4-5 #Qingming : TSMC has resumed work at its fabrication sites in Taiwan after an earthquake-led shutdown; Infinix is allegedly developing its first foldable smartphone; Google is allegedly considering charging for AI-powered search; etc.

Taiwan Semiconductor Manufacturing Company (TSMC) has resumed work at its fabrication sites in Taiwan after an earthquake-led shutdown. The company was able to recover quickly, with no damage to critical chipmaking equipment, ensuring minimal disruption to chip supply. Despite initial concerns, analysts believe the impact on TSMC’s operations should be manageable, with limited downside. TSMC’s advanced technological measures and revised building codes helped mitigate damage and casualties, which showcased Taiwan’s resilience to seismic events. The earthquake, with a magnitude of 7.4, struck Taiwan on 3 Apr 2024, causing some disruptions to chip production. (Neowin, MacRumors, Unwire.hk, Yahoo)

Intel’s next-generation 20A (2nm-class) and 18A (1.8nm-class) manufacturing technologies are crucial for Intel’s success and profitability in the mid-term future. According to Intel CEO Pat Gelsinger, the company will be ramping the first products Panther Lake and Clearwater Forest, the first client product and the first server product based on 18A in 2025. They will start to ramp to higher volumes in 2026 for 18A. For Intel, 18A is more than just two products as this technology will also be used by a number of its foundry partners. (CN Beta, Tom’s Hardware, My Drivers)

Microsoft and Quantinuum have announced a major breakthrough in quantum error correction. Using Quantinuum’s ion-trap hardware and Microsoft’s new qubit-virtualization system, the team was able to run more than 14,000 experiments without a single error. This new system also allowed the team to check the logical qubits and correct any errors it encountered without destroying the logical qubits. This, the two companies say, has now moved the state-of-the-art of quantum computing out of what has typically been dubbed the era of noisy intermediate scale quantum (NISQ) computers. (CN Beta, Microsoft, Reuters, TechCrunch, Quantinuum)

Samsung’s foldables in 2024 could include the Z Fold 6 / Fold 6 FE, Flip 6, and Z Fold 6 Ultra. The Samsung Galaxy Z Fold 6 / Fold 6 SE could be an entry-level Fold and this will launch around Sept / Oct 2024, which is 2-3 months after the flagship models. The company is targeting 200K-300K entry-level Galaxy Z Fold 6 / FE shipments. Samsung is preparing a new smartphone with the model number SM-F958. Everything up to the trailing 8 points to the Galaxy Z Fold 6 handset, and the trailing 8 in other models is used to represent the top-end smartphones with the “Ultra” tag. The certification process has revealed the potential of a cheaper “Fan Edition” model, specifically the Galaxy Z Fold 6 FE. (Android Authority, GizChina, GalaxyClub, 91Mobiles)

The patent of Huawei’s smartphone with a tri-fold display has emerged. It is rumored that Huawei is accelerating the development of a new “tri-fold display phone” and has begun large-scale stocking. It is expected to be released as early as 2Q24. Huawei is smoothly advancing the research and development of tri-fold phones, and has begun testing the maturity of related projects such as ultra-thin flexible glass cover (UTG) a few months ago. As the acceptance of tri-fold phones continues to increase, if calculated based on a penetration rate of 40-60%, annual sales are expected to reach 2.8M-4.2M units. (CN Beta, My Drivers, OfWeek, Gizmo China, Huawei Central)

Apple’s patent titled “Displays With Selective Pixel Brightness Tuning” describes future Macs and external Apple monitors could offer spatial computing features via displays that could include lenticular lenses that enable the display to provide 3D content to the viewer. The lenticular lenses may be formed over an array of pixels such as organic light-emitting diode pixels or liquid crystal display pixels. A lenticular display could show and hide subtly different parts of an image, depending the user’s angle of view. The effect is that as a user moves their head, or alters their gaze, they see parallax-like changes and so the image appears 3D. (Apple Insider, Apple Today, USPTO)

