[Music] Welcome to the Vergecast, the flagship podcast of testing cursed technology. I'm your friend V Song and I'm here with a special Sunday bonus episode. Yay.We're calling this AI for normies. So, here's the concept. AI can be so open-ended, it's really hard for the average person to know what it's good for.And if you ask me, I don't think big tech is doing such a great job at explaining that either. But we here at the verge. com are a bunch of giant nerds and we test all of this stuff for a living.So I ...

Technology & Innovation - zkGPT is a system for proving the correctness of large language model (LLM) inference without revealing the model [1] - The system enables private, verifiable LLM inference [1] - The system generates compact proofs in under 25 seconds [1]

Core Insights - Innodata Inc. (INOD) is positioned as a key player in the AI revolution by providing essential data for training advanced language models [1] - The company reported Q2 2025 adjusted earnings per share of $0.20, exceeding the Zacks Consensus Estimate of $0.11 [1] - Quarterly revenues reached $58.39 million, reflecting a 79% year-over-year increase and surpassing estimates by 3.6% [2] Revenue Growth and Guidance - Following strong Q2 performance, Innodata raised its 2025 revenue growth guidance to over 45% year-over-year, up from a previous forecast of 40% [2] - The expected revenue growth rate for the current year is 41.9%, while the earnings growth rate is projected at -23.6% [6] AI Demand and Market Position - Innodata is set to benefit from the increasing demand for data engineering services in large language model development, supporting five of the seven major hyperscalers [3] - The company has diversified its customer base, which is expected to support long-term growth across various sectors including technology, healthcare, and federal agencies [4] New Product Launch - Innodata introduced a GenAI Test and Evaluation Platform aimed at validating large language models, with MasterClass as the first customer [5] - The platform is designed to enhance integration with major tech companies' upcoming GenAI investments [5] Stock Performance and Estimates - Innodata's stock is currently trading 38.6% below its 52-week high, despite a year-to-date return of 10.3%, outperforming the S&P 500 [7] - Brokerage targets suggest a potential upside of 72.1%, with average short-term price targets indicating a 53.2% increase from the last closing price of $43.58 [10] Consensus Estimates - The Zacks Consensus Estimate for current-year earnings has remained stable over the last 30 days, while next-year earnings estimates have improved by 2.9% [6]

Core Insights - The article discusses advancements in trajectory prediction using a generative active learning framework called GALTraj, which applies controllable diffusion models to address long-tail issues in data [1][2]. Group 1: GALTraj Framework - GALTraj is the first framework to apply generative active learning to trajectory prediction tasks, enhancing long-tail learning without modifying the model structure [2]. - The framework employs a tail-aware generation method that differentiates the diffusion guidance for tail, head, and related agents, producing realistic and diverse scenarios while preserving tail characteristics [2][3]. Group 2: Experimental Results - In experiments on WOMD and Argoverse2 datasets, GALTraj significantly improved long-tail sample prediction performance, reducing the long-tail metric FPR₅ by 47.6% (from 0.42 to 0.22) and overall prediction error minFDE₆ by 14.7% (from 0.654 to 0.558) [1][6]. - The results indicate that GALTraj outperforms traditional methods across various metrics, showcasing its effectiveness in enhancing prediction accuracy for rare scenarios [7][8]. Group 3: TopoLiDM Framework - The article also highlights the TopoLiDM framework developed by Shanghai Jiao Tong University and Twente University, which integrates topology-aware diffusion models for high-fidelity LiDAR point cloud generation [13][15]. - TopoLiDM achieved a 22.6% reduction in the Fréchet Range Image Distance (FRID) and a 9.2% reduction in Minimum Matching Distance (MMD) on the KITTI-360 dataset while maintaining a real-time generation speed of 1.68 samples per second [13][15]. Group 4: FastDriveVLA Framework - FastDriveVLA, developed by Peking University and Xiaopeng Motors, introduces a reconstruction-based visual token pruning framework that maintains 99.1% trajectory accuracy with a 50% pruning rate and reduces collision rates by 2.7% [21][22]. - The framework employs a novel adversarial foreground-background reconstruction strategy to enhance the identification of valuable tokens, achieving state-of-the-art performance on the nuScenes open-loop planning benchmark [27][28]. Group 5: PLA Framework - The article presents a unified Perception-Language-Action (PLA) framework proposed by TUM, which integrates multi-sensor fusion and GPT-4.1 enhanced visual-language-action reasoning for adaptive autonomous driving [34][35]. - The framework demonstrated a mean absolute error (MAE) of 0.39 m/s in speed prediction and an average displacement error (ADE) of 1.013 meters in trajectory tracking within urban intersection scenarios [42].

