Anyone want to help a beginner?

Quick background: I work in IT management, but most of my coding experience was as a teenager in the 80s. In the past couple of years I’ve used AI to try to get back into coding in my little spare time. I’ve still got a lot to learn. I’ve used Replit fairly successfully, but recently tried Firebase Studio because I think Google is heading in a good direction on the AI front, and I like not having to pay a monthly fee.

My first project was just for testing – a small, personal app. It is a more awkward experience than Replit, but I finally got the app fully working. Then I tried to add Google Auth. It got into a loop where I seemed to be authenticated, but then came back to the login screen. I thought being in production might help, so I tried to deploy, but deployment failed. I’ve been paying for Gemini Pro as the AI, but it failed on these last two tasks. Now my cute little app is sitting around worthless. I’d like to figure out how to fix those issues, but Gemini is no help. Any suggestions? Just wait for Gemini 3.0?

2 Likes

I’m in the same boat with one of my apps.

What can help is to ask the AI to add debug console logs at different parts of the application to see what goes wrong.

One can the share the outputs with the AI to see if it can guide him…

Yep, tried that. Had Gemini add log data. We were getting responses on the authentication, but Gemini couldn’t figure out what was wrong. It started going in a loop, changing the code from one thing to another, and then back.

I gave it the logs from the attempt to deploy. It seemed to be something related to backend authentication. It tried to fix it, but failed four or five times.

What I’m going to be trying on my project is to push it to github and then from there use cursor with GPT 5 to try and debug and fix it

I seemed to move one step forward. I ask to remove the firebase authentication - then once it did, there was still a loop in the application with a different error, which it then seemed to be able to resolve…

Yeah, that’s where I was too. I just assume that Gemini should have more training data on Firebase than GPT-5, but it’s worth a try.

Hello I am a student and I need some help

A suggestion: try a simple to-do app, in your prompt, add “please allow Google sign-in”, make sure the sign-in feature is in your app blueprint. If this works well, try to learn and compare the source code between your projects. You may be able to spot what’s causing the endless loop with the confused Gemini.

1 Like

Thanks for taking the time to respond. It’s great knowing there’s some official input on the forum. I’ll get back to the app this weekend and update here.