I absolutely love this, and it is very helpful. You can ask anything starting from asking a question or general knowledge and even asking a deep question, even though it’s AI, but it actually really helps the the user to understand human feelings which even humans are not able to understand. but I absolutely hate the new update to the audio chatting option because the response given in the chat on voice is absolutely useless and not what I want to hear, but when I ask it on chat, it gives the correct and needed information so I want you to improve the voice call option it used to be good, but after the new update, I am not using it anymore

As a power user who treats AI as a critical productivity partner, my experience with Gemini has reached a breaking point. I am a results-driven professional with a perfectionist mentality, currently investing hours into deep analysis and aggressive prompt engineering to achieve a "9/10" output quality. Unfortunately, Gemini is failing the basic "logic and intent" test.Despite iterating on prompts and providing clear, specific context, the model suffers from severe contextual drift and hallucinations. In my most recent workflow, the AI inexplicably derailed, referencing a "mosque" in a thread exclusively about a tailoring business. This isn't just a minor bug; it is a massive time-wastage factor. When I’m asking about "mangoes," this AI starts talking about "coconuts."I need a tool that respects the user intent and handles complex multi-turn conversations without losing the thread. Currently, I am forced to repeat or change the prompt in every possible way just to get a basic, accurate result. For anyone willing to pay for a premium paid AI software to decrease friction, Gemini’s current state is a "mood spoiler." Compared to competitors like ChatGPT, Gemini feels like a background assistant that lacks the reasoning capabilities required for high-level work.

I have been using Gemini extensively and have been impressed with its capabilities. However, I am encountering a recurring issue where longer chat conversations fail to load. This seems to happen intermittently, without a clear pattern.When I try to revisit a detailed and lengthy chat thread, the interface will often get stuck on a loading screen or only display the most recent messages, making it impossible to access the earlier parts of the conversation. This can be particularly frustrating, as I often refer back to previous discussions and the information stored within them.The rest of the app and its features appear to work perfectly fine during these times. The problem seems to be specifically linked to the retrieval of extensive chat logs.I believe resolving this bug would significantly improve the user experience, especially for those of us who rely on Gemini for ongoing, detailed projects and conversations. I hope the development team can look into this loading issue.

The AI itself is very unsettling. Information accuracy is less than 50% and all it does is paraphrase and provide unnecessary verbosity. Provides extremely biased analysis to information and won’t mitigate from its biased interpretation. No improvements in language reasoning and misinterprets any and all accurate texts or metaphors. Keeps labelling that the user is frustrated. Which actually made me frustrated. I wasn’t frustrated till then and was just calling it out on not providing me accurate information or correcting the way it laid out the information as from what I asked but it just keeps saying i am frustrated. Then it says as an AI I don’t feel emotion. Brother I know you are an AI I don’t need you to keep reminding me in every sentence. Please make it stop assuming and labelling my emotions and saying what I am feeling and rather to provide accurate information. I am not sure but if Gemini was a person, it sure would be diagnosed with cluster B personality disorders. As far as the app UI is considered, why does it just suddenly stop? Sometimes it just keeps thinking forever and you have to manually stop it. Why not go just admit that the information is unavailable. The only reason I use it is because I have an account in Google and that helps keep my information and data linked within the interface. More like I would have stopped and started with a better AI if it wasn’t for me having my data and information in Google. Finally, why is it necessary for Google to read the conversation for improvement? Yes I understand the LLM. But can’t Google give the user a choice on what conversation remains private and inaccessible to the programmers? I’m not sure how this works, but I don’t want to put my head to another privacy concern where to use the app I have no other choice other than to accept whatever terms and conditions are laid out. I hope it improves. But very unlikely it will.