The day's stories
ResearchVentureBeat names Rob Strechay as its first Lead Analyst, expanding its enterprise AI research pushVentureBeat has announced the appointment of Rob Strechay as its first Lead Analyst, a strategic move aimed at bolstering its enterprise AI research capabilitieVentureBeat AIAug 19· 61%Product LaunchAmazon makes its AI-powered Alexa+ free on Fire TV, no Prime requiredAmazon has made a significant move by announcing that its AI-powered assistant, Alexa+, will now be available for free on all compatible Fire TV devices in the TechCrunch AIAug 19· 59%Product LaunchOpenAI seeks to one-up Anthropic with new customer privacy protectionsOpenAI has launched a new feature called Private Safety Processing, designed to enhance customer privacy by monitoring potential misuse of AI while ensuring thaTechCrunch AIAug 19· 59%Product LaunchGoogle packs Search and Gemini with new AI study toolsGoogle has unveiled new AI-powered study tools integrated into its Search and Gemini platforms. These enhancements include interactive visuals and customized quTechCrunch AIAug 19· 58%OtherAI was supposed to win people over by now -- it hasn'tDespite significant advancements in artificial intelligence technology, public sentiment towards AI is increasingly negative. Recent polls indicate that many coTechCrunch AIAug 19· 58%OtherResearchers say OpenAI revoked their access to limited cyber programOpenAI has recently revoked access to its Trusted Access for Cyber (TAC) program for several security researchers, citing a technical issue as the reason behindTechCrunch AIAug 19· 58%InfrastructureTerraPower’s nuclear reactor has a secret weapon for powering AI data centersTerraPower, a company co-founded by Bill Gates, is preparing to unveil its inaugural data center project that utilizes its innovative molten salt-cooled nuclearTechCrunch AIAug 19· 57%Product LaunchStripe didn't really buy OpenRouter because of the 'singularity'Stripe has made headlines with its acquisition of OpenRouter for a staggering $7.5 billion, a significant increase from OpenRouter's previous valuation of $1.3 TechCrunch AIAug 19· 57%Model ReleaseLFM2.5 Q4\_0 Checkpoints from Quantization-Aware DistillationLiquidAI has announced the release of updated QAD Q4_0 checkpoints for its LFM2.5 models, which have undergone Quantization-Aware Distillation (QAD). This technHugging Face BlogAug 19· 56%ResearchAI isn't close to curing cancer. This startup says it knows what it will take.Vivodyne, a biotech startup, has introduced HIVE, a modular robotic laboratory designed to cultivate human tissue and produce causal biological data aimed at enTechCrunch AIAug 19· 56%FundingMeet the startup helping Wall Street put a price on AI computeSilicon Data, a startup focused on establishing a pricing mechanism for AI compute, has successfully raised $30 million in Series A funding. This funding will sTechCrunch AIAug 19· 56%OtherCognition CEO denies report that SpaceX tried to acquire the startupCognition CEO Scott Wu has publicly refuted a report from Bloomberg that suggested SpaceX was in talks to acquire the AI coding startup. Wu emphasized that CognTechCrunch AIAug 19· 54%ResearchGxP-Agent: Process-DAG Topology for Reliable Clinical Trial Programming with LLM AgentsResearchers have introduced GxP-Agent, a novel multi-agent system designed to improve the reliability of clinical trial programming. This system employs a direcarXiv cs.AIAug 19· 65%SafetyDiSCO: Defending text-to-image generation through distribution-guided contrastive prompt optimizationResearchers have introduced DiSCO, a novel defense mechanism aimed at improving the safety of text-to-image generative models. This approach focuses on distribuarXiv cs.AIAug 19· 64%ResearchFedPref: Federated Preference Learning for Structured Radiology Report ExtractionThe introduction of FedPref marks a significant advancement in the field of federated learning, particularly for the extraction of structured data from radiologarXiv cs.AIAug 19· 64%ResearchA decodability criterion predicts when hidden-state selection beats majority voting in large language modelsRecent research has introduced a novel method known as CASE (Correctness-Axis SElection) aimed at enhancing decision-making capabilities in