New features and launches in ChatGPT, Cursor, Perplexity and other AI tools, plus changes to pricing and availability.
Google DeepMind just unveiled CodeMender, an autonomous AI agent powered by Gemini Deep Think models that diagnoses and patches software vulnerabilities. Over six months, it has already upstreamed 72 security fixes to open-source projects, though human researchers still review every patch before submission.
Google just launched Jules Tools, a lightweight CLI bringing its AI coding agent straight into your terminal, alongside early access to the Jules API. Developers can now wire Jules into Slack bugs or CI/CD pipelines. Google also rolled out persistent memory for coding preferences, a file selector, and environment variable controls.
Perplexity just made its AI-first browser Comet free and available worldwide, officially scrapping the waitlist. Every new tab now comes packed with an embedded AI assistant, alongside new Background Assistants designed to work through tasks asynchronously. It is a clear push to turn standard web browsing into an agentic workflow.
Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, just unveiled Tinker, an automated tool designed to fine-tune frontier open-source models. Developers can use a few lines of code to run supervised fine-tuning or reinforcement learning on Meta's Llama and Alibaba's Qwen, bypassing messy GPU cluster management entirely. Access is open via application for free during beta.
Sony just pushed a firmware update to its flagship WF-1000XM5 earbuds and WH-1000XM6 headphones, adding support for Google's conversational Gemini Live assistant. The update also brings Auracast-powered audio sharing for Android users, letting friends listen together on supported Pixel and Galaxy devices. Why it matters: Premium noise-canceling cans are steadily evolving into real-time voice-first AI interfaces.
Microsoft just launched the Microsoft Agent Framework in public preview, bringing together research from AutoGen and the enterprise plumbing of Semantic Kernel into a single open-source SDK. It adds native support for Anthropic's Model Context Protocol (MCP) and multi-agent workflows across Azure. Why it matters: Developers no longer have to juggle fragmented tools to build and monitor collaborating AI agents.
Peloton just overhauled its Bike, Tread, and Row machines with the new Cross Training series. The upgraded gear features rotating displays and built-in motion cameras that count reps and nudge your squat form in real time, backed by Peloton IQ for personalized planning. The catch is that both device prices and monthly subscriptions are getting steeper.
Wikimedia Deutschland just launched an AI-friendly vector database to make 30 million Wikidata entries easily digestible for large language models. Built using Jina AI's embedding model and hosted by DataStax, the project turns clunky structured data into rich contextual vectors. The goal is to level the playing field for smaller developers outside Big Tech so their AI agents can easily surface niche facts.
Anthropic just rolled out a deep two-way integration with Slack. You can now summon @Claude directly in channels, DMs, and threads to draft replies and analyze files, or link your workspace so Claude can search message history for context. Public thread replies stay private until you approve them.
Amazon just rolled out new AI features on its updated Kindle Scribe, targeting forgetful readers. The "Story So Far" tool delivers a quick, spoiler-free recap of what you last read, while "Ask This Book" lets you highlight text to query character motives and plot points. Why it matters: It is one of the most practical everyday consumer uses of AI, solving the real pain of dropping a novel and forgetting who is who.
Granola just rolled out Recipes, a feature letting users create and reuse prompt shortcuts directly inside the AI note-taking app. By typing “/” in chat, users can invoke prompts across single or multiple meetings without manually copying transcripts to external chatbots. It comes with pre-made templates sorted by before, during, and after meetings.
PayPal just announced its Honey extension is integrating directly into AI chatbots, starting with ChatGPT. When you ask an AI for buying advice, Honey automatically surfaces real-time pricing, coupons, and seller options, while plugging retail blind spots the bot missed. It’s part of PayPal’s broader push into agentic commerce.
Ring just unveiled its first-ever 4K camera lineup with Retinal Vision, paired with smart features aimed at neighborhood safety. The standout tool, Search Party, lets outdoor cameras use AI to spot missing dogs across the neighborhood, while Alexa+ Greetings acts as a front-door butler to screen visitors. Pre-orders are open now.
Opera just launched Neon, an AI browser that handles tasks like filling out forms, comparing data across sites, and drafting code directly on the page. Unlike cloud-heavy rivals, its "Neon Do" agent runs actions locally on your device for privacy. It's pitched as a paid subscription for power users, starting in early access.
Google Search just upgraded its AI Mode with conversational visual exploration and shopping. Instead of fiddling with filters or struggling to describe a vibe in words, you can now chat naturally, upload photos, and refine results on the fly. It is rolling out in English across the U.S. this week.
Nothing just rolled out Playground and Essential Apps for Nothing Phone 3, letting users build and share custom apps, widgets, and back-display animations entirely through prompts. The feature is launching in alpha on Nothing OS 4, and developers can export the generated code to keep tweaking. It is an early step toward Nothing's goal of an AI-focused operating system.
A comprehensive guide breaks down how Factory Droid shifts software development from single-prompt code edits to coordinated agent teams. By pairing a research-first Spec Mode with native Git worktrees and milestone-driven Missions, Factory focuses on multi-file refactors and migrations that quickly overwhelm standard coding assistants.
Google just added AI-powered ransomware detection to Drive for desktop. When the specialized model spots signs of mass file encryption, it automatically pauses cloud syncing to stop the damage and lets users restore healthy file versions in a few clicks. It is rolling out now in open beta across most Workspace tiers.
Microsoft just launched Copilot Portraits, an experimental feature in Copilot Labs that pairs voice chats with 40 stylized avatars. Powered by Microsoft Research's VASA-1, the system generates real-time lip-sync and facial expressions from a single image. It is currently rolling out to select adult users in the US, UK, and Canada.
WhatsApp just rolled out a batch of new features that let users generate custom chat themes and video call backgrounds using Meta AI. The update also adds Live Photos sharing and Android document scanning, though Meta AI tools aren't available to everyone just yet.
