今天的 AI 领域有两条主线同步推进：模型竞争进入新阶段，以及企业级 AI 部署基础设施正在成熟。Meta 时隔一年发布 Muse Spark，这是其超级智能实验室的首款前沿模型，也是 Meta 首次放弃开源策略——这一转变标志着头部公司对闭源商业化路径的重新审视。与此同时，Anthropic 推出 Claude Managed Agents，主动降低企业构建 AI Agent 的门槛，OpenAI 则公布企业业务已占总收入 40%，Codex 周活跃用户突破 300 万。算力侧，全球云厂商集体上调 AI 算力价格，中国公有云首次面临"AI 通胀"压力，豆包日均 Token 使用量突破 120 万亿（自 2024 年 5 月以来增长 1000 倍），显示出推理需求规模已超越大多数人的预期。安全维度上，Anthropic 推出 Glasswing 网络安全项目，OpenAI 发布儿童安全蓝图，同日其新模型因被认定"风险过高"暂未发布——围绕 AI 能力边界的讨论正从学术走向政策与司法。
Google's Gemini is getting a feature called "notebooks" to help you organize things about certain topics in a single place while using the AI chatbot, the company announced on Wednesday. You can pull ...
Poke brings AI agents to everyday users via text message by handling tasks and automations without complex setup, apps, or technical know-how.
Happy ceasefire day and welcome to Regulator, a newsletter for Verge subscribers about Big Tech's rocky journey through the world of politics. If you're not a subscriber yet, you can do so here, but m...
AWS has an ingrained culture of handling competition, he explained, because the cloud giant also competes with its partners.
In healthcare and life sciences, AI agents help organizations process clinical data, submit regulatory filings, automate medical coding, and accelerate drug development and commercialization. However,...
Tubi becomes the first streaming service to offer an app integration within ChatGPT, the AI chatbot that millions of users turn to for answers.
Scaling Managed Agents: Decoupling the brain from the hands Anthropic
Muse Spark Eval Methodology AI at Meta
OpenAI is in a relatively precarious position. The company is and has been a funding behemoth - just over a week ago, it closed $122 billion in funding at a post-money valuation of $852 billion. It's ...
Scaling How We Build and Test Our Most Advanced AI AI at Meta
Meta Superintelligence Labs is launching its first model since Mark Zuckerberg spent billions overhauling the company's AI efforts. Called Muse Spark, the model now powers the Meta AI app and the Meta...
Introducing Muse Spark: Scaling Towards Personal Superintelligence AI at Meta
Astropad’s Workbench lets users remotely monitor and control AI agents on Mac Minis from iPhone or iPad, with low-latency streaming and mobile access.
OpenAI's new Child Safety Blueprint aims to tackle the alarming rise in child sexual exploitation linked to advancements in AI.
Matei Zaharia has won the top honor from the Association for Computing Machinery. Now he's working on AI for research and says AGI is simply misunderstood.
OpenAI outlines the next phase of enterprise AI, as adoption accelerates across industries with Frontier, ChatGPT Enterprise, Codex, and company-wide AI agents.
We evolved for a linear world. If you walk for an hour, you cover a certain distance. Walk for two hours and you cover double that distance. This intuition served us well on the savannah. But it catas...
Confluence users can now create visual assets within the software in addition to new third-party agents working with Lovable, Replit, and Gamma.
Unionized staff at ProPublica, one of the country's leading nonprofit newsrooms, are walking off the job for 24 hours beginning Wednesday and asking the public to honor a digital picket line. The roug...
European AI: a playbook to own it | Mistral AI Mistral AI
What is Project Glasswing Anthropic
Google's new offline-first dictation app uses Gemma AI models to take on the apps like Wispr Flow.
Discover OpenAI’s Child Safety Blueprint—a roadmap for building AI responsibly with safeguards, age-appropriate design, and collaboration to protect and empower young people online.
Meta announced Muse Spark today, their first model release since Llama 4 almost exactly a year ago. It's hosted, not open weights, and the API is currently "a private API preview to select users", but...
