Category: Artificial intelligence

  • Tesla Optimus Robot Gains AI Communication, Hive Mind, Self-Charging

    Tesla Optimus Robot Gains AI Communication, Hive Mind, Self-Charging

    While Tesla’s Vice President for Optimus robot development has confirmed that the bots showcased at the Robotaxi event were helped by human operators, he also mentioned that a significant update with new features is on the way.

    Exciting New Features Announced

    It looks like that update is happening today, as the Optimus team has just revealed a range of new capabilities, some of which might have been demonstrated during the Robotaxi celebrations.

    To start, the Tesla Bot has gained the ability to communicate with other Optimus humanoids nearby, allowing it to map out unfamiliar environments. Additionally, it can now explore "unseen" areas independently.

    Enhanced Physical Abilities

    The robots have also improved their physical skills, now capable of climbing stairs and handling substantial loads, like a 25-pound tray filled with battery cells.

    Perhaps the most noteworthy feature of the new Optimus robot is its ability to utilize artificial intelligence and interact with humans through a neural network that operates in real-time on the bots. This means it can recognize and fetch various objects on command, marking a significant step towards functioning as a humanoid butler.

    Autonomous Charging Capabilities

    On the topic of autonomy, the Tesla Bot can now independently search for, find, and plug into the nearest charging station, even in unfamiliar settings.

    Tesla’s Optimus charging connector appears to be situated on its back, and the rear cameras will now assist in locating nearby charging stations.

    It remains uncertain how many of these features were actively used during the Robotaxi presentation, as the Optimus bots there seemed to walk and dance autonomously, yet likely required human help to pour drinks at the bar or engage with attendees.

  • Tesla’s Optimus Bots Walked and Danced at Robotaxi Event

    Tesla’s Optimus Bots Walked and Danced at Robotaxi Event

    One of the key Tesla engineers involved in the Optimus robot initiative shed some light on how the Tesla Bots behaved during the unveiling of the Robotaxi.

    As Elon Musk hinted that the Optimus robots would be mingling among guests and even serving drinks at the bar after the event, it appears that their ability to perform more intricate tasks was somewhat exaggerated.

    Insights from Tesla’s Engineer

    Milan Kovac, the engineer behind the Tesla Bot, explained that the roughly twenty Optimus units showcased at the event were able to walk and dance independently for four hours. However, when it came to handling more complex functions, the robots didn’t use any mysterious AI to determine who ordered how many drinks; they were actually "human-assisted." This should not come as a surprise.

    Throughout the event, about 20 bots were constantly active—navigating through a busy crowd, dancing, snapping selfies, and even serving drinks and snacks. While they were indeed assisted by humans to some degree to illustrate our vision for a fantastic future, they managed to walk, balance, and dance on their own for approximately four hours, with only one minor incident (a handbag got in the way of a bot’s movement).

    A Showcase of Innovation

    Overall, it was an impressive display, especially when combined with other announcements like the Cybercab and the self-driving minibus.

    The primary goal of featuring the Tesla Bot was to highlight the advancement in the robotic hand’s capabilities, which increased from 11 degrees of freedom to 22. This change significantly enhances the tactile sensing abilities of the robots. Additionally, the demonstration showcased the autonomy and balance of the Tesla Bots.

    Creating appealing demonstration videos is challenging, but ensuring safe, live functioning of numerous humanoid robots for hours in a public outdoor setting set a much higher standard. This effort led to major advancements in full-body control, locomotion, hardware stability, and overall infrastructure.

    Future Prospects

    Nevertheless, the Optimus bots observed pouring drinks with rather slow and careful motions, as well as interacting with attendees, were all under the guidance of human operators. This suggests that Elon Musk’s vision of having an Optimus as a personal butler for $30,000 each is still quite a way off.

    Kovac did hint at some major progress in the autonomy of the Tesla Bots that the company has made, and he teased that more information will be shared with the robot-enthusiast public "soon."

