https://24-ai.news/hr/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI hr 2026-07-13T00:00:00Z Allen Institute: što je izgradnja agenta Shippy naučila o pouzdanim AI agentima https://24-ai.news/hr/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI hr 2026-07-13T00:00:00Z AMD: GEAK agent automatski optimizirao DeepSeek-V4 MLA kernel na MI355 uz do 9× ubrzanje https://24-ai.news/hr/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI hr 2026-07-13T00:00:00Z Anthropic: Claudeove vrijednosti variraju ovisno o modelu i jeziku https://24-ai.news/hr/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI hr 2026-07-13T00:00:00Z arXiv:2607.09375: Li Auto predstavlja Mach-Mind-4-Flash — 35B MoE s 3B aktivnih parametara https://24-ai.news/hr/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI hr 2026-07-13T00:00:00Z arXiv:2607.09415: samovođeni test-time trening podiže dugokontekstne LLM-ove do 15% https://24-ai.news/hr/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI hr 2026-07-13T00:00:00Z arXiv:2607.09493: dijeljena selektivna memorija za agente — 96% uspješnost uz 97× manji trošak tokena https://24-ai.news/hr/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI hr 2026-07-13T00:00:00Z arXiv:2607.09424: Soofi S — suvereni open-source model za njemački i engleski treniran na Telekomovom cloudu https://24-ai.news/hr/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI hr 2026-07-13T00:00:00Z arXiv:2607.09532: statistički nedetektabilni backdoori u dubokim neuronskim mrežama (ICML 2026) https://24-ai.news/hr/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI hr 2026-07-13T00:00:00Z AWS: OpenAI GPT-5.6 (Sol, Terra, Luna) sada općenito dostupan na Amazon Bedrocku https://24-ai.news/hr/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI hr 2026-07-13T00:00:00Z AWS: Bluesight gradi agentni AI na Bedrock AgentCore — izvještaji 96% brži https://24-ai.news/hr/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI hr 2026-07-13T00:00:00Z Microsoft: formalna verifikacija Rust kriptografije u SymCryptu uz pomoć AI agenata i Leana https://24-ai.news/hr/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI hr 2026-07-13T00:00:00Z Sakana AI: 'pametne stanične kocke' donose kolektivnu inteligenciju u fizički svijet https://24-ai.news/en/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI en 2026-07-13T00:00:00Z Allen Institute: What Building the Shippy Agent Taught Us About Reliable AI Agents https://24-ai.news/en/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI en 2026-07-13T00:00:00Z AMD: GEAK Agent Automatically Optimized DeepSeek-V4 MLA Kernel on MI355 with Up to 9× Speedup https://24-ai.news/en/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI en 2026-07-13T00:00:00Z Anthropic: Claude's Values Vary by Model and Language https://24-ai.news/en/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI en 2026-07-13T00:00:00Z arXiv:2607.09375: Li Auto Presents Mach-Mind-4-Flash — 35B MoE with 3B Active Parameters https://24-ai.news/en/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI en 2026-07-13T00:00:00Z arXiv:2607.09415: Self-Guided Test-Time Training Improves Long-Context LLMs by Up to 15% https://24-ai.news/en/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI en 2026-07-13T00:00:00Z arXiv:2607.09493: Shared Selective Persistent Memory for Agents — 96% Success at 97× Lower Token Cost https://24-ai.news/en/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI en 2026-07-13T00:00:00Z arXiv:2607.09424: Soofi S — Sovereign Open-Source Model for German and English Trained on Telekom's Cloud https://24-ai.news/en/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI en 2026-07-13T00:00:00Z arXiv:2607.09532: Statistically Undetectable Backdoors in Deep Neural Networks (ICML 2026) https://24-ai.news/en/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI en 2026-07-13T00:00:00Z AWS: OpenAI GPT-5.6 (Sol, Terra, Luna) Now Generally Available