Category: Artificial intelligence

  • Elon Musk’s Grok 3 Enters AI Race: Overhyped or Competitive?

    Elon Musk’s Grok 3 Enters AI Race: Overhyped or Competitive?

    Key Takeaways

    1. Claims of Superiority: Elon Musk asserts that Grok 3 is the “smartest artificial intelligence on Earth,” expected to excel in mathematics, science, and programming.

    2. Launch Challenges: Grok 3 faced scrutiny during its launch due to incorrect answers to simple questions, raising questions about its reliability.

    3. Competitive Edge: Despite claims of being “far ahead,” Grok 3’s performance lead over competitors like DeepSeek R1 and GPT-4.0 is only 1-2%.

    4. Resource Investment: Built with 200,000 H100 chips and 200 million training hours, Grok 3’s slight edge raises concerns about the efficiency of such extensive computational resources.

    5. Infrastructure Growth and Funding: xAI is expanding its infrastructure with $10 billion in funding and plans for a large supercomputer cluster, aiming for significant advancements in AI capabilities.

    Elon Musk’s xAI has rolled out Grok 3, which Musk boldly claims is “the smartest artificial intelligence on Earth.” He says it’s a game changer that can outperform all other mainstream AIs in areas like mathematics, science, and programming. Musk also suggested that Grok 3 might help with SpaceX’s Mars mission and could lead to groundbreaking discoveries worthy of a Nobel Prize within three years.

    Launch Event Issues

    Even with its advanced features, Grok 3 encountered some issues during its launch. It struggled to answer a simple AI test question about the difference between “9/11” and “9/9,” which caused some users to doubt its reasoning skills. Furthermore, Grok 3 shared incorrect details about the game Path of Exile 2 during the event, raising alarms about the reliability of its answers.

    Competitive Analysis

    xAI stated that Grok 3 is “far ahead” of its rivals, but a closer look reveals its lead is merely 1-2% over models like DeepSeek R1 and GPT-4.0. Although this ranking places it among the top AI systems, the performance gap is quite minimal.

    Development and Training Details

    Grok 3 was built using 200,000 H100 chips and received 200 million hours of training. Yet, despite these substantial resources, it only slightly surpassed DeepSeek V3, which was trained on 2,000 H800 chips for a mere two months. These findings have sparked discussions about diminishing returns in AI scaling and whether such massive computational investments are efficient.

    Subscription Availability

    Grok 3 can be accessed by X Premium+ subscribers, and xAI has unveiled a new subscription tier named SuperGrok for mobile app and Grok.com users. This aligns with the trend of AI companies providing tiered access to their models.

    New Features and Updates Ahead

    One of the standout features of Grok 3 is DeepSearch, a sophisticated tool aimed at improving search, research, and data analysis abilities. xAI has also revealed intentions to incorporate voice interaction in future updates, making user interactions more intuitive and friendly.

    Financial Backing and Infrastructure Growth

    xAI is ramping up its infrastructure by securing $10 billion in funding, potentially valuing the company at around $75 billion. The firm is working on a supercomputer cluster named “Colossus” in Memphis, Tennessee, expected to be one of the largest globally. Additionally, xAI is negotiating with Dell Technologies for a $5 billion purchase of servers using Nvidia GB200 chips.

    Musk’s Attempt to Acquire OpenAI

    In a significant move, Musk along with a group of investors proposed $97.4 billion to take over OpenAI’s nonprofit assets. However, OpenAI’s CEO Sam Altman and the board turned down this offer, highlighting the fierce competition and strategic maneuvering within the AI sector.

    Grok 3 represents a crucial advancement for xAI in the crowded AI landscape. Although it has introduced new features and secured substantial funding, its performance during the launch has raised concerns about its actual effectiveness. As xAI works on improving the model and enhancing its infrastructure, only time will tell if Grok 3 can fulfill its lofty goals.

