- Horizon AI
- Posts
- 5 Hidden ChatGPT Features You're Probably Missing Out On 👀
5 Hidden ChatGPT Features You're Probably Missing Out On 👀
Anthropic’s Claude Code and Cowork Can Now Control Your Computer 🖥
Gina Acosta
March 24, 2026
In partnership with
Welcome to another edition of Horizon AI,
Using ChatGPT often feels simple: type a prompt, get an answer, move on. But that surface-level use barely scratches what it can actually do. In today's issue, we highlight 5 often overlooked but powerful features that help you learn more effectively, stay organized, and get more done.
Let’s jump into it!
Read Time: 4.5 min
Here's what's new today in the Horizon AI
Claude Code and Cowork Can Now Control Your Computer
Luma AI Launches Uni-1, Challenging Google and OpenAI on Cost and Performance
AI Tutorial: 5 Hidden ChatGPT Features Worth Trying
AI Tools to check out
AI Findings/Resources
The Latest in AI and Tech 💡
AI News
ANTHROPIC
Claude Code and Cowork Can Now Control Your Computer
Anthropic has introduced new capabilities that allow Claude to perform tasks autonomously in its Code and Cowork AI tools by using your computer for you.
Details:
Claude can now open files, browse the web, use apps, and run developer tools on its own, even when the user is away, with "no setup required," according to the company.
The feature is available as a research preview for Claude Pro and Max users and is currently limited to macOS devices.
To use it, the Claude desktop app must be running on a supported Mac and connected to the mobile app.
The system prioritizes integrations with services like Slack and Google Workspace, but will still execute tasks if a connector isn’t available by directly controlling the browser, mouse, keyboard, and screen when needed.
Claude will always request explicit user permission before taking actions such as clicking, scrolling, or navigating through the system.
The company admits the new feature won't always work perfectly: "Complex tasks sometimes need a second try, and working through your screen is slower than using a direct integration," but they are choosing to share it "early" so they can "learn where it works and where it falls short."
TOGETHER WITH OUTSKILL
Build real AI automations this weekend. Live. For free.
Outskill's 2-day AI Mastermind has trained 10M+ learners across 160+ countries. This Saturday and Sunday, the live workshop returns.
Most AI courses talk. This one builds. Over two days, you'll create personalized AI agents, set up workflow automations, and learn to turn AI into a repeatable edge for your work.
You'll also get free bonus materials: a Prompt Bible, an AI monetization roadmap, and a toolkit builder.
The program has 1,570 reviews on Trustpilot with a 5-star average. Registration is free for the next 72 hours.
📅 Sat & Sun, 10 AM–7 PM EST
LUMA AI
Luma AI Launches Uni-1, Challenging Google and OpenAI on Cost and Performance
Luma AI has introduced Uni-1, a new image generation model that is gaining attention for outperforming top competitors on key benchmarks while offering lower costs.
Details:
Uni-1 surpasses models like Google’s Nano Banana 2 and OpenAI’s GPT Image 1.5 on reasoning based benchmarks, while nearly matching top tier systems on object detection and visual understanding tasks.
The model also ranks highly in human preference tests, leading in overall quality, style, editing, and reference based generation.
Unlike most existing models that rely on diffusion, Uni-1 uses an autoregressive approach, generating images step by step in a single system that combines understanding and creation instead of separating the two processes.
This unified architecture allows the model to reason through prompts, plan compositions, and maintain consistency across edits, reducing the need for manual adjustments in professional workflows.
Luma is also competing on pricing, with high resolution image generation costing roughly 10 to 30 percent less than comparable offerings, making it attractive for enterprise use cases.
You can try Uni-1 for free at Luma Labs. The model also powers Luma Agents, the company's agentic creative platform designed to handle end-to-end creative work across text, image, video, and audio.
AI Tutorial
5 Hidden ChatGPT Features Worth Trying
With QuizGPT, you can turn any topic or study material into flashcards or quizzes instantly.
To use it, select the + sign > More > Quizzes.
Then, upload your study material or simply pick a topic and ask it to create a quiz. This AI tool can boost your learning through the well-known effect of Retrieval Practice.
Translation tools usually get the words right, but the tone wrong.
ChatGPT Translate is different because you can control how something sounds. You can ask for the same message to feel more casual, more professional, or more natural depending on the situation.
That makes a big difference when you’re texting, emailing, or communicating in another language and don’t want it to feel off.
Projects
Projects give you a dedicated space for anything ongoing. Your drafts, notes, and reference material stay in one place, so you can pick up where you left off instead of starting over. It's especially useful for long-term work like writing projects, or anything with a recurring structure.
To create a new Project, click “New Project” in the left-hand sidebar. From there, upload relevant files and add custom instructions that will only apply within that Project. You can also drag and drop existing chats into it, or use the menu next to any chat to move or assign it to a project.
Vision
Sometimes explaining a problem takes longer than solving it. With Vision, you can just show ChatGPT what you're looking at.
Whether it's an error message, a confusing setting, or something random you want identified, simply upload a screenshot or photo and ask your question directly. You can also enable advanced voice mode and talk to ChatGPT about what you're seeing.
When using this feature, instead of quick answers, you get depth: multiple sources, context, timelines, and different perspectives. It's the difference between "what is this?" and "what's really going on here?"
If you’re writing, comparing tools, or trying to understand a topic properly, this is the feature that saves you from opening 20 tabs.
AI Tools to check out
🔥 Pippit AI: Create powerful marketing content in a few clicks. No design skills needed.
✨ Dreamina: An AI-powered creative platform that generates high-quality images and videos from text, image, and video inputs, with a set of advanced tools including AI Agents, Canvas Mode, and more.
🧊 Imagen3D: Convert image to 3D model online.
📞 NextPhone: 24/7 AI answering service for service-based businesses
🔍 Cekura: Monitor and test voice & chat AI agents.
TOGETHER WITH MINTLIFY
Are you tracking agent views on your docs?
AI agents already outnumber human visitors to your docs — now you can track them.
AI Findings/Resources
💉 Asking AI for medical advice? There's a right and wrong way, one doctor explains
💡 The internet ruined customer service. AI could save it.
The latest in AI and Tech
The company has appointed ad industry veteran Dave Dugan, formerly of Meta, to lead its global advertising efforts. He will oversee ad strategy as the company looks to deepen relationships with major brands.
Sam Altman-backed fusion startup Helion is in early talks to supply fusion energy to OpenAI, potentially securing a share of its future power output.
The discussions come as Altman steps down as board chair, a move that could allow closer collaboration between the two companies.
Agile Robots has partnered with Google DeepMind to integrate Gemini-based AI models into its robotic systems and improve them with real-world data.
The collaboration will focus on deploying and refining robots for industrial use across sectors like manufacturing, automotive, and logistics.
The company is shifting toward building specialized models, including tools for video pacing and audio that preserves accents, while targeting marketing and enterprise use cases.
Hachette Book Group has pulled the novel Shy Girl from publication in the U.S. and halted its release in the U.K. over concerns it may have been generated using AI.
The decision followed growing speculation from readers and media, including The New York Times, that AI had been used in significant parts of the work. Author Mia Ballard denied using AI, attributing the issue to an editor, and saying she plans legal action.
That’s a wrap!
We'd love to hear your thoughts on today's email!Your feedback helps us improve our content |
Not subscribed yet? Sign up here and send it to a colleague or friend!
See you in our next edition!
Gina 👩🏻💻