{"date":"2026-10-28","title":"What happened in AI — Oct 28, 2026","generated_at":"2026-10-28T22:16:22Z","intro":["The agent harness is turning into a shippable product: Microsoft opened the Copilot harness through Foundry and Cursor released its Agent SDK, while LangChain pushed scheduled and memory-backed agents.","The best evidence today is measurement: 270 guardrail runs and 324 agent sprints put numbers on what actually keeps coding agents reliable."],"highlights":["Microsoft makes the multi-model Copilot harness available through Foundry.","Cursor ships the Agent SDK for event-driven agents defined as code.","270 coding-agent runs test whether guardrails pay off.","324 agent sprints: 79% of committed stories done in-sprint, merge as slow as build.","Google open-sources Agent Mesh for verifiable agent identity.","Google DeepMind announces Gemini 4 Carbon."],"article_count":34,"categories":[{"name":"Agent harnesses and runtimes are becoming products","slug":"agent-harnesses-runtimes","summary":"Vendors are shipping the harness itself as a reusable layer: multi-model, event-driven, scheduled and memory-backed.","articles":[{"title":"Microsoft makes its Copilot agent harness available through Foundry","summary":"Microsoft exposes the Copilot harness behind Copilot Code, Copilot Studio and the CLI through Foundry as a multi-model agent layer.","source":"hackernews_ai","url":"https://blogs.microsoft.com/blog/2026/10/28/introducing-the-microsoft-copilot-harness-unlocking-the-potential-of-truly-multi-model-agentic-work/","published":"Wed, 28 Oct 2026 18:04:17 +0000"},{"title":"The Agent SDK","summary":"Cursor releases the Agent SDK it uses internally, for agents defined as code that act on events such as PR reviews or incidents.","source":"hackernews_ai","url":"https://cursor.com/blog/agent-sdk","published":"Wed, 28 Oct 2026 14:02:20 +0000"},{"title":"How to Build Proactive, Always-On Agents with the Schedules SDK in Managed Deep Agents","summary":"LangChain's Schedules SDK for Managed Deep Agents frames proactive agents as cron-triggered runs.","source":"langchain_blog","url":"https://www.langchain.com/blog/proactive-agents","published":"2026-10-28T07:15:00+00:00"},{"title":"How to build evidence-grounded agent memory with OpenWiki and Graphs","summary":"LangChain shows an evidence-grounded memory layer built on OpenWiki and graphs, so agents stop rebuilding context per question.","source":"langchain_blog","url":"https://www.langchain.com/blog/building-agent-memory-openwiki-graphs","published":"2026-10-28T17:24:00+00:00"}]},{"name":"Measured results on coding-agent guardrails and process","slug":"coding-agent-measurement","summary":"Today's strongest posts report run counts and scores, not opinions, on what makes coding agents reliable.","articles":[{"title":"Do Guardrails Make AI Coding Agents Better? What 270 Runs Showed","summary":"270 unattended runs across three models and seven tasks test whether instruction files and pre-commit gates pay for themselves.","source":"hackernews_ai","url":"https://secondtimeisabug.com/blog/coding-agent-guardrails-270-runs/","published":"Wed, 28 Oct 2026 01:46:12 +0000"},{"title":"Agent Scrum: Running a team of AI coding agents in two-hour sprints","summary":"Across 324 two-hour agent sprints, 79% of committed stories finished in-sprint, and approval-to-merge took as long as the build.","source":"hackernews_ai","url":"https://daveremy.github.io/agent-scrum/","published":"Wed, 28 Oct 2026 18:47:09 +0000"},{"title":"Getting an analytics agent from 50% to 96% accuracy","summary":"Benchouse took an analytics agent from 50% to 96% accuracy and found that more documentation sometimes made it drop the expected calculation.","source":"hackernews_ai","url":"https://benchouse.ai/blog/getting-an-analytics-agent-from-50-to-96","published":"Wed, 28 Oct 2026 16:55:43 +0000"},{"title":"Artificial Analysis v5 and the benchmarker's dilemma","summary":"Interconnects argues independent benchmarks are hard to keep meaningful at the frontier, framed around Artificial Analysis v5.","source":"interconnects","url":"https://www.interconnects.ai/p/artificial-analysis-v5-and-the-benchmarkers","published":"Wed, 28 Oct 2026 12:58:39 GMT"}]},{"name":"Agent identity and sandbox claims get testable","slug":"agent-identity-sandboxes","summary":"Two open-source projects target trust boundaries: verifiable agent identity and runnable contracts for sandbox claims.","articles":[{"title":"Agentmesh – open-source zero-trust agent network from Google","summary":"Google's Agent Mesh gives each agent node a verifiable identity and a way to publish and call tools across a private network, including behind NAT.","source":"hackernews_ai","url":"https://github.com/google/agentmesh","published":"Wed, 28 Oct 2026 13:12:00 +0000"},{"title":"I ran each AI coding-agent sandbox's own tests against the others","summary":"sandbox-conformance turns sandbox claims such as no key reads or no network into runnable contracts with per-sandbox results.","source":"hackernews_ai","url":"https://github.com/carlostapiacl/sandbox-conformance","published":"Wed, 28 Oct 2026 03:35:56 +0000"}]},{"name":"Frontier model and serving infrastructure","slug":"models-serving-infrastructure","summary":"A new Google frontier model was announced, alongside serving-side moves on freshness and routing.","articles":[{"title":"Gemini 4 Carbon: Our next frontier model","summary":"Google DeepMind announces Gemini 4 Carbon as its next frontier model; the announcement text in the feed carries no further detail.","source":"google_deepmind_blog","url":"https://deepmind.google/blog/gemini-4-carbon-our-next-frontier-model/","published":"Wed, 28 Oct 2026 17:00:58 +0000"},{"title":"Airbnb Cuts Recommendation Staleness from 2 Days to Seconds with Chronon","summary":"Airbnb used Chronon to cut recommendation feature staleness from two days to seconds.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/10/airbnb-chronon-recommendation/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Wed, 28 Oct 2026 13:53:00 GMT"},{"title":"Requesty: AI Gateway and LLM Router for 600 Models in EU Datacenters","summary":"Requesty offers one API over 600+ models with routing, failover, spend limits and per-request observability in EU datacenters.","source":"hackernews_ai","url":"https://www.requesty.ai","published":"Wed, 28 Oct 2026 08:20:54 +0000"}]}],"video":{"youtube_id":"USz96WWDxOQ","title":"LLM Digest Daily Explainer: Oct 28, 2026","duration_s":98}}