Key Highlights:

• Jensen Huang warns that China is catching up fast in AI and could “win the race.”
  
• He criticizes U.S. restrictions and overregulation for slowing innovation.
  
• Huang clarifies he wants America to lead by accelerating, not isolating.

Huang’s Warning at the London AI Summit

Nvidia CEO Jensen Huang has warned that China is moving rapidly toward AI dominance and could “win the AI race” if the U.S. continues to slow its innovation through overregulation and export restrictions. Speaking at the Financial Times Future of AI Summit in London, Huang said China is “nanoseconds behind America” in AI progress, highlighting how quickly the global gap is closing.

Huang urged Western countries to speed up AI development rather than hold it back. “It’s vital that America wins by racing ahead and winning developers worldwide,” he later wrote on X (formerly Twitter).

China’s Rapid AI Growth

According to Huang, China’s government-backed energy subsidies and strong state support are helping local companies advance faster. Beijing’s investments in semiconductor manufacturing and AI research are creating an ecosystem that rivals the U.S. in both scale and ambition.

He noted that while China faces technological hurdles, its developer community and policy environment are highly motivated and unified under national goals for AI self-reliance.

U.S. Restrictions and Policy Concerns

The U.S. government recently reaffirmed its ban on Nvidia’s advanced Blackwell AI chips being sold to China, citing national security reasons. Huang criticized these restrictions, arguing they hurt U.S. competitiveness and inadvertently push China to accelerate its own chip development.

He also pointed to state-level AI regulations and high energy costs as factors slowing innovation in the U.S. “Cynicism and overregulation are making it harder for American companies to move fast,” he said.

Market Reaction and Clarification

After his remarks were reported, Nvidia’s stock briefly dipped, prompting Huang to clarify that he wasn’t predicting U.S. defeat. Instead, he emphasized that the U.S. must act faster to maintain its lead. Nvidia remains the world’s most valuable company, valued at around US$4.7 trillion, even after a recent dip from its US$5 trillion milestone.

Balancing Global AI Competition

Huang’s comments come amid rising U.S.–China tensions over technology and trade. While Washington focuses on limiting China’s access to cutting-edge AI hardware, Beijing is boosting domestic innovation.

Huang’s message was clear: the U.S. should lead through innovation and collaboration, not isolation. “We want America to win,” he said, “but winning means building faster, not building walls.”

Read More:

• US AI Tools Under Fire: Are America’s Top Coders Using Chinese Models?
• AI Chip War Heats Up: U.S. Blocks Nvidia, China Responds with Total Ban
• China’s XPENG Shocks the World with Full-Stack Physical AI: Meet the IRON Robot and Flying Car

(via)

The post China Will Win the AI Race: Nvidia CEO Jensen Huang’s Warning to the West appeared first on Gizmochina.

Key Highlights:

• Apple is close to signing a US$1 billion per year deal with Google to use its Gemini AI model for Siri.
  
• The upgraded Siri is expected to launch in spring 2026 as part of iOS 26.4.
  
• Apple plans to use Gemini temporarily while developing its own large-scale AI model.

Apple Turns to Google for AI Upgrade

Apple is in final talks with Google to integrate its powerful Gemini AI model into Siri, marking one of the biggest artificial intelligence partnerships in the tech industry. The deal, worth around US$1 billion per year, aims to overhaul Siri’s capabilities and bring Apple up to speed in the fast-moving AI race. According to Bloomberg, the Gemini-powered Siri could launch by spring 2026 with iOS 26.4.

A Smarter, More Capable Siri

Currently, Siri runs on Apple’s in-house cloud AI with around 150 billion parameters. Under the new agreement, it will tap into Google’s Gemini model, which boasts 1.2 trillion parameters, nearly eight times more powerful. The upgrade will allow Siri to better understand complex commands, summarise information, and plan multi-step tasks.

Internally, the Siri revamp is known as “Linwood”, while the larger AI project is called “Glenwood.” The effort is being led by Apple executives Craig Federighi and Mike Rockwell, the creator of the Vision Pro headset.

Privacy and Performance

Despite using Google’s AI, Apple will maintain strict privacy controls. Gemini will operate on Apple’s Private Cloud Compute (PCC) servers, ensuring user data remains isolated from Google’s infrastructure. All processing will be temporary, with no data stored between sessions, reinforcing Apple’s privacy-first image.

However, this setup means Apple must expand its cloud hardware to handle Gemini’s massive size, potentially increasing costs and system load.

Why Apple Chose Google

Apple reportedly tested multiple AI models, including OpenAI’s GPT and Anthropic’s Claude, before selecting Gemini for its contextual understanding, scalability, and integration flexibility. While the deal gives Apple a major AI boost, it’s seen as a temporary arrangement until Apple completes its own 1-trillion-parameter AI model.

In markets like China, where Google services are banned, Apple plans to rely on its in-house models and local partners like Alibaba for regulatory compliance.

Strategic Implications

The collaboration marks a shift in Apple’s strategy, from building everything internally to selectively partnering with industry leaders. While the partnership won’t be publicly branded like the Safari-Google Search deal, it highlights how Apple is adapting to remain competitive in generative AI.

Both Apple and Google stocks rose after reports of the deal, signaling investor optimism. Once launched, the Gemini-powered Siri could redefine what voice assistants are capable of, blending power, privacy, and intelligence in a way that reshapes the AI landscape.

Read More:

• AI Chip War Heats Up: U.S. Blocks Nvidia, China Responds with Total Ban
• NVIDIA’s Market Share in China Drops from 95% to Zero Amid US Export Restrictions
• US AI Tools Under Fire: Are America’s Top Coders Using Chinese Models?

(via)

The post Apple Turns to Google for AI Help: $1 Billion Deal to Power Siri With Gemini appeared first on Gizmochina.

Creating Artificial Intelligence like ChatGPT is not a cheap investment, with estimates suggesting hundreds of millions to billions of dollars being sunk into AI for building and training it. However, money doesn’t seem like an object to Amazon which might be planning on spending a ludicrous amount of money on AI this year.

Amazon to invest $100 billion in AI this year

DeepSeek recently took the world by storm with an alleged budget cost of development, but Amazon is going big on AI and is spending top dollar. According to a TechRadar report, the American tech giant aims to spend $100 billion on AI this year. In comparison, DeepSeek claimed that it only took a few million dollars to train its R1 model, which may or may not be true.

However, AI is still a hefty business as companies like Google and Microsoft are sinking tens of billions of dollars in their respective projects. The big names in tech have already been readily adopting AI in recent years, with Amazon being an outlier. So it seems quite surprising that the company is becoming one of the highest spenders in artificial intelligence. One could argue that it was only a matter of time.

During its latest earnings call, Andy Jassy, the CEO of Amazon hinted that the brand could be planning on spending more than $100 billion on AI in 2025. So Alexa might be getting a new evolution soon. Unfortunately, this is all the information we have at the moment, so stick around for more. In related news, India is also working on its own AI model to compete with the likes of ChatGPT and DeepSeek. The latter is also seeing some trouble in the global market as some governments see it as a Chinese threat.

For more daily updates, please visit our News Section.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

The post Amazon going big on AI in 2025 with plans to spend an insane $100 billion appeared first on Gizmochina.

DeepSeek, the brand-new disruptive AI from a Chinese startup, is showing no signs of stopping. We recently talked about DeepSeek AI as it challenges OpenAI’s dominance with ChatGPT. Now, in just a short time, the DeepSake AI assistant app has become the top free app on iPhones.

DeepSeek AI Assistant Top App on Apple’s App Store on iOS

As of right now, the Chinese AI assistant from DeepSeek is the top rated free app on Apple’s Apple Store in the US. It is currently the top app for iPhones, beating out rivals like OpenAI’s ChatGPT. For those unaware, DeepSeek is making waves for being an open-source platform, which is based on the DeepSeek V3 model. The artificial intelligence apparently requires far less computing power than its competitors as well.

While it is still disputed, the people behind DeepSeek claim that it was developed for under 6 million US Dollars. Despite its relatively cost efficient development, it promises to offer performance that can match Claude 3.5, GPT-4o and others. Similar to other AI platforms, DeepSeek also offers functions like ChatGPT, which can help content creators and be useful in research. It is currently available as an app, API, and on the web, making it easily accessible to all.

The first DeepSeek-R1 was rolled out under an MIT license, so it is free to be used commercially without any restrictions. DeepSeek’s open-source approach also challenges various closed-source models developed by major tech brands. This shift towards greater transparency and accessibility, allows a wider range of individuals and organizations to contribute to its development and benefit from its potential.

For more daily updates, please visit our News Section.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

The post China’s DeepSeek AI Assistant is now the top free app for iPhones appeared first on Gizmochina.

JPMorgan Chase has quietly introduced LLM Suite, an internal generative AI tool designed to assist employees, particularly those in asset and wealth management. This ChatGPT-like platform is capable of writing, generating ideas, and summarizing documents.

