Gemini felt dumb on November 14, 2025.

I keep seeing the Notebook LLM spitting out three emojis on every reply again, and it’s driving me nuts. Every time I ask a question it just dumps those emojis instead of a useful answer, which feels lazy and unhelpful. I’m frustrated because I can’t get the information I need, and the repetitive behavior just wastes my time.

I tried using Gemini 3.0 on my phone, and it just crashed straight away. Moving to my PC didn’t help – it failed four times in a row, each attempt ending in the same useless error. The whole experience felt like a waste of money, leaving me frustrated and annoyed that a product marketed as cutting‑edge turned out to be nothing but broken garbage.

I tried using Grok’s voice mode and was impressed – the speech sounds smoother and more natural than Gemini’s, and I can tweak the speed (x1.25, x1.5, etc.) to fit my pace. Gemini’s voice felt slow and added filler, making it tedious to listen to. Grok’s simple speed control felt like a huge usability win, and I’m left wondering why Gemini hasn’t added it yet.

I asked Gemini a straightforward SQL question, but instead it started spouting random business ideas. It kept veering off-topic, which was both confusing and annoying. I felt the tool wasn’t listening to my request, forcing me to sift through irrelevant suggestions just to get a simple answer.

I spent the night testing the canvas feature in the Gemini mobile app and noticed the web version’s outputs feel noticeably nerfed compared to before. The phone’s canvas still seems superior, which makes me wonder if the web is really running Gemini 3. I shared links to both versions, hoping someone can confirm why the web side looks weaker while the mobile version feels more capable.

I noticed my prompts in the AI Studio webapp started spitting out emojis, something I’ve never seen before. I shared a screenshot and wondered if my overly‑flamboyant wording triggered an “emoji expert” mode in Gemini Pro 2.5 (or maybe the upcoming Pro 3). The surprise was quirky—not outright broken—but it felt odd and made me question whether the model was over‑reacting to style.

I’m really annoyed with how this release method turned out – calling it terrible wasn’t an exaggeration. I already have a grudge against Gemini 3 and haven’t even gotten a chance to try it yet, so the whole rollout feels like a slap in the face. I’m asking for live mode with gems, separate voices for each gem, and voice mode that can pull up saved info. It’s frustrating, but oddly I’m still willing to stick around and buy all your products forever if these changes happen.

I was trying to use Gemini 3.0, but it just quit working on me. It keeps reverting to the old 2.5 model for coding, which feels like a step backward. I miss the smoother experience I had before—this regression is annoying and makes me question whether I can rely on the tool for my projects.

I dove into the new iOS app, trying out the rumored Gemini 3.0 for a creative writing test. To my surprise, the output was consistently impressive – the prose flowed naturally and felt far richer than anything I’d seen since version 03‑25. As someone who leans on LLMs for stories, I was genuinely excited by how solid and inspiring the tool turned out to be.