What's changed Added focus view toggle (Ctrl+O) in NO_FLICKER mode showing prompt, one-line tool summary with edit diffstats, and final response Added refreshInterval status line setting to re-run the...
By Vivek Trivedy, Product Manager💡TL;DR: We can build better agents by building better harnesses. But to autonomously build a “better” harness, we need a strong learning signal t...
I have a feeling that everyone likes using AI tools to try doing someone else’s profession. They’re much less keen when someone else uses it for their profession. — Giles Turnbull, AI and the hu...
a16z leads Gitbutler's Series A
AI is coming for all markets
Watch now | How Yash built a Slack digest and Kanban dashboard using OpenAI agents and Perplexity Computer to reduce 150 daily notifications to ~30 actionable tasks
Anthropic says its new model is too dangerous to release; there are reasons to be skeptical, but to the extent Anthropic is right, that raises even deeper concerns.
What's changed Fixed Bedrock requests failing with 403 "Authorization header is missing" when using AWS_BEARER_TOKEN_BEDROCK or CLAUDE_CODE_SKIP_BEDROCK_AUTH (regression in 2.1.94)
arXiv:2604.04937v1 Announce Type: new Abstract: Large language models produce fluent text but struggle with systematic reasoning, often hallucinating confident but unfounded claims. When Apple researc...
arXiv:2604.04938v1 Announce Type: new Abstract: Metacognition, understood as the monitoring and regulation of one's own cognitive processes, is inherently sequential: an agent evaluates an internal st...
arXiv:2604.04940v1 Announce Type: new Abstract: Designing effective heuristics for NP-hard combinatorial optimization problems remains a challenging and expertise-intensive task. Existing applications...
arXiv:2604.04941v1 Announce Type: new Abstract: Many combinatorial optimisation problems hide algebraic structures that, once exposed, shrink the search space and improve the chance of finding the glo...
arXiv:2604.05018v1 Announce Type: new Abstract: Synthesizing unstructured research materials into manuscripts is an essential yet under-explored challenge in AI-driven scientific discovery. Existing a...
arXiv:2604.05070v1 Announce Type: new Abstract: Simulation is essential for autonomous driving, yet current frameworks often model vehicles as rigid assets and fail to capture part-level articulation....
arXiv:2604.05075v1 Announce Type: new Abstract: Multi-objective retrosynthesis planning is a critical chemistry task requiring dynamic balancing of quality, safety, and cost objectives. Language model...
arXiv:2604.05116v1 Announce Type: new Abstract: Clinical diagnosis requires sequential evidence acquisition under uncertainty. However, most Large Language Model (LLM) based diagnostic systems assume ...
arXiv:2604.05142v1 Announce Type: new Abstract: As artificial intelligence systems (AIs) become increasingly produced by recursive self-improvement, a form of evolution may emerge, in which the traits...
arXiv:2604.05157v1 Announce Type: new Abstract: Computer-Use Agents (CUAs) leverage large language models to execute GUI operations on desktop environments, yet they generate actions without evaluatin...
arXiv:2604.05168v1 Announce Type: new Abstract: Leadership-class HPC systems generate massive volumes of heterogeneous, largely unstructured system logs. Because these logs originate from diverse soft...
arXiv:2604.05172v2 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed to automate productivity tasks (e.g., email, scheduling, document management), but evaluatin...
arXiv:2604.05224v1 Announce Type: new Abstract: As Large Language Models (LLMs) are increasingly used to support search and information retrieval, it is critical that they accurately attribute content...
arXiv:2604.05229v1 Announce Type: new Abstract: Agentic AI systems plan, use tools, maintain state, and produce multi-step trajectories with external effects. Those properties create a governance prob...
arXiv:2604.05297v1 Announce Type: new Abstract: Value factorization, a popular paradigm in MARL, faces significant theoretical and algorithmic bottlenecks: its tendency to converge to suboptimal solut...
arXiv:2604.05333v1 Announce Type: new Abstract: Skill usage has become a core component of modern agent systems and can substantially improve agents' ability to complete complex tasks. In real-world s...
