Coder's Newsletter # 9

đŸ“± Google’s AI Goes Offline—Right Into Your Pocket

Hey Outliers,

Welcome back to Outlier coder’s Newsletter! 🎉 Your front-row seat to AI’s biggest moves, smarter tools, and key research updates.

This week, AI showed it can be everywhere—without needing to be everywhere. From edge apps that work offline to agents that collaborate and build, the focus is shifting from what’s possible to what’s practical.

This Week in AI

  • Google’s AI Edge Gallery puts offline models in your pocket

  • Perplexity Labs turns prompts into full apps, reports, and dashboards

  • Mistral rolls out Agents API to level up real-world AI automation

đŸ”„ What’s hot this week?đŸ”„ 

 đŸš€ Google AI Edge Gallery: Bringing Offline AI Models to Your Smartphone

Google has quietly launched Google AI Edge Gallery, an experimental app that lets users download and run a variety of AI models directly on their Android devices—no internet required. An iOS version is coming soon. This marks a significant step toward privacy-focused, on-device AI that works even when you're offline.

Key Highlights:

  • Run AI Models Locally: Download models like Google’s Gemma 3n from Hugging Face and use them offline for image generation, chat, coding, and more—directly on your phone.

  • Stronger Privacy: All AI processing happens on your device, so your data stays private and never leaves your phone.

  • All-in-One Home Screen: The app offers shortcuts for key AI tasks:

    ‱ Ask Image: Upload an image to get answers, descriptions, or object IDs.

    ‱ AI Chat: Chat with your favorite models offline.

    ‱ Prompt Lab: Use templates for tasks like summarizing, rewriting, and coding.

  • Built for Developers: The app is open source (Apache 2.0), allowing both personal and commercial use. Easily switch between models, test performance, and share feedback with Google.

  • Performance Depends on Device: Model speed and features vary based on your phone’s hardware and the size of the model. Newer devices handle larger models better.

What This Means for You:
Google AI Edge Gallery gives you access to powerful AI features—like image analysis, chat, and coding—wherever you are, even without an internet connection. That means you can ask questions, generate images, or run smart tools on the go, without sending your data to the cloud. It’s a major win for privacy and convenience.

You can experiment with different models, optimize for speed, and create experiences that work seamlessly offline. This tool puts cutting-edge generative AI right in your pocket.

âšĄïž Perplexity Labs: Turning Prompts into Projects, Dashboards, and Apps in Minutes

Perplexity AI has introduced Perplexity Labs, a new tool that lets you turn a simple prompt into a full project—reports, dashboards, spreadsheets, and even web apps—in just a few minutes. It’s a major shift from Perplexity’s Q&A roots to a platform for real productivity and creation.

Key Highlights:

  • From Idea to Execution: Go from prompt to a ready-made dashboard, map, or marketing plan—no manual setup.—handled end-to-end by AI agents.

  • Versatile Deliverables: Create reports with charts, dashboards, spreadsheets, code, and simple web apps.

  • Automated Workflow: Labs runs for up to 10 minutes, using web search, code, and image generation to deliver polished results.

  • All Assets Organized: All files and charts are organized in an “Assets” tab, while the “App” tab lets you launch dashboards and apps right from the interface.

  • Available Everywhere: Available now for Pro users on web and mobile, with desktop apps coming soon. Project Gallery offers ready-to-use templates.

  • Faster Task Completion: Combine research, coding, and visualization to finish projects in minutes, that would otherwise take days.

What This Means for You:
Perplexity Labs brings your ideas to life, letting you go from prompt to project in just a few clicks—no coding, no design background required. Whether you’re building a business dashboard, generating a research report, or prototyping a web app, Labs saves you hours by automating the heavy lifting.

This means more time for strategy and decision-making, and less spent on manual research or repetitive tasks.

 đŸ€– Mistral Agents API: A New Era for Autonomous, Action-Oriented AI

Mistral AI has released the Mistral Agents API, a powerful new framework for building AI agents that plan, act, and manage complex tasks on their own. This puts Mistral in direct competition with platforms from OpenAI, Google, and Anthropic—now with a strong focus on real-world automation.

 Key Highlights:

  • Smarter Agents: Go beyond chat—agents can take actions, remember past conversations, and work together to solve complex problems.

  • Built-in Connectors: Agents can run Python code, perform live web searches, generate images, and access documents from the cloud for advanced tasks.

  • Open Standards: Supports Model Context Protocol (MCP) for easy integration with APIs, databases, and dynamic data sources.

  • Multi-Agent Collaboration: Agents can hand off tasks to each other, enabling complex workflows and teamwork between specialized agents.

