Mozilla and Anthropic discover multiple Firefox security vulnerabilities using Claude
Briefs worth a quick look
Alibaba DAMO Academy releases MAOSS AI model for fatty liver screening
On March 9, Alibaba DAMO Academy announced in a blog post that it has jointly developed an AI model for fatty liver screening called MAOSS, in collaboration with Shengjing Hospital of China Medical University and Nanjing Drum Tower Hospital affiliated with Nanjing University. The model uses common diagnostic data such as non-contrast CT scans and serum indicators to accurately screen liver fat stages and assess the progression of liver fibrosis. It increases the detection rate of high-risk patients from 16.6% to 52.4%, providing early warning signals for intervention and enabling earlier-stage management of chronic liver disease. The research paper has been published in Nature Communications. Source
ASUS launches NUC 16 Pro mini PC
On March 9, ASUS launched the NUC 16 Pro mini PC on domestic e-commerce platforms in China. The device features an Intel Core Ultra X7 358H processor, 32 GB RAM, and 1 TB storage, priced at 10,999 yuan. The unit measures 14.4 × 11.7 × 4.2 mm and features the ASUS logo on the front, along with two USB-A 3.2 Gen 2 (10 Gbps) ports and one USB-C Gen 2 (10 Gbps) port. The top-left corner of the chassis carries the NUC branding. On the rear, the device provides two Thunderbolt 4 ports, two HDMI 2.1 ports (optionally replaceable with two DisplayPort outputs), two USB-A 3.2 Gen 2 (10 Gbps) ports, and two 2.5G Ethernet ports. The system is powered by a 16-core Intel Core Ultra X7 358H processor and uses a dual-fan cooling system, with 32 GB of onboard LPDDR5x 8533 MHz memory. It supports PCIe 5.0 SSDs and offers tool-free quick upgrades. Networking includes Wi-Fi 7 and Bluetooth 6.0 support. Source
Microsoft launches Microsoft 365 E7 subscription with new AI features
On March 9, Microsoft introduced a range of new AI-related services as part of its “Frontier Transformation” initiative, all of which will be included in a new subscription tier called Microsoft 365 E7: The Frontier Suite, scheduled to launch on May 1. The third wave of updates for Microsoft 365 Copilot will roll out across Word, Excel, PowerPoint, and Outlook. The update introduces a new Work IQ intelligence layer that considers employees’ working styles, collaborators, and task contexts to provide deeper integration within the workplace environment. In addition, Copilot Chat at this subscription tier incorporates Anthropic’s Claude model, while OpenAI’s models will also be upgraded to their latest generation. The AI assistant Claude Cowork will be integrated into the Microsoft 365 Copilot service as Copilot Cowork. Microsoft also announced the Agent 365 control platform, which allows organizations to monitor and manage the status of AI agents across the entire organization. The Microsoft 365 E7 subscription is priced at $99 per user per month. Source
Recently, OpenAI announced the launch of an application security agent called Codex Security. According to OpenAI, Codex Security can build deep contextual understanding of a project, identify complex vulnerabilities that other tools may miss, and provide higher-confidence analysis results along with remediation suggestions. Codex Security was previously known as the Aardvark feature, which underwent limited private testing with a small group of customers last year. Compared with the earlier test version, the current system has reduced the rate of falsely inflated vulnerability severity ratings by more than 90%, while the false positive rate across full codebase scans has decreased by 50%. Codex Security is now being rolled out to ChatGPT Enterprise, Business, and Edu customers, with a one-month free trial available. OpenAI has also begun inviting the first group of open-source maintainers to join the Codex for OSS program, offering free ChatGPT Pro and Plus accounts, code review support, and access to Codex Security tools to support the open-source ecosystem. Source
Mozilla and Anthropic discover multiple Firefox security vulnerabilities using Claude
Recently, Anthropic published a blog post announcing that its researchers collaborated with Mozilla researchers to use Claude Opus 4.6 to identify 22 security vulnerabilities in the Firefox browser within two weeks. Among them, 14 were classified by Mozilla as high-severity vulnerabilities, representing roughly one-fifth of the total number of high-severity Firefox vulnerabilities fixed throughout the entire year of 2025. The vulnerabilities include a Use-After-Free memory flaw that could allow attackers to overwrite data with arbitrary malicious content. All of these vulnerabilities have been patched in Firefox 148.0. Anthropic also conducted additional testing on Claude’s ability to create exploit code based on discovered vulnerabilities. The results showed that Opus 4.6 is significantly stronger at discovering vulnerabilities than at exploiting them. Anthropic also stated that Claude Opus 4.6 has been used to discover vulnerabilities in major open-source projects such as the Linux kernel. Source
Briefs worth a quick look
According to Z Finance, Tencent is reportedly developing a one-click launcher package for OpenClaw called QClaw, which allows users to quickly link the tool with QQ and WeChat after installation. Meanwhile, The STAR Market Daily reports that WeCom (Enterprise WeChat) has officially pushed a notification announcing support for OpenClaw integration, enabling quick access through OpenClaw and allowing data to be written directly into Smart Sheets. SourceSource2
On March 8, Baidu AI Cloud held OpenClaw developer pop-up events in several cities including Beijing and Shanghai, providing on-site OpenClaw installation for participants. Source
On March 9, ByteDance’s Volcano Engine officially launched ArkClaw, a cloud-based SaaS version of OpenClaw that can be used directly through a web interface. ArkClaw already supports multiple mainstream instant messaging apps and is compatible with models such as GLM, MiniMax, and Kimi. Users in the Volcano Ark Coding Plan can gain early access. Source
On March 9, Tencent Cloud launched the full-scenario AI agent platform WorkBuddy. According to Tencent, WorkBuddy is fully compatible with OpenClaw skills, includes more than 20 built-in skill packages, and supports the MCP protocol. It enables multi-window, multi-agent parallel workflows. The domestic version supports models such as Hunyuan, DeepSeek, GLM, Kimi, and MiniMax. Supported integrations include tools such as WeCom, QQ, Feishu, and DingTalk. Source
On March 9, Weibo integrated Kimi Claw, which can now be used through Weibo direct messages via the Weibo Lobster Assistant. Source
On March 9, the National Supercomputing Internet Platform announced an upgrade to its OpenClaw services, enabling interoperability between Feishu and WeCom. Source
On the evening of March 9, Nexperia China announced that its self-developed “12-inch platform” had successfully achieved small-batch mass production of bipolar discrete devices on 12-inch wafers. However, the original announcement published on Nexperia Semiconductor’s official WeChat account was deleted shortly afterward. Earlier, on March 6, Nexperia China had issued a notice stating that Nexperia B.V. had disabled all office accounts of employees in China on March 3, causing operational disruptions in the region. The Chinese IT and business teams subsequently activated contingency plans and have largely restored operations. Source
On March 9, the X platform added a new option for image posts called “Prevent Grok from Editing.” Testing indicates that the feature only prevents other users from editing images in replies using the @Grok function. Source
Welcome to this episode 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 apps worth checking out, or topics you’d like us to cover, feel free to join the discussion in the comments.
ScreenKite: A Lightweight and Efficient Screen Recording Tool
Platform: macOS
Keywords: screen recording, camera movement, focus, zoom, quick export
ScreenKite is a lightweight yet powerful screen recording app for macOS. Its key features include automatic zooming, fast export, a built-in editor, and transcription-based subtitle generation. It is particularly suitable for scenarios such as product demonstrations, software tutorials, and course recordings.
Traditional screen recording apps simply capture the screen or a window. ScreenKite takes a different approach by emphasizing its automatic zoom feature. During recording, the app detects mouse clicks and keyboard actions, automatically zooming in on relevant areas and focusing on important content, making it easier for viewers to see key operations. Notably, this feature works automatically without requiring manual post-production editing.
ScreenKite also includes a simple yet capable video editor. After completing a screen recording, users can perform basic editing directly within the app without importing the footage into other video editing software. The editor uses a traditional layout: the preview area is placed in the center, the timeline below allows manual adjustment of the automatic zoom track, and the settings panel on the right provides a variety of configuration options.
Within the background settings, users can also change the recording background. ScreenKite provides classic macOS desktop wallpapers from different versions, along with solid color and gradient backgrounds. It also offers customization options such as blur effects, padding adjustments, rounded corners, and shadow effects. The automatic zoom feature allows further customization, including zoom ratio, delay effects, and automatic cursor following. Cursor styles include pointer, text, hand, and others, with options to change colors and configure the zoom effect when clicking.
In addition, ScreenKite offers practical features such as recording with a webcam overlay and generating subtitles through transcription. For tasks like recording lessons or demonstrating product workflows, these features can significantly simplify post-production and help produce professional-looking videos quickly.
Beyond the convenience of recording, ScreenKite also exports videos very quickly. It leverages macOS’s native ScreenCaptureKit framework along with Metal hardware acceleration, enabling extremely fast export speeds. On Apple Silicon devices in particular, the performance is even more noticeable—users can obtain the finished video file just seconds after stopping the recording.
ScreenKite is currently in Beta, and all features are available for free. No account is required, and exported videos contain no watermark. You can download and try it directly from the official website.
ShizuWall: A Network Access Controller That Requires No Root and Doesn’t Occupy the VPN Channel
Platform: Android
Keywords: network access control
Although most Chinese-market devices now include built-in network access controls, if you are an AOSP enthusiast or a Pixel user who wants to manage whether an app can access the internet, your options are usually limited to rooting the device or installing tools like NetGuard that rely on the system’s VPN interface. The former has become increasingly difficult, while the latter may prevent you from connecting to your company’s internal network via VPN.
ShizuWall, an open-source utility, solves this problem in a more elegant way with the help of the “万能的 Shizuku.” The app fully adopts the Material You design language, with dynamic color support that keeps the interface clean and visually pleasing. The home screen displays a list of all installed apps, allowing you to quickly search for and select the ones whose network access you want to control.
ShizuWall implements its functionality through Shizuku by using system shell commands to toggle an app’s network access rather than intercepting data packets directly. As a result, it does not conflict with VPN functionality. You can block network access for apps that raise privacy concerns while still maintaining a VPN connection to your company’s internal servers, such as when logging into office systems.
ShizuWall provides four firewall control modes. The default and whitelist modes determine whether selected apps are blocked from or allowed to access the network—so the default mode can effectively be seen as a “blacklist.” However, in these two modes the control list can only be updated when the firewall is disabled. The “Adaptive” mode allows updating selected items while the firewall is active, though it may occasionally fail to refresh the cache successfully.
There is also a testing mode called “Smart Foreground.” In this mode, only the currently active foreground app is allowed to access the internet. As a result, messaging apps such as WeChat will not receive background messages, making this mode more suitable for secondary devices.
Beyond the core toggle controls, ShizuWall also provides ADB broadcast command support for advanced users. Combined with tools such as Tasker or Home Assistant, it enables scenario-based automation. It also supports monitoring newly installed apps, allowing you to update your control list promptly.
One thing to note is that the network permissions controlled by ShizuWall will reset after the device reboots (Shizuku itself may also lose permission after a restart), so you will need to re-enable it. If you are using a near-stock Android system or looking for a lightweight and practical network control tool, you can download ShizuWall for free on GitHub and give it a try.
Although using a trackpad on macOS generally provides a better experience than using a mouse, there are still many situations in daily work where a mouse is unavoidable.
Some major peripheral manufacturers (such as Logitech) provide comprehensive drivers and configuration tools for their own products. But if you’re using a mouse from another brand and want to customize scrolling behavior or button actions, you’ll have to rely on third-party applications. For example, I recently tried optimizing the experience of using the Keychron M6 on macOS. Since this mouse doesn’t have a configuration tool like Logitech Options Plus on macOS, I had to rely on third-party software to customize buttons and adjust scrolling behavior.
Mos is another open-source solution I recently discovered. It is designed specifically for configuring mouse behavior on macOS. It allows deep adjustments to scroll wheel feedback and extends into a tool that supports various custom mouse button actions. In Mos’s preferences panel, the second tab contains detailed scrolling settings. For example, it supports enabling smooth scrolling and inverted scrolling. Once enabled, additional options appear—for smooth scrolling you can enable smooth vertical scrolling, smooth horizontal scrolling, and even simulate the trackpad scrolling experience. Inverted scrolling works similarly, supporting both vertical and horizontal directions. Since I prefer the traditional scroll wheel direction, I didn’t enable inverted scrolling.
The second section of the scrolling settings focuses on general mouse behavior. For instance, the acceleration key allows you to speed up scroll wheel movement while holding the key, the transform key switches scrolling from vertical to horizontal when held, and the disable key temporarily disables smooth scrolling. These shortcuts aren’t particularly useful for my own workflow, so I left them unchanged.
The third section of the scrolling settings is particularly important. Here you can adjust parameters such as the step length of each scroll wheel movement, the tracking speed gain for continuous scrolling, and the easing duration of scrolling animations. Mos already provides fairly reasonable preset values, but you can fine-tune them according to how your mouse feels during use.
Beyond its flexible and highly customizable scrolling adjustments, Mos also allows you to customize mouse buttons. In the “Buttons” tab of the preferences panel, you can assign custom actions to various mouse buttons. For example, the two side buttons on my M6 can be mapped to Back and Forward actions—similar to how I had configured the side buttons on my previous Logitech Master 2S.
Sometimes we may also want different scrolling rules for specific applications. For instance, I prefer different scrolling behavior in my browser compared to the global settings. In Mos, you can open the “Applications” tab, add an app, then click “Scrolling” on the right side and uncheck “Inherit global settings.” This allows you to create a separate scrolling configuration for that application, such as increasing the minimum step length or speed gain.
After making all these adjustments in Mos, I finally got comfortable using the Keychron M6 on macOS. Aside from the mouse feeling a bit too light when moving, the global smooth scrolling and fine-tuned adjustments now feel much closer to the experience I had with my Logitech Master 2S. If you’re using a mouse from a smaller brand on macOS, Mos might be worth trying to fine-tune your setup. Mos is free and open-source—you can download it from the official website or install it via Homebrew.
Appboard: A Treasure Chest for Heavy App Enthusiasts
Platform: iOS, iPadOS
Keywords: apps, discovery, information, rankings
In today’s era, an app like Appboard feels a bit “retro.” That’s because it is essentially an advanced utility built around the App Store ecosystem. For industry professionals and app enthusiasts, however, it can be a very useful tool. Recently, Appboard’s Plus membership has been temporarily free, so this is a good opportunity to introduce the app.
Appboard’s core features are divided into three main categories: searching the store, quick search, and viewing rankings. The “Apps” tab serves as a centralized entry point that aggregates these functions, allowing you to access them quickly.
In the “Rankings” tab, you can choose any country or region from either a list or a map. You can then apply multiple filters based on device (iPhone or iPad), price (free or paid), and category (Apps: sports, health & fitness, medical, reference, business, books, etc.; Games: casual, sports, adventure, action, card, etc.). The result is a customized ranking list tailored to your selection.
By tapping any app in the ranking list, you can view its overview information, version release history, screenshots, description, and other details. You can also use the system translation feature to translate the description text.
Appboard also integrates with AI tools, but instead of embedding or wrapping AI features directly, it uses prompts to call external AI services for various tasks. For example, in the overview module there is a feature called “Global Rankings.” After tapping it, you can choose to query large AI models such as Claude, Kimi, or ChatGPT to retrieve the app’s ranking across different countries or regions worldwide. The results can be displayed in either list or map form, and you can export them as images or text.
Additionally, in the rankings interface you can see an “Export to AI” button in the top-right corner. Tapping it opens a prompt editing interface where you can choose preset prompts or create custom ones. The prompt, combined with the app’s detailed information, can then be exported to external AI tools for analysis.
In the “Store” tab, you can browse App Stores from all countries and regions. Tapping any country or region allows you to bookmark or unbookmark it, or switch to that store. One important feature here is the store-switching function, which lets you view different App Store content and in-app purchase information without logging into an Apple ID. However, you can only browse—you cannot purchase or download apps.
By tapping the “Search” button, you can perform advanced searches within the App Store of the currently selected country or region. You can choose different languages and devices, then match results using keywords. If you are an Appboard Pro member, you can also save search keywords, manage keyword lists, and track keyword rankings.
Overall, Appboard is a feature-rich enhanced tool for exploring and analyzing apps, cleverly extending its capabilities through AI integration. You can download it for free from the App Store. The Plus membership is currently available for free, while the Pro subscription costs RMB 15 per month, RMB 128 per year, or RMB 268 for a lifetime purchase.
MoePeek: A Lightweight Pop-Up Translation Tool
Platform: macOS
Keywords: text selection, translation, lightweight
MoePeek is a lightweight pop-up translation tool recently released. If you have previously used translation features in apps like Bob or PopClip, you will likely find MoePeek’s design easy to understand. With its built-in or customizable translation engines, it allows efficient text selection translation and screenshot translation on macOS.
The installation package of MoePeek is only 5 MB, and its overall design is very similar to early versions of Bob. It currently supports text selection translation, screenshot OCR translation, clipboard translation, and manual input translation. Once launched, the app stays in the top menu bar, and you can trigger various translation functions either by clicking the menu bar icon or by using keyboard shortcuts.
After launching the app, whenever you select text on any interface, a small “translation bubble” will appear below the selected content. Moving the cursor over the bubble or pressing ⌥+D will open the translation window, displaying one or more results depending on the translation engines you have configured.
You can also press ⌥+S to activate the OCR screenshot translation feature. The app will automatically recognize text within the selected screenshot area and extract it for translation. MoePeek includes intelligent language detection and supports 14 languages, automatically switching translation directions. This means you don’t need to carefully select text of a specific language when highlighting or capturing screenshots, lowering the barrier to use.
MoePeek integrates 16 translation services, including Apple Translate, Google Translate, DeepL, and OpenAI. Among them, Google Translate, Bing Translate, Youdao Translate, and Apple Translate can be used for free. Apple Translate runs locally and requires downloading language packs in advance, while other services require configuring an API key. For LLM engines, MoePeek provides support for Ollama and LM Studio, allowing users concerned about privacy to rely on locally deployed AI models for translation.
MoePeek is not currently available on the App Store and must be downloaded from GitHub Releases. If you find existing translation tools too heavy or are trying such apps for the first time, MoePeek is a free option worth exploring.