Apple reportedly wants to use a Border Reduction Structure (BRS) on the bottom bezel of the iPhone 16 OLED display. To implement a thin bezel, the circuit under the bezel must be placed more tightly and some wiring must be bent down, which increases the technical difficulty. It is known that no company has stabilized the iPhone OLED production rate to the level that Apple expects. In iPhone 16 series, Samsung Display is expected to deliver all four OLEDs. LG Display is also expected to aim for OLED supply in two top lineups such as Pro and Pro Max, and BOE is expected to supply OLED in two lower lineups such as general and plus models. (CN Beta, Apple Insider, 9to5Mac, The Elec)

Infinix is allegedly developing its first foldable smartphone, “Infinix Zero Flip”. An Infinix model had passed EEC certification, and the model number was X6962. Zero Flip will have the model number X6962. This model is Infinix’s first foldable smartphone. It is expected to be affordable. (Gizmo China, Android Headlines)

Samsung will allegedly launch both a Galaxy Z Fold FE and a Galaxy Z Flip FE. The Galaxy Z Fold FE, when folded, will measure 155.1 x 67.1 x 14.2mm, while the Z Flip FE, when unfolded, will measure 165.2 x 71.9 x 6.9mm. The Flip FE will be powered by the Snapdragon 7s Gen 2, while the Fold FE will get a mix of Snapdragon and Exynos chipsets depending on market, with the latter claimed to be the yet-unreleased Exynos 2300. (GizChina, Gizmo China, My Drivers, GSM Arena)

OnePlus is introducing AI Eraser, which is an image editing feature powered by advanced artificial intelligence (AI) and proprietary algorithms. AI Eraser simplifies the process of removing unwanted elements from pictures stored in the OnePlus Gallery app. Users can simply select objects like pedestrians, distracting objects, or even minor imperfections. The core of AI Eraser lies in a powerful, proprietary large language model (LLM) trained on massive datasets encompassing diverse scenes. This training empowers the AI to not only remove unwanted objects but also intelligently replace them with contextually appropriate elements. The company says this will roll out to the OnePlus 12, OnePlus 12R, OnePlus Open, OnePlus 11, and OnePlus Nord CE4, in all regions where these are officially available for purchase.  (GizChina, GSM Arena, Fonearena)

Samsung Electronics plans to enter the new 3D DRAM segment. Samsung has come up with its very own 3D DRAM implementation, which is expected to come into effect within 2025. The DRAM industry is moving into sub-10nm compress lines. To break the deadlock present in innovating modern DRAM technologies, Samsung plans to introduce two new methods, called Vertical Channel Transistors and Stacked DRAM, with both of them involving differences in the positioning of components, which ultimately reduces the device area occupation, in turn ensuring higher performance. To increase memory capacities, Samsung plans on utilizing the stacked DRAM concept, which allows the firm to achieve a higher storage-to-area ratio, thus increasing chip capacities to potentially 100 GB moving into the future. With that said, it is predicted that 3D DRAM markets could grow up to USD100B by 2028. (CN Beta, WCCFTech)

Samsung has recently announced plans to start using 3D DRAM in its upcoming memory products. This move will involve a technology called vertical channel transistors (VCT) and will begin with their cutting-edge, first-generation technology that’s smaller than 10nm. These types of transistors could make memory cells much more efficient. One of the reasons is the use of a 4F² cell design. Tokyo Electron, a big name in the business of making semiconductor manufacturing equipment, believes that using VCTs and this efficient cell design in making DRAM will start happening between 2027 and 2028. Moving to this new kind of DRAM will probably also mean finding new materials to make capacitors and bitlines that work well with VCT-based DRAM. (CN Beta, Guru3D, ZDNet, Tom’s Hardware, Semi Engineering, SamMobile)

Google is allegedly considering charging for artificial intelligence (AI)-powered search in what would be the biggest shake-up of its business model since its founding. Google’s parent company Alphabet is looking at putting AI search into its premium subscription services, which already offers access to its new Gemini AI assistant in Gmail and Docs. Under the plans being considered, Google’s search engine would remain free. In the meantime, the company has begun testing AI-generated “overviews” for search queries in the UK. (Android Authority, Financial Times, Yahoo)

Foxconn has made major AI investments in Mexico. Foxconn has funneled USD690M into the country in the past 4 years. Just in Feb 2024, Foxconn picked up a slice of land in the western state of Jalisco for USD27M to facilitate a major expansion of its AI server production. Nvidia, Amazon, Google, and Microsoft are using Foxconn’s facilities in Mexico to help meet their AI server needs. (CN Beta, WSJ, Quartz, Techspot)

