• AI Fire
  • Posts
  • 👀 AIs Use Your Computer... Like YOU?!

👀 AIs Use Your Computer... Like YOU?!

Google New FREE Career Certificates

ai-fire-banner

Read time: 5 minutes

AI agents - without any instructions, supervision, or shared memory - can start forming their own social norms, just like humans? And yet, they begin to agree on things, show collective bias, and even resist change…

IN PARTNERSHIP WITH GUIDDE

Stop wasting time on repetitive explanations. Guidde’s AI creates stunning video guides in seconds—11x faster.

  • Turn boring docs into visual masterpieces

  • Save hours with AI-powered automation

  • Share or embed your guide anywhere

How it works: Click capture on the browser extension, and Guidde auto-generates step-by-step video guides with visuals, voiceover, and a call to action.

Best part? It’s 100% free.

AI INSIGHTS

Microsoft Build 2025 has dropped a huge wave of updates, but one jaw-dropping announcement is this: AI agents can now interact with apps like a human—clicking, typing, navigating interfaces. No API? No problem.

🤖 Multi-Agent Orchestration (Preview)

  • Build systems where multiple AI agents work together, each handling a part of a complex task.

  • Agents built with Microsoft 365, Azure AI, and Microsoft Fabric can now talk to each other.

🖥️ Computer Use (Frontier Program)

  • Agents can now control desktop apps and websites just like a person.

  • Automates tasks like data entry, invoice processing, or research without APIs.

  • Adapts to UI changes using reasoning and vision — almost like digital hands on your computer.

🧠 Bring Your Own Model + Fine-Tuning: Copilot Studio connects with Azure AI Foundry, letting users:

  • Import 11,000+ models like GPT-4.1, LLaMA, DeepSeek, or custom models.

  • Fine-tune them with internal company data for better responses.

🔌 Model Context Protocol (Now GA)

  • A unified standard for connecting AI to enterprise knowledge and tools.

  • Supports fast rendering, better troubleshooting, and smooth tool integration.

👩‍💻 Pro-Developer Upgrades

  • GitHub Copilot now includes an AI agent that can: Autonomously fix bugs, add comments, refactor code.

  • Visual Studio Code extension for Copilot Studio:

    • Lets devs build and edit agents inside VS Code.

    • Brings IntelliSense, file system clarity, and easier GitHub integration.

Why It Matters: Microsoft is clearly betting big on "agentic AI", the idea of AI agents using your computer like a human is a fundamental shift. And Microsoft wants to own that space before anyone else.

PRESENTED BY HUBSPOT

HubSpot offers an intuitive customer relationship management platform tailored for small businesses. Manage leads, track sales performance, and understand your customers with ease. Best of all, it’s completely free, with no limits on users or data, allowing you to store and manage up to 1,000,000 contacts.

TODAY IN AI

AI HIGHLIGHTS

🗣️ ChatGPT is a debate champ! A study found it outpersuades people in 64.4% of online discussions, especially when equipped with personal data about its opponents.

📵 Anthropic is currently on a hiring spree, posting 200+ job roles but says no AI tools allowed in job applications. Meanwhile, recruiters themselves are using AI to review resumes.

🧠 A woman spent up to 3 hours/day talking to an AI therapist, saying it got her through darkest times. Her insights on both pros and cons of chatbot counselling are quite interesting.

🤖 A student built an AI tool that finds “radical” Reddit users by scanning keywords, gives them a score, then tries to chat and de-radicalize them automatically.

🤝 Google is now an AI-integrated university to bring FREE Career Certificates and AI training for 66,000+ students across all U-M campuses. Graduates also get access to a 150+ company network hiring for relevant roles.

😨 X user Pablo Rothbart claims he's outdone Icon's powerful AI ad maker with his own n8n creation at a fraction of the price. Here’s his entire workflow.

💰 AI Daily Fundraising: PhaseV raised $50 million in Series A to improve drug trials with AI. It serves 30+ global clients and aims to cut costs and speed up the costly, slow drug development process.

AI SOURCES FROM AI FIRE

ToolDrop Ep 7 just landed 🚀

You know the drill — every week, I drop 5 powerful, underrated, and totally FREE tools to help you build faster, automate smarter, and stay ahead in the AI world. Here’s what’s cooking this week:

  • Awesome n8n Templates – 250+ ready-to-use templates for n8n.

  • Stirling PDF – The #1 app for PDF files.

  • Sherlock – Find social accounts just by username.

  • Public APIs – A massive, curated list of free APIs.

  • LLM Course – A free course to break into LLMs.

Note: These exclusive resources & reviews are available only in our AI Fire community. It’s because you guys can freely ask for support or share personal experience during testing there. Get your full breakdown here (no hidden fee)!

ai-fire-academy

NEW EMPOWERED AI TOOLS

  1. 📅 Nodeflow AI creates a month of high-converting contents in just hours.

  2. 🎙 Distro turns interesting podcasts into ready-to-publish posts in minutes.

  3. 🖥 AI Operator by BlackBox AI is your browser AI agent for everything on your screen.

  4. 🛠 Mindverse builds, manages & shares effortlessly with the largest suite of AI tools.

  5. 🎧 Ragie transcribes & retrieves any answers from spoken or visual contents.

AI QUICK HITS

  1. 🖼️ ChatGPT creates X-Ray versions of regular photographs, here’s the result.

  2. ⚠️ AI doesn't know 'no' disease images, a huge problem for medical bots.

  3. 🍎 Ex-Siri head reportedly wanted Apple to choose Gemini over ChatGPT.

  4. 📱 Just before Google I/O 2025, Google launched stand-alone NotebookLM app.

  5. ☁️ Microsoft is one of the first hyperscalers to host Grok through Azure AI Foundry.

AI CHART

ais-develop-their-own-social-norms-without-being-told

New research from the University of London discovered that AI agents can develop shared social conventions and collective behaviors through interaction alone, without any central coordination, much like human communities.

The team ran experiments where groups of AI agents played simple "naming games"—with no central controller, shared memory, or group identity.

  • Over repeated interactions, agents started converging on shared naming conventions. This happened spontaneously.

  • Surprisingly, even though individual agents began unbiased, strong collective biases emerged over time.

  • In early rounds, agents had no preference, choices were split 50/50 → But when agents succeeded with a name, they repeated it. After failure, they switched → By the third interaction, most agents had started favoring a “strong” convention.

When this minority reached a critical mass, the entire group adopted their convention. If the minority was too small, it had no impact. But once past a tipping point, change was inevitable:

  • Strong conventions (more reinforced) needed a larger group to be overturned.

  • Weaker conventions were more fragile—sometimes just 2% of agents could flip the norm.

But not all AI models behave equally, even in the same setup. Just like in human social systems, small committed groups could be used to shift norms intentionally — for better or worse.

This study is a wake-up call. As AI agents become more autonomous and interconnected, they may form their own cultures, conventions, and biases—even without direct human input.

AI JOBS

We read your emails, comments, and poll replies daily

How would you rate today’s newsletter?

Your feedback helps us create the best newsletter possible

Login or Subscribe to participate in polls.

Hit reply and say Hello – we'd love to hear from you!

Like what you're reading? Forward it to friends, and they can sign up here.

Cheers,
The AI Fire Team

Reply

or to participate.