  • Motorola Moto AI: Personalized Features for an Intuitive Experience

    Motorola Moto AI: Personalized Features for an Intuitive Experience

    Motorola has launched a new set of features powered by AI, all of which are grouped under the name Moto AI. These innovations are designed to improve the smartphone experience, making daily tasks more user-friendly. The company is concentrating on three main areas, with "Assist" being one of them.

    Moto AI’s Understanding Capabilities

    The Moto AI system is capable of comprehending user commands and adapting to their habits to provide personalized recommendations. This is accomplished through the use of Large Action Models (LAMs). Naturally, privacy can be a worry, but Motorola assures users that all information is stored securely and locally. Users can easily manage and access their data when they need to.

    Enhancing Creativity

    Another important focus area is "Create." The new AI suite aims to boost users’ creativity. Moto AI can generate summaries of conversations, photos, and documents using AI technology. It also claims to make photo and video editing easier through advanced editing options.

    Improving Photo Quality

    The final area of focus for Motorola is "Capture," where Moto AI seeks to simplify the process of taking high-quality pictures. It includes features like automatic scene detection, which enhances the quality of images taken. Additionally, the AI suite aids in capturing documents and converting them into various digital formats effortlessly.

    Motorola also aims to provide a smooth experience across multiple devices with Moto AI. This is facilitated by Smart Connect, which enables users to control and connect to different devices using natural language commands. Data sharing between connected devices is also straightforward.

    Motorola has not specifically listed which phones will support Moto AI yet. However, the company has indicated that these features are currently being tested in beta and will gradually be made available via invitations. The latest Motorola flagship models are expected to gain access sooner, such as the 12/256 GB 2024 Razr+ currently priced at $799.99 on Amazon.

    Motorola PR’

  • Lenovo Launches AI Agent and Learning Platform for Next-Gen PCs

    Lenovo Launches AI Agent and Learning Platform for Next-Gen PCs

    Lenovo AI Now utilizes Llama 3.1 to offer a fresh, chat-oriented method for locating documents, files, and various data on the newest Lenovo PCs. The interface is organized into clear sections that aim to be user-friendly, featuring a PC Assistant that includes automatically created shortcuts for commands like "Turn on Battery Saver Mode."

    Local and Cloud Chat Features

    The AI Now program also includes sections for Local and Cloud Chat. Lenovo claims that it uses Microsoft Azure AI Content Safety and has received certifications like the UL Verified Mark for AI Model Transparency, ensuring user safety and security.

    Learning Zone Tools

    This AI can operate fully on the device, much like the Lenovo Learning Zone, a brand new set of tools and services designed to enhance education, whether it’s online or face-to-face. It’s designed to gather resources like audio/video lectures, PDF files, and presentations to create notes and summaries, which can be organized by subject.

    The Learning Zone is also capable of creating its own quizzes to help improve retention and engagement with educational content.

    Availability of Features

    This tool is expected to be available as a free optional download for select Lenovo AI PCs starting in December 2024, while AI Now is anticipated to launch in the first quarter of 2025.

    Lenovo Press Release

    Lenovo presents at Tech World 2024 Smarter AI for All with a comprehensive range of AI devices, solutions, and concepts.


  • Lenovo ThinkSmart Core Gen 2 Launches with Intel Meteor Lake Processors

    Lenovo ThinkSmart Core Gen 2 Launches with Intel Meteor Lake Processors

    Lenovo has introduced the ThinkSmart Core Gen 2, which at first glance appears to be a sizable mini-PC. Nevertheless, the company asserts that this gadget is an early instance of an ‘AI optimized compute device’ crafted to enhance video conferencing environments. In line with this goal, the ThinkSmart Core Gen 2 operates on Windows 11 IoT, rather than the more common Windows 11 Pro.

