Presented by
Good morning, AI enthusiasts. For 50 years, the mouse pointer only tracked where you were pointing. Google DeepMind just made it understand what you're pointing at.
It's a deceptively simple change that could reshape how most people interact with computers, and Google is already weaving it into Chrome and a new laptop category called Googlebook.
In today's recap:
Google DeepMind reinvents the 50-year-old mouse pointer
Google's Googlebook laptop built for Gemini Intelligence
Draft and review contract clauses with Claude for Legal
Meta launches Muse Spark for natural AI voice
4 new AI tools, prompts, and more
GOOGLE DEEPMIND
Google DeepMind reimagines the pointer for AI era
Google DeepMind
Recaply: Google DeepMind just unveiled an AI-enabled mouse pointer powered by Gemini, letting users point at any on-screen element and speak in natural shorthand instead of writing detailed prompts.
Key details:
The pointer captures visual and semantic context around the cursor, letting the AI see exactly which word, paragraph, image, or code block the user needs help with, no detailed prompt required.
Google is integrating the technology into Chrome today and into Googlebook at launch, with two interactive experimental demos live now in Google AI Studio.
Researchers Adrien Baranes and Rob Marchant describe the goal as eliminating "AI detours," the friction of switching to a separate AI window to get help with any task.
Magic Pointer is live in Chrome now for contextual questions on webpages, rolling out soon to Googlebook, with ongoing experiments across Google Labs' Disco platform.
Why it matters: For 50 years, the mouse pointer tracked where you were pointing. Now it understands what. That shift sounds subtle, but it's actually a fundamental redesign of how humans interact with computers. Rather than pulling content into an AI chat window, AI meets users where they're working. Whether this replaces the chat interface entirely is still open, but Google's integration into Chrome and Googlebook means hundreds of millions of users will test the answer soon.
PRESENTED BY ADQUICK
Scale Your IRL Campaigns Like Digital Ads
Out Of Home advertising has long been effective but hard to scale—until now. AdQuick makes it simple to plan, deploy, and measure campaigns with the same efficiency and insight you expect from online marketing tools.
Marketers agree: OOH is powerful for brand growth, driving new customers, and reinforcing messaging. AdQuick makes it easy, intuitive, and data-driven—so you can treat real-world campaigns like any other digital channel.
Google enters laptop market with Googlebook
Recaply: Google just introduced Googlebook, a new laptop category built from the ground up for Gemini Intelligence, with a Magic Pointer, custom AI widgets, and deep Android phone integration.
Key details:
The Magic Pointer uses Gemini to offer contextual suggestions at the cursor, while users can build custom widgets with natural language to create a personalized app dashboard.
The announcement drove 807 upvotes and 1,318 comments on Hacker News, making it the most discussed story of the day, ahead of Starship V3 and other major news.
DeepMind's AI pointer research, published the same day, feeds directly into Googlebook's Magic Pointer experience, with both teams coordinating on the new cursor paradigm for Chrome and laptops.
Googlebook launches this fall through hardware partners featuring a distinctive glowbar design, with premium pricing expected. Why it matters: Google making a laptop isn't new. Chromebook has existed for years. But Googlebook is different, built around Gemini as the operating layer, not as a feature bolted on. If the AI pointer works as described, the laptop becomes the first consumer device where pointing replaces prompting. Android device sync adds a compelling reason for the hundreds of millions of existing Android users to consider it as their next machine.
GUIDES
Draft and review contract Clauses with Claude for Legal
Recaply: In this tutorial, you will learn how to use Claude's Legal plugin and MCP connectors to review contracts for risks, flag missing clauses, and generate redlined versions, cutting hours of manual review down to minutes.
Step-by-step:
Open Claude Desktop, go to the Cowork tab, click Plugins in the left sidebar, search for "Legal," and click Install. The plugin activates immediately and its slash commands become available in your workspace.
Start a new Cowork session and grant Claude access to a folder containing your contract files. Claude reads PDFs, DOCX, and plain text, so drop the agreement you need reviewed into the designated folder.
Type /legal-review and reference your contract. Prompt: "Review this agreement against standard commercial terms. Flag any missing clauses, unusual limitations of liability, and non-standard IP assignment language. Use GREEN/YELLOW/RED risk scoring."
Review Claude's clause-by-clause output, which highlights deviations from your organization's standards. For each YELLOW or RED flag, prompt: "Draft alternative language for [clause name] that protects our interests while keeping the agreement balanced."