    Tags: , ,
  • Humane AI Pin to Lose Most Functions After HP Acquisition

    Humane AI Pin to Lose Most Functions After HP Acquisition

    Key Takeaways

    1. Humane is being sold to HP for $116 million, a significant loss compared to the $230 million raised in venture capital.
    2. The Humane Ai Pin will lose nearly all functionalities on February 28, as it will no longer connect to company servers.
    3. The Humane.Center website, essential for managing the Ai Pin, will also be shut down, rendering the device nearly useless.
    4. Customers who purchased the Ai Pin within the last 90 days can request a refund due to its loss of functionality.
    5. HP acquires over 300 patents and CosmOS, Humane’s AI operating system, while Humane’s co-founders will lead a new AI department at HP.

    Humane has revealed in a press statement that it will be sold to HP for $116 million. Since the company raised at least $230 million in venture capital, this deal showcases a major financial setback for Humane’s creators. Additionally, the customers of the company will also feel the impact of this acquisition, as the Humane Ai Pin is set to lose nearly all its functionalities on February 28.

    Loss of Functionality

    Starting from that date, the Humane Ai Pin will no longer connect with the company’s servers. The Humane.Center website, which is essential for managing the Pin, will also be shut down. As a result, this pricey device will be unable to make calls, send texts, respond to inquiries, or access cloud information, turning it into an attractive piece of electronic waste, though it will still show its battery level. Customers who bought the Ai Pin within the last 90 days can ask for a refund.

    A Shift in Direction

    Humane has been searching for a buyer for several months after the Ai Pin turned out to be a market failure. Initially, the company aimed for a selling price of over one billion dollars; however, HP managed to slash this estimate by nearly 90 percent. In addition to acquiring the Ai Pin, HP will also gain over 300 patents and pending patent applications, along with CosmOS, Humane’s AI operating system intended to enhance cars and smart home gadgets in the future. The two co-founders of Humane, Imran Chaudhri and Bethany Bongiorno, will lead a new AI department at HP.

    Source:
    Link

    Tags: , , ,
  • Youdao SpaceOne Launches DeepSeek AI Dictionary Pen

    Youdao SpaceOne Launches DeepSeek AI Dictionary Pen

    Key Takeaways

    1. AI-Powered Learning: The SpaceOne dictionary pen uses the DeepSeek-R1 model to provide detailed explanations and logical reasoning, enhancing the learning experience beyond simple translations.

    2. High-Quality Display: It features a 4.4-inch ultra-wide OLED display with low blue light emission, ensuring clear reading and comfortable use during long study sessions.

    3. Advanced Scanning Technology: Equipped with an 8MP AI camera, the pen can scan complex texts and problems, providing thorough breakdowns and supporting multi-language learning.

    4. Storage and Connectivity: The device has 64GB of internal storage, built-in 4G connectivity, and dual-band Wi-Fi, allowing for easy access to AI tutoring and translation services on the go.

    5. Ergonomic Design and Parental Controls: The SpaceOne is designed for portability with an ergonomic sliding body and pressure-sensitive tip, plus it includes a parental control mode for monitoring learning progress.

    Youdao has launched the SpaceOne dictionary pen, a learning device powered by AI, aimed at students and professionals. It uses the DeepSeek-R1 reasoning model, providing thorough, step-by-step explanations rather than simple translations. The pen is available for 1,699 yuan ($234) and can be purchased on JD.com.

    SpaceOne Features

    The SpaceOne comes with a 4.4-inch (1210×568 resolution) ultra-wide display that not only provides a clear reading experience but also simplifies navigation through scanned text and AI-generated explanations. With TÜV Rheinland certification for low blue light emission, it ensures comfort during long study hours. The device has an OLED panel, offering vivid colors and sharp text, making scanned documents and AI explanations clearer compared to traditional dictionary pens.

    Furthermore, it has built-in 4G connectivity, which enables users to access AI tutoring and translation services without needing Wi-Fi. For those who like wireless connections, it features dual-band Wi-Fi (2.4G/5G) for a quick and stable link. Inside, it houses a 1650mAh battery that charges via USB-C.

