Matthew Tyson
I input some test data and hit “Add Product.” Gemini insists on an image URL, so I put in “foo” and now it accepts the submit.
Filling in some gaps
I notice there is no product type there, so I’m not sure how I can filter for this new item, to see if it really showed up. Also, the new admin page doesn’t have a link back to the home page, so it’s a back arrow for me. Still, Gemini successfully performed a lot of the grunt work in setting this page up.
Back on the main page, I don’t see my Warg Boots listed. It’s the same list of products as before. So I ask Gemini to show me the contents of the products.tsx file, and it shows me:
export const products = [
{
"name": "Warg Boots",
"description": "Keeps the dire wolf's paws warm and clean.",
"price": 1000,
"imageUrl": "foo",
"id": "1746565585872"
}
];
Which is troubling. It’s cool that Gemini has access to the contents. But this entry should have been added to the list, not replaced it. I return to the main page and hit reload:
Matthew Tyson
Which is an error. It looks like we broke the app. The POST handler appears to have rewritten the product list, instead of updating it, and somewhere something more basic has gone wrong to where the page won’t load.
Things have gone awry, but that’s normal. It is still software development, after all.
Offload the grunt work
Obviously, we are a long way off from a non-programmer frolicking around in Firebase Studio, or any similar AI-powered development environment, and building complex applications. Google Cloud Platform, Gemini, and Firebase Studio are best-in-class tools. These kinds of limits apply to all agentic AI development systems.
Still, I would in no wise want to give up my Gemini assistant when coding. It takes a huge amount of busy work off my shoulders and brings much more possibility into scope by letting me focus on the larger picture.
I wonder how the path will look, how long it will take for Firebase Studio and similar tools to mature. It seems clear that something along these lines, where the AI is framed in a tool that lets it take action, is part of the future. It may take longer than AI enthusiasts predict. It may never really, fully come to fruition in the way we envision.
I mean, really, do you want an AI agent to be able to go into your cloud dashboard and provision managed data stores and users and auth? Talk about spiraling cloud costs.
