{"date":"2026-10-31","title":"What happened in AI — Oct 31, 2026","generated_at":"2026-10-31T22:15:52Z","intro":["AI developer tooling took the hits today: Wiz counts at least nine AI developer platforms breached in October, and EMBRACE found command-execution flaws in the built-in browsers of five coding tools.","Meanwhile inference keeps getting cheaper, with Anthropic tokens down 77% and prompt caching reaching up to 94% cache reads in one measured agent."],"highlights":["Wiz: at least nine AI developer platforms breached in October, driving LLMjacking.","EMBRACE: built-in browsers in five AI coding tools could run commands or read the disk.","Epoch: open-weight GLM 5.3 exploited six Firefox releases for about $1,000 each.","Prompt caching took one agent to up to 94% cache reads, about a quarter of the bill.","Apple Intelligence's Qwen access opens for testing in China on November 4."],"article_count":23,"categories":[{"name":"Agent Builder Security Under Attack","slug":"agent-builder-security-under-attack","summary":"Agent tooling is now the attack surface: platform breaches, embedded browsers and cheap open-weight exploits all landed in one day.","articles":[{"title":"Black October: At least nine AI developer platforms breached this month","summary":"Wiz counts at least nine AI developer platforms breached this month, with attackers monetizing stolen access through LLMjacking at scale.","source":"hackernews_ai","url":"https://www.wiz.io/blog/black-october-for-ai-developer-platforms","published":"2026-10-31T00:11:32Z"},{"title":"How the built-in browsers in five AI coding tools broke basic browser rules","summary":"Forever Security's EMBRACE research found built-in browsers in five AI coding tools could run commands, read the whole disk or spoof the address bar.","source":"hackernews_ai","url":"https://theainewsreport.com/2026-10-30-embrace-ai-coding-assistant-embedded-browser-flaws-explained.html","published":"2026-10-31T03:32:31Z"},{"title":"Open-weight AI models can exploit the patch gap","summary":"Epoch showed GLM 5.3, an open-weight model, exploiting the patch gap to compromise six recent Firefox releases at about $1,000 each.","source":"hackernews_ai","url":"https://epoch.ai/publications/open-weight-models-patch-gap","published":"2026-10-31T08:53:31Z"},{"title":"Google Discloses PageBreak AI Security Agent That Found 500+ XSS Bugs","summary":"Google disclosed PageBreak, a security agent that found more than 500 XSS bugs.","source":"infoq_ai_ml","url":"https://www.infoq.com/news/2026/10/google-pagebreak-security-agent/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering","published":"Sat, 31 Oct 2026 06:20:00 GMT"},{"title":"Show HN: InBox Wall – A local-first, open-source email sanitizer for AI agents","summary":"Open-source, local-first sanitizer that strips hidden CSS, Unicode obfuscation and prompt injection from email HTML before an agent reads it.","source":"hackernews_ai","url":"https://github.com/Annora998/inboxwall","published":"2026-10-31T14:59:36Z"}]},{"name":"Coding Agent Runtimes and Workflow","slug":"coding-agent-runtimes-and-workflow","summary":"New open-source agents and tooling focus on context survival, review, and guardrails rather than raw capability.","articles":[{"title":"DeepSeek open-sources dsh-superagent, a coding agent largely built by agents","summary":"DeepSeek published a research-prototype coding agent whose code was largely written by coding agents; it labels it not production-ready.","source":"hackernews_ai","url":"https://github.com/deepseek-ai/dsh-superagent-prototype","published":"2026-10-31T13:41:11Z"},{"title":"Octomind: A coding agent that recalls archived messages after compaction","summary":"Open-source MCP-native