ChatGPT Adds Apps, Gemini 2.5 AI Agents, Dia Browser Open Access


In this newsletter:

  • OpenAI launches apps inside ChatGPT, confirms hardware plans
  • Google releases Gemini 2.5 “computer use” model for AI agents
  • Browser Company's Dia AI browser, now publicly available on Mac

Plus, you’ll find new AI tools and this week’s top AI news headlines!

In partnership with Guidde:

Create How-To Video Guides in Minutes With AI (Quick & Easy)

Sharing FAQs with customers or SOPs with co-workers? Use AI to create a shareable guide and save time.

Guidde is an AI that records you doing a task and automatically creates a professional video with instructions.

  • Screen-record your task (or upload a PDF)
  • Choose from 100+ voices and 35 languages for captions
  • Upload logos, colors, fonts, etc. for branding

In a few minutes, you’ll have a video guide you can share everywhere. You can even track performance to optimize engagement.

👉 Try Guidde today (it’s free to use)

📱 OpenAI Launches Apps inside ChatGPT, Confirms Hardware Plans

OpenAI unveiled its Apps SDK at Dev Day 2025, transforming ChatGPT into an interactive app platform where developers can build applications that live inside conversations, while CEO Sam Altman and designer Jony Ive confirmed they're creating AI hardware devices.

Apps in ChatGPT:

  • New Apps SDK lets developers build interactive apps inside ChatGPT
  • Apps respond to natural language and appear in conversations
  • Launch partners include Spotify, Zillow, Canva, Coursera, and Expedia
  • Users can call apps by name, or ChatGPT suggests them contextually
  • Apps display interactive UI elements like maps and playlists
  • Built on the Model Context Protocol (MCP) standard
  • Reaches 800 million ChatGPT users globally
  • Available to all logged-in users outside EEA, Switzerland, UK
  • Monetization and app submission process coming later this year

Agent Kit for developers:

  • Visual canvas to design autonomous AI workflows
  • Chat Kit embeddable interface for deploying agents anywhere
  • Evaluation suite to measure and improve agent performance
  • Demo showed procurement agent completing multi-week tasks in minutes
  • Agents can parse requests, check policies, and execute actions

Codex availability:

  • GPT-5-based coding model now generally available
  • Usage jumped 10x since August, serving 40 trillion tokens in 3 weeks
  • 70% more pull requests merged weekly at OpenAI
  • Integrates with Slack for delegating tasks via @Codex tags

Why it matters:

While OpenAI previously attempted app ecosystems with GPTs and the GPT Store, this Apps SDK represents its most aggressive push to transform ChatGPT from a chatbot into an operating system.

By embedding interactive apps directly into conversations and giving developers access to 800 million users, OpenAI is competing with traditional app stores while building a walled garden that keeps users inside its ecosystem, a strategy borrowed from WeChat's super-app model that could redefine how people interact with software.

👀 Read about OpenAI's app platform launch!

💻 Google Releases Gemini 2.5 “Computer Use” Model for AI Agents

Google launched Gemini 2.5 Computer Use in public preview, enabling developers to build AI agents that autonomously navigate websites by clicking, typing, and scrolling like humans, with built-in safety features to prevent misuse.

What's new:

  • Gemini 2.5 Computer Use model available via Gemini API
  • Built on Gemini 2.5 Pro's visual understanding and reasoning
  • Agents can click, type, scroll, and manipulate UI elements
  • Operates in continuous loop with screenshots and action history
  • Powers Project Mariner and AI Mode agentic capabilities
  • Available on Google AI Studio and Vertex AI
  • Browserbase demo environment for testing

How it works:

  • User sends requests with screenshots and action history
  • Model analyzes inputs and generates UI action function call
  • Client-side code executes the action via Playwright or Browserbase
  • New screenshot sent back to the model to restart the loop
  • Process continues until task completion

Why it matters:

While Anthropic and OpenAI already offer computer use capabilities with full OS access, Google's browser-only approach targets enterprises seeking reliable web automation without system-wide risks.

