What's New in AI?

June 20, 2025

The Ferrari in the Slow Lane: Why You’re Only Using 10% of ChatGPT’s Power

If you’re using ChatGPT—like millions of others—you’re already getting a lot. Even the free version is powerful.

But if you’re paying for ChatGPT Plus ($20/month) and still using it like a chatbot—one prompt at a time—you’re basically driving a Ferrari in the slow lane.

The real value of the paid version isn’t just faster responses or access to GPT-4o—it’s the suite of tools that turn ChatGPT from a chatbot into a productivity partner. Features like Tasks, Projects, and custom GPTs are what truly set the paid version apart. They help you organize, automate, and collaborate—yet most people aren’t using them.

(And if you don’t even know what these are, you can join me this week for a quick 45-minute tour of ChatGPT and these features. See the information at the end.)

And now, in just the last two weeks, these tools have gotten even stronger with updates that make them more flexible, more private, and easier to use across desktop, mobile, and voice.

Let’s start with one of the most powerful (and still underused) features: Projects.

(Quick refresher: Projects let you group related chats, files, and goals into one place—so you’re not starting over every time.)

Here’s what just got better:

Upload up to 20 files and ChatGPT will reference them all together
Set clear goals to guide how the model responds
Drag in past chats, or turn any conversation into a project
Use voice and mobile uploads—making Projects much more portable
Switch models mid-project depending on what you need
And importantly: Privacy is now the default—your content won’t be used to train the model unless you opt in

Even if you’ve seen (heard or read) me talk about Projects before, these updates make them much more usable for real work—and a lot more worth exploring if you haven’t yet.

Gemeni Introduces New Creation Tools: Infographics, Podcasts, Quiz and Web Page Creation, with One Click!

Google’s Deep Research feature launched in late 2024, giving users the ability to run multi-step web searches and receive organized reports. But a recent update in early June quietly unlocked something wild: after compiling your research, Gemini can now offer to turn it into an infographic, quiz, podcast, or even a web page—instantly.

I stumbled on this while experimenting for a workshop and actually did a double-take. It’s one of the clearest examples I’ve seen of AI moving from assistant to creator. I’ll be talking more about this (and what it means for your work, nonprofit or not) in my AI Jumpstart for Nonprofits workshop on June 22.

Try it! And show me what you create!