News/AI Models
Google’s NotebookLM Turns Text to Podcasts — Here’s How to Try It
Google's NotebookLM: A new AI tool for content creators: Google has introduced NotebookLM, an innovative AI research assistant that transforms written content into realistic audio conversations, mimicking the format of podcast discussions. Key features and functionality: NotebookLM analyzes uploaded articles and generates an audio conversation that sounds like two podcast hosts discussing the content, offering a novel way to engage with written material. The tool accepts various input formats, including articles, PDFs, and plain text. Users can upload multiple sources to create a more comprehensive discussion. The generated audio currently uses two American-accented voices, though customization options are limited. The...
read Sep 22, 2024What It Really Means for Advanced AI Models to ‘Reason’
AI reasoning breakthrough: OpenAI's latest large language model, o1 (nicknamed Strawberry), represents a significant advancement in artificial intelligence capabilities, particularly in its ability to reason and "think" before providing answers. O1 is the first major LLM to incorporate a built-in "think, then answer" approach, moving beyond the limitations of previous models that often produced contradictory or inconsistent responses. This new model demonstrates markedly improved performance on challenging tasks across various fields, including physics, chemistry, biology, mathematics, and coding. The enhanced reasoning ability of o1 is achieved through a technique similar to chain-of-thought prompting, which encourages the model to show its...
read Sep 22, 2024Sam Altman Says Next-Gen AI Agents Could Invent Breakthroughs, Cure Diseases
Advancing AI autonomy: OpenAI CEO Sam Altman has announced that the next generation of AI systems will be capable of performing tasks independently, without human input, marking a significant leap in artificial intelligence capabilities. This development is being driven by new models like OpenAI's o1, which possesses the ability to "reason" and work through problems before presenting solutions. Altman discussed these advancements during a recent T-Mobile Capital Markets Day fireside chat, highlighting the potential for "level 3 AI" or "agentic systems" that can act autonomously to achieve optimal responses. The evolution of AI is expected to progress to "level 4"...
read Sep 22, 2024Mistral, Google and the Spectrum of ‘Openness’ in AI Models
The rapid advancement of artificial intelligence has sparked debates about the transparency and accessibility of AI models, highlighting the need for a clearer understanding of openness in the field. Recent developments in AI openness: Google and Mistral AI have taken divergent approaches to releasing their AI models, showcasing the varying degrees of accessibility in the industry. Google's Gemini release was accompanied by significant publicity but offered limited testing options, primarily through integration with Bard. Mistral AI quietly shared a Magnet link to one of its models, allowing skilled users to download, use, and fine-tune the model without fanfare. The contrast...
read Sep 21, 2024How AI is Reshaping Journalism Beyond Misinformation Concerns
AI's multifaceted impact on journalism: Artificial intelligence is transforming the news industry beyond just misinformation concerns, enhancing various aspects of journalism from content creation to operational efficiency. While AI-powered deepfakes and bot postings remain significant issues, the technology is being applied more broadly to improve journalistic practices and news delivery. AI's role in journalism extends to automated news writing, content personalization, fact-checking, and streamlining newsroom operations. Automated news writing: AI systems, particularly large language models, are increasingly being used to generate news articles on a variety of topics, addressing resource constraints in newsrooms. AI excels at aggregating large amounts of...
read Sep 21, 2024Researchers Develop AI Models Enabling Robots to Adapt to New Environments
Robotic adaptability breakthrough: Researchers have developed AI models that enable robots to perform tasks in new environments without additional training, potentially revolutionizing the field of robotics and home automation. A team from New York University, Meta, and Hello Robot created five "robot utility models" (RUMs) that allow machines to complete basic tasks in unfamiliar settings with a 90% success rate. The tasks include opening doors and drawers, and picking up tissues, bags, and cylindrical objects. This approach could make it easier and more cost-effective to deploy robots in homes in the future. Data collection innovation: The researchers developed a novel...
