Microsoft has revealed that it’s created a minute-long advert for its Surface Pro and Surface Laptop hardware using generative AI. But there’s a twist: it released the ad almost three months ago, and no-one seemed to notice the AI elements.The ad, which went live on YouTube on January 30th, isn’t entirely made up of generated content. In a Microsoft Design blog post published Wednesday, senior design communications manager Jay Tan admits that “the occasional AI hallucination would rear its head,” meaning the creators had to correct some of the AI output and integrate it with real footage.“When deciding on which shots within the ad were to be AI generated, the team determined that any intricate movement such as closeups of hands typing on keyboards had to be shot live,” Tan says. “Shots that were quick cuts or with limited motion, however, were prime for co-creation with generative AI tools.”Microsoft hasn’t specified exactly which shots were generated using AI, though Tan did detail the process. AI tools were first used to generate “a compelling script, storyboards and a pitch deck.” Microsoft’s team then used a combination of written prompts and sample images to get a chatbot to generate text prompts that could be fed into image generators. Those images were iterated on further, edited to correct hallucinations and other errors, and then fed into video generators like Hailuo or Kling. Those are the only specific AI tools named by Tan, with the chatbots and image generators unspecified.“We probably went through thousands of different prompts, chiseling away at the output little by little until we got what we wanted. There’s never really a one-and-done prompt,” says creative director Cisco McCarthy. “It comes from being relentless.” That makes the process sound like more work than it might have been otherwise, but visual designer Brian Townsend estimates that the team “probably saved 90% of the time and cost it would typically take.”The process echoes recent comments from Microsoft’s design chief Jon Friedman to my colleague Tom Warren, that AI is going to become one more tool in creatives’ arsenals, rather than replacing them outright. As Friedman puts it, “suddenly the design job is how do you edit?”Despite the fact that the video has been online for almost three months, there’s little sign that anyone noticed the AI output until now. The ad has a little over 40,000 views on YouTube at the time of writing, and none of the top comments speculate that the video was produced using AI. Knowing that AI was involved, it’s easy enough to guess where — shots of meeting notes that clearly weren’t hand-written, a Mason jar that’s suspiciously large, the telling AI sheen to it all — but without knowing to look for it, it’s clear that plenty of viewers couldn’t spot the difference. The ad’s quick cuts help hide the AI output’s flaws, but suggest that in the right hands, AI tools are now powerful enough to go unnoticed.See More:

A group of tabletop game makers and other American small businesses are filed a lawsuit on Thursday against President Donald Trump and his administration to attempt to stop his tariffs. The lawsuit alleges the tariffs are “unlawful and unconstitutional” and details how they are creating challenges for the plaintiffs’ businesses.The tabletop companies that are among the plaintiffs in the case include Stonemaier Games (which announced its involvement in the lawsuit earlier this week, as reported by Polygon), Spielcraft Games, Rookie Mage Games, XYZ Game Labs, and Tinkerhouse. The 145 percent tariffs imposed on goods imported from China are a significant part of the burden, as spelled out in examples in the lawsuit:Stonemaier Games, which has its products printed at a factory in China, has imports that are “ready to ship on which it estimates it will pay millions in tariffs pursuant to the actions challenged herein.”Spielcraft Games has paid $4,335.40 in tariffs for “a board game set” that it imported from China this month.Rookie Mage Games has paid $3,120.80 in tariffs for “products imported from China” this month and will have to pay tariffs on “all future imports.”XYZ Game Labs estimates that it will have to pay “substantial tariffs” on goods manufactured in China that it plans to import later this year and on “all future imports.”Tinkerhouse is producing a “tabletop roleplaying game accessory” in China that the company has made a down payment on. The company estimates it will have to pay “substantial tariffs” on when it imports the accessory later this year. It also expects that it has to pay tariffs on “all future imports.”Other plaintiffs in the case include Princess Awesome, a clothing company, Mischief Toy Store, a toy store based in Saint Paul, MN, and others. The lawsuit was filed on Thursday with the US Court of International Trade.The lawsuit from the small companies joins a growing list of parties fighting the tariffs in court. On Wednesday, 12 states sued the administration, with New York Attorney General Letitia James calling the tariffs “unlawful” in a statement. The CEO of Learning Resources, a toy company that also filed a lawsuit against the Trump administration over tariffs, told CBS MoneyWatch that “this path is catastrophic.”However, the future of the tariffs is also murky right now, with The Wall Street Journal reporting yesterday that the Trump administration is thinking about reducing the tariffs on China See More:

Ziff Davis, the owner of several digital outlets like CNET, PCMag, IGN, and Everyday Health, is suing OpenAI over claims of copyright infringement, as first reported by The New York Times. In the lawsuit, the digital media company accuses OpenAI of  “intentionally and relentlessly” creating “exact copies” of its outlets’ works without permission. The company also alleges that OpenAI trained its AI models on its work despite Ziff Davis instructing web crawlers not to scrape its data using a robots.txt file, adding that OpenAI allegedly removed copyright information from the content it sucks up. Ziff Davis currently owns more than 45 media brands and has over 3,800 employees, making it one of the biggest publishers to sue OpenAI so far. In the lawsuit, the company said it publishes nearly 2 million new articles every year, and averages over 292 million user visits each month. Some outlets, including The Verge parent company Vox Media, The Associated Press, The Atlantic, The Financial Times, The Washington Post, have signed content licensing agreements with OpenAI. However, Ziff Davis is joining The New York Times, The Intercept, Raw Story, AlterNet, and a group of Canadian media companies on the list of those suing OpenAI over copyright infringement. Ziff Davis alleges that OpenAI has “copied, reproduced, and stored” its outlets’ work, which it uses to create responses in ChatGPT. “Ziff Davis has identified hundreds of full copies of the body text of Ziff Davis Works in merely the small sample of OpenAI’s WebText dataset that it made publicly available,” the company claims. Ziff Davis is asking the court to stop OpenAI from “exploiting” its works, as well as to destroy any dataset or models containing its content. “ChatGPT helps enhance human creativity, advance scientific discovery and medical research, and enable hundreds of millions of people to improve their daily lives,” OpenAI spokesperson Jason Deutrom said in a statement to The Verge. “Our models empower innovation, and are trained on publicly available data and grounded in fair use.” Ziff Davis declined to comment.

