Epic's Claude Code plugin for Unreal went viral this weekend, but it dates to June. It packages a skill and hook for Unreal MCP, the Experimental UE 5.8 server that also works with Cursor and Codex. Setup steps, toolsets, and Epic's security warnings.
monday.com launched Harmony, a voice AI platform whose agents call leads, qualify them, handle objections and hand warm prospects to reps. It uses small task-specific models, starts at $0.39 a minute plus $250 a month, and runs in the US, UK, Australia and Singapore.
Into the Unknown, a new benchmark from CMU and DOE lab scientists, asked nine frontier agents to write 135 follow-up papers. None passed all three gates, and 114 failed integrity, mostly for claiming work never done or numbers their logs contradict.
Modal, which raised $750M at a $15.75B valuation last week, says Nvidia's Arm-compatible Vera CPU gave up to double the task throughput of an x86 Intel GNR machine in its sandbox benchmark. It will run Vera alongside Intel and AMD in a mixed fleet.
World launched World ID for banking and finance at Money20/20: a Zoom Human Badge that matches an Orb image, a selfie and the live video, three levels of proof, and beta tools that let a verified person delegate proofs to an AI agent. Bain Capital and VanEck Funds are design partners.
We checked free and paid vibe coding courses as of October 2026: DeepLearning.AI + Replit, Vibe Code School, zero2vibecode, Cursor Learn, Anthropic Claude Code 101, Codecademy, Coursera/Scrimba, Great Learning, and Google/Kaggle agent materials. What is free, what costs money, and where to start.
Google is replacing Gems with skills in the Gemini app, starting with personal accounts in November 2026. Here are the dates, what skills cannot do yet, how they differ from tasks and scheduled actions, and how to move a Gem yourself.
Meta's Muse Wonder costs $4 input and $20 output per million tokens on the API, while Muse Wonder Full needs a $50 or $200 Muse Code plan. What Active Memory is, how Contributor mode works, and who can use it today.
A practical vibe-coding workflow for Claude: define the term (Karpathy, Feb 2025), set up Projects or Claude Code with CLAUDE.md, use Anthropic's XML and system-prompt guidance, then run a step-by-step loop with copy-paste prompts. As of October 2026. Prompts untested.
GitHub made Copilot local sandboxing generally available on October 7, 2026 in the CLI, Copilot app, and VS Code Agent Host. It limits filesystem, network, and credential access for agent tools via MXC, at no extra cost. Cloud sandboxes remain usage-billed.
Huntington Bank launched The 1866 Studio, an AI system built on Google Cloud's Gemini platform that helps small business owners launch, market and get loan ready. Here is what it does, how it is built, and how it compares with banks blocking outside agents.
DeepSeek's new Search and Fetch APIs cost $0.75 per 1,000 successful calls. Developers praise the price, but early tests put it behind Exa and Parallel, and Hacker News commenters flag empty results on sensitive queries, no zero data retention, and a data path to China.
Meta's Muse, OpenAI's Dots, Grok and Instinct keep hitting walls at Amazon, Delta, Yelp, Chase and Schwab, and 150+ Chinese apps bar ByteDance's Doubao. Who blocks agents, why, which blocks are accidents, and what it means if you use one.
Microsoft announced controls in Microsoft 365 E7 and Project Perception that score each AI agent's risk and can make it read-only, require human approval, or shut it down. It now detects 60 agent types, up from 25. E7 costs $99 per user per month.
DeepSeek priced Search and Fetch at $0.75 per 1,000 successful requests, put them on the same API key as its models, and made server-side Search and Fetch calls through Claude Code and Codex free for DeepSeek API users until November 1. That is how agent search becomes a commodity.
A Chinese-speaking hacker used Cairn AI agents to try more than 1,200 websites and break into the sites of 100+ companies, including Deere, SharkNinja and Cummins, Reuters reported. NorForge's researcher estimates the tokens cost about $5,000, or roughly $50 per successful hack.
Google's Gemini agent works inside Gmail, Docs and Sheets, runs tasks for hours or days, and can join your team as a coworker agent with its own email. It is in private preview. Google says it comes at no extra charge with Gemini Enterprise; Workspace plans list from $7 to $35 per user per month.
Dots are OpenAI's always-on GPT-6 Astra agents in ChatGPT. Which plans include them, why personal Pro users in the EEA, UK and Switzerland are left out while Business Premium is not, how to set one up, and the safeguards and privacy limits to know first.
Anthropic's Frontier Red Team ran a 26B model across 15 cheap Android TV boxes and found GLM-5.2 can spread end to end in a simplified test. It projects small models could reach that level by the end of 2026.
OpenAI's Agents API runs the Codex harness as a hosted service: you send a task, tools and a sandbox choice, and OpenAI runs the loop. Here is how it works, what it costs, its limits, and how it compares with the Claude Agent SDK and Claude Managed Agents.
