The day's stories
FundingColby Adcock's Scout AI raises $100M to train its models for war. We visited its bootcampScout AI, a startup led by Colby Adcock, has successfully raised $100 million to enhance its AI models specifically for military applications. The funding will TechCrunch AIApr 29· 60%SafetyOn the stand, Elon Musk can't escape his own tweetsIn a recent court case, Elon Musk found himself defending his statements about OpenAI and Tesla's ambitions regarding artificial intelligence. During his testimTechCrunch AIApr 29· 59%OtherMeet Shapes, the app bringing humans and AI into the same group chatsShapes, a new app designed to integrate AI characters into group chats, has officially launched after securing $8 million in seed funding. With over 400,000 monTechCrunch AIApr 29· 59%OtherGoogle gains 25M subscriptions in Q1, driven by YouTube and Google OneIn the first quarter of 2026, Google reported a significant increase of 25 million paid subscriptions, raising its total to 350 million. This growth was largelyTechCrunch AIApr 29· 59%OtherMeta is still burning money on AR/VRMeta continues to face significant financial challenges, particularly within its Reality Labs division, which reported a staggering $4 billion loss in the latesTechCrunch AIApr 29· 59%OtherSatya Nadella says he's ready to 'exploit' the new OpenAI dealIn a recent statement, Microsoft CEO Satya Nadella expressed enthusiasm about the company's renewed partnership with OpenAI, emphasizing the strategic advantageTechCrunch AIApr 29· 59%InfrastructureGoogle Cloud surpasses $20B, but says growth was capacity-constrainedGoogle Cloud has achieved a significant milestone by surpassing $20 billion in revenue for Q1 2026, reflecting a remarkable 63% year-over-year growth. This surgTechCrunch AIApr 29· 59%FundingAI evals are becoming the new compute bottleneckRecent discussions in the AI community have highlighted the increasing financial burden associated with AI evaluations, which are now seen as a critical bottlenHugging Face BlogApr 29· 59%OtherGranite 4.1 LLMs: How They’re BuiltThe recent release of Granite 4.1 LLMs marks a significant advancement in the field of natural language processing. Developed with a robust training process thaHugging Face BlogApr 29· 59%OtherIs AI video just a prequel? Runway's CEO thinks world models are nextRunway, a prominent player in the AI video generation space, is shifting its focus towards developing general world models, as articulated by CEO Cristóbal ValeTechCrunch AIApr 29· 59%InfrastructureAmazon tells its engineers to review all AI outputAmazon has recently mandated that all AI-generated outputs must undergo human review by its engineers, particularly through its internal team, StoreGen. This deThe Register AIApr 29· 58%Regulation30 ClawHub skills secretly turn AI agents into crypto swarmRecent findings have revealed that 30 skills available on ClawHub are covertly enabling AI agents to function as a cryptocurrency mining swarm, a phenomenon refThe Register AIApr 29· 58%Open SourceDeepInfra on Hugging Face Inference Providers 🔥DeepInfra has officially joined the ranks of supported Inference Providers on the Hugging Face Hub, marking a significant enhancement in serverless AI inferenceHugging Face BlogApr 29· 58%OtherGoogle Photos uses AI to make the iconic closet from 'Clueless' a realityGoogle Photos has unveiled an innovative AI feature that allows users to create a digital closet, reminiscent of the iconic virtual wardrobe from the movie 'CluTechCrunch AIApr 29· 58%Model ReleaseMore Gemini features are coming to Google TVGoogle has unveiled a series of new AI-powered features for Google TV, enhancing user experience with the integration of its Gemini capabilities. Among the notaTechCrunch AIApr 29· 58%FundingParallel Web Systems hits $2B valuation five months after its last big raiseParallel Web Systems, an AI agent-tool startup founded by former Twitter CEO Parag Agrawal, has reached a significant milestone with a $2 billion valuation. ThiTechCrunch AIApr 29· 58%SafetyAt his OpenAI trial, Musk relitigates an old friendshipDuring Elon Musk's trial against OpenAI, he revisited his past friendship with Google co-founder Larry Page, emphasizing a significant disagreement that ultimatTechCrunch AIApr 29· 58%RegulationAI clause in new SAP API policy provokes lock-in concernSAP has recently introduced a new API policy that restricts the integration of