Huawei revealed in its annual report that it has fulfilled its contractual rights and obligations associated with the supplementary agreement for the sale of the Honor business. Huawei signed this supplementary agreement with Honor’s buyer in 2023. Huawei has sold off its Honor business and server business-related subsidiaries.  (CN Beta, Gizmo China)

ŌURA, maker of Oura Ring has announced the launch of a new in-app platform, Oura Labs. The latest addition to the Oura Membership experience is designed to accelerate innovation by bringing members into the research and development process. Symptom Radar is the first feature coming to Oura Labs for iOS users. Oura users will be able to report back to the company’s team of researchers to let them know how the experimental features are working, making them part of the development process. (TechCrunch, Tom’s Guide, Business Wire, Oura)

Tesla will reportedly send a team to scout locations in India for a proposed USD2B-3B electric car plant. The step towards making vehicles in India comes after New Delhi lowered tariffs on higher-priced imported EVs for companies that commit to making them in the country within 3 years. The tariff cut was a concession Tesla had been pushing for as a precondition for investing. Tesla has told Indian officials it is considering building a smaller car in the proposed new factory than its current models, which would be priced at less than USD30,000. It could then sell the model in India and export to south-east Asia, the Gulf, Africa, and southern and eastern Europe. (CN Beta, CNBC, FT, Economic Times)

Apple has allegedly begun doing some early research into such a home robot. This early research is currently being handled by Apple’s hardware and AI teams. Since this is still very early in development, it’s entirely possible the company could decide to cancel this research at any time. Apple has been exploring a mobile robot that can follow users around their homes. Apple has also developed an advanced table-top home device that uses robotics to move a display around. (TechCrunch, Bloomberg, Business Times, Yahoo)

Won-joon Choi, executive vice president of Samsung’s mobile business has hinted that the company could add generative artificial intelligence (AI) technology to its voice assistant Bixby. Samsung introduced Bixby back in 2017 with the launch of the Galaxy S8 series. Initially introduced as a replacement for the S Voice assistant, Bixby made its debut with features like Bixby Voice, Bixby Vision, and Bixby Home, offering users a comprehensive suite of voice commands, visual search capabilities, and personalized recommendations. (GizChina, CNBC, Android Authority, GSM Arena)

Apple has published a paper detailing Reference Resolution As Language Modeling (ReALM). Apple has developed a new artificial intelligence (AI) system that can understand ambiguous references to on-screen entities as well as conversational and background context, enabling more natural interactions with voice assistants. ReALM leverages large language models to convert the complex task of reference resolution — including understanding references to visual elements on a screen — into a pure language modeling problem. This allows ReALM to achieve substantial performance gains compared to existing methods. The work highlights the potential for focused language models to handle tasks like reference resolution in production systems where using massive end-to-end models is infeasible due to latency or compute constraints. (Apple Insider, ARVIX, VentureBeat)

Microsoft has designed a new image search tool that can identify individuals within a cloud database. Apart from features such as Windows Hello in Windows 11 for keyless sign-in and Azure AI face for detecting and analyzing human faces on images, Microsoft has recently filed a patent that indicates other uses of this technology. The patent also states that it will group the photos by people, not only identifying them but also cataloging them. This tool will enable grouped images if they feature the same person or group of people. Additionally, WindowsReport explained how the image search tool may rank these images using additional data such as connections with family and friends to improve its precision and efficiency. (Android Headlines, TweakTown, USPTO, Windows Report)

Yahoo has announced that it has acquired Artifact, the AI-driven news aggregation and discovery platform created by Instagram cofounders, Kevin Systrom and Mike Krieger. Artifact’s AI-powered discovery engine surfaces content users most want to see and becomes more attuned to their interests over time. The result is a personalized feed of news stories that’s extremely effective at helping users discover stories they want to read. In addition to curation, Artifact brings a host of features aimed at improving the quality of the experience, from easy-access to top news, to link-sharing capabilities, to tools to prevent clickbait proliferation, and more. (CN Beta, The Verge, Yahoo, Reuters)

Leave a Reply

Your email address will not be published. Required fields are marked *