2026-10-28‹ Previous day Next day ›The week's recap

AI news, Wednesday, October 28, 2026

38 stories

The day's stories

Model ReleaseGoogle DeepMind launches Gemini 4 Carbon, its most advanced AI model yetGoogle DeepMind has unveiled its latest AI model, Gemini 4 Carbon, which is touted as the most advanced iteration to date. This model showcases remarkable enhanGoogle DeepMind Blog · OpenAI News · The Verge AIOct 29· 75%Product LaunchNVIDIA launches Nemotron 3.5 Transcribe for multilingual audio transcriptionNVIDIA has introduced the Nemotron 3.5 Transcribe model, which enhances audio transcription capabilities with support for 40 languages and the ability to differHugging Face Blog · TechCrunch AIOct 28· 67%Product LaunchLegora develops agentic operating system to enhance legal workflowsLegora has announced the development of an agentic operating system designed specifically for the legal sector. This new system integrates OpenAI models, such aOpenAI NewsOct 28· 62%Product LaunchOpenAI and ServiceNow enhance partnership to integrate AI into enterprise workflowsOpenAI and ServiceNow have announced an expansion of their partnership, focusing on the integration of OpenAI's AI models and agents into ServiceNow's platform.OpenAI NewsOct 28· 61%Product LaunchAnthropic launches reporting tools to clarify AI spending for enterprisesAnthropic has introduced new reporting tools designed to help enterprise customers gauge the value of their investments in AI, specifically focusing on its codiTechCrunch AIOct 28· 61%FundingAnthropic's IPO expected to yield billions for VCs and $100 billion for AmazonAnthropic is preparing for an initial public offering (IPO) that is anticipated to bring substantial financial returns for its venture capital investors. DespitTechCrunch AIOct 28· 59%Product LaunchChrome's autofill feature now uses Gmail data to streamline form fillingGoogle has rolled out an upgrade to Chrome's autofill feature, enabling it to access and utilize information from users' Gmail accounts. This enhancement allowsTechCrunch AI · The Verge AIOct 28· 58%ResearchAI integration strategies can enhance workplace collaboration and efficiencyThe recent discussion on the future of firms emphasizes the transformative role of AI in organizational structures. It highlights how AI can complement human roOpenAI NewsOct 28· 58%FundingMicrosoft earns $10 billion from OpenAI partnership in Q3 2023In its latest financial report, Microsoft announced that it generated $10 billion in revenue from its collaboration with OpenAI for the quarter ending SeptemberTechCrunch AIOct 28· 57%FundingMicrosoft reports Azure revenue at $33.2 billion for Q3, a 48% year-over-year increaseMicrosoft has officially disclosed Azure's revenue for the first time, reporting an impressive $33.2 billion for the July-September quarter. This figure marks aTechCrunch AIOct 28· 57%Product LaunchGoogle launches Connected Apps Platform for Gemini and Search integrationGoogle has unveiled its Connected Apps Platform, enabling businesses to integrate their applications with the company's AI services in Search and Gemini. This nTechCrunch AIOct 28· 57%Product LaunchMicrosoft discontinues Copilot support for smart TVs from LG and SamsungMicrosoft has announced the end of support for its Copilot AI assistant on smart TVs, specifically those produced by LG and Samsung. This decision comes nearly The Verge AIOct 28· 57%OtherMarissa Mayer's AI app Dazzle criticized for disconnect with public needsMarissa Mayer, the former CEO of Yahoo, has launched an AI application called Dazzle, designed to assist parents with various tasks. However, the app has faced TechCrunch AIOct 28· 57%SafetyWiz's CTO warns of impending 'vulnpocalypse' due to critical PHP vulnerabilitiesWiz's CTO, Tomer Schwartz, has raised alarms about significant vulnerabilities found in PHP, a programming language that underpins approximately 70% of websitesTechCrunch AIOct 28· 57%Product LaunchGoogle launches AI Search ads globally amid slowing core ad growthGoogle has announced the global expansion of its AI-powered Search ads as it faces a slowdown in growth from its traditional search advertising. Despite reportiTechCrunch AIOct 28· 56%FundingMicrosoft achieves $92.3 billion revenue, driven by Azure and OpenAI contributionsMicrosoft has reported a remarkable revenue of $92.3 billion for the latest quarter, marking its tenth consecutive quarter of record-breaking growth. This impreTechCrunch AIOct 28· 56%SafetySouth Korea's financial sector faces cyberattacks linked to suspected China-based hackerThe financial industry in South Korea is grappling with a recent surge of cyberattacks that have targeted prominent institutions, including Daishin Securities, TechCrunch AIOct 28· 56%RegulationMeta likely to postpone new business terms amid publisher backlash over AI useMeta Platforms is contemplating a delay in the rollout of its new business terms, originally set for November 20th, due to significant pushback from publishers The Verge AIOct 28· 55%Product LaunchMeta launches Waddle World, an AI pet journal app for virtual creature interactionMeta has introduced Waddle World, an innovative app