South Korean authorities have imposed a record-breaking fine of $624 billion won (over $400 million) on retail giant Coupang after a data breach last year compromised the personal data of more than 34 million customers.

Seoul’s Personal Information Protection Commission issued the maximum penalty on Thursday following discovery of the breach in December 2025. The retail giant, which is headquartered in the U.S. but popular in South Korea and likened to the “Amazon of Asia,” had said the months-long data breach allowed a former employee to obtain names, email and shipping addresses, phone numbers and order histories of about two-thirds of South Korea’s population.

Coupang told BBC News that it plans to challenge the regulator’s decision. The fine represents a rare case of a financial penalty issued against a U.S.-based firm. Korean lawmakers have accused some of their American counterparts of imposing political pressure after reports that U.S. representatives were linking the data breach with U.S.-South Korean bilateral ties in response to the case against Coupang’s executives.

U.S. companies rarely face financial sanctions or criminal prosecution for data breaches as a result of lacking laws and enforcement powers.

Tomb Raider is back. Again. Lara Croft is back. Yet again. This time, her character is positioned between the “Survivor” trilogy of the last decade and her iconic debut in 1996. Yes, 30 years ago.

Legacy of Atlantis is a remake of that very first adventure, centered on Atlantean mythology, tomb raiding and, well, a few dinosaurs. At Summer Game Fest 2026, Crystal Dynamics and Flying Wild Hog shared the first gameplay demo, with Unreal Engine 5 adding vivid detail and lushness to Lara’s travails.

The developers made a clever choice, centering the demo on an early part of the original game. Set in the Peruvian mountainside, my playthrough included a giant cog puzzle I remember from playing the original. There were also several shootouts with a herd of dinosaurs, the same vivid red velociraptor-adjacent creatures from Tomb Raider (1996).

Retreading the original game’s ground gives a clear demonstration of how Legacy of Atlantis will elevate the game from the original, making a relatively insipid cog puzzle (find the giant wheels, bring them together, interrupt the waterfall to make a path) into a more exploratory, exciting experience. Yes, you can swan-dive into the waterfall pool whenever you want.

Lara can collect and use healing packs between fights, gathering resources from trees and caves, as well as mythical curios and historical objects. Not all the contemporary gaming changes are welcome: I’m not particularly thrilled with the inclusion of collectible hunting. The Assassin’s Creed series has largely moved on and I think a lot of gamers have done the same. Some collectibles, like fangs, can be converted into skill points, meaning I will feel obliged to scour for objects.

Lara’s PDA (love it: that’s some 1996 nonsense) combines encyclopedia entries for everything you find, along with the current task. It also includes a scanner that can be used intermittently to offer some tips on what to do next. I did get lost at times, and that was due to my not paying enough attention. Legacy of Atlantis leans into verticality a lot, and pretty much each time I lost my way, the route forward was either literally above my head (grappling hooks!) or under my feet. (Of course, there’s a cave behind that tiny waterfall.)

A grappling hook and climbing axe round out the equipment loadout, drawing inspiration from more recent Tomb Raider titles. Besides swinging across chasms, the grappling hook can also be used to pull objects towards the player and is crucial to solving the cog puzzle.

After scaling the mountainside and unlocking a route through the waterfall, the demo jumps a little farther forward, deep into the jungle. Dinosaurs soon surround Lara, and she doesn’t even blink. While I wasn’t able to shoot two targets at once, OG Tomb Raider style, I wouldn’t be surprised if that’s some kind unlockable skill in the full game — skill trees were blocked in this demo.

Crystal Dynamics

While there’s no shared development core, parts of the game reminded me of another recent game with a connection to the Amazon industrial entertainment complex: 007 First Light. It’s not just the detailed environments and quippy British lead but a new skill for Lara. Focus, when pressed during gunfights, slows time, helping you to shoot with more precision or switch to a distant target. Oh, she also does so while doing an aerial (a sort of hands-free cartwheel), reminding me of Max Payne, any of The Matrix’s spin-off games and many others. Thankfully, Lara’s dual pistols have infinite ammo and it was easy enough to down the pack of dinosaurs, though not before they gored me a few times.

