https://24-ai.news/hr/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI hr 2026-06-30T00:00:00Z Claude Science: Anthropicov samostalni radni stol za znanstvenike u life sciences https://24-ai.news/hr/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI hr 2026-06-30T00:00:00Z Claude Sonnet 5: Anthropicov najagentičniji model postaje novi standard https://24-ai.news/hr/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI hr 2026-06-30T00:00:00Z Fable 5 se vraća: vlada ukinula izvozne kontrole, globalni redeploy počinje 1. srpnja https://24-ai.news/hr/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI hr 2026-06-30T00:00:00Z MARS: Tekstualni smjerovi odbijanja štite multimodalne AI modele bez dodatnog treniranja https://24-ai.news/hr/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI hr 2026-06-30T00:00:00Z TRIAGE: Kako dodijeliti zasluge pravim tokenima u agentnom pojačajnom učenju https://24-ai.news/hr/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI hr 2026-06-30T00:00:00Z CNCF Kepler prerađen od temelja: precizno mjerenje energije Kubernetes Podova bez privilegija kernela https://24-ai.news/hr/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI hr 2026-06-30T00:00:00Z Claude Sonnet 5 u GitHub Copilotu i agent mode u JetBrainsu: dupli udar za razvojne timove https://24-ai.news/hr/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI hr 2026-06-30T00:00:00Z Google lansira dva modela istog dana: Gemini Omni Flash i Nano Banana 2 Lite https://24-ai.news/hr/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI hr 2026-06-30T00:00:00Z Google Research predstavlja TabFM: zero-shot foundation model za tablične podatke https://24-ai.news/hr/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI hr 2026-06-30T00:00:00Z LangChain: Kako pokrenuti nepouzdani agentski kod bez vanjskog sandboxa https://24-ai.news/hr/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI hr 2026-06-30T00:00:00Z ONNX v1.22.0 donosi native attention operatore za LLM-ove i WebAssembly podršku https://24-ai.news/hr/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI hr 2026-06-30T00:00:00Z SkillOpt: Microsoft Research tretira instrukcijske datoteke agenata kao trenirajuće parametre https://24-ai.news/hr/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI hr 2026-06-30T00:00:00Z NVIDIA: softverski stack na Blackwellu smanjio trošak tokena DeepSeek V4 pet puta za mjesec dana https://24-ai.news/hr/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI hr 2026-06-30T00:00:00Z OECD: Upotreba GenAI chatbota narasla s 18% na 28% populacije u jednoj godini https://24-ai.news/hr/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI hr 2026-06-30T00:00:00Z Miles: PyTorch-nativni open-source okvir za RL post-training LLM-ova frontier razmjera https://24-ai.news/en/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI en 2026-06-30T00:00:00Z Claude Science: Anthropic's Dedicated Workbench for Life Sciences Researchers https://24-ai.news/en/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI en 2026-06-30T00:00:00Z Claude Sonnet 5: Anthropic's Most Agentic Model Becomes the New Standard https://24-ai.news/en/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI en 2026-06-30T00:00:00Z Fable 5 Returns: Government Lifts Export Controls, Global Redeployment Begins July 1 https://24-ai.news/en/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI en 2026-06-30T00:00:00Z MARS: Textual Refusal Directions Protect Multimodal AI Models Without Additional Training https://24-ai.news/en/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI en 2026-06-30T00:00:00Z TRIAGE: How to Assign Credit to the Right Tokens in Agentic Reinforcement Learning https://24-ai.news/en/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI en 2026-06-30T00:00:00Z CNCF Kepler Rearchitected From the Ground Up: Precise Pod Energy Measurement Without Kernel Privileges https://24-ai.news/en/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI en 2026-06-30T00:00:00Z Claude Sonnet 5 in GitHub Copilot and Agent Mode in JetBrains: A Double Update for Dev Teams https://24-ai.news/en/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI en 2026-06-30T00:00:00Z Google Launches Two Models on the Same Day: Gemini Omni Flash and Nano Banana 2 Lite https://24-ai.news/en/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI en 2026-06-30T00:00:00Z Google Research Introduces TabFM: A Zero-Shot Foundation Model for Tabular Data https://24-ai.news/en/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI en 2026-06-30T00:00:00Z LangChain: How to Run Untrusted Agent Code Without an External Sandbox https://24-ai.news/en/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI en 2026-06-30T00:00:00Z ONNX v1.22.0 Brings Native Attention Operators for LLMs and WebAssembly Support https://24-ai.news/en/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI en 2026-06-30T00:00:00Z SkillOpt: Microsoft Research Treats Agent Instruction Files as Trainable Parameters https://24-ai.news/en/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI en 2026-06-30T00:00:00Z NVIDIA: Software Stack on Blackwell Cut DeepSeek V4 Token Cost Five-Fold in One Month https://24-ai.news/en/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI en 2026-06-30T00:00:00Z OECD: GenAI Chatbot Usage Has Grown From 18% to 28% of the Population in One Year https://24-ai.news/en/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI en 2026-06-30T00:00:00Z Miles: A PyTorch-Native Open-Source Framework for RL Post-Training of Frontier-Scale LLMs https://24-ai.news/de/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI de 2026-06-30T00:00:00Z Claude Science: Anthropics eigenständige Workbench für Wissenschaftler in den Life Sciences https://24-ai.news/de/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI de 2026-06-30T00:00:00Z Claude Sonnet 5: Anthropics agentischstes Modell wird zum neuen Standard https://24-ai.news/de/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI de 2026-06-30T00:00:00Z Fable 5 kehrt zurück: Regierung hebt Exportkontrollen auf, globales Redeployment beginnt am 1. Juli https://24-ai.news/de/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI de 2026-06-30T00:00:00Z MARS: Textuelle Ablehnungsrichtungen schützen multimodale KI-Modelle ohne zusätzliches Training https://24-ai.news/de/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI de 2026-06-30T00:00:00Z TRIAGE: Wie man den richtigen Token in agentischem Reinforcement Learning Verdienst zuweist https://24-ai.news/de/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI de 2026-06-30T00:00:00Z CNCF Kepler von Grund auf neu gebaut: präzise Energiemessung von Kubernetes-Pods ohne Kernel-Privilegien https://24-ai.news/de/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI de 2026-06-30T00:00:00Z Claude Sonnet 5 in GitHub Copilot und Agent Mode in JetBrains: Doppelschlag für Entwicklerteams https://24-ai.news/de/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI de 2026-06-30T00:00:00Z Google startet zwei Modelle am selben Tag: Gemini Omni Flash und Nano Banana 2 Lite https://24-ai.news/de/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI de 2026-06-30T00:00:00Z Google Research stellt TabFM vor: Zero-Shot-Foundation-Modell für tabellarische Daten https://24-ai.news/de/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI de 2026-06-30T00:00:00Z LangChain: Nicht vertrauenswürdigen Agenten-Code ohne externen Sandbox ausführen https://24-ai.news/de/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI de 2026-06-30T00:00:00Z ONNX v1.22.0 bringt native Attention-Operatoren für LLMs und WebAssembly-Unterstützung https://24-ai.news/de/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI de 2026-06-30T00:00:00Z SkillOpt: Microsoft Research behandelt Agenten-Instructiondateien als trainierbare Parameter https://24-ai.news/de/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI de 