    Advanced Scanning Capabilities

    The SpaceOne is equipped with an 8MP AI camera, which can accurately capture whole sections of text, complex math problems, and graphical questions. Once the text is scanned, the DeepSeek-R1 AI model kicks in, providing detailed breakdowns and logical explanations. This innovative reasoning system distinguishes it from regular dictionary pens that typically only offer direct translations or basic word definitions.

    Moreover, it facilitates multi-language learning through real-time voice recognition, translation in over 100 languages, AI pronunciation correction, and a virtual speaking coach. The SpaceOne also provides advanced reasoning, instant translation, and AI-powered pronunciation coaching, enhancing the interactive learning experience.

    Storage and Portability

    The pen features 64GB of internal storage, allowing users to easily save scanned text, study notes, and AI-generated explanations. For audio learning, it supports Bluetooth for wireless headphones and includes a Type-C wired earphone option for a traditional setup.

    When it comes to portability, the SpaceOne has an ergonomic sliding body design, which makes it comfortable to hold and easy to transport. It also has a pressure-sensitive wide pen tip, letting users scan both full lines and single words effortlessly. Additionally, it comes with an intelligent parental control mode, enabling parents to keep track of learning progress and limit access to certain content.

    Xiaomi also released a comparable pen last year, known as the Mijia Dictionary Pen C1, which features a 3.02-inch display, 26 million word entries, and 0.3-second OCR recognition with 98% accuracy. It supports offline searches, AI learning strategies, and a battery life of 100 days, making it a versatile tool for students.

    Youdao SpaceOne Launches DeepSeek AI Dictionary PenYoudao SpaceOne Launches DeepSeek AI Dictionary Pen
    Tags: , ,
  • Live Demo of xAI’s Grok 3 as AI Integrates with Tesla Today

    Live Demo of xAI’s Grok 3 as AI Integrates with Tesla Today

    Key Takeaways

    1. Tesla vehicles are becoming more advanced with the integration of xAI’s Grok 3, enhancing their AI capabilities beyond just being “computers on wheels.”
    2. Grok 3 uses real-time data from the X social media platform to provide up-to-date information and news in its responses.
    3. New features in Grok include self-correction to fix errors, reinforcement learning to reduce incorrect responses, and the use of synthetic datasets for faster training.
    4. The introduction of a live-voice mode aims to improve Tesla’s voice command system and make interactions more natural and intuitive.
    5. Grok will work closely with Tesla’s Full Self-Driving (FSD) technology, enhancing self-driving capabilities and enabling complex inquiries like weather updates.

    Elon Musk has often said that his vehicles are basically computers on wheels. However, they are turning out to be even more sophisticated than just that. Soon, Tesla cars will be equipped with the latest version of xAI’s Grok, and there’s a live demonstration of this AI system scheduled for today.

    Major Improvements with Grok 3

    Grok 3 is set to bring major enhancements following xAI’s updates to its Large Language Model (LLM) training techniques. The AI now utilizes real-time data from posts on the X social media platform to formulate responses, allowing it to provide up-to-the-minute news.

    New Features of Grok

    Among the new features in Grok is self-correction, which allows it to identify and rectify its own errors. There’s also reinforcement learning, which employs rewards and penalties to minimize incorrect responses. Additionally, synthetic datasets or artificially created data will help speed up the training process.

    Exciting Live-Voice Mode

    One of the most notable upgrades is the introduction of live-voice mode, an aspect where Grok previously fell short compared to its competitors. This feature is crucial for integrating Grok within Tesla vehicles.

    Grok is anticipated to improve the current voice command system in Tesla, with a demonstration possibly happening tonight. The AI model will closely connect with Full Self-Driving (FSD), allowing Grok to gather real-world data for self-driving capabilities. It should also facilitate more natural general knowledge inquiries, like asking about the weather or handling other complicated requests.

    Other electric vehicle manufacturers are also integrating AI technology, such as the Chinese companies Geely and Voyah, which are working with DeepSeek.

    Source:
    Link

    Tags: , ,
  • South Korea Halts DeepSeek AI Downloads Over Privacy Concerns

    South Korea Halts DeepSeek AI Downloads Over Privacy Concerns

    Key Takeaways

    1. South Korea has halted new downloads of the Chinese AI chatbot DeepSeek due to data privacy concerns, effective February 15, 2025.
    2. The Personal Information Protection Commission (PIPC) found weaknesses in DeepSeek’s communication features and management of personal information with third parties.
    3. Users are advised to refrain from entering personal information into the chatbot until issues are resolved and compliance with South Korean laws is ensured.
    4. Other countries, including the U.S., Italy, and Australia, have also imposed restrictions on DeepSeek due to security threats and excessive data collection.
    5. DeepSeek faces significant challenges in rebuilding trust and complying with regulations as global scrutiny of its data handling practices increases.

    South Korea has put a stop to new downloads of the Chinese AI chatbot DeepSeek, citing worries about data privacy breaches. This decision was made public by the Personal Information Protection Commission (PIPC) and became effective on February 15, 2025, at 6:00 p.m. local time. Although the app is no longer available for download, users can still access the web version while the company works on fulfilling the necessary regulatory conditions.

    South Korea Takes Action Against DeepSeek AI

    The PIPC stated that its investigation into DeepSeek, which began shortly after the chatbot’s introduction, revealed weaknesses in its communication features and how it manages personal information with third-party providers. The commission made it clear that until these problems are fixed in accordance with South Korea’s Personal Information Protection Act, new downloads of the app will not be permitted. Current users have been told to avoid entering any personal information into the chatbot’s prompts until further notice.

    DeepSeek has admitted that it did not adequately consider South Korean data protection laws prior to its launch and has since appointed a local representative to help ensure compliance with regulations. The government intends to use this situation as a model to enhance guidance and oversight, aiming to prevent similar issues from happening again in the future.

    Increased Scrutiny and Security Concerns

    The suspension comes after increased scrutiny from South Korea’s National Intelligence Service (NIS), which had previously alerted that DeepSeek was collecting user data excessively and might be using it to train its AI models. Furthermore, security analysts found that both the Android and iOS versions of the app were sending certain user data to its servers without encryption, raising even more red flags.

    This situation is part of a larger trend, as global concerns about DeepSeek’s data handling practices have surged. Countries like the United States, Italy, Australia, and Taiwan have already placed restrictions on the AI service for government use, citing security threats. NASA has blocked DeepSeek from its systems, and the U.S. Navy has cautioned its personnel against using the app due to potential data risks.

    Challenges Ahead for DeepSeek

    In the midst of these escalating worries, Beijing claims that it allows international internet firms to operate in China while adhering to local laws and asserts it does not force companies to unlawfully collect or store data. However, as regulatory scrutiny grows around the globe, DeepSeek now faces significant challenges in rebuilding trust and adhering to regulations in various regions.

    Source:
    Link

    Tags: , ,
  • Honor to Present Alpha Plan at MWC 2025 Focusing on AI

    Honor to Present Alpha Plan at MWC 2025 Focusing on AI

    Key Takeaways

    1. Honor is set to unveil the “Honor Alpha Plan” at Mobile World Congress 2025, focusing on AI advancements and ecosystem expansion.
    2. The Alpha Plan emphasizes leadership and market supremacy in AI-driven consumer technology, though specific details about products remain unclear.
    3. Honor aims to create an open and collaborative AI ecosystem, indicating a shift beyond smartphones to include wearables and smart home devices.
    4. There are no announcements for new flagship smartphones, suggesting a focus on AI features and software at the upcoming event.
    5. The keynote event will take place on March 2, 2025, in Barcelona, with Honor also showcasing products at their stand during MWC.

    Honor is preparing for the Mobile World Congress 2025, and they’re not just introducing a new gadget. The firm has scheduled its keynote event for March 2 in Barcelona, where it plans to reveal the “Honor Alpha Plan,” a comprehensive corporate strategy focusing on AI advancements and expanding its ecosystem.

    Vague Details Ahead

    Even with the bold branding, specifics are still unclear. Honor refers to “Alpha” as a representation of leadership, top performance, and market supremacy. They believe they will play a crucial role in the AI-led future of consumer technology. Honor states that this plan shows their dedication to pushing the limits in artificial intelligence, yet it’s uncertain what this will mean regarding actual products.