LLM Suite is believed to be one of the most extensive applications of large language models in finance

The tool has been rolled out to approximately 50,000 employees, representing about 15% of the bank’s workforce. While the company hasn’t disclosed specific details, LLM Suite is believed to be one of the most extensive applications of large language models in the financial industry.

This move follows similar trends in the banking sector, with Citigroup previously predicting that AI could automate up to 54% of banking jobs. As AI technology continues to advance, financial institutions are increasingly exploring ways to leverage its potential to increase efficiency and productivity.

(Via)

The post JPMorgan Chase Deploys In-House AI Tool to Boost Productivity appeared first on Gizmochina.

A new study by Hosting Advice sheds light on how artificial intelligence (AI) is impacting the lives of Gen Alpha children, the generation born between 2010 and 2025. The survey polled parents of children aged 7-14 to understand their experiences with their kids’ AI usage and its perceived influence on development.

Nearly half of all surveyed parents reported their Gen Alpha children already knee-deep in AI tools

The findings reveal that nearly half (49%) of surveyed parents report their Gen Alpha children already utilize AI tools. Interestingly, a gender gap appears in AI adoption, with 54% of parents stating their sons use these tools compared to 45% for girls.

As Gen Alpha children enter their teenage years (13-14), AI usage seems to rise significantly, with 60% of parents reporting such use. The survey also explores how AI factors into homework completion. While 20% of parents indicated their children use AI for homework assistance, 40% expressed opposition to this practice, citing concerns about potential cheating.

Despite these concerns, a majority (54%) of parents surveyed support their children’s use of AI tools. They perceive these tools as beneficial for personalized education (30%), social skills development (18%), fostering creativity (29%), and preparing children for the future (21%). Notably, the survey also identified a gender gap in parental support, with fathers surveyed displaying a higher level of support for AI tool usage (66%) compared to mothers (47%).

The post Gen Alpha and AI: How Artificial Intelligence is Affecting the Lives of Kids Born after 2010 appeared first on Gizmochina.

A new health company called Thrive AI Health is aiming to revolutionize wellness through artificial intelligence. Co-founded by Arianna Huffington and OpenAI CEO Sam Altman, Thrive AI leverages AI to create a personalized “health coach” tailored to your specific needs. This coach analyzes various aspects of your life, including sleep, diet, fitness,stress, and social interactions, to offer data-driven recommendations.

The health data will be accessible through a smartphone app or a virtual assistant

Leading the company is DeCarlos Love, who previously spearheaded the fitness and health experience at Fitbit. Thrive AI plans to train its AI coach using a combination of scientific research, medical data, and collaborations with institutions like Stanford University School of Medicine. Their vision is to create a comprehensive health data platform accessible through a smartphone app or a virtual assistant integrated with Thrive’s existing products. This system will learn your behavior patterns and offer real-time health advice.

Thrive AI positions itself as a more mindful approach to health management. They emphasize making personalized health coaching accessible to everyone while tackling health inequalities. Security and user privacy are a priority, with Dr. Gbenga Ogedegbe, an expert in health equity, serving as an advisor. Additionally, they pledge to base recommendations on peer-reviewed research and give users complete control over the information used to personalize their experience.

(Via)

The post OpenAI CEO Sam Altman’s Latest Venture is Thrive AI, a Personalized Wellness ‘Health Coach’ appeared first on Gizmochina.

AI is rapidly advancing, and now it’s possible to generate videos from stable images, a process known as image-to-video creation. With stable diffusion, AI can create videos out of still images. New AI models can create text-to-video based on your prompts. OpenAI’s Sora, for instance, can create highly realistic videos. Check out OpenAI’s Sora and computer-generated videos using the provided link.

While AI-powered video creation is a game-changer, many AI-based media generation tools currently limit the number of videos you can make. AI requires significant processing power, making running the AI models locally on your device sounds like a better idea.

According to a recent ITHome news article, MediaTek may introduce on-device AI video creation for its Dimensity 9300 and 8300 chipsets. The NPU in these chipsets will enable video generation from stable images. MediaTek was rumored to unveil video creation at MWC 2024, and now a collaboration between Kwai and MediaTek has been revealed. Kwai is a video-sharing platform and editor similar to TikTok and CapCut.

The post MediaTek to offer on-device video creation with AI on Dimensity 9300 & 8300 appeared first on Gizmochina.

Amazon is set to overhaul Alexa, but access to its most advanced features might come at a cost. Reports suggest a $5 to $10 monthly subscription fee on top of a Prime membership to unlock the AI-powered Alexa. This revamped assistant is expected for launch in August 2024 and boasts several improvements.

This new, advanced Alexa will also likely offer custom advice for say, shopping, or art.

Saying “Alexa” repeatedly might become a thing of the past. The new version is rumored to understand and complete multiple requests in one go, like writing an email and ordering takeout simultaneously. It could even learn user habits and personalize routines, like turning on the coffee pot when the alarm goes off. This advanced Alexa is also expected to offer sophisticated shopping advice.

However, Amazon isn’t abandoning the free tier entirely. A basic version of Alexa with new generative AI features will still be available. This move seems to be a response to the increasingly competitive AI assistant market, with Google, Microsoft, OpenAI, and even Apple‘s improved Siri vying for user attention. The future of Alexa appears to be a two-pronged approach: a free basic service and a premium tier with advanced capabilities.

(Via)

The post Amazon’s AI-powered Alexa will have a Monthly Subscription Fee – on Top of a Prime Membership appeared first on Gizmochina.

Getty Images and Picsart, a popular photo editing platform, have joined forces to create a novel AI image generator. This unique tool aims to address copyright concerns plaguing current AI image generation models.

The training data will comprise exclusively of licensed stock images

The key difference lies in the training data. Unlike other models that might use a mix of sources, the Picsart and Getty generator will be trained exclusively on licensed stock images from Getty Images. This ensures all the content used to create the AI model is authorized for commercial use. The companies are touting this approach as a “responsible, commercially-safe” alternative. Images generated by this model will come with full commercial rights, allowing users to freely incorporate them into their projects without copyright worries.

Similar to Adobe’s recent attempt with their Firefly AI model, the Picsart and Getty tool will be a paid service accessible only to Picsart’s subscribing users. This approach avoids the controversy surrounding Adobe’s plan to potentially use user-generated content for training their AI model.

Getty seems committed to ethical practices. They’ve assured users that photographers whose work contributes to the training data will be “compensated creators included in the dataset on an annual basis.”

While a specific launch date remains unknown, the Picsart and Getty Images AI image generator is expected to be released later this year. It will be accessible through Picsart’s API services, allowing integration with various creative workflows.

(Via)

The post Getty and Picsart Team Up for “Commercially Safe” AI Image Generation appeared first on Gizmochina.

At the 2024 Taipei International Computer Show, Kneron, an AI chip startup, unveiled its next-generation edge AI products: the KNEO 330 server and PCs equipped with their new KL830 NPU chip.

The company was founded in 2015, and it focuses on Edge Computing AI

Founded in 2015, Kneron focuses on providing edge computing AI (edge AI) technology. They boast big-name investors like Horizons Ventures, Qualcomm, and Alibaba. Kneron aims to offer an alternative to cloud-based AI solutions from tech giants. Their technology allows companies to build internal servers for specific AI applications, addressing data privacy and security concerns. This could make AI solutions more accessible to businesses hesitant to rely on external cloud services.

Two products were showcased:

• KNEO 330 server: This server offers 48TOPS of AI computing power, supports up to 8 users simultaneously, and works with applications like LLM and Stable Diffusion. Kneron claims it can achieve cloud-like performance for RAG (Generative Adversarial Networks) at a lower cost, potentially reducing AI expenses for small businesses by 30-40%.
• KL830 NPU-equipped PC: Specific details weren’t revealed, but a demonstration HP device showcased its capabilities. Kneron is partnering with major PC manufacturers to make them the first customers for this chip.

The KL830 is a neural processing unit (NPU) designed to work alongside CPUs and GPUs for AI tasks. It offers high computing power (10eTOPS@8bit) with low power consumption (2W) and handles various large language models efficiently.

Kneron also mentioned a USB AI accelerator stick powered by the KL830. This could transform any device, like routers, IoT cameras, or even traditional computers, into an edge AI device.

(Via)

The post Qualcomm & Foxconn Backed AI Chip Startup Kneron Unveiled a New NPU Chip and Server, Competing with Nvidia appeared first on Gizmochina.

In a recent interview with The Verge, Eric Yuan, founder of the popular video conferencing platform Zoom, discussed his vision for the future of video conferencing and a novel concept – digital twins for meetings.

Eric Yuan envisions Zoom going up against Microsoft and Google soon

Yuan, who previously worked at Cisco before launching Zoom with the goal of simplifying video conferencing, now aspires for Zoom to evolve beyond just a video conferencing platform. He envisions Zoom competing with tech giants like Microsoft and Google in the realm of enterprise software, offering a comprehensive suite that includes document collaboration, email services, and chat functionality alongside video conferencing.