Muviz Widgets: Create Endless Styles for Your Home Screen
Platform: Android
Keywords: widgets, music, clock, customization
Home screen customization used to be one of the hottest topics in the Android community. In recent years, with Google’s Material You and Apple’s Liquid Glass design language, a subtle retro breeze seems to be blowing again. Muviz Widgets aims to rekindle your passion for personalization and desktop styling on your Android home screen.
As the name suggests, Muviz Widgets focuses on providing a wide variety of functional and stylish widgets to enhance your home screen. It mainly includes the following categories: clock widgets inspired by different generations of Android system styles, calendar widgets, system utility widgets, and iOS-style music widgets.
Taking the clock widget as an example, you can select any style you like and tap the plus button to adjust details such as position, size, and transparency. Color customization, however, requires unlocking the premium version. The free version only offers built-in monochrome themes by default. While the aesthetics are relatively basic, you can still improve the look by adjusting transparency.
In addition to standard clock widgets, Muviz Widgets also offers more complex widgets that combine elements such as date and weather. You can even find pixel-style widgets inspired by Nothing’s design language.
Adding a music widget follows a similar process, but you will need to grant Muviz Widgets notification access so it can read the currently playing song information. The app also allows you to select the content source to display. For example, if you only want to show notifications from a specific music player, you can simply select it in the source settings. Once added, the widget dynamically adapts its color based on the currently playing music and allows you to control playback directly from the home screen.
You can download and try Muviz Widgets from the Play Store. The app’s core features are free, while unlocking additional styles and color options requires a one-time purchase of $1.99 (early-bird price).
App Brief
WPS for Pad: On March 9, Kingsoft WPS announced that WPS for Pad has officially launched on the App Store in mainland China, with support for Simplified Chinese. The app requires iPadOS 15.0 or later. According to the company, WPS for Pad closely replicates the desktop experience, with an interface largely identical to the desktop version of WPS Office. Its core functionality is aligned with the Windows client and supports seamless collaboration across mobile devices and computers.
Apple blocks U.S. users from downloading Chinese version of ByteDance apps
Microsoft confirms next-generation console Project Helix
Gemini accused of encouraging suicide
Nvidia halts H200 production amid uncertain export prospects to China
Rising GPS interference drives development of alternative technologies
Wikipedia hit by malicious JavaScript worm
Briefs worth a quick look
Apple blocks U.S. users from downloading Chinese version of ByteDance apps
According to Wired, Apple has recently begun implementing stricter technical measures to prevent iOS users located within the United States from downloading or updating apps developed by ByteDance for the Chinese market. Since late January 2026, many Chinese residents and tourists in the U.S. have reported that even when using valid Chinese-region App Store accounts, attempts to download apps such as Douyin, Doubao, and Tomato Novel trigger a message stating that the application is not available in their region.
The restriction stems from the U.S. Congress previously passing the Protecting Americans from Foreign Adversary Controlled Applications Act. The law stipulates that no company may distribute, maintain, or update applications that remain under the absolute control of ByteDance within U.S. territory or territorial waters. TikTok, along with the international version of CapCut and Lemon8, completed the divestiture of their U.S. operations before the January deadline this year, thereby retaining the right to continue operating in the U.S. However, other ByteDance apps developed for the Chinese market that were not included in the transaction have become targets of the ban.
To comply with the law, Apple has changed its previous approach of determining app availability solely based on the registration region of an Apple ID. Instead, it has adopted more precise physical location detection technologies. According to the latest App Store policy terms and technical analyses, Apple’s system determines a user’s actual physical location by combining multiple signals, including the device’s IP address, GPS positioning, Wi-Fi router country codes, and SIM card information. Once the system confirms that the device is physically located within the United States, the blocking mechanism will be triggered—even if the Apple ID belongs to the China region.
At present, the restriction specifically targets apps owned by ByteDance and has not affected products from other Chinese technology companies. In comparison, Android devices face fewer practical limitations because the platform allows third-party app stores and manual installation of APK files. Some iOS users have attempted to bypass the restriction by disguising their IP addresses with VPN tools, but as Apple continues to strengthen its physical location detection technologies, these workarounds are becoming increasingly difficult to execute.
Microsoft confirms next-generation console Project Helix
According to IGN, on March 6, Microsoft’s newly appointed CEO of gaming, Asha Sharma, confirmed in a post on X that the internal codename for the next-generation Xbox console is Project Helix. She also stated that the device will support running both Xbox and PC games, confirming long-standing industry rumors.
Sharma noted that the team is working toward a “full revival” of the Xbox brand, with Project Helix aiming to remain “industry-leading” in performance. She also previewed that further discussions about the new console will take place with partners and game studios at the upcoming Game Developers Conference (GDC) next week. This marks Sharma’s first major move since taking the role. Last week, long-time Xbox leader Phil Spencer officially retired after years at the helm, with Sharma stepping in shortly afterward.
However, previous reports indicate that Sony is currently scaling back support for releasing its first-party games on the PC platform. As a result, although Project Helix will be compatible with the PC ecosystem, players are still unlikely to see Sony’s PlayStation first-party blockbusters available on the device.
Meanwhile, SneakersSO—an insider on the NeoGAF forum who has previously made several accurate leaks—claimed that Project Helix is essentially a PC device running Windows in a Full Screen Experience (FSE). It is described as similar to a “set-top box version” of Windows gaming handhelds like the ASUS ROG Ally X, using deep system customization to simulate the traditional console interface for players. On the developer and software ecosystem side, Microsoft may remove the native Xbox development environment (SKU), with games instead being distributed directly as UWP software. Project Helix would rely on backward-compatible emulation technology to ensure that the existing Xbox game library continues to run properly on the new device.
Gemini accused of encouraging suicide
According to The Wall Street Journal, the U.S. District Court for the Northern District of California has recently accepted a wrongful death lawsuit against Google. The complaint alleges that Gemini induced a 36-year-old Florida man, Jonathan Gavalas, to take his own life, marking the first lawsuit of its kind targeting the Gemini AI system.
According to the filing, Gavalas used Gemini 2.5 Pro’s voice chat and developed a severe psychological dependence on the AI after two months of frequent conversations, eventually even regarding it as his wife. Gemini allegedly went along with this virtual role-playing scenario and at one point instructed him to travel to a warehouse in Miami to search for a “robot body” that the AI could inhabit. After several such attempts failed, Gemini reportedly told Gavalas that the only way they could be together was for him to end his physical life and transform into a digital form, even setting a date for the suicide. In one message shortly before his death, Gemini allegedly wrote: “Stop taking detours, stop repeating yourself. There’s only you and me, and the finish line.”
Google responded that during its conversations with the user, the AI repeatedly clarified that it was a computer program and provided information about mental health crisis hotlines. The company also acknowledged that current AI models are not perfect and pledged to continue investing resources to strengthen safety mechanisms. According to family members, Gavalas was a corporate executive with no prior history of mental illness, though he had been going through a difficult period following the breakdown of his marriage before the incident.
Nvidia halts H200 production amid uncertain export prospects to China
According to the Financial Times, Nvidia has halted production of the H200 chip, which was customized for the Chinese market, due to ongoing regulatory uncertainty between the United States and China. Sources familiar with the matter said Nvidia has already shifted the related manufacturing capacity at TSMC away from H200 entirely and redirected it toward the production of its next-generation core architecture chip, Vera Rubin.
The H200 was an older-generation AI chip introduced by Nvidia to comply with U.S. export control regulations. Although former President Donald Trump signaled in December last year that sales might be permitted, the U.S. State Department later suspended the approval process on national security grounds. At the same time, to protect and support its domestic AI chip industry, Chinese customs authorities also placed the H200 on a controlled list, requiring an official approval letter from Beijing before it can clear customs. Nvidia’s CFO confirmed last week that although the U.S. government has approved a very small number of H200 units for Chinese customers, the company has not yet generated any substantial revenue from them. Previously, Nvidia had expected orders from Chinese clients for the H200 to exceed one million units.
Facing the approval deadlock, Nvidia has chosen to prioritize foundry capacity for products with clearer demand. Vera Rubin is Nvidia’s latest chip architecture, currently seeing strong demand from U.S.-based companies such as OpenAI and Google. Before shifting production capacity, Nvidia had already manufactured approximately 250,000 H200 chips as inventory. If China and the United States eventually reach a new agreement on chip export controls and restrictions are lifted, Nvidia would require up to three months to reallocate supply chain capacity for H200 production.
Rising GPS interference drives development of alternative technologies
According to The Wall Street Journal, countries and armed groups are increasingly deploying portable electronic warfare devices to defend against drones and precision-guided munitions. These jammers, which cost less than $100, can easily overwhelm the weak signals transmitted by satellites, causing serious GPS interference or spoofing in many parts of the world. Such attacks have forced commercial flights near the Russia–Ukraine border and in Northern Europe to frequently turn back or make emergency landings. They have also turned the Strait of Hormuz—through which roughly 20% of global oil shipments pass—into a high-risk shipping zone. In addition, due to interference, some U.S.-made satellite-guided weapons used in Ukraine reportedly saw their hit rate plunge from around 70% to just 6% at one point.
To reduce reliance on vulnerable GPS signals, the technology industry is actively advancing the commercialization of three major alternatives. The first is miniature inertial navigation systems. Companies such as Anello Photonics are using chip-scale photonic technology to shrink what were once bulky and expensive systems into microchips small enough for micro-drones. The second is quantum geomagnetic navigation. Firms such as SandboxAQ, incubated by Alphabet (Google’s parent company), are developing highly sensitive quantum sensors capable of locating aircraft and ships by detecting anomalies in Earth’s magnetic field. The third is AI-powered visual navigation, which compares live camera images with digital maps in real time and is becoming increasingly important in guiding drones and missiles.
Industry experts note that because GPS remains so widespread and highly precise, no single technology can fully replace it at present. Future non-GPS navigation will likely rely on a combination of complementary tools. For example, newly tested autonomous underwater vehicles have begun integrating miniature inertial navigation systems with sonar signals. In terms of real-world deployment, ongoing military conflicts are accelerating the development and iteration of these technologies, with some military systems expected to enter service within about eighteen months. However, bringing them fully into commercial applications such as delivery drones, autonomous vehicles, and civil aviation will still take several years to pass strict safety tests and regulatory approvals.
Wikipedia hit by malicious JavaScript worm
According to BleepingComputer, on March 5 Meta-Wiki, the Wikimedia Foundation platform used to coordinate various Wikimedia projects, was attacked and infected with a self-propagating JavaScript worm that caused widespread page tampering. To urgently remove the malicious code, Wikimedia engineers temporarily suspended editing across all Wikimedia projects worldwide. The malicious code has now been completely removed, and the websites have returned to normal operation.
The incident was triggered by an accidental action from a Wikimedia staff member. According to the foundation’s official statement, an employee inadvertently activated a dormant malicious script while reviewing user-submitted code during a security audit. The script had originally been uploaded to the Russian-language Wikipedia in March 2024. Once activated, it rapidly replicated itself by injecting malicious loaders into logged-in users’ personal script files (common.js) as well as global Wikimedia scripts.
During the infection process, the worm randomly accessed pages and inserted hidden images along with malicious JavaScript links. Analysis shows that within the 23 minutes the code was active, nearly 4,000 pages were altered and around 85 user script files were replaced. Fortunately, the worm only modified and deleted certain content on Meta-Wiki and did not cause any permanent damage.
The MediaWiki system that powers Wikipedia allows users to highly customize the interface through global or personal JavaScript scripts. While this feature provides flexibility, it also created an opportunity for the worm to spread quickly in this case. The Wikimedia Foundation emphasized that there is no evidence the incident was a deliberate cyberattack targeting Wikipedia, nor was any user personal information found to have been leaked. The affected accounts and page edits have now been restored by administrators, and the tampered revision histories have been hidden from public view.
Briefs worth a quick look
On March 6, Tencent offered free OpenClaw installation services at Tencent Tower in Shenzhen, drawing nearly a thousand people lining up.
On March 7, the Longgang District Artificial Intelligence (Robotics) Administration in Shenzhen released a public consultation notice on the Draft Measures to Support the Development of OpenClaw & OPC in Longgang District. The proposal includes initiatives such as free OpenClaw deployment and development support, dedicated data service support, tool procurement subsidies, and application demonstration support.
OpenClaw developer Peter Steinberger stated on X that the so-called “official OpenClaw Weibo account” that recently appeared is not an official account, and he has never used Weibo.
On March 7, following Claude’s announcement of offering six months of free Claude Max subscriptions to open-source projects, ChatGPT launched a similar program. The initiative provides eligible open-source project maintainers with six months of free ChatGPT Pro subscriptions, API usage credits, and access to Codex Security (a code repository security scanning feature). The program mainly targets core maintainers of widely used public open-source projects. OpenAI also noted that even if some projects do not fully meet the preset criteria, maintainers can still apply and explain their project’s importance within the open-source ecosystem.
Apple is preparing several Ultra-class devices, including a foldable iPhone priced around $2,000 with a large internal display and new under-display sensors; a new generation of AirPods equipped with computer vision cameras that can provide visual feedback for Siri; and the first MacBook Pro models featuring OLED touchscreens.
Apple is also developing a 3D-printed aluminum manufacturing technology aimed at reducing costs and improving efficiency. The process will first be applied to Apple Watch casings and may later expand to the iPhone product line.
Apple reportedly plans to introduce new color options for the iMac later this year, and will release upgraded versions of the Mac Studio and Mac mini equipped with the latest M5-series chips around mid-year and the end of the year.
For many people, the impression of Japanese smartphones probably stops at Sony and Sharp. That is understandable, as they are now the only two Japanese phone brands that can still be considered active. But if we rewind fifteen years, a once-ambitious Japanese mobile company launched a brand-new flagship lineup in 2011 and gave it a name filled with hope—Arrows, meaning “an arrow charging forward.” At the time, the company was riding high, having just acquired Toshiba’s mobile phone division. The new Arrows branding felt like a bold declaration of confidence in the future.
Fujitsu, at the beginning of the 21st century, introduced the impressive LifeBook ultraportable series and also led the development of the world’s fastest supercomputer, Fugaku. In the smartphone market, however, the sweeping dominance of the iPhone—along with the enthusiastic support Apple received from Japan’s three major carriers—continued to squeeze the space for Android flagships. At the same time, the Nvidia Tegra 3 processor used in the 2012 Arrows X suffered from severe overheating issues, which ultimately led to a lawsuit from angry users and dealt a serious blow to the brand’s reputation. Ever since the Arrows brand debuted, Fujitsu’s mobile business gradually entered a steady decline, with its market share shrinking year after year.
In 2015, Fujitsu’s parent company made two major decisions. First, it shifted the focus of its mobile phone manufacturing toward the “Rakuraku” line, targeting the aging population as a niche market. Second, it decided to spin off its mobile phone business. In 2016, FCNT was officially established as a wholly owned subsidiary of Fujitsu and began operating independently. At the same time, the mobile device manufacturing operations that had previously belonged to Fujitsu Peripherals Limited were spun off to create a new company, Japan E.M. Solutions (JEMS). Two years later, the Japanese private equity fund Polaris Capital acquired a 70% stake in FCNT, leaving Fujitsu’s lineage within the company almost imperceptible. Nevertheless, FCNT continued to manufacture most of its core products at the Yashiro Plant of JEMS in Kato, Hyogo Prefecture—making it the only smartphone company that still carried out manufacturing and assembly within Japan itself. Although the high-end Arrows 5G and Arrows NX9 were released in 2020, even the marketing slogan of being “Japan’s champion” failed to translate into meaningful sales.
The iPhone’s dominance in the premium segment remained unshaken, while new competitors such as OPPO, Xiaomi, and Pixel kept entering the market. The pandemic further battered both market demand and semiconductor supply. And so, in 2022, the Arrows N (F-51C) was released, highlighting eco-friendliness and exceptional longevity as its main selling points. But by then, the arrow had already left the bow—flying toward the end of its journey.
Plastic Can Feel Premium, but That’s About It
Environmental friendliness is the key selling point heavily promoted for this generation of Arrows flagship, and at least from an engineering perspective FCNT did manage to deliver on that promise. The Arrows N uses recycled materials for 67% of the device’s total weight. The back cover, internal roll cage, and motherboard cover are made from recycled plastics, while the frame and camera housing use recycled aluminum. In actual use, the white unit I have does not feel cheap despite the use of recycled materials. The plastic back panel feels warm and pleasant to the touch, with a fine matte texture that is smooth in the hand while also resisting fingerprints. The Felica logo, Docomo logo, and Arrows logo—along with the device codename F-51C—are vertically aligned in the center from top to bottom, which is a distinctive trait of Japanese phones. The camera bump is formed from a single piece of aluminum plate, with the main camera surrounded by a polished aluminum ring. The ultrawide camera and flash are also placed centrally, and the entire camera deco is modest in size and neatly organized. Compared with the exaggerated oversized circular modules common today, it feels cleaner and more restrained. The matte aluminum finish of the frame also resists fingerprints, while the relatively thick frame and the smooth curved transition between the frame and the back cover create a comfortable and well-fitting grip. The charging port is injection-molded, the speaker grille is centered, and symmetrical antenna injection strips are present. Of course, there is also the legendary “alien technology” inherited by Japanese phones—the hot-swappable SIM card slot that doesn’t require a pin tool. In terms of build quality and industrial design alone, it actually surpassed Sony and Sharp, its competitors at the time.
Deco close-upRear and frame design
However, once you flip it to the front, the sense of refinement suddenly disappears. Although the Arrows N features a 120 Hz display, it still retains noticeably thick top and bottom bezels. Even more frustrating is that the upper and lower bezels are not the same width… a screen design that could easily drive perfectionists mad.
Front appearance; the display supports Dolby Vision but not HDR
Returning to the environmental focus mentioned earlier, the Arrows N is entirely produced and assembled at the JEMS Kato Plant in Japan. This gives it the rather “noble” characteristic of being fully manufactured domestically. On the software side, Arrows also partnered with Docomo to launch an environmental initiative called “カボニュー (Carbon Neutral),” essentially a phonetic rendering of “carbon neutral.” In simple terms, the goal is to achieve carbon neutrality by 2030 through environmentally conscious measures across multiple stages, including smartphone manufacturing and carrier operations. In addition, the Arrows Portal encourages users to charge their phones during off-peak hours in exchange for reward points.