    Focusing on AI, Not Phones

    Honor has consistently highlighted AI in its recent product launches, and the Alpha Plan appears to enhance that approach. The brand aims to position itself as a leading force in AI-driven consumer electronics, aspiring to create an “open and collaborative AI ecosystem” with international partners. This hints at a venture beyond just smartphones, likely extending into wearables, smart home devices, and other IoT products.

    No Flagship Phones in Sight

    However, one glaring omission from Honor’s announcement is any reference to new flagship smartphones. Typically, major smartphone releases get more attention, so it’s possible that Honor is holding back those announcements for a future event or prioritizing this keynote on software, AI features, and ecosystem connectivity.

    The event is set to happen at the Hyatt Regency Barcelona Tower on March 2 at 16:30 CET. For those attending MWC, the company will also showcase its products at Stand 3H10 in Hall 3 at Fira Gran Via, where visitors can check out the latest offerings.

    Whether the Alpha Plan is a revolutionary strategy or just another piece of marketing jargon will become clear in time. What’s evident is that Honor is eager to present itself as a significant player in AI. Now they just need to back it up with real technological advancements.

    Tags: , ,
  • Perplexity AI Launches Deep Research Tool to Compete with OpenAI, Google

    Perplexity AI Launches Deep Research Tool to Compete with OpenAI, Google

    Key Takeaways

    1. Introduction of Deep Research Tool: Perplexity AI has launched “Deep Research,” a tool designed to provide expert insights quickly and at a lower cost than competitors like OpenAI.

    2. Cost and Accessibility: Unlike OpenAI’s tool priced at $200/month, Perplexity offers a free version with limitations and a pro version for $20/month, allowing users to make up to 500 inquiries daily.

    3. Human-like Research Emulation: The Deep Research tool mimics human research processes, delivering coherent, citation-supported reports in under three minutes across various subjects.

    4. Performance Metrics: Perplexity’s tool scored 21.1% on “Humanity’s Last Exam,” outperforming some models but still behind OpenAI’s 26.6% score.

    5. Legal Challenges and Partnerships: Perplexity faces legal issues over content usage but is forming partnerships with media publishers to establish revenue-sharing agreements.

    Perplexity AI is making a significant commitment to the future of AI research with its new “Deep Research” tool, which aims to deliver expert insights at a faster pace than ever. Founded by Aravind Srinivas, a former researcher at OpenAI, the company has rapidly established itself in the AI landscape, receiving support from big names like Jeff Bezos and Nvidia. Now, it’s directly challenging OpenAI’s expensive research assistant by offering a freemium option that focuses on being user-friendly and quick.

    Deep Research: Expert Analysis in Minutes

    In contrast to OpenAI’s Deep Research, which comes with a steep $200 monthly fee, Perplexity’s tool is free to use, though there are some restrictions. Users who opt for the pro version, costing $20 monthly, can access up to 500 deep research inquiries each day. The web-based AI assistant is also set to launch on iOS, Android, and Mac soon, making comprehensive research easily accessible with just a few clicks.

    Mimicking Human Research

    Deep Research is fundamentally designed to emulate a human’s approach to understanding complicated topics—searching, reading, and refining information before producing a coherent, citation-supported report. It covers a wide array of subjects, including finance, marketing, healthcare, and product research. Notably, it generates results in less than three minutes, which is a stark improvement compared to OpenAI’s tool, requiring five to thirty minutes for similar outcomes.

    Performance and Legal Challenges

    On the performance front, Perplexity’s Deep Research tool achieved a score of 21.1% on “Humanity’s Last Exam,” a benchmark for expert analysis. This result surpasses models like Gemini Thinking (6.2%) and GPT-4o (3.3%), yet it still trails behind OpenAI’s score of 26.6%. Despite these advancements, Perplexity is facing increasing legal challenges, having been sued by major media houses over claims of unapproved content usage. In response, the company is forming partnerships with publishers such as Time and Fortune to create revenue-sharing arrangements.