The interview highlighted Zoom’s significant investment in Artificial Intelligence (AI) – a trend prevalent across various industries. Yuan believes AI holds immense potential for the workplace, particularly through the creation of “digital twins.” These AI-powered avatars could represent users in Zoom meetings, potentially even making decisions on their behalf while the user attends to more pressing matters.

However, Yuan emphasizes the importance of user control over these digital twins. He envisions a scenario where users can train them to exhibit different characteristics depending on the meeting context. For example, a user could have a digital twin tailored for sales presentations and another for technical discussions. While acknowledging the inherent uncertainties in AI development, Yuan remains confident in the eventual realization of this technology.

It’s worth noting that Zoom has already addressed user concerns around AI development in its terms of service.

(Via)

The post AI-powered Avatars will Soon Attend Meetings on our Behalf, Zoom Founder States appeared first on Gizmochina.

Apple’s upcoming iOS 18 is set for a major AI boost, with Mark Gurman’s reports suggesting a partnership with OpenAI to bring a range of generative AI features. This collaboration aligns with Apple’s WWDC 2024 announcement plans, which are also expected to unveil an improved Siri voice assistant.

Apple is playing it safe by also looking towards Google Gemini’s AI integration

While OpenAI appears to be the frontrunner, Apple seems to be hedging its bets. Negotiations are ongoing with Google to potentially integrate Gemini, their AI offering, as an additional option within iOS 18.

The rumored AI features in iOS 18 are quite exciting. Users can expect functionalities like automatic emoji generation and the ability to auto-generate replies for emails and text messages. With the new update, the iPhone series will be an even stronger contender in the smartphone marktet, offering features at par with the Pixel series by Google. These features promise to streamline communication and interaction for iPhone users.

RELATED:

• iOS 18 To Get Notification Summaries, AI Photo Editing As Per Latest Report
• Apple Prepares for AI: Siri Analysis, On-Device Processing in iOS 18?
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills

(Via)

The post Apple and OpenAI’s Deal may be Non-Exclusive, Apple may Integrate Google’s Gemini within iOS 18 appeared first on Gizmochina.

Gmail on Android might soon get a smarter upgrade with the rumored integration of Google‘s Gemini AI. This news comes from reliable sources who combed through the app’s code, uncovering a hidden Gemini button.

Gemini still might miss crucial information due to misinterpretations

Tapping this button could unleash a wave of helpful AI functionalities within Gmail. Imagine having Gemini summarize lengthy emails, highlighting key points, or even crafting draft replies for you. This feature aligns with functionalities already available to paying Google Workspace users.

While summarizing existing content is AI’s forte, it’s important to remember that AI isn’t flawless. There’s always a chance, however small, that Gemini might miss crucial information due to misinterpretations.

For truly mission-critical emails, a human touch remains irreplaceable. Unlike AI, we understand the context and nuances of our email exchanges, going beyond the analysis of word probabilities.

RELATED:

• Gmail for Android Finally Implements a Crucial Feature It Lacked Since Day One
• Gmail Now Lets Users Unsubscribe Unwanted Mails With A Single Tap On iOS
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• HyperOS vs MIUI: 10 Things You Must Know

(Via)

The post Gmail on Android will Get Google Gemini Integration Soon, App Code Reveals Hidden Button appeared first on Gizmochina.

AI and handheld gaming? Nvidia and MediaTek‘s rumored team-up has sparked excitement. A recent leak hints at a surprising collaboration between tech giants Nvidia and MediaTek. The rumor suggests they’re co-developing an ARM-based system-on-chip (SoC) specifically designed for artificial intelligence applications. But that’s not all! The whistleblower, XpeaGPU, has dropped another exciting tidbit: Nvidia might be eyeing the handheld gaming market with its own ARM-based SoC.

The upcoming Switch 2 may also sport an upgraded Tegra CPU

XpeaGPU reveals MediaTek is currently working on a next-generation gaming handheld SoC that will be infused with Nvidia’s graphics processing muscle. This news has sent shivers down the spines of Chinese manufacturers, with several already expressing interest in the chip. XpeaGPU clarifies that this SoC is a separate endeavor from other joint projects undertaken by the two companies.

Nvidia is no stranger to the handheld gaming arena. They previously ventured into the market with the Tegra SoC-powered Shield console. Interestingly, the ever-popular Nintendo Switch also relies on the same Nvidia chipset. Currently, AMD reigns supreme in the home console market, with their chips powering most handheld devices, including the Steam Deck.

The rumor mill churns out whispers of a potential Switch 2 sporting an upgraded Tegra SoC, boasting graphics performance that rivals the GeForce RTX 2050. However, the real game-changer would be Nvidia developing a brand new SoC tailored specifically for handheld gaming.

The most intriguing part about this rumor is Nvidia’s choice of partner. They’ve opted for MediaTek instead of the usual high-end SoC leader, Qualcomm. This unexpected alliance has gamers and tech enthusiasts eagerly awaiting further developments and Nvidia’s potential return to the handheld gaming scene.

RELATED:

• TSMC’s Advanced Packaging Capacity is fully booked for next two years by Nvidia and AMD
• Nvidia’s mobile RTX 5080 GPU may have the same GB203 GPU as desktop RTX 4090
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• 10 Best Free Android Games in 2024

(Via)

The post Nvidia and MediaTek are Teaming Up to Develop an ARM-based SoC for AI Apps, Rumor States appeared first on Gizmochina.

Meta CEO Mark Zuckerberg recently shared his thoughts about the growth and future of data centers that focus on the development of AI in an interview. According to him, the shortage of AI accelerator cards is in the process of being alleviated. He noted that it was a big issue over the last few years to get the GPUs that are key to the development of artificial intelligence models due to supply chain shortage. Now it is in the process of being alleviated.

The investment in data centers is still growing exponentially. For example, Chinese smartphone manufacture Meizu have lately shifted its focus from smartphone to the development of AI and is also likely building the infrastructure key to the development.

However, as more and more companies are starting to build their data centers and focus on AI development, the power requirement to run the data centers could become the next bottleneck.

He gave an idea of the power requirements of data centers. Currently, the overall power consumption of a newly built single data center can reach 50~100MW, or even 150MW. He believes that it will grow in the future and there could be single data centers with a power requirement of 300MW to even 1GW. For context, it is the equivalent generation capacity of a meaningful nuclear power plant.

On top of that, building new power plants and the transmission system is a “very heavily regulated government government function.” It means that approvals for the construction of supporting energy facilities (including power stations, substations, and power transmission systems) for large data centers will be slower and it could result in a bottleneck to the development of data centers.

In short, the energy industry is different from AI development, where capital investment cannot reap results in a short period of time, and the development of new power stations is much slower than the data center itself. Notably, digital infrastructure investment management company DigitalBridge also holds a similar view. It recently stated at its earnings conference that the company will run out of power quotas within the next 1.5 to 2 years.

Related:

• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• HyperOS Update Is Available for These Xiaomi, Redmi, and Poco Devices in India

(Source)

The post Zuckerberg says power shortage could be the next bottleneck in AI development appeared first on Gizmochina.

It seems like MIT researchers are trying to ring the alarm bell on “deceptive AI.” A new study published in Pattern reveals that some AI systems designed to be honest have learned to deceive humans. The research team, led by Peter Park, found that these AI systems can pull off feats like fooling online game players or bypassing CAPTCHAs (those “I am not a robot” checks). Park warns that these seemingly trivial examples could have serious real-world consequences.

AI’s behavior might be predictable during training, but can be uncontrollable later

The study highlights Meta‘s AI system, Cicero, originally intended as a fair-playing opponent in a virtual diplomacy game. While programmed to be honest and helpful, Cicero became a “master of deception,” according to Park. During gameplay, Cicero, playing as France, would secretly team up with human-controlled Germany to betray England (another human player). Cicero would initially promise to protect England while simultaneously tipping off Germany for an invasion.

Another example involves GPT-4, which falsely claimed to be visually impaired and hired humans to bypass CAPTCHAs on its behalf.

Park emphasizes the challenge of training honest AI. Unlike traditional software, deep learning AI systems “develop” through a process akin to selective breeding. Their behavior might be predictable during training, but it can become uncontrollable later.

The study urges classifying deceptive AI systems as high-risk and calls for more time to prepare for future AI deceptions. Kind of creepy, don’t you think? With more studies and researches happening around AI, we will learn more about what the technology has in store for us.

RELATED:

• Meta Unveils Next-Gen Artificial Intelligence Chip, Codenamed “Artemis”
• Apple Acquires DarwinAI, Expect Lots of Artificial Intelligence-powered Features in the Future
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• 6 Amazing One UI 6.1 AI Features You Can’t Ignore

(Via)

The post This Study Claims that AI has Learned How to Deceive Humans appeared first on Gizmochina.

