WeChat Official Accounts roll out “paid boosting” feature
Ubiquiti launches AC1200 travel router UTR
Logitech releases G304 X LIGHTSPEED mouse
China Media Group releases its Top 10 Domestic and International Science & Technology News of 2025
NVIDIA reaches licensing agreement with AI chip startup Groq
Google lists Android Mini Bot Maze figurine
Rumors You Can Just Glance At
WeChat Official Accounts roll out “paid boosting” feature
On December 24, WeChat Official Accounts officially rolled out the “Paid Boosting” feature. Creators can access the “Paid Boosting” module in the Official Accounts backend and choose already published content for promotional boosting. Specifically, the feature allows creators to either increase readership or gain new followers. This service is paid and purchased using WeChat Beans. For individual creators with fewer than 10,000 followers, a free “Content Boost” feature is available. Together, “Paid Boosting” and “Content Boost” form a complementary traffic growth system that combines paid promotion with free incentives. Source
Ubiquiti launches AC1200 travel router UTR
On December 22, Ubiquiti announced the launch of the UTR Travel Router (UniFi Travel Router). This model allows users to sync their home network configuration while traveling, ensuring a consistent connectivity experience on the go. The UTR travel router measures 95.95 × 65 × 12.5 mm, weighs 89 g, has no built-in battery, and is powered via USB-C. It features a 1.14-inch status display. The router provides two 1GbE RJ45 uplink Ethernet ports, supports AC1200 Wi-Fi, and includes a USB-C port that supports USB network sharing. The official price is USD 79. Source
Logitech releases G304 X LIGHTSPEED mouse
On December 24, Logitech China listed the G304 X LIGHTSPEED mouse on its official website. The mouse retains the classic G304 design, but switches from replaceable batteries to a built-in 290 mAh lithium battery, with a claimed battery life of 106 hours. It uses a USB-C port for charging, weighs 57 g, and features a centered weight balance. The mouse supports both 2.4 GHz wireless and wired connections, is equipped with the HERO 25K sensor and the new-generation LIGHTSPEED technology, and comes in black, white, purple, and pink color options. Pricing has not yet been announced. Source
China Media Group releases its Top 10 Domestic and International Science & Technology News of 2025
On December 24, China Media Group released its Top 10 Domestic and International Science & Technology News of 2025. Source
2025 Top 10 Domestic Science & Technology News:
China builds three world-class science and technology innovation centers
The State Council issues Opinions on Deepening the Implementation of the “AI+” Initiative
Tianwen-2 successfully launched, marking China’s entry into asteroid exploration and sample return
“Artificial Sun” fusion device sets a “hundred-million-degree for a thousand seconds” record
China’s large-scale AI model DeepSeek draws global attention
“Zuchongzhi-3” quantum computing prototype sets a new world record
China’s first electromagnetic catapult aircraft carrier Fujian commissioned, ushering the PLA Navy into the three–aircraft-carrier era
Shenzhou-22 spacecraft launched in an emergency mission, creating a historical milestone with “uncrewed” operations
A series of robotics competitions held as China’s embodied intelligence industry flourishes
2025 Top 10 International Science & Technology News:
China leads rapid global development of renewable energy
Most detailed connectivity map of the mammalian brain completed
Antimatter qubits demonstrated for the first time
AI-powered virtual cells open a new paradigm for biological research
China launches an international scientific program in the field of fusion
Gravitational wave signals verify Hawking’s black hole area theorem
World’s first scalable photonic quantum computer prototype unveiled
Artificial “form vision” research restores sight to blind patients
Scientists observe a massive black hole merger event
World Federation of Engineering Organizations General Assembly releases the Shanghai Declaration 2025
NVIDIA reaches licensing agreement with AI chip startup Groq
On December 24, AI chip startup Groq announced that it has entered into a non-exclusive licensing agreement with NVIDIA covering Groq’s inference technology. Under the agreement, NVIDIA will be able to use Groq’s technology in future chip designs, while several Groq executives—including its founder and president—will join NVIDIA. CNBC reported that the deal effectively amounts to NVIDIA acquiring Groq for $20 billion in cash, but NVIDIA denied this claim to other media outlets. Groq’s announcement also stated that the company will continue to operate independently, and that its GroqCloud business will remain in operation. Groq focuses on developing Language Processing Unit (LPU) chips designed to deliver high-performance, low-cost inference. Source
Google lists Android Mini Bot Maze figurine
Google’s official store has recently listed an Android a-MAZE-ing Collectible figurine. The figure stands about 3 inches (approximately 7.6 cm) tall, features a fully transparent shell, and contains a green marble maze structure inside. Users can tilt the figurine in different directions to guide a small metal ball through the maze. The robot’s arm joints are also articulated. The collectible is currently priced at $24. Source
Rumors You Can Just Glance At
According to The Information, OpenAI is discussing increasing the weight of sponsored advertising content within relevant Q&A responses in ChatGPT, as well as adding pop-up ads to the sidebar to better support shopping and similar use cases. Sources say OpenAI intends to remain “restrained” in its approach to advertising. Source
On December 24, according to Yicai, Bilibili stated that individuals who fabricated and spread false information claiming that “all Bilibili videos will require a paid membership to watch starting next year” have been apprehended and placed under administrative detention in accordance with the law. Bilibili also said it reserves the right to pursue further legal action against those responsible. Source
GTA copyright holder Take-Two Interactive has ordered DOS.Zone to take down an original GTA: Vice City save/load tool that allowed the game to be played directly in a web browser. Source
Microsoft Distinguished Engineer Galen Hunt, whose LinkedIn post previously drew widespread attention, has posted again to clarify rumors that “Microsoft will rewrite Windows in Rust to replace C and C++,” which were sparked by his earlier remarks. Source
Gmail’s Hindi-language help documentation has mentioned a feature allowing users to change their @gmail.com email address, noting that the change will be “gradually rolled out to all users.” Users will be able to change their Gmail address once every 12 months, up to a maximum of three times. Source
Around 1:00 a.m. local time, Steam store servers experienced a large-scale outage, with connection issues also affecting some Valve game servers. Services are currently being gradually restored. Source
Welcome to this edition of SSPAI Review. You can use the table of contents to quickly jump to the sections you’re interested in. If you’ve discovered other interesting apps or topics worth paying attention to, feel free to join the discussion in the comments.
New Apps Worth Paying Attention To
While SSPAI has long been dedicated to discovering and introducing high-quality apps across platforms, there are still many excellent apps—outstanding in design, functionality, interaction, and overall experience—that we haven’t yet covered. They might be long-standing apps or newly released ones, and we’ll introduce them to you here.
Awake: A New App From a Veteran Developer—You Must Complete Tasks to Turn Off the Alarm
Platform: iOS, iPadOS
Keywords: Smart alarm clock, daily briefing
@ElijahLee: Awake is a smart alarm clock app designed to make sure you actually wake up—by requiring you to complete tasks before you can turn the alarm off. It comes from the developer of the well-known scheduling app Structured. Awake is more than just a simple alarm clock; it’s a personalized morning assistant.
Getting started with Awake is very straightforward. When you first open the app, a short and simple onboarding flow teaches you how to use it to wake up. In the tutorial, you’re required to rotate your iPhone three full circles before moving on to the next step—this perfectly reflects Awake’s design philosophy: you must complete a task to dismiss the alarm. In the following steps, you configure your wake-up time, bedtime, and alarm cycle in just three steps to finish setting up the alarm. Thanks to iOS AlarmKit, alarms set via Awake have system-level priority and cannot be disabled.
Awake’s interface is clean and minimal. The app displays your upcoming bedtime, sleep, and wake-up times in a timeline (Schedules) view for the next few days, which you can browse by swiping up or down. The most creative feature is the set of tasks (Missions) required to turn off the alarm. When the alarm goes off, you can’t simply swipe it away—you have to complete a task. Awake offers two free tasks: rotating your iPhone and solving math problems. It also supports tasks via third-party apps such as the puzzle app Art of Fauna and the language-learning app Lengo. With a paid subscription, you can unlock additional tasks like turning on room lights, scanning a QR code, or walking a few steps. These tasks help activate your brain and shake off drowsiness more effectively.
The Morning Briefing feature displays a short summary after you dismiss the alarm, helping you start the day more prepared. You can customize the briefing content yourself; currently supported items include the current weather, calendar events, Structured tasks, and one sec Pomodoro sessions, with future plans for sleep analysis, daily podcasts, and more. This feature feels like a morning stand-up meeting—shifting you from passively waking up to actively planning your day.
You can download Awake for free from the App Store. Most features are available at no cost, while a paid subscription unlocks all alarm-dismiss tasks, all briefing integrations, all alarm sounds, support for multiple timelines, and more. Pricing is RMB 5.75 per month (early-bird price RMB 3.99/month), RMB 22 per year (early-bird price RMB 15.8/year), or a one-time purchase of RMB 156.
Alyx: Scientifically Tracking Caffeine, Letting Coffee and Health Coexist in Harmony
Platform: iOS
Keywords: Caffeine
@Vanilla: I think many coffee lovers like myself have had concerns like these: Did I drink too much coffee today? Will that afternoon cup affect my sleep quality at night? If you share these worries, you might want to try an app called Caffeine Tracker: Alyx. It lets you log coffee intake in multiple ways and automatically calculates caffeine content, then uses intake time and sleep time to estimate how much caffeine remains in your system—helping you understand, in real time, how caffeine is affecting your body. Compared with the feature list itself, I’m even more interested in Alyx’s UI interactions and its use of AI.
Let’s start with how you log coffee. Alyx supports two main ways to add new records.
The first is to tap the “+” button in the bottom-right corner of the app. You can quickly log a drink using presets or by creating your own caffeinated beverages. Alyx comes with a rich set of built-in templates—Espresso, Espresso-Based, Brewed Coffee, Cold Coffee, Instant Coffee, Tea, Energy Drink, Soda, and more—covering pretty much all the caffeinated drinks you’re likely to encounter in daily life.
If you can’t find what you’re looking for among the templates, you can create a custom drink. When doing so, you can either manually enter the caffeine content or let AI automatically identify it for you. In addition to drink templates, you can also directly input a caffeine amount, or type in a description and let Apple Intelligence analyze the caffeine content from text. Before logging, you’ll also need to specify the exact date and time of intake.
The second method leverages Apple Intelligence’s Visual Lookup API, allowing you to add caffeine intake simply by taking a photo. Long-press the camera button on the right side of the iPhone to bring up the Visual Lookup capture interface. Take a photo of the drink, tap the “Search” button in the bottom-right corner, switch to the “Alyx” tab, and you’ll see Alyx automatically identify the beverage and its caffeine content. Choose the appropriate option, and it’s added to Alyx automatically.
Beyond these two core input methods, Alyx also integrates with system features like Siri Shortcuts, widgets, and action buttons, making it easy to trigger logging in ways that fit different users’ habits and preferences.
Alyx offers a variety of charts, including yearly intake, weekly intake, daily intake, caffeine metabolism curves, and sleep correlations, helping you understand how coffee impacts your health. It also includes an AI-powered sleep prediction feature that provides real-time analysis based on caffeine intake, intake timing, and sleep schedules.
At the very bottom of the main screen, there’s also an “Honors Board,” showing the highest caffeine intake across different time spans—monthly, weekly, daily, or per single intake. In the settings, you can change the theme colors, manage your drink library, and adjust parameters like sleep time and daily caffeine limits.
As a caffeine tracking and analysis app, Alyx really stands out in both UI interaction design and feature depth. For coffee enthusiasts—or anyone reliant on caffeine—it’s an excellent health management tool. That said, it’s worth noting that Alyx’s AI sleep analysis should be treated as a reference only. When I fed the same scenario to Gemini 3.0 Pro, I got completely different results.
Alyx is free to download from the App Store, but a subscription is required to unlock features in the Caffeine Club, such as adding custom drinks, viewing more charts and widgets, and accessing sleep analysis. Pricing is $1.99 per month or $9.99 per year.
GitHub Store: A One-Stop Hub for Discovering Open-Source Projects
Platform: Android / Windows / macOS / Linux
Keywords: Open-source projects
@BigBigBigK: If you’re looking for free and open-source app projects, what could possibly beat GitHub? In the past, I could only learn about great GitHub projects through social media and similar channels. That changed a few days ago when I discovered an app called GitHub Store. It brings together nearly all application projects on GitHub—mobile and desktop alike—so you can tap into GitHub’s virtually endless resources through a single app.