2026-06-30T00:00:00Z NVIDIA: Software-Stack auf Blackwell senkt Token-Kosten für DeepSeek V4 in einem Monat um das Fünffache https://24-ai.news/de/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI de 2026-06-30T00:00:00Z OECD: GenAI-Chatbot-Nutzung stieg von 18 % auf 28 % der Bevölkerung in einem Jahr https://24-ai.news/de/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI de 2026-06-30T00:00:00Z Miles: PyTorch-natives Open-Source-Framework für RL-Post-Training von LLMs im Frontier-Maßstab https://24-ai.news/zh/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI zh 2026-06-30T00:00:00Z Claude Science：Anthropic 为生命科学研究者推出独立工作台 https://24-ai.news/zh/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI zh 2026-06-30T00:00:00Z Claude Sonnet 5：Anthropic 最具智能体能力的模型成为新标准 https://24-ai.news/zh/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI zh 2026-06-30T00:00:00Z Fable 5 回归：政府解除出口管制，全球重新部署于 7 月 1 日开始 https://24-ai.news/zh/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI zh 2026-06-30T00:00:00Z MARS：文本拒绝方向保护多模态 AI 模型，无需额外训练 https://24-ai.news/zh/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI zh 2026-06-30T00:00:00Z TRIAGE：如何在智能体强化学习中将功劳归因给正确的 token https://24-ai.news/zh/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI zh 2026-06-30T00:00:00Z CNCF Kepler 重构：无需内核权限即可精确测量 Kubernetes Pod 的能耗 https://24-ai.news/zh/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI zh 2026-06-30T00:00:00Z Claude Sonnet 5 登陆 GitHub Copilot，智能体模式扩展至 JetBrains：开发团队的双重礼包 https://24-ai.news/zh/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI zh 2026-06-30T00:00:00Z Google 同日发布两款模型：Gemini Omni Flash 与 Nano Banana 2 Lite https://24-ai.news/zh/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI zh 2026-06-30T00:00:00Z Google Research 发布 TabFM：面向表格数据的零样本基础模型 https://24-ai.news/zh/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI zh 2026-06-30T00:00:00Z LangChain：如何在无需外部沙箱的情况下运行不受信任的智能体代码 https://24-ai.news/zh/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI zh 2026-06-30T00:00:00Z ONNX v1.22.0 带来面向大语言模型的原生注意力算子和 WebAssembly 支持 https://24-ai.news/zh/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI zh 2026-06-30T00:00:00Z SkillOpt：微软研究院将智能体指令文件视为可训练参数 https://24-ai.news/zh/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI zh 2026-06-30T00:00:00Z NVIDIA：Blackwell 软件堆栈在一个月内将 DeepSeek V4 的 token 成本降低五倍 https://24-ai.news/zh/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI zh 2026-06-30T00:00:00Z OECD：一年内生成式 AI 聊天机器人使用率从 18% 增至 28% https://24-ai.news/zh/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI zh 2026-06-30T00:00:00Z Miles：面向前沿规模大语言模型的 PyTorch 原生开源强化学习后训练框架 https://24-ai.news/ja/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI ja 2026-06-30T00:00:00Z Claude Science：Anthropicがライフサイエンス研究者向けに独立したワークベンチを公開 https://24-ai.news/ja/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI ja 2026-06-30T00:00:00Z Claude Sonnet 5：AnthropicのAIエージェント特化型最新モデルが新標準を確立 https://24-ai.news/ja/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI ja 2026-06-30T00:00:00Z Fable 5復活：輸出規制解除を受け、7月1日からグローバル再展開を開始 https://24-ai.news/ja/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI ja 2026-06-30T00:00:00Z MARS：追加トレーニングなしでマルチモーダルAIモデルを保護するテキスト拒否方向 https://24-ai.news/ja/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI ja 2026-06-30T00:00:00Z TRIAGE：エージェント強化学習において適切なトークンに適切なクレジットを割り当てる方法 https://24-ai.news/ja/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI ja 2026-06-30T00:00:00Z CNCF Keplerが根本から再設計：カーネル特権なしでKubernetes Podの電力を精密計測 https://24-ai.news/ja/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI ja 2026-06-30T00:00:00Z Claude Sonnet 5がGitHub Copilotに登場、JetBrainsにはエージェントモード：開発チームへの二重の恩恵 https://24-ai.news/ja/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI ja 2026-06-30T00:00:00Z Googleが同日に二つのモデルを公開：Gemini Omni FlashとNano Banana 2 Lite https://24-ai.news/ja/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI ja 2026-06-30T00:00:00Z Google