Category: Technology

Roblox is starting to roll out the mandatory age checks that will require all of its users to submit an ID or scan their face in order to access the platform’s chat features. The updated policy, which the company announced earlier this year, will be enforced first in Australia, New Zealand and the Netherlands and will expand to all other markets by early next year.

The company also detailed a new “age-based chat” system, which will limit users’ ability to interact with people outside of their age group. After verifying or estimating a user’s age, Roblox will assign them to an age group ranging from 9 years and younger to 21 years and older (there are six total age groups). Teens and children will then be limited from connecting with people that aren’t in or close to their estimated age group in in-game chats.

Unlike most social media apps which have a minimum age of 13, Roblox permits much younger children to use its platform. Since most children and many teens don’t have IDs, the company uses “age estimation” tech provided by identity company Persona. The checks, which use video selfies, are conducted within Roblox’s app and the company says that images of users’ faces are immediately deleted after completing the process.

Roblox didn’t provide details on how accurate its age estimation features are, but the company’s Chief Safety Officer, Matt Kaufman, said that it was “pretty accurate” at guessing the approximate age of most of its users. “What we find is that the algorithms between that 5 and 25 years old [range] are typically pretty accurate within one or two years of their age,” he said during a briefing with reporters.

All Roblox users can now voluntarily submit to a face scan or provide an ID to the company to ensure their access to its chat features isn’t interrupted. The company says it will be enforcing age checks for all users by January and that people in the Netherlands, Australia and New Zealand will need to comply beginning in early December. Next year, the company also plans to put age restrictions around users’ ability to access links to outside social media sites and to participate in Roblox Studio. 

Roblox has repeatedly come under fire for alleged safety lapses even as it’s released a flurry of child safety updates in recent years. The company is facing lawsuits from Texas, Louisiana and Kentucky amid accusations that it hasn’t done enough to prevent adults from targeting teens and children on its service.

This article originally appeared on Engadget at https://www.engadget.com/gaming/roblox-begins-asking-tens-of-millions-of-children-to-verify-their-age-with-a-selfie-120000311.html?src=rss 

DJI has released its third action cam in less than four months, the Osmo Action 6. The new GoPro rival offers something we’ve never seen in this type of camera: a variable aperture for more control over exposure. It also has a new, square sensor (like the one we saw on the Osmo 360) that makes it easier to crop video for social media-friendly aspect ratios. I’ve had the Osmo Action 6 for a few days and it looks impressive so far.

Rather than a fixed f/2.8 setting as before, the Osmo Action 6 offers a variable f/2.0-4.0 aperture, which offers several benefits. When at f/2.0, it allows twice as much light as before. That in turn allows for higher shutter speeds in dim conditions, which reduces the motion blur that can hurt image quality and stabilization. It also delivers shallower depth of field to help separate the subject from the background. On sunny days, meanwhile, changing the aperture to f/4.0 lets you lower the shutter speed so video looks less staccato and more natural. 

Oddly, DJI doesn’t let you fully control the f-stop, however. You can either choose a fixed aperture at f/2.8, or Auto aperture with a range of ISOs, varying from f/2.0 – 4.0 up to f/2.8 – 4.0. That’s unfortunate, as it would be nice to be able to select a fixed f/2.0 aperture, for example, to have a shallow depth of field for portraits. (Note that you can pick an f/4.0 fixed aperture as well to get a “starburst” pattern on bright parts of the image.)

The Action 6 has a new, 1/1.1-inch square sensor (9.8 x 9.8mm), as well. It doesn’t boost resolution as some people had hoped, but does allow for more aspect ratio choices. As with the Action 5, video can be shot at up to 4K 120 fps, or 4K at 60 fps in SuperNight Mode. However, it now offers a 4K Custom mode that lets you capture 3,840 x 3,840 square video at up to 60 fps, for cropping at full 4K resolution in both horizontal and vertical formats. It also supports 4:3 4K (3,840 x 2,880) at up to 120 fps for more cropping options when shooting slo-mo. 

The new sensor is also considerably larger than the 1/1.3-inch sensor on the Osmo Action 5 and the 1/1.9-inch sensor on the GoPro 13. That allows for improved low-light performance, according to DJI, and the larger pixels also boost dynamic range to a claimed 13.5 stops. That, combined with DJI’s D-LogM and 10-bit capability, should make footage from the Osmo Action 6 easy to adjust, even on high contrast sunny days.

DJI has also taken a page from Fujifilm by introducing a Film Tone function to get six distinct looks straight out of the camera, including “cool tones,” “classic negative film” and “pleasing skin tones.” The Osmo Action 6 also offers a 2x lossless zoom (a digital zoom, effectively), portrait mode and subject centering and tracking. Finally, the new “Natural Wide” mode gives you a wider field of view without the usual ugly distortion at the corners. 

For closeup photography, the Action 6 is available with an optional macro lens, as well as an FOV Boost lens that increases the field of view from 155 to 182 degrees.

