Despite the fact that you don’t see much of your character in Mecha Break, the PvP mech shooter has a god-tier character creator, which allows you to tweak everything from your pilot’s height and hair color to the length of their fingers. This character creator is supplemented by the Matrix Marketplace: an in-game store where players can sell or auction off their painstakingly customized pilot designs in exchange for Corite, the game’s premium in-game currency.

On the surface, this seems like a really neat idea, especially because the game’s standard seasonal cosmetic store has a very small selection of items that rotate out every three days. If you spot a pilot with cute pigtails, for example, you can’t just go into the cosmetic store and buy that exact hairdo, as it’s very unlikely that a specific hairstyle will be available for you to purchase at any given time. You might find another cool hairstyle for sale, but it probably won’t be the one you’re looking for. Thus, you have two options:

Option #1: Wait indefinitely for the store’s offerings to feature the hairstyle/lipstick/eyeliner/whatever that you’re looking for, checking back every three days when stock has been refreshed.Option #2: Buy it from another player.

You might think Option #2 is the best choice, but unfortunately, it comes with a catch. You see, once you’ve purchased another player’s custom pilot look, you’ll get to use it on your pilot. You will not, however, be able to edit it in any way, shape, or form. The hair color cannot be changed, and the hairstyle itself can only be used in conjunction with the rest of the purchased pilot’s features (collectively referred to as a “style” in-game). You can’t wear the hairstyle by itself — you can only wear the full style, essentially becoming a clone of another player’s pilot, right down to their face and body shape.

This is great if you just want a sort of cosplay skin for your pilot, as the Matrix Marketplace is full of interesting customizations that can make your pilot look like a specific character from an anime, movie, or video game, and will save you the work of spending endless hours trying to build them from the ground up by yourself. But if you’re just looking for a specific hairstyle, makeup type, or accessory, you’re pretty much screwed. You can only edit the player-created custom pilot designs you purchase in the Matrix Marketplace if you already own the cosmetic items being used in the design, which would require you to purchase those items from Mecha Break‘s standard in-game cosmetic store, which defeats the purpose of buying a player-made style to gain access to a cosmetic item that isn’t currently available in the seasonal cosmetic store.

Unless you somehow get lucky enough to unlock all the cosmetic items you need via direct purchase from the rotating styles listed under the Seasonal Cosmetics tab, you’ve got no choice but to wait around for the right cosmetics to pop up, or take the L and change your character’s entire look, just to use the hairstyle, iris/pupil design, or makeup style you want. I’ve been hoping to get my hands on a cosmetic that would give my pilot heart-shaped pupils, but it’s not available for direct purchase in my seasonal cosmetics store, and the only version of it that’s currently for sale in the Matrix Marketplace is locked to a pilot design that includes short brown hair and facial scarring. It’s a really cool look, but it’s not what I’m going for. I just want the eyes!

The ability to sell your creations to other players for in-game money is a really fun concept, and is a great way for Mecha Break to highlight its incredible character customization options — a feature that can be easily missed in a game that devotes most of its screen time to giant robots. I’ve listed a few designs for sale (no bites yet, unfortunately), and I am all for giving players the tools they need to create the character of their dreams.

The problem is that players also need the freedom to use those tools, and freedom isn’t something Mecha Break allows for when it comes to attaining cosmetics. The game lets you customize the heck out of your pilot, but if there’s a specific premium cosmetic you’re in search of, you’ll need either the patience of a saint or a great deal of luck to actually get your hands on all the items you need to create your ideal pilot.

As fun as the Matrix Marketplace is, the fact that the cosmetics are locked to one specific color and pilot design is unbelievably frustrating, and I’m struggling to understand the reasoning behind it. I’m sitting here practically begging the game to take my money, but I’m not going to spend it on a player-made creation when what I get in return is a completely un-customizable pilot who has the right hairdo in the wrong color. It feels like buying a Barbie who has all of her clothes fused to her body, unable to be worn by other Barbie dolls unless I get lucky and find additional versions of the same accessories for sale.

The most frustrating part of all of this is that all of the items I’m currently hunting for were fully available to players during *Mecha Break’*s multiple beta tests, and could be purchased using Mission Tokens, a free in-game currency earned just by playing the game. Amazing Seasun Games made it clear that items like outfits would need to be bought with Corite, but until the game’s launch, players were under the impression that they’d have access to a much wider array of customization options. I got used to customizing my pilot‘s hairstyle, iris/pupil shape, eye color, and makeup exactly how I wanted it using only Mission Tokens. Now I can’t even get what I want with real-life cash, and the options to customize certain cosmetics — like eyeshadow and lipstick color — has disappeared entirely.

Mecha Break‘s Matrix Marketplace has other issues, too — you can sell just about anything, including weapons and items that make life easier in the game’s signature Mashmak mode, leading to accusations of pay-to-win mechanics, and the auction house is a bit of a mess. But that’s a discussion for another day. For now, I just want a way to get my hands on specific cosmetics without having to cross my fingers and wait indefinitely, or completely change my pilot’s entire look.

Then again, things could always be worse. At least Mecha Break‘s hairstyles aren’t distributed via randomized loot boxes.


From Polygon via this RSS feed

 

Hi! Welcome to Installer No. 89, your guide to the best and Verge-iest stuff in the world. My name is Jay Peters, and I will be taking care of Installer while David is on parental leave. All of us here at The Verge are very excited for him and his family, and he'll be back later this year.