    AI Features and Specifications

    Moreover, Lenovo’s promotional materials often highlight the device’s AI features, credited to its specialized NPU. However, this claim holds true primarily due to the incorporation of Intel Meteor Lake processors, specifically the Core Ultra 5 135H and Core Ultra 7 165H equipped with Intel vPro. As a point of reference, the ThinkSmart Core Gen 2 is also fitted with DDR5-5600 RAM and PCIe TLC storage, all housed in a fanless design measuring 185 x 220 x 38 mm and weighing 860 g.

    Connectivity and Software

    In addition, Lenovo has incorporated seven USB ports, three HDMI ports, and a cable management system. Furthermore, every ThinkSmart Core Gen 2 unit will come with ThinkSmart Manager software and Lenovo ThinkShield pre-installed. The ThinkSmart Core Gen 2 is slated to be released later this year, priced at $2,900 in the US, and it will include an IP controller or ThinkSmart controller to accommodate various room configurations.


    Image 1
    Image 1
  • Lenovo Launches AI Buddy: Your Friendly Smart Assistant

    Lenovo Launches AI Buddy: Your Friendly Smart Assistant

    At the 2024 Lenovo Innovation Technology Conference, Lenovo revealed a prototype of a smart home assistant called “AI Buddy.” This device is meant to rival popular smart assistants like Amazon’s Echo, featuring distinct designs and fresh functionalities that differentiate it from others in the market.

    Specifications of Lenovo AI Buddy

    The Lenovo AI Buddy has a unique shape resembling a large, foldable MagSafe charging base. On top, it has a round display that shows animated eyes in the form of emojis to mimic emotions and improve user engagement. The facial expressions are similar to those of the JoyfulRobotics Android Desktop Robot, which was created by a former Xiaomi employee last year. This display also provides essential information such as the time, weather updates, music selections, and personal photos.

    With USB-A and USB-C ports along with a headphone jack at the base, the AI Buddy offers various connectivity options for those who seek versatility in their smart home devices. Its sleek and minimalist design reflects Lenovo’s design philosophy while ensuring that it remains functional in a compact size.

    Partnership Announcement

    During the conference, Lenovo’s CEO Yang Yuanqing unveiled the AI Buddy with Meta’s CEO Mark Zuckerberg, announcing a collaboration between Lenovo and Meta to introduce AI Now, a personal AI assistant tailored for PCs based on Meta’s Llama large model. This advanced technology is integrated into the AI Buddy, providing a more natural and adaptive experience for users. The device utilizes sentiment-based AI to tailor its responses and manage tasks like scheduling, reminders, and daily activities while evolving with user preferences over time.

    The AI Buddy features a rotating display that adjusts to keep eye contact with the user, similar to Amazon’s Echo Show 10. This capability makes interactions feel more engaging and personalized, as the AI Buddy tracks users’ movements within the room.

    Task Management and Security Features

    Thanks to the AI Now integration, the AI Buddy can help users efficiently manage their daily tasks. Lenovo also emphasizes the importance of data security in this device, ensuring that sensitive information is treated with care and that user interactions with the AI Buddy are kept private.

    At the same event, Lenovo showcased other prototypes in addition to the AI Buddy, including the AI Mouse, which features a dedicated AI Now button for seamless integration within Lenovo’s AI ecosystem. Lenovo also hinted at the Lenovo Home AI Brain, aimed at managing, organizing, and protecting family memories by creating AI-generated highlight reels for beloved moments.


    Image 1
  • Samsung to Replace ‘Settings’ App with AI Feature

    Samsung to Replace ‘Settings’ App with AI Feature

    The present movement of bringing AI into smartphones began with simple tasks like text summarization, but it’s advancing rapidly. According to ETNews, a South Korean news outlet, Samsung is developing an AI-driven system that could completely replace the ‘Settings’ app on its Galaxy smartphones.

    Predictive Technology

    This new system aims to anticipate user needs by analyzing real-time interactions. By doing this, it can help users make adjustments quickly without having to delve into the app itself.