Sample 34MP photo from DJI’s Osmo Action 6

From my tests so far, low-light video quality looks considerably cleaner and more color-accurate than the Action 5. In addition, the SuperNight Mode is more useful as shutter speeds are higher and thus stabilization in low light is improved over the Action 5. I’ll have more information in my full review, coming soon. 

There are a number of other improvements as well. DJI has boosted the internal storage to 50GB from 47GB before, so you can largely shoot without a memory card, though a microSD slot is also available. As before, it’s waterproof to 20 meters (66 feet) by itself, or nearly 200 feet with DJI’s waterproof case. And on top of voice control, the Action 6 now supports gestures as well. With the same 1,950mAh Extreme Battery Plus as before, shooting time between charges is 4 hours at 1080p and 2 hours at 4K, both with Rocksteady 3.0 stabilization enabled and the screen off. 

The Osmo Action 6 is compatible with DJI’s increasingly generous accessory lineup, like selfie sticks, bike, helmet and body attachments, suction mounts and more. However, the company upgraded its Quick-Release Adapter Mount to a dual-latch system (first introduced on the Osmo Nano) that “can be attached in either direction without mechanical limitations,” the company said. That’s a big quality-of-life improvement over the old mounts, which could only be attached one way. That said, the old mounts are still compatible with the new camera and vice-versa. 

So what about US availability? It looks like, as with the Osmo 360 and Osmo Nano, that DJI’s Action 6 won’t be available on the company’s US store, but it will be sold at Amazon, B&H Photo Video and other third-party resellers, with pricing tba. In Europe, the Osmo Action 6 is now available at DJI’s store and Amazon at €379 for the standard combo with one battery, or €479 in the Adventure Combo that includes three batteries and the Battery Case 3 charger. 

This article originally appeared on Engadget at https://www.engadget.com/cameras/djis-osmo-action-6-is-the-first-action-cam-with-a-variable-aperture-120058873.html?src=rss 

While unofficial videos from filming The Legend of Zelda hit the internet first, Nintendo also unveiled the first images of its live-action project. One photo shows Link and Zelda amid the hills of New Zealand — if you’ve got characters with Elven ears, you have to film in NZ.

As Engadget’s Nathan Ingraham notes, Zelda appears with a bow and quiver, so hopefully it’s a more modern spin than the damsel-in-distress original. I mean, the game is named for her. Filming has just started, so this is all we have until a proper trailer arrives.

— Mat Smith

Get Engadget’s newsletter delivered direct to your inbox. Subscribe right here!

The news you might have missed

‘Metroid Prime 4: Beyond’ thrillingly brings the series into the modern era

The best air purifiers for 2025

To change PC gaming, Valve’s Steam Machine must solve Linux’s anti-cheat problem

DJI Neo 2 drone review

The best budget drone, now even better.

DJI’s latest drone — and it’s hard to keep track — is the Neo 2. It’s a relatively inexpensive, lightweight model for creators and hobbyists, and it’s loaded with new features, like LiDAR obstacle detection, improved video quality, and longer battery life. Gesture control is a major new Neo 2 feature that isn’t on other DJI drones. Wave a hand up, down or sideways to climb and bank, move two hands apart or together to make it approach or move away, clench a fist to stop it and hold out your palm to land. Like a bird! Manual flight controls on your smartphone is another new feature not on other DJI drones. It’s also only slightly more expensive than its predecessor.

Continue reading.

The 22 best gift ideas for the remote worker in your life

‘Hi, hey, hello, sorry, your camera is off…’

There have never been more remote workers in my life, so finding a sweet present to improve their nine-to-five has often been my approach. We’ve rounded up a few techy ideas that should make their days a little more delightful — and we didn’t even suggest noise-cancelling headphones. Because they probably already have them.

Continue reading.

Meta and Google delay undersea cables over security concerns

It’s due to a lot of things, apparently.

Meta and Google are facing delays with long-promised undersea cable projects, according to a report by Bloomberg. A Meta spokesperson blames the delays on a “range of operational factors, regulatory concerns and geopolitical risk.” Meta’s 2Africa subsea cable system is designed to give fiber internet to the African continent. It’s also intended to connect Europe with Asia and Africa. The Google-backed Blue-Raman intercontinental cable system in the region has also been delayed, following its initial announcement in 2021. The company has not provided an updated timeline.

Continue reading.

This article originally appeared on Engadget at https://www.engadget.com/general/the-morning-after-engadget-newsletter-121532744.html?src=rss 

Productivity setups are a personal thing. Some people swear by to-do lists and apps, some people like scheduling and planners. For those in the latter camp, Google has made a minor but valuable addition to the Workspace suite. Going forward, you’ll be able to block off segments of time in Google Calendar for working on specific tasks. You can add descriptions and set your preferences for that task in regard to do not disturb and visibility settings. It’s a nice step up from scheduling meetings with yourself when you have to focus on a particular assignment (which several of us at Engadget have been known to do).