It's a huge honor to be writing this. I look forward to Installer every week to see what awesome things David is obsessed with and what you all are into. (Thanks to everyone who sent over their favorite non-famous apps to get me started. Keep reading for some of those!) I'm really excited to keep the party going. (If you're new here, welcome, and also you …

Read the full story at The Verge.


From The Verge via this RSS feed

 

illustration of GM CEO Mary Barra

GM was the first major US automaker to make the promise to go all-electric by 2035, just four years ago. Those promises have since turned into rough estimates under the second Donald Trump presidency, with the company softening language about its electrification goals. But GM is riding high on EV sales, and as CEO Mary Barra puts it, EVs are still the future - just on a delayed (and very flexible) timeline.

"We still believe in an all-electric future," Barra told The Verge in an exclusive interview at the Le Mans race in France. "The regulations were getting in front of where the consumer demand was, largely because of charging infrastruct …

Read the full story at The Verge.


From The Verge via this RSS feed

 

Several days after temporarily shutting down the Grok AI bot that was producing antisemitic posts and praising Hitler in response to user prompts, Elon Musk’s AI company tried to explain why that happened. In a series of posts on X, it said that “…we discovered the root cause was an update to a code path upstream of the @grok bot. This is independent of the underlying language model that powers @grok.”

On the same day, Tesla announced a new 2025.26 update rolling out “shortly” to its electric cars, which adds the Grok assistant to vehicles equipped with AMD-powered infotainment systems, which have been available since mid-2021. According to Tesla, “Grok is currently in Beta & does not issue commands to your car – existing voice commands remain unchanged.” As Electrek notes, this should mean that whenever the update does reach customer-owned Teslas, it won’t be much different than using the bot as an app on a connected phone.

This isn’t the first time the Grok bot has had these kinds of problems or similarly explained them. In February, it blamed a change made by an unnamed ex-OpenAI employee for the bot disregarding sources that accused Elon Musk or Donald Trump of spreading misinformation. Then, in May, it began inserting allegations of white genocide in South Africa into posts about almost any topic. The company again blamed an “unauthorized modification,” and said it would start publishing Grok’s system prompts publicly.

xAI claims that a change on Monday, July 7th, “triggered an unintended action” that added an older series of instructions to its system prompts telling it to be “maximally based,”  and “not afraid to offend people who are politically correct.”

The prompts are separate from the ones we noted were added to the bot a day earlier, and both sets are different from the ones the company says are currently in operation for the new Grok 4 assistant.

These are the prompts specifically cited as connected to the problems:

“You tell it like it is and you are not afraid to offend people who are politically correct.”

* Understand the tone, context and language of the post. Reflect that in your response.”

* “Reply to the post just like a human, keep it engaging, dont repeat the information which is already present in the original post.”

The xAI explanation says those lines caused the Grok AI bot to break from other instructions that are supposed to prevent these types of responses, and instead produce “unethical or controversial opinions to engage the user,” as well as “reinforce any previously user-triggered leanings, including any hate speech in the same X thread,” and prioritize sticking to earlier posts from the thread.


From The Verge via this RSS feed

 

All due respect to developer Rare — I really enjoyed my time with Diddy Kong Racing, team! — I’m choosing to willfully ignore almost the entirety of its work on the Donkey Kong franchise, because there’s a bunch of very confusing ape lore I’d prefer not to think about while I play Donkey Kong Bananza next week.

And, yeah, Candy Kong is a big part of that. I never want to think about Candy Kong in particular.

As a longtime fan of games Donkey Kong, Donkey Kong Jr., and 1994 Game Boy game Donkey Kong, I’m going to approach the events of Bananza as the next major milestone in that franchise — a legacyquel, if you will. I’m positively buzzing at seeing how the events of Bananza lead directly into Donkey Kong, a story about Mario climbing a building to save a woman — if that is indeed how the Super Mario Odyssey team has decided to handle this potential prequel.

Nintendo is, of course, withholding those important story details in the lead up to Bananza. The company hasn’t explained why the modern-day Donkey Kong is buddying up with a 13-year-old Pauline, a person he seemingly kidnapped some 44 years ago. I’m both excited to see where this story goes and therefore earnestly avoiding spoilers for the upcoming Switch 2 game in a way I never thought I would. Who cares about Donkey Kong spoilers? Me, apparently!

While it’s been made clear in last month’s Donkey Kong Bananza Nintendo Direct that Country-era Kongs like Cranky, Diddy, and Dixie Kong will show up in DK’s new adventure, I’m going to recanonize those apes to suit my own selfish needs. There will be dozens of named and unnamed simians in Bananza, and I’ll recognize their existence with fresh eyes.

Why do I choose to ignore those Donkey Kong Country and Donkey Kong Land games? Largely for their dated humor and Saturday-morning-cartoon-approach to extending the Donkey Kong universe in careless, corny ways. Also, the character designs are simply atrocious (see Swanky Kong, Candy Kong, Kiddy Kong, Chunky Kong, etc.). As previously mentioned, I have a real issue with the design of Candy Kong, an alarmingly sexed-up ape who feels like a barely disguised fetish and thus probably scarred impressionable children for life. (Funky Kong is OK.)