agent runtime that recalls archived messages after context compaction.","source":"hackernews_ai","url":"https://github.com/muvon/octomind","published":"2026-10-31T04:54:13Z"},{"title":"An AI agent read a Shift_JIS page as 208,953 replacement characters","summary":"An OpenClaw fix for pages in legacy charsets that reached agent input as 208,953 replacement characters.","source":"hackernews_ai","url":"https://github.com/openclaw/openclaw/pull/188766","published":"2026-10-31T09:05:23Z"},{"title":"Show HN: Governor, a firewall for AI agent spend and tool calls","summary":"Firewall that checks each agent tool call and spend against ordered rules and signs every decision into an audit log.","source":"hackernews_ai","url":"https://bsvkey-governor-demo.fly.dev/","published":"2026-10-31T00:27:46Z"},{"title":"DiffNotes: Lets your coding agent explain its PRs inline","summary":"Has the coding agent explain its own pull request inline in the diff on GitHub, GitLab and Bitbucket.","source":"hackernews_ai","url":"https://github.com/igorkasyanchuk/diffnotes","published":"2026-10-31T14:52:18Z"},{"title":"Lakebase branching for CI/CD: lessons from 100 PR databases","summary":"Snorkel AI uses copy-on-write database branching to give each of 100 PRs an isolated copy of production data in CI/CD.","source":"databricks_blog","url":"https://www.databricks.com/blog/lakebase-branching-cicd-lessons-100-pr-databases","published":"Sat, 31 Oct 2026 21:25:00 GMT"}]},{"name":"Inference Cost and Serving Efficiency","slug":"inference-cost-and-serving-efficiency","summary":"Token prices keep falling while serving techniques shrink memory and latency, shifting where the margin is.","articles":[{"title":"Cheaper AI inference is making it harder to profit from renting out compute","summary":"Darkbloom's idle-Mac inference network reports earnings of $1.23 a day as Anthropic token prices fall 77%.","source":"hackernews_ai","url":"https://www.tomshardware.com/tech-industry/artificial-intelligence/firm-that-lets-you-rent-out-your-idle-mac-for-ai-inference-workloads-warns-of-plummeting-token-costs-darkbloom-network-sees-earnings-drop-to-usd1-23-a-day-as-anthropic-tokens-fall-77-percent","published":"2026-10-31T17:13:50Z"},{"title":"Prompt caching: same agent, same rules, about a quarter of the bill","summary":"A recipe agent measured on a real API went from no cache reads to up to 94% in four changes, cutting the bill to about a quarter.","source":"hackernews_ai","url":"https://blog.getwerk.ai/posts/prompt-caching-a-quarter-of-the-bill","published":"2026-10-31T13:09:52Z"},{"title":"Qwen-Based EdgeRock-4B Caps Its KV Cache at 72 MiB a Request","summary":"Qwen-based EdgeRock-4B caps its KV cache at 72 MiB per request.","source":"search_cn_open_weight_labs","publisher_name":"FourWeekMBA","publisher_domain":"fourweekmba.com","url":"https://news.google.com/rss/articles/CBMikgFBVV95cUxQSTJZNVNIOHRsZ1dGQWEtZEZsOEYtd0hfNkpFTFJWMUNYV0E4bFFTRDdSUWJudzVZQnptbUI4bkRmREQzU3RfSHVKMFlMZUFjTXRqRkF5dFdRdXNiRlhYaHVuZ2lUZHAtUE1uajNmVTNOT1ExaEUtdDI0Zk44MGk5MlZSeGhnT3BVS2wtRENEcUN1QQ?oc=5","published":"Sat, 31 Oct 2026 13:57:22 GMT"},{"title":"Non-autoregressive NC LLM spits tokens 21x faster than text diffusion","summary":"Stanford's Hazy Research reports Next-Concept Modeling generating tokens 21x faster than text diffusion.","source":"hackernews_ai","url":"https://hazyresearch.stanford.edu/blog/2026-10-12-ncm","published":"2026-10-31T06:31:05Z"},{"title":"GPUCost – Cloud GPU Prices Normalized by VRAM and LLM VRAM Sizer","summary":"Compares cloud GPU rental prices per GB of VRAM and FP8 TFLOP, with a VRAM sizer and self-hosting break-even calculator.","source":"hackernews_ai","url":"https://gpu-cost-r34p.vercel.app/","published":"2026-10-31T16:31:00Z"}]},{"name":"Models