Best of theverge.com on Kiwi News
Most upvoted by Kiwi News readers
56 stories from theverge.com have been shared on Kiwi News since 20 April 2023, with 119 upvotes and 29 comments in total. The latest was shared on 27 April 2026. The 50 most upvoted are listed below.
Text under a title is an AI summary of the linked article.
- Apple Removes E2E iCloud Encryption in the UK
Apple has stopped offering Advanced Data Protection, its end-to-end encrypted iCloud storage, to new users in the UK, and existing UK users will eventually have to disable it. Apple spokesperson Julien Trosdorf told The Verge that Apple is "gravely disappointed" that ADP protections will not be available to UK customers. The move follows a report by The Washington Post that the UK Home Office, led by Home Secretary Yvette Cooper, demanded backdoor access to encrypted files of users worldwide. Trosdorf said Apple has never built a backdoor or master key and never will.
- Major AWS outage takes down Fortnite, Alexa, Snapchat, and more
A major AWS outage on Monday morning took down multiple online services for several hours, including Amazon, Alexa, Snapchat, Fortnite, ChatGPT, and Epic Games Store. AWS's dashboard first reported issues affecting the US-EAST-1 Region at 3:11AM ET, and Amazon attributed the outage to resolution issues in its domain name system (DNS). Amazon later said at 6:35AM ET that the underlying issue "has been fully mitigated," and provided a final update at 6:53PM ET stating that all AWS services returned to normal operations by 6:01PM ET.
- Telegram will now hand over your phone number and IP if you’re a crime suspect
Telegram's updated privacy policy says it may now disclose a user's IP address and phone number to authorities when it receives a valid judicial order confirming the user is a suspect in a case involving criminal activities that violate its Terms of Service. The change follows the arrest of Telegram CEO Pavel Durov last month, whom French authorities charged with enabling illegal activity on the platform. Durov says the changes aim to deter criminals from abusing the app's search function to sell illegal goods, and that the company's previous law enforcement policy applied only to terror suspects, as reported by 404 Media. Telegram says it will disclose in its quarterly transparency reports whether it provided user information to authorities.
- No one’s ready for this. (our basic assumptions of photos changing)
The Verge's Sarah Jeong argues that the Pixel 9's Magic Editor makes convincing fake photos in under 10 seconds, and that this will flip the default assumption about photos from authentic to faked. The article says the Reimagine tool produced realistic images such as an explosion beside a brick building and a cockroach in takeout. It notes that Reimagine edits carry only a removable metadata line, while Google's SynthID watermark is applied to Pixel Studio outputs. Google communications manager Alex Moriconi said the company's guardrails and policies limit abusive content, but the article contends that standard content moderation will not preserve the photograph's value as evidence.
- X will let people you’ve blocked see your posts
X will let users see posts from accounts they have blocked, while still preventing those accounts from engaging with their posts. In a reply on Monday, X owner Elon Musk said the "block function will block that account from engaging with, but not block seeing, public post." Currently, X shows a "You're blocked" message when a blocked user tries to view the profile of the person who blocked them, and it hides their posts, replies, media, followers, and following list. A source at X told The Verge that the change is being made because people can already view posts from blocked users through another account or while logged out, though Verge staff said they found that X blocks profile viewing for logged-out users.
- US officials recommend encrypted messaging to evade hackers in telecom networks
US officials are urging Americans to use encrypted apps for calls and texts to reduce the risk that private information reaches foreign adversaries still inside US telecommunications networks, according to NBC News. FBI and CISA officials spoke with reporters on the lasting effects of a hack tied to Chinese group Salt Typhoon, which affected AT&T, Verizon, T-Mobile and Lumen Technologies. Jeff Greene, executive assistant director for cybersecurity at CISA, said the scale of the hack was so great that agencies could not predict when a "full eviction" of malicious material would occur. He stated that encryption would make intercepted data impossible to use, and the article notes that Signal and WhatsApp offer end-to-end encryption.