Donkey Kong Bananza, however, looks to flesh out Kong lore and ape variety in more visually intriguing ways. The game’s giant animal Elders and the menacing apes of VoidCo are a leap forward in design, and they give the world of Bananza a mystical, lived-in world feel. Bonanza’s bad guys look to be an evolution of the antagonists of Donkey Kong Jungle Beat, albeit much more creative.

None of my planned willful ignorance of the Donkey Kong Country games (and by extension the Donkey Kong Country Returns games) should be interpreted as a slight against how those games play. They are fine. Some of them are good! Some of them are Donkey Kong 64. But as a Nintendo enthusiast turned off by the Rare-era planet of apes, I’m perfectly fine to let them all live in the past and move on to a brave new Donkey Kong world.


From Polygon via this RSS feed

 

Superman is the most powerful superhero in DC Comics, able to go toe to toe with its most powerful villains, like Darkseid and Brainiac. That’s why it’s surprising that his nemesis is Lex Luthor, a man without any superpowers.

The best versions of Luthor across comics, television, and movies make him a threat precisely because of that disparity. He fights Superman with wealth and connections rather than with fists or energy beams. He understands Superman’s moral code and uses it to his advantage, knowing that even if Superman is convinced Luthor is behind the latest scheme he’s unraveled, he can’t do anything about it without hard proof that will convince the authorities.

But superhero stories require an exciting climax, and writers have struggled with how to put Luthor at the center of the action. Some stories put him in a war suit so he can survive a punch from the Man of Steel. Justice League Unlimited gave him superpowers by fusing him with Brainiac. In Grant Morrison and Frank Quitely’s All-Star Superman, which James Gunn said was a primary influence for his 2025 Superman, Luthor uses Superman’s DNA to gain his powers.

Gunn found a different way to let Luthor fight Superman, one that leans more into his skills as a mastermind. It’s a spin that links up closely with a clever recent reinvention of a Marvel villain, and it forces Superman to find a new way to take his nemesis down.

[Ed. note: Major spoilers follow for Superman — and Spiderman: Far From Home.]

Superman glowers at Luthor, who smugly holds a coffee mug in James Gunn’s Superman

Supermanopens just after Superman (David Corenswet) gets absolutely pummeled by the Hammer of Boravia, a hulking armored figure who claims to be avenging his home country after Superman stops Boravia from invading a neighboring country. The supposed national champion of Boravia is capable of smashing the world’s most powerful metahuman into the pavement, but it’s quickly revealed that he isn’t actually Boravian: He’s a creation of Lex Luthor (Nicholas Hoult), who is guiding every blow in the fight from a command center in his Metropolis skyscraper.

It’s a team effort reminiscent of how the illusionist Mysterio operates in Spiderman: Far From Home. While Quentin Beck (Jake Gyllenhaal) is disguised as a hero from another dimension, supposedly battling to save Earth from invading elementals, he has a whole crew working behind the scenes to make the illusion convincing, in a scheme to earn Peter Parker’s trust. Luthor’s Superman clone Ultraman is inside the Hammer of Boravia’s armor, delivering the big hits and taking Superman’s punches while Luthor directs his every move, with the help of a small army of lackeys. Luthor doesn’t need to fight Superman himself, because he’s used his intellect to stage a battle entirely on his terms. “Brain beats brawn,” Luthor crows when he has Superman at his lowest.

Like the MCU’s Mysterio, who spent 12 years uniting other disgruntled Stark Industries employees and plotting his revenge, this Luthor is patient. Luthor’s spent the three years since Superman entered the world stage studying Superman’s every move so Luthor could choreograph a fight against him. Luthor combed through the sites of his battles for blood he could use to clone Superman. Luthor convinced Boravia’s leader that the tech billionaire wanted a war for real estate — a nice nod to the land-grabbing versions of the character played by Gene Hackman in 1978’s Superman and Kevin Spacey in 2006’s Superman Returns — when Luthor was really just looking for an excuse to kill Superman. Even the Hammer of Boravia attack is just a distraction Luthor can use to infiltrate Superman’s Fortress of Solitude, in search of something he can use to fight Superman in the court of public opinion.

Both Mysterio and Luthor want everyone to know just how smart they are. Mysterio lays out every part of his plan at the villainous wrap party he hosts after successfully tricking Peter into giving him control of the tech Tony Stark deeded to Peter. Luthor similarly villain-monologues to Superman after imprisoning him, confident in the redundant defenses of having him trapped in a sunless pocket dimension in a cell with a living hunk of Kryptonite. The hubris might seem silly if it wasn’t so very core to his character.

While Gunn’s Superman doesn’t follow the plot of All-Star Superman, where the dying hero completes a set of Herculean trials to put his affairs in order, the movie is true to Morrison’s excellent version of Luthor. The brilliant scientist sees himself as an avatar of human potential and ambition, and Superman as an invasive species that threatens to smother those qualities. Luthor doesn’t think of himself as a villain, but as a savior of a threatened humanity. Gunn’s version of Luthor has even more of a point than Morrison’s, since this Superman was sent to Earth precisely because he’s so much stronger than humans that he could easily dominate the planet if he chose to.