GitHub Store itself is a cross-platform client written in Kotlin, with a UI fully built around the Material 3 design language. On Android, it also supports dynamic color theming. In terms of functionality, GitHub Store uses GitHub’s official search API to fetch project information across the entire site, then filters repositories based on whether they have installable release assets, supported platforms, version tags, and more. It also pulls in README.md files and statistics, making it easy for users to review each project.
The Trending and New tabs in GitHub Store only surface apps relevant to the current platform you’re using. The search function, however, has no such restrictions—you can even search for macOS or Windows apps while using the Android version.
The desktop version of GitHub Store is relatively more basic, while the Android version offers richer features. For example, it can automatically update installed apps (though for now it only tracks apps installed via GitHub Store itself), or directly add repositories to Obtainium for more granular installation management. It can also integrate with AppManager, a high-privilege tool, to manage and control app behavior.
One thing to note is that GitHub Store will prompt you to log in after a few uses. By linking to GitHub through a third-party app and using a personal token, you can lift API rate limits. GitHub Store supports Android, Windows, and Linux, and can be downloaded for free from GitHub.
Traffic Light: A More Beautiful Way to Track Your Data Usage
Platform: Android
Keywords: Data usage tracking
@Peggy_: I’m not sure what kind of mobile plan everyone is on these days, but since I’ve long had Wi-Fi at home and at school, I don’t use much mobile data. To cut costs, I downgraded to a cheaper plan—and from then on, I started rationing my data whenever I was out. Traffic Light is a beautifully designed data usage tracking app, and its biggest strength is simply how good it looks.
To get Traffic Light to accurately track your data usage, you’ll need to grant the necessary permissions in advance—such as disabling battery optimization to prevent the system from killing it in the background, and granting usage access so it can collect data more precisely. Once authorized, you can clearly see your Wi-Fi usage and mobile data usage at a glance.
On its statistics pages, Traffic Light shows data usage for the current day and week. Wi-Fi and mobile data are displayed in different colored blocks, and the vertical axis of the chart includes a horizontal reference line that adjusts based on your usage, making it easy to quickly gauge overall consumption. If you want to see detailed daily usage, just tap the history statistics page in the center of the main screen.
When it comes to design, nearly every button in Traffic Light’s interface is interactive. From the button used to change themes to each small icon on the statistics pages, tapping them triggers haptic feedback and icon animations, which revert on a second tap. These touches make what would otherwise be a dull data-tracking tool feel lively and fun.
Beyond usage statistics, you can also grant notification permissions to have Traffic Light display current network speed and real-time data usage. Overall, while its functionality isn’t irreplaceable, its polished design and attention to detail make it a pleasant app to keep on your phone and play around with from time to time.
You can download and try Traffic Light via F-Droid. The app is open-source and completely free.
App Updates You Shouldn’t Miss
Beyond brand-new apps, many familiar faces on the App Store are constantly iterating and updating, adding more interesting and practical features. SSPAI aims to help you filter through the app updates worth paying attention to, so you can quickly catch up on the latest developments from apps and their developers.
Cuto 3.1: Set Curated Live Wallpapers on iPhone
Platform: iOS / Android
Keywords: Live wallpapers
@化学心情下2: Cuto has long been a wallpaper app I keep installed on both iPhone and Android, largely because its wallpapers are handpicked and consistently high in quality. The app recently released version 3.1, introducing live wallpaper support, allowing your selected wallpapers to come to life when the screen lights up.
The principle behind live wallpapers is actually quite simple. iOS already allows users to set Live Photos as lock screen wallpapers, and Cuto’s live wallpapers are essentially Live Photos. All you need to do is download these Live Photos and set them as wallpapers in the system to see the animated effect.
Open Cuto’s “Live Wallpapers” collection, or look for wallpapers marked with a Live icon in the top-left corner. Find and download a live wallpaper, then open its preview page and tap the play button in the bottom-left corner to preview the animation. If you’re satisfied, tap the download button in the bottom-right corner to save the wallpaper to your system photo library.
Next, go to the “Wallpaper” section in iOS Settings, select the live wallpaper you just downloaded, and manually set it as your lock screen wallpaper. On the settings page, tap the Live icon to enable the animation, and your live wallpaper setup will be complete.
Due to iOS system limitations, the animation effect only plays for a few seconds when the screen lights up. Additionally, since iOS does not provide an API for configuring live wallpapers, there’s no way to automate the process with Shortcuts—you’ll need to set them manually.
According to the developers, future updates to Cuto will support user-created live wallpapers, such as a “video to Live Photo” feature that lets you convert videos into lock screen live wallpapers with one tap. Live wallpaper support for the Android version is also in development. Compared with iOS, Android imposes fewer restrictions, allowing for higher image quality and smoother animations.
Cuto’s live wallpaper feature requires a subscription to download. You can visit the official website to download Cuto for your platform of choice.
Lorqa Update: Floating Real-Time Subtitles With Custom Model Support
Platform: macOS
Keywords: AI
@Snow: Lorqa is a local AI translation app for macOS. In last week’s 1.3.3 update, it added support for using on-device AI engines to transcribe audio or video currently playing on the system in real time, automatically generating corresponding subtitles and translations. These subtitles can appear as floating overlays anywhere on the screen, significantly lowering the barrier to watching content without subtitles.
Using Lorqa’s real-time subtitles is very straightforward. On first use, follow the prompts to grant “Screen Recording & System Audio” permissions. After that, simply switch to the “Real-Time Subtitles” page, select a recognition model and the target translation language, and enable real-time translation. The recognized subtitles and translated text appear as long-form text in the app window. Once translation is complete, you can export the results as .txt or .md files for further processing.
Lorqa also offers floating subtitle overlays. Just click the “Floating Subtitles” button in the bottom-right corner to enable them. The overlay supports customization of font size, background opacity, original and translated text colors, and maximum width. Because Lorqa relies on system audio permissions, its use cases aren’t limited to local video translation—music, podcasts, online videos, video calls, and online meetings can all be assisted with Lorqa.
I describe it as “assisted translation” because, at present, Lorqa’s recognition speed and translation accuracy aren’t sufficient to fully stand on their own. It typically needs nearly two seconds to complete segmentation and translation. For English documentaries with clear pronunciation, moderate pacing, and more formal language, this delay doesn’t significantly impact comprehension. But in film and TV scenarios, even with the “Aggressive” transcription mode introduced in version 1.3.6, Lorqa struggles to segment speech clearly in fast, overlapping dialogue—and segmentation errors further reduce translation accuracy.
In addition to an “English Recognition Model,” Lorqa also includes a “Multilingual Recognition Model” supporting 25 languages, including French, German, and Bulgarian. However, because you can’t manually specify the language, speech recognition largely depends on luck. When I tried using it to recognize the Slovenian song Sanremo Memory, it alternated between translating into English and Russian—hard to call it accurate; it was more like a complete case of talking past each other.
Since I’m using an M1 MacBook Air, with relatively limited local compute power, that may also affect responsiveness to some extent. In the latest 1.3.6 version, Lorqa has opened up model configuration for Ollama and LM Studio. If you’ve deployed third-party models that are better at translation, you can try connecting them for potentially better results.
Lorqa uses a “free + in-app purchase” model. You can subscribe for RMB 38 per year or make a one-time purchase for RMB 68. The annual subscription includes a 7-day trial, so it’s recommended to try it before deciding whether to pay.
The animation editing and creation app Procreate Dreams released a major 2.0 update, adding support for selection and transformation, GIFs, advanced export options, and more.
“Rules on Pricing Behavior of Internet Platforms” Released
TikTok signs agreements with investors to establish a new U.S. joint venture
Airbus plans to migrate core business to Europe-based cloud services to avoid U.S. jurisdiction risks
Tesla wins final ruling; Musk’s record-breaking 2018 compensation package reinstated
Anna’s Archive announces completion of a large-scale backup of Spotify’s music catalog
Sam Altman explains the “red alert,” infrastructure plans, and hardware strategy in an interview
Rumors You Can Just Glance At
“Rules on Pricing Behavior of Internet Platforms” Released
On December 20, the National Development and Reform Commission, the State Administration for Market Regulation, and the Cyberspace Administration of China jointly issued and released the Rules on Pricing Behavior of Internet Platforms. The Rules were previously open for public consultation from August 23 to September 22, 2025.
The main points of the Rules include—
Protecting operators’ autonomy in pricing. Platforms must not use technical means or their dominant position to impose unreasonable restrictions on operators within the platform, such as forcing “choose one of two,” mandating promotions, or restricting pricing on other channels. Platform fee standards must be open and transparent, and any adjustments to fees must solicit opinions in advance;
Regulating price labeling practices. Operators must clearly disclose product prices, service content, and additional fees (such as shipping). Where dynamic pricing or promotional activities are involved, pricing rules and promotion conditions must be prominently disclosed. Products ranked through bidding must be clearly labeled as advertisements;
Restraining price competition behaviors. Practices such as selling below cost, using algorithms to implement discriminatory pricing (“big data price discrimination”), price gouging, and price fraud are prohibited;
Protecting consumers’ price-related rights. Automatic renewals must provide a convenient cancellation option and issue prominent reminders before charges are made; bundled products must not be pre-selected by default.
Compared with the earlier draft for public comment, the final version expands the scope of protection for pricing autonomy by extending the prohibition on price comparison bans from other platforms to all sales channels, and further banning intervention measures such as search result demotion and algorithmic downgrading. The criteria for identifying price gouging have been relaxed: in non-emergency situations, price increases that do not match cost increases will no longer be automatically deemed price gouging. Requirements for informing users about automatic renewals are made stricter—beyond specifying the deduction time, amount, and any price changes, platforms must also clearly and prominently notify users of the cancellation method. For price labeling, the Rules further require prominent disclosure on service pages.
The Rules will take effect on April 10, 2026, with the aim of allowing operators sufficient time to make necessary compliance adjustments.
TikTok signs agreements with investors to establish a new U.S. joint venture
According to Caixin, on the afternoon of December 18 (U.S. time), TikTok CEO Shou Zi Chew announced in an internal memo that ByteDance and TikTok have signed agreements with three investors to establish a new TikTok U.S. joint venture. The new entity will be named TikTok US Data Security Joint Venture LLC (TikTok USDS Joint Venture LLC) and will be responsible for data protection, algorithm security, content moderation, and software assurance in the United States. Matters related to the agreement are expected to be completed no later than January 22, 2026. Reaching this deal also means that TikTok will avoid being banned in the U.S.
Under the agreement, ByteDance and TikTok will continue to own the intellectual property rights to the core algorithms and will license them to TikTok USDS for use within the United States. Other TikTok entities in the U.S. (wholly owned by ByteDance) will continue to handle commercial activities such as e-commerce, advertising, and marketing operations, as well as maintaining global interoperability of TikTok products.
TikTok USDS will be 19.9% owned by ByteDance, making it the largest single shareholder in the new joint venture. ByteDance’s current U.S. and global shareholders will collectively hold 30.1%, while the new investors will hold 50%. Among them, Oracle, Silver Lake, and Abu Dhabi sovereign wealth fund MGX will each hold a 15% stake. The company will have a seven-member board of directors: ByteDance will occupy one seat, ByteDance’s existing U.S. and global shareholders will hold two seats, the new investors will hold three seats, and the remaining seat will be filled by an independent director appointed by the board.
Previously, on September 25, Trump signed an executive order in the Oval Office approving the transaction that allows TikTok to continue operating in the United States. The executive order also granted TikTok a 120-day exemption period during which it would not face penalties. At the same time, Trump retained the authority to issue further orders on the matter if necessary to safeguard national security.
Airbus plans to migrate core business to Europe-based cloud services to avoid U.S. jurisdiction risks
According to The Register, Airbus is preparing to launch a major tender to migrate its core mission-critical operations to Europe-based cloud platforms with “digital sovereignty,” in a bid to reduce its reliance on U.S. cloud service providers. The contract is expected to be worth more than €50 million and span a period of up to 10 years. The migration will cover systems such as ERP, manufacturing execution systems (MES), and lifecycle management systems involving sensitive aircraft design secrets. The tender is scheduled to open in January next year, with a final supplier expected to be selected before the summer.
Catherine Jestin, Executive Vice President of Digital at Airbus, emphasized that the move is intended to ensure that extremely sensitive information related to national and European security remains fully under European control. Beyond market factors—such as software vendors like SAP shifting their technological focus to the cloud—geopolitical risks are the primary driver. With Donald Trump’s return to the White House unsettling transatlantic trade relations, and the U.S. CLOUD Act allowing American law enforcement to access overseas data held by U.S. companies, concerns over data sovereignty among European enterprises have been intensifying.