- AI-generated art can’t be copyrighted after Supreme Court declines to review
The US Supreme Court declined to hear Stephen Thaler's appeal over whether AI-generated art can be copyrighted, leaving in place rulings that such works are ineligible for protection. Thaler, a Missouri computer scientist, sought copyright for the image "A Recent Entrance to Paradise," created by an algorithm he built, after the US Copyright Office rejected it in 2019 and again in 2022 for lacking "human authorship." Judge Beryl A. Howell ruled in 2023 that "human authorship is a bedrock requirement of copyright," and a federal appeals court in Washington, DC upheld that ruling in 2025. The Copyright Office's newer guidance says AI-generated artwork based on text prompts is not protected by copyright.
- Bluesky is rolling out age verification in the UK
Bluesky is adding age verification in the UK to comply with the Online Safety Act, letting users verify their age by scanning their face, uploading an ID, or entering a payment card. The verification runs through Kid Web Services (KWS), an Epic Games-owned tool that developers use for age checks and parental controls. Users who don't verify, or who are under 18, will still have accounts with some features limited, including direct messaging and access to "adult-appropriate content." The article notes that another part of the law takes effect on July 25th, requiring sites with porn and harmful content to provide "highly effective age assurance," with fines of up to £18 million or 10 percent of worldwide revenue.
- Bluesky crosses the 15 million user mark
Bluesky has passed 15 million users, according to a stat-tracking site built by Bluesky developer Jaz using the Bluesky API. The platform added about a million new users in the last week amid a surge of signups after the US presidential election, and Bluesky COO Rose Wang told The Verge that the majority of new users have come from the US. Bluesky recently rolled out video posting, pinned posts, custom fonts and "anti-toxicity" features, such as detaching posts when someone quotes them. Meta's Threads, which has 275 million monthly users, remains larger, though Bluesky offers user-created algorithmic feeds in addition to its own.
- Major Windows BSOD issue takes banks, airlines, and broadcasters offline
Thousands of Windows machines are crashing at boot with a Blue Screen of Death, taking banks, airlines, TV broadcasters and supermarkets offline. CrowdStrike, a cybersecurity provider, says a faulty update to its Windows security software caused the problem, and that a fix has been deployed. Sky News could not broadcast its morning bulletins for hours, and Ryanair reported a "third-party" IT issue affecting flight departures. The FAA said it was assisting airlines including Delta, United and American Airlines with ground stops. CrowdStrike CEO George Kurtz said Mac and Linux hosts are not affected and that the event is not a security incident or cyberattack.
- Social media is doomed to die by Ellis Hamburger
Ellis Hamburger, a former Snapchat employee, argues that social media apps are doomed to drift away from their original purpose of connecting friends because of their dependence on investors and advertisers. He points to Snapchat's push notification asking users to wish a friend happy birthday as a sign that the app had adopted growth tactics it once rejected. He writes that Snapchat Stories was intended from the start to include ads, and he says users also contribute to the cycle by favoring broadcasting and growing their friend lists. He suggests newer apps and protocols, such as Mastodon, Locket and BeReal, may offer a different path, though the article notes investors have little interest in small-scale social products.
- The birth of id Software
John Romero recounts how, in September 1990 while working at Softdisk, he saw a demo by John Carmack that showed smooth horizontal side-scrolling on a PC, which he considered a turning point for PC gaming. The demo, a Super Mario Bros. 3-style program with Romero's character Dangerous Dave, was made by Carmack and Tom, a colleague. Romero says Jay Wilbur suggested pitching it to Nintendo, which turned down the request three weeks later. Romero then describes how fan letters from Scott Miller, who was using multiple names, led to contact with Miller and his company Apogee.
- The fediverse, explained
The fediverse is an interconnected ecosystem of social apps built on open protocols, most prominently ActivityPub, which lets users move their posts, followers and data between platforms. David Pierce, writing for The Verge, explains that the fediverse is not a single network but many apps, such as Mastodon, Pixelfed, Lemmy and PeerTube, that can read and interact with each other's content. He says he expects ActivityPub to become the standard, noting Threads has promised to federate and Flipboard is moving to the protocol. The article notes that Mastodon, though the largest fediverse app, remains small compared with Facebook.
- Bored Ape NFT event attendees report ‘severe eye burn’
Attendees of ApeFest, a Bored Ape Yacht Club NFT event held in Hong Kong from November 3rd to 5th, reported eye pain, vision problems and sunburnt skin after the event. One attendee posted on X that they were diagnosed with photokeratitis, a condition caused by unprotected ultraviolet exposure, and another blamed UV from the stage lights, leading to speculation about improper lighting. Yuga Labs, the company behind the project, said it is contacting those affected to understand the root cause. It said the 15 people it has been in direct communication with represent less than one percent of the approximately 2,250 attendees and staff at its Saturday night event.
