@beaware.social I'll give Gemini credit for one thing: It's a whole lot more "free".

ChatGPT cuts you off after about 5 questions these days. I can pound away at Gemini, feeding it error messages to try & fix some broken BS code I don't understand for *hours*.

@RxBrad @beaware.social with the new models?

Cause that's not my experience. Every free model I've used has been the out-of-date models that don't know the difference between up and down.

However, I've never tried to code with it because historically that's a bad idea.‍️

@BeAware @beaware.social Whatever "Gemini 2.5 Flash" is....

Last time I used it was to try and setup a web scraper in Docker to turn a specific webpage into JSON, every 4 hours.

ChatGPT didn't get anywhere before I hit the limit. CoPilot ran me in circles for hours. Gemini got me squared away in a few minutes.

You need to know just enough about the code to tell the AI *specifically what part* it's doing is stupid.

@RxBrad @beaware.social ahh...the model that came out yesterday is supposed to be better at most things and requires a subscription.

Though if I knew anything about coding, I'd probably do it myself because I know about it.

I tend to use AI to find out about things I don't know about and if I don't know about something, I cannot correct errors so I need it to be accurate the first time because I won't know it's wrong.

If I have to correct the older models for the first few messages, it's doing me no good. So I pay to use the smarter ones.