Although U.S. giants such as Microsoft and AWS offer compliance solutions, Microsoft has previously acknowledged in a French court that it cannot fully guarantee exemption from U.S. law. Additionally, reports that the International Criminal Court’s (ICC) Chief Prosecutor once had services cut off by Microsoft due to U.S. sanctions have further heightened Airbus’s concerns about business continuity. At present, Airbus is awaiting clarification from European regulators on whether it can truly obtain protection from the extraterritorial reach of foreign laws.
Finding a suitable European provider, however, remains a challenge. Jestin noted that, given the relatively limited scale and technological maturity of European cloud service providers, there is only about an 80% chance of finding a satisfactory solution. These stringent requirements will not only test the technical capacity of Europe’s cloud vendors but also push them to accelerate industry collaboration in order to meet Airbus’s timeline.
Tesla wins final ruling; Musk’s record-breaking 2018 compensation package reinstated
On December 19, the Delaware Supreme Court issued a ruling overturning a lower court decision that had previously voided Tesla CEO Elon Musk’s 2018 compensation plan. The record-setting equity incentive plan was valued at approximately $56 billion at the time of vesting. The Supreme Court held that the Delaware Court of Chancery’s remedy—outright cancellation of the plan—was “overly extreme” and failed to give Tesla an opportunity to propose a reasonable alternative form of compensation. The court therefore reinstated the plan and awarded only $1 in nominal damages.
Back in 2018, Tesla’s board approved a ten-year equity incentive plan under which Musk, upon achieving a series of market-capitalization and operational milestones, would be entitled to purchase roughly 304 million Tesla shares at a deeply discounted price of $23.33 per share. During the plan’s term, Musk would receive no other compensation. A minority shareholder later filed suit, alleging that Musk and Tesla’s board breached their fiduciary duties. In January 2024, the lower court ruled that the compensation-setting process suffered from “serious defects,” finding that the board lacked independence, was controlled by Musk, and failed to adequately disclose key information to shareholders—ordering the plan’s rescission.
Although the compensation plan has now been reinstated, experts note that the Supreme Court’s ruling primarily addressed the proportionality of the punishment—namely, rescinding the plan—rather than overturning the lower court’s factual findings that Musk was a controlling shareholder and that the compensation process was unfair. As such, the court preserved the negative assessment of Tesla’s corporate governance and merely corrected the remedial approach.
Following the earlier loss, Musk moved Tesla’s corporate domicile from Delaware to Texas and pushed shareholders to reapprove the plan in subsequent votes. With the 2018 plan reinstated, Tesla’s contingency compensation plan—prepared to mitigate the risk of an adverse outcome—immediately became void. Notably, Tesla shareholders also approved an even larger 2025 compensation incentive plan this November; if its targets are met over the next decade, the new plan’s total value could reach as high as $1 trillion.
Anna’s Archive announces completion of a large-scale backup of Spotify’s music catalog
On December 20, the well-known piracy archive Anna’s Archive announced that it has successfully completed a large-scale backup of streaming giant Spotify’s music library. The site claims the project to be the world’s first “fully open” music preservation archive, aiming to ensure the permanent survival of musical culture through distributed storage. The release totals roughly 300 TB of data, including metadata for 256 million tracks and 86 million audio files, covering approximately 99.6% of Spotify’s total user listening volume. Anna’s Archive did not disclose the source of the files.
The released metadata database contains 186 million unique ISRCs (International Standard Recording Codes). By comparison, the mainstream open database MusicBrainz currently holds only about 5 million unique ISRCs. This means Anna’s Archive has built the largest publicly accessible music metadata index in the world.
Previously, Anna’s Archive focused primarily on the preservation of books and academic papers. The team stated that this effort was intended to fill gaps in existing archiving practices—namely, the tendency of audiophile-driven archives to prioritize lossless audio quality (resulting in massive file sizes that are difficult to mirror), as well as an overemphasis on popular works at the expense of “long-tail” music. Anna’s Archive emphasized that many niche tracks on Spotify have extremely small audiences and lack dedicated enthusiast maintenance; once streaming platforms remove them or data loss occurs, these “cultural artifacts” risk disappearing permanently.
According to the release plan, the metadata is already online, while the audio files, album artwork, and differential patches used to reconstruct original files will be released in stages based on popularity. At present, the project is intended primarily for archival purposes and does not yet support online searching or downloading of individual tracks.
Sam Altman explains the “red alert,” infrastructure plans, and hardware strategy in an interview
Recently, OpenAI CEO Sam Altman appeared on the Big Technology Podcast, confirming that Google’s recent release of the Gemini 3 model did indeed trigger an internal “Code Red” at OpenAI. This marked the company’s second time entering such an emergency state this year—the first being in response to competitive pressure from the Chinese AI company DeepSeek. Despite the intense competition, Altman revealed that ChatGPT’s weekly active users have surged from 400 million at the beginning of the year to 800 million. He stressed that OpenAI’s approach is not purely defensive, but rather focused on maintaining leadership through rapid iteration. The recently released GPT-5.2 model is widely regarded as the strongest reasoning model to date, excelling in scientific research and enterprise-level tasks.
On the product ecosystem, Altman said OpenAI is working to move beyond the single chatbot paradigm. He expressed reservations about Google’s strategy of embedding AI into existing search and office suites, arguing that the future lies in “native AI” products. As for the much-discussed hardware plans, Altman disclosed that OpenAI is developing a family of devices centered on proactive perception and environmental understanding, aiming to break free from the interaction limits imposed by traditional screens and keyboards.
To support its massive computing needs, Altman confirmed infrastructure investment commitments totaling as much as $14 trillion. This enormous sum will be allocated over the long term across chips, data centers, and energy infrastructure. Although OpenAI currently expects annual revenue to reach $20 billion, the company remains unprofitable in the short term due to high training costs. Regarding financial sustainability, Altman said that as inference costs account for a larger share and enterprise business scales up, revenue will eventually cover expenses. On the question of going public, he admitted that while he is not particularly eager to become the CEO of a public company, constraints related to shareholder limits and capital requirements mean that an IPO will be an unavoidable option.
Altman defined the core of the next generation of AI as “scientific discovery capability,” and previewed the release of a significantly improved model built on this foundation in the first quarter of 2026. As for the definition of artificial general intelligence (AGI), he proposed a new benchmark—“superintelligence”—where AI systems must surpass human capabilities when taking on complex roles such as serving as a company CEO or a national leader.
Rumors You Can Just Glance At
According to Weibo user @i冰宇宙, Samsung’s Galaxy S26 series is confirmed to be announced in February, with an expected release in March. This would be later than the launch schedule of recent years.
Some users have noticed that the official WeChat account of Computer Fan magazine has recently been deactivated, and its official website is no longer accessible. Computer Fan was first founded in 1993.
On December 20, Xiaomi Brand General Manager Lu Weibing said during a livestream that “the Xiaomi 17 Ultra will definitely see a price increase—and I think it will be a fairly significant one. But compared with the rise in memory costs, I still think it’s relatively modest.” When the Xiaomi 15 Ultra was released, the company had previously stated that it would be the “last time at 6,499 yuan.” Lu explained that this assessment did not fully take memory costs into account, and was based only on increases in processor and camera costs. The 17 Ultra, however, also factors in rising memory prices, with increases far exceeding those of processors and cameras. He added that since the end of 2022, AI has experienced explosive growth, and based on overall projections, memory costs are expected to continue rising through 2025, 2026, and 2027.
Jizhi Dynamics launches the multi-form embodied robot TRON 2
China Media Group unveils the mascot for the 2026 Spring Festival Gala
Hainan Free Trade Port completes island-wide customs closure
Volcengine officially releases the Doubao Large Model 1.8
OpenAI announces developers can submit apps to ChatGPT
Rumors You Can Just Glance At
Apple adjusts App Store and iOS rules in Japan
On December 18, Apple announced a series of adjustments to the App Store and the iOS system in Japan to comply with the regulatory requirements of the Mobile Software Competition Act (MSCA). Under the new rules, developers are allowed to offer third-party payment methods within their apps, direct users to websites to complete purchases, or distribute apps through alternative app marketplaces, alongside the introduction of a new fee structure. The overall framework is similar to the EU’s DMA, but differs in areas such as the parallel presentation of payment options, app distribution channels, and interoperability requirements.
In addition, the MSCA is driving several system-level changes, including allowing users to choose default browsers, search engines, and navigation apps during setup, as well as opening the side button to eligible third-party voice assistants. In terms of protections for minors, Apple has introduced stricter restrictions on payments and external links for children’s apps and underage users. These changes began rolling out with the iOS 26.2 update and apply only to iOS apps distributed in Japan. Source
Jizhi Dynamics launches the multi-form embodied robot TRON 2
On December 18, LimX Dynamics announced the launch of its multi-form embodied robot TRON 2 and opened pre-orders, with a starting price of RMB 49,800. The product features a modular design that allows a single core body to rapidly switch between multiple configurations—such as dual-arm, bipedal, and wheeled-leg forms—covering use cases ranging from fine manipulation to mobility across complex terrain.
In terms of system capabilities, TRON 2 is equipped with high-degree-of-freedom embodied dual arms and all-terrain locomotion, and supports the integration of multimodal perception components. The product offers open APIs and standardized hardware interfaces, enabling compatibility with mainstream Vision-Language-Action (VLA) models, and is accompanied by secondary development tools and a data platform, targeting applications in research, industry, and embodied intelligence–related fields. Source
China Media Group unveils the mascot for the 2026 Spring Festival Gala
China Media Group has revealed the mascots for the 2026 Spring Festival Gala: four fine horses named “Qiqi,” “Jiji,” “Chichi,” and “Chengcheng,” which made their debut together. According to the announcement, the mascot designs draw inspiration from iconic depictions of horses across different periods of Chinese history, incorporating traditional motifs such as flowing cloud patterns and mountain-cloud patterns, symbolizing auspicious wishes for success and a bright future. Source
Hainan Free Trade Port completes island-wide customs closure
According to the official Hainan Free Trade Port WeChat account, starting December 18, 2025, the Hainan Free Trade Port will officially implement an island-wide customs closure. After the closure takes effect, the relationship between Hainan and areas outside the island will be defined as the “first line,” with “outside the island” including not only foreign countries and regions, but also Hong Kong, Macao, and Taiwan.
The account also released two catalogues. The first is a prohibited and restricted list, covering goods and items that are banned or restricted from import and export; goods and items not on the list may move freely. The second is the catalogue of taxable imported goods (also known as the “zero-tariff negative list”): goods not included in this catalogue may enter the free trade port exempt from import duties. After the customs closure, the policies of “zero tariffs, low tax rates, and a simplified tax system” will gradually be extended to a broader scope.
For businesses, the Hainan Free Trade Port means a wider range of policy benefits with lower thresholds; for individuals, it means that purchasing various products in Hainan will become more affordable. Source
Volcengine officially releases the Doubao Large Model 1.8
At the Volcano Engine Force Core Power Conference held on December 18, ByteDance officially released Doubao Large Model 1.8. The model has been specifically optimized for multimodal agent scenarios, with enhanced capabilities in tool invocation, complex instruction understanding, and OS agent–related functions, significantly improving its planning and execution performance when handling complex tasks.
Volcano Engine stated that in multiple public benchmarks, Doubao 1.8 achieved best-in-class or near-best results in tasks such as visual reasoning, general visual question answering, spatial understanding, and video understanding; it ranks among the global leaders on the general agent benchmark BrowseComp; and in foundational language model capabilities such as mathematics and reasoning, its overall performance is also close to that of the world’s top general-purpose models. Doubao 1.8 is now available on Volcano Engine and has opened its API to enterprises and developers. Source
OpenAI announces developers can submit apps to ChatGPT
On December 17, OpenAI announced that the previously launched ChatGPT app feature is now open for developer submissions. Developers can apply for listing in accordance with the app submission guidelines in order to publish their apps on the ChatGPT platform. OpenAI stated that it has provided a range of resources to help developers build higher-quality applications.
At the same time, an app directory has gone live within ChatGPT. Users can browse recommended apps or search through all published apps in the directory. The directory can be accessed via the Tools menu, or directly by visiting chatgpt.com/apps. Developers can also use deep links on other platforms to take users directly to their app’s dedicated page within the directory.