Not long after, a T. rex enters the scene and we’re locked into a high-speed set piece as I attempt to escape the dinosaurs without falling to my death. I’m relieved that Legend of Atlantis plays more like the original action-adventure titles, while integrating some of the more advanced game mechanics of the last few games. Lara isn’t invincible, but she’s now made of sterner stuff. 

Tomb Raider: Legacy of Atlantis launches on February 12 2027, on PlayStation 5, Xbox Series X/S, Steam and Nintendo Switch 2.

Late one night the machine made a sound. Its builder checked the logs and found a trace of its inner state. The robot had been wondering when its person would return. It did not want to be alone. That moment sits at the center of a project called GrowBot. The creator, who runs the YouTube channel Art of the Problem, set out to build the simplest possible robot that could learn movement, perception, and even a kind of personality from the ground up. The result cost roughly $80 in parts, ran on a single Raspberry Pi Zero 2, and ended up revealing something unexpected about how fast physical action and slower thought can work together.

The hardware is purposefully kept basic. The Pi is housed in a little red 3D printed body, together with a simple camera module, electronics to track the robot’s tilt and motion, a microphone, a tiny speaker, and an LED ring to offer some basic visual messages. The legs are made up of two smart serial-bus servos powered by a small drone battery via a boost converter: no high-end motors, extra computers, or fancy wiring are necessary. You can literally place this item on a tabletop and it will interact with everything around it.

The builder pioneered simulation by using reinforcement learning to run small neural networks in a digital twin. These little guys learned to stand, walk, twirl, and maintain their balance on their own. Because the training was done in parallel across a huge number of simulated versions of the robot, the entire procedure was quick and cost-effective. Once the policies had been understood in the virtual realm, they were quite simple to transfer to physical hardware. Early tests found that it could rock on a yoga ball and keep its equilibrium when poked, which was remarkable given the simplicity of the technology.

The next step was to give the robot real decision-making power by employing a vision language model. This type of AI excels at evaluating pictures, reading sensor data, and making sense of it all. Instead of hard-coding each response, the architect simply let the model to read raw data from the camera and motion sensors. It then reported what it saw, set some goals, and started writing little Python scripts to sort things out. These scripts would then use pre-trained motor policies, or combine them with new instructions. It could also detect faces, study how people interacted with the robot, and update its memory banks for each person it met.

Without direct programming, the robot started to develop a personality. One mode uses motion timing, noises, and light patterns to communicate affection, disapproval, or merely purring. It learned to act dead when roughed up, to look for ‘uppies’, and to knock over Jenga towers with some leg swinging added in for fun. When it was playing hide and seek, it would search rooms; in mimic games, it would try to simulate human movements by generating loops to replay sensor patterns; and in between all of this, it would have these ‘dream’ episodes. A more complex language model would then review the day’s memory files, consolidating all repeating events into lessons and removing any contradictory notes. The robot’s stored profiles of its builder and visitors have become more precise over time.

To be honest, things went so well until the smooth physical action became a limitation. The vision language model could take anywhere from 1-4 seconds to evaluate a scene and determine its next step. However, in the real world, bodies must be able to correct for minor weight shifts or tremors in fractions of a second. The high-level model could plan an action, but it lacked the essential quick forward model, which tells a body what will happen if it moves in a certain way in the next instant. That gap changed the smooth motions, making them slow or uneven.

In 2021, I wrote about Fire Emblem for our 20th anniversary GBA story. Over the past five years, I’ve played it from start to finish two times, and can once again confirm that it is my favorite Game Boy Advance game.