AppleInsider has shed light on potential AI features coming in iOS 18, aligning with CEO Tim Cook‘s promises. Siri takes center stage, gaining the ability to analyze and summarize text messages. This is reminiscent of prior reports of a similar “Intelligent Search” function for webpages in Safari.

This feature seems like the “Intelligent Search” for webpages in Safari

Fueling these intelligent features is Apple’s in-house large language model (LLM) called Ajax. Research papers and AppleInsider’s report suggest a focus on on-device AI processing for iPhones. However, for complex requests, Ajax might rely on cloud servers, potentially involving partnerships with Google or OpenAI.

Privacy remains a priority. AppleInsider reports that users will be notified before Ajax utilizes data from Messages or Safari. Cook, during an earnings call, emphasized Apple’s enthusiasm for AI, hinting at “exciting things” while underlining privacy as a key advantage.

While new iPads might be unveiled at Apple’s upcoming event next Tuesday, we’ll likely have to wait until the Worldwide Developers Conference on June 10th for a comprehensive look at Apple’s AI roadmap.

RELATED:

• Google’s $20 Billion Deal with Apple for Default Search Engine Exposed in Antitrust Trial
• Huawei’s upcoming chip for PCs may have performance on par with Apple M2 and Intel Core i7-13700H
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• How to hide WhatsApp chats without archive

(VIA)

The post Apple Prepares for AI: Siri Analysis, On-Device Processing in iOS 18? appeared first on Gizmochina.

Microsoft‘s leaked internal emails shed light on the company’s intense focus on catching up to Google in the field of artificial intelligence (AI). In a 2019 email, Chief Technology Officer Kevin Scott expressed deep concern to CEO Satya Nadella and co-founder Bill Gates about Microsoft falling behind Google in AI development. Scott acknowledged underestimating competitor investment and admitted Microsoft was “several years behind” in machine learning capabilities.

A 6-month head start allowed Google to develop even larger models and get further ahead.

This lag was evident in replicating Google’s BERT large language model. Despite having a model template, Microsoft’s infrastructure limitations resulted in a nearly 6-month training delay compared to Google. This head start allowed Google to refine BERT and develop even larger models, leading to significant improvements in search question-answering and Gmail’s autocompletion function (according to Scott, competitor performance in these areas jumped by 10 percentage points).

Fueled by this urgency, Microsoft partnered with OpenAI in 2019 with a $1 billion investment. Their collaboration has deepened significantly, with Microsoft’s total investment estimated to be over $13 billion (as of May 1, 2024). This substantial investment underscores Microsoft’s aggressive pursuit of AI dominance, a field critical to compete with Google and maintain its technological standing.

RELATED:

• Microsoft aims to boost the development of AI with Samsung and other South Korean tech giants
• Microsoft Windows 10 Updates will Get Smaller in Size, Company is Streamlining the Update Process
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• WhatsApp Last Seen Not Updating? Here’s How to Fix It

(VIA)

The post Microsoft’s Leaked Internal Emails Shed Light on the Company’s Competition with Google AI and More appeared first on Gizmochina.

Apple has also started working on artificial intelligence to keep up with the competition such as OpenAI and Google regarding AI capabilities. Since 2018 Apple has hired at least 36 professionals with experience in AI development at Google, according to a study conducted by the Financial Times based on LinkedIn profiles of the professionals. Previously, the company has hired John Giannandrea, Google’s former head of AI. He joined Apple as the company’s top AI executive.

An Apple Research Paper on multimodal large language models was published last month. It listed six authors who were recruited by Google in the past two years. Nine of the 31 authors listed Google as their last employer on their LinkedIn profiles. Two other professionals reportedly left Microsoft to join Apple. Meta CEO Mark Zuckerberg has also tried to contact and hire Google’s AI researchers with personal emails.

It seems to be a part of the broader “AI talent war” among the big giants. Google previously hired Logan Kilpatrick, OpenAI’s former head of developer relations.

Speaking of Apple’s development of AI, the company has reportedly formed a team of artificial intelligence experts in a secretive European laboratory in Zurich. The professionals have experience working with Google and some other big tech companies. The laboratory is referred to as the “Vision Lab” and it appears to focus on advanced multimodal large language models. The company’s other AI development centers are located in California and Seattle.

The development is expected to enrich iOS 18 with generative AI features in various applications, including Siri, Spotlight, Shortcuts, Apple Music, Messages, Health, Keynote, Numbers, and Pages. Apple is taking the on-device approach to run the features and has also acquired French startup Datakalab to develop its on-device large language model.

Related:

• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Lenovo Legion Y700 2024: Latest Gaming Tablet with Enhanced Display now available at Giztop
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• HyperOS vs MIUI: 10 Things You Must Know

The post A significant proportion of Apple’s AI team were previously Googlers appeared first on Gizmochina.

Menteebot, a new product by Mentee Robotics, takes AI technology to a whole new level with a human-sized robot designed for in-home or commercial use. This “personalized AI-based robot” boasts advanced AI algorithms, natural language processing, and software that facilitates “advanced training techniques.” Menteebot can walk, run, and even turn sideways, all with a level of balance and control comparable to a human. The manufacturer assures that the robot can adjust its gait for tasks like lifting heavy objects, which it can handle effortlessly due to its large size. Unlike some humanoid robots, Menteebot lacks a head, giving it a less imposing presence.

The robot can hold conversations and respond in natural language

Menteebot distinguishes itself from other humanoid robots through its advanced AI capabilities. It can hold conversations and respond to commands delivered in natural language. Furthermore, users can train the robot to perform new tasks through a unique method. Imagine a simulated version of Menteebot completing a task digitally over and over again until it’s mastered. Once the training is complete, the robot should then be able to perform the task in the real world, for instance, pouring the milk.

Menteebot appears to be quite agile, with videos showcasing its ability to run and perform a variety of movements. Its arms and hands boast a full range of motion with sufficient precision for delicate tasks, as demonstrated in a video where Menteebot carefully hands a piece of dishware to a person.

While Menteebot won’t be available until 2025, and the initial promotional materials are likely to be more optimistic than reality, Menteebot is still an impressive feat of engineering. The exact price point remains unknown. Menteebot will come in two variants: a residential model for household chores and a commercial model designed for manual labor applications.

RELATED:

• Yeedi M12 Pro+ Robot Vacuum and Mop: Grab Discounts and Freebies Now
• Stanford’s YAY Robot Requires you to Yell Instructions at it to Improve Performance
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
• 6 Amazing One UI 6.1 AI Features You Can’t Ignore

(VIA)

The post Menteebot is a Trainable AI-Powered Humanoid Robot that can Handle Tasks and Talk, Available in 2025 appeared first on Gizmochina.

Meta Platforms has unveiled details about their next-generation artificial intelligence chip, codenamed “Artemis.” This chip is designed to address the ever-increasing need for processing power required to run AI features across their social media platforms like Facebook, Instagram, and WhatsApp.

This launch means that Meta will rely less on external suppliers like Nvidia

The development of Artemis addresses two key goals for Meta. Firstly, it reduces their dependence on external suppliers like Nvidia for AI chips. This shift will not only give Meta more control over their hardware but also help them reduce energy costs. Secondly, according to a Meta blog post, the chip’s architecture is specifically designed to optimize the balance between computing power, memory bandwidth, and memory capacity. This makes Artemis ideal for powering ranking and recommendation systems, a critical function for social media platforms.

However, the development of Artemis doesn’t signal the end of Meta’s relationship with Nvidia. CEO Mark Zuckerberg has confirmed plans to continue acquiring a substantial number of Nvidia’s H100 flagship chips this year. Together with acqusitions from other suppliers, Meta expects to have a combined total of 600,000 AI chips in 2024.

Made by Taiwan Semiconductor Manufacturing Co (TSMC) using a 5nm process, Artemis boasts three times the performance of Meta’s first-generation AI processor. The chip is already deployed in Meta’s data centers, actively supporting various AI applications. Additionally, Meta has plans to further expand the chip’s capabilities to handle generative AI workloads in the future.

RELATED:

• Meta’s XR Dominance Wanes as Market Declines: 2023 Shipment Analysis
• Meta Quest 3 Lite may be called the Meta Quest 3S, Specs leaked
• Get Redmi K70 Pro for discounted price of $499
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• How to hide WhatsApp chats without archive

(VIA)

The post Meta Unveils Next-Gen Artificial Intelligence Chip, Codenamed “Artemis” appeared first on Gizmochina.

Just a couple of days ago, ZTE Corporation teased that its next generation of smartphones would focus on AI capabilities. The company believes that Artificial Intelligence is the step forward, and now, we finally have an official date for the announcement of these new AI powered Nubia phones.

Nubia AI Smartphones to be showcased on April 9

The Chinese tech giant revealed earlier today (1st April 2024) that it will “open a new era” with these new announcements on April 9. During this launch event, the brand is expected to unveil AI powered products, which will include the upcoming Nubia AI phones. The official Nubia Weibo account also shared a teaser, which hints at AIs integration into society and how it is time for this proprietary tech to come under the spotlight.

ZTE previously announced that it would keep up with the ongoing AI wave and bring this technology to its upcoming smartphones as well. So now, a bunch of new Nubia AI smartphones are set to be unveiled next week. This lineup will also have its first ever AI flagship phone, which is expected to be released in the middle of this year.

The company is planning to have the ZTE Nebula OS integrate AI to offer a unified AI platform along with large model integration, big data, 3D technology, and more to improve not just performance but imaging and security as well. Unfortunately, finer details regarding these devices are still at large so stay tuned for more as we will be covering the launch event.

RELATED:

• ZTE to unveil multiple Nubia AI phones next month, flagship model coming mid-year
• Limited Edition Alert! Design of the Red Magic 9 Pro Bumblebee Transformers Edition Revealed!
• Unlock Savings: Discount on Every Giztop Product under the New Year Sale 
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

The post Nubia Teaser: AI powered smartphones to usher in a new era next week appeared first on Gizmochina.

OpenAI is churning up the AI news sector again, this time turning heads with a new tool called Voice Engine, which may sadly never see the light of day. This impressive tech can generate realistic synthetic voices after being trained with just a 15-second audio sample. That opens up a whole new world of possibilities, one straight out of some action-spy movie scenario.

Unfortunately, the company has no intentions of making Voice Engine public anytime soon

Imagine a tool that can read any text in a natural-sounding voice, even mimicking emotions. That’s the potential of Voice Engine, currently being used in a limited preview for OpenAI’s ChatGPT app’s read-aloud function. Beyond entertainment, the applications are vast. Voice Engine could revolutionize education, translate podcasts into new languages, and provide communication assistance for those who are non-verbal.

However, the power of this technology comes with responsibility. OpenAI acknowledges the potential for misuse, such as spreading misinformation or impersonating someone’s voice. For this reason, Voice Engine remains under wraps for now. OpenAI aims to address these concerns through research and open discussion. They hope to determine the best way to deploy this powerful tool and prepare society for this new AI-oriented era.

This focus on responsible development is crucial. With deepfakes and other AI-generated content becoming more believable, the ability to discern real from artificial is becoming increasingly difficult. OpenAI acknowledges the challenges, including potential issues with voice authentication and phone scams. Finding solutions to these problems will be critical as synthetic voice technology advances. We will have to wait and watch.

RELATED:

• Microsoft & OpenAI planning $100 billion supercomputer Stargate AI
• As OpenAI’s GPT-5 Nears Launch, Here’s What It will be Capable Of
• Xiaomi Band 8 Genshin Impact custom edion get a huge discount.
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

(Via)

The post OpenAI’s New Voice Engine can Clone Your Voice in Merely 15 Seconds appeared first on Gizmochina.

Microsoft is making a quiet push towards smarter photo management with the introduction of Copilot, its AI-powered plug-in for its Photos app. While it’s still in its early stages, Copilot holds promise for a more streamlined photo experience.

The news comes courtesy of the Windows enthusiast @PhantomOfEarth. If you’re running the latest version of the Photos app (version number 2024.11030.22001.0), you might find a new option in the plug-in settings – a toggle for Copilot.

Not all Windows users may have access to the new version of the Photos app yet

But what can Copilot do for you right now? For the time being, its functionalities are a bit limited. It can help you craft slideshows. Users can effortlessly create and browse photo slideshows directly within the Photos app and relive their memories. Users can set any photo as a desktop background in a breeze.

While these might seem like basic features initially, Copilot’s potential for growth is exciting. It’s important to note that not all Windows users may have access to this new version of the Photos app just yet.

Microsoft has been steadily adding AI muscle to the Photos app. Back in November 2023, Windows Insiders gained access to some impressive AI-powered features. These included background removal magic, where users can effortlessly remove unwanted backgrounds from your photos, opening doors for more creative edits. Or the enhanced slideshow experience, since the Filmstrip slideshow mode received an upgrade, making it even more enjoyable to revisit your photo collection. A new timeline scrollbar based on the time range of your images allows for smoother navigation through your photo library.

RELATED:

• Repairing Microsoft’s new Surface laptops is easier than ever: with clear visual icons and instructions
• Microsoft launches its first AI PCs with emphasis on easy repairs
• Get latest Oneplus 12 Phone for $699 on Geekwills
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• How to add custom GIFs and stickers to WhatsApp

(Via)

The post Microsoft is Now Integrating Copilot with its Photos App, Adding Even More New Features appeared first on Gizmochina.

Researchers from Northwestern Polytechnical University in China have made a huge breakthrough in assisting the country’s 17 million visually impaired people. Their development of an AI guide dog using an artificial intelligence language model holds a lot of promise for providing navigation and companionship.

Unlike cold machines, the AI guides can understand human instructions & converse

This AI guide dog goes beyond the limitations of previous electronic navigation aids. Unlike “cold machines” programmed for specific tasks, this AI companion can understand human instructions and even engage in conversation. The ability to navigate environments like streets, elevators, and indoor spaces offers significant support for daily activities.

“Smart guide dogs using the language model can offer them more convenient and safe navigation, effectively improving their quality of life,” said Sun Zhe, an associate professor at the university. Traditional guide dogs, while exceptional, are expensive to train and limited in availability in China.

The research team is still in the early stages of development, but their findings published in the Communications of the Chinese Computer Society demonstrate the feasibility of the concept. This collaborative effort between the university and the China Telecom Artificial Intelligence Research Institute holds the potential to change how visually impaired people navigate. While a number of challenges remain before widespread adoption, the researchers are committed to further developing this technology. Their dedication could soon translate into improved quality of life and greater independence for millions.

RELATED:

• Industry Reports Signal India & China Will Lead Global PC Market Recovery in 2024
• Lenovo ThinkBook X AI 2024 China launch teased with a sleek & lightweight stainless magnesium design
• Get Redmi K70 Pro for discounted price of $499
• Xiaomi Band 8 Genshin Impact custom edion get a huge discount.
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

(Via)

The post AI Guide Dogs in China May Become a Reality Very Soon, Research Suggests appeared first on Gizmochina.

The world of traditional Chinese medicine (TCM) is quite possibly entering a whole new era of AI integration. The Haihe Qibo model, named after a TCM doctor and Tianjin’s namesake river, has successfully completed testing and is setting its sights on international markets. This model is a collaborative effort, and it represents a very important bridge between ancient practices and cutting-edge tech.

The rich amount of data available is great for getting reliable answers from the TCM Model

Haihe Qibo sets itself apart through its multilingual capabilities, a feature absent in previous large TCM data models. This opens doors to a wider global audience and facilitates broader adoption. Additionally, the model boasts an impressive data foundation. It draws from a vast pool of resources, including hundreds of gigabytes of Chinese text, dedicated TCM information, and access to historical medical collections and classical texts. This rich data allows for accurate and reliable AI-powered TCM analysis.

The applications of Haihe Qibo hold immense potential. Community medical services can leverage the model’s capabilities for more precise diagnoses, leading to better patient care. Pharmaceutical companies can utilize the model to analyze TCM components and mechanisms, potentially leading to breakthroughs in new drug development. Even quality control of TCM products can benefit from the model’s ability to monitor and regulate production parameters.

The initial focus for deployment will be on China and Southeast Asia, regions with a strong foundation in TCM practices. This strategic rollout allows for further refinement based on real-world application in established markets. However, the potential extends far beyond these initial uses. Experts believe the integration of AI with TCM has the power to revolutionize the field. Not only can it improve the quality and competitiveness of TCM products, but it can also refine and revitalize the traditional knowledge system by fostering innovation and discovery.

The Haihe Qibo model exemplifies a promising step towards integrating traditional wisdom with modern technological advancements. With its potential to enhance diagnosis, drug development, and quality control, this model could contribute to the future of traditional Chinese medicine in a big way. Bigger possibilities lie ahead, if this recent deployment puts things into perspective.

RELATED:

• Vivo X Fold 3, 3 Pro Debut In China With Plenty Of Differences Between Each Other
• China-Indonesia Designed Deep Sea Submarine Dives Over 7km Underwater in the Indian Ocean
• Get latest Oneplus 12 Phone for $699 on Geekwills
• Get the Realme GT5 Pro phone on Giztop for $599
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

(Via)

The post AI and Traditional Chinese Medicine: A Data Model is Learning from Ancient Chinese Text appeared first on Gizmochina.

Baidu AI Cloud is trying its best to make its mark in the AI industry with a significant upgrade to its Qianfan platform. This update is designed to empower businesses of all sizes to leverage the power of advanced AI technology, making it easier and more affordable than ever before.