on Amazon Bedrock https://24-ai.news/en/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI en 2026-07-13T00:00:00Z AWS: Bluesight Builds Agentic AI on Bedrock AgentCore — Reports 96% Faster https://24-ai.news/en/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI en 2026-07-13T00:00:00Z Microsoft: Formal Verification of Rust Cryptography in SymCrypt with AI Agents and Lean https://24-ai.news/en/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI en 2026-07-13T00:00:00Z Sakana AI: 'Smart Cellular Bricks' Bring Collective Intelligence to the Physical World https://24-ai.news/de/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI de 2026-07-13T00:00:00Z Allen Institute: Was der Bau des Shippy-Agenten über zuverlässige KI-Agenten gelehrt hat https://24-ai.news/de/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI de 2026-07-13T00:00:00Z AMD: GEAK-Agent optimierte DeepSeek-V4-MLA-Kernel auf MI355 automatisch mit bis zu 9× Beschleunigung https://24-ai.news/de/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI de 2026-07-13T00:00:00Z Anthropic: Claudes Werte variieren je nach Modell und Sprache https://24-ai.news/de/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI de 2026-07-13T00:00:00Z arXiv:2607.09375: Li Auto stellt Mach-Mind-4-Flash vor — 35B MoE mit 3B aktiven Parametern https://24-ai.news/de/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI de 2026-07-13T00:00:00Z arXiv:2607.09415: Selbstgeführtes Test-Time-Training verbessert Langkontext-LLMs um bis zu 15 % https://24-ai.news/de/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI de 2026-07-13T00:00:00Z arXiv:2607.09493: Geteiltes selektives persistentes Gedächtnis für Agenten — 96 % Erfolg bei 97× geringeren Token-Kosten https://24-ai.news/de/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI de 2026-07-13T00:00:00Z arXiv:2607.09424: Soofi S — Souveränes Open-Source-Modell für Deutsch und Englisch, trainiert auf Telekom-Cloud https://24-ai.news/de/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI de 2026-07-13T00:00:00Z arXiv:2607.09532: Statistisch undetektierbare Backdoors in tiefen neuronalen Netzen (ICML 2026) https://24-ai.news/de/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI de 2026-07-13T00:00:00Z AWS: OpenAI GPT-5.6 (Sol, Terra, Luna) jetzt allgemein verfügbar auf Amazon Bedrock https://24-ai.news/de/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI de 2026-07-13T00:00:00Z AWS: Bluesight baut agentische KI auf Bedrock AgentCore — Berichte 96 % schneller https://24-ai.news/de/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI de 2026-07-13T00:00:00Z Microsoft: Formale Verifikation von Rust-Kryptografie in SymCrypt mit KI-Agenten und Lean https://24-ai.news/de/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI de 2026-07-13T00:00:00Z Sakana AI: Intelligente Zellbausteine bringen kollektive Intelligenz in die physische Welt https://24-ai.news/zh/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI zh 2026-07-13T00:00:00Z Allen Institute：Shippy智能体的开发为构建可靠AI智能体带来了哪些启示 https://24-ai.news/zh/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI zh 2026-07-13T00:00:00Z AMD：GEAK智能体自动优化DeepSeek-V4 MLA内核，MI355上实现最高9倍加速 https://24-ai.news/zh/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI zh 2026-07-13T00:00:00Z Anthropic：Claude的价值观因模型和语言而异 https://24-ai.news/zh/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI zh 2026-07-13T00:00:00Z arXiv:2607.09375：理想汽车发布Mach-Mind-4-Flash——35B MoE模型，3B激活参数 https://24-ai.news/zh/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI zh 2026-07-13T00:00:00Z arXiv:2607.09415：自引导测试时训练将长上下文LLM提升最高15% https://24-ai.news/zh/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI zh 2026-07-13T00:00:00Z arXiv:2607.09493：智能体的共享选择性持久记忆——96%成功率，token开销降低97倍 https://24-ai.news/zh/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI zh 2026-07-13T00:00:00Z arXiv:2607.09424：Soofi S——在德国电信云上训练的德英主权开源模型 