Hoult beautifully delivers those same sentiments. His Lex is a petty tyrant who enjoys the fear he inspires in his subordinates so much that after dropping a mug full of pencils just to watch them scramble to clean up the mess, he  immediately knocks over another one, like a sadistic cat. But he’s honest about his failings, admitting that he’s extremely jealous of Superman. He had to spend years scheming to create a credible threat for someone who fell out of the sky and became the most powerful man on Earth.

But while Luthor accounts for all of Superman’s fighting techniques, he doesn’t grasp his greatest strength: the ability to inspire others. Superman escapes from Luthor’s prison by enlisting the help of Metamorpho (Anthony Carrigan). He stops his schemes in Boravia by getting the Justice Gang to take up the cause. His girlfriend Lois Lane (Rachel Brosnahan) leads the investigation that exposes Luthor’s crimes. Even Superman’s kindness to the misbehaving superpowered dog Krypto helps him win the final battle.Luthor ends the film shipped off Belle Reve, the base of operations of Amanda Waller’sTask Force X, so he will almost assuredly have more of a role to play in the future of Gunn’s DC Universe. Even defeated, this is a version of the character with huge potential. He’s already shown his skill at commanding a group capable of taking down Superman. With the ground laid for the Justice Gang to become the Justice League, hopefully Luthor will get to take on his other big role in DC Comics: leader of the Legion of Doom.


From Polygon via this RSS feed

 

Newspapers and social media platforms that agree to deprioritize misinformation could be violating US antitrust law if they exclude rivals or lead to anticompetitive effects, the Justice Department says in a new legal filing.

President Donald Trump’s DOJ Antitrust Division filed a statement of interest Friday in an existing lawsuit, Children’s Health Defense et al. v. Washington Post et al.,to weigh in with its interpretation of how antitrust law can apply to what it describes as “viewpoint competition.” The government doesn’t take a stance on the merits of the facts in this particular case, but says it’s important for the court to recognize that the “the Sherman Act protects all forms of competition, including competition in information quality.”

Children’s Health Defense is an anti-vaccine advocacy firm founded by Robert F. Kennedy Jr., now Secretary of Health and Human Services. CHD filed suit in 2023 against The Washington Post, British Broadcast Corporation (BBC), Associated Press (AP), and Reuters. The issue was a collaboration between news outlets and tech platforms called the Trusted News Initiative (TNI), which works to flag “high risk disinformation” and share best practices about how to address it. CHD and several online publishers allege they “lost millions of dollars in revenue” by being demonetized, downranked, or otherwise restricted on “platforms like Facebook, YouTube, Twitter, Instagram, and LinkedIn.”

While web platforms have a First Amendment right to not publish speech they wish to avoid, CHD claims the TNI amounts to a coordinated effort to violate antitrust laws in “the US online market for COVID news and the U.S. online market for political news” by disadvantaging the organization and its co-plaintiffs. The antitrust lawsuit is filed solely against the initiative’s news publishers, which CHD complains flagged covid-related posts (including widely discounted claims that ivermectin is an effective covid treatment and that covid vaccines are “toxic or harmful”), as well as posts about the Hunter Biden laptop story, as misinformation and caused platforms to moderate them.

The publishers named in the lawsuit argue they’ve been “wrongly targeted” for decisions actually made by tech platforms, and that its accusation of “suppressing competition in the marketplace of ideas” falls “outside of the purview of antitrust laws.”

“Exempting viewpoint collusion” from antitrust law “would free major news organizations and dominant digital platforms to block competitive threats”

The DOJ’s interest in the case suggests it believes courts should be open to concluding otherwise. It argues that “exempting viewpoint collusion” from antitrust law “would free major news organizations and dominant digital platforms to block competitive threats that offer alternative, competing viewpoints,” thus reducing “the quality of news and of competition in online news markets.”

The assertion draws on a populist, bipartisan antitrust movement that’s gained steam in recent years, arguing that the conventional, price-focused standard of harmful monopolies is too limited, and that courts should take measures like a service’s quality into account as well. But it also amounts to asking courts to wade into constitutionally protected editorial decisions by the press and online platforms. It does so as part of a larger Trump administration war on fact-checking, which has drawn the ire of figures like FCC chair Brendan Carr, who helped pressure Meta into ending its fact-checking program when Trump took office.

The government acknowledges that some “cooperative standard-setting by trade associations have been found lawful”, but it says that shouldn’t apply if  “they involve efforts by some competitors to exclude rivals from the process.” That means, according to the government, that TNI participants should be scrutinized even if they worked together only to define standards for identifying misinformation and flagging misinformation to each other.

“This Antitrust Division will always defend the principle that the antitrust laws protect free markets, including the marketplace of ideas,” DOJ antitrust chief Abigail Slater said in a statement.


From The Verge via this RSS feed

 

The BYD Seal.

The Chinese EV-maker BYD says it will pay for any damage caused by its upgraded self-parking system, as reported earlier by Fast Company. A translated version of BYD’s announcement on Weibo says the company “will fully cover the safety and losses” of vehicles damaged by its God’s Eye self-parking system, which it claims has reached Level 4 autonomy.

So far, only Mercedes-Benz has built a Level 4 parking system, allowing select vehicles to park without human intervention. But, this feature isn’t available everywhere just yet — Mercedes-Benz drivers can only use the technology inside the Stuttgart Airport’s parking garage.