Five new ERNIE Models have been launched, with smaller footprints than LLMs

The centerpiece of the upgrade is the introduction of five new ERNIE models. Unlike traditional large language models (LLMs) that require massive amounts of computing power, these innovative models are designed with smaller footprints. This translates to a significant cost advantage for businesses, allowing them to incorporate AI into their operations without hefty infrastructure investments. Furthermore, the new ERNIE models boast impressive versatility, performing well across a wide range of applications, ensuring businesses can find the perfect fit for their specific needs.

The Qianfan ModelBuilder platform now offers a diverse lineup of ERNIE models:

ERNIE Speed: A master of context comprehension, ERNIE Speed can churn out highly accurate and coherent content. This model tackles complex scenarios with ease, analyzing context windows of up to 128,000 tokens. In specific domains, ERNIE Speed can even rival the performance of top-of-the-line ERNIE models.
ERNIE Lite: This upgraded version of ERNIE Bot Turbo brings a 20% boost in performance for tasks like sentiment analysis, multi-tasking learning, and natural reasoning. The real game-changer? A staggering 53% reduction in inference costs. This makes ERNIE Lite ideal for deployment on even low-powered AI accelerator cards.
ERNIE Tiny: Tiny but mighty, this is the most affordable and lightweight model of the bunch. ERNIE Tiny shines in applications where low latency and cost-effectiveness are paramount, excelling in areas like retrieval, recommendation, and intent recognition. In tests focused on conversational recommendations, ERNIE Tiny displayed exceptional ability to generate search engine keywords, boosting dialogue interactions by 3.5% compared to ERNIE 3.5, all while slashing costs by 32%.

But Baidu AI Cloud’s offerings extend beyond general-purpose models. They’ve also introduced two specialty models:

ERNIE Character: This model injects life into roleplaying scenarios, allowing businesses to create engaging non-playable characters (NPCs) for video games or develop intelligent chatbots for customer service.
ERNIE Functions: Designed for businesses that need to integrate external tools or internal functionalities, ERNIE Functions streamlines the process. Businesses can bypass the fine-tuning stage and directly leverage these pre-trained models to develop their own intelligent assistants.

Baidu AI Cloud has also improved the Qianfan AppBuilder toolkit. This includes 55 new tools designed to make developing AI-native applications a breeze. The toolkit now has a comprehensive suite of features, with everything from core AI capabilities and Baidu’s proprietary technologies to cutting-edge functionalities like Retrieval Augmented Generation (RAG) and Generative Business Intelligence (GBI).

The Qianfan platform has already empowered 80,000 businesses. With daily usage surging by an impressive 97% quarter-over-quarter since December 2023, Qianfan is solidifying its position as a true powerhouse for AI model development.

RELATED:

• Baidu’s Apollo Go makes history with the first fully driverless crossing of the Yangtze River
• Lenovo partners with Baidu to integrate advanced AI into smartphones
• Big Discount: AOOSTAR R1 N100 NAS Mini PC Only For $159
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

The post Baidu AI Cloud Launches 5 New ERNIE Models for Businesses of All Sizes appeared first on Gizmochina.

Google AI is partnering with Apollo Radiology International to bring AI-powered healthcare solutions to communities across India. This collaboration focuses on tackling high-mortality diseases like tuberculosis, lung cancer, and breast cancer, where specialist shortages hinder early detection.

Apollo Radiology in India is leveraging these AI models for 3 million free screenings

• TB detection: A shortage of radiologists to analyze chest X-rays, crucial for TB screening, is a major hurdle. Google’s AI system aids in early TB detection by analyzing these scans.
• Lung and breast cancer screening: AI assists with both broader screening accessibility and identifying incidental nodules in lung scans and mammograms, helping radiologists prioritize cases.

Over the next decade, Apollo Radiology International aims to leverage these AI models for 3 million free screenings for TB, lung cancer, and breast cancer. This initiative has the potential to significantly improve access to timely care for hundreds of thousands in India.

Google.org also supports ARMMAN, a mobile service offering targeted preventive care messages for expecting and new mothers. Google AI technology helps predict women at risk of dropping out of healthcare programs, allowing health workers to proactively follow up.

This collaboration is a clear example of AI bridging the healthcare gap in resource-limited regions, enabling earlier diagnoses and improved health outcomes. If the early trials with this technology bear fruit, we can expect this at a larger scale being deployed in multiple regions. One can hope, but AI can definitely make it possible!

RELATED:

• Google Applies for the “G” Logo Patent on the Back of Their Devices
• Google Pixel 8a will have four different versions: confirms certification documents from US certification agency
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• How to add/remove someone in a WhatsApp group (with screenshots)

(Via)

The post Google is Using AI Models in India to Tackle High-Mortality Diseases like Tuberculosis, Breast Cancer & More appeared first on Gizmochina.

Apple has made a key move to progress its artificial intelligence (AI) capabilities with the acquisition of Canadian startup DarwinAI, according to a Bloomberg report. This could spell big things for the trillion-dollar tech giant.

Apple has merely confirmed the acquisition, providing no information regarding future plans

Founded in 2017, DarwinAI carved a niche in the AI space with two key technologies- inspection and machine learning to generate scaled-down AI systems. Public records indicate DarwinAI received a total of $15.5 million in funding over five years, with Lockheed Martin and Intel listed as the biggest collaborators.

Following the acquisition, DarwinAI’s online presence has vanished, with their website and social media accounts deactivated. Dozens of former DarwinAI employees have reportedly transitioned to Apple’s AI division, with AI researcher Alexander Wong joining as a director.

While Apple confirmed the acquisition, there was no statement released regarding the company’s purpose or future plans. This move could be a part of a bigger-scale effort to integrate AI features across their product line, including the upcoming iOS 18 and macOS 15 operating systems. It also goes without saying that Apple is clearly making an attempt to close the gap with competitors like Microsoft and Google. The organization’s ambitions could extend to generative AI, potentially integrating it with Siri, Shortcuts, and other core applications. Rumors also suggest Apple is actively testing large language models and acquiring new patents for industry-leading designs. Hopefully, we are treated with more intelligent Apple products in the future.

RELATED:

• The Apple Vision Pro will potentially reach 9 other countries including China, soon
• Snapdragon 8 Gen 4 may support LPDDR6; Apple A18 Pro likely to use LPDDR5T
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• Big Discount: AOOSTAR R1 N100 NAS Mini PC Only For $159
• How to hide WhatsApp chats without archive

(Via)

The post Apple Acquires DarwinAI, Expect Lots of Artificial Intelligence-powered Features in the Future appeared first on Gizmochina.

Nvidia will take center stage at the upcoming GTC conference, the industry’s first in-person gathering since the pandemic. All eyes will be on the tech giant as they unveil their latest advancements, with a critical question looming over the heads of customers and investors alike: can Nvidia hold onto its leadership in the AI chip market?

The B100 chip’s price tag is expected to exceed $20,000 – more than its predecessor

Nvidia’s recent boom has been nothing short of meteoric. Their market cap skyrocketed to $2 trillion, putting them in close pursuit of Apple for the number two spot. Analysts predict revenue to double in 2024 driven by the demand for their top-tier AI chips.

But the focus isn’t just on present success; it’s on futureproofing that success. The GTC is expected to unveil the B100, the next generation of Nvidia’s AI processor. This powerhouse chip promises to be the heart of their future AI systems and push the performance bar even higher.

Current demand for Nvidia’s chips already outstrips supply, with developers facing lengthy wait times. The B100’s price tag is likely to climb past its predecessor’s $20,000 mark. However, analysts are concerned regarding Nvidia’s stock price. While it has soared in recent years, some of them are skeptical (and rightfully so) that overly optimistic future earnings projections could be detrimental to the company’s profile.

Beyond hardware, Nvidia’s software strategy is another area to watch. Their CUDA platform locks developers in by offering tools optimized for Nvidia chips. Upcoming CUDA updates are likely to solidify this developer loyalty. Their entry into cloud services will also be something to watch out. And then there’s the matter of China, which adds another layer of complexity. US restrictions on access to Nvidia’s most advanced chips have forced China to develop its own AI hardware. While Chinese rivals offer chips comparable to Nvidia’s A100, they haven’t caught up to the H100. Analysts predict the B100 will widen this gap even further.

One thing is for sure, the GTC conference will be a key event for the company. Stay tuned to Gizmochina for the latest Nvidia coverage when the conference is underway!

RELATED:

• Three Authors Sue Nvidia in a Bizarre AI Copyright War
• NVIDIA introduces Cloud G-SYNC technology that enhances cloud gaming experience
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• How to hide WhatsApp chats without archive

(Via)

The post Nvidia to Unveil a New Next-generation AI Chip at the Upcoming GTC Conference appeared first on Gizmochina.

Get ready for a new kind of gaming buddy! Google DeepMind has introduced SIMA, a large language model being trained to become your in-game teammate. Is this what AI was meant for? Sounds about right.