Researchが TabFM を発表：表形式データ向けゼロショット基盤モデル https://24-ai.news/ja/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI ja 2026-06-30T00:00:00Z LangChain：外部サンドボックスなしで信頼できないエージェントコードを実行する方法 https://24-ai.news/ja/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI ja 2026-06-30T00:00:00Z ONNX v1.22.0：LLM向けネイティブAttentionオペレーターとWebAssemblyサポートを搭載 https://24-ai.news/ja/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI ja 2026-06-30T00:00:00Z SkillOpt：Microsoft Researchがエージェントの指示ファイルを訓練可能パラメータとして扱う https://24-ai.news/ja/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI ja 2026-06-30T00:00:00Z NVIDIA：BlackwellのソフトウェアスタックがDeepSeek V4のトークンコストを1カ月で5倍削減 https://24-ai.news/ja/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI ja 2026-06-30T00:00:00Z OECD：GenAIチャットボットの利用率が1年で人口の18%から28%に増加 https://24-ai.news/ja/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI ja 2026-06-30T00:00:00Z Miles：フロンティア規模のLLMのRLポストトレーニング向けPyTorchネイティブオープンソースフレームワーク https://24-ai.news/ko/news/2026-06-30/anthropic-claude-science-workbench/ 24 AI ko 2026-06-30T00:00:00Z Claude Science: 생명과학 연구자를 위한 Anthropic의 독립형 워크벤치 https://24-ai.news/ko/news/2026-06-30/anthropic-claude-sonnet-5/ 24 AI ko 2026-06-30T00:00:00Z Claude Sonnet 5: Anthropic의 가장 에이전틱한 모델이 새로운 표준이 되다 https://24-ai.news/ko/news/2026-06-30/anthropic-redeploying-fable-5/ 24 AI ko 2026-06-30T00:00:00Z Fable 5 귀환: 수출 규제 철폐, 7월 1일 글로벌 재배포 시작 https://24-ai.news/ko/news/2026-06-30/arxiv-mars-multimodal-refusal-steering/ 24 AI ko 2026-06-30T00:00:00Z MARS: 추가 훈련 없이 텍스트 거부 방향으로 멀티모달 AI 모델 보호 https://24-ai.news/ko/news/2026-06-30/arxiv-triage-role-typed-agentic-rl/ 24 AI ko 2026-06-30T00:00:00Z TRIAGE: 에이전트 강화학습에서 올바른 토큰에 공로를 귀속시키는 방법 https://24-ai.news/ko/news/2026-06-30/cncf-kepler-rearchitected-power/ 24 AI ko 2026-06-30T00:00:00Z CNCF Kepler, 처음부터 재설계: 커널 권한 없이 Kubernetes 파드의 정밀한 전력 측정 https://24-ai.news/ko/news/2026-06-30/github-copilot-sonnet-5-jetbrains-agent/ 24 AI ko 2026-06-30T00:00:00Z Claude Sonnet 5, GitHub Copilot 탑재 및 JetBrains 에이전트 모드: 개발팀을 위한 이중 공세 https://24-ai.news/ko/news/2026-06-30/google-gemini-omni-flash-nano-banana-2/ 24 AI ko 2026-06-30T00:00:00Z Google, 같은 날 두 모델 출시: Gemini Omni Flash와 Nano Banana 2 Lite https://24-ai.news/ko/news/2026-06-30/google-tabfm-tabular-foundation-model/ 24 AI ko 2026-06-30T00:00:00Z Google Research, TabFM 발표: 표 형식 데이터를 위한 제로샷 파운데이션 모델 https://24-ai.news/ko/news/2026-06-30/langchain-untrusted-agent-code-quickjs/ 24 AI ko 2026-06-30T00:00:00Z LangChain: 외부 샌드박스 없이 신뢰할 수 없는 에이전트 코드 실행하기 https://24-ai.news/ko/news/2026-06-30/lfai-onnx-1-22-attention-operators/ 24 AI ko 2026-06-30T00:00:00Z ONNX v1.22.0, LLM용 네이티브 어텐션 연산자와 WebAssembly 지원 도입 https://24-ai.news/ko/news/2026-06-30/microsoft-skillopt-trainable-skills/ 24 AI ko 2026-06-30T00:00:00Z SkillOpt: Microsoft Research가 에이전트 명령 파일을 훈련 가능한 파라미터로 처리합니다 https://24-ai.news/ko/news/2026-06-30/nvidia-inference-stack-lowest-token-cost/ 24 AI ko 2026-06-30T00:00:00Z NVIDIA: Blackwell의 소프트웨어 스택이 한 달 만에 DeepSeek V4 토큰 비용 5배 절감 https://24-ai.news/ko/news/2026-06-30/oecd-genai-chatbot-usage-web-traffic/ 24 AI ko 2026-06-30T00:00:00Z OECD: 생성형 AI 챗봇 사용률, 1년 만에 인구의 18%에서 28%로 성장 https://24-ai.news/ko/news/2026-06-30/pytorch-miles-rl-post-training/ 24 AI ko 2026-06-30T00:00:00Z Miles: 프론티어 규모 LLM을 위한 PyTorch 네이티브 오픈소스 RL 포스트 트레이닝 프레임워크 https://24-ai.news/hr/news/2026-06-29/allen-discoformer-density-score/ 24 AI hr 2026-06-29T00:00:00Z Allen Institute: DiScoFormer — jedan transformer za gustoću i score kroz različite distribucije https://24-ai.news/hr/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI hr 2026-06-29T00:00:00Z AMD: ROCm low-latency GEMM kerneli ubrzavaju LLM inferenciju do 1,79× na Instinct MI355X https://24-ai.news/hr/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI hr 2026-06-29T00:00:00Z arXiv:2606.28270: Agent-Native Immune System — šestoslojna runtime obrana ugrađena u zaključivanje AI agenata https://24-ai.news/hr/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI hr 2026-06-29T00:00:00Z arXiv:2606.28166: Tandem RL — verifikabilne nagrade uz čitljiviji lanac misli i bolji handoff na manji model https://24-ai.news/hr/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI hr 2026-06-29T00:00:00Z arXiv:2606.28061: ToolPrivacyBench — mjeri 'need-to-know' privatnost u LLM agentima s alatima https://24-ai.news/hr/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI hr 2026-06-29T00:00:00Z arXiv:2606.27483: Internalizing the Future — jedinstvena paradigma treniranja za svjetski model planiranja LLM agenata https://24-ai.news/hr/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI hr 2026-06-29T00:00:00Z AWS: multi-tenant AI agent s row-level security i Split-Plane SQL kriptografskim granicama podataka https://24-ai.news/hr/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI hr 2026-06-29T00:00:00Z AWS: Nova 2 Lite i Claude — cost-optimizirani pipeline za obradu skeniranih dokumenata https://24-ai.news/hr/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI hr 2026-06-29T00:00:00Z GitHub: Claude Opus 4.8 fast mode stiže u Copilot preview; Anthropic ukida fast za Opus 4.6 https://24-ai.news/hr/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI hr 2026-06-29T00:00:00Z LangChain: Dynamic Subagents u Deep Agents — agent piše kod koji paralelno dispatcha stotine podagenata https://24-ai.news/hr/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI hr 2026-06-29T00:00:00Z Meta: Brain2Qwerty v2 — neinvazivno dekodiranje misli u tekst s 61% točnosti, bez kirurškog implantata https://24-ai.news/hr/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI hr 2026-06-29T00:00:00Z Microsoft: 2026 Agent Confidence Index — 300 graditelja, prosječno povjerenje u AI agente 64/100 https://24-ai.news/hr/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI hr 2026-06-29T00:00:00Z Microsoft Research: Memora — memorija AI agenata uz do 98% manje tokena i SOTA na dugim razgovorima https://24-ai.news/hr/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI hr 2026-06-29T00:00:00Z NVIDIA: Palantir i NVIDIA Nemotron donose suvereni AI američkim agencijama u air-gapped sustavima https://24-ai.news/hr/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI hr 2026-06-29T00:00:00Z OpenAI: Mapping Europe's AI Workforce — izvještaj o prilikama AI-ja za EU radnu snagu https://24-ai.news/en/news/2026-06-29/allen-discoformer-density-score/ 24 AI en 2026-06-29T00:00:00Z Allen Institute: DiScoFormer — One Transformer for Density and Score Across Distributions https://24-ai.news/en/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI en 2026-06-29T00:00:00Z AMD: ROCm Low-Latency GEMM Kernels Speed Up LLM Inference by Up to 1.79× on Instinct MI355X https://24-ai.news/en/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI en 2026-06-29T00:00:00Z arXiv:2606.28270: Agent-Native Immune System — Six-Layer Runtime Defense Built Into AI Agent Reasoning https://24-ai.news/en/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI en 2026-06-29T00:00:00Z arXiv:2606.28166: Tandem RL — Verifiable Rewards With More Readable Chain of Thought and Better Handoff to Smaller Models https://24-ai.news/en/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI en 2026-06-29T00:00:00Z arXiv:2606.28061: ToolPrivacyBench — Measuring 'Need-to-Know' Privacy in LLM Agents With Tools https://24-ai.news/en/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI en 2026-06-29T00:00:00Z arXiv:2606.27483: Internalizing the Future — A Unified Training Paradigm for World Model Planning in LLM Agents https://24-ai.news/en/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI en 2026-06-29T00:00:00Z AWS: Multi-Tenant AI Agent With Row-Level Security and Split-Plane SQL Cryptographic Data Boundaries https://24-ai.news/en/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI en 2026-06-29T00:00:00Z AWS: Nova 2 Lite and Claude — Cost-Optimized Pipeline for Processing Scanned Documents