OpenAI rolled out an update to its ChatGPT mobile app this week, and the headline feature is a little camera icon sitting right there in the chat bar.
Tap it, point your phone at something, and the bot will describe, translate, or explain whatever's in frame.
It's also the exact kind of feature that makes you wonder who's driving the product roadmap over there.
The rollout hit iOS and Android simultaneously, which is genuinely impressive for a company that once took three months to ship a dark mode toggle.
You can now point your camera at a restaurant menu in a language you don't speak, a weird rash you should absolutely be showing a doctor instead, or your friend's questionable new haircut.
Early users report the vision model is fast and mostly accurate, and the translation use case alone might justify the update for travelers.
If you're standing in a Tokyo subway station squinting at a sign, this beats typing a desperate query into Google Translate and hoping for the best.
But here's the thing nobody at OpenAI seems willing to say out loud.
Every app on your phone now wants camera access, and every one of them swears it's only looking when you tell it to.
ChatGPT's new button is opt-in per use, which is the right call, but the permissions creep is real.
You're one accidental tap away from livestreaming your bathroom counter to a server farm.
The camera feature is free for now, which in OpenAI's language usually means "free until we figure out how to charge you for it." Voice mode followed the same arc.
Expect a tier somewhere between "casual user" and "please stop giving us money" before the year's out.
The update also quietly shuffled a few other things around.
The model picker moved, the history sidebar got a facelift, and the app is now noticeably faster on older phones, which is the most genuinely useful change in the whole release and of course got buried under the camera hype.
It's free, it's fast, and the camera thing is fun for about a week.
Just don't kid yourself that this is some revolutionary leap.
It's a feature that Google Lens has done for years, wrapped in a shinier interface and a chatbot that's slightly less likely to call your houseplant a "leafy green vegetable-like object." Our take: the camera button is a solution in search of a problem for most users, but it's a legitimately great tool for travelers, students, and anyone who's ever needed to identify a mystery object at 2 a.m.
Final Thoughts
Download it, play with it, then go back to using ChatGPT for the one thing you actually opened it for, which is asking whether a hot dog is a sandwich.