已验证来源
每个故事都链接到最有力的可用证据：可用的原始来源，否则明确归因的报告。
简单的英语
发生了什么、为什么重要以及值得关注的内容——无需行话。
无填料
当信号很弱时，我们不会发布任何内容，而是填充提要。
更多故事
9 故事创新
FLARE uses an LLM and Lean to verify optimization reformulations
Researchers introduce FLARE, a system that pairs an LLM-based agent with the Lean proof assistant to check whether proposed mixed-integer linear programming reformulations preserve the original problem. On the paper’s 20-problem, 109-formulation benchmark, the authors report 100% accuracy on the NP-hard subset and…arxiv.org创新
A preprint proposes profiling AI and workplace tasks by cognitive capabilities
Researchers propose a framework that compares AI systems with workplace tasks using shared cognitive-capability profiles, based on evaluations of six AI systems and task requirements gathered from 410 employees.arxiv.org安全
FuzzingBrain-Bench tests whether LLMs can find unexpected software crashes
A new arXiv benchmark evaluates whether large language models can discover distinct crashes in open-source software without being given a predefined vulnerability target. In the authors’ tests, Claude Opus 4.8 triggered crashes in 60 of 77 challenges, but achieved only 196 of a possible 579 points.arxiv.org工业
随着 ChatGPT 用途的扩大，OpenAI 在巴西开始商业运营
OpenAI 表示，它已在巴西启动商业运营，随着公司扩大 ChatGPT、Codex 和企业采用范围，圣保罗团队为企业、开发人员、研究人员和公共机构提供支持。
openai.com创新
PhysElite paper reports that leading multimodal LLM solved 33.7% of Olympiad-level physics problems
A new benchmark of 11,586 bilingual, diagram-based physics problems reports that the strongest of 18 tested multimodal language models achieved only 33.7% answer accuracy.arxiv.org创新
LifePlanner benchmark reports that LLM agents falter at complex planning with social-media data
A new arXiv benchmark evaluates LLM agents on geospatial planning using maps, tools and local social-media posts. Its authors report a sharp decline on complex tasks, with a 40.2% pass rate.arxiv.org工业
Crunchbase News reports Socure raises $156 million and acquires Fravity
Socure raised $156 million at a reported $5.2 billion valuation and agreed to acquire AI fraud-investigation startup Fravity, according to Crunchbase News.news.crunchbase.com工业
MiniMax reports 703% surge in Open Platform and enterprise AI revenue
MiniMax says revenue from its Open Platform and other AI-based enterprise services rose 703.1% year over year to $73.9 million in the first half of 2026. The unaudited results also show higher gross profit but a wider adjusted net loss.minimax.io创新
Ai2 and Providence Swedish bring AI-assisted discovery into cancer research
Ai2 says Providence Swedish will run its AutoDiscovery platform on protected cancer-research data after a breast-cancer signal was validated in a separate dataset and lab work.allenai.org