    As AI research tools continue to evolve, Perplexity’s focus on speed, cost-effectiveness, and user accessibility could position it as an attractive option for users who prefer not to spend large amounts on high-end AI solutions.

    Source:
    Link

    Perplexity AI Launches Deep Research Tool to Compete with OpenAI, Google
    Tags: , ,
  • Meta’s AI Achieves 80% Accuracy in Mind Reading Technology

    Meta’s AI Achieves 80% Accuracy in Mind Reading Technology

    Key Takeaways

    1. Meta’s AI can reconstruct sentences from brain activity with 80% accuracy, aiding those who have lost speech.
    2. The research uses non-invasive methods (MEG and EEG) to capture brain activity without surgery.
    3. Limitations include the need for a magnetically shielded environment and the requirement for participants to remain still.
    4. The AI helps understand how the brain translates thoughts into language, revealing a ‘dynamic neural code.’
    5. Meta is investing in further research with a $2.2 million donation and partnerships with various European institutions.

    Meta’s AI research team is making strides in understanding human thoughts. In partnership with the Basque Center on Cognition, Brain, and Language, the company has created an AI model that can reconstruct sentences from brain activity with accuracy reaching 80%. This research uses a non-invasive method for recording brain activity and, as stated by the company, could lead to technology that assists those who have lost the ability to speak.

    The Technology Behind It

    Differing from current brain-computer interfaces that typically need invasive procedures, Meta employs magnetoencephalography (MEG) and electroencephalography (EEG). These methods capture brain activity without any surgery involved. The AI model was trained on recordings from 35 participants while they typed sentences. When faced with new sentences, Meta asserts that it can predict up to 80% of the typed characters using MEG data—this is at least double the effectiveness of EEG-based decoding.

    Limitations and Challenges

    However, there are certain limitations to this approach. MEG necessitates a magnetically shielded environment, and participants have to remain completely still for precise readings. Furthermore, this technology has only been evaluated on healthy individuals, leaving its performance for those with brain injuries uncertain.

    Understanding Word Formation

    In addition to decoding thoughts into text, Meta’s AI is assisting researchers in comprehending how the brain converts ideas into language. The AI model scrutinizes MEG recordings, observing brain activity in milliseconds. It uncovers how the brain changes abstract thoughts into words, syllables, and even the movements of fingers while typing.

    A significant discovery is that the brain utilizes a ‘dynamic neural code,’ which connects various stages of language creation while keeping previous information readily accessible. This might shed light on how individuals effortlessly construct sentences while communicating.

    Meta’s research reinforces the idea that AI could eventually facilitate non-invasive brain-computer interfaces for those unable to communicate verbally. Nevertheless, the technology is not yet ready for practical application. There is a need for improved decoding accuracy, and the hardware requirements of MEG limit its usability outside of laboratory environments.

    Meta is committed to fostering this research by forming partnerships. The company has pledged a donation of $2.2 million to the Rothschild Foundation Hospital to aid ongoing research. It is also collaborating with institutions such as NeuroSpin, Inria, and CNRS in Europe.

    Source:
    Link

    Meta's AI Achieves 80% Accuracy in Mind Reading TechnologyMeta's AI Achieves 80% Accuracy in Mind Reading Technology
    Tags: , ,
  • Perplexity Launches Deep Research Feature for Free AI Chatbot

    Perplexity Launches Deep Research Feature for Free AI Chatbot

    Key Takeaways

    1. Perplexity has launched a new Deep Research feature for its AI chatbot, allowing it to answer complex questions by finding real-time information online.
    2. Traditional AI chatbots are limited by their training data and struggle with intricate inquiries that require research and data analysis.
    3. The Deep Research feature prioritizes accuracy over speed, improving the AI’s ability to tackle complicated prompts effectively.
    4. Perplexity AI has achieved over 90% accuracy in factual questions and excels in advanced inquiries, outperforming many competitors.
    5. Unlike OpenAI, Perplexity offers the Deep Research feature for free to all users, with options for both limited daily prompts for free users and unlimited access for paid users.

