The day's stories
FundingPramaana Labs raises $27M seed round from Khosla Ventures to bring formal verification to AIPramaana Labs has successfully secured $27 million in seed funding, led by Khosla Ventures, to advance the field of formal verification in artificial intelligenTechCrunch AIJun 17· 60%FundingWorld model maker Odyssey nabs $1.45B valuation backed by Amazon and other big namesOdyssey, an AI startup focused on developing world models that simulate physical environments, has reached a significant milestone with a $1.45 billion valuatioTechCrunch AIJun 17· 59%Model ReleaseGLM-5.2: Built for Long-Horizon TasksHugging Face has officially launched GLM-5.2, a significant upgrade from its predecessor GLM-5.1. This new model is specifically designed to handle long-horizonHugging Face BlogJun 17· 59%FundingCollecting robot training data is dirty, unglamorous work. Some AI labs are already paying XDOF to do it.OpenAI is revitalizing its robotics program, emphasizing the critical need for high-quality training data in the development of robotic systems. To address thisTechCrunch AIJun 17· 59%ResearchMolmoMotion: Language-guided 3D motion forecastingThe recent introduction of MolmoMotion marks a significant advancement in the field of 3D motion forecasting. This innovative model utilizes language guidance tHugging Face BlogJun 17· 58%ResearchOnly 16 percent of Americans think AI will have a positive impact on society, a new study showsA recent Pew Research study has highlighted a significant skepticism among Americans regarding the future impact of artificial intelligence on society. Only 16%TechCrunch AIJun 17· 58%InfrastructureFrom the Hugging Face Hub to robot hardware with Strands Agents and LeRobotHugging Face has announced a significant integration with Strands Robots, allowing for a more streamlined approach to training and deploying robotic agents. ThiHugging Face BlogJun 17· 58%FundingCanadian pension giant joins race to fund India's AI-fueled data center boomThe Canada Pension Plan Investment Board (CPP Investments) has made a significant move in the AI and cloud infrastructure sector by investing up to ₹70 billion TechCrunch AIJun 17· 57%Product LaunchPinterest launches an experimental AI shopping app called 'Ask Pinterest'Pinterest has introduced an experimental AI shopping app named 'Ask Pinterest', designed to improve product discovery through a conversational interface. This lTechCrunch AIJun 17· 57%Regulation"Dangerous" AI models are coming no matter whatAnthropic has recently taken its AI models, Claude Fable 5 and Mythos 5, offline in response to a U.S. government export-control directive. These models exhibitArs Technica AIJun 17· 57%Product LaunchGoogle bets on Gemini to reinvent the smart home speakerGoogle has officially launched its new Home Speaker, which is integrated with the advanced capabilities of its Gemini AI. This new device aims to enhance user eTechCrunch AIJun 17· 56%RegulationWorld leaders want American AI. They just don't want America to be able to turn it off.During the recent G7 Summit, leaders from various democratic nations voiced their concerns regarding the potential for the U.S. to restrict access to American ATechCrunch AIJun 17· 56%FundingAnthropic becomes first AI startup to join the Frontier carbon removal coalitionAnthropic has made headlines by becoming the first AI startup to join the Frontier carbon removal coalition, which is focused on addressing climate change throuTechCrunch AIJun 17· 56%FundingNEA's Tiffany Luck on AI IPOs, personal agents, and the ROI reckoningTiffany Luck from NEA has highlighted a significant shift in the AI investment landscape, moving from a phase of high spending to a more cautious approach focusTechCrunch AIJun 17· 56%RegulationTesco moving 40,000 server workloads off VMware amid Broadcom's “abusive conduct”Tesco has decided to migrate 40,000 server workloads away from VMware as a direct response to what it describes as 'abusive conduct' by Broadcom, the company thArs Technica AIJun 17· 56%ResearchTowards Fast GNN Surrogates for CO2 Migration in Complex Geological FormationsRecent research has focused on the development of a machine learning model that utilizes graph neural networks (GNNs) to predict the migration of CO2 plumes in arXiv cs.LGJun 17· 64%ResearchQuantifying Consistency in LLM Logical Reasoning via Structural UncertaintyRecent research has introduced a novel framework termed 'structural uncertainty' aimed at assessing the consistency of logical reasoning in large language modelarXiv cs.AIJun 17· 63%ResearchProbing, Fusion, and Trustworthiness: A Systematic Evaluation of Foundation Model Representations for Multimodal Cancer AnalysisA recent study published on arXiv systematically evaluates foundation models (FMs) for their effectiveness in multimodal cancer analysis, particularly using whoarXiv cs.LGJun 17· 63%ResearchVerified Detection and Prevention of Concurrency Anomalies in Multi-Agent Large Language Model SystemsRecent research has introduced a formal framework aimed at detecting and preventing concurrency anomalies in multi-agent large language model (LLM) systems. ThearXiv cs.LGJun 17· 63%ResearchThe Critical Role of Model Selection in Causal Inference: A Comparative Analysis of Classification Models within the InferBERT Framework for PharmacovigilanceA recent study published on arXiv highlights the importance of model selection in causal inference, particularly within the InferBERT framework used for pharmacarXiv cs.LGJun 17· 63%ResearchDistributed General-Purpose Agent Networks: Architecture, Key Mechanisms, and PrototypesA recent paper published on arXiv outlines a novel framework for distributed general-purpose agent networks, which allows for heterogeneous agents to work collaarXiv cs.AIJun 17· 63%ResearchCorrect When Paired, Wrong When Split: Decoupling and Editing Modality-Specific Neurons in MLLMsRecent research has introduced DECODE, a novel approach designed to tackle the decoupling failure observed in Multimodal Large Language Models (MLLMs). This faiarXiv cs.LGJun 17· 62%BenchmarkSkillChain-Gym: A Benchmark for Reskilling-Aware Production-Inventory Control under DisruptionsThe introduction of SkillChain-Gym marks a significant advancement in the field of production-inventory control, particularly in the context of workforce dynamiarXiv cs.AIJun 17· 62%ResearchInformative Missingness to Generate Irregular Clinical Time SeriesA recent study introduced a novel diffusion-based method aimed at generating irregular clinical time series data. This approach is particularly significant as iarXiv cs.LGJun 17· 62%BenchmarkMemTrace: Probing What Final Accuracy Misses in Long-Term MemoryResearchers have introduced a new benchmark called MemTrace, aimed at evaluating the long-term memory capabilities of large language model (LLM) agents. Unlike arXiv cs.AIJun 17· 62%ResearchNothing from Something: Can a Language Model Discover 0?A recent study published on arXiv explores the ability of language models to independently discover the concept of zero, a fundamental mathematical structure. TarXiv cs.AIJun 17· 62%ResearchWhen Rules Learn: A Self-Evolving Agent for Legal Case RetrievalRecent research has introduced a self-evolving framework designed to improve legal case retrieval, specifically enhancing the traditional BM25 model. This innovarXiv cs.AIJun 17· 62%ResearchModels Take Notes at Prefill: KV Cache Can Be Editable and ComposableRecent research has introduced significant advancements in prefix caching for machine learning models, specifically focusing on the concept of editable and comparXiv cs.LGJun 17· 62%ResearchBeyond Parallel Sampling: Diverse Query Initialization for Agentic SearchA recent paper has introduced a novel method called DivInit, aimed at enhancing agentic search capabilities in AI systems. This method specifically targets the arXiv cs.AIJun 17· 62%ResearchTreatment Response Optimized Clinical Decision Support AI System via Digital Twin SimulationRecent advancements in Clinical Decision Support AI Systems (CDSASs) have led to the development of an innovative online adaptive framework that integrates TreaarXiv cs.AIJun 17· 61%BenchmarkSpeechDx: A Multi-Task Benchmark for Clinical Speech AIResearchers have unveiled SpeechDx, a new multi-task benchmark specifically tailored for clinical speech AI applications. This benchmark encompasses 12 distinctarXiv cs.AIJun 17· 61%ResearchSkill-Constrained Model Predictive Control for Resilient Manufacturing Supply ChainsA recent study published on arXiv presents a novel approach to managing manufacturing supply chains through a skill-constrained model predictive controller. ThiarXiv cs.AIJun 17· 61%ResearchNoise-Driven Escape from Metastable Phases explains Grokking in Deep Neural NetworksRecent research has unveiled a significant relationship between noise-driven processes in deep neural networks (DNNs) and a phenomenon known as grokking. The starXiv cs.LGJun 17· 60%ResearchDiagnosing and Repairing Shape-Prior Shortcuts in Long-Range Single-Shot Fringe Projection ProfilometryRecent research has focused on enhancing long-range single-shot fringe projection profilometry (FPP), a technique used for precise 3D shape measurement. The stuarXiv cs.LGJun 17· 59%