https://24-ai.news/zh/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI zh 2026-07-13T00:00:00Z arXiv:2607.09532：深度神经网络中统计上不可检测的后门（ICML 2026） https://24-ai.news/zh/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI zh 2026-07-13T00:00:00Z AWS：OpenAI GPT-5.6（Sol、Terra、Luna）现已在Amazon Bedrock上正式可用 https://24-ai.news/zh/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI zh 2026-07-13T00:00:00Z AWS：Bluesight在Bedrock AgentCore上构建智能体AI——报告速度提升96% https://24-ai.news/zh/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI zh 2026-07-13T00:00:00Z Microsoft：借助AI智能体和Lean，对SymCrypt中的Rust密码学进行形式化验证 https://24-ai.news/zh/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI zh 2026-07-13T00:00:00Z Sakana AI：「智能细胞砖块」将集体智能带入物理世界 https://24-ai.news/ja/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI ja 2026-07-13T00:00:00Z Allen Institute：Shippyエージェントの開発から学んだ信頼できるAIエージェント構築の教訓 https://24-ai.news/ja/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI ja 2026-07-13T00:00:00Z AMD：GEAKエージェントがMI355上でDeepSeek-V4 MLAカーネルを自動最適化、最大9倍の高速化を実現 https://24-ai.news/ja/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI ja 2026-07-13T00:00:00Z Anthropic：Claudeの価値観はモデルと言語によって異なる https://24-ai.news/ja/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI ja 2026-07-13T00:00:00Z arXiv:2607.09375：理想汽車がMach-Mind-4-Flashを発表——35B MoEモデル、アクティブパラメータ3B https://24-ai.news/ja/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI ja 2026-07-13T00:00:00Z arXiv:2607.09415：自己誘導テスト時トレーニングで長コンテキストLLMを最大15%向上 https://24-ai.news/ja/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI ja 2026-07-13T00:00:00Z arXiv:2607.09493：エージェントの共有選択的永続記憶——成功率96%、トークンコスト97分の1 https://24-ai.news/ja/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI ja 2026-07-13T00:00:00Z arXiv:2607.09424：Soofi S——Deutsche Telekomのクラウドでトレーニングされた独英主権オープンソースモデル https://24-ai.news/ja/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI ja 2026-07-13T00:00:00Z arXiv:2607.09532：深層ニューラルネットワークにおける統計的に検出不能なバックドア（ICML 2026） https://24-ai.news/ja/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI ja 2026-07-13T00:00:00Z AWS：OpenAI GPT-5.6（Sol、Terra、Luna）がAmazon Bedrockで一般提供開始 https://24-ai.news/ja/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI ja 2026-07-13T00:00:00Z AWS：BluesightがBedrock AgentCoreでエージェンティックAIを構築——レポート速度96%改善 https://24-ai.news/ja/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI ja 2026-07-13T00:00:00Z Microsoft：AIエージェントとLeanを使ってSymCryptのRust暗号を形式検証 https://24-ai.news/ja/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI ja 2026-07-13T00:00:00Z Sakana AI：「スマートセルラーブリック」が集合知能を物理世界へ https://24-ai.news/ko/news/2026-07-13/allenai-shippy-agent-lessons/ 24 AI ko 2026-07-13T00:00:00Z Allen Institute：Shippy 에이전트 개발로 배운 신뢰할 수 있는 AI 에이전트 구축의 교훈 https://24-ai.news/ko/news/2026-07-13/amd-geak-deepseek-v4-mla-kernel/ 24 AI ko 2026-07-13T00:00:00Z AMD：GEAK 에이전트가 MI355에서 DeepSeek-V4 MLA 커널을 자동 최적화, 최대 9배 가속 https://24-ai.news/ko/news/2026-07-13/anthropic-claude-values-model-language/ 24 AI ko 2026-07-13T00:00:00Z Anthropic：Claude의 가치관은 모델과 언어에 따라 다르게 나타납니다 https://24-ai.news/ko/news/2026-07-13/arxiv-mach-mind-4-flash-li-auto/ 24 AI ko 2026-07-13T00:00:00Z arXiv:2607.09375：Li Auto, Mach-Mind-4-Flash 발표——35B MoE, 활성 파라미터 3B https://24-ai.news/ko/news/2026-07-13/arxiv-self-guided-test-time-training/ 24 AI ko 2026-07-13T00:00:00Z arXiv:2607.09415：자기 유도 테스트 시간 훈련으로 장문맥 LLM 최대 15% 향상 https://24-ai.news/ko/news/2026-07-13/arxiv-shared-selective-persistent-memory/ 24 AI ko 2026-07-13T00:00:00Z arXiv:2607.09493：에이전트를 위한 공유 선택적 영속 메모리——성공률 96%, 토큰 비용 97배 절감 https://24-ai.news/ko/news/2026-07-13/arxiv-soofi-german-english-sovereign-model/ 24 AI ko 2026-07-13T00:00:00Z arXiv:2607.09424：Soofi S——Deutsche Telekom 클라우드에서 훈련된 독일어·영어 주권 오픈소스 모델 https://24-ai.news/ko/news/2026-07-13/arxiv-undetectable-backdoors-dnn/ 24 AI ko 2026-07-13T00:00:00Z arXiv:2607.09532：심층 신경망의 통계적으로 탐지 불가능한 백도어（ICML 