designed as a 'tamagotchi-style' pet journal. In this new platform, users can engage with virtual creatures The Verge AIOct 28· 54%OtherAdam Mosseri predicts increasing negativity towards AI among students and creativesDuring a recent interview, Adam Mosseri, the head of Instagram, addressed the growing concerns and negativity surrounding artificial intelligence, particularly The Verge AIOct 28· 52%ResearchMARQ introduces a modular, LLM-agnostic requirement quality assessorThe recent study presents MARQ, a novel modular requirement quality assessor that is not tied to any specific large language model (LLM). This tool integrates sarXiv cs.AIOct 28· 64%ResearchNew research shows simple agentic loops can outperform complex theorem proving methodsRecent research has highlighted the effectiveness of simple agentic loops in automated theorem proving, achieving state-of-the-art results and solving previouslarXiv cs.LGOct 28· 64%ResearchNew model estimates sports players' time-to-arrival using quantile regressionA recent study has developed a novel method for estimating the time-to-arrival (TTA) of players in various sports, utilizing conditional quantile regression on arXiv cs.AIOct 28· 63%ResearchResearchers develop StruPO, a structure-aware prompt editor for text-to-image generationA team of researchers has introduced StruPO, an innovative structure-aware prompt editor designed to enhance the fidelity of text-to-image generation. By utilizarXiv cs.AIOct 28· 63%ResearchNew framework achieves 99.98% success in autonomous rocket landingsA recent paper introduces a curriculum-driven deep reinforcement learning framework specifically designed for the autonomous vertical landing of reusable launcharXiv cs.LGOct 28· 63%ResearchResearchers improve deep learning models for olive tree disease classificationA recent study has introduced a novel data generation strategy aimed at enhancing the performance of deep learning models specifically for diagnosing olive treearXiv cs.AIOct 28· 63%ResearchNew DISCO framework enhances distributed LLM inference efficiencyA novel framework named DISCO has been introduced to optimize distributed inference for large language models (LLMs). This approach focuses on communication effarXiv cs.LGOct 28· 63%ResearchResearchers develop CEGAN to optimize aerodynamic designs for commercial aircraftA team of researchers has introduced a novel Condition-Enhanced Generative Adversarial Network (CEGAN) aimed at improving the aerodynamic configuration design oarXiv cs.LGOct 28· 62%ResearchKORA architecture reduces AI model invocations by 80% in deterministic scenariosThe newly introduced KORA architecture focuses on deterministic-first execution control for AI workloads, aiming to enhance efficiency in task handling. By prioarXiv cs.LGOct 28· 62%ResearchNew ResCom method enhances low-bit quantization accuracy in language modelsThe recently published paper on arXiv presents ResCom, a novel method designed to enhance the accuracy of low-bit quantization in large language models. This tearXiv cs.LGOct 28· 62%ResearchResearchers develop Skill Dreaming framework to enhance coding agents' abilitiesA new research framework called Skill Dreaming has been introduced to improve repository-specific skills by analyzing pull request (PR) streams. This innovativearXiv cs.AIOct 28· 62%ResearchCohortFlow introduces a steerable agent for efficient clinical cohort curationCohortFlow is a newly developed steerable clinical cohort curation agent designed to streamline the process of updating eligibility criteria for clinical trialsarXiv cs.AIOct 28· 62%ResearchStudy shows knowledge distillation improves diabetic retinopathy grading modelsA recent study has investigated the effectiveness of knowledge distillation in diabetic retinopathy grading by utilizing various teacher architectures, such as arXiv cs.LGOct 28· 61%ResearchNew model quantifies time savings and error reduction from AI assistanceResearchers have developed a novel model aimed at measuring the time saved and defects avoided through the use of AI assistance. This model distinguishes betweearXiv cs.AIOct 28· 61%ResearchStudy finds 2.5D models excel in classifying pulmonary nodule malignancy riskA recent study evaluated the performance of 2D, 2.5D, and 3D residual networks in classifying the malignancy risk of pulmonary nodules. Utilizing a dataset comparXiv cs.LGOct 28· 61%ResearchNew method improves resolution of knowledge conflicts in language modelsResearchers have introduced an innovative approach to address knowledge conflicts in retrieval-augmented language models. This new method emphasizes the importaarXiv cs.AIOct 28· 60%ResearchNew method enhances uncertainty calibration for world models under policy changesA recent paper presents a novel approach to calibrating uncertainty in world models when deployment policies change. The authors introduce a method called weigharXiv cs.LGOct 28· 60%ResearchNew framework emphasizes human roles in science alongside AI capabilitiesA recent paper discusses the growing capabilities of AI in scientific research and the importance of maintaining human involvement in the scientific process. ItarXiv cs.AIOct 28· 59%