The model's strong benchmark performance and built-in safety controls position it well for automating repetitive digital tasks, such as form filling and UI testing. However, its limited scope keeps it behind competitors offering broader computer control.

👀 Read about Google's browser-controlling AI!

🌐 Browser Company's Dia AI Browser, Now Publicly Available on Mac

The Browser Company ended Dia's waitlist, making its AI-first browser publicly available on Mac following a $610 million acquisition by Atlassian last month.

What's new:

  • Dia now available to all Mac users without an invite
  • Requires macOS 14+ with M1 chips or later
  • Free plan with limited AI chat usage available
  • Pro plan costs $20/month with unlimited chat
  • Built-in AI assistant handles tab summaries and web research
  • Skills feature includes fact-checking, outlining & productivity planning
  • Student tools turn notes and lectures into flashcards and quizzes
  • Memory feature learns user preferences over time
  • October updates include redesigned Skills and Arc's Focus Mode

Why it matters:

While Arc browser never achieved mass adoption due to its steep learning curve, Dia takes a Chrome-like approach with AI at the core, targeting mainstream users rather than power users.

Following Atlassian's $610 million acquisition, the company is resuming weekly updates and bringing Arc features to Dia, positioning it against Opera Neon, Google Chrome with AI, and Perplexity's Comet in the increasingly crowded AI browser space.

👀 Read about Dia's public launch!

My Latest LinkedIn & X/Twitter Posts:

  • How brands can track & grow their visibility across AI platforms (view post)
  • 15 steps beginner to expert roadmap to learn AI (view post)
  • How to analyze ad performance in seconds (view post)
  • 3 levels of basic to expert prompting tips (view post)
  • How to connect AI agents to internal data safely (view post)
  • How to use ChatGPT’s top features (view post)
  • How to build AI agents in n8n with one prompt (view post)
  • How to control GPT-5’s thinking for sharper outputs (view post)
  • Top 14 free AI learning resources with direct links (view post)
  • How to use GPT-5’s reasoning modes for better results (view post)
  • How to make high-quality ads with Sora 2 for free (view post)

In partnership with Kit:

Monetize Your Audience With My Favorite All-in-One Tool 💡

As a creator, your time should be spent doing what you love, not juggling a dozen tools just to run your business.

Kit gives you everything you need in one place:

✅ Build and grow your email list (I use Kit for my newsletter)
✅ Easily monetize with paid newsletters and digital products
✅ Automate your emails with triggers and custom workflows
✅ Track what’s working and optimize with powerful insights

It’s not just email. It’s your entire creator business, simplified.

Join a thriving community of successful creators.

Use Kit for Free — and Start Building Smarter

New AI Tools to Boost Your Productivity:

  • HeyNews: Automates newsletter content creation.
  • AIHumanizer: Turns AI‑generated text into human‑like writing.
  • Notis: Turns voice/text into notes, tasks, and content in Notion.
  • AskAnyModel: Chat with AI (questions, writing, coding) via uploads.
  • Geneo: Optimizes your content to rank in AI search engines.
  • Extruct AI: Automates company research and builds lead lists.
  • Prue AI: Automates resume, cover letter, and LinkedIn optimization.
  • X‑Pilot: Converts ideas or docs into full explainer videos.
  • ViralGenie: Automates viral content research, creation & posting.
  • VoiceType: Transcribes and formats your speech into text.
  • SotaVideo: Generates videos from text/images using top AI models.
  • FramEditsAI: Edits or generates images from text prompts.
  • Meet‑Ting: Schedules meetings from email threads.
  • Mujo AI: Automates e‑commerce product image and copywriting.
  • Grist: Generates and edits spreadsheets from text.