2026） https://24-ai.news/ko/news/2026-07-13/aws-bedrock-gpt-5-6-sol-terra-luna/ 24 AI ko 2026-07-13T00:00:00Z AWS：OpenAI GPT-5.6（Sol, Terra, Luna）이 Amazon Bedrock에서 정식 출시 https://24-ai.news/ko/news/2026-07-13/bluesight-bedrock-agentic-healthcare/ 24 AI ko 2026-07-13T00:00:00Z AWS：Bluesight가 Bedrock AgentCore에서 에이전틱 AI 구축——보고서 속도 96% 향상 https://24-ai.news/ko/news/2026-07-13/microsoft-symcrypt-rust-crypto-verification/ 24 AI ko 2026-07-13T00:00:00Z Microsoft：AI 에이전트와 Lean으로 SymCrypt의 Rust 암호화를 형식 검증 https://24-ai.news/ko/news/2026-07-13/sakana-smart-cellular-bricks/ 24 AI ko 2026-07-13T00:00:00Z Sakana AI：「스마트 세포 블록」이 집단 지능을 물리적 세계로 가져옵니다 https://24-ai.news/hr/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI hr 2026-07-11T00:00:00Z arXiv:2607.08093: CausalDS — novi benchmark koji mjeri kauzalno zaključivanje AI agenata u data-science zadacima https://24-ai.news/hr/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI hr 2026-07-11T00:00:00Z arXiv:2607.08625: opsežan pregled izazova kliničkih dijaloških AI sustava — od kliničkog zaključivanja do sigurnosti pacijenata https://24-ai.news/hr/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI hr 2026-07-11T00:00:00Z arXiv:2607.08681: SolarChain-Eval — fizikalno ograničen benchmark za AI agente u decentraliziranim energetskim tržištima https://24-ai.news/hr/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI hr 2026-07-11T00:00:00Z arXiv:2607.07779: Terence Tao među autorima rada o granicama velikih jezičnih modela u frontierskoj matematici https://24-ai.news/hr/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI hr 2026-07-11T00:00:00Z arXiv:2607.08395: Token-Flow Firewall — runtime nadzor koji dugoživuće AI agente štiti i spušta uspješnost napada na 12,5% https://24-ai.news/hr/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI hr 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207 uključuje auto mode na svim enterprise cloudovima, mijenja zadani model na Opus 4.8 i zatvara shell-injection propust https://24-ai.news/en/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI en 2026-07-11T00:00:00Z arXiv:2607.08093: CausalDS — a new benchmark measuring causal reasoning of AI agents in data-science tasks https://24-ai.news/en/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI en 2026-07-11T00:00:00Z arXiv:2607.08625: comprehensive survey of challenges in clinical conversational AI systems — from clinical reasoning to patient safety https://24-ai.news/en/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI en 2026-07-11T00:00:00Z arXiv:2607.08681: SolarChain-Eval — physics-constrained benchmark for AI agents in decentralized energy markets https://24-ai.news/en/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI en 2026-07-11T00:00:00Z arXiv:2607.07779: Terence Tao among authors of paper on limits of large language models in frontier mathematics https://24-ai.news/en/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI en 2026-07-11T00:00:00Z arXiv:2607.08395: Token-Flow Firewall — runtime monitoring that protects long-running AI agents and reduces attack success rate to 12.5% https://24-ai.news/en/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI en 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207 enables auto mode on all enterprise clouds, switches default model to Opus 4.8, and closes shell-injection vulnerability https://24-ai.news/de/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI de 2026-07-11T00:00:00Z arXiv:2607.08093: CausalDS — neues Benchmark misst kausales Schlussfolgern von KI-Agenten in Data-Science-Aufgaben https://24-ai.news/de/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI de 2026-07-11T00:00:00Z arXiv:2607.08625: umfassender Überblick über Herausforderungen klinischer konversationaler KI-Systeme — von klinischem Schlussfolgern bis zur Patientensicherheit https://24-ai.news/de/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI de 2026-07-11T00:00:00Z arXiv:2607.08681: SolarChain-Eval — physikbeschränktes Benchmark für KI-Agenten in dezentralen Energiemärkten https://24-ai.news/de/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI de 2026-07-11T00:00:00Z arXiv:2607.07779: Terence Tao als Mitautor einer Studie über die Grenzen großer Sprachmodelle in der Frontier-Mathematik https://24-ai.news/de/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI de 2026-07-11T00:00:00Z arXiv:2607.08395: Token-Flow Firewall — Runtime-Überwachung schützt langlebige KI-Agenten und senkt Angriffserfolgsrate auf 12,5 % https://24-ai.news/de/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI de 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207 aktiviert Auto-Modus auf allen Enterprise-Clouds, wechselt Standardmodell zu Opus 4.8 und schließt Shell-Injection-Lücke https://24-ai.news/zh/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI zh 2026-07-11T00:00:00Z arXiv:2607.08093：CausalDS — 衡量 AI 智能体在数据科学任务中因果推理能力的新基准 https://24-ai.news/zh/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI zh 2026-07-11T00:00:00Z arXiv:2607.08625：临床对话 AI 系统挑战全面综述——从临床推理到患者安全 https://24-ai.news/zh/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI zh 2026-07-11T00:00:00Z arXiv:2607.08681：SolarChain-Eval — 用于去中心化能源市场 AI 智能体的物理约束基准 https://24-ai.news/zh/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI zh 2026-07-11T00:00:00Z arXiv:2607.07779：Terence Tao 参与撰写探讨大型语言模型在前沿数学中局限性的论文 https://24-ai.news/zh/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI zh 2026-07-11T00:00:00Z arXiv:2607.08395：Token-Flow Firewall — 实时监控持久 AI 智能体并将攻击成功率降至 12.5% https://24-ai.news/zh/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI zh 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207 在所有企业云上启用 auto mode，将默认模型切换为 Opus 4.8 并修复 shell-injection 漏洞 https://24-ai.news/ja/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI ja 2026-07-11T00:00:00Z arXiv:2607.08093：CausalDS — データサイエンスタスクにおける AI エージェントの因果推論を測定する新ベンチマーク https://24-ai.news/ja/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI ja 2026-07-11T00:00:00Z arXiv:2607.08625：臨床対話 AI システムの課題に関する包括的レビュー——臨床推論から患者安全まで https://24-ai.news/ja/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI ja 2026-07-11T00:00:00Z arXiv:2607.08681：SolarChain-Eval — 分散型エネルギー市場の AI エージェント向け物理制約付きベンチマーク https://24-ai.news/ja/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI ja 2026-07-11T00:00:00Z arXiv:2607.07779：Terence Tao が共著——フロンティア数学における大規模言語モデルの限界を論じた論文 https://24-ai.news/ja/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI ja 2026-07-11T00:00:00Z arXiv:2607.08395：Token-Flow Firewall — 長期稼働 AI エージェントをリアルタイム監視し攻撃成功率を 12.5% に低下 https://24-ai.news/ja/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI ja 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207 が全エンタープライズクラウドで auto mode を有効化、デフォルトモデルを Opus 4.8 に変更し shell-injection 脆弱性を修正 https://24-ai.news/ko/news/2026-07-11/arxiv-causalds-data-science-agents/ 24 AI ko 2026-07-11T00:00:00Z arXiv:2607.08093：CausalDS — 데이터 과학 작업에서 AI 에이전트의 인과 추론을 측정하는 새 벤치마크 https://24-ai.news/ko/news/2026-07-11/arxiv-patient-centered-clinical-dialogue/ 24 AI ko 2026-07-11T00:00:00Z arXiv:2607.08625：임상 대화 AI 시스템의 과제에 관한 포괄적 개요——임상 추론에서 환자 안전까지 https://24-ai.news/ko/news/2026-07-11/arxiv-solarchain-eval-energy-agents/ 24 AI ko 2026-07-11T00:00:00Z arXiv:2607.08681：SolarChain-Eval — 분산형 에너지 시장의 AI 에이전트를 위한 물리 제약 벤치마크 https://24-ai.news/ko/news/2026-07-11/arxiv-tao-llm-formal-mathematics/ 24 AI ko 2026-07-11T00:00:00Z arXiv:2607.07779：Terence Tao가 공동 저술한 최전선 수학에서 대형 언어 모델의 한계를 논한 논문 https://24-ai.news/ko/news/2026-07-11/arxiv-token-flow-firewall-agent-security/ 24 AI ko 2026-07-11T00:00:00Z arXiv:2607.08395：Token-Flow Firewall — 장기 실행 AI 에이전트를 실시간 감시하고 공격 성공률을 12.5%로 낮추다 https://24-ai.news/ko/news/2026-07-11/claude-code-2-1-207-auto-mode-opus48/ 24 AI ko 2026-07-11T00:00:00Z Anthropic: Claude Code v2.1.207이 모든 엔터프라이즈 클라우드에서 auto mode를 활성화하고 기본 모델을 Opus 4.8로 변경하며 shell-injection 취약점 수정