- X glitch wipes out most pictures and links tweeted before December 2014
X is failing to display images and expanded links in older posts, with the problem appearing to affect tweets published before December 2014. Tom Coates highlighted the issue on a Saturday, after a Brazilian vtuber, @DaniloTakagi, had pointed it out a couple of days earlier. The glitch removed the picture from Ellen DeGeneres's 2014 Oscars selfie, though that image was restored later that night. The @Support account said on Monday, "We fixed the bug, and the issue will be fully resolved in the coming days." The article links the cutoff to Twitter's 2016 "enhanced URL enrichment" changes, and X did not respond to requests for comment beyond an automated message.
- The moral bankruptcy of Marc Andreessen and Ben Horowitz
Marc Andreessen and Ben Horowitz of venture firm Andreessen Horowitz have declared support for Donald Trump, citing crypto, AI regulation, and tax policy, according to a podcast episode of The Ben & Marc Show. The author argues their stated reasons are weak, since Trump's own record on crypto is not clearly pro-industry, and the complaints about regulation and antitrust conflict with their own investments and behavior. On the proposed tax change, Andreessen calls it "the final straw," describing a plan to tax unrealized capital gains on wealth above $100 million, and Horowitz calls it a "very scary proposal."
- I regret to inform you that LinkedIn’s games are very fun
LinkedIn has added three daily games to its app, Pinpoint, Crossclimb, and Queens, and columnist David Pierce says they are fun and effective at bringing him back every day. He describes the games as closely modeled on The New York Times' puzzles, with Pinpoint resembling Connections in reverse and Crossclimb resembling the Times' mini crossword. Pierce writes that LinkedIn is using the games to become more than a job site, and he notes that Microsoft reported LinkedIn revenue up 10 percent year over year, and that CEO Satya Nadella said engagement hit a record high last quarter.
- Startups face PMF problems, but often overlooked are the distribution problems
Beating Google at search is far harder than building a search engine at all, and Neeva's short life is presented as evidence of this. Sridhar Ramaswamy and Vivek Raghunathan co-founded Neeva in 2018 after Ramaswamy left Google, where he had run its advertising division. Neeva built its own index after starting with Bing's data, and its engine came close to Google in internal metrics and user studies, though in a test of about 500 queries, human raters found Google slightly ahead. Neeva shut down only four years in.
- Flipboard’s Surf app is a big new idea about the future of social
Flipboard CEO Mike McCue believes the open, decentralized social web is the future, and Flipboard's new Surf app is meant to be a browser for it. Surf organizes content into feeds, drawing from ActivityPub platforms such as Mastodon and Threads, Bluesky via the AT Protocol, and any RSS feed. The app is now in public beta, though access is limited at first while curators and creators are brought in. The reviewer found filtering useful and video and news presentation pleasant, but search messy and interaction mostly limited to Mastodon users.
- OpenAI plans to release its next big AI model by December
OpenAI plans to launch Orion, its next frontier model, by December, according to The Verge, which cites sources. Unlike GPT-4o and o1, Orion will not initially be released widely through ChatGPT; OpenAI plans to grant access first to companies it works closely with, according to a source familiar with the plan. A second source said Microsoft engineers are preparing to host Orion on Azure as early as November. OpenAI spokesperson Niko Felix later told The Verge the company doesn't "have plans to release a model code-named Orion this year," though the company plans to release other technology.
- The messy WordPress drama, explained
The WordPress community is in a public dispute between WordPress co-founder Matt Mullenweg and WP Engine, a hosting company that sells its own prepackaged WordPress service. Mullenweg called WP Engine a "cancer" and, in a September 21 blog post, said it contributes just 40 hours per week to the WordPress.org open source project. WP Engine's cease and desist letter alleges that Automattic demanded a "very large sum of money" and threatened a "scorched earth" approach if it did not pay. WordPress.org blocked WP Engine from its servers, then temporarily lifted the block and gave the company until October 1 to create its own mirror or resolve the conflict.
