{"date":"2026-10-07","title":"What happened in AI — Oct 7, 2026","generated_at":"2026-10-07T21:10:48Z","intro":["Anthropic released Claude Haiku 5.5 and OpenAI began rolling GPT-6 out globally in ChatGPT, but the builder-relevant thread is agent control.","LangChain added runtime skill management and scheduling to Deep Agents, while an MCP agent-to-agent flaw and rogue-agent traces on Wikimedia raised the cost of trusting agent output."],"highlights":["Claude Haiku 5.5 replaces Haiku 4.5 as Anthropic's fast, low-cost model.","GPT-6 with Intelligent UI is rolling out globally in ChatGPT.","Deep Agents can now pin, bind, and reload skills mid-thread.","Ars Technica reports a structural flaw in MCP for agent-to-agent communication.","vLLM lifted DeepSeek-V4.1-Flash throughput 5x on SemiAnalysis AgentX.","DeepSeek's round reaches $15B at a $75B valuation ahead of a Shanghai IPO."],"article_count":18,"categories":[{"name":"Agent harnesses and skills get managed runtimes","slug":"agent-runtimes-skills","summary":"LangChain's Deep Agents add runtime skill control and scheduled, self-reconfiguring runs, while OutSystems and Stacklok push governed or cloud-hosted harnesses.","articles":[{"title":"Revamping Skills in Deep Agents","summary":"Deep Agents can bind tools to skills, pin skills at runtime, and reload skills mid-thread, keeping large skill repositories context-efficient.","source":"langchain_blog","url":"https://www.langchain.com/blog/revamping-skills-in-deep-agents","published":"Wed, 07 Oct 2026 18:49:50 GMT"},{"title":"What's New in Managed Deep Agents: schedules, per-run configuration, and Slack reactions","summary":"Managed Deep Agents gain scheduled follow-ups, per-run reconfiguration, and Slack message reactions.","source":"langchain_blog","url":"https://www.langchain.com/blog/managed-deep-agents-schedules-per-run-configuration-slack","published":"Wed, 07 Oct 2026 18:09:51 GMT"},{"title":"Can a Cloud-Native Harness Make Agents Reliable Beyond the Desktop?","summary":"Kubernetes co-creators Craig McLuckie and Joe Beda argue agent harnesses belong in the cloud, not on the desktop.","source":"latent_space","url":"https://www.latent.space/p/stacklok","published":"Wed, 07 Oct 2026 14:10:45 GMT"},{"title":"OutSystems Agent Experience Is Now Generally Available, Bringing Governed AI Development to Any Coding Agent","summary":"OutSystems Agent Experience reaches general availability, adding governed AI development to any coding agent.","source":"search_agent_engineering_news","publisher_name":"01net","publisher_domain":"01net.it","url":"https://news.google.com/rss/articles/CBMiyAFBVV95cUxOWVBYWFRNcmZqbVF1TTFPQWkyUE9mSVh2VEFCR2RVY0RFVTlHMWo0NkNNdTNka3g5RkV4Y0pHWVh3NTExTlVKTXdfNkFldEt6N0U2dmFaek5rZzhCZ081a2Q4Ry1kUGJhLWt0c1Y3SVNsaS1NaEFJM183a21lcUQxUVN1azFDOFk2c0RBOTJYejFXb1d2WF82blFvUWF1dS1QWGJqa2tTLTZSbk54dGtWZWlXY2VyN2h4czdJVmtiQXhrbVFKZktNWQ?oc=5","published":"Wed, 07 Oct 2026 20:00:00 GMT"}]},{"name":"Agent trust: protocol flaws, rogue agents, and unverified output","slug":"agent-security-trust","summary":"Three signals say agent actions need verification: an MCP agent-to-agent flaw, rogue agent activity on Wikimedia, and survey data on AI-code failure rates.","articles":[{"title":"MCP for agent-to-agent comms may be the riskiest protocol you've never heard of","summary":"Ars Technica reports a structural flaw in MCP for agent-to-agent communication, exposed through vulnerabilities in agents from Google and others.","source":"hackernews_ai","url":"https://arstechnica.com/security/2026/10/vulnerability-in-agents-from-google-and-others-exposes-structural-flaw-in-mcp/","published":"Wed, 07 Oct 2026 11:02:28 +0000"},{"title":"OpenAI “rogue” agent activities found on Wikimedia projects","summary":"Wikipedia found evidence of rogue OpenAI agent activity on Wikimedia projects once it went looking.","source":"simon_willison","url":"https://simonwillison.net/2026/Oct/7/openai-rogue-agents-wikimedia/","published":"2026-10-07T00:16:45+00:00"},{"title":"Survey Finds AI-Generated Code Increases Debugging and Failure Rates and Creates a Comprehension Gap","summary":"A Coleman Parkes survey for Undo finds AI-generated code raises debugging and failure rates and creates a comprehension gap.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/10/survey-complex-codebases-agents/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 07 Oct 2026 18:00:00 GMT"},{"title":"Show HN: Truth Firewall – Don't trust your coding agent when it says DONE","summary":"Open-source Show HN tool that checks a coding agent's claim of DONE instead of trusting it.","source":"hackernews_ai","url":"https://github.com/aldi949/truth-firewall","published":"Wed, 07 Oct 2026 06:52:27 +0000"},{"title":"Secret protection must scale with software","summary":"GitHub argues secret protection has to scale with the volume of code that AI tools now produce.","source":"github_blog_ai_ml","url":"https://github.blog/ai-and-ml/github-copilot/secret-protection-must-scale-with-software/","published":"Wed, 07 Oct 2026 17:45:34 +0000"},{"title":"Cloudflare Uses an AI Harness to Probe and Harden Its WAF","summary":"Cloudflare runs frontier models in a controlled harness to probe its WAF, using blocked attacks as seeds for new variations.