ChatGPT now has a new "Work" mode that acts like a helpful assistant who can actually do things—it can move between your apps and files, work on projects for hours at a time, and turn your big ideas into completed work. For someone managing multiple tasks or projects, this means you could hand off tedious work like organizing files, drafting reports, or coordinating information across different programs without having to do it yourself. Worth knowing: if you use ChatGPT, look for the Work mode option in your settings and start with smaller tasks to see how well it handles the kinds of work you want to delegate.
# SensorFM: A Smarter Way to Understand Health Data from Wearables
Researchers have created an AI system called SensorFM that can understand health information from wearable devices—like smartwatches and fitness trackers—and answer questions about your patterns in everyday language. For anyone wearing these devices, this means your gadget could eventually give you genuinely useful insights about your sleep, heart rate, and activity without needing to be a data expert yourself. Worth knowing: as this technology develops, you may soon be able to ask your smartwatch simple questions like "Why is my sleep worse this week?" and get a sensible answer based on your actual data.
Elon Musk's X platform is planning to send you direct messages whenever one of your posts gets flagged with a "Community Note"—a correction added by other users to point out misleading information. The idea is to make sure people actually notice when their posts are being questioned, rather than letting corrections quietly sit beneath the original message where few might see them. Worth knowing: if you post on X, you may soon get notifications about disputed claims, which could prompt you to reconsider what you've shared or check the facts before resharing.
Google's new tool was able to spot that a viral picture of Senator Mitch McConnell in a hospital bed was artificially created by AI, not a real photo. As fake images become easier to make with artificial intelligence, having reliable ways to check whether something is genuine matters more than ever—especially during elections and breaking news. Worth knowing: if you see a shocking image online that seems too dramatic to be true, you can use reverse image search tools (like Google Images) to check where it originally came from, or look for fact-checking sites like Snopes to verify what you're seeing.
A new app called Roost deliberately slows down your messages to travel at the speed of a real carrier pigeon—taking hours or even days to deliver—so you can't get caught in the rush of constant online chat. For those of us who feel exhausted by the pressure to respond instantly to every notification, this app offers a genuinely different way to communicate that feels more thoughtful and less stressful. You can try Roost if you're curious about stepping off the constant-connection treadmill, or simply notice how much calmer your day feels when you give yourself permission to reply to messages later rather than right away.
X (formerly Twitter) is planning to notify you directly when posts you've interacted with get fact-checked by its community labeling system, so you'll know if something you engaged with turned out to be inaccurate. This matters because misinformation can spread quickly, and knowing after the fact that you've shared or liked something false helps you avoid looking unreliable to your friends and followers. Worth knowing: when you see a post with a Community Note label attached, take a moment to read what the note actually says before sharing further—it's your chance to stop spreading the same mistake others made.
Google Photos now has a new tool that uses artificial intelligence to automatically improve your videos—things like brightening dim footage, changing boring backgrounds to something more interesting, or adding artistic effects. If you enjoy taking videos on your phone or have old family recordings that could use a fresh look, this means you can make them more polished without needing special skills or expensive software. You can try this feature in Google Photos by looking for "Video Remix" in the editing tools, though it's currently rolling out gradually to users.
# Roku's Streaming Shortfall
Roku devices are widely used for streaming TV, but they're falling behind competitors like Apple TV and Amazon Fire in features like faster performance, better picture quality, and access to newer apps. If you're relying on an older Roku to stream your shows, you might notice it's slower to respond or missing some of the services your friends with newer devices can access. Worth knowing: before buying a streaming device, check which apps and services you use most often, since not all platforms support them equally.
TikTok's algorithm learns what you like by watching how long you pause on videos, but researchers discovered that when you tell it "not interested," that feedback only works temporarily—the app gradually starts showing you those same videos again unless you keep saying no repeatedly. For anyone who finds their feed cluttered with unwanted content, this matters because it means your clicks to dismiss videos have less power than TikTok suggests, and the platform seems designed to keep testing whether you might change your mind. Worth knowing: if you genuinely want to steer your FYP away from certain topics, consistent, repeated use of the "not interested" button works better than a one-time click—though you shouldn't expect instant results.
Meta is adding a light that turns on whenever someone uses the camera in their AI glasses to record video, so people nearby will know they're being filmed. This matters because privacy concerns about wearable cameras are real—many people feel uncomfortable when they don't know if they're being recorded in public or private spaces. Worth knowing: when these glasses become widely available, you might want to look for that light as a signal that someone is actively recording near you.
ChatGPT's voice mode now responds to what you're saying in real time, interrupting naturally like a human would, and it can deliberately speak more slowly if you ask it to. This makes conversations feel more natural and gives you time to process what you're hearing—especially helpful if you prefer a leisurely pace or have hearing difficulties. You can try this feature by enabling voice mode in ChatGPT and simply asking it to "speak slower" during any conversation.
OpenAI has created a smarter version of ChatGPT's voice feature that listens better, interrupts you less often, and knows when to pause so you can finish your thoughts—much like talking with a real person. For someone who wants to use their voice instead of typing, this means conversations will feel more natural and less awkward, without the AI jumping in before you're done speaking. You can try this new voice mode when OpenAI rolls it out and notice whether it feels like a genuine back-and-forth conversation rather than talking to a machine that's always eager to respond.