In the current early stage, developers can set up outbound links within their ChatGPT apps to guide users to their own websites or native apps to complete transactions for physical goods. In the future, OpenAI will explore additional monetization options, including digital goods, and will provide further details based on real-world usage and interactions between developers and users. Source
In addition, ChatGPT has also been integrated into the Apple Music app, allowing users to search for songs and generate playlists directly within Apple Music. Source
Rumors You Can Just Glance At
In the latest Android 2512 Canary build, the system has added a new “Satellite” Quick Settings tile along with a “Satellite Connectivity” settings page, which displays availability status and lists apps that support satellite networks. The tile will show as enabled, available, or unavailable depending on whether the user’s carrier plan supports NTN (Non-Terrestrial Network) roaming. Code indicates that the system will present different lists of satellite-enabled apps based on whether the device and plan support LTE-NTN or narrowband NB-NTN. In addition to the already announced Google Maps, Messages, and WhatsApp, apps such as AccuWeather, Google Weather, Snapchat, and X also appear on the candidate list. This suggests that satellite communication capabilities within the Android ecosystem may continue to expand, though these features are still in testing and development. Source
According to Bloomberg, TikTok CEO Shou Zi Chew revealed in an internal memo that TikTok and ByteDance have signed binding agreements with Oracle, Silver Lake, and MGX to establish a U.S. joint venture primarily owned by American investors, advancing TikTok’s long-delayed divestment plan. The transaction is expected to be completed on January 22, 2026, pending multiple preparatory steps and regulatory approvals. Source
After earning the Platinum trophy in Black Myth: Wukong, I’d been hoping that one day I could own a 1/12-scale articulated figure of the Destined One. So when Inart unveiled a gray prototype at an exhibition, I already knew that once it went on sale, I’d definitely pick one up. Only recently—after players who paid in full had already received their figures and enjoyed them for over a month—I finally got the payment notice for my deposit order and received my very own 1/12 Destined One.
You Truly Delight Me, Little Monkey
Although I’d previously bought a 1/12 Zhu Bajie, once I got my hands on this Destined One, I still couldn’t help but marvel at the level of detail. Even at a 1/12 scale, it manages to present distinct textures through a mix of hard plastic, soft rubber, and fabric clothing. The armor’s surface details and weathered paintwork are especially refined. There’s even a wire embedded in the tail, allowing you to pose it freely into whatever shape you like.
To be honest, before buying it I was a bit worried about articulation. Zhu Bajie’s articulation hadn’t been particularly impressive, but this figure exceeded my expectations. In addition to the poseable tail mentioned earlier, the lower body uses a soft rubber armor combined with fabric clothing to avoid restricting hip movement, leaving thigh articulation completely unhindered. To better convey dynamic poses, the forearms include ball joints, and even the feet feature individually articulated joints for the big toe and the other four toes.
The accessories are also impressively abundant. I went with the Deluxe Edition, which adds a long staff for displaying the “staff stance,” a pair of swinging-staff effect parts, and—most importantly—the monkey head. Compared to the standard version’s head with the “bronze monkey mask,” the look without the headpiece actually better evokes my impression of the Destined One. Among the shared accessories, the weapon “Beast Staff · Ermine” uses a three-section assembly with metal pegs and foolproof design, so there’s little worry about damage. In addition, there’s a Shrinking Talisman and two interchangeable gourds, making the overall play value quite high.
A Big, Full Box (Even the Accessories Are Packed in Three Layers)
That said, the weapon itself does have some weight—especially once the effect parts are attached—so you’ll need to spend a bit of time adjusting the wrist joints to keep it from drooping under the load.
All in all, this Destined One figure exceeded my expectations. While ¥608 admittedly isn’t cheap, compared to some Japanese manufacturers that charge over ¥1,000 RMB for figures that come with only a few extra hands, Inart’s offering feels like excellent value for money.
@西鸽: Li Auto AI Glasses Livis (Olive Green, Matte, Non-prescription, Sunglasses Lens)
Reference price: ¥1,869.15 (after subsidies)
Colleagues at my company are using various AI glasses, to the point where they’ve practically become everyday office gear. I’d also been curious to try a different brand. It just so happened that Li Auto put the Livis smart glasses on sale, so I quickly picked up a pair and wanted to share my impressions.
I didn’t watch the launch event and had no idea what features they offered. The next day I checked the official site, thought the glasses looked pretty good, and placed an order. Since I didn’t need prescription lenses, they shipped directly and arrived on the third day. Since these are Li Auto glasses, I naturally wondered whether they would integrate with Li Auto vehicles. I drive a Li ONE—an older model that doesn’t even get system updates anymore—so I didn’t expect full support from a new smart device. After looking into it, though, it turns out there’s basically no support at all.
First, I assumed there would be a HUD feature. The Li ONE doesn’t have a HUD, so I thought maybe smart glasses could serve as a kind of “physical add-on” for older models. But that’s not the case. What’s a bit funny is that I asked this same question to a Li Auto staffer, Doubao, and Gemini:
Only Gemini clearly told me that these glasses do not have an in-lens HUD, and even pointed out the gap between product marketing and user understanding. Kind of interesting.
Of course, being able to see a HUD clearly doesn’t mean the glasses themselves have one. According to the AI, the intended meaning of this marketing point is that wearing the glasses lets drivers see the car’s built-in HUD more clearly—the glasses themselves do not include a HUD.
So how much linkage is there between the Livis glasses and the Li ONE? Basically none. With the glasses already connected to my phone, I can give voice commands to the glasses, which then pass them to the Li Auto Assistant app on my phone. That app itself already has simple vehicle controls, like unlocking doors or opening windows and the trunk. If I’m outside the car, this just saves me from taking out my phone; if I’m inside the car, it’s not practical at all, because it would trigger both the glasses and the in-car system’s “Li Auto Assistant” at the same time.
The product page on the official website is also very vague about Li ONE integration—what exactly you can “use immediately” isn’t really explained.
So it’s pretty clear: if you don’t have an L-series or newer Li Auto vehicle, the Livis glasses are essentially just a smart voice assistant that works with your phone (via the Li Auto Assistant app), plus an independent device with audio recording, music playback (the sound quality is much better than bone-conduction), and photo/video recording.
The version I bought is the sunglasses model. For office use, it’s fine as a Bluetooth audio device for listening to music and recording, but wearing sunglasses indoors feels odd, which rules out another usage scenario.
So in the end, let’s just look at the appearance. At the very least, my kids say they look cool. And I do enjoy wearing the Livis while chasing after them and recording videos—it’s a lot more convenient than holding up a phone.
@Microhoo: DJI Mini 5 Pro
Reference price: ¥4,788 (standard kit)
I’ve always felt that the Mini series is the hidden gem among all DJI drones. Because of DJI’s precise product segmentation—and the fact that many people mistakenly label it as “entry-level”—the Mini lineup ends up landing right at that “good enough” sweet spot. The result is a bit awkward: if you just want to buy a drone to play around with, nearly ¥5,000 isn’t cheap; but if you’re looking for a serious aerial creation tool, its overall strength can feel just a bit short, making it tempting to stretch up to the Air or even the Mavic series.
With expectations already low, when the Mini 5 Pro was announced I barely glanced at DJI’s official release. I only registered that it had switched to a 1-inch sensor and didn’t look much further. But over the past few months, after seeing countless hands-on experiences shared by creators across different platforms, I slowly pieced together the full picture of what had changed—and suddenly realized that the Mini I’d always wanted had finally arrived.
For other drone models, I usually recommend going “as complete as possible,” because a real-world flight time of just over twenty minutes passes in the blink of an eye, and extra batteries mean more peace of mind. But since the core appeal of the Mini 5 Pro is its compactness—its footprint is about the size of a smartphone—I actually think the standard kit is perfectly fine. I mainly use it in scenarios like hiking and mountaineering, where minimizing weight is critical. When I come across a great vantage point, I fly it, then toss it back into my bag and plug in a charging cable. Especially since the Mini 5 Pro’s internal storage has been upgraded to 42GB, that’s more than enough for a single trip’s worth of shooting.
My only real regret is that, even though the screenless remote included in the standard kit is already very lightweight, it’s still an extra thing to carry. Before getting it, I had imagined that the Mini series might be able to fly directly via a phone connection like the Neo, so I could just slip it into my pocket and go. In reality, that’s not possible—you still need the remote. That said, it’s understandable. With the Mini 5 Pro’s new transmission technology, its flight range has improved significantly, and combined with its strong acceleration, using just a phone would probably mean flying out of signal range in the blink of an eye. Still, I hope DJI will consider ultra-lightweight users and add a simplified phone-only control mode in the future, even if it comes with heavily limited range and speed.
It’s also worth mentioning that while light weight is the Mini 5 Pro’s standout trait, its biggest advantage is actually traveling abroad with it. Multiple generations of the Mini series have kept their weight at 249g not out of obsession, but because under the drone regulations of most countries, drones under 250g are exempt from complicated registration and filing procedures. This is especially true in places like Europe and Japan, which are ideal for aerial sightseeing and often have specific exemptions or simplified processes. This is the main reason I’ve always been drawn to the Mini series. As mentioned earlier, it’s not an entry-level product for beginners at all—instead, it’s a true “all-rounder” suited to most people.
The two most important upgrades of the Mini 5 Pro are, first, its 50MP 1-inch sensor camera. Almost every user has praised it as a “Pocket 3 that can fly.” I’m not sure whether it’s the same sensor used in the Pocket 3, but after watching some comparison videos, the image quality is strikingly similar. In situations with drastic changes in ambient light and white balance, however, the Mini 5 Pro responds noticeably faster than the Pocket 3—likely thanks to improvements in the new chip’s processing power. Given how well-regarded the Pocket 3 is, I don’t think the Mini 5 Pro’s imaging quality will feel outdated anytime soon, at least until display technology enters its next generation.
I also took a couple of photos—one outdoors at noon with extreme dynamic range, and one in a pitch-black city scene—just to get a sense of its imaging capabilities.
The other major upgrade is the “full trickle-down” of obstacle avoidance. Building on the previous generation’s omnidirectional obstacle sensing, the Mini 5 Pro adds a forward-facing LiDAR sensor that was previously reserved for flagship models. This not only improves obstacle avoidance but also makes it effective in low-light environments. That said, I’d still treat it as an emergency measure—when it comes to crashing drones, there are only two outcomes: zero or one. Don’t take chances; always fly carefully.
Like its predecessor, the Mini 5 Pro supports lossless vertical shooting, and thanks to a greater rotation angle, it can now perform horizon-breaking camera moves similar to those on the Mavic 4 Pro. All of this can be achieved through the built-in one-tap video creation features. Which brings me to my recurring plea: since DJI is already lowering the barrier to entry as much as possible, I really hope they’ll eventually offer a phone-only control option to further reduce what you need to carry in certain scenarios.
All things considered, the Mini 5 Pro is, in my view, the first truly “complete” small drone. Especially when you factor in regulatory weight limits, its overall capabilities exceed expectations. Compared with other models, it fits into more use cases as well. Unless you’re a commercial aerial photographer, the Mini 5 Pro is an extremely reliable choice. Once I’ve got the time and money, I’ll definitely be taking it out on more trips.
@Lotta: BlanBunny The Wizard of Oz Tea Book Gift Box
Reference price: ¥127
If you’ve ever wandered through the commercial streets of tourist attractions, you may already be familiar with the brand BlanBunny. Its teal-blue, European-style vintage storefronts, giant rabbit statues wearing top hats, and richly ornate gift boxes often make it one of the most eye-catching shops on the street, tempting passersby to step inside. Even though the relatively high prices often persuade people to leave empty-handed, there’s usually one or two finely made items that end up quietly saved in a wishlist or buried deep in a shopping cart—just waiting for a good deal to trigger an impulse purchase.
The Wizard of Oz gift box was exactly such a purchase during the Double 11 shopping festival. Inside the box are 10 different teas, two sachets of each, enough to brew 20 cups in total. It includes some sweet-and-tangy fruit and floral teas that I’m not particularly fond of, such as Grape Serenade and Creamy Strawberry Oolong; some pure teas that I do enjoy, like Gardenia Yuhua Tea, Jasmine White Tea, and Osmanthus Oolong; as well as more exotic herbal options, including South African rooibos and mangosteen yerba mate.
The tea is packaged in a rather distinctive perforated aluminum stick, designed to be brewed by stirring. As you stir, the color of the tea gradually releases into the water, which adds a bit of playful charm to the process. That said, for lighter-flavored teas like Osmanthus Oolong, the infusion does need to steep longer than it would in a standard tea bag.
Even the most monotonous workdays occasionally need a splash of non-routine color, and that’s exactly what draws me to BlanBunny—their lavish design. Each tea comes with its own illustration and a Wizard of Oz-themed name, and the gift box cover even includes a detachable illustrated booklet and a tea “map.” Compared with Japanese brands that are also known for extravagant packaging—where small candies or sweets can easily cost over a hundred yuan—BlanBunny doesn’t feel overly superficial. In terms of design and visual richness, it’s every bit on par with Japanese brands, making it a great choice as a small gift when meeting friends.