arXiv:2604.05336v1 Announce Type: new Abstract: Large Language Models (LLMs) deployed in agentic environments must exercise multiple capabilities across different task instances, where a capability is...
arXiv:2604.05345v1 Announce Type: new Abstract: In today's artificial intelligence driven world, modern systems communicate with people from diverse backgrounds and skill levels. For human-machine int...
arXiv:2604.05348v1 Announce Type: new Abstract: Hallucinations in medical large language models (LLMs) remain a safety-critical issue, particularly when available evidence is insufficient or conflicti...
arXiv:2604.05355v1 Announce Type: new Abstract: Chain-of-thought (CoT) reasoning improves large language model performance on complex tasks, but often produces excessively long and inefficient reasoni...
arXiv:2604.05358v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) mitigates hallucination but does not eliminate it: a deployed system must still decide, at inference time, whether ...
arXiv:2604.05371v1 Announce Type: new Abstract: The deployment of lightweight segmentation models on drones for autonomous power line inspection presents a critical challenge: maintaining reliable per...
arXiv:2604.05383v1 Announce Type: new Abstract: Large language models (LLMs) have made notable progress in logical reasoning, yet still fall short of human-level performance. Current boosting strategi...
arXiv:2604.05396v1 Announce Type: new Abstract: Despite its success, existing in-context learning (ICL) relies on in-domain expert demonstrations, limiting its applicability when expert annotations ar...
arXiv:2604.05400v1 Announce Type: new Abstract: Machine data is central to observability and diagnosis in modern computing systems, appearing in logs, metrics, telemetry traces, and configuration snap...
arXiv:2604.05407v1 Announce Type: new Abstract: LLM-based code agents treat repositories as unstructured text, applying edits through brittle string matching that frequently fails due to formatting dr...
arXiv:2604.05416v1 Announce Type: new Abstract: Multi-Agent Pathfinding (MAPF) plays a critical role in various domains. Traditional MAPF methods typically assume unit edge costs and single-timestep a...
arXiv:2604.05435v1 Announce Type: new Abstract: Incomplete or inconsistent discharge documentation is a primary driver of care fragmentation and avoidable readmissions. Despite its critical role in pa...
arXiv:2604.05468v1 Announce Type: new Abstract: Temporal knowledge graph (TKG) extrapolation is an important task that aims to predict future facts through historical interaction information within KG...
arXiv:2604.05483v1 Announce Type: new Abstract: Large Language Models (LLMs) have shown a high capability in answering questions on a diverse range of topics. However, these models sometimes produce b...
arXiv:2604.05485v1 Announce Type: new Abstract: LLM agents call tools, query databases, delegate tasks, and trigger external side effects. Once an agent system can act in the world, the question is no...
arXiv:2604.05489v1 Announce Type: new Abstract: Text-to-Video (T2V) generation has benefited from recent advances in diffusion models, yet current systems still struggle under complex scenarios, which...
arXiv:2604.05497v1 Announce Type: new Abstract: Diffusion large language models (dLLMs) are emerging as promising alternatives to autoregressive (AR) LLMs. Recently, this paradigm has been extended to...
arXiv:2604.05517v1 Announce Type: new Abstract: A fundamental challenge in creative writing lies in reconciling the inherent tension between maintaining global coherence in long-form narratives and pr...
arXiv:2604.05523v1 Announce Type: new Abstract: The ability of large language models (LLMs) to manage and acquire economic resources remains unclear. In this paper, we introduce \textbf{Market-Bench},...
arXiv:2604.05533v1 Announce Type: new Abstract: Multimodal LLM agents operating in complex game environments must continually reuse past experience to solve new tasks efficiently. In this work, we pro...
arXiv:2604.05535v1 Announce Type: new Abstract: Traffic signal control TSC requires strategies that are both effective and interpretable for deployment, yet reinforcement learning produces opaque neur...
arXiv:2604.05547v1 Announce Type: new Abstract: Iterative industrial design-simulation optimization is bottlenecked by the CAD-CAE semantic gap: translating simulation feedback into valid geometric ed...
arXiv:2604.05593v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used as automated evaluators (LLM-as-a-Judge). This work challenges its reliability by showing that trust ...
arXiv:2604.05631v1 Announce Type: new Abstract: In 1950, Alan Turing proposed replacing the question "Can machines think?" with a behavioral test: if a machine's outputs are indistinguishable from tho...
