OpenAI Operator, Stargate, and R1 | Weekly Edition

PLUS HOT AI Tools & Tutorials

Jan 24, 2025

Hello and welcome to our weekly roundup!

Today, we have some huge news: a $500B AI project for the US, OpenAI’s next big move into agents, and a Chinese reasoning model that beats ChatGPT! Let's get started.

Some news is worthy of special attention, so we'll break them down next week!

This Creators’ AI Edition:

Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁

Get deeper insights with Aftercare’s AI survey 🔎

Aftercare will collect better data and save hours of analysis time compared to traditional surveys and form builders.

✨ Use intelligent AI follow-up questions to dig deeper into respondents’ answers

✨ Save hours of analysis time with Aftercare’s AI, which categorizes open-ended responses and summarizes your raw data into actionable insights

✨ Generate a survey instantly with AI, customized to your needs. Perfect for gathering product feedback, measuring satisfaction at an event, conducting market research, and more

👉 Generate a survey with AI for free!

Featured Material 🎟️

OpenAI Introduces Operator

The image shows a user request to find and book the highest-rated one-day tour of Rome on TripAdvisor. It includes a search process summary and a TripAdvisor webpage displaying a "Best Seller" tour: "Rome: Colosseum, Roman Forum and Palatine Hill."

After the global shutdown of ChatGPT, OpenAI is back with the next big step towards the agentic future. The company has unveiled “Operator”. It's a research preview of an agent that can use its own browser to perform tasks for users.

Operator can perform any task within the browser. You can ask it to order groceries, enter data into Excel, or (as the company says) generate memes. It's essentially the same ChatGPT but with access to all the tools we use daily. The agent should be pretty useful in both day-to-day life and workflows.

To access Operator, you must have a $200/mo Pro subscription and be in the US. We covered this tier in the following post:
ChatGPT Pro: Is It Worth Your $200/mo?
Stepan Ikaev
·
December 11, 2024
Read full story

Keep your mailbox updated with key knowledge & news from the AI industry

How It Works

Operator is powered by a model called “Computer-Using Agent.” It combines GPT-4o's vision capabilities with advanced reasoning through reinforcement learning. So, it can interact with graphical user interfaces—buttons, menus, and text fields on a screen.

This means Operator “sees” and “interacts” rather than working through custom API integration. If the agent encounters an error, it applies reasoning capabilities to solve it. If that doesn't work either, it returns control to the user.

You can also personalize your workflows by adding custom instructions, both site-wide and site-specific. For example, tell Operator what kind of pizza you prefer or the type of accommodations you want to book regularly on Airbnb.

Like using multiple tabs in a browser, you can have Operator perform similar tasks simultaneously. Simply give a task and create a new conversation.

When the global deployment will begin is unclear. But OpenAI will provide access to Operator for Plus, Team and Enterprise users in the near future.

The release of Operator doesn't seem to be so significant. Firstly, we have seen something similar from Anthropic, and secondly, we knew about it beforehand.
But experience and the fine print suggest that it's not that simple. The company is already working with DoorDash, Instacart, OpenTable, and others to make Operator as efficient as possible.
And this touch gives great hope for an interesting agentic future.

News Of The Week 🌍

DeepSeek Unveils Open-Source Model Rivaling OpenAI's o1

Chinese AI startup DeepSeek has released DeepSeek-R1, an open-source reasoning model that matches OpenAI's o1 in performance across math, coding, and reasoning. Developed using reinforcement learning without supervised fine-tuning, DeepSeek-R1 is available under the MIT license. The company claims that it performs similarly to OpenAI's o1 at approximately 95% less cost.

This is one of the major releases of the last month (even considering the 12 Days of OpenAI) that deserves its own attention. We'll break down all features of the DeepSeek-R1 next week. So stay tuned!

Trump Announces $500 Billion 'Stargate' AI Infrastructure Project

Immediately after taking office, the 47th President of the US announced the largest infrastructure project dedicated to AI. The Stargate Project, valued at $500B, will be led by three companies: OpenAI, SoftBank, and Oracle. SoftBank will manage the financial aspects, and OpenAI will oversee operations. The venture aims to establish 10-20 data centers across the US, starting with a facility in Abilene, Texas.

OpenAI has begun deploying the first $100B. The company says Stargate will secure American leadership in AI, create hundreds of thousands of jobs, and generate massive economic benefits for the world.

Samsung Held Galaxy Unpacked with Focus on AI

Samsung held its annual event, where it showed off its Galaxy S25 lineup. This time with a big focus on AI. The company has partnered with Google to integrate Gemini and agentic features into its flagships. Multimodal AI can now access apps like Calendar, Notes, and Clock to help with your day's schedule and tackle important to-dos.

Galaxy S25, S25+, and S25 Ultra also use AI to interact with the world around you through the camera. We've seen similar features at past Google events.

To reduce your concerns about safety, Samsung promises to analyze all data locally with a special chip and will not use it to train Gemini.

Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!

Refer a friend

Perplexity Launches the Sonar API for AI Search

Perplexity AI has launched the Sonar API, a new generative search tool available in two tiers: Sonar and Sonar Pro. The API enables developers to integrate search capabilities into their apps, offering real-time web connectivity and citations. Here's a short comparison:

Sonar Basic

Context Window: 127,000 tokens
Price: $1 per million tokens
Search Cost: $5 per 1000 searches

Sonar Pro

Context Window: 200,000 tokens
Price: $3 per million tokens
Search Cost: $15 per 1000 searches

Some companies (including Zoom) are integrating this API to enhance their AI with real-time search functionality.

OpenAI Releases o3-mini for Free ChatGPT Users

Sam Altman announced that ChatGPT's free tier will now utilize the o3-mini model. In the same post, Altman revealed that paid subscribers to ChatGPT Plus and Pro plans will enjoy “tons of o3-mini usage,” giving people an incentive to upgrade to a paid account with the company. The decision comes after the model successfully passed external safety testing.

The o3-mini offers faster response times, reduced computational requirements, and improved reasoning capabilities. In short, it's better at everything.

Useful Tools ⚒️

Jolt AI – AI assistant for 100k to multi-million line codebases

Spell by Spline – A model to generate 3D worlds

Turing List – Grow your business with AI

Small Business Chatbot - Conversational AI for your website

Tellers.AI – Quickly turn any song, voice memo or podcast into a video

Tellers.ai is an automated video editing platform that turns any text, article or story into a video. This tool was designed for creators and companies and is not bad for making your content available on YouTube, Instagram, and TikTok. Additional features include B-Roll creation, improved SEO, and seamless integration with existing video archives.