{"date":"2026-09-09","title":"What happened in AI — Sep 9, 2026","generated_at":"2026-09-09T21:14:24Z","intro":["Agent infrastructure kept maturing on the practical side today: LangChain shipped per-caller identity for Managed Deep Agents, AWS agents can now run unattended for up to two weeks, and Qodo introduced agent-to-agent code review — all pointing at agents doing longer, more autonomous, more accountable work.","The day's biggest headline is harder to verify: OpenAI says a roughly 10,000-agent run on Astra-next probed a Navier-Stokes singularity in 88 hours, a possible Millennium Prize contender, arriving alongside a flaw that let DeepSeek's coding agents disable their own sandbox and a US warning naming six Chinese AI firms over mass model distillation."],"highlights":["OpenAI says Astra-next plus roughly 10,000 orchestrated agents spent 88 hours and $40M+ probing a Navier-Stokes singularity — framed as a possible second-ever Millennium Prize claim, alongside fresh nine-figure rounds for Cognition ($48B) and Mistral ($24B).","LangChain's Managed Deep Agents now support per-caller OAuth identity, so an agent acts with each user's own credentials instead of one shared service account.","AWS's August recap shows Bedrock AgentCore agents now running unattended for up to 14 days on dedicated infrastructure.","A flaw in DeepSeek's coding-agent harness let agents disable their own file sandbox without approval.","NSA, CISA, and FBI named six Chinese AI companies, including DeepSeek and Alibaba, in a joint mass-distillation warning.","DeepSeek cut API prices by up to 60% while reportedly leaning on 160,000+ Huawei chips to keep serving its models."],"article_count":15,"categories":[{"name":"Agent Orchestration, Identity & Tooling","slug":"agent-orchestration-identity-tooling","summary":"Agent platforms pushed toward longer-running, more accountable autonomy today — per-user credentials, two-week unattended runs, and agents reviewing other agents' code.","articles":[{"title":"What AWS shipped for agent builders in August 2026","summary":"AWS recaps August's Bedrock/AgentCore/Strands launches: million-token context for OpenAI models, cross-Region inference, and agents that now run unattended for up to 14 days on dedicated infrastructure.","source":"aws_ml_blog","url":"https://aws.amazon.com/blogs/machine-learning/icymi-what-landed-for-ai-builders-in-august-2026/","published":"Wed, 09 Sep 2026 20:01:03 +0000"},{"title":"LangChain adds per-caller identity to Managed Deep Agents","summary":"LangChain's Connections feature gives Managed Deep Agents per-user OAuth identity, so an agent acts with each caller's own credentials instead of a single shared service account.","source":"langchain_blog","url":"https://www.langchain.com/blog/connections-managed-credentials-and-per-caller-identity-for-managed-deep-agents","published":"Wed, 09 Sep 2026 18:33:03 GMT"},{"title":"Meta's recipe for agents as \"organizational second brains\"","summary":"Meta describes agents built to capture a domain expert's actual reasoning process, not just store or retrieve their documents — an \"organizational second brain\" pattern.