Hold on tight, because the future of AI interaction is about to get a whole lot more organized! Google's Gemini, aiming to be your ultimate AI assistant, is cooking up something big behind the scenes: a feature called "Projects." Think of it as creating dedicated, super-focused workspaces for your AI chats and file sharing. It's like giving each of your tasks its own private room, keeping everything neat and tidy.
TL;DR:
- Gemini is actively developing "Projects" within its Android app, promising a more organized AI experience.
- Initially, it seemed Projects might be web-only creations, but now it looks like you'll be able to start them directly within the app.
- Google is also tinkering with the app's menu layout, potentially making some options easier to find.
AI tools like Gemini are incredibly versatile, handling everything from brainstorming ideas to drafting emails. But sometimes, all that flexibility can feel a bit… scattered. You might want to keep certain conversations and files completely separate from others. That's where "Projects" comes in. ChatGPT already offers a similar feature, providing smart workspaces to silo your conversations and files for specific tasks. Earlier this year, eagle-eyed observers spotted signs that Gemini was working on its own version (https://www.androidauthority.com/gemini-chatgpt-projects-apk-teardown-3600434/), and now we have even more to share.
Back in September, when we first caught wind of this, all we had was a "Projects" label in the Gemini app's side menu, along with some text suggesting you'd need to create these projects on the web before using them on your phone.
But here's where it gets controversial... The latest updates in version 16.47.49.ve.arm64 of the Google app suggest a different approach. Instead of being forced to start in a browser, you'll now see a "New Project" button right inside the Gemini app on Android. This is a game changer for mobile users!
Unfortunately, we're still in the dark about many details. While we can tap the "Create project" button, it's not yet functional. It's a little frustrating, because we're eager to understand Google's full vision for this feature. And this is the part most people miss... We did, however, uncover a new line of code that sheds some light on how Projects will work:
<string name="assistant_project_source_files_header_subtitle">Add up to 10 files that Gemini can reference throughout your project.</string>
This confirms our suspicion that Projects will let you gather specific source materials that you don't want Gemini to use for other tasks. Imagine having a project dedicated to writing a novel, and being able to upload character sketches, plot outlines, and research documents that Gemini can access only within that project.
But right now, we're still left with a ton of questions. Will that 10-file limit be set in stone, or will it vary depending on your Google AI subscription? Will there be different project types optimized for different tasks? We're anxiously waiting for the chance to actually test this feature and see how it all comes together.
Beyond Projects, there are minor UI tweaks being tested, too.
For example, the NotebookLM shortcut isn't currently visible in the account switcher. However, developers seem to be considering adding it there, potentially making it easier to access your NotebookLM documents.
Staying in the same area of the interface, Gemini's scheduled actions (https://www.androidauthority.com/gemini-scheduled-actions-rollout-3565152/) are currently accessible through the account switcher, under settings. But that might change. Perhaps to make them more noticeable, Google is considering moving them to the side menu.
None of these changes are live just yet, and Google ultimately decides when (or if) they'll be released to the public.
⚠️ Remember, an APK teardown gives us a glimpse into potential future features based on unfinished code. It's always possible that these features won't make it into the final product.
Now, here's a question for you: What kind of projects would you create with Gemini? Do you think a 10-file limit is reasonable, or should it be higher? Share your thoughts in the comments below! Do you think this feature will truly change the way we interact with AI, or is it just a minor improvement?
