͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏

Forwarded this email? Subscribe here for more

🌐What Is OpenAI's New Web-Browsing Agent?

Plus: Is This the Ultimate Test for AI Intelligence?

Yaro Celis

Jan 24

READ IN APP

Get in Front of 50k Tech Leaders: Grow With Us - Let Us Feature Your Business.

Happy Friday, Team!

Exciting updates this week: OpenAI has launched new browsing agents, and Gemini 2025 is introducing customizable agents for our use. Anthropic has also rolled out new releases, and we’re diving into the latest Ultimate AI Test—a fascinating step toward understanding if AI is approaching AGI. Plus, we’re sharing the latest news and trends. Drop your thoughts in the comments sections! Thanks, and have a great weekend!

OpenAI’s New Web-Browsing Agent.
🧰 AI Tools
Is This the Ultimate Test for AI Intelligence?
What Is OpenAI's New Web-Browsing Agent?
📰 News and Trends.
Is Gemini 2025 a Response to OpenAI’s Operator?

Give a gift subscription

Is This the Ultimate Test for AI Intelligence?

Researchers Struggle to Create AI-Proof Tests as Models Like OpenAI and Google Gemini Advance.

The Center for AI Safety and Scale AI have developed Humanity’s Last Exam, a new test designed to measure AI intelligence and potentially serve as a litmus test for reaching artificial general intelligence (AGI). This rigorous exam includes 3,000 multiple-choice and short-answer questions spanning fields like analytic philosophy and rocket engineering, submitted and refined by top experts. AI systems' performance on these questions is viewed as a bellwether for AGI.

Testing AI has evolved from SAT-style exams to Ph.D.-level challenges, but even the best models from OpenAI, Google, and Anthropic failed this latest test. OpenAI’s o1 model scored the highest at just 8.3%. Researchers expect rapid improvement, with scores potentially exceeding 50% by year-end. While AI continues to excel in specialized tasks, its struggles with basic reasoning highlight the importance of human adaptability and common sense in a rapidly advancing landscape.

Our Partners:

RemoteBase - Hire the top 1% of tech talent in the world. We have a thorough multi-step vetting process, that gives you a strong internal community of Front-End, Back-End, Full-Stack, and Dev-Ops engineers. Our clients are matched with the required engineering teams within 24 hours.
OpenVC University - Jumpstart your AI career in 100 days with OpenCV University’s AI Challenge – master AI, Computer Vision, and Deep Learning with a 35% discount!
FRENZ Brainband by Earable - Silence your thoughts and drift into a deeper sleep. —AI-powered, scientifically-backed sleep tech transforming restless nights into restful slumber. Exclusive savings with code 'celis'!*

Let's Become Partners.

What Is OpenAI's New Web-Browsing Agent?

Operator is OpenAI’s latest innovation, designed to perform web-based tasks for users through its own browser. Currently available as a research preview for Pro users in the U.S. with plans to roll it out to Plus, Team, and Enterprise users, bringing seamless task execution to ChatGPT. Operator leverages advanced AI capabilities to handle repetitive browser tasks, saving time and enhancing digital workflows.

What is Operator?

AI Agent for the Web: Operator uses a new model called the Computer-Using Agent (CUA), which combines GPT-4’s vision and reasoning to interact with browser GUIs. It can type, click, scroll, and self-correct when facing challenges. CUA in the API: Developers will soon be able to build their own agents using CUA.
Capabilities: Tasks include filling forms, booking travel, ordering groceries, and even creating custom workflows like restocking items or booking reservations.
Applications: Operator collaborates with platforms like Instacart, Uber, and DoorDash to streamline user and business workflows. It’s also being piloted in public-sector projects like making city services more accessible.

Safety and Privacy Features

User Control: Operator requests input for sensitive actions like logging in or finalizing payments.
Task Safeguards: It avoids high-stakes actions, requires confirmation for major tasks, and closely supervises actions on sensitive websites.
Data Privacy: Users can opt out of data sharing for training, delete browsing data, and monitor Operator’s behavior to prevent misuse.
Adversarial Protections: Operator detects malicious attempts like hidden prompts or phishing and can pause tasks if needed.

Make sure you don’t provide sensitive information or data to your operators. Also keep in mind that besides the risks of exposing your sensitive data, OpenAI and tech companies will keep and use your data to further their models.

Current Limitations

Early Development: Operator may struggle with complex tasks (e.g., creating slideshows) and is still learning.

The goal of this it transforms AI from a passive tool to an active assistant, simplifying everyday tasks and enabling businesses to deliver innovative customer experiences, tread carefully.

📰 AI News and Trends

ByteDance released Doubao-1.5-pro, an upgrade to its flagship AI model, which it claims outperforms OpenAI’s o1 in AIME (FC)
The X Games Team With Google On AI Judging System For Ski And Snowboard Events (Deadline)
Check Anthropic's Citations, enabling Claude to reference source documents directly, improving trust and accuracy. available for Claude 3.5 via API and Google Cloud (Anthropic)
Musk and Altman Clash Over $500B Stargate AI Project Backed by OpenAI, SoftBank, and Oracle (AP)
Four Hours. the amount of time that ChatGPT was offline on thursday morning leaving thousands without access and leaving coders and millions of others without their primary assistant (NYT)

🌐 Other Tech news

Eggs are hard to find and expensive you do find them, blame the bird flu. With the news of the country’s first human death due to bird flu, vaccination isn’t yet recommended (Wired)
Why You Might Suddenly Be Following Trump on Instagram and Facebook and why you cant easily unfollow them (NYT)
Popular podcasters played a key role in Donald Trump’s 2024 re-election by mobilizing young male voters through discussions on voting, immigration, war, and cultural issues, often embedding conservative messages in entertainment. An analysis of 2,000+ videos found these podcasts, with predominantly male audiences, amplified grievances against progressive policies, drawing millions of views and reshaping political influence in the digital era (Bloomberg)
Dogecoin ETF Registered as Elon Musk’s D.O.G.E Gains Traction (Coindesk)

Is Gemini 2025 a Response to OpenAI’s Operator?

Google's Gemini app is set for major updates in 2025, with highlights including a redesigned overlay that reduces its footprint to a bar, Gemini 2.0 with improved performance (optimized for tasks like coding and reasoning), and Deep Research arriving on mobile.

Gemini Live will integrate Project Astra, enabling live camera and screen-sharing capabilities, and Extensions for tasks like managing messages and calendars. An immersive trip planner leveraging Gmail, Maps, and Search data for custom itineraries is also in the pipeline. Additional features include PDF querying, smart home controls, and Wear OS support, replacing the current Assistant app for seamless interactions on smartwatches.

🧰 AI Tools

Download our list of 750+ Tools free here.

🚀 Showcase Your Innovation in the Premier Tech and AI Newsletter (link)

As a vanguard in the realm of technology and artificial intelligence, we pride ourselves in delivering cutting-edge insights, AI tools, and in-depth coverage of emerging technologies to over 55,000+ tech CEOs, managers, programmers, entrepreneurs, and enthusiasts. Our readers represent the brightest minds from industry giants such as Tesla, OpenAI, Samsung, IBM, NVIDIA, and countless others.

Reserve a Spot

Explore sponsorship possibilities and elevate your brand's presence in the world of tech and AI. Learn more about partnering with us.