https://24-ai.news/en/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI en 2026-06-29T00:00:00Z GitHub: Claude Opus 4.8 Fast Mode Arrives in Copilot Preview; Anthropic Retires Fast Mode for Opus 4.6 https://24-ai.news/en/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI en 2026-06-29T00:00:00Z LangChain: Dynamic Subagents in Deep Agents — Agent Writes Code That Dispatches Hundreds of Subagents in Parallel https://24-ai.news/en/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI en 2026-06-29T00:00:00Z Meta: Brain2Qwerty v2 — Non-Invasive Thought-to-Text Decoding at 61% Accuracy, Without Surgical Implants https://24-ai.news/en/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI en 2026-06-29T00:00:00Z Microsoft: 2026 Agent Confidence Index — 300 Builders, Average Confidence in AI Agents 64/100 https://24-ai.news/en/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI en 2026-06-29T00:00:00Z Microsoft Research: Memora — AI Agent Memory With Up to 98% Fewer Tokens and SOTA on Long Conversations https://24-ai.news/en/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI en 2026-06-29T00:00:00Z NVIDIA: Palantir and NVIDIA Nemotron Bring Sovereign AI to US Agencies in Air-Gapped Systems https://24-ai.news/en/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI en 2026-06-29T00:00:00Z OpenAI: Mapping Europe's AI Workforce — Report on AI Opportunities for the EU Workforce https://24-ai.news/de/news/2026-06-29/allen-discoformer-density-score/ 24 AI de 2026-06-29T00:00:00Z Allen Institute: DiScoFormer — ein Transformer für Dichte und Score über verschiedene Verteilungen https://24-ai.news/de/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI de 2026-06-29T00:00:00Z AMD: ROCm Low-Latency-GEMM-Kernel beschleunigen LLM-Inferenz bis zu 1,79× auf Instinct MI355X https://24-ai.news/de/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI de 2026-06-29T00:00:00Z arXiv:2606.28270: Agent-Native Immune System — sechsschichtige Runtime-Abwehr in der KI-Agenten-Kognitionsschleife https://24-ai.news/de/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI de 2026-06-29T00:00:00Z arXiv:2606.28166: Tandem RL — verifizierbare Belohnungen mit besser lesbarer Gedankenkette und verbessertem Handoff auf kleinere Modelle https://24-ai.news/de/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI de 2026-06-29T00:00:00Z arXiv:2606.28061: ToolPrivacyBench — misst „Need-to-Know”-Datenschutz in LLM-Agenten mit Werkzeugen https://24-ai.news/de/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI de 2026-06-29T00:00:00Z arXiv:2606.27483: Internalizing the Future — einheitliches Training-Paradigma für World-Model-Planning von LLM-Agenten https://24-ai.news/de/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI de 2026-06-29T00:00:00Z AWS: Multi-Tenant-KI-Agent mit Row-Level-Security und Split-Plane-SQL als kryptographische Datengrenzen https://24-ai.news/de/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI de 2026-06-29T00:00:00Z AWS: Nova 2 Lite und Claude — kostenoptimierte Pipeline zur Verarbeitung gescannter Dokumente https://24-ai.news/de/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI de 2026-06-29T00:00:00Z GitHub: Claude Opus 4.8 Fast Mode kommt in Copilot-Preview; Anthropic stellt Fast für Opus 4.6 ein https://24-ai.news/de/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI de 2026-06-29T00:00:00Z LangChain: Dynamic Subagents in Deep Agents — Agent schreibt Code zum parallelen Dispatch hunderter Unteragenten https://24-ai.news/de/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI de 2026-06-29T00:00:00Z Meta: Brain2Qwerty v2 — nichtinvasive Dekodierung von Gedanken in Text mit 61% Genauigkeit, ohne chirurgisches Implantat https://24-ai.news/de/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI de 2026-06-29T00:00:00Z Microsoft: 2026 Agent Confidence Index — 300 Entwickler, durchschnittliches Vertrauen in KI-Agenten 64/100 https://24-ai.news/de/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI de 2026-06-29T00:00:00Z Microsoft Research: Memora — KI-Agenten-Gedächtnis mit bis zu 98% weniger Token und SOTA bei langen Gesprächen https://24-ai.news/de/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI de 2026-06-29T00:00:00Z NVIDIA: Palantir und NVIDIA Nemotron bringen souveräne KI für US-Behörden in Air-Gapped-Systemen https://24-ai.news/de/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI de 2026-06-29T00:00:00Z OpenAI: Mapping Europe's AI Workforce — Bericht über KI-Chancen für EU-Arbeitskräfte https://24-ai.news/zh/news/2026-06-29/allen-discoformer-density-score/ 24 AI zh 2026-06-29T00:00:00Z Allen Institute：DiScoFormer——单个Transformer跨分布同步估计密度与得分函数 https://24-ai.news/zh/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI zh 2026-06-29T00:00:00Z AMD：ROCm低延迟GEMM内核在Instinct MI355X上将LLM推理提速最高1.79倍 https://24-ai.news/zh/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI zh 2026-06-29T00:00:00Z arXiv:2606.28270：Agent原生免疫系统——嵌入AI智能体推理中的六层运行时防御 https://24-ai.news/zh/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI zh 2026-06-29T00:00:00Z arXiv:2606.28166：Tandem RL——可验证奖励提升思维链可读性并改善向小模型的迁移 https://24-ai.news/zh/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI zh 2026-06-29T00:00:00Z arXiv:2606.28061：ToolPrivacyBench——衡量LLM工具调用智能体的「知其所需」隐私保护 https://24-ai.news/zh/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI zh 2026-06-29T00:00:00Z arXiv:2606.27483：内化未来——LLM智能体世界模型规划的统一训练范式 https://24-ai.news/zh/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI zh 2026-06-29T00:00:00Z AWS：多租户AI智能体结合行级安全与Split-Plane SQL实现加密数据隔离 https://24-ai.news/zh/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI zh 2026-06-29T00:00:00Z AWS：Nova 2 Lite与Claude——扫描文档批量处理的成本优化双阶段流水线 https://24-ai.news/zh/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI zh 2026-06-29T00:00:00Z GitHub：Claude Opus 4.8快速模式进入Copilot预览；Anthropic终止Opus 4.6的快速模式 https://24-ai.news/zh/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI zh 2026-06-29T00:00:00Z LangChain：Deep Agents中的Dynamic Subagents——智能体编写代码并行调度数百个子智能体 https://24-ai.news/zh/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI zh 2026-06-29T00:00:00Z Meta：Brain2Qwerty v2——无需外科植入物，以61%准确率将大脑信号转化为文字 https://24-ai.news/zh/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI zh 2026-06-29T00:00:00Z Microsoft：2026智能体信心指数——300位构建者对AI智能体的平均信心得分64/100 https://24-ai.news/zh/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI zh 2026-06-29T00:00:00Z Microsoft Research：Memora——减少最多98%Token消耗并在长对话上达到SOTA的AI智能体记忆框架 https://24-ai.news/zh/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI zh 2026-06-29T00:00:00Z NVIDIA：Palantir与NVIDIA Nemotron为气隙隔离系统中的美国政府机构带来主权AI https://24-ai.news/zh/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI zh 2026-06-29T00:00:00Z OpenAI：欧洲AI劳动力地图——关于AI对欧盟劳动力机遇影响的报告 https://24-ai.news/ja/news/2026-06-29/allen-discoformer-density-score/ 24 AI ja 2026-06-29T00:00:00Z Allen Institute: DiScoFormer — 一つのTransformerが様々な分布の密度とスコアを同時推定 https://24-ai.news/ja/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI ja 2026-06-29T00:00:00Z AMD: ROCm低レイテンシGEMMカーネルがInstinct MI355XのLLM推論を最大1.79倍高速化 https://24-ai.news/ja/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI ja 2026-06-29T00:00:00Z arXiv:2606.28270: Agent-Native Immune System — AIエージェントの推論に組み込まれた6層ランタイム防御 https://24-ai.news/ja/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI ja 2026-06-29T00:00:00Z arXiv:2606.28166: Tandem RL — 検証可能な報酬により思考の連鎖の可読性と小モデルへのハンドオフを改善 https://24-ai.news/ja/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI ja 2026-06-29T00:00:00Z arXiv:2606.28061: ToolPrivacyBench — ツールを使うLLMエージェントの「need-to-know」プライバシーを計測 https://24-ai.news/ja/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI ja 2026-06-29T00:00:00Z arXiv:2606.27483: Internalizing the Future — LLMエージェントの世界モデル計画のための統一的なトレーニングパラダイム https://24-ai.news/ja/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI ja 2026-06-29T00:00:00Z AWS: 行レベルセキュリティとSplit-Plane