read Sep 20, 2024LinkedIn is Training its AI Models on Your Data — Here’s How to Opt Out
LinkedIn's AI training initiative: LinkedIn has implemented a new policy that allows the company to use user data for training generative AI models, with users automatically opted in without explicit consent. The professional networking platform introduced a new privacy setting and opt-out form before updating its privacy policy to reflect this change. LinkedIn states that it uses generative AI for features such as writing assistance, but the extent of data usage and potential applications remains unclear. This move follows a recent admission by Meta that it has been scraping non-private user data for AI model training since 2007. Opting out...
read Sep 20, 2024New AI Model from MIT Reveals the Structures of Crystalline Materials
AI breakthrough in crystallography: MIT chemists have developed a new generative AI model called Crystalyze that can determine the structures of powdered crystalline materials from X-ray diffraction data. The model could significantly accelerate materials research for applications in batteries, magnets, and other fields by solving structures that have remained unsolved for years. Crystalyze uses machine learning trained on data from the Materials Project database, which contains information on over 150,000 materials. The AI model breaks down the structure prediction process into subtasks, including determining lattice size and shape, atom composition, and atomic arrangement within the lattice. How Crystalyze works: The...
read Sep 20, 2024The Brains and Brawn of AI Models and How to Understand Their Output
Recent insights from a talk by Devavrat Shah shed light on conceptual frameworks for understanding and regulating artificial intelligence systems. The mind and muscle of AI: Cognitive output, whether from humans or AI, can be viewed as a combination of learning capability (mind) and mechanistic automation (muscle). The 'mind' component represents the learning aspect, involving data interpretation and logical reasoning. The 'muscle' refers to the brute-force application of assessment to data, or what Shah terms 'mechanistic automation'. This conceptual framework helps in distinguishing between AI systems that simply process large amounts of data and those that demonstrate more sophisticated learning...
read Sep 20, 2024Microsoft’s New GRIN-MoE AI Model Excels at Math and Coding
Microsoft's GRIN-MoE AI model has emerged as a powerful contender in the field of artificial intelligence, particularly excelling in coding and mathematical tasks while offering enhanced scalability and efficiency for enterprise applications. Innovative architecture and approach: GRIN-MoE, which stands for Gradient-Informed Mixture-of-Experts, employs a novel technique to selectively activate only a small subset of its parameters at a time, resulting in improved performance and resource efficiency. The model uses a Mixture-of-Experts (MoE) architecture, routing tasks to specialized "experts" within the system. GRIN-MoE utilizes SparseMixer-v2 to estimate the gradient for expert routing, overcoming traditional challenges in MoE architectures. With 16×3.8 billion...
read Sep 20, 20241X’s New AI Model Uses Generative World Data to Train Robots
Robotics breakthrough in generative world models: 1X Technologies has developed a new generative model to enhance robot training efficiency in simulated environments, addressing a key challenge in robotics. The model aims to bridge the "sim2real gap" by learning to simulate the real world using raw sensor data collected directly from robots. This approach differs from traditional methods that rely on hand-authored digital twins and rigid body simulators, which often have inaccuracies in physics and geometry. The data used to train the model comes from thousands of hours of video and actuator data collected from 1X's EVE humanoid robots performing various...
read Sep 19, 2024How Data, Machine Learning and AI are Transforming Industries
The digital age love triangle: Data, machine learning, and artificial intelligence form a powerful alliance that is reshaping industries and driving innovation across various sectors. This unique relationship has the potential to revolutionize decision-making processes, unlock hidden insights, and tackle complex challenges in ways previously thought impossible. The synergy between these three elements is increasingly becoming the cornerstone of progress in fields such as healthcare, finance, marketing, and transportation. Data as the foundation: Data serves as the critical base upon which machine learning and artificial intelligence build their capabilities, providing the raw material for analysis and decision-making. The growing recognition...
