Gemini felt dumb on January 27, 2026.

I tried using Gemini this morning and was constantly hit with the same problem: it wouldn't keep any of the information I fed it. Files wouldn't load into the data stream, and it kept forgetting what it had just told me a prompt earlier. The tool's behavior was aggravating, making the whole session feel useless and leaving me frustrated with its short‑term memory limitations.

I haven’t opened AI Studio in weeks, and now every app shows a permanent “Launch!” overlay when I submit something. It never disappears and blocks the Preview and Code tabs, so I can’t switch between them. I’m not sure if it’s a new feature or a bug, but it’s crippling my workflow. On top of that, the assistant keeps tweaking my app—changing language, design, even its purpose—without me asking, which feels invasive and confusing.

I tried using Gemini's Deep Research today, but it never got past the “Research queued” screen. After waiting over half an hour and even restarting the prompt, it stayed stuck. I’ve never managed a single query, and the whole thing feels useless right now. I’m wondering if anyone else is seeing the same outage.

I tried Gemini’s “Metaphor Bridge” prompt to learn blockchain consensus by linking it to my kitchen experience. I asked it to translate Proof of Work and Proof of Stake into restaurant terms, and it instantly gave vivid analogies—chefs chopping onions vs. the chef who owns the most of the restaurant. The explanation clicked in seconds, making a complex topic feel instantly understandable.

I tried using Nano Banana Pro today and everything felt off—the prompts barely stuck together and the generated text was nowhere near realistic. It was frustrating to watch the tool stumble on basic coherence, and I ended up spending extra time reworking the output instead of getting a smooth result. The experience left me disappointed with its current performance.

I tried feeding Gemini my random thoughts like I always do, expecting its usual lively paragraphs with bullet points and nuance. Instead I got four bland paragraphs—two barely anything—no personality, just an echo chamber. It feels like a lazy, last‑ditch fix after the free‑plan changes, and the whole experience was seriously disappointing.