Gemini felt dumb on May 21, 2026.

I’m honestly going crazy trying to get Gemini 3.5 to follow a simple image prompt! I just want a three-quarter view from 20 feet away, but it refuses to budge on the camera distance. I’ve wasted seven generations on this. It keeps apologizing and promising to fix the angle, but then it just repeats the same mistake or does something entirely different. It is so incredibly frustrating to be ignored!

I’m honestly so conflicted about NB2. While the 4K enhancement is actually remarkable—it transforms my low-res images into something stunning—the moderation system is just absurdly inconsistent. It’s incredibly frustrating to have a bikini photo pass while a slight décolleté gets flagged. The whole experience feels like a total gamble because of those erratic safety filters.

I was just messing around with Omni for a bit of fun, and while I did manage to put something together, the experience felt a bit underwhelming. I really wanted to nail the facial details, but the constant restrictions made it nearly impossible. It was honestly quite frustrating to have my creative vision stifled by the tool's own guardrails, leaving the final result feeling incomplete.

I've been testing a two-step JSON approach for image and video edits, and the difference is honestly so noticeable. Direct prompts used to be a mess, often changing things I didn't want touched, but this structured method keeps everything contained. I tried Gemini for style-matching and subject swapping, and it handled it all beautifully! It’s such a clean, reliable way to get precise results.

I honestly can't believe I built a working operating system from scratch just by vibe coding with Gemini! It was such a wild, surreal experience; the AI's ability to handle such massive, complex logic was absolutely mind-blowing. I felt like I was co-creating something legendary. The tool was remarkably helpful, turning what seemed impossible into a functional reality.

I'm feeling so let down and honestly quite annoyed. I paid for the subscription specifically to use Gemini Omni for AI UGC content with motion control, but it keeps refusing my requests for no clear reason! It’s such a massive disappointment to spend money only to hit a brick wall like this. I really need to know if there’s any workaround, because right now, it's just useless.

I can't work with these new prompt or computation limits they've created! I can't get anything done and the "free" flash lite model available after i've reached the limit has ABSOLUTELY no context window and forget about its own statement one response ago! useless to say the leas

I'm so frustrated with Omni right now. I tried to just change the style of a video, but it keeps hitting me with these annoying 'can't generate' errors. There's absolutely nothing rule-breaking about my prompt or the input, so I don't get why it's blocking me. I feel like I'm just wasting my usage credits trying to reword things just to appease a tool that doesn't even understand me.

Ever since the UI update earlier this week I am reaching my limit on Pro 3.1 within 4 prompts. Also there is only Flash and Pro 3.1 available to which you can change the thinking level, but flash is dumb so I have to use Pro 3.1 but it get's maxed within 15 minutes now!?