    A Growing Trend

    While this may sound a bit like science fiction, many smartphone manufacturers have been heading in this direction for the past few years. For example, smartphones now have cameras that change their image processing based on what they detect—be it a document, landscape, or person. The AI on the device identifies the type of scene you want to capture and automatically makes adjustments.

    There are numerous applications of AI in various apps and user interfaces designed to simplify tasks for users. Samsung, in particular, is working to further elevate the user experience through a deeper integration of AI in its OneUI interface. This latest development aligns with their main strategy.

    Samsung’s Bixby

    AI-enabled voice assistants have been a part of smartphones for quite some time now, yet their capabilities remain quite basic. These smart assistants can help with simpler tasks, like setting alarms, but they often lack comprehensive control over the device. It remains to be seen if this new development will offer a more advanced AI that can handle complex tasks independently.

    Current AI Features in Galaxy Devices

    Samsung’s Galaxy AI currently comes with a range of useful features, such as easier photo editing and facilitating conversations between people who speak different languages. The enhanced "ProVisual Engine" supports the camera, enabling it to produce clearer and more stable images, even in poor lighting.

    Additionally, the AI-powered Note Assist feature streamlines the process of taking notes and retrieving information. While these AI features do save time, their usage is still confined to certain apps and areas of the user interface. Nevertheless, Samsung’s ongoing work could enhance accessibility, leading to a significant improvement in the overall user experience.


    Image 1
  • Open Source AI Video Generator Pyramid Flow Now Online

    Open Source AI Video Generator Pyramid Flow Now Online

    Already gaining traction in YouTube tutorial clips, Pyramid Flow is an innovative AI system trained on freely available datasets, amounting to about 10 million videos. This project is a collaborative effort between AI specialists from Peking University, Kuaishou Technology, and Beijing University of Posts and Telecommunications. Notably, Pyramid Flow is itself open-source. Licensed under the MIT License, it can produce virtual high-resolution (768p) video content, and it particularly excels at 384p. Its developers claim that it can generate a five-second video in under a minute, utilizing an A100 GPU in an unspecified hardware setup.

    Performance Insights

    In various situations, Pyramid Flow performs exceptionally well. Nevertheless, when handling certain text prompts, the output can be inadequate. Like many generative AI tools, there is a degree of unpredictability involved with this model. On the positive side, Pyramid Flow requires significantly less computational power compared to its rivals. Furthermore, since its code is open-source, those who are interested can implement it in local or cloud settings without any licensing concerns.

    Copyright Concerns

    While the AI team behind Pyramid Flow has provided a list of all datasets used for its training, they did not address potential copyright issues that could arise. Some content creators argue that using open-source materials to make virtual videos infringes on the rights of copyright owners. Nevertheless, Pyramid Flow might be beneficial for refining such content without needing to engage third parties.

    Pyramid Flow (on GitHub, via Tech Xplore)

  • Tech Firms Shift from Green Energy as AI Demand Soars

    Tech Firms Shift from Green Energy as AI Demand Soars

    AI usage has rapidly expanded recently, leading tech giants like Microsoft to consider nuclear energy. This shift is driven by the rise of generative AI chatbots, such as OpenAI’s ChatGPT, and integrated AI tools like Microsoft CoPilot in Windows 11. The surge in demand for data center power is so significant that wind and solar energy alone cannot fulfill it.

    Future Power Demand

    According to McKinsey & Company, the demand for power in data centers is expected to rise from 3.7 percent of total power consumption in the US to 11.7 percent by the decade’s end. Morgan Stanley also predicts that global CO2 emissions will increase from 200 million tons to 600 million tons due to the expansion of data centers.

    In Memphis, Tennessee, a data center that trains and runs the Grok 3 AI from X seeks to raise its power needs from 50 MW to 150 MW. This amount of power could supply electricity to around 80,000 homes. Additionally, the facility consumes 30,000 gallons of water daily from underground wells for cooling purposes.