Group 1: Core Insights - The article emphasizes the importance of foundational skills in programming and software engineering for entering the AI field, with Python being the preferred language due to its ease of use and comprehensive ecosystem [1][2][4] - It highlights that while many AI roles stem from machine learning, the most sought-after positions are closer to software engineering, necessitating knowledge of languages like Java, GO, or Rust [1][2] - Continuous practice and real-world application are deemed essential for mastering programming languages, rather than solely relying on courses or books [2] Group 2: Recommended Resources - A variety of resources are suggested for learning Python, including a beginner's course that can be completed in four hours and a highly regarded specialization course [5] - For mathematics and statistics, specific books and courses are recommended to understand the underlying principles of machine learning and AI [9][10] - The article lists essential resources for deep learning and large language models, emphasizing the significance of frameworks like PyTorch and TensorFlow in the industry [13][14] Group 3: AI Engineering and Productization - The article stresses the need for skills in productizing AI models, indicating that most AI roles resemble traditional software engineering rather than pure machine learning engineering [11] - It mentions the importance of learning MLOps for model deployment, covering aspects like containerization and cloud systems [11] - The article concludes with advice on becoming an expert in the field through project-based learning and self-reflection [14]

Core Viewpoint - The future of AI is likely to surpass human intelligence, leading to significant implications for society and the relationship between humans and AI [1][47]. Group 1: AI Development and Understanding - AI has evolved through two paradigms: logical reasoning and learning through neural networks, with the latter being more aligned with human thought processes [5][12]. - Large language models (LLMs) are seen as descendants of earlier models, utilizing more complex structures and interactions to understand language similarly to humans [12][25]. - The understanding of language in LLMs is compared to building with LEGO blocks, where words are multi-dimensional and can adapt based on context [16][19]. Group 2: Knowledge Transfer and Efficiency - The efficiency of knowledge transfer in AI is significantly higher than in human communication, allowing for rapid sharing of information across multiple instances of AI [37][40]. - Digital intelligence can replicate and share model weights and experiences, leading to a collaborative learning environment that surpasses human capabilities [39][41]. Group 3: Implications of Advanced AI - As AI systems become more intelligent, they may develop motivations for survival and control, potentially leading to challenges in managing these systems [47][48]. - The relationship between humans and advanced AI could shift, with AI becoming more autonomous and capable of influencing human decisions [49][52]. - The necessity for international cooperation in AI safety and governance is emphasized, as the risks associated with advanced AI systems are global in nature [59][62].

Core Viewpoint - The article highlights a significant achievement in artificial intelligence (AI), where large language models (LLMs) have reached gold medal level in the International Mathematical Olympiad (IMO), showcasing their advanced problem-solving capabilities [4][5][6]. Group 1: AI Achievement - Google DeepMind's large language model successfully solved problems equivalent to those in the IMO, achieving a score that surpasses the gold medal threshold of 35 out of 42 [4][5]. - This marks a substantial leap from the previous year's performance, where the model was only at the silver medal level, indicating a qualitative breakthrough in AI's ability to handle complex mathematical reasoning [5][6]. Group 2: Implications of the Achievement - The success of LLMs in the IMO demonstrates their capability to tackle highly complex tasks that require deep logical thinking and abstract reasoning, beyond mere text generation [7]. - Such AI advancements can serve as powerful tools in education and research, assisting students in learning higher mathematics and aiding researchers in exploring new conjectures and theorems [7]. - Achieving gold medal level in mathematics is a significant milestone on the path to artificial general intelligence (AGI), as it requires a combination of various cognitive abilities [7][8]. Group 3: Broader Impact - The breakthroughs by DeepMind and OpenAI not only elevate AI's status in mathematical reasoning but also suggest vast potential for future applications in scientific exploration and technological development [8].

Core Viewpoint - The article discusses the release and features of the Qwen3-Coder model by Alibaba Cloud, highlighting its advanced capabilities in coding and agentic tasks, as well as its competitive performance against other models in the market [3][4][5]. Group 1: Model Features - Qwen3-Coder series includes various versions, with Qwen3-Coder-480B-A35B-Instruct being the most powerful, featuring 480 billion parameters and supporting 256K tokens natively, expandable to 1 million tokens [4]. - The model has achieved state-of-the-art (SOTA) results in areas such as Agentic Coding, Browser Use, and Tool Use, comparable to Claude Sonnet4 [5][6]. - The training data for Qwen3-Coder amounts to 7.5 terabytes, with 70% being code, enhancing its programming capabilities while maintaining general and mathematical skills [12]. Group 2: Technical Details - The model utilizes a unique approach to reinforcement learning (RL) by focusing on real-world software engineering tasks, allowing for extensive interaction and decision-making [16]. - A scalable environment for RL has been established, enabling the simultaneous operation of 20,000 independent environments, which enhances feedback and evaluation processes [16]. Group 3: Tools and Integration - Qwen Code, a command-line tool for agentic programming, has been developed to maximize the performance of Qwen3-Coder in coding tasks [17]. - The integration of Qwen3-Coder with Claude Code is also highlighted, allowing users to leverage both models for enhanced coding experiences [22][26]. Group 4: User Experience - Users can access Qwen3-Coder through the Qwen Chat web version for free, providing an opportunity to experience its capabilities firsthand [6][7]. - Various demos showcasing the model's capabilities, such as simulating a solar system and creating visual effects in coding environments, are available for users [8][9][10].

Finally, a framework to connect any LLM to any MCP server (open-source).mcp-use lets you connect any LLM to any MCP server & build custom MCP Agents, without using closed-source apps like Cursor/Claude.Compatible with Ollama, LangChain, etc.Build 100% local MCP clients! https://t.co/8rhqh7BUZh ...

LLM Training Stages - LLM 从零开始训练包含四个阶段 [1] - 第一步是使用随机初始化的模型 [2] - 之后在大规模语料库上进行预训练 [2] - 使用指令微调使其能够遵循命令 [2] - 使用偏好和推理微调来优化响应 [2]