Core Viewpoint - The article highlights the rapid growth and innovative technology of the Israeli startup Decart, which has recently achieved a valuation of $3.1 billion after raising $100 million in Series B funding, showcasing its potential to reshape various industries with its groundbreaking AI video generation model, MirageLSD [4][20]. Group 1: Company Overview - Decart, co-founded by Dean Leitersdorf and Moshe Shalev, has quickly become a unicorn, raising a total of $153 million in just 11 months [4][9]. - The company focuses on real-time AI creative generation, with a team that has expanded from 15 to 60 members within two years [10][11]. - The founders have unique backgrounds, with Leitersdorf being a prodigy with a PhD at 23 and Shalev a veteran from the elite Israeli intelligence unit [5][9]. Group 2: Technology and Innovation - Decart launched MirageLSD, the world's first model capable of generating infinite-length videos in real-time with a latency of less than 40 milliseconds, making it faster than human perception [6][20]. - The technology utilizes a method called "diffusion forcing" to maintain video clarity over long durations and optimize GPU performance for rapid frame generation [17][21]. - MirageLSD allows users to interactively modify video content in real-time, enhancing user experience in gaming, live streaming, and other applications [19][20]. Group 3: Business Model and Market Strategy - Decart's business strategy includes two product lines: a GPU optimization tool for enterprises and a consumer-facing AI game called "Oasis," which has attracted over a million users shortly after launch [11][25]. - The GPU optimization tool significantly reduces operational costs from approximately $100 per hour to $0.25 per hour, generating millions in revenue [11][25]. - The company aims to become a trillion-dollar enterprise, aspiring to create an app that will be downloaded by a billion users, similar to the impact of smartphones [25][26]. Group 4: Future Vision - The founders envision transforming the entertainment and creative sectors through AI, aiming to revolutionize how users interact with technology [28][29]. - They believe that the future of the internet will be significantly influenced by AI, particularly in areas like video content and gaming, which have yet to fully leverage AI capabilities [28][29].

Core Insights - Decart, led by former Benchmark partner Victor Lazarte, recently completed a $100 million Series B funding round, raising its valuation to $3.1 billion in less than two years [1] - The company has seen a sixfold increase in valuation from $500 million to $3.1 billion in just over six months [1] - Decart's core products, Oasis and Mirage, are pioneering real-time generative AI technologies that enhance user interaction and experience [3][4] Product Development - Oasis is a real-time generative AI open-world model that allows users to interactively shape their virtual environment, achieving over 1 million users within three days of launch [4] - Mirage, described as a "world transformation model," enables real-time video-to-video conversion with a response time of under 40 milliseconds, eliminating delays common in previous AI video models [3][4] - Both products represent a shift from static visual content to dynamic, interactive experiences, expanding the potential applications in gaming, virtual reality, and the metaverse [5] Market Position and Strategy - Decart aims to create a consumer application with a user base of one billion, aspiring to reach a market valuation of $1 trillion [8] - The company is preparing to launch an API for Mirage, which will allow developers and businesses to leverage its core technology, fostering an open ecosystem [9] - Decart currently generates revenue from GPU acceleration and anticipates that the Mirage model will become a significant revenue source as costs for content generation are drastically reduced [10] Financial Performance - The company has achieved significant revenue from GPU acceleration, amounting to tens of millions of dollars [9] - The proprietary optimization technology has reduced the cost of content generation from $10 to $1,000 per hour to less than $0.25, positioning Decart competitively in the market [10] - The rapid increase in valuation reflects strong investor confidence driven by broad market demand and Decart's technological advantages [11]

Core Perspective - The article contrasts traditional high-budget AAA games with a new genre of AI-driven text adventure games, suggesting that the latter may better represent the future of the gaming industry [1][3][17]. Group 1: Game Mechanics and Structure - The game "Tower-Crawl" is a web-based text adventure where players create characters and interact through text input, with AI generating responses and narrative developments [4][8][10]. - Unlike traditional games with pre-set scripts, "Tower-Crawl" allows for unique gameplay experiences with each session, as AI generates different outcomes based on player input [13][20]. - The game operates on a model where players pay for input actions, creating a simple monetization strategy compared to traditional game sales [22][25]. Group 2: Industry Trends and Implications - The rise of AI-driven games like "Tower-Crawl" signifies a shift in game development, reducing the need for extensive pre-written content and allowing for rapid creation of diverse gaming experiences [25][29]. - The low development costs associated with AI text adventures make them accessible for independent developers, leading to a surge in the number of such games [26][31]. - The gaming industry is on the brink of an AI revolution, with major companies exploring AI to enhance game development processes and create more interactive experiences [40][44]. Group 3: Future of Gaming - The demand for greater player freedom and creativity is driving the evolution of gaming, with AI poised to fulfill these desires by generating expansive and personalized game worlds [36][48]. - As AI technology improves, it is expected to address current limitations in game continuity and consistency, potentially transforming the landscape of game development [46][50]. - The future may see a shift from traditional game purchases to a model where players buy AI processing power to create their own gaming experiences, leading to a more stratified gaming environment based on financial investment [50][51].