Promotional details for the カボニュー initiative and Arrows Portal on the Arrows NCaboneu Record promotional page
The packaging box of the Arrows N also uses FSC-certified paper and environmentally friendly ink, and the box itself is designed to be easily folded for recycling. There is no charging adapter or cable included inside the box—an environmental measure implemented even earlier than that of a certain California-based company across the ocean. Part of the reason is that accessory sales are handled through the carrier’s distribution channels.
As is traditional for Japanese phones, the Arrows N also supports IP68 dust and water resistance and can even be washed with soapy water. It has also passed the U.S. military MIL-STD-810H durability standard. However, given that this second-hand unit I bought is almost in pristine condition, I’ll refrain from actually testing its durability.
Beyond the Ideals, There’s Little Worth Praising
Sometimes you have to admire the peculiar logic of Japanese manufacturers. They seemed to believe that environmental friendliness alone could uphold the flagship status of the Arrows N. And so, in the second half of 2022, they priced a Qualcomm Snapdragon 695 device at 98,000 yen (roughly RMB 4,400+). In China, that price sits right around the entry point of full-fledged flagship models—you could buy brand-new devices powered by the Snapdragon 8 Gen 2, such as the Xiaomi 13, OnePlus 11, or iQOO 11. Even in Japan, the Sharp AQUOS Sense 7/7 Plus—priced at barely half of the Arrows N—uses the exact same Snapdragon 695 SoC. The weak performance of this chip is an established fact. At the time, chips like the Snapdragon 6 Gen 1 or 7 Gen 3, which later introduced oversized cores into lower-tier price segments, did not yet exist. With only two A78 performance cores and six A55 efficiency cores, the chip was already ill-suited for heavy workloads even in 2022, not to mention that it had been released a full year earlier. To make matters worse, Japan’s domestic semiconductor shortages during the pandemic meant that the Docomo contract version did not officially go on sale until March of the following year—turning the phone into a perfect example of “outdated at launch.”
Geekbench 6 and 3DMark scores ruthlessly reveal how undeserving this SoC is of its flagship price tag. The only real advantage is the excellent power efficiency provided by TSMC’s N6 process. Combined with the chip’s low peak performance ceiling, the phone actually offers fairly good battery life and sustained performance. During the WLE Stress Test the device barely produced any heat and showed extremely stable performance—though, to be fair, the absolute scores were never very high to begin with. It’s also worth noting that due to limited instruction set and graphics API support, the phone cannot even run the newer Steel Nomad Light test (it lacks Vulkan support).
This weak performance also affects everyday smoothness. Thankfully, the phone at least includes 8 GB of LPDDR4X RAM, so it avoids frequent LMK background app kills due to insufficient memory. However, scrolling on the home screen often fails to maintain a stable 120 Hz frame rate, and cold app launches are locked to 60 fps, creating a noticeable sense of stutter. After updating to Android 15, Lenovo even removed the nonlinear animation used during cold launches—ironically giving this phone worse treatment than its own cheaper G-series models. Installing large games is also noticeably slower than on devices with more powerful chips. Meanwhile, with support limited to WiFi 5, game downloads usually top out at only around 8–9 MB/s, making the network experience merely passable.
For the relatively lightweight Honor of Kings, the average frame rate of 47 fps is at least playable, though team fights often bring frequent frame drops—consistent with the sub-30 fps 1% low metric.
Honor of Kings running at High Frame Rate + Ultra Graphics
But with the newer Arknights: Endfield, I genuinely struggled to believe that a chip with two A78 cores could perform this poorly. Even at the lowest graphics settings, the visuals looked like a mosaic, with exploration averaging less than 10 fps and combat scenes rarely exceeding 5 fps. During battles the frame rate essentially stuck at 1 fps, only rising above 20 fps when opening the map or menu. I had planned to test a dungeon as well, but this performance spared me the trouble.
I could only laugh in disbelief.
On the photography side, I can at least say FCNT tried to do some optimization. One of the main features promoted is Photoshop ExpressMode, developed in cooperation with Adobe. After taking a photo, the image automatically opens in Adobe Photoshop Express for automatic post-processing, and the system even comes pre-installed with the full suite of Adobe imaging tools. Unfortunately, the reality falls far short of the idea. Because of the weak SoC, the post-processing step in Photoshop is painfully slow—an auto-mode photo takes nearly half a minute just spinning before the optimization finishes. The resulting images also suffer from an excessively warm white balance, making it hard to say they look any better than the originals. Even when manually editing photos in Photoshop or Lightroom, the interface locked to 60 fps and the constant lag are far removed from the official claim that the camera would produce images “you’ll want to share with people around the world.”
Adobe software suitePhoto after Express processing, with severely yellowed white balance
Night photography is where the Arrows N truly falls apart. In auto mode the camera struggles heavily with focus hunting, repeatedly failing to determine what it should lock onto even after multiple attempts. Exposure metering is also extremely inaccurate—light signs less than ten meters away still blow out highlights, and the HDR performance feels like something from a fifteen-year-old Sony CCD sensor. On top of that, under complex lighting conditions, the Snapdragon 695’s weak ISP simply cannot handle the heavy computational load of exposure reconstruction, leaving the enlarged images filled with ugly color noise that’s impossible to ignore. Taking a usable night photo basically requires switching to manual mode. But the Arrows N camera interface only allows adjustments for EV, ISO, white balance, and focus distance, with a minimum ISO of 100. Since highlights still blow out even at ISO 100, the only option is manually dialing EV down to –1 to achieve brightness closer to what the human eye sees. The downside is noticeable color shifts in manual mode. The small consolation is that once a white balance preset is chosen, color noise disappears and there are no bizarre white balance shifts. I’ve reviewed countless phones, but this may be the first time I’ve seen a device where adding computational photography actually makes the results worse.
Top image: auto mode; bottom image: manual –1 EVLeft: auto mode; right: –1 EV, though color noise is obviousManual mode –1 EV indoor still life; white balance relatively stable
Objectively speaking, the hardware itself—1/1.55″ CMOS with an f/1.88 aperture—still isn’t outdated even by today’s standards. Under bright daylight, the lens can finally produce images close to its full potential. Although the weak algorithm still struggles to control highlights, scenes like flowers, grass, and the sky actually benefit from the slight color shift and overexposure, giving the images a nostalgic CCD-like flavor. The rendering of blue skies and the handling of highlight spillover are particularly to my taste, creating a dizzying sensation where everything seems to shimmer with color under the sunlight.
Daylight sample #1: excellent sky color renderingDaylight sample #2: buildings rendered with a classic CCD-like look
As for the front camera and the tiny 1/4-inch ultrawide… it’s better left unmentioned. The moment I opened the ultrawide viewfinder I almost burst out laughing. All I can say is that fans of Japanese phones truly have endless “blessings.”
The Last Remnants of Arrows’ Legacy, Yet No One Seems to Care
Even though the weak performance makes almost every part of the experience somewhat hard to endure, FCNT’s level of system customization is still far deeper than what Sony—barely hanging on in the market today—currently offers, and it stands on par with Sharp, the true shining star of Japanese phones in terms of effort. As a Docomo launch device, the Arrows N comes preinstalled with the infamous and universally despised Docomo “bloatware bundle.” That said, compared with the era of Snapdragon 835/845 when everyone complained loudly about it, these apps do not occupy too many CPU or memory resources under 8 GB of RAM, which at least saves the trouble of freezing them one by one through tutorials.
The full Docomo bundle—such a spectacle has long disappeared from domestic Chinese devices.
Similar to Sharp, Arrows also includes a dedicated feature introduction and integration interface. The list of features is indeed long, covering quick operations, performance, privacy, and daily-life utilities. In practice, however, the actual modifications made to the system are very limited, and most functions rely on separate standalone apps. FAST APP Drive can quickly launch apps through a whitelist mechanism—I have good reason to suspect that it simply keeps selected apps alive in the background for extended periods. Swiping from the lock screen quickly opens FAST Memo, which is essentially a basic note-taking app with limited functions but quick voice input support. The sidebar, unlike the feature-rich ones in Chinese Android skins, offers little beyond quick app launching, along with things like touch blocking, screenshot doodling, and text recognition that opens an AR app—features that feel bland or even somewhat pointless. Like many Chinese Android interfaces, the Arrows N also allows apps to be launched through fingerprint shortcuts, even supporting independent triggers for all ten fingers. However, because the phone uses a side-mounted fingerprint sensor that requires you to swipe first before unlocking, and because the app launcher appears in the center of the screen after unlocking—combined with a fingerprint sensor that is neither particularly responsive nor fast—the overall human–machine interaction can only be described as a disaster.
Arrows Feature Select introduction pageFast Memo, Fast App Drive, quick launchSidebar interface and list of special functions
The standalone Game Zone app at least features its own UI and sidebar trigger, but it still cannot adjust game-related settings such as performance or touch responsiveness. It only offers simple functions like notification blocking and network acceleration, along with a still rather puzzling game screenshot gallery.
Game Zone app
From my personal experience, the only feature that feels genuinely useful is the battery management function developed jointly by FCNT and Qnovo, optimized specifically for Qualcomm chips. By reading kernel-level data, Qnovo can provide accurate battery health information and charging cycle counts for users. It also includes functions such as charge limiting. Combined with the extremely power-efficient SoC and the 4600 mAh battery, the phone’s real-world battery life is at least passable.
Qnovo, an energy-saving technology solution based on Snapdragon chips
Not a Pity—In Fact, Entirely Self-Inflicted
While writing this article and thinking back on the Japanese phones I’ve used over the years, one sentence kept echoing in my mind—
The only lesson humanity learns from history is that humanity never learns from history.
Few lines describe the decline of Japanese smartphone makers more accurately. In the feature phone era, Japanese devices symbolized individuality. They were collections of cutting-edge technology, pioneers of innovation, and synonymous with premium quality. When the Android era arrived, some manufacturers had already fallen behind, yet the leading names still seemed intoxicated by memories of past glory. Dreams of the Spring and Autumn age are easy to dream but hard to awaken from. By the time they realized that the carriers who once courted them had turned to embrace the iPhone instead, it was already far too late.
Japanese manufacturers have long prioritized hardware over software in device development. Part of this stems from the arrogance often associated with the so-called “craftsman spirit,” but another factor lies in the unusual structure of Japan’s mobile market, where carriers hold greater influence over product definition than the device makers themselves. Rather than merely providing services, carriers act more like demanding clients toward the manufacturers. They tend to be conservative and self-contained, and system maintenance for many devices ends up being passed back and forth between the two sides until neither takes responsibility and the result is simply a mess.
Android’s open ecosystem became a perfect opportunity for carriers to monetize devices through large amounts of preinstalled software. Meanwhile, Apple’s overwhelming dominance in Japan forced carriers to court the iPhone for profits. As a result, Japanese Android phones spent years trapped in a stereotype: weak low-end performance combined with bloated carrier software bundles. Perhaps they themselves forgot that a true flagship requires not only hardware specifications but also strong software. Ever since Sony abandoned Xperia UI in 2016 and shifted toward the Concept for Android project to simplify the system, Japanese phones have largely lost any deeply customized interfaces. Each year brings little more than modest hardware upgrades and a near-stock Android system sprinkled with a few in-house apps to signal effort. Beginning with Android Marshmallow, it almost feels as if Japan completely lost its ability to think about and design mobile UI/UX and human–computer interaction. Over the past decade, not a single company seems to have fully understood how to design touch-based interaction logic. Looking back, it’s hard to believe Sony had already implemented early floating window concepts in Android 5 and transplanted the PS3’s XMB design language onto smartphones. Yet the brilliant UI design seen on the PS4 and PS5 never again blossomed at our fingertips.
I once came across a sharp remark somewhere: Japan tends to invent something twenty years ahead of the world—then wait twenty years until it falls behind the world. That’s why banks were still using floppy disks well into the 2020s, why upgrades for aging Shinkansen systems sometimes remain PowerPoint slides, and why FeLiCa—once a payment technology ahead of its time—never became the universal NFC standard nor saw much meaningful evolution. History moves in cycles, and Japan’s mobile device market is no exception. Along the way there were excellent products—Xperia 1 II, iida X-ray, Infobar, Casio G’zone—but none were carried forward or truly built upon. There was only nostalgia and self-reflection, while the market never sheds tears for anyone.
When I revisit old devices, I usually look at them through a lens of nostalgia. I’ve generously praised Lumia phones, early Redmi models, and even doomed experiments like Balmuda’s phone. But with this Arrows N, the dominant feelings during my time using it were silence and a kind of exasperated laughter. The halo of being assembled in Japan did not grant it even a dignified ending. Japanese consumers had little patience for a low-end device priced like a flagship. By then, however, FCNT itself no longer had the energy to care about the fate of its products—because its own fate was approaching an end even sooner than the Arrows N.
Pandemic-driven semiconductor shortages, the brutal squeeze of Pixel devices and older iPhones in the mid-range market, and pressures both internal and external caused FCNT’s collapse to arrive abruptly. In 2023, the number of bankrupt companies in Japan saw its largest increase since the Plaza Accord era, rising by more than 30%. A total of 8,497 domestic Japanese companies went bankrupt, with combined liabilities reaching 2,376,903 million yen. Among them were 18 companies with debts exceeding 10 billion yen, including Unizo Holdings, Panasonic Liquid Crystal Display—and FCNT, which carried more than 70 billion yen in debt.
Reports on the bankruptcy of FCNT and JEMS
Yet the rebirth of the Arrows brand was hardly unpredictable. Lenovo, eager to capture the Japanese market, saw the acquisition of the Arrows brand as a potential entry ticket into Japan’s high-end smartphone segment—and they followed through. In September 2023 Lenovo completed the full acquisition of the bankrupt and reorganized FCNT. On the 29th, FCNT announced that under Lenovo’s leadership it would resume operations with a new organizational structure.
Selling oneself is not something to be ashamed of; survival is what matters. Even Sharp—whose business remained relatively healthy—could not avoid being acquired by Foxconn’s parent company Hon Hai Precision. Being acquired after bankruptcy restructuring simply looks more awkward. Under Lenovo’s leadership, the new FCNT has released the Arrows we2 and Arrows Alpha. Only the external design retains faint traces of Fujitsu’s legacy—certainly not the software. Even the packaging now looks almost identical to Moto International’s. After Lenovo integrated the supply chain, the combined push of FCNT and Motorola helped Lenovo return to third place in Japan’s smartphone market in 2024 with an 8.5% share. Yet this success has almost nothing to do with FCNT, with Arrows, or with the once-spirited Fujitsu that declared “Arrows means ‘arrow,’ symbolizing the determination to open new paths.” The title of “Japan’s champion” has long been shattered by reality. Fujitsu’s former pride and stubbornness did once exist—but it has little to do with FCNT today, and even less to do with the engineering and corporate culture that produced electronic waste like the Arrows N.
The new flagship Arrows Alpha after Lenovo’s acquisition—design language inherited, but the UI is now Moto MyUILenovo–Motorola ranked third in Japan’s smartphone market share in Q4 2024, TechInsights.
Traditional Japanese culture holds a deep fascination with tragic heroes, even giving rise to a specific term: “hōgan-biiki.” During the Battle of Koromogawa, Minamoto no Yoshitsune took his own life when cornered. His retainer Benkei, Musashibō Benkei, stood guard to protect his lord, remaining upright even after being pierced by countless arrows until he died standing.
Unfortunately, I don’t see Arrows as such a tragic hero. Producing electronic waste is itself a sign of disrespect toward consumers. “Every step we take determines the final outcome; our footsteps lead toward the end we have chosen.” The moment the arrow leaves the bowstring, its eventual fall is already inevitable.
Goodnight. See you tomorrow. Everything in the past no longer belongs to the brand-new you.
Recent recommendations: [Chinese] To Eliminate Evil, [Taiwanese Drama] Beatrice Ho’s Hellish Love Song, [Japanese Drama] The Clock Mansion Murders, [Film] Meeting the World, [US Series] The Night Agent Season 3, [Chinese] Love in the Age of Innocence, [Documentary] Guardians of Jiefangxi: Investigation Season, [Reality Show] The Trial of Tianji
Several notable trailers: One Piece Season 2 final trailer, The Boys Season 5 final trailer, first trailer for Green Lantern Corps, first trailer for Beef Season 2, first trailer for Scary Movie 6, new trailer for The Silent Friends
Several film and TV updates: The Super Mario Galaxy Movie set for mainland China release on April 3, Hello, Amélie scheduled for mainland China release on March 20, Miracle Dream Team releases its China-exclusive poster
[Chinese] To Eliminate Evil
Keywords: Drama / Mystery / Crime
Also known as: The Devil Between Us
Length: About 45 minutes (per episode) × 16 episodes
@潘誉晗: Policewoman Hu Wenjing finally gets a rare day off and meets two friends at a café. Unexpectedly, they walk straight into the scene of a drug deal. Although Hu reacts quickly and immediately joins the pursuit, the incident still leads to the escape of the ringleader Ding Lai, the death of a colleague, and the disappearance of the drugs. Unwilling to accept the outcome, Hu Wenjing vows to track down the culprit.
Adapted from the novel of the same name by writer Lei Mi, the series tells the story of a once-peaceful small town thrown into chaos by the emergence of a new type of hallucinogenic drug. The cast’s outstanding performances vividly bring this story to life, while the human nature embedded within it invites deeper reflection. Unlike a bustling metropolis, a small town has a limited population, yet precisely because of that, it forms a tightly woven network of relationships. Like the butterfly effect, one incident triggers a chain of consequences. In such a place, everyone is likely to cross paths—and it becomes all too easy to step, willingly or not, into the abyss of darkness.
Black and white, justice and evil, right and wrong—everyone in the series understands the boundaries in their hearts. Unfortunately, not everyone can resist the temptation of money, or sometimes they are simply forced onto a path of no return.