  • Enterprise-Ready: Designed for both developers and large companies to power assistants for software, finance, travel, nutrition, and more.

  • Better Performance: Web search integration leads to major accuracy boosts—SimpleQA accuracy rises from 23% to 75% for Mistral Large.

What This Means for you:
With the Mistral Agents API, AI can now do more than just chat—it can complete real tasks, find information, and help you get things done faster. You can use it to automate work, get better answers, and save time. It’s like having a smart assistant you can rely on.

This makes it easy for anyone to use powerful AI in their daily work or build new tools without extra effort.

đŸ—“ïž Save the Date: Webinar on June 7— Sign Up Today!

Find out what top devs are doing differently in 2025’s front-end ecosystem

Click âŹ‡ïž to signup

☄ Trending Bytes

Me: It’s 2025.
Google AI: Not in my multiverse. 🌀

💡 AI Model Spotlight

Model Name

Parent Company

Release Date

Key Highlights

Google AI Edge Gallery

Google

May 27, 2025

On-device AI, Offline functionality, Open-source under Apache 2.0

DeepSeek-R1-0528

DeepSeek

May 29, 2025

87.5% math accuracy, Open-source license, Vibe coding support

Perplexity Labs

Perplexity

May 29, 2025

Automated reports, Spreadsheet creation, Web app creation

FLUX.1 Kontext

Black Forest Labs

May 29, 2025

Text & image prompts, Iterative editing,Flow matching architecture

Mistral Agents API

Mistral AI

May 28, 2025

Built-in connectors, Persistent memory, Agentic orchestration

Droids

Factory AI

May 28, 2025

End-to-end SDLC, Autonomous agents-Enterprise integration

Claude Voice Mode

Anthropic

May 27, 2025

Five voice options, Real-time transcription, text-voice switch

Claude 4 Series

Anthropic

May 22, 2025

Long coding , Compact + focused

Veo 3

Google

May 20, 2025

Realistic video, Audio sync

Imagen 4

Google

May 20, 2025

Sharper images, Better text

Gemini 2.5 Pro & Flash

Google

May 20, 2025

Deep Think, Coding + logic

GitHub Copilot Coding Agent

Microsoft

May 19, 2025

Assigns issues, Drafts PRs, Requires approval

Microsoft Discovery Platform

Microsoft

May 19, 2025

R&D agents, Scientific discovery support

Microsoft 365 Copilot Tuning

Microsoft

May 19, 2025

Agent tuning , Identity via Entra,Org data

Windows AI Foundry + MCP

Microsoft

May 19, 2025

MCP support, On-device AI, Windows 11 tools

Codex

OpenAI

May 16, 2025

Autonomous coding assistant, Integrated into ChatGPT, Enhances developer productivity

AlphaEvolve

Google DeepMind

May 15, 2025

Tackles complex math, Algorithm design capabilities, Builds on AlphaFold’s legacy

SWE-1 Model Family

Windsurf

May 15, 2025

Code + planning, 3 sizes (full, lite, mini), Tool-native

CodeRabbit for VS Code

CodeRabbit

May 14, 2025

Real-time reviews, Editor-native, Free access

Psyche Decentralized AI Network

Nous Research

May 14, 2025

40B parameter model, Multi-head Latent Attention (MLA) , On-chain compute

Mistral Medium 3

Mistral AI

May 7, 2025

90% of Claude Sonnet 3.7 performance, Cost-effective deployment, Open-source availability

Figma AI Suite

Figma

May 7, 2025

Claude Sonnet 3.7 integration, Modular AI tools, Focus on creative workflows

Gemini 2.5 Pro (I/O Edition)

Google

May 6, 2025

Superior coding performance, Advanced multimodal reasoning, Early access release

IBM Agent Catalog

IBM

May 6, 2025

Wide range of AI agents, Focus on interoperability, Support for various industries

Olmo 2 1B

AI2

May 1, 2025

Efficient performance, Open-source release, Suitable for resource-constrained environment

📋 Feedback

How would you describe your experience with this edition of the newsletter?

Login or Subscribe to participate in polls.

That’s it for now—keep pushing AI forward! 🚀

You received this email because you are subscribed to Outlier.ai. The content of this email is for informational purposes only and may not be reproduced or distributed without written permission. AI research is rapidly evolving, and while we strive for accuracy, we encourage readers to verify details from official sources.
Please note that all emails exchanged with Outlier.ai may be subject to monitoring for compliance with company policies. Information contained in this email is confidential and intended solely for the recipient. No legally binding commitments are created by this email.
All trademarks used in this email are the property of their respective owners. You are receiving this email because you have authorized Outlier.ai to send you updates. For more details, visit the Outlier.ai website. Terms & Conditions apply.