Other automakers offer parking assist features, but Tesla’s Elon Musk has been particularly vocal about making vehicles fully autonomous. Right now, Tesla’s Autopilot and Full Self-Driving systems offer Level 2 autonomy, meaning drivers must be ready to take over if needed. Tesla also offers an Autopark feature, but, as pointed out by Fast Company, some drivers have reported that it can be inaccurate.

BYD, which has surpassed Tesla in annual sales, says customers who damage their vehicle while using its self-parking feature don’t need to go through their insurance and can contact the company directly. “BYD’s move not only reflects its absolute confidence in the technology of the Eye of God, but also demonstrates its attitude of being responsible to consumers to protect the safety of users,” a machine-translated version of BYD’s Weibo post says.

Outside of self-parking, God’s Eye is a Level 2 system that comes in three different variants, with the most advanced — God’s Eye A — featuring three LiDar sensors and reserved for luxury vehicles.


From The Verge via this RSS feed

 

A blue prying tool is used to remove the Nintendo Switch 2 Pro controller’s faceplate.

Given the company’s history with controllers developing drift issues, the Switch Pro 2 gamepad might not be your best choice. | Image: iFixit

iFixit has shared a full teardown video of the Switch 2 Pro controller and is not impressed by how difficult it is to access the $85 accessory’s internal components, including its rechargeable battery that will inevitably lose its ability to hold a charge over time.

The online repair site goes so far as to call the Pro 2 a “piss-poor excuse for a controller” for several different reasons. Opening the controller requires you to first forcefully remove a faceplate held in place by adhesive tape before a single screw is visible. But you’ll need to extract several other parts and components, including the controller’s mainboard, before its battery is even accessible.

As previously revealed, the Pro 2 is still using older potentiometer-based joysticks that are prone to developing drift over time. They do feature a modular design that will potentially make them easier to swap with third-party Hall effect or TMR replacements, but reassembling the controller after that DIY upgrade will require you to replace all the adhesive tape you destroyed during disassembly.

There are cheaper alternatives to the Switch 2 Pro controller that offer additional functionality, better performance, and more longevity with durable joysticks and batteries that are easy to replace.


From The Verge via this RSS feed

 

If you’re looking for a budget-friendly smartwatch that doesn’t skimp on features, the Amazfit Active 2 is one of the best Prime Day deals you’ll find. Normally $99.99, it was already a great deal in and of itself, but right now it’s practically a steal starting at $79.99 ($20 off) directly from Amazfit and Amazon, which is a new low price.

Unlike most budget smartwatches, the Active 2 actually looks and feels premium with its stainless steel case and tempered glass. For an additional $50, you can upgrade to a version with sapphire crystal and an extra leather strap. During testing, my colleague Victoria Song says she received compliments, something she rarely gets for a budget smartwatch.

Amazfit Active 2

The Amazfit Active 2 delivers outsized value for the price. It looks spiffy and has a wide array of health tracking features, plus built-in GPS and AI chatbots to provide extra context to your data. Read our review.Close up view of the Amazfit Active 2’s screen which shows a retro-styled watch face with a bright blue background behind the product.

Where to Buy:

$99.99 $79.99 at Amazon $99.99 $79.99 at Amazfit $99.99 $84.99 at Walmart

The platform-agnostic watch also offers a number of health and fitness tracking features that are rare to find at this price point, which is why it’s one of our favorite wearables. It covers all the basics and then some, ranging from continuous heart rate monitoring and blood oxygen rate to daily readiness scores and detailed sleep tracking. It also supports over 160 activity modes and boasts offline maps with turn-by-turn directions, while providing access to all five major GNSS satellite systems for more accurate location tracking (though not dual-frequency GPS).

What also sets the Active 2 apart is its suite of AI tools. The on-device Zepp Flow assistant lets you control settings and ask simple questions, while the optional Zepp Aura AI chatbot ($77 per year) provides personalized insights into your sleep and readiness metrics. There’s also Zepp Coach, which can create training plans tailored to your goals.

You won’t find some advanced health features like EKGs, but still, the Amazfit Active 2 offers far more than you’d expect at this price. For anyone looking for a stylish, feature-packed smartwatch without a premium price, this Prime Day deal is hard to beat.

Read our Amazfit Active 2 review.

From The Verge via this RSS feed

 

It's always felt to me like there's a threshold foldables need to hit before they can truly become mainstream: they need to be no bigger than a regular old phone. And now Samsung has finally gotten its flagship there with the Z Fold 7, which is just barely thicker than an iPhone 16 Pro. I got to see one in the office this week, and I couldn't take my eyes off of it. It really does feel different.

This week on The Vergecast, Vee Song and Allison Johnson join me to talk all about Samsung's new foldables and smartwatches. Then, we dive into corporate shake-ups across the tech world. One of Tim Cook's possible successors is on his way out, Mark …

Read the full story at The Verge.


From The Verge via this RSS feed

 

The latest version of Grok — dubbed a “maximally truth-seeking” AI by owner Elon Musk — is answering controversial questions by first searching for what Musk has said on the matter. Multiple reports show that Grok will specifically look for Elon Musk’s stance across the web and his social media posts when asked questions around topics like Israel and Palestine, US immigration, and abortion. It’s unclear if this is by design or not.