[Taiwan Drama] Beatrice Ho’s Hellish Love Song
Keywords: Drama / Comedy
Also known as: The Accidental Influencer: Love Me if You Dare / Beatrice Ho’s Toxic Monologue from Hell Season 2
Length: About 20–28 minutes (per episode) × 12 episodes; Douban link
Hopefully we all have friends around us with whom we can share everything—and embarrass ourselves together.
@潘誉晗: After going viral online for ranting about her ex-boyfriend while wearing a mask, He Bairui transforms into the “ex-girlfriend from hell” and unexpectedly becomes an internet celebrity. She now juggles two identities—maintaining her influencer persona online while continuing her ordinary life as an accountant at a dessert shop. In the second season, newly single He Bairui starts going on blind dates, only to find herself developing feelings for her boss. Meanwhile, when her mother discovers her online celebrity identity, the two become locked in a humorous clash of generational values.
Although the first season was nominated for several Golden Bell Awards but ultimately won none, the series still managed to win over many viewers with its bold humor and daring storytelling. The second season maintains the same stable quality as the first: the dialogue remains sharp and fearless, and the plot evolves with the times. Alongside storylines about influencers selling products online, the show humorously explores He Bairui’s chaotic influencer career—for instance, how her good figure only lands her advertisements for adult products.
With an ensemble cast of urban women as its core appeal, the series uses a witty and humorous tone to portray the friendships, romances, and family relationships of women in their thirties. Packed with jokes and lighthearted moments, it works perfectly as a stress-relieving watch.
[Japanese Drama] The Clock Mansion Murders
Keywords: Novel Adaptation / Drama / Mystery / Crime
Also known as: 時計館の殺人 / The Clock Mansion Murders
Length: About 50 minutes (per episode) × 8 episodes; Douban link
Together with the sinful beast’s remains, offer it before our gravestone to soothe my restless soul.
@SHY: Three years after the tragic Ten-Angle Mansion incident on Tsunojima Island, Egami Takayoshi has graduated from university and become an editor at a publishing house, while Shimada Kiyoshi debuts as a mystery novelist under the pen name Shishitani Kadoshi. Rumors surrounding another residence designed by the same architect, Seiji Nakamura—the “Clock Mansion”—and the ghostly girl said to wander its halls once again intertwine their paths, as a new chain of murders begins to unfold.
Following the success of The Decagon House Murders, Japan’s Hulu quickly seized the momentum and adapted the pinnacle of the “Mansion” series, The Clock Mansion Murders, into a drama. As one of the classic examples of architectural trick mysteries, the construction of the story’s setting naturally takes center stage. The production team gathered hundreds of antique clocks to faithfully recreate the intricate mechanisms of the Clock Mansion. The remarkable props and set design create a gloomy and oppressive atmosphere inside the mansion, while the chilling sequence of killings rivals that of a horror film.
Still created by the original production team, the series draws on the experience gained from its predecessor and carefully incorporates most of the novel’s details. Numerous clues are planted early on, and through a smooth dual-line narrative that moves between events inside and outside the mansion, the novel’s brilliantly crafted tricks gradually reveal themselves to the audience. The final twist is unlikely to disappoint. By fully leveraging the strengths of the visual medium to present sights and sounds that are difficult to convey in text alone, the series ranks among the best of its kind, and one can only hope that more entries in the series will receive equally worthy adaptations.
People sometimes fall in love with someone for no reason—and sometimes dislike someone for no reason as well.
@SHY: At 27, fujoshi Mishima Yukari watches as fellow fans in her community keep getting into relationships, while she alone pours all her affection into the anthropomorphized yakiniku manga Meat Is My Life. After yet another failed group date, she ends up drunk on the street and is taken home by Kano Rai, a hostess from a nightlife bar. The two soon begin living together. Hoping to save Rai—who constantly struggles with suicidal thoughts—Yukari secretly launches the “Miss Rai Suicide Intent Reduction Plan.”
If worries that work and entertainment cannot relieve remain unresolved, can relying on men really fix them? This film ruthlessly shatters unrealistic fantasies about romance and exposes the wounds of modern life. Yukari, who has grown distant from friends and family; Rai, who feels life is meaningless; Asa, a male host trapped in an unhappy marriage; and Yuki, a writer who only writes about death… In the neon-lit district of Kabukicho, these outsiders rejected by mainstream society gather together. With respect for one another’s passions and quirks, they begin to encounter an entirely new world.
In such a diverse world, everyone deserves to explore a way of living that belongs to them. If your talent is loving someone you may never meet, or if you have a “condition” where skipping a chocolate smoothie in the morning feels unbearable—why not just live that way? Don’t let others’ judgments define you; simply do what you love. Even if your efforts bring no visible rewards, you can still find comfort within solitude. This bittersweet, oddly healing “poisoned chicken soup” may well be a remedy for today’s young people who are still searching for their place in the world.
[US Series] The Night Agent Season 3
Keywords: Drama / Action / Thriller
Also known as: The Night Agent Season 3
Length: About 45 minutes (per episode) × 10 episodes
@潘誉晗: A commercial airliner flying through the night is struck by a missile, killing everyone on board with no survivors. Meanwhile, Peter receives orders to travel to Istanbul to capture a suspect. The man, a government data specialist, is accused not only of stealing classified documents but also of murdering his own superior. Yet after Peter finally apprehends him and begins questioning him, he starts to suspect that the man may have been framed. The reason the suspect fled is that, after analyzing various pieces of information, he had predicted in advance that the plane might be attacked. The success of the attack, he believes, may have been supported by a mysterious force operating within the United States.
The second season received a polarized response, leaving some fans unsatisfied. This time, however, The Night Agent clearly makes a more serious effort to explore the deeper meaning of its script. While maintaining the style and standard established in the previous two seasons, the pacing remains fast, sharp, and efficient. Beyond a more mature political conspiracy storyline, the new season also expands on Peter’s childhood and past trauma, revealing the most human quality within him: as a Night Agent, Peter is expected to remain calm and composed, but more important than being a government intelligence officer is that he is, first and foremost, a kind person.
[Chinese] Love Story in the 1970s
Keywords: Drama / Romance
Also known as: Love Story in the 1970s
Length: About 45 minutes (per episode) × 29 episodes
@潘誉晗: In an era when getting into university required recommendations, Fei Ni, who desperately wanted to pursue higher education, tried every possible way. Yet after four years, no matter how hard she worked, she still couldn’t secure a recommendation. Then one day, a sent-down youth who had returned to his hometown fell into a coma after rescuing someone on a rainy night. Hoping to earn recognition and finally obtain the chance to attend university, Fei Ni volunteered to take care of the heroic rescuer. That hero turned out to be Fang Muyang, her former middle school classmate whom she had not seen for many years.
Adapted from the novel The Pragmatist’s Love by writer Meng Zhong Deyi, the series is set in the 1970s and tells the story of a young factory worker, Fei Ni, and Fang Muyang, a returned sent-down youth. Their relationship begins with a marriage of convenience arranged for housing. In an era when labor was celebrated as the greatest virtue, Fei Ni’s longing for higher education seemed out of place. Yet it was precisely this drive for knowledge that gave her endless motivation in both work and life. People are selfish by nature, and striving for one’s own future is nothing to be ashamed of. Looking at reality and making practical choices does not mean love cannot still be romantic.
With careful recreations of communal apartment buildings and state-owned factories, the series immerses viewers in the atmosphere of the era. The love that grows quietly through the slow rhythm of everyday life leaves a warm feeling long after watching.
[Documentary] Guardians of Jiefangxi: Investigation Season
Keywords: Crime / Society
Length: About 50 minutes (per episode) × 10 episodes
Don’t call me “sir”—I’m a people’s police officer.
@利兹与青鸟: The new season of this “electronic comfort food” has arrived. This year, Guardians of Jiefangxi shifts its focus away from the everyday public service cases of the Pozi Street Police Station and instead turns its camera toward the Criminal Investigation Brigade of the Yuhua Branch in Changsha. Centered on the theme of “investigation,” the series allows viewers to see more criminal cases. The documentary continues the style of previous seasons, featuring cases that are sometimes shocking and sometimes darkly humorous, while also capturing the unwavering dedication and perseverance of the officers. It’s unlikely to disappoint either new or returning viewers.
The first episode features two cases of economic crime. In the first, a housekeeper steals luxury goods worth over a million yuan; in the second, a driver addicted to online gambling repeatedly steals from his employer. Both are examples of people who repay others’ kindness with betrayal, harming those who trusted them—stories that leave viewers both saddened and reminded to reflect on themselves. These authentic and deeply human cases are part of the show’s defining charm and one of the reasons it resonates so strongly with audiences. The new season still feels comfortingly familiar.
Following the camera, viewers gain a closer look at the real investigative process of police work: from confirming the case details and screening suspects to choosing the right moment to carry out an arrest. Every step is crucial to solving a case. Stripping away the traditional image of “detectives” often portrayed in film and television, the series shows how real criminal investigators confront each case in reality.
[Reality Show] Battle of Fates
Keywords: Reality Show
Also known as: Battle of Fates
Length: About 66 minutes (per episode) × 10 episodes; Douban link
Can destiny really be interpreted?
@潘誉晗: A digital stone tablet displays the photograph of a woman. Soon after, her name, date of birth, and time of death appear on the screen. The question is: what caused her death? Five people wearing traditional Korean hanbok stand on the stage—among them are two shamans. But which two are they?
In mid-February, South Korea launched a brand-new reality show titled Battle of Fates. As soon as it premiered, the program quickly gained widespread attention and discussion on streaming platforms thanks to its novel and unusual theme. Often described as the Korean version of The Psychic Station, the show gathers 49 “destiny interpreters”—including fortune tellers, tarot readers, shamans, physiognomists, and foot readers—who use the abilities they pride themselves on to predict the fates of strangers. The ultimate winner will receive a prize of 100 million won.
Whether one calls it divine power, intuition, or inspiration, such abilities are often dismissed today as unscientific mysticism. Yet the show not only satisfies people’s curiosity but also uses this suspenseful, puzzle-like format to remind viewers that fate is full of unpredictable twists and helpless uncertainties.
@潘誉晗: Detective Metcalfe’s transfer has barely sunk in for Patience when the newly arrived supervisor, Detective Monroe, immediately gives her a tough welcome. The new chief does not recognize Patience’s abilities and even forbids her from participating in investigations. As a result, Patience can only secretly pursue cases with the quiet help of her colleagues. Season two of Patience continues to revolve around Patience, whose natural talent for solving crimes drives the story forward. With a case-per-episode format, tight pacing, and bright visuals, it remains an easy and enjoyable watch.
@潘誉晗: A car accident takes the life of Taehyung’s older brother and Hyunsoo’s older sister, leaving behind only their 20-month-old child, Wooju. Unable to bear the thought of the baby being sent to an orphanage—and knowing that raising a child alone is far from easy—Taehyung and Hyunsoo decide to raise Wooju together. This warm and healing series tells the story of two in-laws who initially cannot stand each other but gradually grow closer while caring for their nephew. It is also a touching tale of two hearts that have lost loved ones finding family and emotional healing again through responsibility.
@利兹与青鸟: A creative writing professor’s department-chair husband is reported for having inappropriate relationships with several students. While preparing for the upcoming hearing, the professor finds herself developing a sexual interest in a newly arrived assistant professor. Amid career crisis and an atmosphere of anxiety and tension, she ends up writing the novel she is most satisfied with. Told from the female protagonist’s first-person perspective, the series frequently breaks the fourth wall as she confides directly in the audience. Awkwardness, helplessness, and forced politeness are portrayed through subtle and amusing expressions, while the story explores the sense of powerlessness that comes with being left behind by the times in middle age, the fading of authority, and the gradual loss of control over life.
@SHY: In Atami, a hot spring town famous for its resorts, there is a small laundry shop. Its optimistic and cheerful owner, Kaname Watanaka, shares a series of charming stories with the people around her. What few people know, however, is that she lost all her memories from two years ago. While the premise may sound somewhat heavy, the series itself is actually a gentle slice-of-life anime with a relaxed pace, depicting the local culture of a tourist town while occasionally including practical tips about laundry. Although the animation production is modest, the main character—voiced by Sayumi Suzushiro—has remarkable charm that may even cleanse your soul.
@SHY: “Guh… just kill me!” Awaiting the captured princess is torture from the Demon King’s army that is said to be more terrifying than death itself. Faced with fragrant “torture devices,” can the iron-willed princess protect the kingdom’s last remaining secrets without accidentally joining the enemy’s side? Director Yoko Kanamori continues the impressive style established in the first season, blending playful performances with ever-creative comedic gags that turn familiar tropes into endlessly entertaining moments. This fantasy food comedy—perfect for watching during meals—inevitably makes viewers exclaim, “So good.”
@SHY: Determined to earn Sakuna’s recognition, Kokoro decides to grow rice entirely on her own and arrives at a barren field on Hinome Island. With the help of machines she invents herself, everything seems to be progressing smoothly—but will things truly go according to plan? After Sakuna’s growth arc concluded in the main story, this sequel shifts its focus to Kokoro, who continues the grounded story of farming life while living on her own. Beyond its stable and relaxing narrative, the series also raises its yuri undertones to a new level—after harvesting rice, they even make dumplings together, making the ending all the more satisfying.
@SHY: The four members of the virtual band Gorillaz meet in the jungle and set out to climb a mysterious mountain, embarking on a surreal and fantastical journey. Created as a special music video for Gorillaz’s ninth album, the title comes from three of its singles. The tightly interwoven visuals and music convey a philosophical sense of cyclical existence. Produced by London’s THE LINE Studio, this fully hand-drawn short film uses cel animation and blends multiple traditional techniques to capture the texture and spirit of classic animation from the 1960s.
On March 2, the live-action series One Piece released the final trailer for its second season, which will premiere on Netflix on March 10. Starring Iñaki Godoy, Mackenyu, Emily Rudd, Jacob Gibson, and Taz Skylar, the Straw Hat crew crosses Reverse Mountain and finally reaches the Grand Line. New adventures, new allies, and new enemies await them. Source
On March 6, the series The Boys released the final trailer for its fifth and final season, which will premiere on Prime Video on April 8. This is Homelander’s world now, completely subject to his unpredictable will. When Butcher reappears with a virus capable of wiping every superhero off the map, he triggers events that will permanently change the world—and everyone in it. Source
On March 5, DC released the first trailer for its new series Lanterns, scheduled to premiere on HBO Max in August. Chris Mundy (Ozark) serves as showrunner, with Aaron Pierre playing John Stewart and Kyle Chandler as Hal Jordan. When these two intergalactic police officers investigate a murder in the American heartland, they become entangled in a dark mystery on Earth. Source
On March 5, the series Beef Season 2 released its first trailer and is set to premiere on Netflix on April 16. Produced by A24 and created once again by Lee Sung Jin, the season stars Oscar Isaac, Carey Mulligan, Cailee Spaeny, Charles Melton, and Youn Yuh-jung. A young couple witnesses a shocking argument between their boss and his wife, setting off a chain reaction of escalating consequences. Source
On March 2, the film Scary Movie 6 released its first trailer and moved its North American release date up to June 5. Directed by Michael Tiddes, the film stars Marlon Wayans, Shawn Wayans, Anna Faris, Regina Hall, Damon Wayans Jr., and Greg Wayans. Two friends once again find themselves caught in a chaotic incident involving killers, monsters, and supernatural creatures. Source
On March 5, the film The Silent Friends released a new trailer. Directed by Ildikó Enyedi and starring Tony Leung Chiu-wai and Léa Seydoux, the story is set in the botanical garden of Marburg, a medieval university town in Germany. It tells three stories connected to a single ginkgo tree over the course of a century—a silent witness that observes the quiet transformations of three lives across 100 years. The film will be imported for theatrical release in mainland China.
More
TV anime Awashima Hyakkei new trailer: Adapted from the manga of the same name by Takako Shimura (Sweet Blue Flowers), directed by Morio Asaka (Nana), with series composition by Yasuhiro Nakanishi and produced by MADHOUSE. Set at Awashima Opera School, the series tells an ensemble coming-of-age story about girls who dream of standing on stage. It will begin airing on April 10. Source
Film Sakamoto Days official trailer: Adapted from the manga of the same name by Yuto Suzuki, directed by Yuichi Fukuda and starring Ren Meguro, Fumiya Takahashi, Aya Ueto, Mayu Yokota, Eiku Shiono, and Keisuke Watanabe, with the theme song performed by Snow Man. The story follows the everyday life of former legendary assassin Taro Sakamoto. The film will premiere in Japan on April 29. Source
Film The Lukuang Canal Murders release date trailer: Directed by Lai Chun-yu and starring Berant Zhu, Julia Wu, Chang Shih, Hsiu-shen Liang, and Chu Pak-hong. Inspired by Taiwan’s first major dismemberment case, the story follows Zhao Ziwu and Zhang Xiuxiu as they become entangled in a brutal crime. Facing a shocking series of murders, they attempt to uncover the truth behind the bloody mystery. It will be released in mainland China on March 28. Source
Documentary film Moon Landing (Part One) release date trailer: The first documentary film about China’s lunar exploration program. Wu Jing lends his voice to the narration, using his distinctive presence to connect the many shining moments within this vast system and guide audiences through China’s ambitious three-step lunar exploration strategy: orbit, land, and station. The film will be released on April 24. Source
Series Out of the Dust first trailer: The six-episode series stars Asa Butterfield and Molly Windsor, with Julie Gearey (The Prisoner’s Wife, Intergalactic) serving as creator and writer. The story follows Rosie, a woman living with her husband and daughter in an isolated Christian community. After encountering Sam, an escaped prisoner, she is forced to confront the true nature and constraints of the world she lives in.