read Sep 19, 2024Why API Launches from AI Video Startups Runway and Luma Are a Big Deal
AI video generation reaches new milestone: Runway and Luma Labs, two leading AI video generation companies, have released APIs for their respective platforms, potentially revolutionizing the way developers and consumers interact with AI-generated video content. Key developments: Runway Gen-3 Alpha Turbo and Luma Dream Machine APIs are now available in limited access, with more models expected to be supported in the coming months. The APIs will allow developers to integrate generative video AI models into their own applications, significantly expanding the reach and potential use cases for AI-generated video. This development could lead to innovative applications, such as Chrome extensions...
read Sep 19, 2024Hollywood Giant Lionsgate to Provide Library to Runway for AI Training
Landmark AI partnership reshapes Hollywood: Lionsgate and Runway have announced a groundbreaking collaboration to create an AI model trained on Lionsgate's extensive film and TV library, potentially transforming content creation in the entertainment industry. The deal's core components: This partnership aims to provide Runway with legally clear training data while equipping Lionsgate with advanced tools to enhance content creation and potentially reduce production costs. Lionsgate, known for franchises like John Wick and The Hunger Games, views AI as a means to boost efficiency in content production. Michael Burns, Lionsgate's vice chair, emphasized the potential for developing "cutting edge, capital efficient...
read Sep 18, 2024Scientists are Designing “Humanity’s Last Exam” to Assess Powerful AI
AI experts launch unprecedented challenge for advanced artificial intelligence: Scientists are developing "Humanity's Last Exam," a comprehensive test designed to evaluate the capabilities of cutting-edge AI systems and those yet to come. The initiative's scope and purpose: The Center for AI Safety (CAIS) and Scale AI are collaborating to create the "hardest and broadest set of questions ever" to assess AI capabilities across various domains. The test aims to push the boundaries of AI evaluation, going beyond traditional benchmarks that recent models have easily surpassed. This project comes in response to rapid advancements in AI, such as OpenAI's new o1...
read Sep 18, 2024Hume Launches ‘EVI 2’ AI Voice Model with Emotional Responsiveness
Hume's EVI 2: A leap forward in AI voice technology: Hume, an AI startup, has unveiled Empathic Voice Interface 2 (EVI 2), a significant upgrade to its AI voice model and API, offering enhanced naturalness, emotional responsiveness, and customizability. Key improvements and features: EVI 2 brings substantial enhancements over its predecessor, addressing critical aspects of AI voice interaction. The new version boasts a 40% reduction in latency, with response times ranging from 500 to 800 milliseconds, greatly improving real-time conversation capabilities. EVI 2 demonstrates improved emotional intelligence, better understanding and responding to the emotional context of user inputs. The system...
read Sep 18, 2024Tencent, Johns Hopkins Unveil New Text-to-Audio AI Model
Breakthrough in AI-generated audio: Tencent AI Lab and Johns Hopkins University researchers have unveiled EzAudio, a revolutionary text-to-audio (T2A) generation model that produces high-quality sound effects from text prompts with remarkable efficiency. Key innovations driving EzAudio: The model's architecture, called EzAudio-DiT (Diffusion Transformer), introduces several technical advancements to enhance performance and efficiency. EzAudio operates in the latent space of audio waveforms, departing from traditional spectrogram-based methods and eliminating the need for a neural vocoder. The model incorporates a new adaptive layer normalization technique called AdaLN-SOLA, long-skip connections, and advanced positioning techniques like Rotary Position Embedding (RoPE). These innovations allow for...
read Sep 18, 2024Lionsgate Runway Deal Is a Test of Hollywood’s Relationship With AI
AI enters Hollywood's mainstream: Lionsgate, the studio behind major franchises like John Wick and The Hunger Games, has partnered with AI startup Runway to create a custom AI tool for film and TV production. The partnership grants Runway access to Lionsgate's archives to develop an AI model for pre- and post-production work, potentially revolutionizing aspects of the filmmaking process. This move marks a significant step in the integration of AI technology into mainstream Hollywood production, potentially setting a precedent for other studios to follow. The tool aims to assist filmmakers, directors, and creative talent in augmenting their work and generating...