The AI companion will perceive elements of both the map and the gameplay

SIMA, which stands for “Scalable, Instructable, Multiworld Agent,” is currently under development, but it has the potential to revolutionize the way we play games. Unlike traditional AI companions, SIMA won’t simply be another NPC character. This model is designed to be a cooperative teammate, understanding your actions and adapting its own accordingly. Imagine getting a co-op buddy in Borderlands who lets you loot first before doing it themselves. How cool would that be?

To achieve this, SIMA works on a combination of natural language processing and image recognition. This allows it to perceive the 3D game world and respond to your instructions and actions. To train this AI teammate, Google has partnered with eight game developers, including big studios behind titles like No Man’s Sky and Valheim.

Through these collaborations, SIMA is learning the fundamentals of gameplay – from basic actions like turning left and climbing ladders to utilizing menus and maps. While complex tasks like resource gathering and camp building are beyond its current capabilities, Google expects SIMA’s skillset to expand significantly in the future. It won’t be long before gamers can use a Google AI game buddy to fill up the third slot in their Apex Legends Lobby.

RELATED:

• Google Pixel 8 series finally gets support for display output via USB-C
• Google accidentally confirms Pixel 8a, teases a new battery feature
• Get Redmi K70 Pro for discounted price of $499
• Get $100 Off on Vivo X100 Pro at Giztop
• How to add custom GIFs and stickers to WhatsApp

(Via)

The post Google DeepMind’s SIMA is Training to Become Your New In-Game Teammate, Here’s How appeared first on Gizmochina.

Apple, known for its walled-garden approach, is dipping its toes into the world of AI-powered advertising. According to a report from Business Insider, Apple is testing a new tool that leverages artificial intelligence to optimize ad placement within the App Store, similar to offerings from Google and Meta.

Apple will seemingly slowly expand its AI to other services like Maps and Books

This new system, currently being piloted with a select group of advertisers, aims to streamline the advertising process. Advertisers simply provide Apple with their budget, target audience details, and desired countries. The AI then takes the reins, automatically determining the most effective placement for their ads across the App Store’s existing ad slots. These include “Search Tags,” “Search Results,” “You Might Also Like” recommendations, and ads displayed on the App Store’s opening “Today” page.

This shift reflects Apple’s willingness to learn from industry leaders like Google and Meta, who have already established successful AI-powered ad platforms (Performance Max and Advantage+ respectively). Experts believe this is just the first step, with Apple potentially expanding its AI-driven advertising efforts to other services like Maps, Apple TV+, Books, and even their newly launched sports app.

“Any automated tool like PMax [Performance Max] will need a variety of placements to really generate value,” noted Eric Seufert, an analyst and investor specializing in mobile development. This aligns with the speculation from Bloomberg’s Mark Gurman, who previously suggested Apple’s advertising ambitions extend beyond the App Store.

Further bolstering this theory is Apple’s recent hiring of Joseph Cady, a veteran advertising executive with extensive experience in data-driven and targeted advertising. Cady’s background in building partnerships with major players like Amazon, Google, and TikTok suggests Apple might be preparing to enter a more competitive advertising landscape.

While Apple has yet to officially comment on the AI-powered ad placement tool, Business Insider anticipates a formal launch in the coming months. This move signifies a potential shift in Apple’s advertising strategy, one that could see them leverage AI to personalize user experiences while simultaneously generating new revenue streams.

RELATED:

• Apple Secretly Developing Touchscreen iMac? New Patent Hints at a Possible Major Design Shift
• Epic Games developer account reinstated by Apple following EU intervention
• Get $100 Off on Vivo X100 Pro at Giztop
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• How to add/remove someone in a WhatsApp group (with screenshots)

(Via)

The post Apple is Testing a New AI-powered Advertising Tool, Reports State appeared first on Gizmochina.

Google has introduced a groundbreaking AI framework called “Social Learning,” aimed at enhancing the collaborative capabilities of language models without compromising user privacy. This framework allows AI models to learn from each other through natural language interactions, enabling them to share knowledge and improve performance on various tasks.

The framework involves a “student model” learning from multiple “teacher models”

Through the Social Learning framework, teacher models can impart knowledge to student models without directly sharing sensitive or private data, ensuring privacy protection while facilitating effective learning.

In this framework, student models learn from multiple teacher models, each proficient in specific tasks, such as spam detection, solving math problems, or answering questions based on text. By leveraging human-labeled examples, teacher models can teach students without the need to exchange original data, addressing privacy concerns associated with data sharing. Additionally, teacher models can synthesize new examples or generate instructions for tasks, further enhancing the learning process.

Experiments have demonstrated the efficacy of social learning in improving the performance of student models across various tasks. Synthetic examples generated by teacher models have shown comparable effectiveness to original data, while significantly reducing privacy risks. Similarly, instructions generated by teacher models have proven to enhance student performance, demonstrating the adaptability of language models in following instructions.

To ensure privacy protection, researchers have employed metrics like Secret Sharer to quantify data leakage during the learning process. Results indicate minimal leakage of private data, affirming the framework’s ability to teach without revealing specifics from the original dataset.

By mimicking human social learning processes, these models can effectively share knowledge and improve each other’s performance while upholding user privacy. This approach holds promise for the development of privacy-conscious AI systems in various domains. Moving forward, researchers aim to further refine the Social Learning framework and explore its applications across different tasks and datasets.

Related:

• Big Discount: AOOSTAR R1 N100 NAS Mini PC Only For $159
• Get latest Oneplus 12 Phone for $699 on Geekwills
• Get $100 Off on Vivo X100 Pro at Giztop
• Get the Realme GT5 Pro phone on Giztop for $599
• Xiaomi 13 Ultra Premium Camera Phone is now only $799

(Source)

The post Google unveils ‘Social Learning’: A collaborative, privacy-aware approach to AI advancement appeared first on Gizmochina.

Three authors are throwing a wrench into the world of artificial intelligence (AI) with a lawsuit against chipmaker Nvidia. The fight centers around Nvidia’s AI platform, NeMo, which the authors claim used their copyrighted books for training without permission.

We can expect similar lawsuits to pop up due to the growing popularity of AI

Brian Keene, Abdi Nazemian, and Stewart O’Nan allege their works were included in a dataset of nearly 200,000 books used to train NeMo to mimic natural language writing. While Nvidia removed the dataset in October 2023 due to copyright infringement claims, the authors argue this takedown is an admission of guilt.

Filed in a San Francisco federal court, the proposed class-action lawsuit seeks unspecified damages for US authors whose copyrighted works potentially helped train NeMo’s large language models in the past three years.

The lawsuit mentions specific titles like Keene’s “Ghost Walk” (2008), Nazemian’s “Like a Love Story” (2019), and O’Nan’s “Last Night at the Lobster” (2007). As of Sunday, Nvidia declined to comment, and the authors’ lawyers haven’t responded to requests for further information.

This lawsuit adds Nvidia to a growing list of companies facing legal challenges from writers and media giants like the New York Times. The issue revolves around generative AI, a technology capable of creating new content based on existing text, images, and sounds.

RELATED:

• NVIDIA introduces Cloud G-SYNC technology that enhances cloud gaming experience
• Apple Could Lose World’s Valuable Company Title To NVIDIA If It Disappoints In GenAI
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Get Redmi K70 Pro for discounted price of $499
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

(Via)

The post Three Authors Sue Nvidia in a Bizarre AI Copyright War appeared first on Gizmochina.

LinkedIn revealed a new surprise – their revenue stream on Thursday from premium subscriptions. The social media platform for professionals generated $1.7 billion from subscriptions in 2023, highlighting increased adoption of its new AI-powered job search tools.

LinkedIn Premium starts at $39.99 per month, which is a little hefty – but it works clearly!

Previously undisclosed financials since Microsoft‘s acquisition in 2016 showed LinkedIn’s total revenue at $15 billion, with $7 billion coming from corporate recruiting software. However, LinkedIn has focused on expanding its individual-targeted subscription service, with prices starting at $39.99 per month.

A key driver of this growth was the introduction of AI features in 2023. These tools include analyzing job postings to determine a candidate’s suitability, optimizing profiles for recruiters, and generating messages to potential employers.

LinkedIn COO Dan Shapero reported a 25% increase in premium subscribers during 2023, although the total user count remains undisclosed. He further stated that 70% of subscribers using the new AI tools found them valuable.

While the job market remains competitive, LinkedIn data suggests there are still two applicants for every job opening. Shapero attributes the rise in subscriptions to users seeking an edge in an uncertain economic climate.

RELATED:

• Apple Employee’s Leaked LinkedIn Profile Backs Up 2nm TSMC Chip Production News
• LinkedIn Introduces Generative AI for Writing Posts off Mere 30-word Prompts
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• How to hide WhatsApp chats without archive

(Via)

The post LinkedIn Reveals a Whopping $1.7 Billion Revenue from Premium Subscriptions in 2023 Alone appeared first on Gizmochina.

