I have noticed in the past month or two, but especially in the past week or so that Gemini has been rushing everything and skipping steps and doing things I don’t ask, or editing wrong file versions.
I have also noticed the more I point out its mistakes, the more it tries to overcompensate and messes even more things up.
I have also noticed that it randomly (even in not very long chats) seems to have a memory wipe and act as if it was cut and pasted from someone else’s session. For example, we will for an entire hour be referring to a toggle button as “Live Toggle Button” then on the 15th time randomly start talking about a button in some program we haven’t even been discussing that has nothing to do with my project or workspace in any way.
Does this have to do with Google playing around with their new TurboQuant code behind the scenes and treating us all like beta testers, or what?
Something is making Gemini get worse and worse and this past week especially last few days its a major downgrade.
I used to alternate between Gemini and Claude, but with the token limits being completely broken lately it LITERALLY takes Claude’s entire token limit to ALMOST fix Geminis errors (what it catches anyway), and I almost always have to go back to Gemini and have it finish doing Claudes fixes.
It’s almost so bad to the extent that it is more efficient to not use Gemini at all, or only use it for rough draft planning, and then in the hour or two of Claude use I have (used to last more around 5 or 6 hours until recently) just get quality work done.
It’s bad enough the antigravity GUI and wrapper are so poorly coded, but at least I could deal with it (barely) and get stuff done. Now everything is on the absolute verge of literally being unusable.
I really don’t understand why Google is going so downhill. If they literally just slow down and do quality control fixes then they would own most the market, despite Claude being so much better.
Then on the other hand Claude is amazing, but their GUIs are just all ALMOST good, but all so BAD.
How do these big companies have such a hard time just doing things right? With all the funding they have I literally cannot imagine how (even after taking hundreds of variables into consideration) they fail so badly.