large language modelarXiv cs.AIAug 19· 64%ResearchTraining and Evaluating Ethical Reinforcement Learning Agents on Per-Episode DistributionsRecent research has focused on enhancing the ethical performance of Reinforcement Learning (RL) agents by comparing various training methods. The study specificarXiv cs.LGAug 19· 63%ResearchDUET: Dual-Teacher On-Policy Distillation via Same-Weight Disagreement for Prohibition ComplianceResearchers have introduced a new method called DUET, which stands for Dual-Teacher On-Policy Distillation. This innovative approach aims to improve compliance arXiv cs.LGAug 19· 62%ResearchDiscrete Diffusion Language Models Are Training-Free Multi-Label ClassifiersA recent paper has introduced a novel approach to multi-label text classification through the use of discrete diffusion language models, specifically a method carXiv cs.LGAug 19· 62%ResearchEfficient Neural-Network-Based High-Resolution Radiative Transfer for CO___ Retrieval, and Application to Interferometric SensingA recent study has introduced an efficient feedforward multilayer perceptron (MLP) surrogate model aimed at optimizing high-resolution radiative transfer simulaarXiv cs.LGAug 19· 62%ResearchSkillEffect: Checked Lowering for Memory-Bounded Agent ToolsA new runtime system called SkillEffect has been introduced to address memory management in agent tools. This system employs a checked-lowering mechanism that earXiv cs.AIAug 19· 62%ResearchThe Problem Is the Problem: Towards Scalable Mathematical DiscoveryA recent paper on arXiv presents a novel approach to mathematical discovery through a system called FAR. This human-AI collaboration paradigm focuses on automatarXiv cs.AIAug 19· 62%Product LaunchKernelArc: A Multi-Agent Framework for GPU Kernel OptimizationKernelArc, a new multi-agent framework specifically designed for optimizing GPU kernels, has been released. This innovative framework demonstrates significant parXiv cs.AIAug 19· 62%ResearchRuntime Governance for Agentic AI: Action-Boundary Control with Trusted Provenance and Fail-Closed ExecutionA recent study has introduced Aegis, a novel runtime governance system designed to enhance the safety of agentic AI systems. This system acts as a trusted decisarXiv cs.AIAug 19· 62%ResearchiFuzz-Meta: An Interpretable Fuzzy Learning Framework Bridging Top-Down and Bottom-Up Knowledge IntegrationThe recent paper on iFuzz-Meta introduces a novel fuzzy rule-based learning framework that aims to enhance the interpretability of neural networks by bridging tarXiv cs.LGAug 19· 61%ResearchValid Per-Field Selective Risk Control for Document Extraction: Three Failure Modes, a Validity Ladder, and When Conditioning PaysA recent study published on arXiv presents a novel approach to managing selective risk in document extraction systems. The researchers identified three distinctarXiv cs.LGAug 19· 61%ResearchMemory Is Communication: The Frontier Between Remembering and SignalingRecent research has introduced the concept of a remembering-signaling frontier, which explores the balance between memory retention and communication among bounarXiv cs.AIAug 19· 61%ResearchSMOPD: Selective Token-Entropy Masking for Dirty-History Multi-Turn On-Policy Self-DistillationThe recent paper on SMOPD presents a novel approach to enhance multi-turn on-policy self-distillation in machine learning models. By implementing a technique tharXiv cs.LGAug 19· 60%ResearchRandomly initialized autoencoders: fixed points and edge-of-chaosA recent paper published on arXiv delves into the dynamics of randomly initialized autoencoders, focusing on their fixed points and stability within the framewoarXiv cs.LGAug 19· 60%ResearchEarly Cycle Charge Trajectory Generative Prediction and Full Life Cycle Health Management of Iron-Chromium Flow Batteries Based on FlowBD-E1A recent study has introduced FlowBD-E1, a novel generative forecasting framework designed to predict the charge voltage and current trajectories of iron-chromiarXiv cs.LGAug 19· 60%ResearchThe Price of Thinking: Reasoning Effort as a Model-Specific API ContractA recent study published on arXiv delves into the relationship between reasoning effort and API contracts for AI models. The researchers found that when AI modearXiv cs.AIAug 19· 59%