This option began appearing for Google’s Rapid Release domains earlier in November, while standard release domains will see a gradual release beginning December 1. Once it’s fully rolled out, it will be available by default, whether you’re a Google Workspaces customer or just using a personal Google account.

This article originally appeared on Engadget at https://www.engadget.com/apps/google-calendar-now-lets-you-schedule-tasks-231026000.html?src=rss 

Game publisher Paradox Interactive has split with Colossal Order, the studio behind Cities: Skylines II. Colossal Order, which also made the original Cities: Skylines back in 2015, will move on to something new after completing a few last updates. The intellectual property will remain with Paradox, which said today that work on Cities: Skylines II will be moved to Iceflake Studios. The Finland-based internal team will be responsible for all development, including future content updates and the long-awaited console version, starting in 2026.

The move comes after about two years of trying to salvage the citybuilding sequel and the reputations of all parties. While the game was initially expected to debut in October 2023 on PC and consoles, the launch was limited to PC, with higher minimum specs than previously announced and with a lot of bugs. The ongoing issues caused tension and toxicity between Colossal Order and the Cities: Skylines audience. Although the game has reached a more stable state, new content has been slow to arrive and there is still no sign of a console launch as the studio exits their project.

This article originally appeared on Engadget at https://www.engadget.com/gaming/cities-skylines-ii-dev-parts-ways-with-publisher-paradox-210229692.html?src=rss 

Creators on Facebook and Instagram have long griped about accounts that lift their videos without permission. Now, Meta is rolling out a new tool that allows creators to more easily track when their videos have been reposted by others.

The company introduced a new tool for creators called “content protection,” which can automatically detect when a creator’s original reel is reposted, either fully or partially, on Facebook or Instagram. Creators who are enrolled will be able to see which accounts have shared their work and will be able to take a range of actions on the clip. 

Available actions include “track,” which allows the creator to add a label indicating the clip originally came from their account. In addition to the link back, creators will also be able to keep tabs on the number of views it’s getting. Creators can also opt to block a clip entirely, which will prevent anyone else from being able to view the reel. (Meta notes choosing this option won’t impose additional penalties on the account that lifted the original content.) Finally, creators can choose to “release” the video, which removes it from their dashboard so they will no longer have any visibility into how it’s performing.

The dashboard tracks instances of reused content,

The dashboard also provides some other details that could help creators decide how to respond. For example, they can see whether the video using their content is being monetized, which may influence their decision to track with attribution or block entirely. On the other hand, if a reel was lifted from an account with few followers, they may opt to simply keep an eye on it. 

Meta has already offered Facebook creators some of these abilities in the past through its rights manager platform, but the company says making the features available directly in the Facebook app will make it accessible to more people. Notably, the company is only offering content protection to creators who share reels on Facebook. So even though the feature will detect copycats on Instagram, it will only do so if the original video has been posted to Facebook. 

Meta says content protection is rolling out now to creators in its monetization program “who meet enhanced integrity and originality standards” as well as those already using rights manager. Creators can also apply for access directly.

This article originally appeared on Engadget at https://www.engadget.com/social-media/facebook-rolls-out-new-tools-for-creators-to-track-accounts-stealing-their-content-201020255.html?src=rss 

Google’s DeepMind just released WeatherNext 2, a new version of its AI weather prediction model. The company promises that it “delivers more efficient, more accurate and higher-resolution global weather predictions.” To that end, it should be able to provide accurate forecasts up to two weeks out, including information on temperature, pressure and wind.

It should also be able to better predict tropical storm tracks, according to researchers. This means that predictions of a hurricane’s path should be accurate up to three days out. The previous model only predicted things up to two days ahead of the storm.

The model also brings hourly forecasts into the mix. All told, Google says this new model is eight times faster than the previous iteration. This should help businesses like energy traders make more precise decisions, according to a report by Bloomberg.

“It gives you a more granular forecast,” DeepMind AI researcher Akib Uddin said. “Many other industries are quite interested in these one-hour steps. It helps them make more precise decisions. Their goal is, how can they make their business more resilient to weather?”

The improvements here stem from a new approach to weather models, as explained in this recently-published research paper. Older methods demanded the use of machine learning models that were built for image and video generation. This required repeated processing to ensure an accurate result. The new model only requires a single processing step, which also reduces reliance on costly AI computing systems.

AI may not be great at everything, despite what proponents want you to believe, but it is great at predicting the weather. Newer models typically outperform traditional prediction methods, even those that rely on supercomputers.

These models aren’t perfect, however, as even Google has acknowledged that WeatherNext 2 will likely struggle to predict outlier rain and snow events. This is due to gaps in the training data. “It’s one limitation of our forecast, but one that we are working on improving,” DeepMind research scientist Ferran Alet told Bloomberg.

AI weather prediction is becoming a crowded field. Companies like NVIDIA, Microsoft, AccuWeather and Huawei are all throwing their hats in the ring in one way or another.

This article originally appeared on Engadget at https://www.engadget.com/ai/deepmind-releases-a-new-weather-forecasting-model-for-more-accurate-predictions-175451228.html?src=rss