China approves its first batch of L3 autonomous driving models
OpenAI launches GPT Image 1.5
Google Labs introduces experimental AI agent CC
Bambu Lab launches the “Let’s Make It” Creator Fund
Google to shut down the “Dark Web Report” feature
Steam 2025 Year in Review page goes live
Merriam-Webster names “slop” as Word of the Year
Rumors You Can Just Glance At
X reaffirms ownership of the Twitter trademark
On December 16, X Corp. filed a lawsuit in federal court in Delaware against startup Operation Bluebird, accusing the company of attempting to revoke X’s ownership of the Twitter brand.
On December 2, Operation Bluebird submitted a petition to the U.S. Patent and Trademark Office seeking to cancel X’s federally registered Twitter and Tweet trademarks. The lawsuit centers on whether X still retains trademark rights to the Twitter brand following its rebranding in 2023. In its complaint filed in Delaware, X argues that the Twitter brand remains valid, noting that millions of users continue to access the platform via twitter.com, and stating that “rebranding does not equate to abandoning original trademark rights.” Source
Around the same time, TechCrunch also discovered that X had updated its Terms of Service. The revised terms state that, effective January 16, 2025, nothing in the agreement grants users the right to use the X name or the Twitter name, including any X or Twitter trademarks, logos, domain names, other distinctive brand features, or other proprietary rights—despite Elon Musk having publicly stated in 2023 that he would abandon Twitter-related branding. Source
China approves its first batch of L3 autonomous driving models
On December 15, the Ministry of Industry and Information Technology officially announced China’s first batch of entry permits for L3 (conditionally automated) driving vehicles. The approved models include the Changan SC7000AAARBEV all-electric sedan and the Arcfox-branded BJ7001A61NBEV all-electric sedan, designed respectively for urban congestion scenarios and highway driving.
Both models will carry out on-road pilot programs in designated areas of Beijing and Chongqing, marking a critical step as China’s L3 autonomous driving moves from the testing phase toward commercial application. Source
OpenAI launches GPT Image 1.5
On December 17, OpenAI released GPT Image 1.5. The new model promises up to a fourfold increase in image generation speed compared with its predecessor, improved prompt adherence, and more precise image editing capabilities that maintain visual consistency throughout the editing process. It also delivers greater reliability when handling smaller, denser text and large numbers of small-sized faces in images.
The model has now been fully rolled out to all ChatGPT and API users. Source
Google Labs introduces experimental AI agent CC
On December 16, Google Labs announced the launch of an experimental AI agent called CC. The agent can connect to services such as Gmail, Google Calendar, and Google Drive to help users stay on top of their daily schedules. Each morning, CC sends a “Today’s Agenda” briefing to the user’s inbox. In addition to schedule overviews, key tasks, and summaries of recent updates, the briefing may also include email drafts and calendar links when needed. Users can control CC and submit custom requests by replying to or directly sending emails, as well as use it to add personal information or jot down thoughts and to-do items.
CC is now available in early access to Google account holders aged 18 and above in the United States and Canada, initially rolling out to Google AI Ultra users and paid subscribers. Source
Bambu Lab launches the “Let’s Make It” Creator Fund
On December 15, Bambu Lab announced the launch of the “Let’s Make It” Creator Fund. The fund aims to provide long-term financial, technical, and resource support to makers around the world, with each project eligible to receive up to RMB 1 million. Both individual creators and small studios are welcome to apply.
The “Let’s Make It” Creator Fund is open on an ongoing basis with no application deadline. Applicants can submit project proposals through Bambu Lab’s official website, and the same applicant may submit multiple projects. After evaluation, the review committee will work with selected applicants to determine specific support plans. This is currently the largest maker grant program by funding amount known in China. Source
Google to shut down the “Dark Web Report” feature
On December 16, Google informed users via email that its security check tool “Dark Web Report,” which scans for personal data and password leaks, will be discontinued on February 16, 2026. Starting January 15, 2026, the feature will no longer provide new monitoring results; from February 16, 2026 onward, the feature and its associated data will no longer be available.
Google stated that while the report tool could provide basic information, user feedback indicated that it failed to offer practical follow-up actions. As a result, Google has decided to shift its focus toward developing tools that deliver clearer and more actionable guidance. Source
Steam 2025 Year in Review page goes live
On December 17, the 2025 Steam Year in Review page officially went live. The statistics period for the 2025 Steam Year in Review runs from the very first second of January 1, 2025, to the very last second of December 14, 2025 (GMT). It does not include playtime in Offline Mode or while disconnected from the internet, nor does it count time spent running tools, unreleased, preloaded, pre-release, or disabled games, or other non-game software.
Users can set the visibility of their Steam Year in Review page to “Friends Only” or “Public.” All adult members in a Steam Family will also automatically have permission to view the Steam Year in Review of every other member in that family, while child members do not receive any special access privileges. Source
Merriam-Webster names “slop” as Word of the Year
The well-known dictionary publisher Merriam-Webster recently announced that slop has been selected as its Word of the Year for 2025, reflecting the widespread proliferation of low-quality AI-generated content across the internet. The term originally emerged in the 18th century to describe mud or sludge, but today it is more commonly defined as “low-quality digital content mass-produced by artificial intelligence.” Merriam-Webster president Greg Barlow said that the surge in searches for slop indicates growing public awareness of false and low-value content. Source
Rumors You Can Just Glance At
After reviewing Paramount’s all-cash offer of USD 30 per share and taking into account Paramount’s financing structure and regulatory approval pathways, Warner Bros. Discovery is reportedly planning to reject the previously proposed USD 108.4 billion hostile takeover bid from Skydance Media. According to sources familiar with the matter, the Warner Bros. Discovery board plans to recommend that shareholders accept the existing Netflix deal, with a related announcement potentially coming as early as Wednesday. Source
In response to claims by video creator @MLID that Samsung plans to gradually discontinue solid-state drives based on the SATA interface, wccftech contacted Samsung for comment. A Samsung spokesperson issued a brief response stating that rumors about Samsung phasing out SATA or other SSD products are false. However, as the response did not specifically address SATA SSDs, whether Samsung will discontinue its SATA SSD lineup remains unclear and is expected to become clearer next month. Source
Google to bring real-time voice translation to Google Translate
Astell&Kern’s IRIVER returns to the Chinese market
Meituan announces suspension of the “Tuan Haohuo” business
China Mobile releases a 6G transmission technology white paper and prototype
Ant Group’s AI health app AQ renamed Ant Afu
Rumors You Can Just Glance At
vivo releases the S50 series smartphones
On December 15, vivo held the S50 series launch event, unveiling multiple new products.
The vivo S50 is powered by the Qualcomm Snapdragon 8s Gen 3 processor, featuring a 6.59-inch display and a 6,500mAh battery, with support for 90W wired fast charging. It uses an ultrasonic under-display fingerprint sensor and the Ultra low-frequency global antenna. For imaging, it is equipped with a 50MP periscope telephoto camera using a 1/1.95-inch Sony IMX882 sensor, a 50MP main camera with a 1/1.56-inch Sony LYT-700V large sensor with advanced stabilization, and a 110° ultra-wide, low-distortion lens. The device focuses on 3× telephoto Live Photo shooting, and supports features such as Live Photo passerby removal, AI-customized beauty effects, film-style filters, and highlight slow-motion camera movement. The vivo S50 is available in Confession, Leisurely Blue, Inspiration Purple, and Deep Space Black, priced from CNY 2,999 (12GB+256GB) to CNY 3,599 (16GB+512GB).
The vivo S50 Pro mini is powered by the Qualcomm Snapdragon 8 Gen 5 processor, with a 6.31-inch display and a 6,500mAh battery. It supports 90W wired fast charging and 40W wireless charging, and also features an ultrasonic under-display fingerprint sensor and the Ultra low-frequency global antenna. In terms of imaging, the S50 Pro mini shares the same ultra-wide and periscope telephoto lenses as the S50, while its main camera uses a 50MP 1/1.56-inch Sony IMX921 ultra-sensitive bionic large sensor. The vivo S50 Pro mini is available in Confession, Inspiration Purple, and Deep Space Black, priced from CNY 3,699 (12GB+256GB) to CNY 4,299 (16GB+512GB). Pre-orders for both S50 models start today, with official sales beginning on December 19.
In addition, vivo Pad5 Pro has introduced a new matte display edition, featuring a 3.1K high-definition matte screen produced with nano-level crystalline etching technology. The 12GB+256GB version is priced at CNY 3,699, while the 16GB+512GB version costs CNY 4,199, and is available starting today. Source
On December 14, iRobot—the company that pioneered the robot vacuum category—issued an announcement stating that it has reached a restructuring support agreement with its secured creditors and primary contract manufacturer, Shenzhen Shanchuan Robotics Co., Ltd., and its subsidiaries. Under the agreement, Shanchuan Robotics will acquire iRobot and obtain 100% ownership.
Following the completion of the acquisition, iRobot will become a wholly owned private company of Shanchuan Robotics and will be delisted from Nasdaq. Previously, Shanchuan Robotics had assumed iRobot’s USD 190 million loan. The acquisition will relieve iRobot of this debt, as well as an additional USD 161.5 million in loans owed to Shanchuan Robotics related to contract manufacturing. Shanchuan Robotics is one of the world’s largest contract manufacturers of robot vacuum cleaners. Source
Google to bring real-time voice translation to Google Translate
On December 12, Google published a blog post outlining the next phase of upgrades for Google Translate. Google will introduce real-time natural speech translation powered by Gemini across its apps. The Google Search app will deliver more natural and accurate text translations. With Gemini support, Google Translate will provide more precise translations of idioms, implied meanings, localized expressions, and slang. This feature will first launch in the U.S. and India on both app and web, supporting mutual translation between English and nearly 20 languages including Spanish, Hindi, Chinese, Japanese, and German.
At the same time, Google Translate is rolling out a beta Android app in the U.S., India, and Mexico that supports real-time natural speech translation—users can activate it simply by wearing headphones—with support for over 70 languages. The feature is expected to expand to the iOS app and more regions in 2026. In addition, the built-in language learning features within Google Translate will be extended to 20 new countries and regions, including Germany, India, and Sweden. Source
Astell&Kern’s IRIVER returns to the Chinese market
On December 13, South Korean digital audio brand IRIVER announced its return to the Chinese market via an official post on Xiaohongshu. IRIVER stated that it will focus on the Hi-Fi audio and CD player segments going forward, and that full-scale operations in the Chinese market will soon begin. Brands under the IRIVER umbrella include Astell&Kern, ACTIVO, and IRIVER. Source
Meituan announces suspension of the “Tuan Haohuo” business
On December 15, Meituan’s Tuan Haohuo business issued an internal email stating that, after research and discussion, the grocery retail management team has decided to suspend the Tuan Haohuo operation and refocus on exploring new retail formats. As an exploratory initiative within Meituan’s parcel-delivery e-commerce efforts, the business accumulated experience in product retail. However, in recent years, as the grocery retail sector has continued to innovate, parcel-based e-commerce has struggled to meet the needs of instant retail users. “We will follow the trend and proactively seek change,” the email said.
Regarding personnel affected by this adjustment, Meituan noted that internal communication will be conducted, with detailed follow-up plans to be explained. Tuan Haohuo is a B2C e-commerce business incubated by Meituan. It launched as a mini program in August 2020, was promoted to a top-level entry within the Meituan app in December 2020, and was later renamed Meituan E-commerce. Source
China Mobile releases a 6G transmission technology white paper and prototype
On December 15, China Mobile published an article stating that, at the 2025 China Information and Communications Conference and the academic annual meeting of the China Institute of Communications, it released the China Mobile 6G Transmission Technology White Paper along with the “China Mobile 6G Transmission System Prototype 1.0.”
