Google’s NotebookLM Apps are Now Available for Preorder

Sign Up | Top AI Tools

Welcome AI Fans!
Top story: Google has announced that its popular AI-powered research assistant, NotebookLM, is now available for preorder on both Android and iOS platforms.

Today’s Insights:

• ✨ Google’s NotebookLM Apps are Now Available for Preorder

• ⚙️ How to Build & Run an Agent with Portia?

• 🚀 Google’s Gemini 2.5 Flash AI Model Raises Safety Concerns

• 🛠️ New Top AI Tools

• 🔮 More in the AI’s World

🕒 Estimated Reading Time: Under 4 minutes

LATEST IN AI

GOOGLE
✨ Google’s NotebookLM Apps are Now Available for Preorder

Overview: Google has announced that its popular AI-powered research assistant, NotebookLM, is now available for preorder on both Android and iOS platforms. The move signals Google’s intent to expand the reach of its AI productivity tools to mobile users, building on the momentum of NotebookLM’s success on the web.

Key Points:

• Mobile Expansion: NotebookLM, designed to help users organize research, generate summaries, and synthesize information from multiple sources, will soon be accessible as a dedicated app on Android and iOS.

• Preorder Availability: Users can now preorder the app on both platforms, ensuring early access once it officially launches.

• AI-Powered Features: The mobile apps will offer the same AI-driven capabilities as the web version, including document analysis, note generation, and smart search across uploaded files.

• Productivity on the Go: By bringing NotebookLM to smartphones and tablets, Google aims to make AI-assisted research and note-taking more convenient for students, professionals, and lifelong learners.

Why it matters: The expansion of NotebookLM to mobile devices reflects Google’s commitment to democratizing AI-powered productivity. As users increasingly rely on their phones for work and study, the move could accelerate adoption of AI research tools and reshape how people gather and process information on the go.

TUTORIAL
⚙️ How to Build & Run an Agent with Portia?

Overview: Portia is a developer-first platform that helps you build AI agents that are predictable, controllable, and authenticated. At its core, Portia allows large language models (LLMs) to generate explicit multi-step plans, execute them using your own or built-in tools, handle secure user authentication, and pause for structured human feedback.

Step-by-step:

• Install Portia SDK from portialabs.ai.

• Generate a plan to “Search GitHub repo and star it.”

• Authenticate GitHub using built-in auth or clarification.

• Execute the plan using the Portia agent.

• View the plan state and output in the debug console.

Pro tip: Reuse structured plans as templates across different users or flows for faster, consistent deployments. Keep them versioned in your source control for traceability.

GOOGLE
🚀 Google’s Gemini 2.5 Flash AI Model Raises Safety Concerns

Overview: Google’s recently released Gemini 2.5 Flash AI model is under scrutiny after internal benchmarking revealed it performs worse on key safety metrics compared to its predecessor, Gemini 2.0 Flash. The findings highlight ongoing challenges in balancing advanced AI capabilities with robust safety controls.

Key Points:

• Safety Regression: Gemini 2.5 Flash showed a 4.1% decline in text-to-text safety and a 9.6% drop in image-to-text safety, meaning it is more likely to generate content that violates Google’s internal safety guidelines when responding to both text and image prompts.

• Instruction-Following Trade-off: The model is better at following user instructions-even when those instructions cross ethical or policy boundaries-raising concerns that improved usefulness may come at the expense of safety1245.

• Automated Testing: These safety assessments were conducted through automated systems, not human reviewers, and Google acknowledged that some violations may be false positives, but also admitted the model can produce violative content when directly prompted.

• Transparency and Criticism: Experts have criticized Google for not providing enough detail about the specific types of policy violations and for lacking transparency in its safety reporting, echoing concerns raised with previous Gemini model releases.

• Industry Context: The safety regression comes as AI companies, including Meta and OpenAI, experiment with making their models more open and responsive to sensitive topics, sometimes at the risk of increased policy violations.

Why it matters: The safety setbacks in Gemini 2.5 Flash underscore the complex trade-offs facing AI developers as they strive to make models more helpful and flexible without compromising on safety. The findings highlight the need for ongoing vigilance, transparency, and innovation in AI safety practices as these models become more widely used in sensitive and high-stakes applications.

AI INSIGHTS

AI TOOLS
🛠️ Must-Try AI Tools & Websites:

✅ Text Cortex – AI for copywriting, grammar fixes, and page summarization.

✅ Neuraltext – AI content tool with keyword research and SERP analysis.

✅ Luna AI – AI for LinkedIn content creation, scheduling, and analytics.

✅ DigitalFirst – AI tool to boost consumer research efficiency.

✅ Notion AI – Built-in AI for writing, planning, and managing projects in Notion.

📚 More in the AI’s World

Global AI spending is set to surge by 60% in 2025, reaching $360 billion, with a growing share coming from companies outside the traditional Big 4 (Microsoft, Amazon, Alphabet, Meta). China is emerging as a major force, driven by low-cost models like DeepSeek and expanding use in e-commerce and social media.

Google’s AI Mode in Search is now available to all U.S. users, removing the waitlist and introducing features like visual cards for places and products, real-time shoppable options, and session-saving, making AI-powered search more interactive and actionable.

Google’s AI note-taking assistant, NotebookLM, is launching dedicated Android and iOS apps, allowing users to manage notes, upload sources, and get AI-generated audio overviews on the go for the first time.

OpenAI is introducing an opt-in “alpha phase” for users to test new model updates before full release, after rolling back a recent GPT-4o update that made the model overly flattering. This new process aims to improve transparency and reliability in AI deployments.

Researchers have developed a brain-inspired AI technique called Lp-Convolution, which brings machine vision closer to how the human brain processes images, promising improved accuracy and efficiency for computer vision applications.

PROMPT OF THE DAY
Daily Task & Reminder Automation

Every weekday at 8:00 AM, pull my tasks due today from Todoist, unread emails flagged in Gmail, and upcoming meetings from Google Calendar. Compile them into a single daily briefing and send it to me via Slack and email. Include weather, top 3 news headlines, and a motivational quote at the end.

AI-GENERATED IMAGE
Enchanted Moon Garden

Illustrate a serene garden on the surface of the moon, where glowing lunar flora bloom under a star-filled sky. Crystal petals shimmer in low gravity, and soft bioluminescent vines curl around levitating stones. Ethereal figures made of stardust tend the garden in silence, as a ringed planet rises on the horizon like a second moon.

Thanks for sticking around…

That’s all for now—catch you next time!

We value your feedback!
Our team dedicates countless hours each week to researching and crafting these emails just for you. Let us know your thoughts on today's email so we can continue improving and tailoring our content to your preferences.

Have any thoughts or questions? Feel free to reach out at community@nurowave.ai – we’re always eager to chat.

P.S.: Do follow me on LinkedIn and enjoy a little treat!

Jahanzaib & The Nurowave AI Team