The day's stories
FundingGeneral Intuition's $2.3B bet that video games can train AI agents for the real worldGeneral Intuition has successfully raised $320 million, achieving a valuation of $2.3 billion, to focus on developing AI agents capable of learning from video gTechCrunch AIJun 25· 61%RegulationThe White House is asking OpenAI to slow roll the release of its new model over safety concernsThe White House has intervened in the release of OpenAI's upcoming model, GPT 5.6, requesting that the company limit access to a select group of partners. This TechCrunch AIJun 25· 59%ResearchWhich tokens does a hybrid model predict better?Recent research has focused on the comparative performance of the Olmo Hybrid model versus the traditional Olmo 3 transformer model in predicting tokens. The fiHugging Face BlogJun 25· 59%FundingNetris raises $15M Series A from a16z to help AI neoclouds go live fasterNetris has successfully secured $15 million in a Series A funding round led by Andreessen Horowitz. This investment aims to bolster Netris's network automation TechCrunch AIJun 25· 58%InfrastructureAmazon ups India bet with fresh $13B AI infrastructure investmentAmazon has announced a significant $13 billion investment aimed at enhancing its AI and cloud infrastructure in India. This decision follows a strategic meetingTechCrunch AIJun 25· 57%Product LaunchAnthropic's Claude is winning over paid consumers, a market owned by ChatGPTAnthropic's AI model, Claude, is experiencing significant growth in the paid consumer market, reporting a 75% increase in revenue since January 2026. While ClauTechCrunch AIJun 25· 56%ResearchDatabricks’ former AI chief thinks he can cut AI’s power bill by 1,000xUnconventional AI, founded by Naveen Rao, the former AI chief at Databricks, has introduced its first model, Un-0, which leverages a novel oscillator-based compTechCrunch AIJun 25· 56%FundingPatronus AI lands $50M to build ‘digital worlds’ that stress-test AI agentsPatronus AI has successfully secured $50 million in a Series B funding round to develop simulated digital environments designed for stress-testing AI agents. ThTechCrunch AIJun 25· 56%Product LaunchNotion killing Skiff-influenced email app since most users use AI agents insteadNotion has decided to shut down its Notion Mail email client, which was developed following the acquisition of Skiff. The decision comes as a result of low userArs Technica AIJun 25· 54%ResearchDiagnosing and Mitigating Compounding Failures in Agentic Persuasion via Taxonomic Strategy RetrievalRecent research has introduced a novel approach called Taxonomic Strategy RAG (TS-RAG) aimed at improving the performance of foundation-model agents in persuasiarXiv cs.AIJun 25· 64%ResearchTRUSTMEM: Learning Trustworthy Memory Consolidation for LLM Agents with Long-Term MemoryThe recent paper on TrustMem presents a novel framework designed to improve the trustworthiness of memory consolidation in large language model (LLM) agents. ByarXiv cs.AIJun 25· 64%ResearchTransferability for General Reasoning: An Automated Curriculum for Multi-Domain RLVRA recent paper published on arXiv presents a novel approach to multi-domain reinforcement learning through the introduction of the Transfer-Aware Curriculum (TAarXiv cs.AIJun 25· 63%ResearchProject Auto-World: Towards Automated Benchmarking of Neural Relational ReasonersProject Auto-World is an innovative initiative focused on automating the benchmarking process for neural relational reasoners. By utilizing large language modelarXiv cs.AIJun 25· 63%ResearchLLM Evolution as an Industry-Scale Ecosystem: A Lifecycle Perspective on Continual LearningRecent research has introduced a framework for the lifecycle of large language models (LLMs) within an industrial context, focusing on the importance of continuarXiv cs.LGJun 25· 63%ResearchHolographic Memory for Zero-Shot Compositional Reasoning in Knowledge Graphs: A Mechanistic Study of Where and Why It FailsA recent study published on arXiv delves into the use of Holographic Reduced Representations (HRR) for zero-shot compositional reasoning within knowledge graphsarXiv cs.LGJun 25· 63%ResearchBeyond Shapley: Efficient Computation of Asymmetric Shapley ValuesRecent research has unveiled a polynomial-time algorithm designed for the efficient computation of Asymmetric Shapley Values (ASV) within certain causal graph sarXiv cs.AIJun 25· 63%ResearchThe Hitchhiker's Guide to Agentic AI: From Foundations to SystemsThe recently released guide titled 'The Hitchhiker's Guide to Agentic AI' serves as a comprehensive reference for the development of autonomous AI systems. It carXiv cs.AIJun 25· 62%ResearchElo-Disentangled Player-Style Embeddings for Human Chess via Rating-Conditioned Residual Move ModelRecent research has introduced a novel approach to understanding and predicting human chess styles through the use of Elo-disentangled player-style embeddings. arXiv cs.AIJun 25· 62%ResearchSupervised Reinforcement Learning for the Coordination of Distributed Energy ResourcesA recent paper introduces a novel Supervised Reinforcement Learning (SRL) framework designed to enhance the coordination of distributed energy resources (DERs).arXiv cs.LGJun 25· 62%BenchmarkMacroLens: A Multi-Task Benchmark for Contextual Financial Reasoning under Macroeconomic ScenariosThe introduction of MacroLens marks a significant advancement in the field of financial reasoning, particularly in the context of macroeconomic scenarios. This arXiv cs.LGJun 25· 62%ResearchWhen Do Conservation Laws Survive Learned Representations? Certified Horizons for Latent World ModelsA recent study published on arXiv has introduced the concept of 'certified horizons' in the context of latent world models. This research investigates how long arXiv cs.LGJun 25· 62%ResearchOn-Device Neural Architecture SearchA recent paper presents a novel approach to Neural Architecture Search (NAS) that is designed to function directly on deployment devices. This on-device NAS aimarXiv cs.LGJun 26· 62%ResearchA Spectral Phase Diagram for Binary Few-Shot Classification: Intrinsic Dimensionality, Geometric Saturation, and Representational DiagnosisRecent research has introduced a novel saturation index for binary few-shot classification, which aims to optimize the process of collecting labeled examples. TarXiv cs.LGJun 25· 61%ResearchDo vision-language models search like humans? Reasoning tokens as a reaction-time analog in classic visual-search paradigmsA recent study explores the capabilities of vision-language models (VLMs) in performing visual search tasks, comparing their behaviors to those of humans. By adarXiv cs.AIJun 25· 61%RegulationThe Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI PrescribingThe transition of AI systems to autonomous medication prescribing presents significant challenges regarding trust, liability, and uncertainty. A recent study hiarXiv cs.AIJun 25· 61%ResearchTo Isolate or to Score? Model-Adaptive Assessment for Cost-Efficient Multi-Agent RAGA recent study published on arXiv explores a novel approach to multi-agent document assessment in retrieval-augmented generation (RAG) by introducing the Model-arXiv cs.AIJun 25· 61%ResearchFrom Meta Idea to Advanced Mathematical Discovery -- Human-AI Co-Discovery of Sign-Embedding Quantum AlgorithmsA recent study highlights the potential of AI-assisted exploration in advancing mathematical research, specifically in the realm of sign-embedding quantum algorarXiv cs.LGJun 25· 60%