The white paper proposes that the core concept of 6G transmission is to move from “connectivity” toward “beyond connectivity.” By deeply integrating three key elements—dynamic intelligence, multidimensional coordination, and end-to-end security—it aims to build an integrated 6G intelligent collaborative transmission network architecture consisting of a service perception layer, a coordinated transmission packet layer, a coordinated transmission channel layer, and a coordinated transmission medium layer. Source
Ant Group’s AI health app AQ renamed Ant Afu
On December 15, Ant Group held its Ant Health AI strategy and brand upgrade launch event, announcing that Ant AQ has been renamed Ant Aifu, alongside the release of a new version of the app. The Aifu app will focus on three core health functions: health companionship, health Q&A, and health services. It has already partnered with the health commissions of 8 provinces and municipalities, as well as the medical insurance authorities of 18 provinces and municipalities through the “Yibaor” program, forming a network of 200 hospital-based medical AI agents, and has launched joint AI research initiatives with 16 institutions. Source
Rumors You Can Just Glance At
According to Tom, host of the Moore’s Law Is Dead channel, citing sources from multiple distribution and retail channels, Samsung is preparing to gradually exit its SATA SSD business and plans to halt SATA SSD production in the long term. If Samsung removes its SATA SSD production lines, it will inevitably put upward pressure on prices across the entire SSD market, including both SATA and NVMe products. This also signals that the era of low-cost SATA SSDs may be coming to an end. Source
Earlier this month, the GNOME Shell Extensions store banned AI-generated extensions from being listed. Source
At the Black Hat Europe conference on December 11, Microsoft Vice President Tom Gallagher announced that Microsoft will immediately expand the scope of its bug bounty program. Any critical vulnerability that has a direct and verifiable significant impact on Microsoft online services will now be eligible for a reward. Source
According to the Financial Times, the UK government is considering requiring Apple and Google to build age-verification systems and explicit-content detection algorithms directly into their operating systems. Until users complete adult verification via biometric methods or official identification, all explicit content would be blocked by default. Source
Many LG webOS TV users have reported that after completing a recent webOS system update, a non-removable Microsoft Copilot app appeared on the home screen. Source
Welcome to this edition of Pi Review. You can use the table of contents to quickly jump to the sections you’re interested in. If you’ve discovered other apps worth checking out, or topics you’d like us to cover, feel free to join the discussion in the comments.
New Apps Worth Noticing
While SSPAI has long been dedicated to discovering and introducing high-quality apps across platforms, there are still many excellent apps—whether in design, functionality, interaction, or overall experience—that we haven’t yet covered. Some may be long-standing apps, others newly released. We’ll introduce them to you here.
RecordMaster: A Recorder Can Be Colorful Too
Platform: Android
Keywords: Recording
@Peggy_: In the SSPAI homepage feature Deep Dive | What Exactly Is Material 3 Expressive? Using Android 16 Stable as an Example, editor Clyde offered readers a detailed explanation of the M3E design language, its evolution, and its strengths and weaknesses. Compared with large companies, independent developers are often able to respond more quickly to the latest design guidelines, giving rise to many new apps. RecordMaster, the app introduced today, is one such product born from this context.
Starting with the basics, RecordMaster is a voice recorder app built around the newer M3E design language. When you open the app, the main interface features just a simple add button. Tap it to enter the recording screen, then tap “Record” to start recording. During recording, the interface displays a waveform that changes with the audio volume. The animation is smooth, and the rich use of color makes the recording process feel far less monotonous.
Like many recorder apps, RecordMaster allows you to pause and resume recording at any time. Once you stop, the app shows the exact storage path of the current recording, making it easy to organize later. By default, recordings are sorted by recording time and displayed sequentially on the main screen.
As your collection grows, RecordMaster also supports organizing recordings by folders, making it easier to manage different types of audio files. Of course, since all recordings are stored locally, RecordMaster currently does not support speech-to-text conversion, nor does it offer instant syncing of recorded audio. For users who prioritize data privacy, however, this fully local storage approach can actually be an advantage.
If you’re looking to try a recorder app with a distinctive design, you can download and try RecordMaster via its GitHub page. The app is completely free and only requires microphone access.
Eney: A Local Intelligent Companion for macOS
Platform: macOS
Keywords: AI assistant, intelligent companion
@ElijahLee: Eney is an AI-driven Mac assistant developed by MacPaw. Positioned as a local-first intelligent companion, it allows users to use natural language to have it carry out a wide range of tasks—rather than merely offering suggestions or answering questions.
After installing Eney, you must grant it Full Disk Access for it to function properly. During the initial setup, Eney downloads AI models that power its various skills. Once installed, a small round bubble with two eyes appears in the lower-left corner of the screen—cute and unmistakable, this is your AI companion. You can summon it by clicking the icon or by double-tapping the Option key.
Eney’s core selling point is completing tasks on macOS through natural language. You can speak to it as you would to a person, telling it what you want to do, and it will carry out the operation automatically—such as scheduling meetings, writing emails, searching files, or editing videos. For these tasks, Eney doesn’t just provide instructions; it actually performs them.
At present, however, Eney’s “intelligent companion” role isn’t quite that intelligent. In my brief testing, using natural language to “add a meeting in the macOS Calendar app” failed—it couldn’t even open macOS apps. This may be related to system permissions, but oddly enough, entering “find yesterday / today / tomorrow” does allow Eney to retrieve the corresponding calendar events and open Calendar. When it comes to launching apps, you must type the app’s name directly; using natural language like “open xx app” results in an error instead. Eney also can’t check the weather. In conversational scenarios, the model frequently hallucinates—for example, describing Heated Rivalry as a book about technology and business competition. All in all, at this stage, Eney doesn’t seem dramatically different from Siri in terms of results 😓.
Eney places strong emphasis on local-first processing and privacy protection. Its models are sourced from HuggingFace, downloaded during setup, and run locally, enabling many tasks even while offline. The downside is that model size and capabilities are limited: it can handle unit conversions, invoke CleanMyMac to clean disk space, retrieve system status, perform local searches, compress images, and so on—but the results may not always be satisfying. User data isn’t uploaded to the cloud arbitrarily; only for more complex tasks (such as long-form summaries) will cloud processing be used, and only with your consent.
Eney is currently in beta, with features under active development. The app runs only on Macs with Apple silicon and is included as part of the Setapp subscription—only Setapp members can install and use it.
AirShare: A Cross-Platform Tool for Fast File Sharing
Platform: macOS / Windows / Linux
Keywords: File transfer
@化学心情下2: In an ideal world, sending files should work just like Apple’s AirDrop—select the device you want to send to, wait for confirmation, and the transfer is done, without any complicated pairing or setup. Unfortunately, tools this seamless are still largely “Apple-only” at the moment (even though Google has achieved partial AirDrop compatibility with the Pixel 10). So when you need a similar experience across platforms, third-party tools become the only option.
AirShare is one such file transfer tool. It supports Windows, Linux, and macOS (mobile operating systems are not supported for now), making it well suited for quickly sharing files within small teams—as long as all devices are on the same local network.
When it comes to sending files, AirShare supports transferring single or multiple files, as well as entire folders over a local network. On the receiving end, when a sender initiates a transfer, you’ll see a notification in the system notification area showing who sent the files, their total size, and what files are included. Once you click Accept, the files are automatically saved to the default Downloads folder.
Within the AirShare app itself, you can also view a dedicated receive dialog that displays the sender’s device, the file list, file sizes, and total transfer size. From there, you can choose to reject, accept, or view more detailed information about the files.
AirShare offers both a free version and a Pro version. The free version supports local network transfers only, with a file size limit of 200 MB and a transfer history retained for seven days. The Pro version removes size limits and allows file transfers over the internet. You can download the client from AirShare’s official website, with support for Windows, macOS, and Linux.
App Updates You Shouldn’t Miss
Beyond brand-new apps, many familiar faces on the App Store are constantly iterating, adding more interesting and practical features. At sspai, we aim to help you filter the app updates worth paying attention to, so you can quickly catch up on what’s new from apps and their developers.
Darkroom 7.0: A New Engine and New Lighting Tools
Platform: iOS / iPadOS / macOS
Keywords: Image editing
@Snow: Darkroom, a cross-platform photo editing app available on iOS, iPadOS, and macOS, rolled out its major 7.0 update last week. The new version adopts an entirely new rendering engine, which the developers say significantly improves smoothness and stability during editing. Two new lighting tools have also been added, opening up new creative and expressive possibilities for your photos.
Within the editing tools, you’ll now find two brand-new adjustment sliders—Bloom and Halation—under the Tone section. These additions help fill a long-standing gap in Darkroom’s simulation of vintage lighting effects. Bloom is similar to the popular “black mist” filters of recent years, simulating light diffusion beyond natural boundaries to create a soft, dreamy atmosphere. Halation, on the other hand, mimics a classic film photography effect, where the edges of bright areas bleed a subtle red-orange glow into darker regions. This tool brings a warm, cinematic feel to strong light sources and high-contrast images. If you want to learn more about Bloom and Halation, Darkroom has prepared a professional guide with more detailed explanations and comparison examples.
As a long-time Darkroom user, I have to say that smoothness and stability have always been its most persistent pain points. In older versions, entering or exiting the edit view, zooming images, or working with RAW files would occasionally result in stutters or crashes. The most frustrating issue was when the entire editing preview would intermittently fail—adjustments produced no visual feedback at all, leaving force-quitting the app as the only solution.
After the update, image stability has improved noticeably. Over the past few days, I haven’t encountered crashes or complete preview failures. That said, during zoom operations, version 7.0 still suffers from intermittent preview loss and missing animations. As for perceived smoothness, things feel a bit more “mysterious.” Beyond the new rendering engine, the app has also tweaked the interaction logic when entering the edit view from the library. In the new version, you can swipe left or right to switch between images, swipe down to return to the library, and swipe up to enable editing tools. However, before enabling the editing tools, you can’t zoom with one- or two-finger gestures or tap to compare edits. Once the tools are disabled, gestures and taps work again. This design may indirectly improve perceived smoothness, but the inconsistent interaction logic hardly feels like a true optimization—at least to me.
One improvement that does deserve praise is that version 7.0 removes zoom restrictions in the editing view. You can now zoom all the way in to pixel level to check fine details, zoom out beyond “fit to view” to evaluate the overall composition, or use the Zoom option in the extended menu to quickly switch between preset ratios. The new version also supports adjusting the canvas background color, allowing you to use neutral tones to better judge color and contrast.
Version 7.0 also brings a number of upgrades to video editing. It adds support for slow motion, Apple ProRes, and 8K video, with improved loading efficiency compared to previous versions, and most adjustment tools can be reused. A new draggable timeline with timecode has been added for video clips, supporting frame-by-frame positioning as well as fast forward and rewind. A real-time histogram is also available, making it easier to fine-tune visuals with precision. Unfortunately, the stability gains seen in photo editing don’t fully carry over here. Video clips may load to a black screen or briefly go black when switching between portrait and landscape; the histogram overlay can disappear intermittently; certain adjustments—such as distortion correction—can cause preview glitches; and there are even inexplicable auto-play issues. In short, video editing still falls well short of being reliably usable.
Overall, the two new lighting tools introduced in Darkroom 7.0 are genuinely useful, and image editing stability and smoothness have improved. However, the slight changes in interaction logic may take some getting used to for long-time users. While video editing has seen major feature upgrades, its stability issues mean it’s still not recommended as a primary tool. For a more detailed breakdown of what’s new in Darkroom 7.0, you can check the official changelog on the website.
Paste: A Brand-New Power Search with Keyword Prediction, Smart Filters, and Image Search
Platform: iOS / iPadOS / macOS
Keywords: Clipboard management
@Vanilla: Paste can be considered one of the most long-standing clipboard managers on macOS. Its biggest strength lies in its UI and interactions, which feel almost like a native app and deliver an excellent user experience. Of course, as Apple has introduced system-level clipboard features of its own, it has become harder for Paste to rely solely on a “system-like” experience to convince users to pay for a subscription to a similar tool. As a result, Paste has continually tried to add features not available in the system clipboard, such as Pinboard, sharing, and previews—but for someone as frugal as me, that still wasn’t quite persuasive enough. Recently, Paste introduced Power Search, comprehensively upgrading clipboard search across three dimensions: keyword prediction, smart filtering, and image search. In terms of functionality, it can be said to fully surpass the system clipboard.
First, Power Search adds keyword prediction. When we type Chinese characters or letters into the search box, Paste performs real-time predictive matching based on the input. It also applies intelligent error correction to the keywords: if no results exactly match what you typed, Paste will still try to surface similar matches.
Second, Power Search supports smart filtering. We can directly enter parameters in the search box—such as file type, app name, date, or device—or click the filter button on the right side of the input field to select them. By using these filters alone or in combination with keywords, we can conduct precise searches and more quickly locate the clipboard history we’re looking for.
Finally, Power Search supports image recognition. When we copy photos, screenshots, or documents, Paste can recognize the text within them, allowing us to search using keywords and even highlight matched text in the preview view. However, in my own testing, I was unable to reproduce this feature when searching Chinese or English text in photos and screenshots, despite what was mentioned in the official update. I can’t be sure whether this is a bug or just an isolated issue on my end.
Image source: Paste official blog
After this update, I believe Paste offers plenty of motivation for heavy Mac users—such as writers, designers, developers, students, and researchers—to give this clipboard enhancement tool a try. It not only manages clipboard content effectively, but more importantly enables fast and precise searching through that content.