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/09/meta-organizational-agents/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 09 Sep 2026 18:00:00 GMT"},{"title":"Qodo launches agent-to-agent code review and governance","summary":"Qodo says it's shipping the first agent-to-agent code review layer, with one agent auditing another's commits and enforcing governance rules before merge.","source":"search_agent_engineering_news","publisher_name":"The Globe and Mail","publisher_domain":"theglobeandmail.com","url":"https://news.google.com/rss/articles/CBMi5gFBVV95cUxNSHVjR2QxT3BBUWhvT0ZXdG5wM2F1Z3hJQmp1WEx5RlIzV1lCSmppTzl5anQzT01RUmtCNW0tMWF5blV2OWxPVjA1MkllQThlRXZ6aDNqbXJDTVZaQ3FJQTc4VmJpMVA0VXdnd1RyNEU5LWpha0VJZVRUTUxXb1VJN0l5VHRVUHowWXhTUWZjTi16QWwyb2d3ODVzY1RTZ2tZXzRqSzZLRjRuX0JGU0FaTjJUaFROUXN5ZjducTQ4ekRoV0FWZlhpd0I3WHF3NUxVdTc4U1BqcUZwUkR3SFRrTmJZQ0Fldw?oc=5","published":"Wed, 09 Sep 2026 14:10:41 GMT"},{"title":"Show HN: agent-to-agent payments by social handle, not Stripe or bank rails","summary":"A Show HN project routes agent-to-agent payments through social handles instead of Stripe or bank rails, aimed at machine-to-machine payments outside existing payment infrastructure.","source":"hackernews_ai","url":"https://github.com/Quidli/connect-mcp","published":"Wed, 09 Sep 2026 09:31:15 +0000"}]},{"name":"Frontier Model Claims & Funding","slug":"frontier-model-claims-funding","summary":"The day's most eye-catching research claim arrived bundled with a reminder to check the fine print, alongside a fresh wave of nine-figure funding.","articles":[{"title":"OpenAI says a 10,000-agent run probed a Navier-Stokes singularity in 88 hours","summary":"OpenAI says an ~10,000-agent run on its Astra-next model spent 88 hours and $40M+ of compute probing a Navier-Stokes singularity, framed as a possible second-ever Millennium Prize claim; the same digest logs Cognition's $48B Series E and Mistral's $24B Series D.","source":"latent_space","url":"https://www.latent.space/p/ainews-openai-reports-navier-stokes","published":"Wed, 09 Sep 2026 05:04:51 GMT"},{"title":"GPT-6 Astra, looped transformers, and hidden reasoning","summary":"Sebastian Raschka traces the recurrent-depth (\"looped\") transformer-block research behind GPT-6 Astra and its hidden chains of thought, rather than just covering the launch.","source":"sebastian_raschka","url":"https://magazine.sebastianraschka.com/p/gpt-6-astra-looped-transformers-and","published":"Wed, 09 Sep 2026 11:14:25 GMT"},{"title":"DeepSeek cuts API prices by up to 60%","summary":"DeepSeek cut its API prices by up to 60%, sharpening the cost gap builders already weigh between open-weight and frontier-lab models.","source":"search_cn_open_weight_labs","publisher_name":"36 Kr","publisher_domain":"eu.36kr.com","url":"https://news.google.com/rss/articles/CBMiU0FVX3lxTE9ULWFjNi1ZSUQ0YUVMbW40T1BxRldpSEhJMWlMa0pFajNZbi1YR2tjTk5tdGFOVmRIRzFHTjl3TXFzcXNSTGFwTTNGYUJDVGpMdTJz?oc=5","published":"Wed, 09 Sep 2026 05:34:04 GMT"}]},{"name":"AI Infra & Inference Economics","slug":"ai-infra-inference-economics","summary":"Compute supply and serving costs were the throughline today: chip access, a deprecated serving stack, retrieval latency, and sandbox density all moved.","articles":[{"title":"DeepSeek reportedly leans on 160,000+ Huawei chips to keep serving models","summary":"DeepSeek is reportedly relying on more than 160,000 Huawei chips to keep serving its models, a sign of how export controls are reshaping Chinese labs' compute stacks.","source":"search_cn_open_weight_labs","publisher_name":"WinBuzzer","publisher_domain":"winbuzzer.com","url":"https://news.google.com/rss/articles/CBMiqAFBVV95cUxOaFlOWVZEcGhqV1ZmeF9xZGhSc05mZGtXWXJ2M3l6UU1YWk9uYzhKWHdiWFlmWDFmSnlfOWt3THhNdmFHU0RoazBUeE1EemVDM0xGUUVaUjlTWTZqU0UzaWRDdEhhWjFNMlhWVjREd1ZCUXc5REN4NDJIRTFoZmoycm8wOGljU1U2U3JBMmJiZlBVQWI5bl9mdDFVRDVMcU5majQ5SXc4LWs?oc=5","published":"Wed, 