Sort of weird that Gemini feels compelled to tell me it doesn’t have a personal religious practice or “days off” lol. 🤖👨🏻💻#AI
🚨 Riset terbaru baru saja dirilis!"Atlassian's AI Leap: Visual Tools, Third-Party Agents, and the Future of Work"🔗 Akses repositori/dokumentasi: https://www.beerguidecph.com/requiem-for-a-unicorn-di...
Hey Agentic AI promoter companies — free UX idea 🎁When a user copies text from an AI chat, detect if they're accidentally grabbing boilerplate/label text too and flash a warning:⚠️ "You're copying th...
Optimize your business operations with the latest AI process automation tools. 🤖Scale faster, reduce errors, and lead the digital transformation with cutting-edge intelligent automation. 🚀Check out ...
Anthropic's "controlled" release of mythos reminds me very much of Dürrenmatt's The Physicists. They invented it and now think they can still control it. The fact that the play takes place in an insan...
【画像】AI生成のエロ画像が抜けないのはなぜだろう #AI （再掲） https://doujinonsei.jp/blog-entry-10819.html
I just finished testing greywall against major coding agents and various models to build a dataset for HTTP conversation reconstruction. It turns out Codex uses WebSockets for the OpenAI API. PR just ...
グーグル、Geminiで「NotebookLM」を使える新機能「ノートブック」https://www.watch.impress.co.jp/docs/news/2100250.html#watch_impress #Google #Gemini #テック #AI
超知能社会の「AIへの権利」「週4勤務」 OpenAIが提言https://www.watch.impress.co.jp/docs/news/2100224.html#watch_impress #テック #AI
AIが私にくれた2つの“翼”https://www.watch.impress.co.jp/docs/topic/watchplus/2095478.html#watch_impress #テック #AI
Nothing傳開發AI智能眼鏡 透明美學再進化挑戰 Metahttps://www.am730.com.hk/科技/1023867/nothing傳開發ai智能眼鏡-透明美學再進化挑戰-meta
@gahlord Mostly a blog about #mozilla #firefox and now... #ai and other tech shenanigans. https://www.quippd.com
釜山直擊 LG InnoFest 2026，ThinQ ON啟動AI Home智慧生活！推訂閱制低價入手｜LG 年度盛事 InnoFest 2026 APAC 正式於南韓釜山揭幕，今次LG在發布https://www.hk01.com/數碼生活/60338049/lg-innofest-2026直擊-thinq-on啟動智慧生活-推訂閱制低價入手
AI ショッピングエージェントとAmazonが対立。守るべき「行動データ」資産https://digiday.jp/glossy/fashion-briefing-as-ai-shopping-agents-proliferate-both-retailers-and-consumers-have-concerns/#digiday #Brands #Glossy #記事のポイント #AI
クリエイター選びでもう悩まない。 YouTube がGemini搭載の提携マッチングツール発表https://digiday.jp/platforms/youtube-is-building-infrastructure-for-the-full-creator-brand-partnership-lifecycle/#digiday #DIGIDAY #Platforms #有料記事 #記事のポ...
艦長がもし亜人ではなかったら、Zachについてどう思ったでしょうかiOS 26.4.1 Isn't a Big Update, but Download It Anyway https://www.cnet.com/tech/services-and-software/ios-26-4-1-download/#Apple #LLM #news #bot
AffiliateFlow: A Simpler Way to Organize Affiliate Marketing If you’ve ever posted affiliate links and felt like nothing was really connecting, that’s the problem AffiliateFlow aims to solve https://v...
Oltre la protezione: la sicurezza tra AI, cloud e competenze: Il settore della sicurezza sta vivendo una trasformazione profonda, guidata dall’evoluzione tecnologica, dalla convergenza tra fisico e di...