Core Viewpoint - The article discusses the innovative AI startup Decart and its groundbreaking video model MirageLSD, which enables real-time, zero-latency video generation, revolutionizing live streaming, gaming, and video communication [4][5][7]. Group 1: Technology and Features - MirageLSD is the first AI model to achieve zero-latency, infinite real-time video generation, allowing for continuous video streams without time limitations [4][5]. - The model operates at a speed 16 times faster than previous models, generating video at 24 frames per second and allowing for ongoing prompts, transitions, and edits during video generation [6][28]. - It addresses the "error accumulation" issue found in traditional autoregressive video models, ensuring temporal coherence while generating content frame by frame [9][11]. Group 2: Innovations and Mechanisms - The model employs a custom real-time stream diffusion model (Live-Stream Diffusion) that generates each frame based on previously generated frames and user prompts, rather than relying on the entire video sequence [14]. - It utilizes Diffusion Forcing technology to independently denoise single frames during training, ensuring coherence in frame generation [15]. - The model incorporates a historical enhancement strategy to preemptively correct potential errors by simulating artifacts during training [16]. Group 3: Performance and User Interaction - MirageLSD's architecture includes an improved Transformer model and a specially designed visual encoder, which enhances processing speed and reduces latency [18][20]. - The system features a dynamic input mechanism that processes player inputs with ultra-low latency, allowing for immediate responses to changes in the environment [22]. - Users can perform actions like changing outfits or transforming objects with minimal delay, showcasing the model's interactive capabilities [23]. Group 4: Company Background and Future Developments - Decart, the company behind MirageLSD, was founded in 2023 and previously launched the Oasis model, which also supports real-time interactions [25][26]. - The team plans to regularly release upgrades and new features for MirageLSD, including facial consistency, voice control, and precise object manipulation to enhance user experience [28].

Core Viewpoint - The article discusses a novel multimodal instruction data synthesis method called Oasis, which eliminates the need for complex prompt design by relying solely on images for data generation, thereby enhancing efficiency and quality in data synthesis [1][6]. Research Motivation - The traditional multimodal data synthesis methods face issues such as lack of diversity, insufficient quality, and high reliance on manual input, which Oasis aims to address [7][8]. Method Introduction - Oasis operates through three main steps: constructing a hooking prompt for autoregressive sampling, classifying the sampling results to retain instruction-type outputs, and conducting quality control and response generation [11][12]. Data Characteristics Analysis - The Oasis dataset, Oasis-500k, was synthesized from approximately 500,000 images, demonstrating scalability as data volume increases linearly with the number of images [21][22]. - The average instruction length for Oasis data is 76.80, while the average response length is 71.16, indicating richer information content compared to LLaVA-NeXT [24]. - The language diversity in Oasis data includes English (78.52%), Chinese (18.66%), and several other languages, showcasing its broad applicability [27]. Experimental Results - Oasis shows significant performance improvements over baseline models, with average accuracy increases of 3.1% for Vicuna1.5, 1.8% for Qwen2.5, and 3.2% for Llama3 [38]. - The addition of 500k Oasis data resulted in an average score increase of 5.2%, confirming the effectiveness of data scaling [41]. Effectiveness of Oasis - Oasis demonstrates strong capabilities in synthesizing domain-specific data, particularly in OCR tasks, leading to notable performance enhancements in relevant benchmarks [43]. Quality Control Mechanism - The quality control mechanism for instructions is essential, as it significantly improves model performance, with a noted increase of over 7% in specific tasks [50].

Core Viewpoint - The rapid advancement of world models, particularly with the introduction of interactive world models like Matrix-Game, signifies a pivotal moment in AI development, enabling more immersive and controllable virtual environments [4][50]. Group 1: Development of World Models - The Oasis project marked the first real-time, interactive open-source world model, showcasing a significant leap in understanding physical and game rules [1]. - Microsoft's MineWorld further enhanced visual effects and action generation consistency in interactive world models [2]. - The recent launch of Matrix-Game by Kunlun Wanwei represents a major milestone in interactive world generation, being the first open-source model in the industry with over 10 billion parameters [10][50]. Group 2: Features of Matrix-Game - Matrix-Game allows for fine-grained user interaction control, enabling players to experience seamless movement and environmental feedback in a game world [17]. - The model demonstrates high fidelity in visual and physical consistency, generating realistic interactions and maintaining visual coherence during gameplay [19][20]. - It exhibits multi-scene generalization capabilities, allowing for the generation of diverse environments beyond just Minecraft, including cities and historical buildings [25][26]. Group 3: Evaluation and Performance - Kunlun Wanwei introduced a comprehensive evaluation framework called GameWorld Score, assessing visual quality, temporal consistency, controllability, and understanding of physical rules [29]. - In comparative assessments, Matrix-Game outperformed other models like Oasis and MineWorld across all evaluation dimensions [31]. - The model achieved over 90% accuracy in action control, demonstrating its robustness in responding to user inputs [35]. Group 4: Technological Innovations - Matrix-Game's success is attributed to its innovative data collection and model architecture, utilizing a large dataset for training that includes both unlabelled and labelled data [41][42]. - The architecture focuses on image-to-world modeling, allowing the model to generate interactive video content based solely on visual inputs without relying on language prompts [44][45]. - The model's ability to maintain temporal coherence during video generation is a significant advancement, addressing previous challenges in long-sequence content generation [45]. Group 5: Broader Implications - Matrix-Game's capabilities extend beyond gaming, impacting content production in various fields such as film, advertising, and XR [51]. - The development of spatial intelligence through models like Matrix-Game is crucial for advancing embodied intelligence and enhancing machine understanding of the three-dimensional world [49][50]. - Kunlun Wanwei aims to create a comprehensive AI creative ecosystem, facilitating innovation and expression in a new dimension of interaction [52].