🧠 Google launches Personal Intelligence

Good Morning! Google just launched Personal Intelligence, a feature that connects Gemini to your Gmail, Photos, Search, and YouTube to create a truly personalized AI assistant. Plus, you’ll learn how to use Manus 1.6 to build apps, connect to your existing software and even operate inside your real browser to complete tasks end to end.

In today’s AI newsletter:

• How to Use an AI Agent That Actually Does Real Work

• Google Launches Personal Intelligence for Gemini

• Gemini Helps Prove a New Math Theorem

• AI Learns From Evolution to Design New Drugs

• 4 new AI tools worth trying

AI NEWS

🧠 Google Launches Personal Intelligence for Gemini

Google says Gemini will now connect your Gmail, Photos, Search, and YouTube data to provide truly personalized AI assistance, marking the company's biggest push into context-aware AI.

• Gemini can now reason across your entire Google ecosystem to surface proactive insights

• Uses Gemini 3 to pull specific details from emails, photos, videos, and search history

• Privacy-first: Off by default, you choose which apps to connect and can disconnect anytime

• Currently rolling out to Google AI Pro and Ultra subscribers in the US only

• Coming to free tier and AI Mode in Search "soon"

This is Google's direct response to Apple Intelligence. By keeping all data within its own ecosystem, Google is betting that your existing digital footprint is the secret to making AI actually useful. If successful, this could finally deliver the truly personal AI assistant we've been promised; one that understands your context without you having to explain it every time.

AI MATH

📐 Gemini Helps Prove a New Math Theorem

Google announced that a new result in algebraic geometry was proved with substantial help from Gemini 2.5 Deep Think and an internal math-specialized system called FullProof. According to the authors, the solution was not in the training data and shows genuine originality.

• Gemini and FullProof contributed key arguments to the main proof, not just verification

• Authors say the ideas don’t closely match existing proofs or known sources

• Co-author Dr. Ravi Vakil called Gemini’s insight “the kind I’d be proud to produce myself”

• Follows recent breakthroughs like GPT-5.2 helping solve multiple Erdős problems

The gap between AI-assisted and human-only mathematical discovery is shrinking fast. With Gemini now contributing original insights in deep algebraic geometry, AI is moving from “helper” to co-discoverer in frontier science.

AI RESEARCH

🧬 AI Learns From Evolution to Design New Drugs

Developed with NVIDIA, Eden was trained on DNA collected across 28 countries, learning how nature solved biological problems over billions of years. Instead of guessing, the AI studies evolution itself to design safer gene therapies and powerful new antibiotics.

• Trained on evolutionary DNA data spanning 1M species across the planet

• Designed a new gene-editing method that inserts DNA without cutting it (safer than CRISPR)

• 63%+ of AI-designed gene therapies worked in lab tests for diseases like muscular dystrophy

• Created new antibiotics, with 97% effective against drug-resistant superbugs

Most medicines are slow, expensive, and reactive. By teaching AI to learn directly from evolution, Basecamp could radically speed up treatments for genetic diseases and tackle the growing crisis of antibiotic resistance, where current drugs are failing fast.

HOW TO AI

🗂️ How to Use an AI Agent That Actually Does Real Work

In this tutorial, you’ll learn how to use Manus 1.6 that goes far beyond chatting and suggestions. This tool can build apps, connect to your existing software, design visuals, and even operate inside your real browser to complete tasks end to end.

🧰 Who is This For

• Builders who want AI to execute, not just suggest

• Creators and founders tired of juggling multiple AI subscriptions

• Non-technical users who want apps, automations, and designs without code

• Anyone curious about where autonomous AI agents are headed next

STEP 1: Build a Real App From a Simple Prompt

When you start a new session in Manus, you don’t write code or configure tools. You simply describe what you want to build. For example, you can ask it to create a simple goal-tracking mobile app. Within minutes, Manus generates a working prototype, complete with frontend, backend logic, a built-in database, and an editable file structure.

You can instantly preview the app on your phone using a QR code, edit the code if you know how, or let Manus handle everything for you. You can even ask it to integrate services like Stripe or export the project to GitHub or app stores. This step alone replaces multiple development tools.

STEP 2: Connect Manus to the Apps You Already Use

Manus becomes far more powerful once you connect it to your everyday tools. Inside a chat, you can enable access to apps like Gmail, Google Calendar, Google Drive, or Notion. Once connected, Manus can read data, trigger actions, and complete workflows across these platforms.

For example, you can ask it to schedule a meeting, send emails, or fetch information from your calendar. You don’t need to understand any technical setup. You simply grant access once, and Manus handles the rest in the background.

STEP 3: Run Massive Workflows Using MCP and Automations

To unlock even more capabilities, Manus connects with automation platforms through MCP. This allows it to work with thousands of apps and actions across spreadsheets, analytics tools, task managers, and more. You can ask it to pull data from Google Sheets, update dashboards, track social media growth, or sync information across multiple platforms automatically.

Instead of manually switching between tools, Manus acts like a digital operator that executes multi-step workflows from a single instruction, saving hours of repetitive work.

California AG Rob Bonta opens an investigation into xAI over the proliferation of nonconsensual, sexualized images generated by Grok, and urges xAI to act.

OpenAI strikes a multibillion-dollar agreement to buy 750 MW of computing capacity from Cerebras over three years; sources: the deal is worth more than $10B.

X says Grok will no longer edit images of real people into bikinis or revealing clothing, and is locking image generation behind a paywall as regulators step in.

Verizon says it is “aware of an issue impacting wireless voice and data services for some customers”, as thousands of users report no cellular reception.

🤖 Claude Cowork: Anthropic’s feature that turns Claude into a fully autonomous virtual assistant

🎥 Veo 3.1: Google’s video AI with better references and vertical videos

📱 Gridfy: Turn raw data into interactive website widgets

💬 Slackbot: Slack’s smarter built-in AI assistant

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas