Vercel Workflows hits general availability after processing over 100 million runs and 500 million steps across 1,500+ customers, offering developers a durable execution model that embeds workflow logic directly into TypeScript or Python code — no separate orchestration services required.
Roblox supercharges its AI Assistant with agentic Planning Mode, automated playtesting, and new 3D mesh generation tools, creating a self-correcting development loop that helps creators build, test, and debug games faster than ever.
Billions of dollars are flooding into humanoid robotics as tech giants and startups race to build AI-powered machines, with experts calling it a potential platform revolution — but serious challenges around dexterity, training data, job displacement, and overhyped expectations threaten to slow the boom.
Cisco's Jeetu Patel warns that AI agents are becoming a dangerous new cybersecurity threat as the developer population could explode from 150 million to 3 billion within a year, urging enterprises to prioritize trusted, governed AI access before nation-states and bad actors exploit autonomous agents at scale.
Google DeepMind launches Gemini Robotics-ER 1.6, a major upgrade featuring enhanced spatial reasoning, multi-view understanding, and a new industrial gauge-reading capability built with Boston Dynamics — now available to developers via the Gemini API.
Anthropic launches a sweeping Claude Code overhaul featuring a multi-session desktop UI, drag-and-drop controls, an integrated terminal, and new Claude Code Routines that automate coding tasks on a schedule — a direct enterprise power move against OpenAI's rumored superapp combining Codex, Atlas, and ChatGPT.
Claude Mythos Preview becomes the first AI to complete a complex 32-step corporate network attack simulation, succeeding 3 out of 10 attempts and achieving a 73% success rate on expert-level cybersecurity challenges that no AI could previously solve, alarming security experts who urge organizations to strengthen basic cyber defenses immediately.
AI training compute has exploded 1 trillion times since 2010, while deployment costs have collapsed 900x annually, and with global capacity forecast to hit 100 million H100-equivalents by 2027, human-level autonomous AI agents capable of handling complex, multi-week tasks are rapidly approaching reality.
Microsoft is testing a powerful OpenClaw-like AI agent for enterprise Microsoft 365 Copilot users, featuring enhanced security controls and the ability to autonomously complete complex, multistep tasks — with a full reveal expected at Build 2026 in June.
OpenAI is merging ChatGPT, its Atlas browser, and Codex into a single desktop superapp for 900 million users, rolling out incrementally over the next few months as the company transforms its AI platform into a powerful agentic productivity tool for knowledge workers.