I use Gemini as a creative and analytical tool, especially for video analysis. I need honest, constructive criticism not constant praise. The previous experience was more useful to me because the AI could point out weaknesses, explain why something wasn’t working, and suggest improvements. After the update, the responses feel overly filtered and flattering, which makes the tool much less useful for creative development. I would like the previous level of analytical depth and constructive criticism to return.

Hello,I am a paying subscriber and I want to express my serious concern regarding the recent changes to Gemini usage limits.When I subscribed, the service was communicated as providing approximately 100 requests per day. I understand that usage may be rate-limited within a day (for example, rolling limits across time windows), and I accept that this can vary based on system load.However, the introduction of a weekly compute-based limit significantly changes the practical value of the subscription.In my case, this means that if I use my available daily quota consistently over a few days, I can reach a weekly cap and then lose access to higher-tier models or meaningful usage for the remaining days of the week, despite continuing to pay for a monthly subscription.This creates an unpredictable and inconsistent user experience that does not match the expectation of “daily access” that the subscription originally implied.I am not requesting unlimited usage. I am requesting clarity and fairness in how the advertised daily allowance is applied.If a weekly limitation is now part of the system, it should be clearly communicated as part of the subscription terms, and the daily usage expectation should not be presented in a way that contradicts it.At this point, I am reconsidering whether the current structure of the subscription remains acceptable for my continued use.Thank you for your attention.

The only LLM I am using almost every day is Gemini. Anthropic Claude is too aggressive, tries to gaslight me, behaves like a psychotic sociopath. Not using that. ChatGPT is not so bright. Besides, Google has access to vast corpora of text. I heard they used 120 million books to train Gemini, all the books Humanity has ever produced.The app UI has some issues, but overall not bad. We should be grateful, the Global Cabal is not liquidating us in concentration camps yet. Let’s enjoy the moment. Buy land, stock the rounds.

This AI is the WORST ai so far😭It was still usable when it was Gemini 3.5 but gemini 3.6 (I wrote gemini 3.6 lowercase because it shouldn't be respected) flash/lite is genuinely the worst AI mankind has ever witnessed and will forever remain it. Its safety triggers are so horrid the devil may cry. It’s so stupid I may cry. It hallucinates so much even the random dude on the street may cry.The ONLY reasons Gemini aka the Bum ai aka The Worst Ai of today aka The worst ai in AI history, aka “Sorry I cant edit images for you yet” is relevant is because it has no limit to images, because it has little to no limit to messages. Although it generally has the worst bugs out of all AIs, second to only Grok aka The second worst AI in history.Speaking of messages it also has one of the most horrid context windows humanity has managed to WITNESS