    The Energy Challenge

    The energy requirements of AI models stem from the vast number of calculations needed to answer user queries. Researchers from the University of California, Riverside, in collaboration with the Washington Post, found that generating a simple 100-word email using OpenAI’s GPT-4 AI necessitates a bottle of water for cooling and enough electricity to run 14 light bulbs for an hour.

    Constructing power plants and electrical transmission systems is a slow process. Many energy companies are already dealing with shortages of power distribution units, switchgear, and transformers, leading to delays that can exceed a year. Power generation in various regions near current data centers is either at capacity or nearing it, causing rolling blackouts in areas like California.

    Nuclear Energy as a Solution

    In response, tech firms are increasingly looking to nuclear power to satisfy their electricity needs for AI data centers. These nuclear plants can produce large quantities of energy without requiring as much land as solar and wind farms. Additionally, nuclear energy isn’t reliant on sunlight or wind conditions.

    Microsoft is not only funding the development of a new nuclear power facility but has also invested in restarting a reactor at the notorious Three Mile Island nuclear power plant, which was the site of a nuclear meltdown in 1979. This incident released radioactive gases into the atmosphere, marking it as the most severe nuclear disaster in the US, although it is less catastrophic compared to the Chernobyl and Fukushima disasters.

    Waste Disposal Concerns

    Nuclear power stations in the US produce highly dangerous radioactive waste. Regrettably, the government has yet to determine a long-term disposal solution for this waste following the cessation of funding for the Yucca Mountain nuclear waste repository during the Obama administration.

    For those looking to make a positive impact on the environment, purchasing a solar panel kit (like one available on Amazon) can help charge laptops and phones using solar energy. AI enthusiasts may also consider running AI LLM models on solar-powered laptops at home, instead of relying on nuclear-powered data centers.

    Sources include McKinsey & Company, WSJ, Washington Post, Constellation Energy, MIT Technology Review, Time, CBS Evening News on YouTube, Nuclear Energy Institute, and The Register.


    Image 1
    Image 1
    Image 1
  • Humans Outperform AI, Says Apple-Funded Study

    Humans Outperform AI, Says Apple-Funded Study

    Earlier this month, a group of six AI experts supported by Apple released a study introducing GSM-Symbolic, a new benchmark for AI that "allows for more controllable assessments, giving important insights and more dependable metrics for evaluating the reasoning abilities of models." Unfortunately, it appears that large language models (LLMs) still face significant limitations and are missing even the most fundamental reasoning skills, as shown by initial tests using GSM-Symbolic with AI systems from major companies like Meta and OpenAI.

    Issues with Current Models

    The research pointed out a major issue with current models, which is their lack of consistency when faced with similar questions. The findings indicated that minor changes in wording, which wouldn’t change the meaning for a human, often result in varied responses from AI systems. No specific model was identified as performing notably well.

    The report stated, "In particular, the effectiveness of all models drops [even] when just the numerical values in the question are modified in the GSM-Symbolic benchmark." It also found that "the weakness of mathematical reasoning in these models [shows] that their performance worsens significantly as the number of clauses in a question goes up."

    Study Details

    This 22-page study is accessible here (PDF file). The final two pages include problems with some irrelevant details added at the end, which shouldn’t change the answer for a human. Yet, the AI systems considered these parts, leading to incorrect answers.

    In conclusion, AI systems remain trapped in pattern recognition and still do not possess general problem-solving skills. This year saw the introduction of several LLMs, including Meta AI’s Llama 3.1, Nvidia’s Nemotron-4, Anthropic’s Claude 3, the Fugaku-LLM from Japan (the largest model ever trained solely on CPU power), and Nova by Rubik’s AI, which was launched earlier this month.

    Upcoming Publication

    Tomorrow, O’Reilly will publish the first edition of "Hands-On Large Language Models: Language Understanding and Generation" by Jay Alammar and Maarten Grootendorst. It is priced at $48.99 for the Kindle edition and $59.13 for the paperback version.