Ask just about anybody, and they’ll tell you that new cars are too expensive. In the wake of tariffs shaking the auto industry and with the Trump administration pledging to kill the federal EV incentive, that situation isn’t looking to get better soon, especially for anyone wanting something battery-powered. Changing that overly spendy status quo is going to take something radical, and it’s hard to get more radical than what Slate Auto has planned.Meet the Slate Truck, a sub-$20,000 (after federal incentives) electric vehicle that enters production next year. It only seats two yet has a bed big enough to hold a sheet of plywood. It only does 150 miles on a charge, only comes in gray, and the only way to listen to music while driving is if you bring along your phone and a Bluetooth speaker. It is the bare minimum of what a modern car can be, and yet it’s taken three years of development to get to this point.But this is more than bargain-basement motoring. Slate is presenting its truck as minimalist design with DIY purpose, an attempt to not just go cheap but to create a new category of vehicle with a huge focus on personalization. That design also enables a low-cost approach to manufacturing that has caught the eye of major investors, reportedly including Jeff Bezos. It’s been engineered and will be manufactured in America, but is this extreme simplification too much for American consumers?Simplify, Then Embrace DamageIf you haven’t seen the leaks and the reports of weirdly wrapped trucks hiding in plain sight, the Slate Truck is the first product from Michigan-based Slate Auto. Think “American kei truck” and you’re not far off. It’s a machine designed to be extremely basic, extremely customizable, and extremely affordable. Those are not your typical design goals, but then the Slate Truck isn’t the fruit of your typical design process. Wander through any automotive design studio anywhere in the world and you’ll inevitably come across a mood board or two, sweeping collages of striking photos meant to align the creative flows of passers-by. They’re a tool for helping a disparate design team to create a cohesive product, but where many such mood boards feature glamour shots of exotic roads and beautiful people, front and center in the Slate’s mood board was something different: a big, gray shark, covered in scrapes and scars.“It looks like a shark that has definitely been in more than one brawl and clearly has come out ahead because it’s still swimming,” says Tisha Johnson, head of design at Slate and who formerly spent a decade at Volvo. That aesthetic, of highlighting rather than hiding battle scars, is key to the Slate ethos.Instead of steel or aluminum, the Slate Truck’s body panels are molded of plastic. Or, as Slate calls them, “injection molded polypropylene composite material.” The theory is that this makes them more durable and scratch-resistant, if only because the lack of paint means they’re one color all the way through. Auto enthusiasts of a certain age will remember the same approach used by the now-defunct Saturn Corporation, a manufacturing technique that never caught on across the industry. Slate continues the theme through to the upholstery, too, a heathered textile that was designed to get better looking as it wears. The idea is to lean into the aged aesthetic. But not everybody will dig the shark theme, and so the Slate Truck is designed to be customizable to a degree never seen before on a production vehicle. Johnson says this is in contrast to the overly curated experience offered by many brands. She says over-curation by automotive designers results in situations like premium, luxury cars that are only available in a palette of disappointingly bland colors: “There’s usually only a fraction that you actually want, and those are always more expensive,” she says.Disparaging other brands for offering limited color choices might seem disingenuous coming from the designer of a vehicle available in a single shade. The Slate Truck, though, was designed to take advantage of the current trend of vinyl-wrapping cars. Its simple shape and minimal trim pieces mean that even amateurs can do the job. Slate will offer DIY kits that newbies can slap on in an afternoon and replace just as quickly based on mood.However, the biggest benefit of this monochromatic thinking might come in production.1/9Bare-Minimum ManufacturingIt’s probably no surprise to you that building cars is expensive. Elon Musk loves to bemoan just how complicated the process can be whenever Tesla is late shipping its next new model, but he’s far from alone in that assessment. What is a little less commonly known is just how expensive it is to paint those cars. Creating a facility that can reliably, quickly, and cleanly lay down a quality coat of color on automotive body parts is a complicated task. That task has only gotten more complicated (and thus expensive) in recent years, with greater environmental regulations and consumer expectations forcing manufacturers to find ways to offer more vibrant hues with less ecological impact. Mercedes-Benz just announced it’s building a “Next Generation Paintshop” at its Sindelfingen plant in Germany, and estimates place the thing’s cost at nearly $1 billion. By eliminating paint, and thus eliminating the paint shop, Slate’s manufacturing process is massively simplified. So, too, the lack of metal body parts. “We have no paint shop, we have no stamping,” says Jeremy Snyder, Slate’s chief commercial officer who formerly led Tesla’s global business efforts. Vehicle factories tend to have high ceilings to make room for the multiple-story stamping machines that form metal body parts. Injection molding of plastic is far easier and cheaper to do in limited spaces — spaces like the factory that Slate has purchased for its manufacturing, reportedly near Indiana. “The vehicle is designed, engineered, and manufactured in the US, with the majority of our supply chain based in the US,” Snyder says. The simplification goes simpler still. Slate will make just one vehicle, in just one trim, in just one color, with everything from bigger battery packs to SUV upgrade kits added on later. “Because we only produce one vehicle in the factory with zero options, we’ve moved all of the complexity out of the factory,” Snyder says.While most buyers will rightly fixate on the cost of the truck, the bigger story here might just be this radically simplified approach to manufacturing. “From the very beginning, our business model has been such that we reach cash flow positivity very shortly after start of production. And so from an investment standpoint, we are far less cash-reliant than any other EV startup that has ever existed, as far as I know,” Snyder says.As Slate tries to dash to production without tripping over the headstones of failed EV startups that litter the countryside, that leanness is key. It’s helped them attract some major investors. “The greatest industry magnates to invest in our company,” Snyder says. He declined to name names, but according to a TechCrunch report, one of those magnates is Jeff Bezos. “We don’t have a direct connection to Amazon,” Snyder clarified, but he didn’t rule out some corporate cooperation. “Who knows? Who knows if you’ll be able to purchase on Amazon? I don’t know.”1/6BYODThose vinyl wraps are literally just the first layer of what Slate’s designers are positioning as a, well, blank slate. They want owners to personalize every aspect of the vehicle, including its silhouette.Need room for more than two passengers? Slate has an SUV upgrade kit that will bolt onto the back of the truck, adding extra rollover crash protection and rear seats with seat belts to match, all in a package that’s easy to install at home. No, this isn’t a Subaru Brat redux. The seats will be forward-facing, and the whole setup is supposed to be strong enough to meet crash test regulations. In fact, Slate’s head of engineering, Eric Keipper, says they’re targeting a 5-Star Safety Rating from the federal government’s New Car Assessment Program. Slate is also aiming for a Top Safety Pick from the Insurance Institute for Highway Safety. This will be, in large part, thanks to a comprehensive active safety system that includes everything from automatic emergency braking with pedestrian detection to automatic high beams.A mandatory part of today’s safety features is a digital rear-view camera. Typically, this view pops up on a modern car’s central infotainment screen, but the Slate doesn’t have one of those. It makes do with just a small display behind the steering wheel as a gauge cluster, which is where that rearview camera will feed. You’ll have physical knobs for controlling the in-cabin temperature controls plus the typical turn stalk and other switchgear, but that’s about it.The truck not only lacks a touchscreen for infotainment duties, it lacks any form of entertainment at all beyond whatever fun you can get from the 201-horsepower, rear-drive configuration. There’s no radio, no Bluetooth, and no speakers of any kind beyond for those required to play basic warning chimes. Many will consider this a cost-cutting step too far, but the interior was designed for ease of upgrading, with easy mounting space for anything from a simple soundbar to a full sound system. There’s an integrated phone mount right on the dashboard, but there’s nothing stopping you from bringing something even larger. I expect the low-cost Android tablet and 3D-printing communities to have a field day coming up with in-car media streaming solutions.The rather extreme omission of any kind of media system in the car is jarring, but it, too, has secondary benefits. “Seventy percent of repeat warranty claims are based on infotainment currently because there’s so much tech in the car that it’s created a very unstable environment in the vehicle,” Snyder says. Eliminating infotainment, the theory goes, necessarily boosts reliability. And reliability will be key because Slate is taking DIY to new extremes on the maintenance front, too.Sales and ServiceThe right to repair your devices is a massively important topic for everyone from smartphone users to smart tractor operators. Traditionally, auto manufacturers haven’t exactly gone out of their way to make DIY maintenance easy, partly because their dealers make so much money hawking cabin air filters and unnecessary coolant flushes.As an EV, the maintenance schedule for Slate Truck should be minimal (most EVs don’t need much more than an annual tire rotation), but for any warranty concerns, the company will encourage users to do the fixes themselves. At least when it’s safe to do so. “If you’re not going to break the vehicle and you’re not going to injure yourself, meaning high voltage, you can do service and warranty service on your vehicle yourself and have the videos and the helpline to support you to do that work,” Snyder says.That support network will be called Slate University and it’ll teach you everything you need to know. Don’t fancy yourself a shade tree mechanic? Or maybe you don’t have a tree to park under in the first place? Slate has a partnership with already-established nationwide service centers, where owners can take their trucks for any needed fixes. Upgrades can be performed here as well, including installing an extended-range battery that will bring the truck’s maximum range up to 240 miles. “At start of production, we will have coverage across the country for servicing your vehicle,” Snyder says. Snyder declined to say who will provide the service, but it seems reasonable to expect something along the lines of a Midas, Monro, Meineke, or perhaps some other nationwide service chain that begins with the letter M.And finally, how can you buy one? It should come as no surprise that Slate will follow Tesla’s footsteps by offering direct sales. No nationwide network of dealerships is planned. Instead, a limited set of pickup centers will pop up as needed based on preorder data. Or, if you don’t mind paying a little more, home delivery will be available.Preorders cost just $50 on Slate’s site, and deliveries are expected to start in late 2026. Slate hasn’t said exactly how much the truck will cost, only that it’ll be less than $20,000 after federal incentives — assuming those incentives are still in place in 18 months’ time.The bigger question, though, is whether consumers will actually be into such a simplified vision of what a car can be. The Slate Truck is a rolling rejection of the current, bloated state of American motoring, but it’s consumer demand that’s driven the market down this dark alley. Are those consumers ready for a rolling digital detox? See More:

Bluesky experienced a major outage on Thursday evening that lasted around an hour. On my desktop browser, my Bluesky feed wouldn’t load at all, just showing a spinning loading circle. I saw something similar with the mobile app; I could log in, but when I did, the feed didn’t show. Other Verge colleagues were also experiencing issues. Now, though, it looks like things have fully recovered. In a 6:55PM ET message on its status page, Bluesky said that that there were “major PDS networking problems” and that “we are investigating a major outage with Bluesky hosted PDS instances.” In a 7:38PM ET message, it said that “we have identified a likely root cause and are rolling out a fix to the Bluesky PDS fleet.” At 7:54PM ET, Bluesky CTO Paul Frazee said that “we are back” and that Bluesky is “still clearing out the situation.” Bluesky’s status page also says that all systems are operational. Okay we are back. Very sorry for the outage everyone. We're still clearing out the situation. Will keep you updated— Paul Frazee (@pfrazee.com) 2025-04-24T23:54:02.277Z Bluesky didn’t immediately reply to a request for comment. In November, Bluesky had some issues with slow-loading feeds and notifications because one of its internet providers had downtime because of a cut fiber cable. That outage happened on what Bluesky said at the time was its “highest traffic day ever.” Updates, April 24th: Bluesky said it was rolling out a fix, and now it appears the site is back to normal.

Legacy search brand Yahoo has been working on its own web browser prototype, and says it would like to buy Google's Chrome if the company is forced by a court to sell it. The information came out during the fourth day of the Justice Department's remedies trial to rectify Google's search monopoly. The DOJ has - among other proposals - requested Judge Amit Mehta break up Google by requiring it sell its Chrome browser, which the agency says is a key distribution channel for its popular search engine that's amassed too much power for anyone else to compete. Yahoo isn't the only company interested in buying Chrome. While DuckDuckGo's CEO said they wouldn't be able to afford it, witnesses from Perplexity and OpenAI both expressed interest in the popular browser on the stand this week. Witnesses from Perplexity and OpenAI both expressed interest in Chrome Yahoo seems to agree that owning a web browser is a very important portal for search, and is looking to run its own, either through developing or buying one. Yahoo Search General Manager Brian Provost testified that about 60% of search queries are done through a web browser; many people search directly from the address bar. That's … Read the full story at The Verge.