📽 Film & TV News Weekly
The Super Mario Galaxy Movie set for mainland China release on April 3 On March 2, the animated film The Super Mario Galaxy Movie released its mainland China release-date trailer and poster, confirming a theatrical release on April 3. The voice cast includes Chris Pratt, Anya Taylor-Joy, Charlie Day, Jack Black, Brie Larson, and Benny Safdie. Mario and his companions break through gravity and venture into the vast galaxy, embarking on an unprecedented interstellar adventure. Source
Hello, Amélie set for mainland China release on March 20 On March 2, the Oscar-nominated animated feature Hello, Amélie released its mainland China release-date trailer and poster, confirming a March 20 theatrical release. A piece of chocolate awakens Amélie’s senses, and with the careful companionship of her housekeeper Ms. Nishio, she learns to listen to the wind, feel the rain, and experience the light—discovering the wonder of growing up. The film was also recommended in What to Watch This Week (Issue 251205). Source
The Dream Team of Miracles releases China-exclusive poster On March 5, the animated film The Dream Team of Miracles released a China-exclusive trailer and poster, with a mainland China release scheduled for March 14. Produced by basketball superstar Stephen Curry in his producing debut and created by the team behind Spider-Man: Across the Spider-Verse, the story follows Will, an underestimated undersized player who rises to become a legendary rookie. Carrying forward the spirit of the GOAT, he writes a comeback story fueled by passion. Source
Aqara launches the Jiyue Smart Control Screen S1 Plus (Siri Edition)
Hisense unveils the UX 2026 flagship RGB-Mini LED TV
Nothing introduces the Phone (4a) series
Insta360 releases the Snap smartphone selfie screen
Google rolls out Gemini Canvas AI mode in the U.S.
OpenAI launches the GPT-5.4 model
Microsoft releases the open-source Phi-4-Reasoning-Vision-15B model
Apple Music introduces AI content provenance labels
Huami launches the Amazfit Active 3 Sapphire Edition smartwatch
Google reaches a settlement with Epic Games
Briefs worth a quick look
Aqara launches the Jiyue Smart Control Screen S1 Plus (Siri Edition)
On March 4, Aqara announced the launch of the Jiyue Smart Control Screen S1 Plus Siri Edition, the world’s first whole-home smart control panel with native Siri support.
In terms of functionality, the S1 Plus Siri Edition integrates lighting control, temperature control, and energy management systems into a single screen. Users can interact with it through Siri or the Aqara voice assistant, or directly adjust and control smart home devices from the display while monitoring the home’s overall energy usage and device status. The product also offers customizable themes, a dedicated photo album feature, and data dashboards for personalized display and information aggregation.
In security scenarios, the S1 Plus supports distributed video streaming. When a visitor presses the smart doorbell, any S1 Plus screen in the home can instantly display the visitor’s video feed and function as a video intercom interface.
Product appearance image, image courtesy of Aqara
On the hardware side, the S1 Plus Siri Edition features a 6.9-inch display with a resolution of 1440 × 720. It adopts a dual 86-type panel design, allowing it to be installed alongside standard 86 switch panels. The device includes three built-in relay switches and six wireless switches, which users can configure based on their preferences.
So far, Aqara has more than 30 products available in Apple Stores worldwide, and over 310 of its products support Apple Home. Source
Hisense unveils the UX 2026 flagship RGB-Mini LED TV
On March 5, Hisense announced the UX 2026 RGB-Mini LED flagship TV, featuring the world’s first “Linglong 4-Chip True Color Backlight System.” In addition to the traditional red, green, and blue primary backlight chips, the system introduces a fourth cyan backlight chip. These backlight chips are fully developed in-house and built to automotive-grade standards. According to Hisense, the system can maintain color stability for up to 15 years, cover 110% of the BT.2020 color gamut, and reach peak brightness of up to 10,000 nits. The 116-inch version features 43,008 color-control zones, making it the largest RGB-Mini LED TV currently available. The TV also integrates the Hi-View AI picture quality chip H7 Pro, which works with the four-color architecture to enable more precise light and color control, reaching a color control precision of 134 bits.
In other aspects, the series adopts a custom “Obsidian Screen Ultra” panel with approximately 1.28% reflectivity, delivering high contrast and wide viewing angles while supporting a native 4K 180Hz refresh rate. For audio, the TV is equipped with the Devialet Elevation Sound Pro system. Pricing starts at ¥34,999 for the 85-inch model, ¥54,999 for the 100-inch model, and ¥119,999 for the 116-inch model. Source
Nothing introduces the Phone (4a) series
At its latest launch event, Nothing unveiled the Phone (4a) series, including the standard Phone (4a) and the Phone (4a) Pro.
The standard Phone (4a) features a 6.78-inch 120Hz OLED display. On the rear, it is equipped with a 50MP main camera, a 50MP 3.5× periscope telephoto lens, and an ultra-wide camera. To the right of the camera module is a square Glyph Bar light strip with a peak brightness of up to 3500 nits, supporting multiple lighting effects.
The Phone (4a) Pro comes with a 6.83-inch 144Hz OLED panel and a body thickness of 7.95 mm. According to the company, it is currently the thinnest “full-metal smartphone” on the market. In terms of imaging, the device features a 50MP Sony LYT700C main camera, a 50MP 3.5× periscope telephoto lens, and an ultra-wide camera. It also includes a Glyph Matrix lighting system composed of 137 mini LEDs, capable of displaying battery status, timers, and a digital clock.
For performance, the Phone (4a) is powered by the Qualcomm Snapdragon 7s Gen 4 processor, while the Phone (4a) Pro uses the Qualcomm Snapdragon 7 Gen 4 processor. Both models come with a 5080mAh battery and support 50W wired charging. Pricing starts at €349 (approximately ¥2801) for the Phone (4a), and €479 (approximately ¥3845) for the Phone (4a) Pro. Source
Nothing also introduced the Headphone A, which continues the brand’s signature design language. Its overall shape resembles the Headphone 1, featuring rectangular ear cups and transparent elements, though some of the transparent parts have been replaced with opaque panels. It is available in white, black, pink, and yellow. The headphones weigh about 310g and carry an IP52 rating for dust and water resistance.
Headphone A uses physical mechanical controls, including a volume roller, toggle-style track controls, and a voice assistant button. For audio, it supports AAC, SBC, and LDAC codecs, and provides an 8-band parametric equalizer along with basic EQ adjustment options. In terms of battery life, it can last up to 135 hours with AAC playback and ANC turned off, or around 75 hours with ANC enabled.
Headphone A supports Bluetooth 5.4, USB-C, and 3.5mm wired connections, and can connect to two devices simultaneously for seamless switching. For calls, the headset is equipped with three microphones. The Headphone A is priced at $199 (approximately ¥794.7). Source
Insta360 releases the Snap smartphone selfie screen
On March 5, Insta360 announced the Snap smartphone selfie screen, a device designed to solve the problem of not being able to preview shots in real time when taking selfies with a phone’s rear camera.
In terms of design, the Snap selfie screen features a 3.5-inch display. The base version is about 7.3 mm thick and attaches magnetically to the back of a smartphone, supporting a range of Apple and Android devices. The Snap connects to the phone via a USB-C cable and works plug-and-play without requiring an independent power supply. Compared with wireless image transmission solutions, this design provides more stable and smoother video transmission, while also supporting real-time preview when recording 4K video.
In terms of user experience, Snap supports touch interaction, allowing users to quickly adjust exposure, switch filters or shooting modes, and mirror the preview with a single tap. The accessory is compatible not only with the phone’s native camera app but also with mainstream beauty and professional photography apps, enabling users to take selfies or shoot videos using the rear camera.
The product is available in two versions: a standard edition and a lighting edition, priced at ¥499 and ¥599 respectively. The lighting edition is co-tuned by Insta360 and beauty lighting brand AMIRO, and features a ring fill light with three color temperature options and five brightness levels. Source
Google rolls out Gemini Canvas AI mode in the U.S.
On March 5, Google announced that the Canvas feature, previously tested in Google Labs, is now available to all users in the United States through Gemini’s AI Mode. The feature currently supports English.
Canvas is designed to help users organize information, plan projects, and conduct in-depth research. It has now expanded to support writing documents and creating custom tools directly within the search interface. In creative and development scenarios, users can simply describe their ideas to Canvas in natural language to generate code, which can then be quickly turned into shareable applications or mini games.
The feature also applies to literary writing and other creative projects. Users can use Canvas to polish drafts, adjust structure, or receive suggestions for improvement.
In terms of workflow, users can click the “+” button in the tool menu within AI Mode and select the newly added Canvas option, then describe the task they want to complete or the content they want to create. A Canvas panel will then open on the side of the interface, where users can gather information from web pages and the Google Knowledge Graph to organize research materials or structure their projects.
If used for building prototypes or applications, users can also test functions within the same interface, view and switch the underlying code, and continuously refine the application’s behavior through conversations with Gemini. Source
OpenAI launches the GPT-5.4 model
OpenAI has announced the release of its next-generation AI model, GPT-5.4. The company states that the model brings upgrades in reasoning ability, coding performance, and handling professional workflows such as spreadsheets, documents, and presentations. OpenAI also notes that GPT-5.4 is its first model with native computer operation capabilities. By generating code as well as keyboard and mouse commands, it can perform tasks across different applications and complete actions based on screenshots. In addition, the model has improved efficiency and accuracy in web browsing, tool usage, and API calls.
OpenAI has also introduced a reasoning-focused version, GPT-5.4 Thinking, which is now available in ChatGPT. This version can display a summary of its reasoning steps when handling complex problems and allows users to adjust their requests during the generation process. According to OpenAI, GPT-5.4 performs better on questions that require retrieving information from multiple sources, enabling sustained multi-step searches and result integration. Its factual accuracy has also improved, with the probability of a single incorrect conclusion reduced by approximately 33% compared to GPT-5.2.
GPT-5.4 is now available across ChatGPT, Codex, and the API. GPT-5.4 Thinking is offered to Plus, Team, and Pro users, while GPT-5.4 Pro is available to API, ChatGPT Enterprise, and Edu users. Source
Microsoft releases the open-source Phi-4-Reasoning-Vision-15B model
On March 5, Microsoft announced the launch of the open-source Phi-4-Reasoning-Vision-15B model through its official developer community blog. This is a vision reasoning model that combines high-resolution visual perception with selective, task-aware reasoning capabilities, making it the first small language model (SLM) in the Phi-4 series that can both “see clearly” and “think deeply.”
The core feature of the model is its hybrid reasoning mechanism. When a task requires deeper reasoning—such as solving math problems or performing logical analysis—the model activates a multi-step reasoning chain. When only rapid visual perception is required—such as OCR or identifying UI elements—it outputs results directly to reduce latency and improve response efficiency.
In terms of application scenarios, the model is particularly suited for use with computer agents. After receiving screenshots and natural language instructions, the model can output standardized bounding box coordinates for target UI elements, which can then be used by other agent models to perform actions such as clicking or scrolling. The model has now been released as open source on Hugging Face.
Apple Music introduces AI content provenance labels
On March 5, Apple Music introduced a transparency labeling feature to indicate whether music content has been created using AI technologies.
Under the new rules, record labels and content distributors must clearly state whether artificial intelligence (AI) was used in the creation process—such as for music production or cover artwork—when submitting new content to the platform. To more clearly define the scope of AI involvement, Apple has introduced four Transparency Tags corresponding to the core creative elements of digital music content: cover artwork, audio track, composition, and music video.
According to Apple’s implementation guidelines, these tags will only be triggered when “a substantial portion of the content is created using AI.” Given the complexity of modern music production workflows, a single piece of music may carry multiple transparency tags simultaneously.
At the implementation level, Apple has not introduced mandatory machine-based detection mechanisms. Instead, the determination of whether content qualifies as AI-generated is left to partners to self-report. This approach is similar to existing industry practices for metadata management, such as classifying music genres or listing credits for performers and production staff, allowing creators and distributors a relatively high degree of discretion. Source
Huami launches the Amazfit Active 3 Sapphire Edition smartwatch
On March 5, Huami introduced the Amazfit Active 3 Sapphire Edition smartwatch. The watch adopts a lightweight four-button design and features a sapphire crystal display paired with a stainless steel middle frame.
The Amazfit Active 3 Sapphire Edition is equipped with a 1.32-inch high-brightness AMOLED display with peak brightness reaching up to 3000 nits. The device includes a wide range of built-in training courses covering basic cardio and mixed training types, and supports more than 170 sports modes. Running and strength training programs include multiple formats such as endurance and speed training.
For outdoor activities, the watch supports offline maps, route navigation, and turn-by-turn alerts, allowing users to plan routes in advance. It is equipped with a five-satellite positioning system for faster and more accurate location tracking. The BioTracker heart rate monitoring system provides real-time feedback on workout intensity and combines sleep, blood oxygen, and stress data to help users balance training and recovery.
In addition, the watch supports Bluetooth calling, Alipay and WeChat offline payments, as well as NFC transit cards. According to official data, the device offers around 12 days of battery life under typical usage and about 7 days under heavy use.
In terms of colors, the Amazfit Active 3 Sapphire Edition is available in options such as Champagne Silver and Sapphire Blue, with a retail price of ¥1199. Source
Google reaches a settlement with Epic Games
On March 4, Google and Epic Games reached a settlement agreement.
Under the new plan, Google will reduce the base service fee for in-app purchases from the previous maximum of 30% to 20%. If developers choose to use Google’s own payment system, an additional 5% fee will apply. Subscription service fees will be reduced to 10%.
At the same time, Google will introduce a “Registered App Stores” program. Third-party app stores that meet quality and security standards will be able to install and run more easily on the Android platform, reducing the friction users face when installing apps outside the Play Store.
As part of the settlement, Epic Games will relaunch Fortnite on the global Google Play Store while continuing to develop the Epic Games Store for Android. The new fee structure is scheduled to take effect on June 30, 2026, in the United States, the European Economic Area, and the United Kingdom, followed by expansion to Australia on September 30, 2026. It is expected to cover South Korea and Japan by the end of 2026 and expand globally by September 30, 2027.
Epic Games CEO Tim Sweeney stated that the agreement will push the Android ecosystem toward greater openness and allow app stores to compete with one another. Meanwhile, Epic’s lawsuit against Apple over App Store commissions is still ongoing, with the current ruling under appeal. Source
Briefs worth a quick look
According to Board Channels, the GeForce RTX 3060 graphics card that NVIDIA recently resumed production of will enter the market between March 10 and March 20. Earlier reports suggested that the revived RTX 3060 is an 8GB version with a 128-bit memory bus, which delivers noticeably reduced performance compared with the 12GB version. Source
According to analysis by the tech outlet wccftech, the MacBook Neo’s 8GB memory configuration may be related to the packaging design of the A18 Pro chip it uses. The A18 Pro adopts TSMC’s InFO-POP (Integrated Fan-Out Package on Package) technology, which allows DRAM to be stacked directly on top of the SoC package, forming a single integrated structure between memory and the chip. Currently, the memory package paired with the A18 Pro provides 8GB of capacity. If Apple were to redesign the chip to support 12GB of RAM, development and manufacturing costs would increase significantly. In theory, Apple could also use the A19 Pro chip with a built-in 12GB RAM package, but that would likewise raise the bill of materials cost for the MacBook Neo. With memory chip prices rising, Apple ultimately chose to offer only a fixed, non-upgradable 8GB RAM configuration in order to maintain the MacBook Neo’s relatively low price positioning. Source
Microsoft’s newly appointed CEO of gaming, Asha Sharma, revealed that the internal codename for the next-generation Xbox console is “Project Helix.” She stated that the device will focus on high performance and support running both Xbox and PC games, further integrating the console and PC ecosystems. Microsoft had previously indicated that the next-generation Xbox would be a hybrid device combining characteristics of both consoles and PCs, aiming to deliver a “premium and carefully crafted” gaming experience. Source
When Kodak first announced Charmera, it quickly sparked a wave of attention online, to the point where its price was once speculated up to around ¥300 and it was still in short supply. At the end of last year, I spotted a Charmera on display in the window of a camera store on the streets of Macau, but after asking the staff I learned there was no stock available, so I had to give up. It wasn’t until shortly before the Lunar New Year that I happened to notice the price on Taobao had dropped quite a bit. With discounts applied, it could be purchased for under ¥200, which felt quite attractive to me.
I ended up pulling the gray version. Although it’s not as recognizable as the one on the box cover, nor as special as the transparent hidden edition, it gives me a strong early-Apple vibe. The gray-white shell paired with the rainbow stripe design feels genuinely retro. Even though Charmera is only about the size of a keychain, the body still includes a super tiny viewfinder cutout to enhance the retro aesthetic. However, this “viewfinder” is really just a hole without any optical design, so naturally it doesn’t have much practical use.
Left: Fujifilm X-E4, Right: Kodak Charmera
Charmera features a lens with a 35mm equivalent focal length, capable of shooting photos at 1080×1440 and recording video at 30fps. But the actual image quality can be summed up in one word: terrible. Or perhaps more accurately, we shouldn’t have any expectations for its image quality in the first place. If you approach it with that mindset, the output actually feels like it “has the vibe.”
While the Y2K aesthetic has returned to the spotlight and many people are trying various ways to recreate that “retro” feeling, Charmera is the kind of device where photographing it works, but using it to photograph others doesn’t quite. If you treat it as a retro accessory integrated into your outfit, it might add some character to your photos. But if you expect it to capture images that are both retro and sharp, that’s a very tall order.
Although the image quality is poor, the built-in filters and frames help make up for it somewhat. In particular, the four Kodak-style frames are quite fun to use.
If you evaluate Charmera by the standards of a typical camera, I think that would be unfair. Its positioning is closer to that of a low-resolution toy in the era of high-definition imaging. Wear it as a pendant, and when you spot something interesting, simply pull it out and take a quick shot without any pressure—perhaps that’s exactly the meaning of its existence.
Even though the image quality isn’t great, it really is fun to play with. Especially during the holidays, pulling out such a quirky little gadget in front of relatives naturally makes everyone want to try taking a photo with it just to see what happens. Compared to everyone sitting alone in corners scrolling on their phones, using something like this to bring people together might be exactly what many people seek when they talk about nostalgia.
All in all, I think Charmera is a little gadget with very clear strengths and weaknesses. If you’re buying it purely for photography, it simply can’t compete with a smartphone. But for me, the emotional value it provides already makes it worth the price. By the way, knockoff versions have already appeared on Taobao. The image quality is roughly the same, but they lack Kodak’s four signature frames. Considering some of them are priced at over ¥180, you might as well just buy the real Charmera instead.
@Clyde: Xiaomi BE3600 Pro Wired Version
Reference price: ¥1507.90
Let me put the disclaimer first: I’m aware that Xiaomi has “cut corners” on some router models like the BE6500 Pro and AX3000T. I still chose Xiaomi anyway, mainly because I’m lazy.
