AI Agents & Automation
-
Your coding agents get desks and robot bodies in Termi Protocol
Multi-agent coding has a blind-spot problem. Run three agents on one repo and you’re staring at three scrolling terminals with no idea who’s touching what. Termi Protocol turns that black box into a 3D room you can actually watch. What it actually is It’s a desktop app that gives your real CLI agents — Claude… Continue reading
-
Microsoft Frontier Company: $2.5B to put 6,000 engineers inside customers’ offices
Microsoft just admitted the quiet part out loud: selling AI models isn’t the hard part — getting them to actually work inside a business is. On July 2, it launched Frontier Company, a $2.5 billion unit built to fix exactly that. What it actually is Frontier Co. is a services organization, not a product. Microsoft… Continue reading
-
Google Search Information Agents turn 1 billion AI Mode users into agent operators
Google is rolling out Information Agents inside Search, starting with AI Pro and Ultra subscribers this summer. These aren’t chatbots you poke at — they run 24/7 in the background, watching news, shopping, finance, and whatever topics you care about, then ping you with a summary and next steps at the moment it actually matters.… Continue reading
-
Alibaba Page-Agent (in-page JavaScript GUI agent) skips screenshots — it reads the DOM instead
Everyone building web agents right now points a multimodal model at screenshots — browser-use, UI-TARS, OpenAI’s computer use. Alibaba’s new open-source Page-Agent went the opposite way: no screenshots, no vision model, no browser extension, no headless browser. It’s plain JavaScript that lives inside your page, reads the live DOM as text, and clicks and types… Continue reading
-
Microsoft Copilot merges into one app and ships AutoPilot background agents
Microsoft is collapsing two Copilots into one. Come August 2026, the consumer Copilot and the enterprise Microsoft 365 Copilot become a single app — same identity, capabilities that flex based on your subscription and data permissions. Low-use features like Copilot Podcasts get cut. This is Microsoft’s answer to the AI super app race Anthropic and… Continue reading
-
Devin Security Swarm (Cognition) finds 36 of 50 real GHSA vulnerabilities — 3 that every other tool missed
Devin isn’t just writing code anymore. On July 1, Cognition launched Devin Security Swarm, an AI agent product that hunts exploits: a swarm of parallel Devin agents sweeps your entire codebase (Cognition calls it Agentic MapReduce), reproduces every finding in an isolated sandbox to prove it’s actually exploitable, then ships a remediation PR. No pattern… Continue reading
-
694 stars in one day: Vibe-Trading (HKUDS) wants to do for trading what vibe coding did for code
The HKU lab behind LightRAG and AI-Trader has done it again. Vibe-Trading (HKUDS) jumped 694 GitHub stars in a single day, now around 17K total. This lab is an open-source hit factory at this point. What it is A trading agent that plugs into Claude Code or Codex CLI as a skill. You describe a… Continue reading
-
ZCode (Zhipu / z.ai GLM-5.2 coding agent) hit HN’s 1 spot — China’s open answer to Claude Code
Zhipu (now z.ai) just shipped ZCode, a desktop coding agent tuned for its GLM-5.2 model. It topped Hacker News at 260 points. The pitch: this is China’s open-weight camp taking a direct swing at Claude Code. What it actually is ZCode is a desktop ADE — think Claude Code, but built around GLM-5.2 and designed… Continue reading
-
Pluno skips the UI and talks to APIs — 14x faster than Claude’s browser extension
Most browser agents work like a slow human: screenshot the page, find the button, move the cursor, click, wait, screenshot again. Pluno throws that whole loop out. It’s a browser extension that ignores the interface entirely and talks straight to the private APIs sitting underneath web apps like HubSpot, Notion, and Stripe. You describe the… Continue reading
-
Claude Sonnet 5 scores 63.2% on SWE-bench Pro at a third of Opus 4.8’s price
Anthropic shipped Claude Sonnet 5 on June 30, and the pitch is blunt: this is now the cheapest way to run agents that don’t fall apart. It’s a mid-tier LLM built for agentic workflows — planning, calling tools, driving a browser or terminal, and grinding through long coding tasks without a human babysitting every step.… Continue reading
Top AI Product 