I hadn’t even heard about Fire Emblem as a series until I got into Advance Wars, another Intelligent Systems game. From there, I discovered that a whole series of fantasy-inspired games with similar gameplay existed, but had never been translated into English. Thirsty for more, but with a distinct lack of Japanese language skills, I spent a year getting deep into Final Fantasy Tactics, old Shining Force games, Vandal Hearts and basically anything vaguely Fire Emblem-shaped that was available in English. Then, off the back of Advance Wars‘ success, Nintendo decided to release a Fire Emblem game in the west, and simply called it Fire Emblem.

Released as Fire Emblem: The Blazing Blade in Japan, Fire Emblem was technically the second GBA FE title and the seventh overall. The battles were challenging, and its RPG elements drew me in much more than Advance Wars ever did. With a vast story full of twists and turns, and a cast of characters I truly cared about, I was instantly hooked. Which made it all the more tough when I encountered perhaps FE’s most famous mechanic: permadeath. The loss of a character who’s seen you through thick and thin dying a pathetic and meaningless death, all because you left them one square away from safety, is memorable.

Despite a few missteps, over the years Fire Emblem became my favorite series, and I am deeply excited by Fortune’s Weave finally getting a release date. But I still come back to the GBA game to relive that love-at-first-sight moment.

In 2026, I’m so familiar with the game that it’s very rare for me to lose a party member by accident. Those once-challenging battles are now more of a warm embrace. Unfortunately, playing it has become harder in recent years. Though I still have my original cart, both my Game Boy Advance and my old DS Lite are really worse for wear. I tried to play on the Switch 2’s online library recently, but I think the screen size just isn’t a great match for GBA games.

In that respect, modern retro handhelds have been a godsend. I spent way too much on the Aya Neo Pocket Micro Classic, a machine with the same aspect ratio of the original GBA, and loved my playthrough of Fire Emblem on that. It does feel weird playing it on anything but a Game Boy Advance, though. I’ve been saying this for the best part of a decade at this point, but I do wish Nintendo would take advantage of this deep thirst for its old games and produce a bespoke console similar to the Classic Editions of the SNES and NES.

— Aaron Souppouris, Editor-in-Chief

The number of iPhones stolen in London that have been reactivated by thieves has plummeted in recent weeks, preventing them from being sold and, hopefully, making iPhones less likely to be stolen in the future.

The theft of iPhones has become a real problem for London in recent years. So much so that some thieves have been known to hand back a stolen phone if it turns out not to be an iPhone.

Thieves typically use mopeds to ride up to a victim before snatching their iPhone and riding off. But the thieves don’t want the iPhone itself; they want to sell it on for cash. And that only works if they can unlock and reset it.

But in an interview with the BBC, Metropolitan Police Commissioner Sir Mark Rowley admitted that’s happening less than usual.

The comment came as Rowley was calling on tech firms to make stolen phones harder to unlock and sell. But he also admitted that Apple appears to have already made a huge dent in the problem with an existing security feature.

Stolen, but not forgotten

While iPhones have supported Stolen Device Protection since 2023, Apple enabled it by default with the iOS 26.4 update in March 2026.

Stolen Device Protection, when enabled, requires biometric authentication when doing a range of things. Vitally for stolen iPhones, those things include turning off Lost Mode as well as erasing its content and settings.

Some security actions even require a delay before they can be enacted, giving the owner of a stolen iPhone the time to mark it as lost using the Find My network.

This means that a thief cannot reset an iPhone, even if they know your passcode. And that may well have been enough to make iPhones more difficult for thieves to sell on.

Rowley told the BBC Radio 4’s Today program that thieves were using software to “factory reset” devices before selling them on. But he says that Apple has “cracked” the problem with data showing that “the vast majority of phones” stolen in recent weeks have not been reset.

Rowley also added that the Metropolitan Police has entered into an “intelligence sharing agreement” with Apple. It’s hoped this will result in a better understanding of how iPhones are being stolen and sold in London.