- Don’t ever hand your phone to the cops
Digital IDs in Apple Wallet and Google Wallet make it tempting for police to ask for phones, but the article argues you should never voluntarily hand your phone to an officer. It notes that California's digital IDs are for select TSA checkpoints and participating businesses, not police stops, while Louisiana and Colorado have digital IDs that can be used during traffic stops. Brett Max Kaufman of the ACLU says that handing over an unlocked phone creates a complicated question about the scope of consent. The article also says courts disagree on whether police can compel biometric unlocking, and that the Fifth Amendment typically protects passcodes but often not biometrics.
- How BeReal missed its moment
Casey Newton argues BeReal missed its moment because the team shipped only a couple of user-facing features in the past nine months. A New York Times report by Callie Holtermann said daily users fell 61 percent from about 15 million in October to under six million in March, per Apptopia, while the company's unsigned blog post claims 20 million daily users. Its latest change, a UK test giving users two bonus posts a day, looks to Newton like an engagement tactic that few users asked for.
- X will start charging new users in two countries $1 per year
X will start charging new web users in New Zealand and the Philippines $1 a year as part of a "Not A Bot" program, according to a post on X's help center published Tuesday. X said the subscription is meant to bolster its efforts to reduce spam, platform manipulation and bot activity, and new web users in those countries must also verify their account with a phone number. Users who do not subscribe will only be able to take "read only" actions, such as viewing posts and watching videos. The program is separate from X's main $8 monthly subscription.
- Why note-taking apps don’t make us smarter
Casey Newton argues that note-taking apps and productivity software will not make people smarter, because thinking happens in the brain and resists automation. He describes how his own use of Roam Research, Obsidian and Mem, with journaling and bidirectional links, failed to produce the insights he expected. He attributes much of the problem to constant digital distraction, citing Gloria Mark's finding that the average time on a single screen fell from 2.5 minutes to about 47 seconds. He expects that a chat interface over his Notion link database could help as a research assistant, though today's chatbots cannot yet meet a reporter's standard.
- An interview with Midjourney founder David Holz
Midjourney, a 10-person research lab that runs an AI image generator through a Discord server, is expanding access so anyone can create their own Discord server with its image generator, according to an interview with founder David Holz. Holz says the company is focused on expanding human imagination rather than deepfakes, and that its community of roughly a million people co-creates images in shared spaces. He says the company has banned words to stop users from generating shock imagery and that the model's default style makes deepfakes hard to produce.
- This is Instagram’s new Twitter competitor
Instagram is reportedly building a text-based app for conversations that would compete with Twitter, according to a leaked marketing slide that Lia Haberman shared in her ICYMI Substack newsletter. The app is reportedly codenamed P92 or Barcelona, and users could sign in with their Instagram credentials, with followers, handle, bio, and verification carrying over. Text posts of up to 500 characters with attached links, photos, and videos would be supported, and the slide says the app will be compatible with apps like Mastodon, letting their users follow and interact with public profiles. Meta declined to comment, but in March it told Platformer it was exploring a standalone decentralized social network for text updates.
- Discord’s username change is causing discord
Discord is replacing its four-digit discriminator usernames with unique, Twitter-style handles, with users selecting new names in a staggered rollout over several months, prioritized by registration date. Co-founder and CTO Stanislav Vishnevskiy said the discriminators were too confusing, noting that over 40 percent of users don't know their number and that "almost half" of friend requests fail, largely due to mistyped numbers. Several users and server admins quoted in the article object that the change pushes Discord toward the social media norms of public platforms, eroding the anonymity and pseudo-identity culture that many users valued.
- Apple Intelligence is here, but it still has a lot to learn
Apple Intelligence has launched with features the reviewer finds underwhelming, and the author argues that its time-saving promise is not yet realized. The reviewer says notification summaries and the Clean Up tool in Photos work reasonably well, though Clean Up is only about as good as Google's older Magic Eraser, which runs on-device on a four-year-old iPhone 12 Mini. Email summaries, smart replies and the upgraded Siri are described as offering little beyond the existing features, with bigger Siri updates planned for 2025. The reviewer adds that Apple's "Built for Apple Intelligence" marketing set expectations that the current release does not meet.
- Big Tech quietly sponsors Trump’s military parade party
Major tech and defense companies are sponsoring festivities for Trump's military parade in Washington, DC, while US taxpayers cover the cost of the soldiers, tanks and planes. According to America250, Oracle, Amazon, Coinbase, Lockheed Martin and Palantir are among its highest-profile sponsors, and the nonprofit is handling spectator logistics for the June 14 parade. Amazon, Palantir and Oracle did not confirm their involvement, while Lockheed Martin confirmed its sponsorship. Meta was listed as a top-level sponsor until America250 removed its logo after The Verge asked, and Meta said it is not currently a sponsor.
- Here’s some cool stuff you can do with Bluesky
Bluesky offers a range of features and settings that differ from X and Threads, including user-selectable algorithmic feeds, moderation tools and customizable reply threads, according to this guide. The article notes that Bluesky has passed 15 million users and recently added direct messages, pinned posts and 60-second videos. It describes custom feeds, Starter Packs, custom domain handles that require a TXT record, third-party Labelers for moderation and tagging, and App Passwords for third-party apps such as Graysky. It also warns that a handle's domain becomes available for others to claim once a custom domain is removed.
- Software brain is changing the world, but most people still aren’t buying.
The Decoder host argues that "software brain," a way of seeing the world as databases, algorithms and loops, explains the gap between tech industry enthusiasm for AI and growing public distrust of it. He cites polls showing Gen Z is increasingly hopeful less and angry more about AI, with only 18 percent of Gen Z hopeful in a recent Gallup poll. He argues that tech leaders see this as a marketing problem, citing OpenAI's reported $200 million spend on the TBPN podcast, while the real issue is that people's direct experience of AI does not match the industry's view of reality. He also contends that the law is not deterministic and cannot be treated like code.
- Sir Tim Berners-Lee doesn’t think AI will destroy the web
Sir Tim Berners-Lee, inventor of the World Wide Web, discusses the web's future on Decoder with host Nilay Patel, the article's author. Berners-Lee says he remains optimistic but worries the web has strayed from its original vision of a democratizing force for knowledge and creativity. He argues that market dynamics have led to monopolies like Chrome, Google Search and a single dominant social network, and criticizes TikTok's addictive algorithms. The interview also covers his memoir This Is For Everyone and his work at Inrupt on the Solid standard.
- Spotify adds lossless streaming after 8 years of teasing
Spotify is rolling out lossless audio to all Premium subscribers in 50 markets over the next two months, without a new higher-priced tier. The article lists launch markets including Australia, Germany, Japan, the US and the UK, and says Spotify has confirmed prices will not change as part of the rollout. Spotify's lossless option tops out at 24-bit / 44.1 kHz FLAC, while Apple Music, Tidal and Qobuz offer HiRes FLAC up to 24-bit / 192 kHz. The article notes YouTube is now the only major streaming music service without lossless support.
- DeepSeek’s top-ranked AI app is restricting sign-ups due to ‘malicious attacks’
DeepSeek's AI Assistant, which surged to the top of Apple's US App Store charts, is temporarily restricting new user sign-ups. An incident report page says registrations are limited "due to large-scale malicious attacks on DeepSeek's services," though the article notes it is unclear how the limits are applied. An earlier update had said only registration with a mainland China mobile phone number was supported, but that requirement was later deleted, and Verge staff were able to create accounts through Google and Apple ID sign-in. Existing users can log in as usual, according to DeepSeek.
- Toyota’s futuristic Woven City in Japan is ready for its first residents
Toyota says its Woven City in Japan, built on the site of a former car factory at the base of Mount Fuji, is nearly ready for its first residents, with an official launch planned for 2025. Toyota Chairman Akio Toyoda described the $10 billion project at CES as a "living laboratory" where residents can help inventors test new ideas. In fall 2025, Toyota plans to welcome the first 100 residents, all employees of Toyota or its subsidiary Woven by Toyota, with the first phase eventually housing 360 people. The article notes that Toyota's plans for the city's relation to its vehicle business remain unclear.