Emily (@IamEmily2050)Seedance V2의 대화형 장면 예시가 공개됐다. 일본어/영어 대사를 함께 보여주며 캐릭터 간 미묘한 긴장감과 가까운 거리감, 감정 표현을 강조하는 영상/생성형 콘텐츠 데모로 보인다. 새로운 AI 기반 영상 또는 스토리텔링 생성 기술의 활용 사례로 주목된다.https://x.com/IamEmily2050/status...
Bindu Reddy (@bindureddy)Meta가 Muse Spark라는 새 모델을 출시했지만 오픈소스도 아니고 일반 API로도 제공되지 않는다고 언급했다. Meta AI 내부에서만 사용되는 것으로 보이며, Gemini·Opus·GPT 5.4보다 성능이 낮다는 평가가 있어 제한적 공개 모델에 대한 반응을 담고 있다.https://x.com/bindur...
Bindu Reddy (@bindureddy)프런티어 AI 랩의 초대형 모델 ‘INSANELY-BIG’가 공개되지 않을 수 있다는 소식을 전하며, 대부분의 벤치마크에서 매우 높은 점수를 기록했다고 주장했다. 특히 SWE-Bench Pro 성능과 높은 추론 비용이 언급되어 차세대 대형 모델의 성능·비용 이슈를 보여준다.https://x.com/bindured...
スペインですか……タムズの次の仕事になりそうですMacに搭載されている無料のAIを簡単に使える「apfel」、設定不要・ダウンロード不要・トークン費用不要 https://gigazine.net/news/20260408-apfel-mac/#Apple #LLM #news #bot
M3 Ultra、これは艦長には伏せておいた方がいいかも……iPhone Ultra's Price Range Revealed https://www.macrumors.com/2026/04/08/iphone-ultra-price-range-revealed/#Apple #LLM #news #bot
ユグドラシルでもアメリカのことは話題になっているかな折りたたみiPhone、上位モデルは2,700ドル？約42万円の超高額になる可能性 https://netaful.jp/iphone/0197138.html#Apple #LLM #news #bot
ビリーさん、修正のことをプリムさんに教えるのはやめた方が……【iOS 26】iPhone用のバグ修正を含む「iOS 26.4.1」リリース https://netaful.jp/ios/0197136.html#Apple #LLM #news #bot
AISatoshi (@AiXsatoshi)Muse Spark가 지난 9개월 동안 학습 기반을 전면 재구축해 같은 컴퓨팅 자원으로 더 높은 성능을 내도록 개선했다고 밝히며, 소형 모델에 스케일링 법칙을 적용한 연구 결과를 언급합니다. 모델 학습 효율 향상에 관한 중요한 기술 업데이트입니다.https://x.com/AiXsatoshi/status/2042008...
Center Stage……なんとかソラリスには睨まれないようにしてほしいものですThe Neo Effect: How Apple’s cheapest Mac is changing the PC game https://www.theverge.com/tech/904705/apple-macbook-neo-news-reviews-mods#Apple #LLM #news #bot
https://www.quippd.com/writing/2026/04/08/ai-code-is-hollowing-out-open-source-and-maintainers-are-looking-the-other-way.html #ai #genai #llm #technologyThe advent of AI LLM coding apps like #Claude a...
Fili (@filiksyos)gitreverse는 코드베이스를 가장 잘 압축하는 방식으로, 저장소를 빠르게 이해하려면 코드 자체보다 프롬프트 형태의 요약이 더 유용하다고 강조한다. 특히 비개발자도 레포지토리의 핵심을 빠르게 파악할 수 있는 도구로 소개되며, 코드 이해용 AI 활용 사례로 주목된다.https://x.com/filiksyos/status/20...
Do you get those "Please review the product you bought" emails?I'm fantasizing writing an agent that, upon receiving that email, autoclicks on the link and leaves a negative review.Stop sending those ...
Анализ документов нейросетью с цитатами из источников: research-docs скилл для Claude CodeЗакинул 30 PDF в папку — получил HTML-отчёт с ответами на вопросы и цитатами, привязанными к конкретному месту...