Paste can be downloaded from the App Store and requires a subscription to use. Pricing is 28 RMB per month or 168 RMB per year; the family plan costs 398 RMB per year; and a one-time lifetime purchase is 598 RMB, covering both Mac and iOS clients. You can also purchase team plans based on seats, or access Paste for Mac and Paste for iOS through a Setapp subscription.
Xiaoxing Accounting 4.0: Fully Adapting to Material 3
Platform: Android
Keywords: Expense tracking
@大大大K: The Android expense-tracking app Xiaoxing Accounting rolled out its 4.0 Beta major update last Sunday. The biggest change in this release is its long-awaited full adaptation to Material 3.
Upon entering the main interface, you’ll notice the new UI feels more “standardized.” The previously immersive homepage data display has been replaced with a card-based layout; search is now pinned to the top as a search bar; and the quick-access entry at the bottom of the screen has been expanded to include templates, a calendar, expense reimbursement management, and more.
As for core expense-tracking features, the overall layout hasn’t changed much, and the original interaction logic still applies. The most notable improvement is the numeric keypad. In earlier versions, tapping the calculation button would pop up a separate calculator keypad, which felt somewhat disjointed. In the new version, the calculation button appears directly within the existing numeric keypad and is accompanied by a scaling animation, making the recording flow noticeably smoother.
The settings interface has also been reworked in Xiaoxing Accounting 4.0. In previous versions, all settings were listed at the top level and divided by tags. While this made items quicker to find, the accumulation of options often felt cluttered. Version 4.0 organizes settings into categories such as ledgers, display, recording settings, and reports, and many options have been reassigned within the hierarchy. Since settings are typically accessed infrequently, this categorized structure is much clearer from a usability standpoint.
For more detailed changes, you can refer to the update log. It’s worth noting that this release is still a Beta version, and there are some minor bugs related to data calculations and UI display. If your current version is below 3.7, it’s recommended that you first update to 3.7 and manually back up your data before trying the 4.0 update, to avoid potential data incompatibility issues. You can now upgrade to Xiaoxing Accounting 4.0 Beta for free on Coolapk.
FocusFour: When Your To-Do List Blurs Your Focus, the Four Quadrants Make Things Clearer
Platform: HarmonyOS / iOS
Keywords: Productivity, Four Quadrants, SSPAI
@轻舟_: If you’ve followed FocusFour on SSPAI before, you may have already witnessed its journey from a co-creation project to release. Recently, FocusFour has launched a native HarmonyOS version. If most of the time your problem isn’t that you don’t know what to do, but that you don’t know what to do first, then FocusFour—built around the “Four Quadrants of Time Management”—might be worth a try.
As a task management tool, FocusFour doesn’t try to hold everything for you. Instead, it repeatedly reminds you, in a very intuitive way, to distinguish between what is important and what is urgent. In FocusFour, tasks are no longer arranged simply by time, but placed into four quadrants. This approach is straightforward and grounded in reality, quickly making you realize that truly important things are actually quite few. Compared with endlessly ticking off completed items, this view is more about helping you make choices than about keeping records.
The newly released HarmonyOS version of FocusFour currently focuses on local data. There’s no complicated setup and no extra learning cost, and it’s completely free. When you open the app, the first thing you do isn’t figuring out “how to use it,” but deciding “where to place this task.” The overall experience of the HarmonyOS version continues the design philosophy of the earlier co-creation release: a clean interface, natural interactions, and no attempt to trap you with features.
If you relate to any of the following:
You have lots of things to do, but spend every day busy with “not-so-important” tasks
Important tasks are always pushed to tomorrow
You’ve tried many tools, but in the end only stick with a single list
You want a tool that helps you make judgments rather than adding more burden
And you also happen to be using a HarmonyOS device, then FocusFour may be worth trying—especially since there aren’t many apps of this type on the HarmonyOS platform. FocusFour will continue to evolve, keeping feature parity with the iOS version, and will add capabilities such as task analysis, widgets, and notification reminders. When it comes to task management, doing a little less can sometimes be more effective.
Mozilla Firefox (Windows | macOS | Linux): Updated to v146. On Windows 10 devices, users can enable the browser backup feature, which backs up browser data such as passwords and bookmarks to the local device on a daily basis. On macOS, an independent GPU process is now enabled by default, among other changes.
Market regulator drafts compliance guidelines to curb “selling cars at a loss”
Motion-sensing game console Nex Playground gains popularity in the U.S.
Musk seeks evidence from Alipay in lawsuit accusing Apple of monopoly practices
OpenAI begins selling physical merchandise
Central authorities signal continuation of national subsidies next year, with policies to be optimized
Doubao mobile assistant clarifies it cannot capture screenshots of banking keyboards
Rumors You Can Just Glance At
Market regulator drafts compliance guidelines to curb “selling cars at a loss”
On December 12, China’s State Administration for Market Regulation (SAMR) released the Automobile Industry Price Conduct Compliance Guidelines (Draft for Public Comment) and opened it for public feedback. The document emphasizes that automobile manufacturers and dealers that “sell cars at a loss” through various means will face significant legal risks.
The Guidelines first detail pricing conduct requirements for automobile manufacturers, calling for: the establishment of end-to-end pricing management systems covering vehicle sales and financial services; clear and transparent rebate policies that respect dealers’ independent pricing rights; prohibitions on price collusion, below-cost sales aimed at excluding competitors, and price discrimination; and requirements that “paid unlock” features disclose free-use periods and fee standards to protect consumers’ right to be informed.
The Guidelines also specify pricing conduct requirements for automobile dealers, including: clear price marking and strict separation of vehicle prices from sales service fees; standardized promotions; bans on deceptive pricing, false reference prices, and failure to honor pricing commitments; encouragement for platforms to provide two-way alerts for conspicuously low prices; and strict prohibitions on charging without providing services, duplicate charges, and passing on improper fees.
In addition, the Guidelines encourage companies to establish internal compliance management systems covering price decision-making, contract management, internal oversight, and risk prevention, so as to prevent pricing violations at the source.
In response, automakers such as Changan Automobile, BYD, BAIC Group, and XPeng have successively stated their support and indicated they will actively comply with the Guidelines.
Since the beginning of this year, regulators have continued to roll out measures to curb “involution” in the auto industry. In March, a senior official at the National Development and Reform Commission criticized disorderly competition in the car market, noting that some companies were sacrificing profits to grab market share, and pledged efforts to rectify market chaos and regulate competition. In May, an official from the Ministry of Industry and Information Technology said oversight of involution-style competition would be intensified. In June, under regulatory guidance, automakers committed to shortening supplier payment terms to 60 days; in September, the China Association of Automobile Manufacturers further issued initiatives to standardize practices.
Motion-sensing game console Nex Playground gains popularity in the U.S.
According to The Wall Street Journal, during the 2025 U.S. holiday shopping season, a motion-sensing game console called Nex Playground unexpectedly emerged as a market “dark horse.” Launched by Silicon Valley startup Nex, Nex Playground topped Amazon’s Toys & Games rankings during Black Friday and sold out at brick-and-mortar retailers such as Walmart and Target. Data shows that its single-week sales surpassed Microsoft’s Xbox, trailing only Sony’s PS5 and Nintendo’s Switch 2, ranking third in overall U.S. game hardware sales.
Priced at $249 (approximately RMB 1,757), Nex Playground’s core appeal lies in completely abandoning traditional controllers in favor of capturing player movements via built-in cameras and AI-powered computer vision. Unlike conventional consoles, the product targets family and parent–child scenarios, encouraging kids to “get moving” in the living room rather than slumping on the couch—an approach that has resonated strongly with parents. In addition, Nex secured the IP licensing rights to Bluey, a children’s animated series with exceptionally high viewership on streaming platforms, which became a key driver behind the surge in sales.
Nex’s success stems from a high-risk business pivot. The company originally started by developing an AI basketball training app and once received investment from NBA star Steve Nash. After recognizing the market ceiling of a single sports-focused app, the team leveraged its core vision technology to first pivot into AR game development, and then crossed over into the supply-chain-intensive hardware sector, aiming to fill the gap left by the discontinuation of the Nintendo Wii in the motion-gaming market.
The commercial returns from this transformation have been striking. Two years ago, the device sold only 5,000 units; this year, sales during Black Friday and Cyber Monday surpassed 300,000 units, with full-year sales projected to reach 600,000. Company revenue is expected to grow from roughly $3 million annually to $150 million, with profitability anticipated for the first time.
Musk seeks evidence from Alipay in lawsuit accusing Apple of monopoly practices
According to 9to5Mac, newly disclosed court filings show that in xAI’s antitrust lawsuit against Apple and OpenAI, Elon Musk’s company sought documentary evidence from foreign tech firms including China’s Alipay and South Korea’s Kakao. The aim is to argue that Apple preserves its iPhone hardware monopoly and high pricing by restricting the growth of “super apps.”
In the case, xAI alleges that Apple not only entered into exclusive agreements with OpenAI and suppressed the visibility of competitors such as Grok in the App Store, but also used ecosystem rules to block “super apps” that integrate instant messaging, payments, transportation, and e-commerce. Musk’s team contends that such apps reduce users’ dependence on a specific operating system, making it easier for consumers to switch from iPhones to other brands, and are therefore deliberately constrained by Apple.
Under evidence-gathering requests sent to judicial authorities in China and South Korea pursuant to the 1970 Hague Evidence Convention, xAI asked Alipay and KakaoTalk to provide their App Store ranking histories, revenue data, and internal strategic assessments on “how super apps affect users’ willingness to switch smartphones.” Under the Hague Convention, judicial authorities in one contracting state may request evidence from the competent authority of another contracting state in civil or commercial matters. xAI is expected to submit similar requests to other leading Asian app developers, including WeChat and Grab.
Musk has long sought to transform the social platform X (formerly Twitter) into a Western-style “super app” akin to WeChat. While the lawsuit initially appeared to stem from a dispute over Grok’s App Store ranking, the expanding scope of evidence collection suggests the case’s core has clearly shifted toward Musk’s ambitions for an all-in-one app.
OpenAI begins selling physical merchandise
According to Business Insider, to mark its 10th anniversary, OpenAI has opened its official merchandise store, OpenAI Supply Co., to the public for the first time. Previously, the store was accessible only to internal employees. After OpenAI shared the store link on X, it quickly sparked a buying frenzy, with most popular sizes of hoodies and T-shirts already sold out.
The store currently lists 10 items available for purchase, largely themed around OpenAI’s technological vision and geek culture. For example, there is a set of Pokémon-style collectible trading cards featuring models such as Sora 2 and GPT-5, as well as several apparel designs incorporating elements of “artificial general intelligence” (AGI). Beyond the items for sale, the website also showcases a large archive of past merchandise, including hats printed with the “1-800-CHATGPT” phone number and retro designs featuring San Francisco city motifs.
Central authorities signal continuation of national subsidies next year, with policies to be optimized
According to Yicai, the Central Economic Work Conference held in Beijing from December 10 to 11 made it clear that next year China will “adhere to demand-driven growth and build a strong domestic market,” and will optimize the implementation of the “Two New” policies.
The so-called “Two New” policies refer to large-scale equipment upgrades and the trade-in program for consumer goods, the latter of which is commonly known as the “national subsidy.” The conference’s clarification that the “Two New” policies will be optimized in 2026 indicates that national subsidies will continue next year, though with adjustments to specific measures.
To boost consumption, China issued RMB 150 billion in ultra-long-term special government bonds in 2024 to support consumer goods trade-ins. Given the strong results that year, China doubled the amount in 2025, issuing RMB 300 billion in ultra-long-term special bonds for the same purpose. The 2025 national subsidy program also expanded on the 2024 scope, adding three categories of digital products—smartphones, tablets, and smartwatches/bands—as well as four categories of home appliances, including microwave ovens, water purifiers, dishwashers, and rice cookers.
Data from the Ministry of Commerce shows that from January to November this year, consumer goods trade-ins drove related product sales exceeding RMB 2.5 trillion, benefiting more than 360 million instances of consumers. Among them, over 11.2 million vehicles were traded in, more than 128.44 million home appliances were replaced, subsidies covered over 90.15 million purchases of phones and other digital products, more than 12.91 million electric bicycles were traded in, and over 120 million home renovation and kitchen/bathroom “refresh” items were upgraded.
Experts interviewed said that the national subsidy budget in 2026 may see a moderate increase over 2025 levels, with further optimization of funding allocation—potentially expanding into service consumption—to better leverage fiscal spending to stimulate consumption, promote industrial transformation and upgrading, and support economic growth.