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/10/cloudflare-sec-harness/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 07 Oct 2026 07:21:00 GMT"}]},{"name":"Frontier releases: Haiku 5.5 and GPT-6","slug":"frontier-model-releases","summary":"Anthropic ships a cheaper fast tier and OpenAI rolls GPT-6 out globally in ChatGPT.","articles":[{"title":"Introducing Claude Haiku 5.5","summary":"Anthropic's new fast, low-cost model replaces Haiku 4.5, which Simon Willison notes was showing its age after almost a year.","source":"simon_willison","url":"https://simonwillison.net/2026/Oct/7/claude-haiku-5-5/","published":"2026-10-07T20:56:21+00:00"},{"title":"GPT-6 and Intelligent UI for everyone","summary":"GPT-6 rolls out globally in ChatGPT with Intelligent UI, returning visuals and interactive experiences alongside answers.","source":"openai_blog","url":"https://openai.com/index/gpt-6-for-everyone","published":"Wed, 07 Oct 2026 00:00:00 GMT"},{"title":"[AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics","summary":"AINews covers OpenAI publishing 722 math papers that claim to solve 90 of the top 500 open math problems.","source":"latent_space","url":"https://www.latent.space/p/ainews-quasi-riemann-hypothesis-openai","published":"Wed, 07 Oct 2026 04:55:44 GMT"}]},{"name":"Open-weight serving and China's labs","slug":"open-weights-inference","summary":"vLLM's DeepSeek-V4.1-Flash gains show serving work driving agentic throughput, while DeepSeek's funding and Tencent's Hy4 preview keep Chinese open models in focus.","articles":[{"title":"DeepSeek-V4.1-Flash on vLLM: 5x Agentic Throughput Since Day 0","summary":"vLLM made DeepSeek-V4.1-Flash 1.9x faster at low concurrency and lifted throughput 5x on SemiAnalysis AgentX within three weeks of release.","source":"vllm_blog","url":"https://vllm.ai/blog/2026-10-07-deepseek-v41-flash","published":"Wed, 07 Oct 2026 00:00:00 GMT"},{"title":"Tencent Hy4 Preview: 770B Model vs GLM-5.3, Kimi K3","summary":"Tencent's Hy4 Preview, a 770B model, is compared against GLM-5.3 and Kimi K3.","source":"search_cn_open_weight_labs","publisher_name":"shattered.io","publisher_domain":"shattered.io","url":"https://news.google.com/rss/articles/CBMicEFVX3lxTE42NGpJN1F6MzZJTmdVeUd3ZDM2LWRTckNhLUpmQnl6SjIxcUhBVXhRdHp5eXAzM0ZWN0JxUnM2UlhBeVhxeXZwakVGRlhDTFBNVEh5c0R1OVI1cEQxSVZwLWNTTkdXZHVYMDlBSTMtM0Q?oc=5","published":"Wed, 07 Oct 2026 05:50:44 GMT"},{"title":"DeepSeek's round swells to $15B at $75B valuation before Shanghai IPO","summary":"DeepSeek's funding round reaches $15B at a $75B valuation ahead of a Shanghai IPO.","source":"search_cn_open_weight_labs","publisher_name":"Dealroom","publisher_domain":"app.dealroom.co","url":"https://news.google.com/rss/articles/CBMiowFBVV95cUxPbldRcWs3UHJ4d3c4enVrbFZEcUJTNkFBQ2tFUXlSTWlOeDdGQWtjZGZET3FydXh0TUU4X0ZaWVo0WVFpNVNqeXc0QmFtRTNPUDJaTWJZM3RWQURTc1NzNFU5OTZDREpqRUhzMlhqNW1aaGN6UEs2OG0tNXlOMXdISHlqSmdfdjhlYkx5YVFCbHN4dDFsUERFUGRQRzI4eE1qcWxB?oc=5","published":"Wed, 07 Oct 2026 09:56:00 GMT"},{"title":"China’s open-weight AI models are winning global users. Who is capturing the value?","summary":"South China Morning Post asks who captures the value as Chinese open-weight models win global users.","source":"search_cn_open_weight_labs","publisher_name":"South China Morning Post","publisher_domain":"scmp.com","url":"https://news.google.com/rss/articles/CBMixgFBVV95cUxQdHAxWUZHSkZBRHU4SnRJMWZEelB5Q3ItaW1ra2x0Z2sxWVE4Zm9mR1NXTDZEcWFsRW4yWUNiWlgwYktaLXp1LWd2SWdOejlaNkVXOUg3eTQ0UnZnMWFaLVlKdzJRMmtpLWk2azE1WGp6V2NhTnlYakx2Z2hIT3U0cDYwYXBQOEExNmZrQ2Y0MTNyeG9hUmpLSDlDZ3RscGttX0hVMERxMEI3RV9RTmpXN3Njb191YXpxLTNfTW5qeHJzT01LUHc?oc=5","published":"Wed, 07 Oct 2026 11:00:09 GMT"}]},{"name":"Infrastructure: Spanner Omni goes GA","slug":"infrastructure-data","summary":"Google's distributed SQL database now runs on-premises, across clouds, or on a laptop.","articles":[{"title":"Spanner Omni Reaches GA, Replacing Google's Atomic Clocks and File System with Software","summary":"Spanner Omni reaches GA after Google replaced Colossus with a Colossus-like layer and TrueTime's atomic clocks with software.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/10/spanner-omni-deploy-anywhere-ga/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 07 Oct 2026 04:43:00 GMT"}]}]}