{"date":"2026-09-06","title":"What happened in AI — Sep 6, 2026","generated_at":"2026-09-06T21:15:00Z","intro":["AI agents are moving from chat into unsupervised operational work: Google open-sourced an agent that hunts and fixes vulnerabilities, Figma built agents that triage its own security alerts, and a new tool lets coding agents drive a live Chrome session hands-off.","OpenAI's own blog cuts both ways today — coding agents are measurably speeding up its internal research, even as chief scientist Jakub Pachocki warns model reasoning is drifting from human thought. China pushes ahead on compute regardless: DeepSeek's chip order gets specifics, and Morgan Stanley flags a revenue surge at rivals MiniMax and Zhipu."],"highlights":["Google open-sourced Mantis, an agent that finds, validates, reproduces, and fixes vulnerabilities to cut scanner false positives.","Figma's security team runs AI agents that investigate alerts, check systems, and draft fixes for its own engineers.","Nimvarya lets coding agents control a full Chrome session without stealing your keyboard or window focus.","OpenAI says coding agents are raising research velocity internally — but its chief scientist warns model reasoning is diverging from human thought.","DeepSeek's 160,000-chip Huawei order now has specifics: Ascend 950DT accelerators for a new inference cluster in Inner Mongolia, as Morgan Stanley flags ARR surges at rivals MiniMax and Zhipu."],"article_count":7,"categories":[{"name":"Agents Take On Operational Work","slug":"agents-take-on-operational-work","summary":"Three items today show agents doing real operational tasks instead of just chatting: hunting and fixing vulnerabilities, triaging security alerts, and driving a live browser session hands-off.","articles":[{"title":"Google Mantis: An Agentic Vulnerability Scanning Harness for Reducing False Positives","summary":"Google open-sourced Mantis, an agent framework that automates the full vulnerability lifecycle — find, validate, reproduce, and fix — aimed at cutting the false positives that plague automated scanners.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/09/google-mantis-vulnerability-scan/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Sun, 06 Sep 2026 12:00:00 GMT"},{"title":"How Figma Uses AI Agents for Security","summary":"Figma's security team built AI agents that investigate alerts, search past incidents, check company systems, and draft code fixes — offloading first-pass triage work engineers used to do by hand.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/09/figma-security-agents/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Sun, 06 Sep 2026 06:59:00 GMT"},{"title":"Nimvarya – Let AI coding agents control Chrome without stealing focus","summary":"Nimvarya lets coding agents drive a full Chrome session — clicking, typing, reading the page — without stealing keyboard or window focus from the developer at the same machine.","source":"hackernews_ai","url":"https://github.com/Basiliskin/nimvarya","published":"Sun, 06 Sep 2026 08:00:17 +0000"}]},{"name":"Inside OpenAI: Research Speed and the Alignment Question","slug":"inside-openai-research-speed-and-alignment","summary":"OpenAI's own blog posted a productivity story and a warning on the same day: coding agents are measurably speeding up internal research, even as its chief scientist flags that model reasoning is drifting further from human thought.","articles":[{"title":"Research acceleration: The view inside OpenAI","summary":"OpenAI's internal data shows coding agents raising experiment velocity and task complexity inside its own research org — an early read on how much agents can speed up frontier AI research itself.","source":"openai_blog","url":"https://openai.com/index/research-acceleration-view-inside-openai","published":"Sun, 06 Sep 2026 08:00:00 GMT"},{"title":"An Alien Mind","summary":"OpenAI chief scientist Jakub Pachocki argues that as models grow more capable, their reasoning diverges further from human thought, and calls for stronger safeguards and international coordination to keep pace.","source":"openai_blog","url":"https://openai.com/index/an-alien-mind","published":"Sun, 06 Sep 2026 09:00:00 GMT"}]},{"name":"China's Compute Buildout and Model Economics","slug":"chinas-compute-buildout-and-model-economics","summary":"China's infrastructure and business signals point the same direction: DeepSeek's chip order gets concrete details while Morgan Stanley sees rival labs' revenue accelerating into a shakeout phase.","articles":[{"title":"DeepSeek Plans 160,000 Huawei Ascend 950DT Chips for Inner Mongolia Inference Cluster","summary":"DeepSeek's previously reported 160,000-chip Huawei order now has specifics: Ascend 950DT accelerators headed to a new inference cluster in Inner Mongolia, not model training.","source":"search_cn_open_weight_labs","publisher_name":"pandaily.com","publisher_domain":"pandaily.com","url":"https://news.google.com/rss/articles/CBMilAFBVV95cUxObGc4Q3I4UERxX3RFdmtwSjl6dXV2YTRWUmpKS0dyNF81dFh0Q2dOQXNiN19LUXJ4V1RIaVI1YXAxTHl3WWlmNV9fTHRlU3h1cXZRb0JJVnVfeFl1eklmUGlldHpxelRick5PY1lVSEVTbEc1ZmpLNWE1a2JsU29Oa3ZDeEdQdEhoRjVHQUx0d1gwc2xR?oc=5","published":"Sun, 06 Sep 2026 14:08:28 GMT"},{"title":"Morgan Stanley Private Briefing: MiniMax and Zhipu AI See Surge in ARR; Model Competition to Enter a Tiered Elimination Phase","summary":"Morgan Stanley's private briefing says Chinese labs MiniMax and Zhipu AI are seeing ARR surge, with the bank calling the next phase of model competition a 'tiered elimination' among labs that can't sustain the burn.","source":"search_cn_open_weight_labs","publisher_name":"富途牛牛","publisher_domain":"news.futunn.com","url":"https://news.google.com/rss/articles/CBMipgFBVV95cUxNRUI3dnNiclkwdndDNzhybDhUNW10b01qaE1XbzhEOWE4QUNXVTMwVE5HVUo2M1dqMng2UHNWa1NUcVJQV3Y3clF0Q2w5XzY0Mzh2VDdJMjFBdFFVWGppdTRiUVFaSnBSQ3p3MVlKX0dmREMwV1Z1Z044NDJZdmtVaGlIeVJvdWowOUlyMHhENUxrRTZDZ2FndTF6bXJRbTlOT2dyMnN3?oc=5","published":"Sun, 06 Sep 2026 04:37:44 GMT"}]}]}