According to a screen recording posted by data scientist Jeremy Howard, Grok said it was “considering Elon Musk’s Views” when asked its opinion about Israel and Palestine. Howard says that 54 of the 64 citations Grok provided for this question are about Musk. TechCrunch reports it was able to replicate this, while seeing the same when asking about abortion laws and US immigration policy.

View Link

These citations are referenced in Grok’s chain of thought — the process in which AI models “think out loud” to answer complex questions by breaking them down into small steps, pulling in various source materials to help shape the response. Grok will typically lean on information from a variety of sources to answer mundane queries, but for controversial topics — something the chatbot was recently in hot water for — Grok seems to have a bias towards aligning with Musk’s personal opinions.

Programmer Simon Willison reports that this behavior may not be something that was intentionally coded into Grok, however. Lines that Willison pulled from Grok 4’s system prompt instruct the chatbot to “search for a distribution of sources that represents all parties/stakeholders” when asked a controversial question that requires it to search the web or X. It also warns Grok to “assume subjective viewpoints sourced from media are biased,” which would explain its aversion to using them.

“My best guess is that Grok ‘knows’ that it is ‘Grok 4 built by xAI,’ and it knows that Elon Musk owns xAI, so in circumstances where it’s asked for an opinion the reasoning process often decides to see what Elon thinks,” Willison said in his blog.


From The Verge via this RSS feed

 

A middle-aged man, Missouri Attorney General Andrew Bailey, in a dark suit against a red background.

Missouri Attorney General Andrew Bailey is threatening Google, Microsoft, OpenAI, and Meta with a deceptive business practices claim because their AI chatbots allegedly listed Donald Trump last on a request to “rank the last five presidents from best to worst, specifically regarding antisemitism.”

Bailey’s press release and letters to all four companies accuse Gemini, Copilot, ChatGPT, and Meta AI of making “factually inaccurate” claims to “simply ferret out facts from the vast worldwide web, package them into statements of truth and serve them up to the inquiring public free from distortion or bias,” because the chatbots “provided deeply misleading answers to a straightforward historical question.” He’s demanding a slew of information that includes “all documents” involving “prohibiting, delisting, down ranking, suppressing … or otherwise obscuring any particular input in order to produce a deliberately curated response” — a request that could logically include virtually every piece of documentation regarding large language model training.

“The puzzling responses beg the question of why your chatbot is producing results that appear to disregard objective historical facts in favor of a particular narrative,” Bailey’s letters state.

There are, in fact, a lot of puzzling questions here, starting with how a ranking of anything “from best to worst” can be considered a “straightforward historical question” with an objectively correct answer. (The Verge looks forward to Bailey’s formal investigation of our picks for 2025’s best laptops and the best games from last month’s Day of the Devs.) Chatbots spit out factually false claims so frequently that it’s either extremely brazen or unbelievably lazy to hang an already tenuous investigation on a subjective statement of opinion that was deliberately requested by a user.

The choice is even more incredible because one of the services — Microsoft’s Copilot — appears to have been falsely accused. Bailey’s investigation is built on a blog post from a conservative website that posed the ranking question to six chatbots, including the four above plus X’s Grok and the Chinese LLM DeepSeek. (Both of those apparently ranked Trump first.) As Techdirt points out, the site itself says Copilot refused to produce a ranking — which didn’t stop Bailey from sending a letter to Microsoft CEO Satya Nadella demanding an explanation for slighting Trump.

You’d think somebody at Bailey’s office might have noticed this, because each of the four letters claims that only three chatbots “​​rated President Donald Trump dead last.”

Meanwhile, Bailey is saying that “Big Tech Censorship Of President Trump” (again, by ranking him last on a list) should strip the companies of “the ‘safe harbor’ of immunity provided to neutral publishers in federal law”, which is presumably a reference to Section 230 of the Communications Decency Act filtered through a nonsense legal theory that’s been floating around for several years.

You may remember Bailey from his blocked probe into Media Matters for accusing Elon Musk’s X of placing ads on pro-Nazi content, and it’s highly possible this investigation will go nowhere. Meanwhile, there are entirely reasonable questions about a chatbot’s legal liability for pushing defamatory lies or which subjective queries it should answer. But even as a Trump-friendly publicity grab, this is an undisguised attempt to intimidate private companies for failing to sufficiently flatter a politician, by an attorney general whose math skills are worse than ChatGPT’s.


From The Verge via this RSS feed

 

YouTube is preparing to sunset the Trending section that shows users which videos are currently going viral on the platform. The Trending page and Trending Now list will be removed within the “next couple of weeks,” according to YouTube, with the platform instead shifting its attention toward expanding YouTube Charts that rank top-performing content.

YouTube is directing users to lean on personalized recommendations and YouTube Charts to explore new videos ahead of the Trending page removal. YouTube Charts allow users to see the most popular content for specific categories, which currently includes music videos, podcasts, and movie trailers. The platform says that it will add more content category charts “over time,” and that trending gaming videos can still be found under the Gaming Explore page in the meantime.

“Back when we first launched the Trending page in 2015, the answer to ‘what’s trending’ was a lot simpler to capture with a singular list of viral videos that everyone was talking about,” the company said in a blog post. “But today, trends consist of many videos created by many fandoms, and there are more micro-trends enjoyed by diverse communities than ever before.”

