Hi, I’m a developer. I’ve been developing websites and extensions for years for private clients and some companies.
I completed a personal project I started years ago about 3-4 months ago using the “Gemini 3 Flash” model, even with the free version of my account. This is because I know how to develop on my own, so I made my job easier by leveraging my experience. But the speed of the model itself—the critical point I want to highlight is this…
Search “Bookmark Domain Updater” to understand more about the difference between creating it yourself and with little AI that you have offered.
Since you introduced the new “Models,” everything hasn’t worked as it used to. I’ll therefore list the various errors I’ve encountered:
1 - The Model Quotas don’t make sense; 2 or 3 requests and they’re finished.
2 - Too many redundant errors.
3 - If the Template fails, you lose your tokens, and clicking “Continue” spends the same tokens, which makes no sense, because in this way the “Model” in use is used up in about twenty minutes.
4 - If the “model” keeps crashing and I “DON’T” click Continue but Undo, the files it had previously modified will have errors or rewriting problems, forcing me to save the files before clicking Undo. And guess what? If I don’t write “Continue” but “Continuaaa” in my native language, adding a vowel or two? The “model” starts working like magic, which means there’s a reading error when I click that button (more investigation is needed).
I’ve only written these four errors, which are really important to me, and I imagine to the entire community. Please don’t reply to me with the same canned messages as if you were a bot.
The community deserves more respect.