Meet OpenAI's Operator: The AI Agent That Handles Web Tasks for You

In partnership with

Want to advertise? Click Here | Subscribe to our newsletter

Hey everyone,

Just a quick heads-up: OpenAI launched a crazy innovation that can put your PC on autopilot. It can now handle tasks for you.

What’s inside today’s newsletter:

• 🌐 Tech Pulse: Anthropic’s new Citations feature aims to reduce AI errors

• 📁 Unfold AI Tricks: Sharpen Your Writing with DeepL Write

• 5 New Tools on the Block

• 🔎 Spotlight: OpenAI's Operator The AI Agent That Handles Web Tasks for You

• 🤝 Who is getting funded?

• ✅ To-dos

🌐 Tech Pulse

• Anthropic’s new Citations feature aims to reduce AI errors

  • Anthropic's Citations feature allows AI models to provide detailed references to the exact sentences and passages from source documents used to generate responses.

  • Citations is available in Anthropic's API and Google's Vertex AI platform, and is particularly useful for document summarization, Q&A, and customer support applications. It is available for Claude 3.5 Sonnet and Claude 3.5 Haiku, and may incur charges depending on the length and number of source documents.

• Perplexity now has a mobile assistant on Android

We are excited to launch the Perplexity Assistant to all Android users. This marks the transition for Perplexity from an answer engine to a natively integrated assistant that can call other apps and perform basic tasks for you. Update or install Perplexity app on Play Store. pic.twitter.com/FSRDLtVWzB

— Aravind Srinivas (@AravSrinivas) January 23, 2025

• Perplexity Assistant is a new AI-powered agent that uses reasoning, search, and apps to assist with daily tasks, and it is available on Android devices. It can perform multi-app actions like hailing a ride or searching for a song.

• Perplexity Assistant is multimodal, using the phone’s camera to answer questions, and it maintains context from one action to another, which allows it to research and make reservations at restaurants, for example.

📁 Unfold AI Tricks: Sharpen Your Writing with DeepL Write

Overview:

Need help refining your writing? DeepL Write helps transform rough drafts into polished prose, offering suggestions for clarity, tone, and style.

Duration: 5-10 minutes

Skill Level: Beginner

Steps:

1. Set Up DeepL Write:

• Head over to the DeepL Write platform (write.deepl.com).

• Paste your text into the editor. This can be anything from an email draft, article, or even a creative story.

2. Analyze and Improve:

• Review the suggestions provided by DeepL Write for grammar, clarity, and word choice.

• Explore the different tone options (formal, casual, etc.) to match your audience.

3. Customize and Finalize:

• Accept or modify the AI’s suggestions to fit your personal style.

• Refine specific sentences that need more impact or clarity.

Pro Tip:
To get the best results, feed DeepL Write a moderately complete draft instead of a fully polished piece. The tool works best when there’s room for improvement!

💼 From our Partners

The gold standard of business news

Morning Brew is transforming the way working professionals consume business news.

They skip the jargon and lengthy stories, and instead serve up the news impacting your life and career with a hint of wit and humor. This way, you’ll actually enjoy reading the news—and the information sticks.

Best part? Morning Brew’s newsletter is completely free. Sign up in just 10 seconds and if you realize that you prefer long, dense, and boring business news—you can always go back to it.

Join 4.3 Million Readers Now

🔧 New Tools on the Block

1. Trae: Adaptive AI IDE that helps you ship faster

2. Shimmer 2.0: ADHD Coaching, now AI-enhanced

3. Jolt AI: AI assistant for 100k to multi-million line codebases

4. Opengrep: The open-source code security engine

5. GoCodeo: An AI coding agent extension for VSCode

Note: Want to sponsor your tool in our newsletter? CLICK HERE

🔎 Spotlight: OpenAI's Operator The AI Agent That Handles Web Tasks for You

Summary:

This OpenAI introduces a computer-using agent (CUA), a model enabling AI to interact with the digital world using a universal interface of screen, mouse, and keyboard rather than relying on specific APIs. CUA, powered by GPT-4o's vision and reinforcement learning, achieves state-of-the-art results on various benchmarks, showcasing its ability to perform multi-step tasks and adapt to challenges. However, safety is prioritized, with multiple layers of mitigation against misuse, model mistakes (including adversarial attacks), and frontier risks. The research preview, accessible through the Operator, aims to gather user feedback to refine CUA's capabilities and ensure its safe deployment. The ultimate goal is to broaden AI's accessibility and application by enabling interaction with any human-designed software.

Why it matters:

• Universal Interface: CUA uses a universal interface of screen, mouse, and keyboard, which allows it to interact with any software designed for humans. This capability means it can perform digital tasks without relying on specific operating systems or web APIs, greatly increasing its flexibility. This adaptability opens up a wide array of potential applications across different computer environments.

• Advanced Capabilities: CUA combines advanced GUI perception with structured problem-solving, enabling it to break complex tasks into multi-step plans and adapt to challenges. It also uses chain-of-thought reasoning to improve its task performance. These capabilities allow it to navigate complex digital tasks, handle errors, and adapt to unexpected changes.

• Real-world Application: CUA is available through Operator, a research preview of an agent that can perform web-based tasks for users. The real-world feedback gathered from this preview will help refine CUA’s abilities and safety measures. This iterative approach is important for making CUA reliable and useful in various real-world scenarios, moving beyond specialized APIs.

Source: OpenAI, The Verge, TechCrunch

🤝 Who is getting funded?

1. Neko, the body-scanning startup co-founded by Spotify’s Daniel Ek, snaps up $260M at a $1.8B valuation (TechCrunch)

2. Mistral AI plans IPO (TechCrunch)

 😂 Tech Memes

✅ To-dos

Do you have a topic in mind? We’re always open to suggestions—let us know what you want to learn next!

If you like today’s issue, consider subscribing to us.

That’s a wrap! Catch you in tomorrow’s edition. 👋

—Harman