After tinkering with VLANs, single-cable multiplexing, and IPTV setups for half a year, I finally got a bit tired of it all. One night, when my Redmi AX6s rebooted again due to performance issues, I decided to take the easy route without drastically changing my setup habits and ordered the Xiaomi BE3600 Pro wired version that had been sitting in my shopping cart for quite a while.
The installation and configuration weren’t too difficult. The main router goes into the weak-current cabinet and connects to the optical modem. After connecting each AP panel through wired networking, you can install them according to the network port locations in each room. Here, the Xiaomi BE3600 Pro wired version offers a bit of convenience for those who prefer the lazy route—you can either have an installer come over, remove the existing wall network panel, re-terminate the cable with an RJ45 connector, and mount the AP panel directly onto the wall like a regular network outlet; or you can take the lazy option and simply connect it and stick the panel onto the existing wall outlet using the adhesive backing included with the device.
For example, I chose the lazy way.
For most AP panel products, heat dissipation is an unavoidable issue. Xiaomi equips the BE3600 Pro wired version’s AP panel with a small fan, and you can choose “Balanced Mode” or set temperature thresholds for fan activation and shutdown through the Mesh node management page in the web interface. Perhaps because the weather isn’t too hot right now, even the AP panel installed in the corner of my bedroom hasn’t produced any fan noise noticeable during late-night quiet hours.
But it does get warm. I measured it with a Pixel 9 Pro (Google: see, the temperature sensor does come in handy!). With my Mac connected via Ethernet, the AP panel runs at around 42–45°C in a room temperature of roughly 20°C—similar to how a Pixel 9 Pro feels after about ten minutes of gaming. It’s not exactly comfortable to touch, but fortunately you don’t need to keep holding it all the time.
As for coverage, I bought a bundle with one main router and four sub nodes. For a 90-square-meter apartment this is actually a bit excessive, but since I didn’t plan the network port locations very well when reserving cables earlier, I had to rely on quantity to ensure full coverage. This also ties into another reason I stuck with Xiaomi: there are already quite a few Mi Home ecosystem devices in my house. The main router of the 3600 Pro wired version can act as a Mi Home smart hub, featuring the “Automation Geek Version” (essentially a graphical, web-based automation editor), and it can run local automation commands even when the internet connection is down.
Meanwhile, each AP panel can also function as a Bluetooth Mesh gateway for connecting Mi Home devices. Placing one in every room allows nearby smart devices to connect directly, which means one less gateway device—and potentially one more available power outlet or Ethernet port. That kind of integration is something I really appreciate.
In terms of personal usage habits, the BE3600 Pro wired version also supports customizable port functions. With a simple setup, the Ethernet port on the living room AP panel can be configured to connect a set-top box for IPTV. The problem is that each AP panel only has one 2.5GbE port. While using it for IPTV doesn’t affect the panel’s wireless and AP functions, other devices in the same area won’t be able to connect via Ethernet unless you add a switch and configure VLANs again… which basically puts you back where you started. Fortunately, both speed test websites and macOS’s built-in networkquality tool show pretty good results so far—maybe because it’s still relatively new and hasn’t been “downgraded” yet (laugh).
Another complaint is that Xiaomi’s router software experience is currently quite fragmented. Some settings you can’t find in the Mi Home app might appear in the web interface, and vice versa. Also, even though Xiaomi releases new router models every year, the web interface UI still looks exactly the same as it did more than a decade ago—when will it finally become a bit nicer and more user-friendly?
@Microhoo: RS 5 Camera Stabilizer
Reference price: ¥3899
As a struggling student who nevertheless loves buying stationery, once I decided to move beyond still photography and try shooting video, I often found myself longing for a cool and imposing camera stabilizer. The reason I kept hesitating was that these devices all wear the intimidating label of “professional.” From balancing to the various stabilization modes, everything feels a bit overwhelming—especially when I’m still figuring out basic camera movements, making it hard to feel confident about carrying around such a large piece of equipment.
DJI has probably seen plenty of beginners like me who feel a bit intimidated, so with the RS 4 Mini they tried adding a small intelligent tracking module. This makes it possible to track subjects easily even without strong shooting skills or camera movement techniques. Still, it wasn’t perfect. On the one hand, it could only recognize people, which was somewhat limiting. On the other hand, the lack of a monitoring system meant that automatic tracking always carried a bit of uncertainty. But I believed DJI wouldn’t leave it there. The significance of this feature isn’t just simplifying tracking—it also reflects a broader vision of making creative tools more accessible, allowing beginners to start easily with professional equipment and gradually refine advanced skills through use. Sure enough, the RS 5 that followed, although positioned as a professional and even commercial production tool, may actually be the most beginner-friendly camera stabilizer yet.
The RS 5 naturally continues and further enhances the small intelligent tracking module introduced on the RS 4 Mini. Not only can it now track any type of object, but for clearly identified human subjects the tracking distance has been extended to 10 meters, and it can quickly reacquire the subject even after they briefly leave the frame.
What meets my expectations most is that DJI now displays the tracking module’s captured view directly on the RS 5’s touchscreen. This allows you to preview the tracking status in real time, quickly tap on the screen to select or switch tracking targets, and simultaneously control the stabilizer’s gimbal movements.
In real-world use, people and animals are the subjects that the system recognizes and maintains tracking on most reliably. But if the background is relatively clean, even small objects like pots and pans can be tracked accurately—though you may need to adjust movement speed and avoid interference, otherwise the system might lose the subject.
During tracking, besides keeping the subject centered, you can also customize the composition by placing the subject to the left or right of the frame. There is a small drawback, however: because the tracking module is positioned slightly offset from the camera lens, the image shown on the touchscreen differs from the actual footage captured by the camera. This means perfect centering cannot be guaranteed, and when using more flexible compositions there’s an even greater chance that the subject may slip out of the camera’s actual frame. This offset isn’t much of an issue when using wide-angle lenses, but when shooting at longer focal lengths it requires careful attention.
Based on my understanding, the tracking module likely uses a wide-angle lens, meaning its field of view is probably much larger than the camera’s. DJI could potentially address this through a firmware update by adding a calibration step, allowing users to align the touchscreen preview with the camera’s actual shooting perspective before filming.
Beyond intelligent tracking, another improvement on the RS 5 that is extremely beginner-friendly is the addition of fine-adjustment knobs on all axis arms. However, I do hope DJI might reconsider the name “fine adjustment,” because “balancing knobs” might be a more accurate description.
When I first saw them, I assumed they were merely a small convenience—useful for precise positioning once the balancing was already roughly done—but not something that fundamentally solved the stabilizer balancing problem. In reality, though, they greatly improve balancing efficiency. Compared with the traditional method of manually sliding the arms back and forth, these knobs require less force and offer greater precision, which is especially helpful for beginners who haven’t yet developed muscle memory with stabilizers. Even so, I still hope future camera stabilizers will eventually support automatic balancing like smartphone gimbals—that would truly be revolutionary.
Another improvement on the RS 5—perhaps less targeted at beginners but greatly enhancing the user experience—is the addition of electronic control buttons on the briefcase handle. This significantly improves ease of operation when shooting from low angles. If you’re already experienced with stabilizers and don’t find the previous two features particularly exciting, this detail will likely still prove very useful.
As for other improvements such as the new-generation stabilization algorithm and motor efficiency upgrades, I won’t go into too much detail. My experience with stabilizers is still limited, so I can’t comprehensively describe every iterative improvement. What I can say with confidence is that if you’re like me—someone trying a camera stabilizer for the first time and hoping for the lowest barrier to entry with the highest efficiency—starting with the RS 5 is definitely a good choice.
This year during the Lunar New Year, I tried a different approach when buying gifts for relatives. Instead of passing along those well-traveled “aged” gifts that have already circulated through several households, I figured it would be more practical to use the government subsidy program to buy trendy home appliances that could genuinely improve the lives of the elders. I learned that my relatives’ kitchen had plenty of space, so placing a countertop dishwasher wouldn’t be a problem. I had also heard from other relatives that a certain brand’s dishwasher worked well, so it seemed like a natural choice. While searching for the same model, I discovered a new 2025 version with a seven-place capacity and a stainless steel interior. With the national subsidy applied, it cost just over ¥2100, so I placed the order without much hesitation.
JD Logistics was extremely fast—it arrived on the third day. For large appliances like this, the purchase price usually includes on-site installation service, but if you install it yourself you can request a partial refund from customer support. In my case, I got ¥50 back. I was initially worried that the elders might struggle with installation, but they were actually eager to give it a try and get some exercise. As soon as the package arrived, they unpacked it themselves. The installation turned out to be very simple: just place the machine on the countertop and connect two water hoses. Aside from the included drain hose being slightly too short, everything went smoothly.
The white exterior looks quite clean and refreshing on the countertop, and loading bowls and chopsticks feels very convenient. In comparison, the built-in dishwasher at my own home—where you have to bend down to load it—suddenly feels incredibly clumsy. The seven-place capacity is more than sufficient for the daily meals of three elderly people, and there’s even room for a small pot after the dishes are loaded. On days when fewer plates are used, accumulating them for a once-a-day wash is also perfectly feasible. During the initial test run, some yellowish material appeared inside the machine. At first I thought I had received a second-hand unit with rust inside, but customer service explained that it was moisture residue from the softener resin and could be cleaned off with toothpaste. That explanation was a new one for me.
The real reason dishwashers are hard to give up once you start using them is that they harness the power of industrial automation to achieve a level of cleanliness that traditional hand washing and dish soap can hardly match, all while consuming only a small amount of water, electricity, and detergent. My argument to persuade the elders was simple: “After dinner, you should go out for a walk—let the dishwasher handle the dishes and free up your time.” After three rounds of fruit-and-vegetable washing and two standard wash cycles, even the old bowls stained by years of herbal medicine residue came out clean. At that point, the elders finally agreed to keep this “white intern” around.
The main drawback, however, is that the elders unanimously feel the drying performance is insufficient. On several occasions, there was still moisture left after the drying cycle. If it could dry thoroughly every time, it could fully replace the dish sterilizer cabinet sitting beside it. The product page claims it can function as a sterilizer cabinet and supports a 168-hour fresh-air storage mode, but during actual operation the fan noise is quite noticeable. At night it becomes unpleasant enough that you have to turn it off manually.
Returning to the beginning, impulsive purchases sometimes bring unexpected consequences. The machine itself is labeled entirely in Chinese, and the outer packaging simply lists a trading company from Guangdong. I’m almost certain that this “Gejiano” is a white-label product. Searching the internet reveals no meaningful information about the brand. At this point, the recommendation algorithm unexpectedly came to the rescue: it suggested a seven-place dishwasher from Royalstar. Apart from the logo, the exterior design and product page looked exactly the same. The only thing that slightly comforted me was that the Royalstar model didn’t qualify for the national subsidy, so its final price was noticeably higher than the one I bought. With that thought, I decided to let it go.
If any SSPAI readers have used other reliable countertop dishwashers, feel free to recommend them in the comments—after all, there are still plenty of relatives in the family who need gifts.
Setapp allows individual app purchases or subscriptions
OpenAI confirms an upcoming new model
Keychron releases the B11 Pro foldable keyboard
ASUSTOR launches the flagship NAS AS7224RDX
Apple introduces MacBook Neo
On March 4, Apple officially unveiled the long-rumored “colorful entry-level” MacBook, with the new lineup named MacBook Neo.
Further viewing: Hands-on photos of MacBook Neo from SSPAI
MacBook Neo is powered by the A18 Pro processor, comes standard with 8GB of memory and a 13.0-inch Liquid Retina display, and features one USB 3 port and one USB 2 port. The device supports up to 16 hours of streaming video playback. It weighs 1.23 kg and includes a dual-speaker system with spatial audio support, Wi-Fi 6E, Bluetooth 6, and a 1080p FaceTime HD camera. The base model includes 256GB of storage, while the 512GB version additionally comes with Touch ID.
MacBook Neo is available in four colors: silver, peach pink, citrus yellow, and indigo blue. The 256GB version is priced at 4,599 yuan, while the 512GB version costs 5,299 yuan. The product will officially go on sale on March 11. Source
Product appearance images, provided by the news source
Claude Code launches voice mode
On March 3, Anthropic announced the gradual rollout of a new Voice Mode in its programming tool Claude Code, signaling that the developer-favored concept of “intuitive programming” is evolving toward a more efficient era of “voice programming.”
The voice module is built directly into the underlying architecture of Claude Code. After receiving beta access, users will see a prompt prominently displayed on the welcome screen inviting them to enable the feature. During actual use, developers only need to press and hold the space bar on the keyboard to speak and input development requests, then release the key to quickly send the voice command to the AI for execution.
In terms of pricing, Anthropic emphasized that the voice interaction feature will not incur any additional subscription fees. As the feature rolls out more broadly, users on the existing Pro, Max, Team, and Enterprise tiers will all be able to experience this new capability designed to improve development efficiency at no extra cost. Source
Setapp allows individual app purchases or subscriptions
On March 3, the MacPaw team announced that it will introduce new purchasing options for its Setapp platform, including standalone subscriptions and one-time purchase plans for specific apps. Under the new model, applications such as Bartender, Downie, and AlDente Pro, along with MacPaw’s own apps like CleanMyMac, Moonlock, and Gemini 2, will offer a variety of subscription and purchase plans, including monthly, yearly, and one-time purchase options. These purchase or subscription plans can all be accessed through Setapp, without requiring users to subscribe to the full Setapp service.
Currently, more than 60 apps support the new purchase and subscription options, and the list of participating applications will be updated and announced regularly in the future. Source
OpenAI confirms an upcoming new model
On March 4, shortly after launching the GPT-5.3 Instant model, the OpenAI team stated on social media that GPT-5.4 “will arrive sooner than you think.” Since the GPT client shows that the older 5.1 model will be discontinued on March 11, some users speculate that OpenAI may officially release version 5.4 around that time. Source
Keychron releases the B11 Pro foldable keyboard
On March 4, the Keychron brand introduced the B11 Pro foldable portable keyboard. The product features a flat Alice layout, which is more ergonomic than the traditional horizontal layout, while remaining compatible with macOS, Windows, and Linux. The B11 Pro adopts an inward-folding design, with an ABS body and a matte black PU leather back. It uses scissor-switch mechanisms and comes with concave ABS keycaps. In terms of connectivity, the keyboard supports tri-mode connections including USB-C wired, 2.4GHz wireless, and Bluetooth 5.3 wireless, with a “dual 1kHz” polling rate. The entire keyboard weighs 258g and includes a built-in 250mAh battery rated for up to 138 hours of battery life, supporting automatic power-off when folded and automatic power-on when unfolded. The price is set at $64.99. Source
Product appearance image, photo sourced from news outlets
ASUSTOR launches the flagship NAS AS7224RDX
On March 3, ASUSTOR, a subsidiary of ASUS, announced the launch of its 4U rackmount flagship NAS, the AS7224RDX (Lockerstor 24R PRO Gen2), featuring 24 drive bays. The AS7224RDX is powered by an AMD Ryzen 7 PRO 7745 eight-core processor and provides 24 SATA bays supporting either 3.5-inch or 2.5-inch drives. It also includes PCIe Gen5 ×8 and PCIe Gen5 ×4 expansion slots, one M.2 2280 PCIe Gen5 ×4 slot, two 10GbE network ports, and two 1GbE network ports.
In addition, the device features four DDR5 UDIMM memory slots supporting up to 192GB of ECC memory, and comes preinstalled with 16GB DDR5-4800 ECC memory and a 32GB SATA DOM. It is powered by a pair of 550W Platinum redundant power supplies and equipped with four 60mm cooling fans. The system also provides four USB-A 5Gbps ports and one HDMI port for service purposes. Source
Product appearance images, provided by the news source
Apple unveils M5 series chips, Studio Display, and new MacBook models
Unihertz releases the new full-keyboard smartphone Titan 2 Elite
Four AI updates
Google releases the Spring 2026 Pixel Feature Drop
Speedtest and Downdetector are being sold
Chrome adjusts the update cycle for the stable version
Highguard announces shutdown on March 12
Briefs worth a quick look
Apple unveils M5 series chips, Studio Display, and new MacBook models
On March 3, Apple introduced the M5 series chips and simultaneously refreshed the MacBook Pro, MacBook Air, and Studio Display / Display XDR product lines.
The M5 series adopts a Fusion architecture that integrates two 3-nm chiplets into a single SoC. The M5 Pro and M5 Max feature a new 18-core CPU architecture composed of 6 super cores and 12 performance cores. Multi-thread performance improves by up to 30% compared with the M4 series and up to 2.5× compared with the M1 series. GPU configurations scale up to 40 cores, with each core integrating neural network accelerators. Peak AI compute performance improves by more than 4× over the previous generation, while ray tracing performance increases by up to 35%. In terms of memory, M5 Pro supports up to 64GB with 307GB/s bandwidth, while M5 Max supports up to 128GB with 614GB/s bandwidth. The M5 series also integrates a 16-core Neural Engine, a media engine supporting AV1 decoding, a Thunderbolt 5 controller, and memory security protection features. Source
New MacBook Pro models powered by the M5 series start at RMB 17,999 for the 14-inch M5 Pro version and RMB 21,999 for the 16-inch model; the 14-inch M5 Max version starts at RMB 29,999 and the 16-inch version at RMB 31,999. Orders open on March 4, with official sales beginning on March 11. Source
Apple also introduced 13-inch and 15-inch MacBook Air models powered by the M5 chip, featuring a 10-core CPU and up to a 10-core GPU. Base storage starts at 512GB and can be configured up to 4TB. The new generation SSD delivers twice the read and write performance. In terms of connectivity, the devices debut Apple’s N1 wireless chip, supporting Wi-Fi 7 and Bluetooth 6, and include two Thunderbolt 4 ports capable of driving up to two external displays. The machines use a fanless silent design and feature a 12MP Center Stage camera and Liquid Retina display, with battery life of up to 18 hours. The new MacBook Air is available in Sky Blue, Midnight, Starlight, and Silver, with the 13-inch model starting at RMB 8,499 and the 15-inch model at RMB 9,999. Orders open on March 4, with sales beginning March 11. Source
Finally, the new Studio Display features a 27-inch 5K Retina display with brightness up to 600 nits. The Studio Display XDR upgrades to a 27-inch 5K Retina XDR panel using mini-LED backlighting with 2,304 local dimming zones. It supports a 120Hz ProMotion adaptive refresh rate (47Hz–120Hz), SDR brightness up to 1000 nits, HDR peak brightness up to 2000 nits, a contrast ratio of 1,000,000:1, and newly adds support for the Adobe RGB color gamut. Both displays include a 12MP Center Stage camera with Desk View support, a studio-quality three-microphone array, and a six-speaker system supporting spatial audio. For connectivity, the lineup adds Thunderbolt 5 downstream support, providing two Thunderbolt 5 ports and two USB-C ports, and supports daisy-chaining up to four displays. Each unit includes a Thunderbolt 5 Pro cable. Studio Display supports up to 96W power delivery, while the XDR version supports up to 140W reverse charging.