Export the redlined version by prompting: "Generate a tracked-changes DOCX with all suggested revisions and a one-page risk summary." Use this as the basis for your next legal review meeting or send directly to counsel for sign-off.
Pro tip: Connect your document management system (Box or Egnyte) via MCP connectors to let Claude pull contracts directly from your existing folders without manual uploads, and set up a playbook with your organization's standard clause language so Claude always redlines against your specific terms.
TOGETHER WITH HCAPTCHA
Protect your platform, your way.
hCaptcha is the leading privacy-first, bot management platform.
Define rules and monitor behavior to stop fraud and abuse with privacy-first, frictionless bot protection.
From startups to the top online services and Fortune 100 enterprises, our systems stop bots and fraud without slowing down real users.
Want to learn more? Book a demo today.
META
Meta's Muse Spark arrives for natural AI voice
Meta
Recaply: Meta just launched AI Voice Conversations powered by Muse Spark, letting users interrupt, switch topics, and swap languages mid-conversation, with real-time image generation and Reels recommendations built into replies.
Key details:
Muse Spark powers natural voice conversations in the Meta AI app, with a Live AI mode letting users point their camera at the world and ask about what they see in real time.
Muse Spark is available on meta.ai and rolling out across WhatsApp, Instagram, Facebook, and Messenger, with Ray-Ban Meta and Oakley Meta glasses rollout beginning within weeks.
According to Meta's announcement, the company plans to open-source future versions of Muse Spark, with an API private preview already open to select partner companies.
Available now on meta.ai in the US and Canada, with glasses integration rolling out over the coming weeks and Meta Ray-Ban Display support arriving this summer.
Why it matters: Meta's Superintelligence Labs has been quietly building for months, and Muse Spark is its first public product. The voice-first, interruption-friendly design sets it apart from GPT-4o Voice and Gemini Live, which still feel transactional. Meta's real edge is in its distribution. Billions of users across four apps, plus Ray-Ban and Oakley glasses with real-world camera context for live AI. Voice AI isn't a features race anymore. It's a distribution race, and Meta may be better positioned for it than anyone.
TOOLS
Trending AI Tools
🎥 Perceptron Mk1 - Frontier video understanding model matching GPT-4o at $0.15/M input
✨ QuiverAI SVG - Paper's new AI SVG generation model for exploring logo ideas and illustrations in vector format
🤖 Qwen3.6-27B - Alibaba Cloud's open-source dense model
🎨 Krea 2 - Krea AI's first foundation model built for aesthetic diversity and stylistic control
NEWS
What Matters in AI Right Now?
Isomorphic Labs secured $2.1B in Series B funding led by Thrive Capital, with Alphabet, GV, MGX, Temasek, and the UK Sovereign AI Fund participating. CEO Sir Demis Hassabis said the capital will scale its IsoDDE drug design engine and advance drug candidates toward clinical trials.
Anthropic released 20+ new MCP connectors and 12 practice-area plugins connecting Claude to the software legal teams run on, including contract lifecycle systems, e-discovery platforms, and document management tools. Legal professionals have become Claude Cowork's most engaged users of any knowledge-work function since the platform's launch earlier this year.
The Claude Platform on AWS is now generally available, giving AWS customers access to the full Claude API with AWS authentication, billing, and commitment retirement, with Anthropic operating the service and new features shipping the same day they go live on the native Claude API.
Google introduced Gemini Intelligence for Android, bringing automation that handles multi-step tasks like filling shopping carts from email syllabuses and turning voice notes into polished text. The Rambler feature and custom widget creation are rolling out to select Samsung and Google phones this summer.
Lovable published AIUC-1, a 51-requirement standard covering six principles for AI coding agent security, safety, and reliability. Lovable is among the first platforms to pursue certification, with a third-party audit scheduled for summer 2026.
About 65% of U.S. doctors, or roughly 650,000 physicians, actively use OpenEvidence, an AI clinical decision tool, across 27 million encounters in April alone, NBC News found. Another 1.2 million use it internationally, though few patients are aware their doctor consults AI during visits.
JudgmentLabs launched with $32M in Seed and Series A funding led by Lightspeed Venture Partners, building infrastructure that helps teams improve AI agents by analyzing production data. Co-founder Alex Shander said the platform turns agent behavioral logs into a repeatable improvement loop.
🧡 Enjoyed this issue?
🤝 Recommend our newsletter or leave a feedback.
How'd you like today's newsletter?
Cheers, Jason