and Platform Moves","slug":"models-and-platform-moves","summary":"Qwen's distribution widens while Anthropic quietly adjusts Claude Code behavior.","articles":[{"title":"Alibaba, Apple partner to bring Qwen to Apple Intelligence for users in China","summary":"Alibaba and Apple will bring Qwen to Apple Intelligence for users in China.","source":"search_cn_open_weight_labs","publisher_name":"Global Times","publisher_domain":"globaltimes.cn","url":"https://news.google.com/rss/articles/CBMiYkFVX3lxTE9wd25LX1dZZXEtR3ZUTi1nOXVFUW9LSXluR3c4c3pGaDhYTmxkaGRrdndJWXlFY0l5bFgxWmZwZnlmRklIeGZUblNSTDBZV296UklWMVJ6elFnR21lVHdMakx3?oc=5","published":"Sat, 31 Oct 2026 03:15:00 GMT"},{"title":"Apple officially announced: Apple's intelligent access to Alibaba Qwen will be open for testing on November 4.","summary":"Apple Intelligence's Qwen access opens for testing on November 4.","source":"search_cn_open_weight_labs","publisher_name":"富途牛牛","publisher_domain":"news.futunn.com","url":"https://news.google.com/rss/articles/CBMisgFBVV95cUxQemVab0FPVDhmZFBXcmlxRG4zVHZaQV9adVNYV3lOVjZKa3g0ZGxGZWVTdWdOZTQ1eEJiSkZybkVzUVpXUUY1Y2FWTm1yZHkyampKS2V3clp2QnBrYWFMaWh6YmtrNGVBUENDTW42ODhyMERqWmM5UGxycXRaSkJpOUhid0hUYWZrWWJaQVVWOXhDMWpWeVMwejF2S2pKSnBOTjVDM2J6c1dhVFVVbmVDdzlR?oc=5","published":"Sat, 31 Oct 2026 01:46:24 GMT"},{"title":"Qwen Code Now Sizes Unlisted Qwen4 Models at 1M Tokens","summary":"Qwen Code now sizes unlisted Qwen4 models at 1M tokens of context.","source":"search_cn_open_weight_labs","publisher_name":"FourWeekMBA","publisher_domain":"fourweekmba.com","url":"https://news.google.com/rss/articles/CBMiigFBVV95cUxNaGdQQXZfM0VzeVBIVVZaSTdGV3JpTzJsZ2FsUFktaUs3UTQ1eDVMOXRYQy1oVVFrQVpycnBGNk4zeUo1VTdzTS1rdzhYTzFWbGdsUENfYjhpM1E5dG1lYU9Nc2QyVFVjRzhGLW1LSjdlNngzZXp1NkFQWThHb0E0ek1mOE9QeGFYblE?oc=5","published":"Sat, 31 Oct 2026 12:31:08 GMT"},{"title":"Anthropic orders AI tool Claude to stop remaking video games","summary":"Anthropic confirmed an Opus 5.5 update that steers Claude Code away from recreating someone else's specific game.","source":"hackernews_ai","url":"https://www.gamedeveloper.com/programming/anthropic-appears-to-have-ordered-ai-tool-claude-to-stop-remaking-video-games","published":"2026-10-31T14:41:44Z"}]},{"name":"Evals, Policy and Business Signals","slug":"evals-policy-and-business-signals","summary":"Builders get a lightweight eval tutorial and a new crawler opt-out, while Google's AI Mode revenue trade-off hits its stock.","articles":[{"title":"To Bin or Not to Bin: A Tutorial on Fast Feedback Loops for AI Capabilities","summary":"Hands-on tutorial taking one small problem from first prompt to an agent with automatic scheduled checks.","source":"hackernews_ai","url":"https://alexhans.github.io/posts/to-bin-or-not-to-bin.html","published":"2026-10-31T13:20:47Z"},{"title":"Google Adopts New AI Training Control","summary":"Google plans to adopt an IETF proposal for an AI training control signal, alongside Bing, OpenAI and Anthropic.","source":"hackernews_ai","url":"https://www.seroundtable.com/google-ai-training-control-42430.html","published":"2026-10-31T06:01:40Z"},{"title":"[AINews] Google admits AI Mode hurts Search revenue; -15% $500B drop worst since 2004 IPO. AI Bear Market coming?","summary":"Google acknowledged AI Mode hurts Search revenue; the stock fell 15%, about $500B, its worst drop since the 2004 IPO.","source":"latent_space","url":"https://latent.space/p/ainews-google-admits-ai-mode-hurts-search-revenue-15-500b-drop-worst-since-2004-ipo-ai-bear-market-c/","published":"Sat, 31 Oct 2026 05:00:48 GMT"}]}],"video":{"youtube_id":"fVvQLEvTcv8","title":"LLM Digest Daily Explainer: Oct 31, 2026","duration_s":111}}