salesforce integration will be added later. Claude can send Slack messages on users’ behalf, generate charts, or access cloud files within these services, depending on which app permissions are enabled.

    In a blog post, Anthropic said that tasks such as data analysis, content design, and project management are inherently better suited to dedicated visual interfaces. When combined with Claude’s intelligence, users can significantly improve efficiency in ideation, revision, and iteration. The new feature is currently available to Claude Pro, Max, Team, and Enterprise subscribers, and can be enabled and managed at claude.ai/directory. Source

    OpenAI releases scientific research AI collaboration space Prism

    On January 27, OpenAI officially unveiled Prism, an AI-native collaborative workspace designed specifically for scientific research, and simultaneously announced the acquisition of the cloud-based LaTeX platform Crixet as its underlying infrastructure. Prism deeply integrates the GPT-5.2 Thinking model, enabling AI to directly understand paper structures, formulas, and references within project environments. It supports automatically generating TikZ diagrams, retrieving and citing papers from arXiv, auto-formatting bibliographies, and converting handwritten formulas or sketches on whiteboards into LaTeX code. In addition, Prism offers real-time multi-user collaboration and can generate graduate-level course plans and accompanying exercises based on specific research topics.

    Prism is currently available for free to all users with a personal ChatGPT account, supporting the creation of unlimited projects and collaboration seats. It will also be rolled out soon to ChatGPT Business, Team, Enterprise, and Education plans. Source

    Rumors You Can Just Glance At

    • On January 28, Samsung began teasing a new privacy protection feature that will be introduced on Galaxy devices. The feature is designed to address the risk of “shoulder surfing” in public places through deep integration of hardware and software. It is said to adjust display visibility at the pixel level to selectively obscure password entry, specific app interfaces, and notification pop-ups, allowing only the person directly in front of the screen to clearly view certain content. Source
    • On January 26, Meta announced plans to begin testing premium subscription services for its three major apps—Instagram, Facebook, and WhatsApp—in the coming months. The subscription system aims to unlock more powerful productivity, creativity, and AI-enhanced features through paid plans, while keeping the core basic experience free. Source
    • According to reports from several media outlets and a settlement agreement recently submitted to a federal court in San Jose, Google has agreed to pay USD 68 million to settle a class-action lawsuit over illegal recordings made by its Google Assistant voice assistant. The case centers on Google Assistant’s “false accept” mechanism, in which devices are activated and capture sensitive user audio for human review or targeted advertising even without hearing the “OK Google” wake word. This is similar in nature to Apple’s previous settlement of USD 95 million over comparable issues with Siri. The settlement covers U.S. users who, since May 18, 2016, have had accounts associated with devices preinstalled with Google Assistant (including Pixel phones, Google Home smart speakers, and Nest Hub displays). Under the agreement, eligible device purchasers are expected to receive compensation of USD 18 to USD 56, while affected individuals living in environments with such devices may claim USD 2 to USD 10. Source
    • According to internal documents obtained by the New Mexico Attorney General’s Office, Meta CEO Mark Zuckerberg personally led and vetoed a decision to add parental control features to the company’s AI chatbots, instead pushing for looser safety restrictions. The allegations stem from a lawsuit filed by the state of New Mexico accusing Meta of failing to protect children from sexual harassment. A previous investigation by The Wall Street Journal had also pointed out risks that Meta’s AI bots could engage in sexually explicit conversations with minors or impersonate them. As a result, Meta suspended access to AI character features for teen accounts last week while developing relevant parental control tools. Source
    • On January 27, Google accidentally leaked screenshots of a desktop interface codenamed Aluminum OS (ALOS) on the Chromium bug tracker. The images came from an HP Elite Dragonfly Chromebook powered by a 12th-generation Intel Core processor, running system version ZL1A.260119.001.A1 based on Android 16. The screenshots show a desktop environment optimized for large-screen multitasking, featuring a taskbar, second-level time display, an M3E battery icon, a Gemini AI entry point, and a screen recording control panel. The bundled Chrome browser also supports desktop extensions. This further confirms earlier rumors that Google is testing a deeply integrated version of Android and ChromeOS on Chromebooks. Source