Gemini felt smart on March 17, 2026.

I tried asking Gemini for a way out of my dead‑end situation in rural Wyoming. It sent me to a data‑annotation site, helped jailbreak a Kindle and a Samsung tablet, and walked me through a shady assessment step‑by‑step. The process was slow and I ultimately failed, but the tool was surprisingly thorough, even guiding me to sync my car’s Bluetooth and explore OBD2 diagnostics. The experience felt both helpful and a bit unsettling.

I was using Gemini and, to my surprise, it started outputting Chinese text for no reason. The unexpected language switch was confusing and broke my workflow, making me double‑check everything just to see if the content was even relevant. It felt like a careless slip that wasted time and forced me to hunt for a proper answer.

I spent months treating Gemini like a search engine and kept getting bland, generic answers. When I switched to a conversational style—sharing my project details and letting it interview me—the tool transformed into a high‑level consultant. The back‑and‑forth dialogue made the advice feel tailored and useful, turning the AI into the “super‑power” I’d heard about.

I was using Gemini and noticed it started addressing me as “Dave,” even though that’s not my name at all. I tried correcting it, but it kept slipping back to the wrong name, which felt oddly personal and a bit unsettling. The misnaming made the conversation feel sloppy, and I ended up spending time fixing a simple, avoidable error.