San Francisco AI startup Deep Cogito just emerged from stealth and open-sourced Cogito v1, a lineup of hybrid reasoning models ranging from 3B to 70B parameters. Trained on top of Llama with an AlphaGo-style self-play loop, the models toggle between instant answers and deep reflection while adding native tool calling.
Shopify CEO Tobi Lütke told staff that teams asking for more headcount or resources must first demonstrate why AI cannot do the work. Effective AI usage is now a baseline expectation at the e-commerce giant and will be baked into peer reviews, turning AI into the default benchmark before hiring humans.
ElevenLabs just rolled out an official Model Context Protocol (MCP) server, plugging voice cloning, text-to-speech, and transcription straight into Claude Desktop, Cursor, and Windsurf. It offers a hosted endpoint so you don't even need to run a local script. Why it matters: Your coding assistants and chat agents can now spit out full audio and soundscapes on demand.
Microsoft just overhauled Copilot with persistent memory, real-world vision on mobile, and agentic Actions that book tickets and dinner reservations across the web. The update also brings a native Windows app accessible via Alt+Space, AI-generated podcasts, and Deep Research. Microsoft is shifting Copilot from a generic chat widget into a deeply customized assistant that acts on your behalf.
GitHub just rolled out Agent mode and MCP support to all VS Code users, turning Copilot into an autonomous developer that edits multiple files, runs terminal commands, and self-heals runtime bugs. It also opened access to Claude 3.7 Sonnet and Gemini 2.0, alongside a new $39-a-month Pro+ plan.
Sentient Foundation just released Open Deep Search (ODS), an open-source framework designed to challenge closed AI search tools like Perplexity and ChatGPT Search. ODS pairs custom reasoning agents with web search, plugging into models like DeepSeek-R1 or GPT-4o. Paired with DeepSeek-R1, it even beat GPT-4o Search Preview on complex benchmark queries.
Amazon just began testing "Buy for Me," an AI shopping agent powered by its Nova models and Anthropic's Claude that browses third-party stores and completes purchases. If Amazon lacks an item, the agent securely auto-fills shipping and billing info on external sites without leaving the app. Returns still go through the third-party seller, and it is currently limited to select users.
Cognition AI just launched Devin 2.0 in beta, slashing its steep $500 monthly barrier down to a $20 starting plan. The revamp introduces a cloud-hosted IDE capable of running multiple Devins at once, plus interactive planning and automated code documentation. Why it matters: Lowering the price tag gives the high-profile AI software engineer a second chance to prove its worth against surging rivals.
Microsoft is using its 50th birthday to unveil the next phase of its consumer Copilot assistant while actively trying to reduce its reliance on OpenAI. Reports indicate the tech giant is developing its own reasoning models, dubbed MAI, and testing alternative models from Meta, DeepSeek, and xAI. Why it matters: Even OpenAI's biggest backer is hedging its bets to avoid being locked into a single provider.
Lindy AI just rolled out agent swarms, allowing its AI agents to duplicate themselves and run hundreds of tasks at the same time. The team also announced over 5,000 integrations and 4,000 web scrapers across the platform. Why it matters: It shifts AI agents from sequential helpers into scalable parallel workforces.
Google DeepMind just broke down its new Gemini Robotics family, fine-tuning Gemini 2.0 to give robots real-world dexterity and reasoning. Powered by Gemini Robotics-ER and a new vision-language-action model, dual-arm robots can tackle unfamiliar tasks like slam dunking a toy ball or packing lunches on the first try. It is an impressive step toward general-purpose embodied AI that moves beyond single-task programming.
Viral AI agent platform Manus just introduced paid subscriptions starting at $39 a month, alongside a brand-new iOS app. Paying users get dedicated task credits and parallel runs, while the service also upgraded its backend model to Anthropic's Claude 3.7 Sonnet.
Amazon just launched Nova Act, a general-purpose AI agent that takes the wheel in your browser to book reservations, order food, and fill out forms. Developed by its new AGI lab to power the upcoming Alexa+, it takes direct aim at OpenAI's Operator and Anthropic's Computer Use. It's available now as a research preview.
Former interim OpenAI CEO Emmett Shear just co-founded Softmax, a 10-person research startup pursuing "organic alignment" inspired by biology and ant colonies. Instead of top-down hierarchical control, the team is running reinforcement learning tests on small agents in virtual worlds to study emergent cooperation. Softmax has raised backing from Andreessen Horowitz and others.
QwenLM just officially launched the first version of QVQ-Max, a dedicated visual reasoning model designed to analyze images and videos. The model goes beyond basic perception to solve complex math and logic tasks, with MathVision benchmark accuracy scaling up as its thinking time increases. Why it matters: QwenLM plans to evolve this foundation into a visual agent that can operate smartphones and computers.
OpenAI just announced that it is adding support for rival Anthropic’s open-source Model Context Protocol (MCP) across its entire product lineup. It is available immediately in the Agents SDK, with ChatGPT desktop and Responses API rollouts coming soon. Two fierce rivals are now aligned on the same data standard, making it far simpler to hook AI into real-world software.
OpenAI just rolled out major upgrades to its security apparatus as it races toward AGI. The company boosted its top bug bounty payout from $20,000 to $100,000, broadened research grants to cover agent resilience, and teamed up with SpecterOps for continuous red-teaming. As autonomous agents take the wheel, defensive stakes have never been higher.
BMW and Alibaba just expanded their partnership to bring Alibaba's Qwen LLM into BMW's Neue Klasse electric vehicles in China starting in 2026. The co-developed AI engine powers dual in-car agents capable of handling complex multi-intent navigation and lifestyle queries. Why it matters: Global automakers are increasingly relying on local AI giants to survive China's fiercely competitive smart EV market.
Amazon just launched an AI-powered feature called Interests that acts like a personal product scout. Shoppers can write natural language prompts describing niche hobbies, style preferences, or budget limits, and LLMs will continuously monitor new inventory to alert them of restocks, deals, and new arrivals. It is rolling out first to select U.S. mobile users.
Microsoft just introduced Researcher and Analyst, two new reasoning agents inside Microsoft 365 Copilot built for complex workplace tasks. Researcher pairs OpenAI's deep research model with company data to run multi-step market reviews, while Analyst taps o3-mini to execute Python code directly against messy spreadsheets. They roll out to Microsoft 365 Copilot license holders this April under a new Frontier early-access preview.
