{"date":"2026-10-13","title":"What happened in AI — Oct 13, 2026","generated_at":"2026-10-13T21:10:26Z","intro":["Google launched Gemini 4 Argon on DeepMind and Google Cloud, but the more actionable news for builders came from agent teams quantifying what works: LangChain's harness changes lifted a coding agent from Top 30 to Top 5 on Terminal Bench without a model change.","Open models are closing the gap on agent tasks, and the infrastructure underneath is being rebuilt for agent volume, from Anthropic's new version control system to Modal's mixed-hardware RL."],"highlights":["Gemini 4 Argon launches on DeepMind and Google Cloud, with a safety write-up.","LangChain: harness engineering took its coding agent from Top 30 to Top 5 on Terminal Bench.","LangChain evals: GLM-5 and MiniMax M2.7 match closed frontier models on core agent tasks.","Anthropic replaced its source control with an in-house system after a tenfold rise in merges per day.","Modal keeps RL stable across mixed-precision, mixed-accelerator samplers.","GitHub's secret classifier cuts median scan time by 86%."],"article_count":36,"categories":[{"name":"Agent harnesses, evals, and production builds","slug":"agent-harnesses-evals-production","summary":"Teams are showing measured gains from harness changes and eval-driven builds rather than model swaps.","articles":[{"title":"Improving Deep Agents with harness engineering","summary":"Self-verification, tracing, and context optimization moved LangChain's coding agent from Top 30 to Top 5 on Terminal Bench.","source":"langchain_blog","url":"https://www.langchain.com/blog/improving-deep-agents-with-harness-engineering","published":"2026-10-13T16:04:23Z"},{"title":"Open Models have crossed a threshold","summary":"LangChain's evals say GLM-5 and MiniMax M2.7 now match closed frontier models on file operations, tool use, and instruction following, at lower cost and latency.","source":"langchain_blog","url":"https://www.langchain.com/blog/open-models-have-crossed-a-threshold","published":"2026-10-13T16:11:23Z"},{"title":"Coinbase Built, Evaluated, and Post-Trained a Fraud Agent","summary":"Coinbase details how it built, evaluated, and post-trained a fraud-detection agent.","source":"hackernews_ai","url":"https://www.coinbase.com/blog/part-one-how-coinbase-built-evaluated-and-post-trained-a-fraud-agent","published":"2026-10-13T19:16:20Z"},{"title":"Search agent traces by behavior with AI Search in LangSmith","summary":"LangSmith AI Search finds traces by describing the behavior in natural language instead of filtering on fields.","source":"langchain_blog","url":"https://www.langchain.com/blog/agent-traces-ai-search-langsmith","published":"2026-10-13T12:59:12Z"},{"title":"Quoting Geoffrey Huntley","summary":"Geoffrey Huntley's fix for agent drift: enforce diagnostics-library use with a pre-commit hook on the git diff, not AGENTS.md edits.","source":"simon_willison","url":"https://simonwillison.net/2026/Oct/13/geoffrey-huntley/","published":"2026-10-13T13:32:00Z"}]},{"name":"Gemini 4 Argon launch","slug":"gemini-4-argon-launch","summary":"Google shipped Gemini 4 Argon across DeepMind, Google Cloud, and robotics, with a safety write-up the same day.","articles":[{"title":"Gemini 4 Argon: Built to solve your hardest problems","summary":"Google DeepMind introduces Gemini 4 Argon, pitched at its hardest-problem workloads.","source":"google_deepmind_blog","url":"https://deepmind.google/blog/introducing-gemini-4-argon/","published":"2026-10-13T16:01:52Z"},{"title":"Gemini 4 Argon available on Google Cloud, built to help solve your business’ hardest problems","summary":"Gemini 4 Argon is available on Google Cloud for enterprise use, tied to the Gemini agent in Gemini Enterprise.","source":"google_cloud_blog","url":"https://cloud.google.com/blog/products/ai-machine-learning/gemini-4-argon-available-on-google-cloud/","published":"2026-10-13T16:00:00Z"},{"title":"Developing Gemini 4 Argon responsibly and safely","summary":"DeepMind publishes the safety and responsibility approach behind the Gemini 4 Argon release.","source":"google_deepmind_blog","url":"https://deepmind.google/blog/developing-gemini-4-argon-responsibly-and-safely/","published":"2026-10-13T19:43:07Z"}]},{"name":"Training and developer infrastructure","slug":"training-developer-infrastructure","summary":"Infrastructure posts focus on stable RL across mixed hardware and on scaling source control for agent-driven commit volume.","articles":[{"title":"Toward RL on mixed-precision, mixed-accelerator samplers","summary":"Modal trained an agentic-coding model with fully asynchronous RL on mixed-precision, mixed-accelerator samplers; off-policy correction kept training stable and on par with homogeneous samplers.","source":"modal_blog","url":"https://modal.com/blog/toward-rl-on-mixed-precision-mixed-accelerator-samplers","published":"2026-10-13T12:00:00Z"},{"title":"Redesigning source control at Anthropic","summary":"Anthropic migrated to an in-house version control system in five months after merges per day grew more than tenfold.","source":"anthropic_engineering","url":"https://www.anthropic.com/engineering/redesigning-source-control","published":"2026-10-13T00:00:00Z"},{"title":"GitHub’s model for secret detection: Finding more secrets with less noise","summary":"GitHub's context-aware secret classifier raises weighted acceptance by 11.5% and cuts median scan time by 86%.","source":"github_blog_ai_ml","url":"https://github.blog/ai-and-ml/githubs-model-for-secret-detection-finding-more-secrets-with-less-noise/","published":"2026-10-13T17:56:33Z"}]},{"name":"Open-weight and frontier model moves","slug":"open-weight-frontier-model-moves","summary":"Open-weight agent tooling from Alibaba and a new Meta frontier model add competitive pressure on closed labs.","articles":[{"title":"Alibaba's Qwen team open sources QwenGyre to train AI agents that run for hours","summary":"Alibaba's Qwen team open-sources QwenGyre to train agents that run for hours.","source":"search_cn_open_weight_labs","publisher_name":"Startup Fortune","publisher_domain":"startupfortune.com","url":"https://news.google.com/rss/articles/CBMiqgFBVV95cUxOWjRmbC1iVXVLRGp0cjM1ZU5Zd01HSFU3bkhtU0hWN0NEQ0kxRl9DODQwMVZKYnRoRElTYkN6RkN3NGVFXzBfaDFZazBFY0RYemNRT25uNGx1UmVSY21BR3JtRjYydzhRVGtPRlI0M1VOTzRNQUZER1RIdkdzWlpYY1V3eTVpM1pDVl9KMHBtTjBIT0k1U3dsM3RGOTB2MWdJVGs4ZC1naFVUdw?oc=5","published":"2026-10-13T01:20:21Z"},{"title":"[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model","summary":"Latent Space covers Meta's Muse Wonder models and what they mean for MSL as a frontier lab.","source":"latent_space","url":"https://www.latent.space/p/ainews-meta-muse-wonder-and-wonder","published":"2026-10-13T09:37:04Z"}]}]}