SQLの暗号化データ境界を持つマルチテナントAIエージェント https://24-ai.news/ja/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI ja 2026-06-29T00:00:00Z AWS: Nova 2 Liteとクロード — スキャン文書処理のコスト最適化パイプライン https://24-ai.news/ja/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI ja 2026-06-29T00:00:00Z GitHub: Claude Opus 4.8 fastモードがCopilotのプレビューに登場、AnthropicはOpus 4.6のfastモードを廃止 https://24-ai.news/ja/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI ja 2026-06-29T00:00:00Z LangChain: Deep AgentsのDynamic Subagents — エージェントが数百のサブエージェントを並列ディスパッチするコードを記述 https://24-ai.news/ja/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI ja 2026-06-29T00:00:00Z Meta: Brain2Qwerty v2 — 外科的インプラントなしで脳信号を61%の精度でテキストにデコード https://24-ai.news/ja/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI ja 2026-06-29T00:00:00Z Microsoft: 2026 Agent Confidence Index — 300人のビルダーがAIエージェントへの平均信頼度64/100と評価 https://24-ai.news/ja/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI ja 2026-06-29T00:00:00Z Microsoft Research: Memora — AIエージェントの記憶を最大98%少ないトークンとSOTAの長期会話で実現 https://24-ai.news/ja/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI ja 2026-06-29T00:00:00Z NVIDIA: PalantirとNVIDIA Nemotronがエアギャップシステムの米国政府機関に主権AIを提供 https://24-ai.news/ja/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI ja 2026-06-29T00:00:00Z OpenAI: EUの労働力へのAIの機会をまとめた「Mapping Europe's AI Workforce」レポート https://24-ai.news/ko/news/2026-06-29/allen-discoformer-density-score/ 24 AI ko 2026-06-29T00:00:00Z Allen Institute: DiScoFormer — 하나의 트랜스포머로 다양한 분포의 밀도와 점수 추정 https://24-ai.news/ko/news/2026-06-29/amd-rocm-low-latency-gemm-inference/ 24 AI ko 2026-06-29T00:00:00Z AMD: ROCm 저지연 GEMM 커널로 Instinct MI355X에서 LLM 추론 최대 1.79배 가속 https://24-ai.news/ko/news/2026-06-29/arxiv-agent-native-immune-system/ 24 AI ko 2026-06-29T00:00:00Z arXiv:2606.28270: Agent-Native Immune System — AI 에이전트 추론에 내장된 6계층 런타임 방어 https://24-ai.news/ko/news/2026-06-29/arxiv-tandem-rl-verifiable-rewards/ 24 AI ko 2026-06-29T00:00:00Z arXiv:2606.28166: Tandem RL — 더 읽기 쉬운 사고 연쇄와 소형 모델 핸드오프를 갖춘 검증 가능한 보상 https://24-ai.news/ko/news/2026-06-29/arxiv-toolprivacybench-agent-privacy/ 24 AI ko 2026-06-29T00:00:00Z arXiv:2606.28061: ToolPrivacyBench — 도구를 사용하는 LLM 에이전트의 『필요한 만큼만』 프라이버시 측정 https://24-ai.news/ko/news/2026-06-29/arxiv-world-model-planning-agents/ 24 AI ko 2026-06-29T00:00:00Z arXiv:2606.27483: Internalizing the Future — LLM 에이전트 세계 모델 계획을 위한 통합 에이전트 학습 패러다임 https://24-ai.news/ko/news/2026-06-29/aws-multitenant-agent-row-level-security/ 24 AI ko 2026-06-29T00:00:00Z AWS: 행 수준 보안과 Split-Plane SQL 암호화 데이터 경계를 갖춘 다중 테넌트 AI 에이전트 https://24-ai.news/ko/news/2026-06-29/aws-nova-2-lite-claude-document-pipeline/ 24 AI ko 2026-06-29T00:00:00Z AWS: Nova 2 Lite와 Claude — 스캔 문서 처리를 위한 비용 최적화 파이프라인 https://24-ai.news/ko/news/2026-06-29/github-copilot-claude-opus-4-8-fast/ 24 AI ko 2026-06-29T00:00:00Z GitHub: Claude Opus 4.8 fast mode가 Copilot 프리뷰에 출시; Anthropic은 Opus 4.6의 fast 모드 종료 https://24-ai.news/ko/news/2026-06-29/langchain-dynamic-subagents-deep-agents/ 24 AI ko 2026-06-29T00:00:00Z LangChain: Deep Agents의 Dynamic Subagents — 에이전트가 수백 개의 하위 에이전트를 병렬로 디스패치하는 코드 작성 https://24-ai.news/ko/news/2026-06-29/meta-brain2qwerty-noninvasive-text/ 24 AI ko 2026-06-29T00:00:00Z Meta: Brain2Qwerty v2 — 수술적 임플란트 없이 61% 정확도로 생각을 텍스트로 비침습적 해독 https://24-ai.news/ko/news/2026-06-29/microsoft-agent-confidence-index-2026/ 24 AI ko 2026-06-29T00:00:00Z Microsoft: 2026 Agent Confidence Index — 300명의 개발자, AI 에이전트에 대한 평균 신뢰도 64/100 https://24-ai.news/ko/news/2026-06-29/microsoft-memora-agent-memory-tokens/ 24 AI ko 2026-06-29T00:00:00Z Microsoft Research: Memora — 토큰 최대 98% 감소와 장문 대화 SOTA를 달성하는 AI 에이전트 메모리 https://24-ai.news/ko/news/2026-06-29/nvidia-palantir-nemotron-sovereign-ai/ 24 AI ko 2026-06-29T00:00:00Z NVIDIA: Palantir과 NVIDIA Nemotron이 에어갭 시스템의 미국 기관에 주권 AI 제공 https://24-ai.news/ko/news/2026-06-29/openai-europe-ai-workforce-mapping/ 24 AI ko 2026-06-29T00:00:00Z OpenAI: Mapping Europe's AI Workforce — EU 노동력을 위한 AI 기회에 관한 보고서