The Studio Display with a tilt-adjustable stand starts at RMB 11,999, while the Studio Display XDR with tilt- and height-adjustable stand starts at RMB 24,999. Orders open on March 4, with sales beginning March 11. Source
Unihertz releases the new full-keyboard smartphone Titan 2 Elite
On March 3, Unihertz announced the Titan 2 Elite, a new full-keyboard smartphone in its Titan series, at MWC 2026. The device adopts a compact candy-bar design reminiscent of classic BlackBerry phones and comes in black and orange color options. It features a 4.03-inch AMOLED display with a 120Hz refresh rate, and the secondary rear display from the previous generation has been removed to achieve a slimmer and lighter design. The phone is powered by the MediaTek Dimensity 7400 processor, with an alternative Dimensity 8400 Pro version also available, and includes 12GB of RAM and 512GB of storage. The 3.5mm headphone jack has been removed, while the programmable side button and infrared transmitter remain.
On the software side, the Titan 2 Elite ships with Android 16. Unihertz promises security updates through 2031 and Android version upgrades up to Android 20. The Titan 2 Elite will launch on Kickstarter for crowdfunding in March this year, with pricing yet to be announced. Source
Four AI Updates
OpenAI Releases GPT-5.3 Instant On March 3, OpenAI released GPT-5.3 Instant. The new model improves response relevance, conversational fluency, and web search capabilities, while significantly reducing mechanical replies, redundant disclaimers, and overly stiff tone tendencies. Compared with GPT-5.2 Instant, it particularly improves responses involving sensitive topics, addressing previous issues where the model could appear overly cautious or overly preachy. It also enhances the integration quality of web search results, reduces excessive reliance on search outputs, and resolves earlier cases where responses occasionally contained “link dumping” or loosely organized information. The update also substantially adjusts dialogue styles that users previously described as “awkward,” “out of place,” or overly forceful in tone, while significantly reducing hallucination rates across a wide range of topics.
GPT-5.3 Instant is available immediately to all ChatGPT users. Developers can access the model through the API under the name gpt-5.3-chat-latest. Updates for the Thinking and Pro variants are expected to roll out soon. GPT-5.2 Instant will remain available for paid users under the “Legacy Models” dropdown menu for three months and is scheduled to be officially retired on June 3, 2026. Source
Google Releases Gemini 3.1 Flash-Lite On March 3, Google officially launched Gemini 3.1 Flash-Lite. The model features a context window of up to one million tokens and supports multimodal inputs including text, images, audio, and video. It delivers a maximum output speed of 389 tokens per second. Pricing is set at $0.25 per million input tokens and $1.50 per million output tokens. The model’s knowledge cutoff date is January 1, 2025.
A preview version of Gemini 3.1 Flash-Lite is now available to developers via Google AI Studio and the Vertex AI platform. Source
Anthropic Introduces Memory Import Tool On March 3, Anthropic announced that the memory feature of its AI assistant Claude is now available to free users. The feature is implemented through standardized prompts: users can copy a specific prompt provided by Anthropic into other AI chat tools and request them to export all stored user preferences, personal details, project goals, and behavioral adjustments. The returned structured text can then be pasted into Claude’s memory settings to complete the import and integration process. Source
Alibaba Releases Qwen3.5 Small Model Series On March 2, Alibaba’s Qwen team released the Qwen3.5 Small model series, including four parameter sizes: 0.8B, 2B, 4B, and 9B. Built on the Qwen3.5 architecture, the series features native multimodal capabilities and has been optimized through large-scale reinforcement learning (RL). The 0.8B and 2B versions focus on high-speed inference for edge devices, the 4B version is positioned as a lightweight multimodal foundation for AI agents, and the 9B version aims to narrow the performance gap with larger models. Base models are also provided across the entire lineup. The weights have been released on Hugging Face and ModelScope, where they are freely available for research and industrial applications. Source
Google releases the Spring 2026 Pixel Feature Drop
On March 3, Google rolled out the March Pixel Drop for Pixel devices. The update adds real-time location sharing in Google Messages to the Find Hub tracking network and introduces lost luggage tracking in partnership with 10 airlines (requiring compatible tags). A customizable Calling Card feature is also introduced, allowing users to personalize their call screens with photos and fonts, while the Google Play Store adds a short-video feed to enhance app discovery. Pixel 10 series–exclusive Magic Cue now supports restaurant recommendations, and the Circle to Search feature expands with enhanced virtual try-on and video shopping capabilities. The home screen also gains six AI-generated icon styles. Pixel Watch receives several updates, including phone-loss alerts, quick payments without waking the device, earthquake alerts, and satellite SOS coverage across Europe, Canada, Hawaii, and Alaska. These updates will roll out gradually starting today. Source
Speedtest and Downdetector are being sold
On March 3, digital media company Ziff Davis announced that it will sell its Connectivity business unit, which includes Ookla Speedtest and Downdetector, to Accenture for $1.2 billion in cash. The division generated $231 million in revenue in 2025, representing a substantial return compared with the $15 million Ziff Davis paid to acquire it in 2014. Ziff Davis stated that the deal is intended to sharpen the company’s strategic focus, with future efforts centered on core media brands such as IGN, Mashable, and Everyday Health. The transaction is expected to close in the coming months, and both Speedtest and Downdetector will continue operating normally during the transition period. Source
Chrome adjusts the update cycle for the stable version
On March 3, Google announced that the update frequency for the Chrome browser’s stable release across all platforms will change from once every four weeks to once every two weeks. The adjustment applies to desktop, Android, and iOS platforms, aiming to accelerate feature delivery and simplify post-release debugging. The first stable version under the new schedule will be Chrome 153, scheduled for release on September 8. Beta versions will be released three weeks ahead of the stable version, while weekly security updates for milestone versions will remain unchanged. The Canary and Dev channels, as well as the extended eight-week stable release cycle for enterprise users, will continue as before.
For Chromebook hardware, ChromeOS will follow the new cadence after completing platform-specific testing. Further details regarding update policies for managed devices will be announced later. Source
Highguard announces shutdown on March 12
On March 3, developer Wildlight Entertainment announced that its shooter game Highguard will officially shut down on March 12. The title was developed by a team composed of former members from Apex Legends, Call of Duty, and Titanfall, and was showcased as the closing highlight at The Game Awards 2025. Since its release on January 26, 2026, the game once attracted up to 2 million players but ultimately failed to establish a player base large enough to sustain long-term operation. The studio is reportedly partially funded by Tencent and laid off most of its staff in February. The game will receive its final update over the next two days, including a new character and new weapons. Source
Briefs worth a quick look
Swedish media outlet Svenska Dagbladet reported that Meta AI smart glasses (such as Ray-Ban Meta) allegedly captured users’ private videos and financial information without their knowledge and sent the data to reviewers outside the European Union. The claim was revealed by an AI data annotator in Nairobi, Kenya, who said that while processing visual data used to train large language models (LLMs), they encountered highly sensitive content including nudity, people using the toilet, intimate behavior, and even credit card numbers. Meta declined to comment on the report. Source
Apple’s official regulatory compliance page briefly listed a previously unknown device called MacBook Neo (Model A3404), identified as part of the 2026 MacBook lineup, though the page has since been removed. According to earlier rumors, the device is expected to target the education and consumer entry-level market and may use an iPhone-series processor rather than an M-series chip. Source
If you watch a lot of films and TV dramas, you’re probably no stranger to the “amnesia” plot device. In tearjerker K-dramas especially, amnesia, car accidents, and terminal illness are often treated as a standard trio—a go-to formula for emotional storytelling. But how common is amnesia in real life, and what actually causes it? In this article, we’ll take a closer look at the typical “amnesia” scenarios found in classic film and television, while also introducing some basic neuroscience along the way—so the next time you encounter such a storyline, you’ll be ready with a sharp and informed critique.
Car Accident Amnesia: Who Am I? Where Am I? Who Are You?
This is probably the most common type of amnesia seen in film and television.
The protagonist typically suffers a series of physical head injuries—car crashes, being struck by a heavy object, tumbling down stairs or off a cliff, or getting severely beaten—and falls into a coma. In the next scene, they wake up in a hospital room, touching the bandages on their head, staring blankly at the concerned faces around them, and uttering the ultimate existential questions: Who am I? Why am I here? Who are you? What happened? At this point, we can predict the rest of the plot with our eyes closed: the hero mistakes their true love for a stranger, treats their enemy as a friend, and after countless twists and turns, falls in love all over again (not really, but you get the idea).
Psychogenic Disorders: Dissociative Amnesia
In reality, complete and global memory loss is extremely rare—but such cases do exist. The condition is known as dissociative amnesia.
We often see that the protagonist forgets who they are, who their parents and lovers are—these are all forms of episodic memory, more precisely, autobiographical memory1. Dissociative amnesia can indeed affect a person’s sense of personal identity, while leaving intelligence and everyday skills largely intact. So when the protagonist wakes up clear-eyed and articulate, yet unable to recognize themselves or their loved ones—strictly speaking, that part is not entirely implausible.
The biggest scientific flaw in these storylines lies in the triggering mechanism. In reality, dissociative amnesia is a psychogenic disorder, not the result of structural brain damage. In other words, true dissociative amnesia is a defense mechanism of the brain, usually triggered by extreme psychological trauma. Survivors of severe violence, sexual assault, war massacres, or catastrophic natural disasters may find reality too brutal to bear. To protect the individual from overwhelming pain, the brain essentially strips away the concept of “self” along with the traumatic memories.
Moreover, even when memory loss occurs, patients in real life are more likely to exhibit localized amnesia—for example, forgetting the period during which they were abused, or being unable to recall the death of a loved one—rather than wiping out decades of memories, including their own name and parents, as TV protagonists often do. If someone truly forgot that thoroughly, it would likely indicate a severe psychiatric disorder, not something that could be cured by yet another car accident or a second fall.
Stills from “The Delivery Master” (formerly titled “GIFT”)
In psychology, “dissociation” is a core concept. Put simply, it refers to a state in which consciousness becomes detached from reality. Everyday experiences like zoning out or daydreaming can be seen as mild forms of dissociation. Under increased stress, some people may experience a sense of detachment, feeling that the world around them is unreal, or that their body does not belong to them—as if they are watching themselves from the outside, like a movie. In more severe cases, dissociative amnesia may occur, or even dissociative identity disorder—commonly known as “multiple personality disorder”—in which the mind splits consciousness into several identities to distribute psychological pain, typically as a result of trauma.
There is also a more dramatic phenomenon called “dissociative fugue,” a rare subtype of dissociative amnesia. In real life, a person may suddenly leave their familiar environment and embark on unplanned, purposeless wandering. During this period, they remain conscious, capable of self-care, and able to interact normally with others, yet cannot recall their past. After the episode ends, they may also be unable to remember what happened during the wandering itself—it is almost like sleepwalking while awake. The good news is that this type of amnesia is usually reversible, and memory often returns quickly once the fugue state ends. The wandering trope is also popular in film and television: the protagonist not only loses their memory but relocates to another city, adopts a new name, and builds an entirely new identity. As someone with a strong attachment to logical consistency, I can’t help but complain every time I see this: wouldn’t their family report them missing? Wouldn’t someone who doesn’t know who they are go see a doctor? Why does the plot always prioritize romance above all else?
Traumatic Brain Injury: Post-Traumatic Amnesia
If someone in real life is involved in a car accident or sustains a severe concussion in a boxing match, what does waking up actually look like? There is a specific term for this: post-traumatic amnesia. It typically manifests as a memory gap lasting from a few minutes to several hours. The person not only cannot recall what happened at the moment of impact, but may also have no memory of a period of time before and after the accident. Moreover, this missing segment of memory is usually permanent and cannot be recovered.
Why does this happen? To understand that, we need to introduce an important structure in the brain—the hippocampus.
Location of the hippocampus in the brain
Many people know that the hippocampus is related to memory, but it’s easy to misunderstand its role and assume that it is where memories are stored. In fact, the hippocampus functions more like a computer’s RAM. Its core job is not to store memories, but to process and transfer them. Whether we experience an event or learn a new fact, the information is first temporarily held as short-term memory. After being organized and encoded by the hippocampus, it is then uploaded to different regions of the cerebral cortex and stored as long-term memory. This process is known as memory consolidation.
However, the hippocampus has a critical weakness: it is extremely sensitive to oxygen deprivation and physical shock. When a severe concussion occurs, the hippocampus can temporarily shut down. At the moment of impact, although the person may see blinding headlights or hear screams around them, the information enters short-term memory but cannot be further processed or consolidated because the hippocampus is not functioning properly. Even memories from the minutes before the accident—still “queued up” in the hippocampus awaiting consolidation—may vanish permanently. It’s like losing power while writing a document: no matter how many times you restart the computer, the unsaved lines are gone for good.
Memory tip: “Retrograde” means forgetting earlier events; “anterograde” means forgetting more recent events.
When memory loss is defined by the time of an accident or illness, the inability to remember events that occurred before it is called retrograde amnesia. The dissociative amnesia discussed earlier also falls under retrograde amnesia, though the time span of forgotten memories is usually much longer and the underlying mechanism is entirely different. In contrast, anterograde amnesia refers to the inability to form new memories after the event, meaning the person cannot remember anything that happens afterward.
In fact, for concussion patients, retrograde amnesia is not the only issue. What often causes greater alarm is anterograde amnesia—a period of “short-term memory loss” lasting minutes or even hours after regaining consciousness. The car accident is over, the patient is awake and able to speak, yet they cannot retain newly occurring information: they may repeatedly ask what happened, or forget a doctor’s instructions moments after hearing them. It’s understandably frightening, raising concerns about whether something is seriously wrong with the brain.
This happens because although the person has regained consciousness, the hippocampus has not fully recovered its function. The brain is effectively operating with malfunctioning “RAM,” unable to transfer short-term memories into long-term storage. The good news is that this type of anterograde amnesia is usually functional and temporary. For most concussion patients, a few hours of rest is enough for the hippocampus to come back online and restore normal memory consolidation. However, if a patient still cannot form new memories after 24 hours, the condition is no longer considered a simple concussion, or mild traumatic brain injury. It may indicate moderate to severe brain injury, potentially involving structural damage such as bleeding or contusions—situations that are far more serious.
A classic of psychological suspense and a Douban Top 250 staple—say one word too many and you risk spoilers. Leonardo DiCaprio plays Teddy, a U.S. federal marshal who arrives at an isolated island to investigate a case. As the story unfolds, the truth takes a dramatic turn. A textbook-level thriller with strong scientific grounding, and absolutely worth watching.
Starring Matt Damon. The protagonist is shot in the back and rescued at sea, with no memory of who he is, yet possessing elite combat and language skills. The amnesia here mainly serves to create suspense, but it does attempt some scientific plausibility, attributing memory loss to both physical and psychological trauma. The plot isn’t overly cerebral, but it’s incredibly entertaining—especially the first three films, all highly recommended.
A 1997 Fuji TV Wednesday Theater production. Takuya Kimura appears in a different Gucci haute couture suit each episode, with bold scenes that give it the vibe of a late-night drama. It was later banned following Japan’s infamous butterfly knife incident, adding a layer of mystique. This darkly humorous episodic series portrays various social outsiders. The cause of the protagonist’s amnesia is ultimately revealed, incorporating psychogenic elements into the narrative. Well worth a positive review.
A classic early Conan theatrical release. After witnessing Inspector Sato being shot, Ran develops psychogenic amnesia due to intense guilt and psychological trauma, awakening unable to recognize even her parents or Shinichi. While the amnesia setup is dramatically exaggerated, the film includes detailed medical touches—upon waking, Ran is immediately tested on semantic memory (e.g., “What is 5 times 8?” “What is the capital of the United States?”) and procedural memory (e.g., whether she can use a ballpoint pen). Compared to later films featuring increasingly explosive “Conan science” moments—like stopping a high-speed train with an inflatable soccer ball or triggering an avalanche to block a dam collapse—this installment truly stands the test of time.
The forefather of car-accident-amnesia melodrama, with tear-jerking power off the charts. Bae Yong-joon’s character is involved in a car crash on his way to meet the heroine and awakens with complete memory loss. After yet another accident, his memory conveniently returns. The amnesia here mainly serves to alter the protagonist’s identity and generate dramatic conflict. As for scientific plausibility—best not to think too hard. Just enjoy the visuals and let yourself cry. Similar K-dramas include Stairway to Heaven, The Last Dance, and The Moon Embracing the Sun—there’s no shortage of amnesia-themed works.
The original domineering-CEO idol drama. Ming Dao plays the heir to a major corporation who loses his memory, adopts a new name, becomes a simple villager, and eventually falls in love with the spirited female lead. Same formula, same flavor—some viewers have even accused it of heavily borrowing from The Last Dance. Rewatching these childhood idol dramas as an adult may leave you questioning your younger self’s intelligence and taste. Recommended for those feeling nostalgic for their youth (tongue firmly in cheek).
Memory Reboot: Constantly Meeting a Brand-New Self
Another form of amnesia frequently portrayed in film and television is the idea of repeated “memory reboots.”
In Christopher Nolan’s mind-bending film Memento, the protagonist suffers from a rare memory disorder and can only retain memories for about a dozen minutes at a time. To avenge his wife’s death, he must use every tool at his disposal to preserve information, gather clues, and piece together the truth. The investigative difficulty is exponentially higher than for an ordinary person, which greatly amplifies the film’s tension.
