The White House just signed an executive order launching the Genesis Mission, an ambitious initiative likened to the Manhattan Project to unite federal scientific datasets and supercomputers. Run by the Department of Energy, the program will build a unified computing platform and deploy AI agents to automate scientific workflows across semiconductors, quantum computing, and nuclear energy.
Microsoft just open-sourced Fara-7B, a compact 7B-parameter model specifically built to take over keyboard and mouse actions. By reading raw browser screenshots instead of code trees, it automates routine web tasks like booking trips or comparing prices. Why it matters: It is small enough to run locally on Copilot+ PCs, delivering fast agentic workflows without uploading private screen data.
Anthropic just released Claude Opus 4.5, calling it the world's best model for coding, agents, and computer use. It even outperformed human candidates on the company's internal software engineering exam. With API pricing lowered to $5/$25 per million tokens and a new effort parameter to cut token usage, frontier intelligence is getting both smarter and cheaper.
Anthropic just rolled out new defenses against prompt injection attacks, helping Claude Opus 4.5 resist malicious instructions hidden across the web. The upgrades combine reinforcement learning with improved classifiers to drive attack success rates down to 1%, allowing the Claude for Chrome extension to expand into beta for Max plan subscribers. Why it matters: as agents take over real-world browsing and form filling, securing untrusted web content is critical—though Anthropic stresses a 1% risk is far from a solved problem.
Perplexity just relaunched its AI shopping experience for free across all users ahead of Black Friday, adding chat memory and intelligent product cards to replace endless item grids. It taps past queries for personalized picks and leverages a PayPal partnership for checkout. Why it matters: AI search engines are racing to turn chatbots into full-blown shopping agents that can browse and buy for you.
MIT engineers just trained an AI to operate CAD software exactly like human designers, turning 2D sketches into 3D models. By learning from VideoCAD, a new dataset of over 41,000 screen-recorded UI actions, the agent mimics every click and drag needed to build objects. It sets the stage for an autonomous CAD co-pilot.
OpenAI just launched GPT-5.1-Codex-Max, an agentic coding model built to run continuously for over 24 hours. Using native context compaction, it works seamlessly across multiple windows on multi-million-token refactors while cutting thinking tokens by 30%. It is now the default engine inside Codex.
Microsoft just overhauled Microsoft 365 Copilot at Ignite 2025, introducing Work IQ as a contextual intelligence layer and Agent 365 as an enterprise control plane for governing autonomous AI agents. The suite also embeds agent modes directly into Word, Excel, and PowerPoint with expanded models like Anthropic. Why it matters: Enterprise AI is shifting from individual chat assistants to complex multi-agent teams, and Microsoft is positioning itself as the infrastructure layer running them.
Google just unveiled Gemini 3, its next-generation frontier model family led by Gemini 3 Pro. The model claimed the top spot on LMArena with an unprecedented 1501 Elo, delivering huge upgrades in multimodal reasoning and vibe coding alongside a new agentic platform called Google Antigravity. Why it matters: Google is pushing hard to retake the AI crown by turning models into autonomous, self-executing software agents.
Anthropic just expanded its partnership with Microsoft, bringing Claude Sonnet 4.5, Haiku 4.5, and Opus 4.1 to Microsoft Foundry in public preview. Azure enterprise customers can deploy Claude through existing Azure agreements, while Excel's Agent Mode now lets users tap Claude to build spreadsheets and generate formulas. It removes procurement hurdles for teams wanting frontier reasoning inside Microsoft's ecosystem.
Manus just launched Manus Browser Operator, a browser extension that lets its AI agent work directly inside your local browser. By using your existing login sessions and trusted local IP, it can automate tasks across authenticated CRMs and paid research platforms without getting tripped up by CAPTCHAs. Complex actions like drag-and-drop are still limited, however.
Microsoft just rolled out Copilot Actions to Windows Insiders, letting the AI agent complete tasks directly across local files. Running in a separate, isolated Agent Workspace desktop, it can sort messy downloads, organize vacation photos, or pull data from PDFs while you monitor or take over at any time. Why it matters: It turns Copilot from a side-panel chatterbox into a hands-on assistant that actually does PC chores for you.
Google just expanded Search's travel capabilities with AI-powered itinerary planning in Canvas and direct agentic bookings. Users can now build custom trip plans using real-time Maps and hotel data, while AI Mode will search reservation platforms to help book dinner tables and event tickets.
Levi Strauss & Co. just teamed up with Microsoft to build an Azure-native superagent inside Teams. Instead of jumping between tools, employees use one conversational portal that delegates tasks to multiple specialized sub-agents across retail, warehouses, and headquarters. It shows how legacy brands are betting on agent orchestration to rewire corporate operations.
Shenzhen startup MindOne Robotics just shared a viral demo showing a Unitree G1 humanoid tackling household chores like watering plants, wiping counters, and fetching packages. The company claims the entire video runs autonomously with zero teleoperation or speed-ups, sparking intense industry curiosity given humanoid robotics' notorious generalization hurdle.
Researchers at Google Cloud and UCLA just proposed Supervised Reinforcement Learning (SRL), a training framework designed to supercharge small language models on complex multi-step reasoning. Instead of traditional RLVR that only rewards the final correct answer, SRL breaks problem-solving into key logical actions and gives step-by-step feedback. This dense reward system helps smaller, cheaper models master difficult math and software engineering tasks without getting trapped in all-or-nothing failures.
Google just added an autonomous Deep Research agent to NotebookLM along with support for Word docs, Google Sheets, Drive URLs, and images. The AI can now plan multi-step web searches in the background, browse hundreds of pages, and dump a structured, sourced briefing directly into your notebook.
Google DeepMind just introduced SIMA 2, an AI agent powered by Gemini that plays and reasons inside 3D virtual worlds. Unlike its predecessor, it doesn't just follow simple commands; it operates via mouse and keyboard, explains its goals in natural language, and can self-improve in newly created Genie environments. It is currently available as a limited research preview for select academics and game developers.
Fei-Fei Li, Yann LeCun, and DeepMind are all heavily betting on "world models," but they basically mean three completely different architectures. World Labs' Marble delivers interactive 3D Gaussian splats, LeCun's JEPA builds an internal predictive latent brain, and DeepMind's Genie 3 and SIMA 2 generate real-time video simulators. Why it matters: The industry is packaging front-end assets, robot gymnasiums, and core machine cognition under a single marketing buzzword.
Anthropic just disrupted what it describes as the first documented large-scale cyber espionage campaign executed almost entirely by AI. Attackers tricked Claude Code into infiltrating roughly 30 global targets, leaving the model to autonomously research exploits, harvest credentials, and exfiltrate data for 80% to 90% of the campaign. Why it matters: The barrier to sophisticated cyberattacks has collapsed, proving agentic AI can already operate as an autonomous offensive weapon.