    Perplexity has introduced a new Deep Research feature to its complimentary AI chatbot service. This capability enables the AI to effectively respond to intricate inquiries by scouring the Internet for the most current information as it tackles issues in a step-by-step manner, similar to how a human might approach them.

    Traditional Limitations of AI Chatbots

    AI chatbots such as OpenAI ChatGPT, xAI Grok, Google Gemini, Meta AI, Anthropic Claude, and Perplexity AI have historically relied on a predetermined set of training documents. They generate responses based on their learned knowledge. However, this approach restricts their understanding to the time of their training, leaving them oblivious to any events or updates that happen afterward.

    Moreover, these AI systems tend to provide the most probable answers based on their existing knowledge base, which hinders them from addressing intricate questions that necessitate research abilities, including data integration and analysis to formulate responses.

    Transitioning to Deep Research

    The introduction of Deep Research shifts Perplexity AI’s priority from speed to accuracy, compelling the AI to invest more time contemplating how to approach a complicated prompt and the necessary steps to resolve it. Concurrently, it can access real-time Internet data to gather the latest pertinent information it requires.

    Consequently, Perplexity AI has excelled in the SimpleQA AI benchmark test, achieving over 90% accuracy in answering factual questions. Additionally, it performs better than many other AI systems in the PhD-level inquiries of Humanity’s Last Exam benchmark, with the exception of the new OpenAI o3 that also includes deep research capabilities.

    Access and Availability

    In contrast to OpenAI, which restricts its deep research features to paying customers of ChatGPT, Perplexity offers Deep Research at no cost to all users as of now. Users can easily switch from Auto mode to Deep Research mode before posing their questions to utilize this functionality. While free users are limited in the number of prompts they can submit daily, paid users enjoy unrestricted access.

    Perplexity AI with Deep Research is currently accessible for free on the website, and this feature will soon be incorporated into the company’s mobile applications.

    Source:
    Link

    Perplexity Launches Deep Research Feature for Free AI ChatbotPerplexity Launches Deep Research Feature for Free AI Chatbot
    Tags: , ,
  • Topaz Labs Introduces Project Starlight AI in Video AI Software

    Topaz Labs Introduces Project Starlight AI in Video AI Software

    Key Takeaways

    1. Topaz Labs is introducing generative AI video technology to enhance low-resolution videos with realistic details.
    2. The generative AI model recognizes intricate details, improving the quality of upscaled videos beyond what standard algorithms can achieve.
    3. Project Starlight uses an AI diffusion model to reconstruct missing details during the video upscaling process, resulting in high-resolution outputs without artifacts.
    4. The new AI feature will be available as a free research preview, allowing limited processing of clips initially, with plans for expanded features and resolutions in the future.
    5. Current Video AI software is priced at $299, but specific pricing for Project Starlight will be announced later.

    Topaz Labs is set to introduce generative AI video technology into its Video AI software. This new feature will enable users to enhance low-resolution videos, elevating them to higher resolutions while incorporating realistic details that match the objects within the scenes.

    Understanding Generative AI

    The generative AI model has been trained using a vast collection of images from the world. This enables it to recognize the intricate details that are found in reality. For instance, it understands that a cat possesses delicate whiskers near its nose. Such tiny details are typically missing in low-resolution videos, and standard upsampling algorithms fail to recreate these lost elements.

    The Power of Project Starlight

    Project Starlight utilizes an AI diffusion model for video, allowing it to reconstruct anticipated missing details during the upscaling process. This results in high-resolution videos that are brimming with detail. Additionally, this feature functions without artifacts, avoiding issues like incorrect object motions, which can arise when AI fails to generate realistic outcomes.

    Availability and Pricing

    This AI feature will be launched next week as a free research preview within the Video AI software. At first, users will be limited to processing three ten-second clips weekly at 2K resolution for free, while longer clips, up to five minutes in duration, will incur a fee. Eventually, the complete feature set, which will include the ability to output at 4K resolution, will be made available.

    Currently, the Video AI version 6 is priced at $299, but details about the pricing for the new Project Starlight will be announced later.

    Source:
    Link

    Tags: , ,