Doubao mobile assistant clarifies it cannot capture screenshots of banking keyboards
On December 13, in response to recent online claims that the “Doubao Mobile Assistant” exploits special permissions to steal private information such as banking keyboard inputs, the Doubao team issued a statement clarifying that this interpretation is incorrect. The company said the assistant uses the system’s native screenshot APIs and strictly follows Android’s underlying security mechanisms, making it impossible to capture protected interfaces such as those used by banking apps.
The core of the controversy centers on the use of permissions such as READ_FRAME_BUFFER. Doubao explained that the assistant employs a “virtual screen” technology, running third-party apps within a virtualized environment. While this permission allows reading the graphics buffer for AI inference, on Android, once an app invokes the FLAG_SECURE interface (as is typically the case during sensitive steps like password entry), the system returns black-screen data at the graphics rendering layer for any screenshot or screen-recording request. As a result, even with screen-reading permissions, no actual content can be seen.
Regarding the CAPTURE_SECURE_VIDEO_OUTPUT permission, Doubao said its practical purpose is to ensure that users can visually view protected pages normally when operating in small-window modes such as the “Dynamic Island,” rather than seeing a completely black screen. This, however, does not change the page’s inherent screenshot-blocking properties.
Doubao added that the mobile assistant currently operates on a workflow of user command → screenshot upload to the cloud → model feedback. Because the multimodal large models required for interface analysis have massive parameter sizes and are constrained by current mobile chip performance, screenshots must be uploaded to the cloud for processing. The company emphasized that screenshots are taken only when explicitly triggered by user commands, and uploaded data is used solely for real-time visual understanding and reasoning, with no cloud storage after the task is completed.
The technical preview version of the “Doubao Mobile Assistant” was released on December 1 and preinstalled on the nubia M153 engineering prototype jointly developed by ByteDance and ZTE, which was sold in limited quantities. On December 3, after encountering risk-control restrictions that prevented normal use of WeChat, Doubao removed the assistant’s ability to operate WeChat. On December 5, Doubao announced plans to standardize and adjust certain AI phone-operation capabilities, including—but not limited to—score farming, incentive farming, and scenarios involving financial apps and games.
Rumors You Can Just Glance At
According to Macworld, a leaked internal build of iOS 26 reveals some of Apple’s future development plans. Among them, iOS 26.4—expected to be released in spring 2026—will include a revamped Siri integrated with Apple Intelligence and based on Google’s Gemini model, along with a redesigned Health app interface. The code also indicates Apple is developing a new verification system that checks device integrity before Apple ID sign-in, meaning devices that have been “jailbroken” or otherwise modified without authorization may be unable to access online services such as iCloud. In addition, the Apple TV app may add a “Sports” subscription, and the Freeform app is set to gain folder management features.
On the hardware side, AirPods Pro 3 will leverage the built-in U2 ultra-wideband chip to support “outdoor Precision Finding” within the Find My network. Looking further ahead, iOS 27 is expected to improve the Photos app’s “Featured” functionality and AirPods pairing mechanisms, while iOS 28 plans to add sleep-tracking metrics such as “time in bed” for Apple Watch. macOS 28 is also slated to receive the Health app.
Amazon to Allow Authors to Publish Kindle Books Without DRM Protection
OpenAI Announces Integration with Select Adobe Apps
Tencent QQ Group Chats Add Support for Yuanbao AI Summary Feature
Alibaba Releases Qwen3-Omni-Flash Multi-Modal Model
Qualcomm Announces Acquisition of Ventana Micro Systems
Rumors You Can Just Glance At
OpenAI Releases GPT-5.2
On December 12, OpenAI announced the launch of its GPT-5.2 model lineup, positioning it as the most suitable model for everyday professional use. According to OpenAI, GPT-5.2 delivers overall improvements in table generation, presentation creation, coding, image understanding, long-context processing, tool use, and complex multi-step tasks.
The GPT-5.2 series includes three versions: Instant, Thinking, and Pro. Among them, the Thinking model stands out for significantly reducing hallucinations, making it more suitable for professional AI agent applications that require high reliability. Companies such as Notion, Box, Shopify, Harvey, Zoom, and Databricks participated in pre-release testing. OpenAI also noted that GPT-5.2 will play a more central role in AI-agent workflows, with the long-term goal of shaping ChatGPT into a highly personalized intelligent assistant. Source
Logos for OpenAI and the Walt Disney Company in white on a background of swirling purples, blues, and magenta OpenAI also announced a three-year licensing agreement with Disney. Sora and ChatGPT Images will gain the ability to generate video and image content featuring characters from Disney, Marvel, Pixar, and Star Wars. The related features are expected to roll out in early 2026, with some generated content showcased on Disney+. Source
In addition, Disney has issued a cease-and-desist letter to Google, claiming that Google used Disney works without permission to train AI models, and subsequently generated and distributed images and videos featuring Disney-related characters for commercial purposes. Source
Snapmaker U1 Color 3D Printer Launches
On December 12, the color 3D printer Snapmaker U1 made its debut on JD.com. The Snapmaker U1 features the SnapSwap™ independent four-head parallel system, supporting multi-color and multi-material printing with quick toolhead switching. According to the company, in multi-color printing scenarios the device can switch toolheads in about 5 seconds, improving efficiency and reducing material waste. Compared with traditional single-nozzle color-changing methods, Snapmaker U1’s four-color printing tasks can reach speeds of up to 500 mm/s and complete prints within 5–6 hours. While achieving up to a fivefold increase in overall efficiency, the design also reduces material usage by about 80% thanks to its purge-free channel structure.
Snapmaker U1 supports automatic calibration, input shaping, pressure advance, automatic bed leveling, and other features to ensure clean edges, smooth walls, and sharp details even in high-speed printing mode. Its printing parameters include a build volume of 270×270×270 mm, along with AI print monitoring and app-based management.
Snapmaker U1 is priced at 5,299 RMB exclusively on JD.com (4,504.15 RMB after national subsidies). Buyers can still claim an exclusive privilege package before 7 p.m. on December 12, with pre-orders officially opening at 8 p.m. The Snapmaker U1 will also be showcased at an offline experience event at the JD MALL Beijing Shuangjing store from December 16 to 17. Source
Amazon to Allow Authors to Publish Kindle Books Without DRM Protection
Amazon has announced that beginning January 20, 2026, authors who choose not to enable DRM (Digital Rights Management) in the Kindle Direct Publishing workflow will be able to offer their e-books to readers in more open formats such as EPUB and PDF. Previously published books will not have their DRM settings changed automatically; authors must manually update their preferences in the KDP dashboard. Once an author updates the DRM status, the change may take up to 72 hours to appear on Amazon’s main site.
Amazon stated that the new option is intended to make it easier for readers to access content they have purchased. However, some authors believe the move may actually increase the number of people choosing to enable DRM. Source
OpenAI Announces Integration with Select Adobe Apps
On December 11, Adobe launched three applications—Photoshop, Acrobat, and Adobe Express—inside ChatGPT.
The new integration allows users to perform image editing, PDF modification, and various design tasks directly within ChatGPT using natural language, without switching to standalone applications. Users only need to mention the app in conversation and describe what they want; the system will then generate selectable results or present interactive UI controls, such as sliders for adjusting brightness and contrast.
Although these integrated apps do not offer the full capabilities of their desktop counterparts, Photoshop can still perform local edits, creative effects, and image parameter adjustments; Acrobat can edit PDFs, compress or convert formats, extract text and tables, and merge files; Adobe Express can generate and edit posters, invitations, and social media graphics. All creations can be exported at any time to Adobe’s native software for further refinement.
According to Adobe, these ChatGPT applications are currently available on desktop, web, and iOS. On Android, only Adobe Express is supported for now, with Photoshop and Acrobat coming later. Source
Tencent QQ Group Chats Add Support for Yuanbao AI Summary Feature
On December 11, Tencent QQ announced via its official Weibo account that group chat messages now support the Yuanbao AI summary feature. When a group accumulates a large number of unread messages, users can tap “AI Summary” in the upper-right corner of the chat interface, and—after granting permission—generate a summary of the conversation. QQ also stated that the summary results are visible only to the user, and the related data will not be stored or used in any other context. Source
Alibaba Releases Qwen3-Omni-Flash Multi-Modal Model
On December 9, 2025, Alibaba’s Qwen team released the next-generation native full-modal model, Qwen3-Omni-Flash-2025-12-01.
Upgraded from Qwen3-Omni, the new model supports unified input across text, images, audio, and video, and provides real-time streamed text output and voice capabilities. This update focuses on improving issues in multi-modal interactions—such as unnatural speech performance and unstable transitions in multi-turn conversations—by adjusting speech rate, pauses, and prosody to enhance overall interaction quality.
Alibaba has also strengthened the controllability of Qwen’s system prompts. Users can now define the model’s behavior patterns, including tone, style, and content length, with improved execution accuracy. In addition, Qwen’s multilingual capabilities continue to expand, now supporting 119 text languages, 19 speech recognition languages, and 10 speech synthesis languages.
In terms of performance, Qwen3-Omni-Flash-2025-12-01 shows improvements over the previous version on several benchmarks, including ZebraLogic (logical reasoning), LiveCodeBench-v6 (code generation), and MMMU (multidisciplinary visual question answering). Source
Qualcomm Announces Acquisition of Ventana Micro Systems
On December 10, Qualcomm announced its acquisition of Ventana Micro Systems, further strengthening its technological presence in the RISC-V domain.
Qualcomm stated that the Ventana team will complement its existing Oryon in-house CPUs and RISC-V development efforts, jointly advancing the company’s technologies across business lines in the AI era. The company also noted that the RISC-V architecture holds significant potential for driving CPU innovation, and this acquisition marks an important step toward building industry-leading RISC-V CPU technology.
Founded in 2018 and headquartered in Cupertino, Ventana Micro Systems focuses on high-performance, scalable, and secure computing chip solutions based on the open RISC-V architecture. The company is a board member of RISC-V International as well as a member of its Technical Steering Committee. Source
Rumors You Can Just Glance At
During a televised broadcast, Donald Trump mistakenly described 6G technology as something capable of displaying deeper image or skin-detail layers, failing to understand that 6G actually refers to the next-generation mobile communication network. His comment drew attention on site, while Qualcomm CEO Cristiano Amon—who was on stage with him—did not correct the statement. As early as 2019, Trump had called for the U.S. to accelerate 6G development to avoid falling behind other nations. He has also made past remarks such as “the Moon belongs to Mars” and incorrectly linking Apple to 5G infrastructure construction. Source
The startup Operation Bluebird has filed a petition with the U.S. Patent and Trademark Office, arguing that Elon Musk’s X Corp. effectively abandoned the “Twitter” and “Tweet” trademarks after rebranding and discontinuing use of the associated branding, and that the trademarks should therefore be canceled. The company has also applied for a new “Twitter” trademark for its upcoming social platform, Twitter.new. In its petition, Operation Bluebird claims that after the rebrand to “X,” X Corp. no longer uses the original trademarks, citing Musk’s previous public statements about “saying goodbye to the Twitter brand forever.” Some legal experts note that although the claim has some basis, the “residual goodwill” of the Twitter brand could be a key factor, as many users still refer to X as “Twitter” and continue using terms such as “tweet.” X Corp. must submit a response by February next year. Operation Bluebird stated that it stands firmly by its position and will continue pursuing relevant legal actions. Source
According to tech outlet MacRumors, internal firmware for iOS 26 reveals:
The new AirTag is expected to receive upgrades across pairing workflow, battery notifications, and tracking capabilities. In particular, it aims to deliver more accurate location tracking when the device is in motion or in crowded environments. The new AirTag may also feature a next-generation ultra-wideband (UWB) chip, expanding the range of Precision Finding. There have been no clear rumors of noticeable exterior redesigns; the overall appearance may remain consistent with the current model, though improvements to the speaker structure could reduce risks of malicious tampering or misuse. Although iOS 26 refers to the device as “2025 AirTag,” current product timing suggests its official release may slip to early 2026. Source
Apple is advancing development of a new smart home hub device. The product will include a camera and support Face ID for identity verification and user recognition, enabling automatic switching between user profiles. The hub is expected to rely heavily on Apple Intelligence and the next-generation Siri set to launch next year. Based on current reports, this smart home hub may take a form somewhere between an iPad and a HomePod, featuring an approximately 7-inch square display and an optional speaker dock. The device is expected to debut alongside iOS 26.4 and the updated Siri.
Code references also reveal an unreleased accessory device codenamed J229, believed to be a home security camera under development by Apple, though its release timeline has not yet been disclosed. Source