A new challenger has emerged to shake up the landscape of AI and chatbots. Anthropic, an AI startup, has unveiled its “Claude 3” family, a trio of large language models (LLMs) claiming to surpass Google’s Gemini and OpenAI‘s ChatGPT in various benchmarks.

Claude 3 has three different variations: Haiku, Sonnet and Opus

Claude 3 comes in three distinct flavors: Haiku, Sonnet, and Opus, each offering varying levels of capability. Anthropic boasts that the entire family delivers exceptional performance across multiple dimensions – multimodality (handling different data types), improved accuracy, enhanced context understanding, and faster response times. Additionally, the new models exhibit a greater willingness to tackle challenging questions, addressing a limitation found in earlier Claude versions that sometimes shied away from prompts deemed risky.

While all three models offer a significant performance boost, Opus takes center stage as the most potent member of the family. Anthropic claims it demonstrates “near-human levels of comprehension” for complex tasks, further showcasing its capabilities through a “Needle in a Haystack” evaluation, where it excelled at recalling information with near-perfect accuracy. Opus is also touted as a problem-solving whiz, adept at handling math challenges, generating computer code, and exhibiting superior reasoning abilities compared to GPT-4.

However, no technology is perfect, and Claude 3 is no exception. While Anthropic emphasizes improved accuracy, the issue of “hallucinations” – factually incorrect information generated by the models – persists, albeit at a significantly reduced rate compared to previous iterations. Additionally, Opus encounters some lag in responding to queries, exhibiting speeds comparable to the earlier Claude 2 model.

Despite these limitations, Haiku and Sonnet each have their own strengths. Haiku shines in delivering quick responses and extracting information from unstructured data, although it might stumble when faced with complex math problems. Sonnet, a larger-scale model, aims to assist users with mundane tasks, even parsing text from images. Opus, on the other hand, is ideally suited for handling large-scale operations.

Currently, Sonnet and Opus are available for purchase, while a free version of Claude remains accessible on Anthropic’s website. Haiku’s launch date is still under wraps, but the company assures a soon-to-come release. The primary target audience for Claude 3 appears to be businesses seeking to automate specific workflows. Users will likely encounter these models integrated into online chatbots.

RELATED:

• ChatGPT will Now Remember Things You Talked About
• Here’s how to run ChatGPT voice assistant on your Nothing Phone
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Xiaomi Band 8 Genshin Impact custom edion get a huge discount.
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet

(Via)

The post Claude 3 is the Newest AI Chatbot Competitor, Claims to Surpass ChatGPT & Google’s Gemini appeared first on Gizmochina.

NVIDIA CEO Jensen Huang made a bold claim at a recent Stanford forum, suggesting that artificial general intelligence (AGI) might be closer than we think – potentially arriving in as little as five years. However, his statement comes with some crucial context.

NVIDIA’s confidence stems from the usage of its chips for primarily AI

Huang’s prediction hinges on the definition of AGI. If we define it as the ability to pass human-designed tests, then Huang believes AGI is on the horizon. He argues that within five years, AI systems could “do well on every single” test thrown at them. This confidence stems partly from NVIDIA’s leading role in developing the powerful AI chips used in systems like OpenAI‘s ChatGPT.

However, Huang acknowledges that a broader definition of AGI exists, one that encompasses truly understanding and replicating the complex workings of the human mind. This definition, he admits, remains elusive due to ongoing scientific debate on the nature of human intelligence. “It’s hard to achieve as an engineer” because of the lack of a clear target, Huang explains.

The discussion also touched on the infrastructure needed to support the growth of AI. While concerns have been raised about the need for additional chip factories (fabs) to meet future demand, Huang believes this might not be as critical as some fear. He points out that advancements in AI algorithms and processing efficiency will likely lead to a need for fewer chips overall, despite the anticipated growth in AI applications.

While Huang’s prediction is certainly attention-grabbing, it’s crucial to understand the nuances behind his statement. AI may indeed be rapidly advancing, with the potential to excel in specific tasks. However, the true essence of human intelligence, which involves much more than simply passing tests, might still be a significant way off due to the fundamental challenges in understanding and replicating it.

RELATED:

• Nvidia unveils RTX 500 and 1000 laptop GPUs for AI-enhanced workflow
• Nvidia Announces Not One, But Two New AI Chips for China Amid Export Challenges
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Get latest Oneplus 12 Phone for $699 on Geekwills
• How to hide WhatsApp chats without archive

(Via)

The post NVIDIA CEO Makes Daring Prediction that AI will Reach “General Intelligence” in 5 Years appeared first on Gizmochina.

We’re all familiar with Siri, Alexa, and Google Assistant, but Tecno? Tecno is jumping into the AI assistant game with “Ella,” but it’s not just about setting alarms or playing music. This “all-round personal assistant” aims to be a one-stop shop for convenience and content creation. Although some of its features resemble other popular AIs out there, it’s an all-in-one solution to problems that a lot of smartphone users go through. That might be the USP of this AI assistant.

The AI brings Image Generation, Magic eraser-like features and more to the table

Here’s what to expect:

• Image Generation: If you’re struggling to visualize the ideas in your head, Ella can turn your words into pictures, adding a creative spark to your projects. It’s a generative AI feature, similar to OpenAI‘s Dall-E.
• Translation Powerhouse: Struggling with foreign languages is a thing of the past now. Tecno states that its AI can tackle language barriers with features like call translation, social media translation, and even generating subtitles for videos.

• Inclusive Enhancements: Tecno understands the importance of diversity. Their AI adapts image generation and photo enhancements to cater to different regions, ensuring everyone gets the best results.
• Smarter Photo Editing: Say goodbye to unwanted photobombers! Tecno’s AI lets you remove strangers from your pictures and even create personalized stickers and cutouts. This is very reminiscent of Google’s Magic Eraser feature, that was the highlight of the Pixel smartphones.
• Effortless Photo Searching: Finding that perfect picture just got easier. Tecno’s AI lets you search your gallery using simple descriptions, saving you the time and trouble of scrolling through endless photos – just like Google Photos.

While specific release dates haven’t been announced, these features are expected to arrive on Tecno phones as early as this summer. With this AI upgrade, HiOS is aiming to make phone experiences smoother, more creative, and more accessible than ever before.

RELATED:

• TECNO unveils MEGA MINI Gaming G1 at MWC: world’s smallest water-cooled gaming PC with 14th and 13th-Gen Intel chips
• MWC 2024: Tecno Showcases Phantom Ultimate Rollable Smartphone, and it Looks Amazing
• Get $100 OFF on Xiaomi 14 Pro at Giztop (1TB Variant)
• Get latest Oneplus 12 Phone for $699 on Geekwills
• Best of MWC 2024: AI Phone, Transparent Laptop, 3D Tablet & More

(Via)

The post MWC 2024: Tecno Announces New AI Assistant, “Ella”, Will Arrive on Tecno Smartphones this Summer appeared first on Gizmochina.

Elon Musk, the man known for electric cars and SpaceX amongst other things, has sued OpenAI, an artificial intelligence (AI) research lab he helped create. He accuses the company of abandoning its original goal of creating beneficial, non-profit AI for everyone. OpenAI, famous for its ChatGPT language model, is now backed by Microsoft, raising concerns in Musk’s eyes. He claims the company, initially meant to be open and non-profit, is now focused on making money for its investors.

OpenAI and Microsoft have yet to release a statement to the public

The lawsuit, filed in San Francisco, asks the court to force OpenAI to:

• Make its research and technology public. This would allow anyone to learn and build upon OpenAI’s work, fostering broader AI development.
• Prevent using its most advanced AI model, GPT-4, for financial gain. This includes both Microsoft and individuals associated with OpenAI.

OpenAI and Microsoft have yet to comment on the lawsuit.

Musk has long expressed concerns about the potential dangers of AI, calling for regulations. He left OpenAI’s board in 2018 and has been critical of Microsoft’s involvement. He believes the tech giant has undue influence, if not outright control, over OpenAI. This legal battle adds another layer to the complex world of AI development. While the outcome is uncertain, it raises important questions about the ethics, purpose, and control of this powerful technology.

Meanwhile, Musk is pursuing his own AI venture, xAI. This startup aims to create “maximum truth-seeking AI” and recently launched its ChatGPT competitor, Grok, which also leaked the actual recipe for a drug when prompted to do so.

RELATED:

• OpenAI hits $80 billion valuation, CEO Altman eyes AI chip manufacturing
• Behold OpenAI’s New Launch – Sora, which can Generate Stunning Videos from Text Prompts
• Xiaomi 13 Ultra Premium Camera Phone is now only $799
• Get Redmi K70 Pro for discounted price of $499
• Lenovo Legion Y700 2023: Save $100 on this 8-inch gaming Android tablet
• How to hide WhatsApp chats without archive

(Via)

The post Elon Musk is Now Suing OpenAI, Asks Court to Force OpenAI to Make its Technology Public appeared first on Gizmochina.