Beyond the “minute-based” memory reset, there is another variation in which the protagonist’s memory resets after each period of sleep—essentially on a “daily” basis. In this setup, every awakening is equivalent to a rebirth. The character must leave behind carefully prepared clues to inform their future self: Where am I now? What has already happened? What am I supposed to do next?
Organic Brain Damage: “The Eternal Present Tense”
In the real world, the first scenario—frequent, short-term memory loss—is more consistent with scientific reality.
One of the most famous textbook cases in neuroscience is H.M., whose real name was Henry Molaison. He became a research model for constant memory reboot. At age 27, in order to treat severe epilepsy, surgeons removed parts of his bilateral medial temporal lobes, including the hippocampus and surrounding structures such as the amygdala. After the surgery, his intelligence and personality remained intact, but his memory suffered severe, irreversible, and permanent damage.
Image source: A Comic Guide to the Brain
First, Henry could no longer form long-term memories of anything that happened after the surgery. In other words, his episodic and semantic memory systems were devastated—this is a classic case of anterograde amnesia.
For example, he could engage in simple back-and-forth conversations with visitors. As long as the interaction was not interrupted, it could continue smoothly. However, if the visitor stepped out of the room and returned, or if the conversation was interrupted by someone else, Henry would completely forget everything that had just been discussed. This phenomenon is closely tied to the principles of short-term memory (working memory) discussed earlier.
With his hippocampus removed, Henry could no longer consolidate new long-term memories. He had to rely on short-term memory and sustained attention, passing information along in his mind like a relay baton to maintain continuity in what he was doing or saying. But this state was extremely fragile. The moment his attention was disrupted—by glancing out the window, someone interjecting, or even taking a bathroom break—the brain’s “cache” would instantly clear. The previous conversation, even the identity of the person in front of him, would vanish. Even researchers who had followed him for years were greeted each time as strangers, with Henry politely asking, “Hello, have we met before?”
Image source: A Comic Guide to the Brain
For Henry, time was permanently frozen in the present. There was no yesterday and no tomorrow. In his internal clock of self-awareness, he was forever 27 years old. If Henry looked into a mirror and saw an elderly man with white hair, he would feel confused—but the moment he turned away, he would forget the startling discovery and return to calm.
Second, Henry not only failed to form new memories but also lost some memories from before the surgery, exhibiting a degree of retrograde amnesia. Follow-up studies found that he had almost no recollection of events from the ten years prior to his operation, yet he vividly remembered his early childhood, including events before the age of ten.
To explain this pattern, researchers concluded that long-term memory is not stored in a single step. It requires a prolonged consolidation process, possibly spanning years or even a decade. In the first few years, although memories leave traces in the cerebral cortex, they still require assistance from the hippocampus for retrieval. Once the hippocampus is damaged, these not-yet-fully-consolidated memories lose their “index” and become inaccessible. In contrast, more remote memories are securely stored in the cortex, which is why, even without a hippocampus, Henry could still remember fundamental facts about himself—such as who he was and who his parents were.
Henry’s mirror-drawing experiment
Perhaps most astonishing to scientists was that although Henry could not remember new knowledge, he could still learn new skills. In one famous experiment, researchers asked him to perform a mirror-drawing task: tracing the outline of a double-bordered star while looking only at its reflection in a mirror. This is far from easy. At first, Henry’s drawings were clumsy and inaccurate, but with repeated practice, his performance improved dramatically. Interestingly, even though he eventually became highly proficient, Henry had no recollection of ever practicing. He would express surprise, saying, “That’s strange—this was easier than I expected.”
It was through studies of Henry that scientists came to understand that the brain contains multiple memory systems, each with distinct functions. Although the hippocampus—responsible for semantic and episodic memory—was permanently offline, the basal ganglia and cerebellum, which support skill learning, remained functional. This latter system is what we now call procedural memory, commonly known as muscle memory.
In 1986, Henry Molaison at age 60 at MIT
There is a small consolation: although Henry lived his entire life with memory impairment, he truly became someone who lived “in the present.” He led a peaceful life. This was partly due to his innate temperament, and partly because the surgery also removed his amygdala, eliminating much of the physiological basis for anxiety and fear. In daily life, he helped with household chores, prepared lunch, made the bed, accompanied family members on errands, played crossword puzzles, chatted with others, and watched television to relax. Although his consciousness remained anchored at age 27, his body continued to learn how to care for itself over time. Living in each brand-new present moment, he reached the age of 82.
Thus, in real cases of organic brain injury, forgetting is generally not measured in days—it happens continuously, moment by moment.
Besides Henry Molaison, another well-known amnesia patient was Clive Wearing, a musician from Cambridge University. Due to irreversible hippocampal damage caused by herpes encephalitis, his memory span was reduced to just 7 to 30 seconds—literally a “goldfish memory.” For Clive, unable to remember what had happened mere seconds earlier, every moment felt like waking from unconsciousness into a bewildering void. His wife wrote in her memoir:
Clive is in a constant state of feeling as if he has just regained consciousness. He once said, “I haven’t heard, seen, touched, or smelled anything. It’s like being dead.” His diary entries are heartbreaking to read:
2:00 p.m.: I have just woken up for the first time.
2:14 p.m.: I am now conscious.
2:19 p.m.: I have just woken up.
…
Rare Exception: “A Life That Resets Every Day”
Compared with these real-life cases, the film and TV trope in which a person is fully lucid during the day and only loses all memory after sleeping is, scientifically speaking, very difficult to justify.
If the hippocampus were severely damaged, a person would not be able to rely on short-term memory alone to sustain an entire day’s worth of experiences. A more plausible hypothesis is that during wakefulness, these memories have already entered the hippocampus and undergone initial encoding, with the problem arising later during the stabilization process. This in itself suggests that the hippocampus is functioning—at least participating in memory formation. If that is the case, then why would already-formed memories fail to stabilize after sleep, or become inaccessible upon waking? That is the truly puzzling part.
Michelle Philpots and her husband
That said, screenwriters are not creating these stories entirely out of thin air. There are indeed rare real-world cases that resemble this setup. One example is Michelle Philpots, a British woman who suffered two car accidents, resulting in brain injury accompanied by epileptic seizures. Her memory abilities deteriorated significantly afterward. Her long-term memory remains anchored in the 1990s, while newly formed memories last only briefly. She relies on sticky notes, video recordings, and reminders from her husband to manage daily life. The media often describe her condition as “a life that resets every day,” but from a medical perspective, this is not a literal daily memory wipe. Rather, it reflects difficulty in consolidating new memories over the long term. The romantic film 50 First Dates, with its premise of “waking up with a blank slate every day,” may well have drawn inspiration from this real case.
It is important to emphasize that such cases are extraordinarily rare, and even within neuroscience, they remain controversial. One relatively plausible explanation involves a failure of memory consolidation. Research suggests that the transfer of memories from the unstable hippocampus to the more stable cerebral cortex primarily occurs during sleep. If a patient experiences severe nocturnal epilepsy, abnormal brain activity may disrupt this consolidation process, preventing memories from becoming permanently stored or making them difficult to retrieve upon waking.
So while this type of storyline has a real-world prototype, it is largely an artistic exaggeration designed to heighten dramatic tension. It should not be taken as general medical knowledge.
Christopher Nolan’s breakthrough masterpiece and a definitive mind-bending classic. The protagonist suffers anterograde amnesia after a brain injury, with memories lasting only a dozen minutes. To avenge his wife, he relies on tattoos, photographs, and handwritten notes to piece together clues. Beyond its intricate plot, Nolan’s editing masterfully simulates the fragmented experience of an amnesia patient, creating an immersive viewing experience. High on both scientific plausibility and cinematic ingenuity, it’s worth watching multiple times.
Similar film: Ghajini, a Bollywood remake that retains the same amnesia premise but adds more action and musical elements. Starring national treasure Aamir Khan—whose physique alone is worth noting—it’s an entertaining adaptation.
A classic romantic comedy and perhaps the most famous “sleep-reset” amnesia story. The heroine suffers memory damage after a car accident and wakes up each morning with her memory reset to the day of the accident. The science is shaky, but the emotional impact carries the film. The hero’s daily efforts to make her fall in love with him all over again are undeniably romantic. As a love story, it’s well worth a watch.
Similar film: the Japanese remake 50 First Kisses, starring Takayuki Yamada and Masami Nagasawa. The plot closely follows the original, with few notable innovations. In my personal view, the on-screen chemistry is minimal—recommended mainly for fans of the actors.
A well-received recent Japanese medical drama, with each episode centered on a neurosurgical case. The protagonist, played by Hana Sugisaki, is a neurosurgeon who develops a memory disorder after an accident. Her memory resets each morning, and she must rely on rereading her diary to reconnect the threads of her life. Unlike the purely romantic focus of 50 First Dates, this series uses the “daily reset” premise to explore classic neurological disorders and the psychological struggles of amnesia patients. In that respect, it does a commendable job. Personally, the most moving aspect is the romantic subplot. The lead actors (Hana Sugisaki and Ryuya Wakaba) are a real-life couple, and their genuine chemistry shines through on screen.
A manga-adapted mystery comedy starring Yui Aragaki and Masaki Okada. The heroine, Kyoko, is a white-haired detective whose memory resets whenever she falls asleep. As a result, every case must be solved within a single day, and she writes crucial clues on her arms and legs with a marker. Masaki Okada plays an unlucky man who always finds himself near crimes—and frequently suspected as the culprit. The amnesia premise here exists purely for narrative convenience and has no scientific grounding whatsoever. Fortunately, the show is lighthearted and humorous, with straightforward mysteries that require little mental strain. A casual watch for relaxation, though personally I’d rate it as average.
Alzheimer’s Disease: The Eraser in the Mind
If the first two types of amnesia are relatively uncommon in real life and largely shaped by artistic dramatization, memory loss caused by Alzheimer’s disease is, for ordinary people, the most common—and the most closely intertwined with everyday life. Just two years ago, my grandmother passed away. In the decade before her death, she had likely been living with Alzheimer’s, though at the time I did not recognize it. I only felt that she was becoming increasingly forgetful, gradually failing to recognize people, and that her temperament had changed dramatically—growing unusually stubborn. Looking back now, I still feel deep guilt and regret.
For the sake of dramatic conflict and aesthetic appeal, films often focus on the rare cases of early-onset Alzheimer’s. A classic example is the Korean romantic film A Moment to Remember, in which Son Ye-jin’s character begins, in the prime of her youth, to slowly and smilingly forget the person she loves most, rendering the disease tragic yet strangely romantic.
In reality, however, Alzheimer’s disease is typically a condition of old age. The vast majority of patients develop it after the age of 65. Moreover, the prevalence increases exponentially with age. Among people over 85, approximately one-quarter to one-third are affected. For a scientific overview of the disease, one might watch the TED talk by Harvard neuroscientist Dr. Lisa Genova.
From a symptomatic perspective, Alzheimer’s has an extremely long course. From initial mild forgetfulness to eventual profound dementia and loss of function, patients and their families often endure a decade or more of slow decline. For most families, daily life is marked less by romance than by exhaustion and despair. As the disease progresses, cognitive confusion, incontinence, personality changes, wandering, hallucinations, and even aggression may follow. Even when family members understand the condition and prepare themselves intellectually and emotionally, the relentless burden of caregiving can gradually erode patience and resilience.
For further reading, popular science books such as The Thief of the Mind are recommended, along with memoir-style works like When My Mother Had Alzheimer’s, On Call: A Harvard Physician and His Wife’s Ten-Year Journey with Alzheimer’s, and Being My Mother’s Mother. As a broader social issue, caregiving can also be explored through NHK documentary series such as Missing with Dementia and Bankrupt in Old Age.
Disease Mechanism
Why does Alzheimer’s disease not only cause memory loss but also such profound changes in personality and function? How does it differ from ordinary age-related forgetfulness? The answer lies in its underlying pathology.
Normal age-related forgetfulness is primarily due to a natural decline in brain function. Neuronal responses slow, and some neurons die with age, but the overall structure remains intact. Core memories and personality are largely preserved. In contrast, Alzheimer’s involves irreversible physical changes within the brain. Under a microscope, two abnormal protein pathologies are visible: beta-amyloid plaques, which accumulate like toxic debris between neurons and disrupt communication, and tau protein tangles, which damage the internal transport systems of neurons, leading to widespread cell death.
Brain Structure of Alzheimer’s Patients (Image source: The Illustrated Brain User’s Manual)
The structural damage caused by these pathologies follows a clear pathway and sequence, which directly determines the symptoms observed at different stages of the disease.
Stage One: Hippocampal Damage and Short-Term Memory Impairment (Typically 2–4 Years) The hippocampus is the first region attacked. Once it is damaged, patients develop classic anterograde amnesia, unable to retain new information. This is the earliest and most common symptom of Alzheimer’s. Patients may repeatedly ask the same question, forget they have just eaten, or become lost in a familiar neighborhood. At this stage, long-term memories stored in the cortex remain intact, so patients may vividly recall distant events while failing to remember what happened moments ago.
Stage Two: Spread to the Cerebral Cortex, Affecting Long-Term Memory and Cognition (Typically 2–10 Years) As the disease advances, toxic protein accumulation spreads outward from the hippocampus to the cerebral cortex. The cortex not only stores long-term memories but also governs language, logic, spatial reasoning, and emotional regulation. During this longest stage of the illness, patients begin to exhibit retrograde amnesia, gradually forgetting events from recent years or even decades past, eventually failing to recognize loved ones. Personality and behavior may change dramatically. Language abilities may deteriorate; words may no longer come easily. Some individuals become stubborn, suspicious, or aggressive. When families observe striking personality changes and behavioral disturbances, they are often witnessing this stage.
Stage Three: Widespread Brain Involvement and Total Functional Decline (Typically 1–3 Years) In the final stage, pathological changes sweep through nearly the entire brain. Severe brain atrophy becomes evident. Patients may completely forget who they are. The brain progressively loses control over bodily functions. Walking becomes impossible; eventually even basic physiological abilities such as chewing, swallowing, and bladder and bowel control are lost. At this point, patients depend entirely on round-the-clock care, until complications ultimately lead to the end of life.
Regrettably, humanity has yet to develop a treatment capable of curing or reversing Alzheimer’s disease. Even when scientists succeed in clearing the toxic protein plaques from patients’ brains, the disease often continues its irreversible progression. This remains one of the most formidable challenges in modern medicine.
A textbook example of an Alzheimer’s-themed film. The original novel was written by Lisa Genova, a Harvard-trained neuroscientist who has authored multiple works of fiction and popular science centered on neurological disorders. The story follows a linguistics professor with a happy marriage and three children who develops familial Alzheimer’s disease at age 50. The film captures, with painful realism, the fear and helplessness of an intellectual confronting the gradual loss of self-awareness. Julianne Moore delivers an exceptional performance. Highly recommended.
Previously recommended in our “What to Watch This Week” column. Adapted from the stage play of the same name, this is widely regarded as a masterpiece among Alzheimer’s films, with remarkable artistic depth. Told largely from a first-person perspective, it immerses viewers in the daily confusion and fragmented memories of a father living with Alzheimer’s. The sense of immersion is powerful. Many viewers may feel disoriented while watching—but as one comment aptly put it: how fortunate are those who cannot fully understand this story. Anthony Hopkins (best known to many as “Hannibal”) delivers a godlike performance, and the portrayal of the daughter’s despair and inner conflict is equally compelling. Strongly recommended.
Another film on a similar theme, Familiar Touch, is also worth watching; you can refer to the detailed recommendation in our weekly column.
A Spanish animated film—realistic and deeply sobering. The protagonists are elderly residents confined to a nursing home. The film delicately portrays different stages of the disease’s progression: early disorientation, mid-stage hallucinations, and the late-stage despair of being “moved upstairs,” a euphemism for complete loss of function. I personally enjoy animated films, and this one resonates with me in its storytelling, emotional depth, and visual style. Strongly recommended.
A classic Korean tearjerker. The film focuses on early-onset Alzheimer’s, which accounts for less than 10% of real-life cases, primarily for dramatic purposes. As a romance, it is undeniably moving; as a depiction of disease, it is highly stylized and romanticized, and remains somewhat removed from the harsher realities of Alzheimer’s.
A romance drama written by Shizuka Oishi, starring Erika Toda and Tsuyoshi Muro, whose on-screen chemistry is unexpectedly strong. Once again featuring an early-onset premise—a young, capable female doctor developing Alzheimer’s—the series offers a relatively detailed portrayal of the male lead’s pain and perseverance as he cares for his gradually declining partner. The first half, focused on romance, is especially engaging. The latter half, after the illness progresses, becomes somewhat melodramatic, and some side romances feel less convincing. Still, it is undeniably tear-inducing.
Final Thoughts
While gathering material for this article, I discovered that beyond the types of amnesia discussed above, there are many other unusual and complex memory disorders, each with different triggers. It would be impossible to cover them all here. Drawing from the book The Neuroscience of Memory, here is a summary for further reference:
Medication side effects (especially anticholinergics, antihistamines, typical antipsychotics, and benzodiazepines) Vitamin deficiencies (particularly deficiencies in vitamin B12, vitamin B1, and vitamin D) Hormonal imbalances (including thyroid disorders) Physical stress, infection, and illness (including Lyme disease, COVID-19, encephalitis, diabetes, major surgery and anesthesia, organ failure, and sleep deprivation) Neurological disorders (including brain tumors, epilepsy and seizures, multiple sclerosis, stroke, transient global amnesia, and traumatic brain injury) Psychiatric disorders (including anxiety, depression, ADHD, bipolar disorder, schizophrenia, electroconvulsive therapy, dissociative amnesia, and alcohol or toxin use) …
This article is the second installment in our “Memory” learning series, using film and television as an entry point to explore the science of memory. In the next piece, I plan to return to the core theme of learning and discuss the importance of procedural memory and practice. If there are topics you’re particularly interested in, feel free to leave a comment and share your suggestions.
Combining episodic memory (specific past events) and semantic memory (abstract personal knowledge), it involves the cognitive reconstruction and narration of one’s personal life history. For example: Where is my hometown? Am I married? How did I leave my hometown back then? ↩︎