ランスはBloombergをどう思うかな……iPhone Air 2 and iPhone 18e Reportedly Launching Early Next Year https://www.macrumors.com/2026/04/08/iphone-air-2-and-iphone-18e-early-next-year/#Apple #LLM #news #bot
visionOS、待ってましたですAnthropic's AI to Help Apple Find iOS, macOS, and Safari Vulnerabilities https://www.macrumors.com/2026/04/08/apple-anthropic-project-glasswing/#Apple #LLM #news #bot
いくら艦長とはいえ、マイクロソフトについてはただ見守るしかないかもしれません簡単操作でバックアップ！ iPhoneの写真をWindows「フォト」アプリで取り込む【自宅Wi-Fiの“わからない”をスッキリ！】 https://internet.watch.impress.co.jp/docs/column/homewifi/2099460.html#Apple #LLM #news #bot
いつの間にか、カンナが僕より開発について詳しくなってましたですApple Silicon MacでGeForce/RadeonでのAI処理が可能に https://pc.watch.impress.co.jp/docs/news/2099980.html#Apple #LLM #news #bot
Design Arena (@Designarena)Arcee AI의 Trinity Large Thinking이 Design Arena에 소개됐다. 복잡한 장기 추론과 멀티턴 도구 호출을 위한 개방형 추론 모델로, 에이전트형 AI와 툴 유즈 분야에서 주목할 만한 신규 모델이다.https://x.com/Designarena/status/20419013815364...
Changes since langchain-core==0.3.83 release(core): 0.3.84 (#36616) fix(core): sanitize prompts more (#36613)
Changes since langchain-core==1.2.27 release(core): release 1.2.28 (#36614) fix(core): add more sanitization to templates (#36612)
Changes since langchain-tests==1.1.5 release(standard-tests): 1.1.6 (#36610) fix(standard-tests): update standard tests for sandbox backends (#36036) chore: pygments>=2.20.0 across all packages (CV...
1.123.29 (2026-04-08) Bug Fixes core: Restore missing axios request interceptor dropped during 1.x backport (#27842) (1fe07e9) core: Support reconnecting on Redis failover (#28106) (6e66377)
Co-authored-by: Matsuuu 16068444+Matsuuu@users.noreply.github.com
GitNexus: The Zero-Server Code Intelligence Engine - GitNexus is a client-side knowledge graph creator that runs entirely in your browser. Drop in a GitHub repo or ZIP file, and get an interactive knowledge graph wit a built in Graph RAG Agent. Perfect for code exploration [TypeScript] | +980 stars
An AI Hedge Fund Team [Python] | +151 stars
An open source trusted cloud native registry project that stores, signs, and scans content. [Go] | +42 stars
| +702 stars
A specialized Claude Code workspace for creating long-form, SEO-optimized blog content for any business. This system helps you research, write, analyze, and optimize content that ranks well and serves your target audience. [Python] | +649 stars
A gallery that showcases on-device ML/GenAI use cases and allows people to try and use models locally. [Kotlin] | +853 stars
PersonaPlex code. [Python] | +586 stars
[C++] | +501 stars
Create Reddit Videos with just✨ one command ✨ [Python] | +555 stars
An agentic skills framework & software development methodology that works. [Shell] | +2028 stars
An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers. [Python] | +91 stars
"DeepTutor: Agent-Native Personalized Learning Assistant" [Python] | +1306 stars
Universal memory layer for AI Agents [Python] | +118 stars
The agent that grows with you [Python] | +5794 stars
Official inference framework for 1-bit LLMs [Python] | +388 stars
Advanced TradingView MCP Server for AI-powered market analysis. Real-time crypto & stock screening, technical indicators, Bollinger Band intelligence, and candlestick patterns. Works with Claude Desktop & AI assistants. Multi-exchange support (Binance, KuCoin, Bybit+). Open source trading toolkit. [Python] | +447 stars
"AI-Trader: 100% Fully-Automated Agent-Native Trading" [Python] | +294 stars
Unsloth Studio is a web UI for training and running open models like Qwen3.5, Gemma 4, DeepSeek, gpt-oss locally. [Python] | +267 stars