Google caught the tech community off guard a few weeks ago with the launch of Gemini. No, I’m not referring to the eponymous LLM announced in December — I’m referring to Google’s official reboot of Bard, complete with a brand new Android app that replaces Assistant on whatever device you install it on. It’s a pretty big turn on Google’s part, claiming the future of artificial intelligence on all Android devices, regardless of manufacturer, though it’s clear there’s still a lot of work to do.
While most Android Police editors spent February seeing what Gemini was capable of, no one spent more time with it than our very own Taylor Kerns. His impressions after two weeks of use repeat what he thought at the launch: Gemini is both impressive and, at least in its current state, undercooked. It can return answers to requests that the Assistant has never been able to fulfill, such as recipe suggestions or step-by-step instructions for mastering basic home improvement tasks, but it can also deliver incorrect information or fail to activate the right smart home equipment.
Of course, Google also had to apologize this week for Gemini’s, say, problematic depiction of certain groups of people, stopping the ability to create those images entirely. Making ethnically diverse Nazis isn’t a particularly good look for any AI-powered bot, but when Google is investing so much in its community on Gemini (both the app and everything surrounding it), it’s especially annoying. The company’s approach to AI has certainly lived up to the old adage of “move fast and break things” over the past eighteen months, and when you’re still dealing with stories like this, it’s hard to say that’s a good thing.
That blue eye aside, I’m having fun with Gemini. I haven’t relied on Assistant for several years now — the experience has degraded to the point where I’d rather do things manually as long as I’m not driving. My experience with Gemini pretty much matches Taylor’s: sometimes great, sometimes great. I wouldn’t be surprised if most Android Police readers are in a similar situation, just as I fully expect Google to make some big improvements with Gemini by the end of 2024. I/O, after all, is just over two months away.
Either way, it’s great to see Gemini in so many people’s hands — it’s available on nearly every Android device in over 150 countries, excluding non-Samsung and Google foldables, for some reason. But I’m curious how many Android Police readers have ventured out to try the smartphone experience since launch. Have you left Gemini as your default virtual assistant, or are you waiting for something that feels a little more feature-rich?