read Sep 18, 2024Library of Congress Is a Go-To Data Source for Companies Training AI Models
The Library of Congress: A new frontier for AI development: The world's largest library has become an attractive resource for AI companies seeking to train their advanced language models using its vast digital archives. The Library of Congress (LOC) houses 180 million items, including books, manuscripts, maps, and audio recordings, with 185 petabytes of digital data. AI companies are increasingly interested in accessing this data to develop and train their most sophisticated AI models. The library's digital collections offer rare, original, and authoritative information in over 400 languages, spanning various disciplines. Surge in data access requests: The Library of Congress...
read Sep 18, 2024Researchers Develop AI Model that is a ‘Shazam’ for Whale Calls
AI-powered whale tracking breakthrough: Researchers have identified the source of a mysterious underwater sound and developed an AI model to track elusive whale species in the Pacific Ocean. The mysterious sound, dubbed "biotwang," was recorded in 2014 near the Mariana Trench and later linked to Bryde's whales. Ann Allen from the National Oceanic and Atmospheric Administration (NOAA) saw an opportunity to use this discovery for tracking migrating Bryde's whales. Technological innovation in marine research: Google researchers collaborated with NOAA to develop an AI model capable of identifying vocalizations from eight different whale species. The AI model analyzes spectrograms, which convert...
read Sep 17, 2024Mistral AI Introduces Free Tier, Reduced Prices to Encourage AI Prototyping
Mistral AI expands AI accessibility: Mistral AI, a leading artificial intelligence company, has announced significant updates to its product lineup, including free offerings, reduced pricing, and enhanced model capabilities. Free tier introduction: Mistral AI has launched a free tier on its serverless platform, la Plateforme, allowing developers to experiment with and prototype AI models at no cost. Users can upgrade to a commercial tier with full data isolation and higher rate limits as needed. The platform offers deployment options across various cloud providers or on the user's own infrastructure. Pricing overhaul: The company has implemented substantial price reductions across its...
read Sep 17, 2024Why Startup Martian Thinks ‘Model Routing’ Is Key to Enterprise AI Success
AI model routing emerges as a critical enterprise solution: Martian, an AI startup, is developing technology to automatically select the best AI models for specific tasks, addressing a key challenge in enterprise AI implementation. Accenture has invested in Martian and plans to integrate its technology into its switchboard services, highlighting the growing importance of AI model optimization in enterprise settings. The integration aims to help organizations choose the most suitable AI models for their specific needs, considering factors such as performance, cost, and compliance requirements. Martian's approach focuses on predicting model behavior by understanding the internal workings of AI models,...
read Sep 17, 2024Early o1 Users Get Warnings from OpenAI for Probing Model’s Inner Thoughts
OpenAI's new AI model sparks controversy: OpenAI's latest "Strawberry" AI model family, particularly the o1-preview and o1-mini variants, has ignited a debate over transparency and user access to AI reasoning processes. The new models are designed to work through problems step-by-step before generating answers, a process OpenAI calls "reasoning abilities." Users can see a filtered interpretation of this reasoning process in the ChatGPT interface, but the raw chain of thought is intentionally hidden from view. OpenAI's decision to obscure the raw reasoning has prompted hackers and researchers to attempt to uncover these hidden processes, leading to warnings and potential bans...
read Sep 17, 2024AI Video Rivalry Heats Up with Dream Machine API Launch
AI video generation landscape evolves: Luma AI's announcement of its Dream Machine API marks a significant development in the rapidly advancing field of AI-generated video technology. The San Francisco-based startup, founded by former engineers from tech giants like Google, Meta, Adobe, and Apple, introduced the API just hours after competitor Runway unveiled its own API. This move intensifies the competition in the AI video technology sector, with major players vying for dominance in this emerging market. Key features of Dream Machine API: Luma AI's offering provides developers with a range of advanced tools for AI-driven video creation and manipulation. The...
read