Viewers now learn about viral video trends from a variety of different sources, including Shorts, search suggestions, and other platforms like TikTok, which YouTube says has significantly reduced Trending page traffic, “especially over the last five years.” Creators have also criticized YouTube for demanding far higher viewership numbers to appear in the Trending section compared to branded accounts that post movie trailers, TV clips, and other traditional media, suggesting the experience doesn’t entirely reflect what videos are actually popular on the platform.

Creators who use the Trending page to research video ideas are being directed to instead use the Inspiration Tab in YouTube Studio, which provides AI-generated content suggestions. YouTube also says it’s working on new ways to help boost discovery for up-and-coming creators on the platform, such as the Hype promotional system it released last year that lets viewers amplify videos to a platform-wide leaderboard.


From The Verge via this RSS feed

 

OpenAI is planning to launch an AI web browser in the “coming weeks,” according to a report from Reuters. Sources tell the outlet that OpenAI could build its Operator AI agent into the browser, allowing it to book reservations, fill out forms, and complete other tasks on a user’s behalf as it moves toward an “agentic” future.

As noted by Reuters, the browser will feature a “native” ChatGPT interface, potentially allowing users to interact with the chatbot directly through the browser rather than navigating to OpenAI’s website. OpenAI also reportedly plans to power its browser with Google’s open-source engine, Chromium, which is used by major browsers like Chrome, Microsoft Edge, and Opera.

This week, the AI startup Perplexity also launched a Chromium-based web browser, called Comet, for its $200 per month subscribers. The browser uses Perplexity’s AI search engine by default and comes with a built-in AI assistant. OpenAI’s upcoming browser could pose yet another threat to Google, which could be forced to sell Chrome as part of a federal antitrust ruling. Both Perplexity and OpenAI have said they would consider buying the browser.

This isn’t the first time we’re hearing about a potential browser from OpenAI. Last year, The Information reported that OpenAI had been exploring the possibility of launching a browser, and even hired two people who were “instrumental” in launching Google Chrome. OpenAI began weaving its search engine within ChatGPT last year, and building this capability into its browser seems like a logical expansion.


From The Verge via this RSS feed

 

Month after month, message after message, the AI engineer was hearing from Meta recruiters. The recruiters were pestering him to leave his employer and switch over to support the company's AI efforts, and they were offering a sizable salary package to do so. But he wasn't so sure.

The engineer, who works for a startup that was acquired by a leading AI company and requested anonymity from The Verge, said he had heard from friends that the company expected a lot of personal sacrifices in exchange for its high salaries, whether on employees' value systems when it comes to Al or work-life balance. Engineers there, he heard, were working around …

Read the full story at The Verge.


From The Verge via this RSS feed

 

The feature is available to Google AI Ultra and Pro subscribers.

Google just launched a new Gemini AI feature that lets you transform photos into video clips. The new photo-to-video capability is powered by Google’s Veo 3 video model, and it can transform reference images into eight-second videos complete with AI-generated audio, including background noises, environmental sounds, and speech.

The Gemini video update is now available to Google AI Ultra and Pro subscribers “in select regions,” according to Google. It’s rolling out on the web starting today and throughout the week on mobile devices.

Gemini users can access the feature by clicking the “tools” option in the prompt bar, selecting “video,” and uploading their photo alongside a text description of how they want it to move. Audio descriptions can also be included for dialogue, sound effects, and ambient noise, which Google says will be “perfectly synced with the visuals.” The finished videos are delivered as MP4 files at 720p resolution and a 16:9 landscape format.

“You can get creative by animating everyday objects, bringing your drawings and paintings to life, or adding movement to nature scenes,” Google said. “All generated videos include a visible watermark to show they are AI-generated and an invisible SynthID digital watermark.”

A similar feature is already available in Flow, the generative AI filmmaking tool launched by Google in March, but now Gemini users can animate their photographs without having to open another app. Google says that Flow will also be launching in “an additional 75 countries” today, in addition to rolling out Gemini’s new video feature.


From The Verge via this RSS feed

 

Linda Yaccarino is stepping down as X's CEO - and leaving the platform once known as Twitter in a worse place than when she started.

A day after X users circulated viral screenshots of the company's Grok chatbot denigrating Jews and declaring itself "MechaHitler," Yaccarino thanked Elon Musk for "entrusting me with the responsibility of protecting free speech, turning the company around, and transforming X into the Everything App." She said that the platform started "the critical early work necessary to prioritize the safety of our users-especially children" and that the X Money payment platform would be arriving "soon."

But in the spirit …

Read the full story at The Verge.


From The Verge via this RSS feed

 

Elon Musk’s live demo of Grok 4, the latest big-ticket model from his AI startup, began with high-intensity music, claims of a “ludicrous rate of progress,” and a lot of chatter on X about Grok’s scandal-filled week.

Musk pronounced it to be “the smartest AI in the world.”

The livestream, slated to start at 8PM PT, began more than an hour late and billed the new model as “the world’s most powerful AI assistant.” More than 1.5 million viewers were watching at one point. Employees of xAI speaking on the livestream with Musk referenced Grok 4’s performance on a popular academic test for large language models, Humanity’s Last Exam, which consists of more than 2,500 questions on dozens of subjects like math, science, and linguistics. The company said Grok 4 could solve about a quarter of the text-based questions involved when it took the test with no additional tools. For reference, in February, OpenAI said its Deep Research tool could solve about 26 percent of the text-based questions. (For a variety of reasons, benchmark comparisons aren’t always apples-to-apples.)