- Apple is ‘concerned’ about AI turning real photos into ‘fantasy’
Apple software chief Craig Federighi said in an interview with The Wall Street Journal that the company wants its AI-powered image editing to preserve photo authenticity, and that Apple is "concerned" about AI's impact on how people view photos as reflections of reality. The iOS 18.1 Photos app gains a "Clean Up" feature that removes objects and people from images, which Federighi and WSJ reporter Joanna Stern noted is far more limited than the editing tools offered by Google and Samsung. Edited images will be tagged "Modified with Clean Up" and embedded with metadata flagging the alteration. Apple Intelligence does not currently let users add AI-generated content to images, and it is unclear whether Apple's metadata system will support the Adobe-driven Content Authenticity Initiative's Content Credentials.
- In iPadOS 18, the whole iPad is a calculator app
The Calculator app in iPadOS 18 is essentially the iPhone version scaled up, but the article's author, David Pierce, argues that the real calculator functionality lies elsewhere in the system. Calculations work in the text boxes of Apple's built-in apps; typing "225/4=" in Messages returns $56.25. Math Notes, used within the Notes app, lets users handwrite equations that are solved automatically, update when variables change, and generate graphs. Pierce notes that it is strict about legible handwriting, can't reuse its answers in other equations, and still has beta-stage quirks.
- OpenAI releases o1, its first model with ‘reasoning’ abilities
OpenAI released o1, the first in a planned series of "reasoning" models trained to answer complex questions, alongside the smaller, cheaper o1-mini. According to the article, o1 does better than GPT-4o at writing code and solving multistep problems, though it is more expensive and slower to use. OpenAI told the article's author that o1 scored 83 percent on a qualifying exam for the International Mathematics Olympiad, compared with 13 percent for GPT-4o, and reached the 89th percentile in Codeforces programming competitions. ChatGPT Plus and Team users received access starting that day, while OpenAI's research lead, Jerry Tworek, said the model "hallucinates less" but that hallucinations are not solved.
- Apple is going to open up iPhone tap-to-pay
Apple will let developers offer in-app NFC contactless transactions using the iPhone's Secure Element starting with iOS 18.1, and users will be able to set a default contactless payment app that opens with a double-click of the side button. Previously, only Apple Pay could appear when the side button was double-clicked. Developers can build apps for in-store payments, car keys, transit, corporate badges, student IDs, hotel keys, loyalty cards and event tickets, and Apple says government IDs will be supported "in the future." The APIs will first be available in an upcoming iOS 18.1 developer beta in Australia, Brazil, Canada, Japan, New Zealand, the UK and the US, but developers must enter a commercial agreement with Apple and pay associated fees.
- Zoom CEO Eric Yuan wants AI clones in meetings
Zoom CEO Eric Yuan describes a vision in which people send AI "digital twins," deepfake avatars of themselves, to attend meetings and make decisions on their behalf. In a Decoder interview with Nilay Patel, Yuan says Zoom is moving from videoconferencing into a broader Workplace platform, positioned against Microsoft and Google in enterprise software, with the Zoom AI Companion as a central feature. He says AI could handle about 90 percent of work within five or six years, though he also says AI cannot replace in-person interaction, and he suggests this could support a four-day or three-day workweek.
- Humane is looking for a buyer
Humane, the startup behind the AI Pin wearable, is reportedly seeking a buyer for its business at a price of between $750 million and $1 billion, according to Bloomberg. The company, led by former Apple employees Imran Chaudhri and Bethany Bongiorno, has raised $230 million from investors including OpenAI CEO Sam Altman. The article says the $699 device was widely panned for slow responses and a user experience that fell short of its promised always-on AI assistant concept, and that Humane was valued at $850 million in 2023, before the product's release.
- Two students find security bug that could let millions do laundry for free
Two University of California, Santa Cruz students, Alexander Sherbrooke and Iakov Taranenko, found a vulnerability in internet-connected laundry machines operated by CSC ServiceWorks, according to TechCrunch. They reportedly exploited an API in the machines' app to start machines without payment and to inflate a laundry account to show millions of dollars. CSC ServiceWorks says it has more than a million laundry and vending machines in the US, Canada, and Europe. The students said they reported the flaw by email and phone in January without a response, and that the company later removed the false balances after they contacted it.
- How the team behind Zelda made physics feel like magic
Nintendo's Tears of the Kingdom developers said at the 2024 Game Developers Conference that the game's acclaimed feel came from a deliberate strategy: a physics-driven world built to support "multiplicative gameplay." Technical director Takuhiro Dohta described two guiding principles, "a vast and seamless Hyrule" and multiplicative gameplay, meaning systems that let players combine actions and objects in their own ways. Physics programmer Takahiro Takayama said making every object physics-driven avoided the need for "dedicated implementation," so developers built systems, such as a system for vehicles, instead of individual programs. The article also notes that every speaker had worked at Nintendo for at least 10 years.
- Five takeaways from the huge US antitrust lawsuit against Apple
The US Department of Justice has sued Apple for monopolizing smartphone markets, arguing that its conduct protects iPhone dominance to the detriment of consumers and third-party developers. The suit was filed on March 21st and touches on iMessage lock-in, App Store developer fees, super apps and cloud gaming. Apple spokesperson Fred Sainz said the lawsuit threatens the principles that set Apple products apart, and the company called it a dangerous precedent. The case is expected to take years, as the earlier government case against Microsoft took over three years to settle.
- Tumblr’s owner is striking deals with OpenAI and Midjourney for training data
Automattic, the owner of Tumblr and WordPress.com, is reportedly in talks with Midjourney and OpenAI to provide training data scraped from users' posts, according to a 404 Media report that says the deals are "imminent." The report cites internal posts describing an "initial data dump" of "all Tumblr's public post content between 2014 and 2023," which apparently included content not publicly visible on blogs. Automattic's statement published after the report says it blocks major AI platform crawlers by default and will share only public content from sites that haven't opted out. It also says it is "working directly with select AI companies" whose plans align with attribution, opt-outs, and control. It is unclear what was done with the data or whether any has been sent to either company.
- Adrian Wojnarowski’s X account was compromised to scam NFT traders
A scam tweet was posted from ESPN reporter Adrian Wojnarowski's account on X on a Saturday evening at about 6:30 PM ET, falsely claiming a "free NFT pack" was available to all customers of NBA Top Shot. The tweet linked to a fake NBA Top Shot site on a .org address, which could attempt to drain assets from users who connected their crypto wallets. The official Top Shot account posted about half an hour later saying there was no free airdrop, and the tweet was removed after being live for nearly an hour.
- The rise and fall of robots.txt
Robots.txt, a plain-text file that lets website owners tell web crawlers which parts of their site they may access, has worked as an informal handshake standard for about 30 years, but the article argues that AI training crawlers are straining that arrangement. Martijn Koster and other web administrators proposed the Robots Exclusion Protocol in 1994, and by summer of that year it was broadly accepted. Google's John Mueller says search crawlers have generally complied with it, while Medium CEO Tony Stubblebine says Medium is blocking AI crawlers because AI companies give writers nothing in return. A study by Ben Welsh found 606 of 1,156 surveyed publishers had blocked GPTBot.
- How BeReal missed its moment
BeReal has lost momentum since its mid-2023 surge, and the author argues its recent product choices show it is not yet solving the core problem of keeping users engaged. Citing Apptopia data reported by The New York Times, the author notes daily users fell 61 percent from about 15 million in October to under six million in March, a figure BeReal disputed in an unsigned blog post claiming 20 million daily users. The author criticizes the team for shipping only a couple of user-facing features in nine months and for launching "bonus BeReals," a UK test that the author says looks like a more restrictive version of Instagram stories. The article also notes the company's policy that statements be attributed only to "the BeReal Crew," not individuals.
- Teen GTA VI hacker sentenced to life in a secure hospital
Arion Kurtaj, an 18-year-old Lapsus$ hacker who helped leak Grand Theft Auto VI footage, has been sentenced to life in a secure hospital prison, according to a BBC report. A British judge ruled that Kurtaj remains a high risk to the public because he still wants to commit cybercrimes. A London jury had earlier found that he carried out cyberattacks against Rockstar Games, Uber and Nvidia, though he was deemed unfit to stand trial because he has autism. Kurtaj leaked 90 videos of GTA VI gameplay in September while on bail, and Rockstar Games said it spent $5 million recovering from the attack.
Most discussed
- X will let people you’ve blocked see your posts
- The moral bankruptcy of Marc Andreessen and Ben Horowitz
- Apple Intelligence is here, but it still has a lot to learn
- Telegram will now hand over your phone number and IP if you’re a crime suspect
- No one’s ready for this. (our basic assumptions of photos changing)