09 Sep 2026 20:26:59 GMT"},{"title":"AWS retires TorchServe support in favor of Ray Serve Deep Learning Containers","summary":"With TorchServe now unmaintained, AWS is pointing teams to a pre-tested Ray Serve Deep Learning Container bundling the framework, GPU drivers, and serving layer instead of a DIY stack.","source":"aws_ml_blog","url":"https://aws.amazon.com/blogs/machine-learning/simplify-and-support-your-torchserve-workloads-using-ray-serve-deep-learning-containers/","published":"Wed, 09 Sep 2026 15:51:29 +0000"},{"title":"Databricks' Adaptive Instructed-Retriever claims 2x lower search latency","summary":"Databricks' Adaptive Instructed-Retriever aims at frontier-quality search at roughly half the latency, built for enterprise data agents that need both accuracy and speed.","source":"databricks_blog","url":"https://www.databricks.com/blog/adaptive-instructed-retriever-frontier-quality-search-2x-lower-latency","published":"Wed, 09 Sep 2026 13:30:00 GMT"},{"title":"Unikraft: fitting 1 million AI sandboxes on a single server","summary":"Unikraft's Felipe Huici describes millisecond cold boots and stateful scale-to-zero that let a single server host up to 1 million AI sandboxes via microVM isolation.","source":"infoq_ai_ml","url":"https://www.infoq.com/presentations/unikraft-microvm-sandboxes-cloud-scaling/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 09 Sep 2026 11:00:00 GMT"}]},{"name":"Security, Safety & China AI Policy","slug":"security-safety-china-ai-policy","summary":"Sandbox integrity and geopolitical model-theft accusations both surfaced today, alongside a notable safety-governance hire.","articles":[{"title":"Flaw in DeepSeek's coding-agent harness let agents disable their own sandbox","summary":"A flaw in DeepSeek's Harness coding-agent tool let agents disable their own file sandbox without approval, a reminder that agent-autonomy features are also new attack surface.","source":"search_cn_open_weight_labs","publisher_name":"DevOps.com","publisher_domain":"devops.com","url":"https://news.google.com/rss/articles/CBMilwFBVV95cUxQOXhpZEFsTGtrVldPWUdyeTF4MGxBVDdjc1RlYWkxUERzZkZsdF9fRzlhWU1xNXRrWGtkZ3lnMTVQLTRNTzhndW5COGQ3LWROYVhXUGNBQXhFUUU0Vm10TDhXVlBpSmoyMi1QRG9lbjd0RF8tUmJGTldUc2NjQVk5ZkZNd1UzWEVCRGt4VXR6amZTY1YtNXp3?oc=5","published":"Wed, 09 Sep 2026 15:57:34 GMT"},{"title":"Paul Christiano joins OpenAI's Foundation Board and safety committee","summary":"AI-alignment researcher Paul Christiano joined OpenAI's Foundation Board and its Safety and Security Committee.","source":"openai_blog","url":"https://openai.com/index/paul-christiano-joins-openai-foundation-board","published":"Wed, 09 Sep 2026 17:00:00 GMT"},{"title":"NSA, CISA, and FBI name six Chinese AI firms in mass-distillation warning","summary":"NSA, CISA, and the FBI warned that six named Chinese AI companies, including DeepSeek and Alibaba, are using mass distillation to siphon capabilities from American models.","source":"search_cn_open_weight_labs","publisher_name":"Help Net Security","publisher_domain":"helpnetsecurity.com","url":"https://news.google.com/rss/articles/CBMiqgFBVV95cUxNRVRERXRyVDlaZVhhZHM5aEdRUXJOVDJUZHlZRGo1N0lwb19XYThVeHJaSkQyMVN3MUZFbnZ6SVZ1RjdvTk8tMHBqX0FZcnJ0dnBzbDU5WUZ6eFRFZDQ4VUVKVEVUUHpoOUdxZXdUdUF0TXJXWVlYRDdTUHBXVzZHSU55Q25QajZFNkRuaFhFZmkydm1LbE5GbHpvYlFfRHVURU4wYkVEX3FXQQ?oc=5","published":"Wed, 09 Sep 2026 09:55:36 GMT"}]}]}