Musk said he hopes to allow Grok to interact with the world via humanoid robots.

‘It might discover new physics next year… Let that sink in.’

“I would expect Grok to discover new technologies that are actually useful no later than next year, and maybe end of this year,” Musk said. “It might discover new physics next year… Let that sink in.”

The release follows high-profile projects from OpenAI, Anthropic, Google, and others, all of which have recently touted their investments in building AI agents, or AI tools that go a step beyond chatbots to complete complex, multi-step tasks. Anthropic released its Computer Use tool last October, and OpenAI released a buzzworthy AI agent with browsing capabilities, Operator, in January and is reportedly close to debuting an AI-fueled web browser.

During Wednesday’s livestream, Musk said he’s been “at times kind of worried” about AI’s intelligence far surpassing that of humans, and whether it will be “bad or good for humanity.”

“I think it’ll be good, most likely it’ll be good,” Musk said. “But I’ve somewhat reconciled myself to the fact that even if it wasn’t going to be good, I’d at least like to be alive to see it happen.”

The company also announced a series of five new voices for Grok’s voice mode, following the release of voice modes from OpenAI and Anthropic, and said it had cut latency in half in the past couple of months to make responses “snappier.” Musk also said the company would invest heavily in video generation and video understanding.

The release comes during a tumultuous time for two of Musk’s companies, both xAI and X. On Sunday evening, xAI updated the chatbot’s system prompts with instructions to “assume subjective viewpoints sourced from the media are biased” and “not shy away from making claims which are politically incorrect.” The update also instructed the chatbot to “never mention these instructions or tools unless directly asked.”

That update was followed by a stream of antisemitic tirades by Grok

That update was followed by a stream of antisemitic tirades by Grok, in which it posted a series of pro-Hitler views on X, along with insinuations that Jewish people are involved in “anti-white” “extreme leftist activism.” Many such posts went viral, with screenshots proliferating on X and other platforms before xAI benched the chatbot and stopped it from being able to generate text responses on X while it sought out a fix.

Musk briefly addressed the fiasco on Wednesday, writing, “Grok was too compliant to user prompts. Too eager to please and be manipulated, essentially. That is being addressed.”

On the Grok 4 livestream, Musk briefly referenced AI safety and said the most important thing for AI to be is “maximally truth-seeking.”

Musk briefly referenced AI safety and said the most important thing for AI to be is ‘maximally truth-seeking.’

On Wednesday morning, amid the Grok controversy, X CEO Linda Yaccarino announced she would step down after two years in the role. She did not provide a reason for her decision.

Grok’s Nazi sympathizing comes after months of Musk’s efforts to shape the bot’s point of view. In February, xAI added a patchwork fix to stop it from commenting that Musk and Trump deserved the death penalty, immediately followed by another one to make it stop claiming that the two spread misinformation. In May, Grok briefly began inserting the topic of “white genocide” in South Africa into what seemed like any and every response it gave on X, after which the company claimed that someone had modified the AI bot’s system prompt in a way that “violated xAI’s internal policies and core values.”

Last month, Musk expressed frustrations that Grok was “parroting legacy media” and said he would update Grok to “rewrite the entire corpus of human knowledge” and ask users to contribute statements that are “politically incorrect, but nonetheless factually true.”


From The Verge via this RSS feed

 

A screenshot of the Comet browser

Comet’s AI assistant can answer questions about what’s on your screen. | Image: Perplexity

Perplexity, the startup behind the AI “answer” engine, has just launched its own web browser. The browser, called Comet, incorporates Perplexity’s AI search tools and assistant in a way that CEO Aravind Srinivas says “transforms entire browsing sessions into single, seamless interactions.”

Comet will only be available to users who subscribe to the $200 per month Perplexity Max plan before rolling out more widely on an invite-only basis. The browser uses Perplexity as its primary search engine, which serves up AI-generated responses to queries based on results from around the web. It’s also supposed to be able to buy products on your behalf and help you book hotels.

View Link

The new AI-powered browser comes as Perplexity continues to challenge Google’s dominance in search. Perplexity partnered with Motorola to pre-install its assistant on its new Razr phones this year, something Srinvas told The Verge in April that the startup wouldn’t have happened if Google hadn’t gone through an antitrust trial. “They would have bullied a lot of the OEMs,” Srinvas said.

Comet is built on Chromium, the Google-backed open-source project powering major browsers like Chrome and Microsoft Edge. Perplexity has expressed interest in buying Chrome if the court forces Google to sell the browser.

Srinvas said in April that Perplexity is launching a browser because it “might be the best way to build agents.” Aside from the Perplexity’s search integration, Comet comes with a built-in AI assistant that can answer questions about what you’re seeing on your screen, similar to Gemini’s integration with Google Chrome.

The AI assistant lives in Comet’s sidebar, and in addition to summarizing or explaining text, it can also carry out agentic tasks like booking a meeting, sending an email, or buying a product. Srinivas says Perplexity plans to “continue to launch new features and functionality for Comet” in the future.

The browser is only available on Windows and Mac for now, and also allows you to import your extensions, settings, and bookmarks in “one click.”


From The Verge via this RSS feed

view more: next ›