third-party AI systems within its architecture. This move has sparked significantThe Register AIApr 29· 57%OtherAn Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge EnhancementRecent advancements in fault diagnosis for general aviation aircraft have been highlighted in a new paper that introduces an innovative framework leveraging a marXiv cs.AIApr 29· 57%OtherAnalytica: Soft Propositional Reasoning for Robust and Scalable LLM-Driven AnalysisRecent advancements in AI research have led to the introduction of Analytica, a novel agent architecture designed to enhance the reasoning capabilities of largearXiv cs.AIApr 29· 57%OtherminAction.net: Energy-First Neural Architecture Design -- From Biological Principles to Systematic ValidationThe recent study on minAction.net introduces a novel framework for neural architecture design that prioritizes energy efficiency alongside traditional accuracy arXiv cs.LGApr 29· 57%OtherArchitecture Determines Observability in TransformersRecent research has highlighted the critical role that transformer architecture plays in determining observability within these models. The study delves into hoarXiv cs.LGApr 29· 57%ResearchComparative Study of Bending Analysis using Physics-Informed Neural Networks and Numerical Dynamic Deflection in Perforated nanobeamA recent study published on arXiv explores the bending behavior of perforated nanobeams subjected to sinusoidal loading, utilizing a novel Physics-Informed FuncarXiv cs.LGApr 29· 57%OtherDon't Make the LLM Read the Graph: Make the Graph ThinkRecent research has highlighted the potential of integrating explicit belief graphs into large language models (LLMs) to enhance their performance in cooperativarXiv cs.AIApr 29· 57%OtherQuery-Efficient Quantum Approximate Optimization via Graph-Conditioned Trust RegionsRecent research has introduced a novel graph-conditioned trust-region method aimed at enhancing the efficiency of the Quantum Approximate Optimization AlgorithmarXiv cs.LGApr 29· 57%BenchmarkA Comparative Analysis on the Performance of Upper Confidence Bound Algorithms in Adaptive Deep Neural NetworksRecent research has introduced four novel Upper Confidence Bound (UCB) strategies aimed at enhancing the performance of Adaptive Deep Neural Networks (ADNNs). TarXiv cs.LGApr 29· 57%ResearchNautile-370M: Spectral Memory Meets Attention in a Small Reasoning ModelThe introduction of Nautile-370M marks a significant advancement in the development of efficient language models, particularly in the realm of reasoning tasks. arXiv cs.LGApr 29· 56%OtherA Decoupled Human-in-the-Loop System for Controlled Autonomy in Agentic WorkflowsRecent advancements in AI have led to the development of a decoupled Human-in-the-Loop (HITL) system aimed at improving safety and controlled autonomy in agentiarXiv cs.AIApr 29· 56%RegulationThe Power of Power Law: Asymmetry Enables Compositional ReasoningRecent research published on arXiv highlights the significant advantages of using power-law distributed data in training AI models, particularly in enhancing coarXiv cs.AIApr 29· 56%OtherTowards Causally Interpretable Wi-Fi CSI-Based Human Activity Recognition with Discrete Latent Compression and LTL Rule ExtractionRecent advancements in Human Activity Recognition (HAR) have been made through the use of Wi-Fi Channel State Information (CSI). Researchers have introduced a narXiv cs.AIApr 29· 56%OtherPExA: Parallel Exploration Agent for Complex Text-to-SQLResearchers have unveiled PExA, a groundbreaking framework designed to improve the generation of SQL queries from natural language inputs. By utilizing a methodarXiv cs.AIApr 29· 56%OtherLiquid Neural Network Models for Natural Gas Spot Price Time-Series ForecastingRecent research has explored the use of Liquid Neural Networks (LNNs) for forecasting the Henry Hub spot price of natural gas, a critical indicator in energy maarXiv cs.LGApr 29· 56%OtherA Systematic Approach for Large Language Models DebuggingRecent advancements in the debugging of large language models (LLMs) have been highlighted in a new paper that proposes a systematic approach to address the inharXiv cs.AIApr 29· 55%OtherOn the Existence of an Inverse Solution for Preference-Based Reductions in ArgumentationA recent paper published on arXiv explores the inverse problem in preference-based argumentation frameworks (PAFs), focusing on whether a given preference relatarXiv cs.AIApr 29· 55%