This Week's Top AI News Headlines:

  • Figma integrates Google’s Gemini AI to Suggest Layouts, Color Palettes, and Text on Demand Within its Design Platform (View Article)
  • Google Ramps Up its Workplace AI Push With Launch of Gemini Enterprise, Enabling Teams to Build and Run Custom AI Agents on Company Data in One Unified Platform (View Article)
  • OpenAI’s Sora App Hits 1 Million Downloads in 18 Days, Fueled by Buzz Around Apple Intelligence and New AI Access Tools (View Article)
  • Google and AWS Launch Competing Enterprise AI Platforms on the Same Day to Challenge Microsoft’s Azure Dominance (View Article)
  • DC Comics Co-Publisher Jim Lee Bans Use of Generative AI on DC Properties, Prompting Industry Debate Over Creative Control and Copyright (View Article)
  • ChatGPT Go, A More Affordable Tier, Launches in 18 New Countries to Expand OpenAI’s Global Reach (View Article)
  • Microsoft Overhauls OneDrive on Windows With a New Gallery View, AI Photo Agent for Smart Searches, On-Device Photo Editing, and Easier File Sharing via “Hero Links” (View Article)
  • Google Expands “AI Mode” to Over 35 New Languages and 40+ More Countries, Bringing it to 200+ Territories Worldwide (View Article)
  • 1Password Adds Secure AI Autofill Agents That Only Work After Device Verification, Blocking Phishing Scripts and Keeping Sensitive Data Locked by Default (View Article)
  • Google Brings AI-Powered Virtual Try-On to Shoes, Just Upload a Full-Length Photo, and the System Swaps Your Current Footwear With Your Desired Pair (View Article)
  • SoftBank Doubles Down On AI Robots, Pledging $4 Billion Over Five Years to Build Mobile Humanoids and Expand Robotics Ecosystem (View Article)
  • OpenAI’s DevDay Highlights Sora App Launch, Sam Altman Talks Copyright Guardrails and AI Safeguards Amid Model Upgrades (View Article)

Work With Me:

If you enjoy this newsletter, please forward it to your friends and colleagues.

Follow me on LinkedIn and X/Twitter to see my latest posts.

Have a wonderful week!

Andrew Bolis

AI Tips for Business Growth - Newsletter by Andrew Bolis

Learn how to use AI to build, market and grow your business. Subscribe to get AI tips, tools and prompts in your inbox to power your marketing, sales and business.

Read more from AI Tips for Business Growth - Newsletter by Andrew Bolis

In this newsletter: Google releases Veo 3.1 with better audio & editing Slack upgrades Slackbot into a full AI assistant Google Gemini can now schedule meetings in Gmail Plus, you’ll find new AI tools and this week’s top AI news headlines! In partnership with Guidde: Turn any PDF or Screen Recording into a How-To Video Guide Guidde is a no-cost AI tool that creates step-by-step, narrated video guides from screen recordings or PDFs. The new Desktop Capture tool allows you to record beyond your...

In this newsletter: Earn $573/day with a simple AI workflow for digital products (full guide) ChatGPT prompts to find a job, learn faster, and 2X productivity How to get in-depth responses using ChatGPT’s extended thinking mode (tutorial) New AI tools to explore and the latest AI news headlines In partnership with Outskill & Growth School: MASTER AI IN 16 HOURS AND BECOME IRREPLACEABLE BEFORE 2025 ENDS! Usually $395, but as a part of their Halloween sale 🎃, you can get in for completely FREE!...

In this newsletter: OpenAI's Sora app lets anyone generate deepfake videos Google launches Gemini-powered AI assistant for home devices Microsoft 365 Premium bundles GPT-5 with Office at the same price Perplexity's Comet AI browser is now available for free to everyone Plus, you’ll find new AI tools and this week’s top AI news headlines! In partnership with Guidde: Create How-To Video Guides in Minutes With AI (Quick & Easy) Sharing information at scale with customers or remote teams? With...