Motorola is kicking off hot foldable summer in style — and a little early this year. This time around it’s offering not two but three Razr models with the introduction of the premium Razr Ultra. It comes with some notable upgrades, and all three phones get some quality-of-life improvements like a sturdier hinge. But the hardware updates are otherwise minimal while Moto leans hard into eye-catching colors and finishes for its fun flip phone.Let’s just say it up front: the Ultra costs $1,299. That’s like, Samsung Galaxy Ultra territory, which is just a lot of money for a phone. For that price you get a generous 16GB of RAM and 512GB of storage in the base model and a Snapdragon 8 Elite chipset, which is the Android flagship processor du jour. The Ultra offers up to 30W wireless charging and 68W wired charging. It also has a slightly larger 7-inch inner screen than the Razr and Razr Plus, which gets a little brighter at up to 4,500 nits compared to 3,000 nits. Oh, and there’s a dedicated AI button, which… more on that in a sec.C’mon, this rules.The Ultra shares the same 4-inch outer screen size as the Razr Plus, but the Razr Ultra has some differentiating hardware in the camera department. The main, ultrawide, and selfie cameras each have 50 megapixel sensors. The Razr and Razr Plus also offer 50-megapixel main cameras, but the Ultra’s sensor uses bigger pixels: 2.0μm binned “quad pixels” versus 0.8μm on the other two.Here’s the thing though: wood grain back panel. The Ultra gets some fun new finishes, including a real wood back panel that’s giving Moto X. More wood gadgets, please. There’s also something called Alcantara that you may have seen once upon a time on a Microsoft Surface. It’s a synthetic fabric with kind of a suede feel and comes in a dark green that I really dig, too. The Ultra also comes in a textured deep red finish and a magenta-ish pink that looks at home on the Razr line.All three phones come with an updated hinge that uses titanium rather than stainless steel, and Motorola says it’s four times stronger than the previous design. It also reduces the inner screen crease when the phone is flipped open. I never found the crease too bothersome to start with, but it practically disappears on these new phones, and you can barely feel it under your finger. It’s kind of spooky.You only see the crease when you really go looking for it.All three Razr models come with an IP48 rating this time, like Samsung’s Z Flip 6 foldables. That means they’re fully water-resistant, but dust is still a concern. That “4” rating means the phones are protected against particles bigger than 1mm, but dust is smaller than that, so you’ll still want to be careful if you bring your Razr to the beach. Otherwise, the standard Razr and Razr Plus are pretty minor refreshes. The Razr Plus still uses a Snapdragon 8S Gen 3 chipset like last year’s model, though the Razr gets a bump up from a Dimensity 7300X to a 7400X chipset.Remember that AI button on the Ultra? Well, Motorola is formally introducing several AI features across the Razr lineup, most of which we’ve seen previously in beta. They’re all housed in an interface called Moto AI, which is available to all three Razrs — though only the Ultra has a physical shortcut button to get there. And honestly they look more promising than the usual AI gimmicks — though there are AI gimmicks here, too. I’m particularly interested to see how “Remember this” works. It’s designed in the same vein as Pixel Screenshots and Nothing’s Essential Space, offering a place to keep screenshots, photos, and notes handy without having to go find them in different apps.You can ask Moto AI to “remember” something and then you can ask about it later. You can also prompt it with “Catch me up” to summarize recent notifications that came in while you were busy. Could be handy! Motorola has also announced a partnership with Perplexity, which you can access through Moto AI, and also helps power some predictive suggestions. But as with all AI features right now, we need to see if it actually does what it’s supposed to before getting too excited.1/8Photo: Allison Johnson / The VergeThere’s no wood grain option on the regular Razr or Razr Plus, boooo, but there are plenty of new colors and finishes to choose from. Last year’s mocha mousse is back on the Razr Plus, and the standard Razr comes in a fun minty green that kept catching my eye at Moto’s launch event.The non-Ultra Razrs cost the same as their predecessors — $999 for the 2025 Razr Plus and $699 for the Razr. All three phones will be available for preorder in the US on May 7th and go on sale May 15th.Photography by Allison Johnson / The VergeSee More:

Threads on the web has a new address. Now, you can access the Meta-owned network by heading to Threads.com instead of Threads.net. The move will likely make it easier for users to find the website, as traffic to an unrelated messaging app startup, “Threads.com,” spiked following the platform’s launch. Meta finally acquired the “Threads.com” domain last September, but didn’t start redirecting users to the site until now. Along with this change, Meta is making a few updates to the site, as it will now display your custom feeds in the same order you’ve arranged them on mobile. You can also find your liked and saved posts from the site’s main menu icon instead of pinning a column for them on your homepage. Threads is also testing a new feature where you can upload a list of people you follow on other apps, starting with X, allowing you to find their profiles on Threads. You can now copy Threads posts as images as well, which will let you paste them into other platforms like Instagram. Additionally, the web app is introducing a new icon on the right side of the screen to add columns, along with a new “+” button in the bottom-right corner that you can use to create posts. These aren’t the only major changes Threads has made this week, as the platform began rolling out ads more broadly across 30 countries, including the US. It also started letting users set their following or custom feeds as their default last month.

Adobe has a new tool that makes it easier for creatives to be reliably credited for their work, even if somebody takes a screenshot of it and reposts it across the web. The Content Authenticity web app launching in public beta today allows invisible, tamper-resistant metadata to be embedded into images and photographs to help identify who owns them.The new web app was initially announced in October and builds on Adobe’s Content Credentials attribution system. Artists and creators can attach information directly into their work, including links to their social media accounts, websites, and other attributes that can be used to identify them online. The app can also track the editing history of images, and helps creatives to prevent AI from training on them.For additional security, Adobe’s Content Authenticity app and Behance portfolio platform — which can also be embedded within Content Credentials — allow creators to authenticate their identity via LinkedIn verification. That should make it harder for people to link Content Credentials to fake online profiles, but given LinkedIn isn’t exactly known for its creative community (yet), it’s also a likely dig at X. Then known as Twitter, X was previously one of the founding members behind Adobe’s Content Authenticity Initiative in 2019, before withdrawing from the partnership and transforming its verification system into a paid subscription reward under Elon Musk’s ownership.Notice how X is the only social media not linked and is still referenced as “Twitter” in this example image. Image: AdobeThe Content Authenticity web app is “currently free” while in beta, according to Adobe, though the company hasn’t mentioned if this will change when it becomes generally available. All you need is an Adobe account (which doesn’t require you to have an active Creative Cloud subscription).Any images you want to apply Content Credentials to don’t need to have been edited or created using one of Adobe’s other apps. While Adobe apps like Photoshop can already embed Content Credentials into images, the Content Authenticity web app not only gives users more control over what information to attach, but also enables up to 50 images to be tagged in bulk rather than individually. Only JPEG and PNG files are supported for now, but Adobe says that support for larger files and additional media, including video and audio, is “coming soon.”The “CR” logo seen in these thumbnails will be invisible when published online unless viewed by one of Adobe’s Content Authenticity inspection tools. Image: AdobeCreators can also use the app to apply tags to their work that signal to AI developers that they don’t have permission to use it for AI training. This is far more efficient than opting out with each AI provider directly — which usually requires protections to be applied to each image individually — but there’s no guarantee that these tags will be acknowledged or honored by every AI company.Adobe says it’s working with policymakers and industry partners to “establish effective, creator-friendly opt-out mechanisms powered by Content Credentials.” For now, it’s one protection of many that users can apply to their work to prevent AI models from training on it, alongside systems like Glaze and Nightshade. Andy Parsons, Senior Director of Content Authenticity at Adobe, told The Verge that third-party AI protections are unlikely to interfere with Content Credentials, allowing creatives to apply them to their work harmoniously.The Content Authenticity app isn’t just for creative professionals, however, as it allows anyone to see if images they find online have Content Credentials applied, just like the Content Authenticity extension for Google Chrome that launched last year. The web app’s inspect tool will recover and display Content Credentials even if image hosting platforms have wiped it, alongside editing history where available which can reveal whether generative AI tools were used to make or manipulate the image.The bonus is that the Chrome extension and inspection tool don’t rely on third-party support, making content easy to authenticate on platforms where images are routinely shared without attribution. With increasingly accessible AI editing apps also making manipulations harder to detect, Adobe’s Content Authenticity tools may also help to prevent some people from being misled by convincing online deepfakes.See More:

The first time I met Jon Friedman, corporate vice president of design and research at Microsoft, was in 2019 when he was spearheading a new open approach to design at Microsoft to make its hardware and software look like it came from one company. That’s a largely successful work in progress, so these days Friedman has, like a lot of Microsoft employees, been focused on AI.I spoke to Friedman last week, ahead of the redesigned Microsoft 365 Copilot app announcement, to get a better understanding of the impact AI is having on human designers and creators. “The change is quite dramatic, and the opportunity is quite dramatic,” says Friedman. “The job changed for us. The entire industry and Microsoft were trying to figure out what we can do with this new material that helps adapt computing to people, and it turns out you can do a lot and you can do a lot fast.”In recent months, Friedman’s own job has changed because of AI, so he’s focused on the content as well as its design. ”Suddenly the design job is how do you edit,” says Friedman. “Even my job over the past 6–8 months has become an editor-in-chief job of the product, not just the design leader.”As Microsoft attempts to build personalized AI agents or AI-powered apps, the design system has to adapt to the very real challenges of meeting this vision, and the tools to build that system are constantly changing, too. Whether it’s Adobe apps or Figma, AI features are increasingly being built into creative tools that push designers into an era of editing and using AI as a tool, rather than having to create everything from scratch.The fear and uncertainty around how AI will impact human creativity has resulted in a very real backlash against the use of generative AI in advertising, movies, video games, and elsewhere. While it was initially obvious when an artist or designer used AI to create art, the tools are advancing so quickly now that it’s becoming far more difficult to spot “AI slop.” Microsoft’s visual design team has been experimenting with using generative AI tools as part of their production process since 2022, and the team revealed this week that it recently made a Surface ad with generative AI. Designers created several top-down stills of a Surface device being used in different locations, and then fed them into generative AI tools like Hailuo and Kling to add motion for the video.The result is a 55-second ad for the business versions of the Surface Laptop and Surface Pro, which has been available on YouTube for two months without anyone in the comments spotting Microsoft’s use of generative AI. Microsoft says it turned to AI for the ad due to “an immensely tight deadline alongside various resource and budget constraints.”Friedman, like many working in tech, believes that AI will increasingly be used like this, as another tool for creators and designers instead of something that will simply replace them. He used DALL-E to illustrate his son’s book about managing childhood anxiety last year, and recently used a variety of AI tools to edit videos and create a website. ”As a maker, I loved it. I have this intent, and I don’t have all the capabilities, but I have a lot of the capabilities, and sometimes, I want to sit there and move pixels around, and other times, I want to get a big idea out,” says Friedman. “I think there’s this thing that’s going to happen in the world where disciplines are going to meld together in new ways. A creative maker’s strength is not just going to be in their specific skill set, but their ability to think big and creatively and then take that creative thinking and drive towards better outcomes in the world.”Friedman thinks that engineers will eventually have their own AI designer and designers will have their own AI engineer, so that these two disciplines will overlap in different ways.I’ve been experimenting with a variety of AI tools and models for months, and while I’m a relatively competent writer, I’ve found that AI tools can often help in the early stages of a project or a pitch. I don’t use AI tools to help me write articles, but that’s something I’m skilled at, so I don’t need AI for help there. I have used AI tools to help me edit my personal photos and videos though, because I’m not fully confident at that.I do feel a sense of guilt using AI tools, though. I’m still unsure exactly where I stand in the use of generative AI debate. As a technology enthusiast, it’s exciting to see what AI can unlock, but as a journalist, I’m as nervous about the future of AI as artists and designers are. I’m unsure whether my job will even exist in 10 years. As a young child I also witnessed my father’s career disappear when the print presses rapidly moved to digital. Back then, you either had to re-skill and learn digital imaging tools like Photoshop or get left behind. “I think what we’re going to see is resistance at first because people fear what it means for them on some level,” says Friedman. “I think there’s a little bit of it’s hard to get over the hump of what it means for you, but then once you see it as another tool in your toolbox, it’s really powerful, for both the things you love to do and the things you may not love to do. For me, it supercharged my makership.”I’m hopeful AI will supercharge human creativity, but at the same time Microsoft is preparing for a world of automated AI agents that it dreams of being “digital colleagues.” The line between AI as a tool and AI as a replacement for skilled human jobs looks like it’s about to be tested, leaving us all to wonder what comes next.The pad:Microsoft 365 Copilot redesigned with new search, image, and notebook features. Microsoft is launching a newly designed version of its Microsoft 365 Copilot app soon. The app has mainly been used by businesses as a hub for accessing Office documents, but Microsoft has been gradually pushing people to use its AI tools over the past year. The Microsoft 365 Copilot app will soon have an updated AI-powered search, a new Create feature for generating images using OpenAI’s GPT‑4o model, and Notebooks that operate like miniature AI projects.Bethesda officially reveals Oblivion Remastered, available now. The worst kept secret in gaming was finally made official this week: Oblivion Remastered. After developer Virtuos leaked images of the game earlier this month, The Elder Scrolls IV: Oblivion Remastered got shadowdropped on Tuesday for Xbox Series X / S, PS5, and PC. It’s already proving popular on the Steam charts, as fans of The Elder Scrolls series jump in to play a remastered version of the nearly 20-year-old title. When the world connected on Skype. Microsoft is shutting down its Skype service next month, after more than two decades. Rest of World has put together a list of stories from readers about how Skype helped them stay connected to friends and family. Skype really did help change the way we communicate.Microsoft’s Xbox app is now available on LG smart TVs. If you want to play Xbox games on your LG TV, but you don’t own an Xbox console, there’s now an app for that. The Xbox TV app is now rolling out to select LG smart TV models, including LG’s 2022 OLED TVs, select 2023 smart TVs, and newer models and smart monitors that are running webOS24 or higher. You’ll need an Xbox Game Pass Ultimate subscription, a compatible LG TV, and a Bluetooth controller to start streaming Xbox games to your TV.You can now run Windows 11 on an iPad if you’re in the EU. The Digital Markets Act (DMA) has opened the door to allow emulation apps to run on iPadOS, and YouTuber NTDev has demonstrated how you can now run things like Windows 11 on an M2 iPad Air.Windows for the Game Boy, with a playable version of Minesweeper. RubenRetro has created a version of Windows for Nintendo’s Game Boy. It even has an MS Paint app that works with the Game Boy Printer, a playable version of Minesweeper, and a built-in media player. I love it.Microsoft Copilot can now ‘see’ what’s on your screen in Edge. Microsoft is making Copilot Vision, the AI feature that can interpret what’s on your screen, available free to use within the Edge browser. You can ask Copilot Vision to compare items you’re looking at on a webpage, help with recipes, or even use it for cover letter brainstorming. The more powerful mobile version of Copilot Vision still requires a Copilot Pro subscription, but this screen viewing experience is now available free of charge.Xbox will soon let you purchase games in its mobile app. Xbox is finally rolling out the ability to purchase games and DLC, as well as sign up for Game Pass directly from its mobile app. To be able to offer this, Microsoft has had to remove remote play within the Xbox mobile app, in order to comply with app store restrictions. Microsoft originally wanted to offer both the ability to buy a game and stream it immediately inside its Xbox mobile app for Android, thanks to a court ruling. But Google appealed the ruling, leaving Microsoft to have to adapt its Xbox mobile app while it waits for a final decision.Xbox has some new Doom Slayer accessories. We’re just a few weeks away from Doom: The Dark Ages, and Microsoft is preparing for the launch of this new Xbox game with not one, but two special edition Xbox controllers. You can pick up a regular Xbox controller with the Doom Slayer’s armor for $79.99, or go for a transparent red top Xbox Elite 2 controller for $199.99. There’s even a $54.99 Xbox Series X wrap, too.Microsoft provides a progress report on its security initiative. Security has been the number-one priority for nearly a year now at Microsoft, and the company is now providing an update on its progress with the Secure Future Initiative (SFI). Microsoft is now using a “Secure by Design UX Toolkit” that has been tested by 20 product teams and rolled out to 22,000 employees to help build more secure products. Microsoft also says its latest policies, investigation methods, and behavioral-based detection models have prevented $4 billion in fraud attempts. New Copilot Plus PC features are now being tested. With Recall’s imminent launch, Microsoft is now turning its attention to improving Click to Do on Windows 11 Copilot Plus PCs. Two new text actions are currently being tested, with a reading coach that will help you practice reading out loud and an immersive reader that creates a distraction-free zone around any text you select on your screen.How secure is Microsoft’s Recall feature? Security researcher Kevin Beaumont has been trying out the version of Recall that’s about to be released, and he’s found that it’s greatly improved this time around, but there are still some concerns. Microsoft says Recall requires a fingerprint or facial recognition to use it, but Beaumont found you can open Recall using the four-digit PIN unlock option with Windows Hello. The sensitive data filtering also doesn’t appear to work reliably, which is something I’ve witnessed myself. I’m always keen to hear from readers, so please drop a comment here, or you can reach me at notepad@theverge.com if you want to discuss anything else. If you’ve heard about any of Microsoft’s secret projects, you can reach me via email at notepad@theverge.com or speak to me confidentially on the Signal messaging app, where I’m tomwarren.01. I’m also tomwarren on Telegram, if you’d prefer to chat there.Thanks for subscribing to Notepad.See More:

Adobe has launched two new versions of its text-to-image generative AI model alongside a host of new Firefly features and Creative Cloud app updates coming to Photoshop and Illustrator. The fourth-generation Firefly Image models follow a similar precedent that OpenAI and Google have set for their chatbot releases, providing users with a choice between one model that’s suited for speed and efficiency, and another for more demanding tasks.Adobe says that Firefly Image Model 4 is its “fastest, most controllable, and most realistic Firefly image model yet,” allowing users to generate images in up to 2K resolution with more control over style, format sizes, and camera angles. The updates from its predecessor are designed to improve the quality of image outputs while allowing them to be generated “quickly and efficiently.” For image prompts that require more “detail and realism,” Adobe is also launching Firefly Image Model 4 Ultra, which is more capable of rendering “complex scenes with small structures.”Adobe hasn’t specified which Firefly Image 4 models were used to create these examples, but they are impressive even if the shadows on the child’s face are a little odd. Image: AdobeThe new Firefly image models are generally available now via the Firefly web app alongside Adobe’s text-to-video and text-to-vector models, which previously launched in public beta. A brand new Firefly web platform tool that’s also launching in public beta today is Firefly Boards — a FigJam-like collaborative generative AI moodboarding app that was introduced during Adobe’s Max event in October as “Project Concept.” Adobe also says that a Firefly mobile app is “coming soon” for iOS and Android devices. The Firefly web app now also gives users access to third-party AI models when generating images or video. Users can select between OpenAI’s new GPT image model or Google’s Imagen 3 for images, or Google’s Veo 2 model for video alongside Adobe’s own AI models, with support for Luma, Pika, Runway, fal.ai, and Ideogram “coming soon,” according to Adobe.Notice how Adobe is trying to simultaniously ebrace third-party models while also making its own as distinct as possible. Image: AdobeThe company describes these third-party offerings as being available for “experimentation” rather than publishable work, however, and clearly marks its own models as being “commercially safe.” That’s not terribly surprising given Adobe trains its AI models on public or licenced content, while OpenAI, Google, and Runway can’t claim the same.Adobe is also rolling out a smattering of updates for its popular Creative Cloud apps. Illustrator’s generative shape fill and text to pattern tools are now generally available after being introduced in beta last year, while Photoshop is making it easier for users to make color adjustments, and automatically select details like hair, clothing, and specific facial features. The Actions panel in Photoshop is also being updated to make improved editing suggestions “based on a creator’s unique style,” and will form the foundation of the creative AI agent that Adobe is planning to build into the photo editing platform.See More:

It is our lot in life to poop. We eat stuff, digest it, and then the waste must be expelled. That’s just the way it is, which means we are destined to spend a not-insignificant portion of our time sitting on the porcelain throne. So why not make that as pleasant an experience as possible? And if your toilet seat can make the experience better, cleaner, and more comfortable, then does it not deserve its own throne?Behold the Brondell Swash Thinline T44, the best bidet seat I’ve ever sat on. And for $700, it had better be. 8Verge ScoreBrondell Swash Thinline T44 bidet toilet seat$559$70020% off$559The GoodBest-looking, lowest-profile bidet toilet seat out thereCleans incredibly thoroughlyUltra comfortableThe BadDifficult to change some of the lower-level settingsWish the air dryer dried betterOccasional compatibility issues with Squatty PottyHow we rate and review productsThe Thinline T44 is a bidet toilet seat. These sorts of things have been around since 1980, when the Toto Washlet debuted in Japan, and they’re a lot more convenient than the old standalone bidets they’ve more or less replaced. At its most basic, the bidet toilet seat fits on your toilet bowl, connects to the water supply line, and washes your butt. They can be as cheap and simple as something that blasts you with cold water when you open a valve. The Thinline T44, from San Francisco-based company Brondell, is on the other end of the spectrum.I am not a bidet rookie here. I reviewed one (actually from Brondell) 13 years ago, and I’ve been a convert ever since. I also just spent three weeks traveling around Japan, where nearly every toilet (including those on high-speed trains) has a bidet seat attached. Nearly all of them are made by Toto, and they range from the most basic water sprayers to full toilets with integrated bidets.This is the nicest photo anyone’s ever taken of a bidet seat. Image: BrondellIt’s pretty sleek for a bidet toilet seat. Image: BrondellThe T44 has a heated seat, instantly heated water (temperature adjustable) to clean your backside, a separate jet to clean your front side (a feature not generally recommended by medical professionals), and a hot air dryer, so you need very little toilet paper. The jets can also be repositioned or set to oscillate, and you can adjust the level of water pressure, too. It features a nightlight (three different colors), a deodorizing air filter, and self-cleaning jets. All of this is managed via a sleek remote control, with a cradle you can stick to the wall within easy arm’s reach.None of those features may sound novel if you’re familiar with bidet toilet seats. The difference is that they work better than any I’ve tried in the past, including units from Bio Bidet, Kohler, Tushy, Brondell, and, of course, Toto. RelatedThe T44 has a wide range of pressure settings, from very gentle to stronger than most of us would (or should) choose. The cleaning is very thorough. The seat heats quickly and more evenly, with fewer hotspots compared to older units I’ve tried from Brondell and others. The water temperature stays more consistent than other models I’ve used, too, which occasionally hit you with a chilly stream. The hot air dryer’s fan is stronger and does a better, targeted job of drying your bum (though it’s still a far cry from a Dyson Airblade hand-dryer, and the hairy-butted among us will probably need two or three dry sessions and still require a bit of TP). When I was traveling through Japan, I was surprised to find that the vast majority of bidet seats don’t have air dryers at all, which I think is an essential feature. They were only included on some extremely high-end units, and even on those, the dryers lacked horsepower.The T44 is pretty easy to hook up to water and power. It also has a nightlight. Photo: Brent Rose / The VergeThe thing that really stands out about the T44, though, is how little it stands out. While it still looks like a cyber-toilet from outer space, it’s much thinner than your typical bidet toilet seat at just 3.9 inches high. That’s a lot taller than your standard toilet seat, but much lower-profile than most bidet seats. For example, Toto’s best-selling Washlet C5 is nearly 6 inches tall at the back. This makes the T44 look more like a modern, luxurious toilet seat than a piece of medical equipment. It’s also more comfortable to sit on.Installing the Swash Thinline T44 is easier than you would expect. Simply remove your old seat and close the water supply valve. Flush the toilet to empty the tank and then disconnect the water supply line. Install the included hose with the T-connector between the valve and the tank. Mount the seat on the toilet, connect the hose to the bidet, turn the water back on, check for drips, plug in the power, and you’re good to go. Actually, that is one thing to consider: not all bathrooms have power outlets, mine included. My toilet and shower are separated from the sink and vanity, where the outlets are. I had to get a very thin extension cord and use some guides to keep it tightly pressed to the doorjamb, otherwise my door wouldn’t close. Even with that, I consider the installation simple and worth it.Finally, a toilet seat with a remote control. Image: BrondellThe cleaning wand has instant warm water and several power levels. Image: BrondellThere is room for improvement. While I love that it has two user-preset buttons, setting those modes requires you to use a secret series of button presses you can only find in the manual. Very unintuitive. You need an even more elaborate series of button presses to change the default drying duration. This is where a Bluetooth radio and a simple phone app would be handy, if only for adjusting settings. There are also a couple weird little bugs and quirks. For example, if you select your user profile, it will wash you with all of your desired settings. But if you want a little more washin’, and you hit that button combination again, something gets lost and it will forget to include oscillation or something. I also had some issues while using the T44 with my feet on a Squatty Potty type thing (yes, I know, how many toilet accessories can one man have? Leave me alone!) Something about the angle of my legs or weight distribution would sometimes trip up the occupancy sensor, so I couldn’t get it to wash me until I stood up for a minute or so to allow it to reset. I went back and forth with Brondell’s excellent technical support team, but we never got to the bottom of it. The support team couldn’t replicate it in their testing, so it might just be my weird body or something. The issue became very rare when I moved the mounting bracket forward so the seat was a bit more flush with the bowl’s rim, but it’s something to consider.How many toilet accessories can one man have? Leave me alone!I hope Brondell steals some of the bidet seat features I tried in Japan for the next generation. For instance, a lid that automatically opens when you step into the bathroom. Being able to raise or lower the seat with a push of a button on the remote control. Some were even connected to the flush mechanism; the toilet would flush, and the lid would close when you push a button on the remote. Are these mandatory features? No, but they’re nice to have. These minor gripes aside, I absolutely love this thing. It’s so much more comfortable to use than toilet paper, and you feel much, much cleaner. Not to mention that it’s mostly a hands-free experience, so once you push the button on the remote, you are free to continue doom-scrolling on your phone. While that’s true of most bidet seats, the extra powerful jets make this one work a lot better than any other unit I’ve tested, and it’s easily the best-looking and most comfortable to sit on. Sure, I wish they could turbo-charge the air dryer to reduce my toilet paper usage even more, but, ultimately, the Swash Thinline T44 is the best your butt can get.Happy poopin’.See More:

LinkedIn is expanding its free verification system to the wider web, allowing external sites and platforms to integrate LinkedIn verification rather than building their own tool. Adobe is among the first companies to sign up.Adobe is integrating LinkedIn verification into its new Content Authenticity app and existing Behance portfolio platform, allowing creators who’ve gone through LinkedIn’s verification to display a “Verified on LinkedIn” badge on their profiles. If verified creators use Adobe’s digital Content Credentials tools, their identity will also appear alongside their work whenever it’s shared on LinkedIn.“It’s getting progressively cheaper and easier to pretend you’re someone you’re not online,” Oscar Rodriguez, LinkedIn’s vice president of trust, told The Verge. “You’re also able to do so in a way that looks more credible than ever before. Obviously authenticity is super important for LinkedIn, the platform is founded on this premise of trust.”“Online platforms across the board are facing the same issues around inauthenticity, so we believe that this collaboration with Adobe will be critical in the sense of empowering LinkedIn members and partners to be able to understand specific attributes of someone’s identity that have been verified.”This is how the “Verified on LinkedIn” badge looks on Adobe’s Behance portfolios. Image: LinkedIn / AdobeLinkedIn introduced verification in 2023, allowing users to confirm specific details such as their identity, workplace, or education history using government-issued ID or company emails. The company says that over 80 million people have verified themselves using the tools since then. Alongside Adobe, other early adopters of the expanded verification system include enterprise platforms TrustRadius, G2, and UserTesting.This week social media network Bluesky introduced its own verification system for “authentic and notable” accounts, aping its rival Twitter with a blue checkmark design. Twitter verification was once the de facto standard online — the company even partnered with Adobe on a Content Authenticity Initiative to attach attribution to images — before its verification program was wound down following Elon Musk’s purchase and the checkmark instead became exclusive to paying X Premium subscribers.See More:

Nintendo’s delayed Switch 2 preorders finally kicked off in the US at 12AM ET on Thursday at Best Buy, Target, and Walmart, and almost immediately, there were some big problems. After nearly an hour of frustration, many of us here at The Verge secured preorders, but it seems like preorders are out of stock at the retailers.Target seemed to open first, but trying to order was rough. Many of us were able to get to the checkout screen, but ran into issues from there that stopped a purchase, like being asked to update our addresses over or Target losing our payment details. Some users online who did appear to get Target orders also reported that they were canceled. Although stock appeared to float in and out after preorders were live, Target now says both the console and the bundle with Mario Kart World are out of stock.Walmart’s website, for me for a while, had a vague queue. “This deal is almost gone,” according to the page. But eventually, many of us at The Verge (including me!) were able to snag our Switch 2 preorders from the retailer. Like with Target, Walmart eventually showed a handful of the consoles and bundles in stock here and there before they both sold out.Best Buy, which promised to launch preorders at 12AM ET, showed a gray “coming soon” button on the Switch 2 and Mario Kart World bundle pack for about 30 minutes. As of 12:31AM ET, I was in a line for the bundle, but Best Buy kicked me out of it shortly before 2AM ET and said that the bundle was out of stock. Although I wasn’t able to buy one from Best Buy, I saw many reports online of success there.Nintendo originally announced that preorders would go live in the US on April 9th ahead of a June 5th launch for the console, but the company delayed the preorders over tariff concerns without sharing a date for when they would kick off.The preorder problems follow Nintendo president Shuntaro Furukawa issuing a supply warning in Japan on Wednesday. Nintendo also announced this week that invitations to its direct sales in the US might not be available to everyone ahead of launch.Last week, Nintendo finally announced that preorders would start on April 24th and that the price would still start at $449.99 in the US. Accessories got price hikes. Preorders started earlier this month in some other territories, including the UK.See More:

In a recent update to its website about its direct sales program for the Switch 2, Nintendo says that due to “the very high demand” for the console, “delivery by June 5 is not guaranteed” and that “your invitation email may arrive after the Nintendo Switch 2 launch.” Nintendo has already said that the initial invitations for eligible registrations on its My Nintendo Store website would go out on May 8th. The requirements to be eligible are quite strict, seemingly to thwart scalpers. But now, as it did for Japan yesterday, the company is signaling that US demand for consoles through its own store is quite high. If you were banking on getting a Switch 2 through the platform, you might want to be prepared to potentially wait. In its update, Nintendo suggests visiting “our participating retail partners,” but the preorders that have happened so far, which kicked off at 12AM ET on Thursday at Best Buy, Target, and Walmart, have been messy. Ahead of those preorders, the list of retail partners on Nintendo’s website included Best Buy, Target, Walmart, and GameStop. As I write this, only GameStop, which opens online preorders at 11AM ET on Thursday, has a logo on the page.

Perplexity’s iOS App just got an update enabling support for the company’s conversational AI voice assistant. Now Apple users can activate the assistant in the app and ask it to perform tasks like writing emails, setting reminders, and making dinner reservations.You can even navigate away from the app and continue to speak to Perplexity, although it doesn’t yet support screen sharing like it does on Android. Meanwhile, some of the conversational AI stuff Apple has promised for the Apple Intelligence-powered Siri could still be more than a year away. And unlike Apple Intelligence, Perplexity’s assistant can do these things on older devices, like my iPhone 13 mini.When the Android version launched in January, Perplexity’s spokesperson told The Verge that it would come to iPhones and iPads once “Apple gives us the right permissions,” and apparently that has been worked out. I downloaded Perplexity for the first time on my iPhone today and asked it to set a reminder for me to start cooking at 7PM. I was greeted with a pop-up asking me to give it permission to see my reminders. I accepted, and it added the reminder as I expected. I tried crafting a text message and, as expected, Perplexity asked for access to my contacts. I declined, but I appreciated that the assistant’s next step was to prompt me to just say the intended phone number instead of failing completely.I also tried booking a table at a restaurant, as shown in the video demo posted by Peplexity, and its mobile assistant opened Open Table, then entered the dates and times I said out loud. The actions aren’t end-to-end, though; you’ll still need to complete the process yourself in the window. Perplexity can also open my Uber app and set me up with a ride.Perplexity’s voice assistant has other limitations on iOS. You can’t ask it to look at your camera and “see” what you see for context, as other AI assistants can, like ChatGPT and Grok. But you can still use the standard text-based chatbot to ask questions about a picture. And you can’t ask it to set your iPhone’s scheduled alarms for you — you’ll still need Siri for that.See More:

Following a brief delay stemming from the Trump’s administration’s tariff policy, Nintendo has finally opened up preorders for the Switch 2. Most major retailers — including Walmart, Best Buy, and Target — are now taking reservations, offering gamers a chance to secure Nintendo’s forthcoming console ahead of its release on June 5th. Nintendo is also accepting sign-ups for those looking to secure a Switch 2, though, if you do go that route, be aware that there are strict requirements.RelatedOverall, Nintendo appears to be hopeful that it has built up sufficient inventory of its next-generation portable gaming system to meet US demand — something Nintendo of America President Doug Bowser reiterated during a recent CNBC appearance. Still, there’s no guarantee, especially given that Nintendo has admitted it “cannot fulfill” all the preorders it has already received through its own Japanese store.Either way, if you’re trying to secure a Switch 2, we suggest beginning with the retailers below. We’ve also included a few tips and tricks further down for streamlining the checkout process, if needed.In this articleLanding pages and online start timesNintendo Switch 2$449$449Preorders start at 12AM ET on April 24th / 9PM PT on April 23rd.$449 at Walmart$499 at Walmart (Mart Kart World Bundle)Nintendo Switch 2$450$450Preorders start at 12AM ET on April 24th / 9PM PT on April 23rd.$450 at Best Buy$500 at Best Buy (Mario Kart World bundle)Nintendo Switch 2$450$450Preorders start at 12AM ET on April 24th / 9PM PT on April 23rd.$450 at Target$500 at Target (Mario Kart World bundle)Nintendo Switch 2$450$450Preorders start at 11AM ET / 8AM PT on April 24th.$450 at GameStop$500 at GameStop (Mario Kart World bundle)What Switch 2 configurations are available?At launch, the Nintendo Switch 2 will be available in two distinct configurations: a $449.99 standard version and a $499.99 bundle that includes the latest Mario Kart title, Mario Kart World. Nintendo is on record saying the bundle will only be available through fall 2025 (or while supplies last), which is a shame, considering picking up the console / game combo saves you $30 on what is likely to be the Switch 2’s biggest launch title. If you don’t manage to secure the bundle, you can pick up Mario Kart World separately for a whopping $79.99.It’s also worth noting that prices could change down the line. Nintendo recently raised the price of the Switch 2 Joy-Con and Pro Controller in response to US tariffs, and has said that “other adjustments to the price of any Nintendo product are also possible in the future depending on market conditions.” The starting price of the Nintendo Switch 2 remains unchanged, but at this point, it feels like the only way to guarantee a console at $449.99 is to purchase one at launch.A few essentials tips to keep in mindIf the the rollout of Sony’s PlayStation 5 is any indication, retailers aren’t always well-equipped to handle the massive influx of interest when it comes to console preorders. Inventory for the Switch 2 is likely to be limited as well, but if you’re struggling to get through the checkout process with a console, we have a few pointers:Use mobile apps: If you own a smartphone, try preordering your Switch 2 via the respective app for Walmart, GameStop, Best Buy, etc. In the past, we’ve found that the dedicated mobile apps for various retailers tend to be faster than using a traditional browser; they also provide an additional means of procuring a console, thus increasing your chances of securing a Switch 2 at launch.Skip the accessories: There will be a raft of new accessories for the Switch 2 at launch, and while it might be tempting to stack your cart with add-ons, preordering Nintendo’s first-party peripherals takes time that’s better spent on securing a console. Accessory preorders don’t typically sell out nearly as fast as consoles, either, meaning you can always circle back for the extras.Bookmark this page: We’ll be updating this page for the next several hours, allowing you to see at a glance which retailers are still accepting preorders on either the standalone console or the Mario Kart World bundle. A number of other social media accounts — including Wario64 and Matt Swider of The Shortcut — are likely going to be tracking preorder inventory as well, if you’re looking for additional publications / users to follow.Image: NintendoWhere can I preorder the Switch 2?Right now, the only confirmed retailers for the Nintendo Switch 2 are Walmart, GameStop, Best Buy, and Target. We anticipate Amazon — and Newegg / Sam’s Club? — opening preorders at some point, but based on previous console launches, the rollout at Amazon is likely to be sporadic and unpredictable. Either way, we’ll update this post accordingly if Amazon and other retailers begin offering preorders.WalmartWalmart is now accepting preorders for the standalone Switch 2 ($449) and the Mario Kart World bundle ($499), along with games like Donkey Kong Bananza ($69) and accessories like the Switch 2 Pro Controller ($84). Walmart’s website also indicates that, if you preorder the console before 8AM ET on June 4th, you should receive it before 9AM ET on June 5th — the console’s launch date.Best BuyBest Buy is currently accepting online preorders for the standalone console and the Mario Kart World bundle, and will being accepting in-store preorders at the “start of regular store hours” on April 24th. What’s more, if you do manage to reserve a console through Best Buy, the retailer has announced that most stores in the US will open at 12AM ET, 11PM CT, 10PM MT, and 9PM PT on June 5th for preorder pickup.Additionally, if you’re a My Best Buy Plus or Total member, you can currently score $20 in credit for every $150 you spend (up to $100) on Nintendo games and gear (excluding all Switch and Switch 2 hardware). That privilege will run you at least $49.99 a year, though both membership tiers also grant you access to a number of exclusive perks, including limited-time discounts and free two-day shipping.TargetTarget, like most other major retailers, has opened preorders for both the standalone Switch 2 and the Mario Kart World bundle, along with Nintendo Switch 2 Pro Controller ($84) and the Switch 2 Edition of The Legend of Zelda: Breath of the Wild ($69.99), among other titles.GameStopGameStop is the lone outlier on our list. The retailer confirmed on X that in-store preorders for the Switch 2 will begin when “doors open” on April 24th — not at midnight — with online preorders starting the same day at 11AM ET. The retailer has already published dedicated landing pages for the standalone console and the Mario Kart World bundle, and the company has announced that “all stores” will open on June 5th at 12AM ET / June 4th at 9PM PT for the console’s launch.If you’re looking to bring the cost down, GameStop will be offering up to $175 off a Switch 2 when you trade in a Switch OLED, or up to $125 off when you trade in a regular Switch. In both instances, you’ll need everything that came with your console — specifically the Switch, its dock, two working Joy-Con controllers (with wrist straps), an HDMI cable, and a power cord. GameStop also confirmed to Gizmodo that exact valuation will depend on the condition of your console, including whether your Joy-Cons are experiencing the dreaded stick drift.Update, April 24th: Updated to reflect the fact that Switch 2 preorders have opened.See More:

The Switch 2 isn’t the only hardware launching on June 5th. Nintendo also plans to release a slate of new accessories for its upcoming gaming console, the bulk of which are now available for preorder at Walmart, Best Buy, and Target ahead of their release. From a redesigned Pro Controller and cases to a camera that lets you host video chats with friends, each is designed to level up your gaming experience.RelatedSadly, President Trump’s on-again, off-again tariff policy has resulted in slight price bumps on a number of Switch 2 accessories — including Nintendo’s latest Joy-Con controllers. There’s no telling whether we can expect more price hikes down the line, either, meaning if you do manage to preorder a Switch 2 and want to pick up a second controller or another accessory to go with it, now might be the time to act.$84$94Joy-Con 2 Charging Grip$40$40Nintendo Switch 2 Camera$54$54Where to preorder controllersJoy-Con 2 controllersNintendo’s Joy-Con controllers are bigger than the original and snap onto the the Switch 2 magnetically, eliminating the need for sliding rails. The right Joy-Con also features the new “C” button, which you can use to pull up Nintendo’s new in-game chat system, GameChat, which will be free until March 31st, 2026. While the controllers don’t feature Hall effect joysticks, Nintendo says they’ve been redesigned for smoother movement. The Joy-Con 2s also feature new optical sensors, which let you them like mice in compatible games.The Switch 2 comes with a pair of Joy-Cons, but if you want a second pair, you can currently grab them for around $94 at Walmart, Best Buy, and Target. If the initial batch of preorders do manage to sell out, GameStop will be opening online preorders at 11AM ET on April 24th.$94Nintendo Switch 2 Pro ControllerThe removable Joy-Cons are great for portable and wireless play, but their small, curveless design can feel cramped during longer sessions if you have larger hands. That’s why the Nintendo Switch 2 Pro controller might be worth a look, especially if you mostly game in TV mode. The upgraded version of one of our favorite Nintendo Switch controller features a few welcome improvements, including a pair of rear GL and GR buttons players can use to remap other controls. Nintendo has also finally added a 3.5mm audio jack, allowing you to pair it with a headset or a set of wired headphones. The gamepad even features the aforementioned “C” button.You can preorder the Nintendo Switch 2 Pro controller for around $84 from Walmart, Best Buy, and Target. GameStop preorders, meanwhile, are slated to kick off online at 11AM ET on April 24th.$84Joy-Con 2 Charging GripUnlike the standard Joy-Con 2 grip that comes with the Switch 2, Nintendo’s so-called Charging Grip features customizable GL and GR buttons — just like the Switch 2 Pro Controller — and can be used to charge your Joy-Cons via an included USB-C cable. Otherwise, it appears pretty similar in functionality of the original Switch’s charging grip, allowing you to combine the Joy-Con 2 controllers for a more traditional gamepad.You can preorder the Joy-Con 2 Charging Grip starting at $39.88 from Walmart, Best Buy, and Target — or from GameStop, if you don’t mind holding off until online preorders open at 11AM ET on April 24th.Joy-Con 2 Charging Grip$40$40Joy-Con 2 Wheel (set of two)If you plan on picking up the new Mario Kart World, the Joy-Con 2 Wheels are worth a look. Simply snap the Joy-Con 2 controllers, which connect magnetically, into the steering wheel attachments before booting up your preferred racing title. The two-wheel set is currently available for preorder for around $24.88 from Walmart, Best Buy, and Target, with GameStop preorders opening at 11AM ET on April 24th.$25Where to preorder casesNintendo Switch 2 Carrying Case and Screen ProtectorThe Switch 2 is a pricey gadget you’ll want to keep safe, so it’s a good idea to invest in a carrying case like the Nintendo Switch 2 Carrying Case & Screen Protector. We’ll likely see third-party brands release their own version in the coming weeks, but Nintendo’s proprietary offering is likely to be one of the few available at launch.The bundle — which includes a console-only carrying case, screen protector, and cleaning cloth — is available for preorder for starting at $39.88 from Walmart, Best Buy, and Target. Like with other accessories, GameStop says it will open online preorders at 11AM ET on April 24th, in case you miss out on the initial preorder window.$40Nintendo Switch 2 All-In-One Carrying CaseNintendo is also launching an All-In-One Carrying Case on June 5th, which is more spacious than the model mentioned above. The higher-end case features separate compartments for stowing the console and various goodies, including the Switch 2 dock, Joy-Con 2 controllers, cables, and up to six game cards. You can preorder it now from Walmart, Best Buy, and Target for around $84 (or at GameStop starting at 11AM ET on April 24th).Where to preorder the Switch 2 CameraSupposedly, you can video chat with friends via GameChat using any compatible USB-C camera, but Nintendo offers its own solution in the form of the Nintendo Switch 2 Camera. The 1080p camera is designed to sit on a stand, allowing you to display your face as a picture-in-picture overlay or as a dedicated video feed. A few titles, such as the upgraded version of Super Mario Party Jamboree, let you see your friends on-screen while you play. Some mini-games even incorporate the camera into the game experience itself.The camera features an adjustable wide-angle lens, designed to capture groups of people in a single shot, as well as a “high-sensitivity image sensor” that automatically adjusts brightness levels. It also features a built-in privacy shutter and includes a USB-C charging cable, as well as support for TV, tabletop, and handheld modes.You can currently preorder the it for around $54 from Walmart, Best Buy, and Target. If you strike out with any of the aforementioned retailers, it will also be available for preorder online at GameStop starting at 11AM ET on April 24th.See More:

Anker’s new all-in-one Nebula X1 smart projector has everything you’d want from a “cinema-grade outdoor entertainment system.”It produces a bright 4K image from a triple-laser light engine capable of 3,500 ANSI lumens, automatically finds the best place to project an image optimized for the environment, has Google TV built-in to unlock 4K Netflix streams, and features two battery-powered Wi-Fi satellite speakers to create a legitimate surround-sound experience. The only thing missing is a portable screen.The X1 can project a 200-inch image from a distance of 13 to 22 feet thanks to its optical zoom. It’s also fitted with a gimbal that can tilt the output up to 25 degrees to automatically find the best placement on a screen or wall where the image is keystone corrected, focused, and adapted to the ambient light and surface color. A button on the remote control can initiate all that spatial adaptation whenever you want and the X1’s memory feature can preserve settings for repeat locations. And its 5,000:1 native contrast and Rec.2020 color support should help make the X1’s Dolby Vision certification more than just a marketing stunt.1/5It’s the X1’s sound system, however, that’s the real star of this show. The projector includes four built-in speakers (two 15W and two 5W). Those speakers can be combined with a pair of optional battery-powered Wi-Fi speakers good for up to eight hours of audio. Each satellite speaker contains a 40W pair of front drivers, a 20W upward-firing driver, and a 20W side-firing driver to add width and height to audio. The result is a 4.1.2 surround-sound experience when the audio drivers inside the X1 are switched into subwoofer mode. And while the satellite speakers feature IP54 resistance to dust and rain, the X1 projector itself offers no such assurances.Lastly, the X1 is also liquid cooled — an industry first, according to Anker. That not only reduces the size of the unit by omitting the high-powered fans, it also drops the noise to a relatively quiet 26dB. There’s even the option for a two-pack of wireless microphones to get your karaoke on.Anker’s Nebula X1 won’t come cheap. The projector alone will cost $2,999 starting on May 21st, or $3,298 for the X1 bundle, which includes the projector, two mics, a travel case, and a pair of those satellite speakers.See More:

Over the past 24 hours, the auto industry has experienced some of the most extreme whiplash in the saga of the Trump administration’s tariffs, ending on Wednesday night with two contradicting policy proposals coming out of the White House: China may be granted exemptions on auto part tariffs, but Canada’s car tariffs might increase.Last night, the Wall Street Journal reported that Trump was considering slashing his 145 percent tariffs on China, lowering some of them potentially to 50 percent – a report that may have assuaged rattled investors. It seemed especially credible given Trump himself hinted at decreases during a press event earlier that day, saying: “145% is too high. It will come down substantially.” Yet Treasury Secretary Scott Bessent denied the WSJ report the next morning, saying the U.S. would not lower tariffs unilaterally. “This is the equivalent of an embargo, and a break between the two countries in trade does not suit anyone’s interests,” he told reporters. Within hours, however, the Financial Times reported that Trump was indeed planning to eliminate his recently-imposed tariffs on steel, aluminum, and car parts imported from China, and the White House confirmed to CNBC shortly thereafter that some unilateral exemptions were indeed under consideration. While not a complete reversal — a 25 percent tariff on foreign-made cars and a 25 percent tariff on all imported car parts would still be intact — it would have offered some relief to carmakers, who faced the possibility of absorbing the cost of multiple tariffs stacked on top of each other.Alas, more confusion ensued. Shortly after the FT report was published, and automotive stocks started trending upwards from the news, Trump told reporters in the Oval Office that Canada — not China — might see auto tariffs further increase. “They took a large percentage of the carmaking, and I want to bring it back to this country,” he said. “I really don’t want cars from Canada. So when I put tariffs on Canada — they’re paying 25 percent, but that could go up in terms of cars — when we put tariffs on, all we’re doing is saying, ‘We don’t want your cars, in all due respect, we want really to make our own cars,’ which is what we’re doing in record numbers.” RelatedThe chaotic jumble on auto tariffs is the latest incidence of the Trump administration vacillating on who they are tariffing, what they are tariffing, and how much those tariffs are. But even if new proposed exemptions are “destacked” from the current tariffs, as officials characterized it to the Financial Times, the tariffs in their existing form threaten to devastate the American auto industry. In a letter sent to the administration on Tuesday, a coalition of powerful U.S. auto industry players cited a Center for Automotive Research report which estimated that a 25 percent auto tariff would increase costs to the industry by up to $107 billion. “Tariffs on auto parts will scramble the global automotive supply chain and set off a domino effect that will lead to higher auto prices for consumers, lower sales at dealerships and will make servicing and repairing vehicles both more expensive and less predictable,” the coalition wrote.See More:

Nvidia is updating its G-Assist AI assistant on Windows to take it beyond optimizing game and system settings. G-Assist originally launched last month as a chatbot primarily focused on improving PC gaming, but it’s now getting plugin support so you can extend the AI assistant to control Spotify, check if a streamer is live on Twitch, and look at stock or weather updates. The new ChatGPT-based G-Assist plugin builder lets developers and enthusiasts create custom functionality for Nvidia’s AI assistant. G-Assist will be able to connect to external tools and use APIs to expand the capabilities of what Nvidia offers right now. Nvidia has published sample plugins on GitHub that can be compiled and used by G-Assist: Spotify — hands-free music and volume control Google Gemini — allows G-Assist to invoke Gemini for cloud-based complex conversations Twitch —  you can use this plugin to checks if a streamer is live with voice commands like, “Hey, Twitch, is [streamer] live?”  Peripheral Controls — change RGB lighting or fan speed on Logitech G, Corsair, MSI and Nanoleaf devices Stock Checker —  provides real-time stock prices Weather Updates — provides current weather conditions in any city These plugins all run locally using a small language model on Nvidia’s RTX GPUs, and developers will also be able to share their own custom plugins through GitHub. G-Assist uses a local small language model that requires nearly 10GB of space for the assistant functions and voice capabilities. The AI assistant works on a variety of RTX 30-, 40-, and 50-series desktop GPUs, but you’ll need a card with at least 12GB of VRAM. If you’re interested in trying out G-Assist or building a plugin, the app is available as an optional part of Nvidia’s main app for Windows.

Perplexity Chief Business Officer Dmitry Shevelenko said he didn’t want to testify in a trial about how to resolve Google’s search monopoly because he feared retribution from Google. But after being subpoenaed to appear in court, he seized the moment to pitch a business opportunity for his AI company: buying Chrome.If Judge Amit Mehta sees things the way the Justice Department does, he could force Google to spin out its popular web browser — including the free open source Chromium browser that many other web browsers are built on. Google says this remedy is playing with fire, and could result in a new Chromium owner charging for the product or failing to keep it running in an adequate way, causing ripple effects across the browser industry.But Perplexity would be happy to take on the project, Shevelenko testified. When an attorney asked if Perplexity believes anyone besides Google could run a browser at the scale of Chrome without diminishing its quality or charging for it, Shevelenko responded, “I think we could do it.”It’s not the first time the AI upstart has put its name in the ring to take over a major technology platform under siege by the US government. The less than three-year-old company also thinks it could buy TikTok, which is facing an ever-extending ban in the US over national security concerns based on its ownership by China-based ByteDance.Although Shevelenko wasn’t a voluntary witness — he was called by the DOJ to demonstrate how Google’s search monopoly affects new generative AI companies — he seemed prepared to lay out his grievances against Google. He described the “jungle gym” of Android settings a user needs to navigate to set Perplexity as their default AI assistant, and admitted he even had to call on a colleague to help him do it (Shevelenko also admitted he’s a long-time iPhone user). Even once Perplexity’s assistant is set as the default over Google’s on an Android phone, he testified, it still doesn’t have the same standing as Google’s, since a user needs to press a button to activate it, rather than using a wake word like “Hey Google.”Companies under contract with Google essentially have a “gun to [their] head”During the DOJ’s questioning, Shevelenko walked through an anonymized list of phone makers that Perplexity has spoken with about a deal to make it a preloaded search engine or default assistant in the US, but failed to strike a bargain with. In talks with one company whose name was hidden from the public, Shevelenko said the discussion kept “hitting a wall because they were fearful of losing Google revenue share or Google agreements.” As it stands, Shevelenko testified, companies under contract with Google essentially have a “gun to [their] head” because Google can turn off significant revenue share if they do something Google dislikes.Mehta found in the first phase of the case that Google used exclusionary agreements with phone and browser companies to lock up distribution channels, and many of those agreements prevent the companies from making certain kinds of deals that could displace Google’s status on their products. Shevelenko said another unnamed company — which based on public reporting appeared to be Motorola — agreed to preload Perplexity on its devices, but would not make it the default assistant “despite both parties wanting it to be.” He testified that Perplexity and the phone-maker “tried every creative workaround” to evade Google’s restrictions, but even though the company thought Perplexity’s assistant would be “great for their users,” they just couldn’t find a way to “get out of their Google obligations” and dislodge it as the default.Shevelenko credited the judge’s monopoly ruling as the reason it could enter these talks at all. The fact that Google is “under pressure,” he said, is why wireless carriers, phone manufacturers, and browser makers are comfortable even discussing the possibility of a deal. But despite his opportunistic pitch, Shevelenko seemed ambivalent of Google being actually forced to sell Chrome. Ahead of his testimony, he shared a post from the Perplexity team on LinkedIn that argues Google should not “be broken up.” His company seems more concerned with ending Google’s search distribution agreements. And in court, Shevelenko warned it would be concerning to see a company like OpenAI buy Chrome (as one of its executives testified Tuesday it would be open to) and discontinue Chromium’s open source model or fail to adequately support the product. “There’s all the self-serving incentive to be here today and shout about how evil Google is, and I think we want to be reasonable,” he said. Google, he continued, builds good products that others are able to iterate on. “We wouldn’t want a remedy that cripples Google’s ability to keep doing that.”See More:

Senator Elizabeth Warren (D-MA) sent Apple CEO Tim Cook a letter asking for information regarding Cook and Apple’s conversations with President Donald Trump’s administration about tariffs and any special exemptions from the tariffs, Bloomberg reports. Warren’s letter, obtained by Bloomberg, points out that Apple “appeared to have invested heavily in influencing the Trump Administration” through things like Cook’s $1 million donation to Trump’s inauguration and Cook’s presence at the inauguration. It discusses The Washington Post’s recent report about Cook’s discussions with White House officials, including a reported conversation with Commerce Secretary Howard Lutnick about how tariffs could impact the price of iPhones. It also notes Trump’s tariff exemptions for smartphones and other electronics and quotes WaPo saying that “Apple appears to be the only company that benefits from virtually all of [them].” Warren’s letter concludes with a list of questions for Apple, which includes inquiries on things like the nature of Cook’s discussions with the Trump administration, when Apple learned about the exemptions, and asking about the “the value of the additional profits that Apple will obtain as a result of President Trump’s tariff exemption.”  The letter does not include a deadline for Cook to respond. Apple didn’t immediately reply to a request for comment from The Verge.

Microsoft is launching a newly designed version of its Microsoft 365 Copilot app today. The app, which is used by businesses as a hub for Office documents and now Microsoft’s AI tools, is moving a little closer to the regular Copilot consumer features with updated AI-powered search, a new Create feature for generating images using OpenAI’s GPT‑4o model, and Notebooks that operate like miniature AI projects.While Microsoft 365 Copilot won’t have the slick-looking home interface found on the consumer version, it now defaults to the chat interface, which is starting to look similar between both versions. The chat interface now has memory and personalization, much like the consumer version of Copilot, so that it starts to build up an understanding of your work and preferences to respond accordingly.The new Microsoft 365 Copilot interface now starts with chat first. Image: Microsoft“We re-built the app, Microsoft 365 Copilot, from the ground up,” explains Jon Friedman, corporate vice president of design and research at Microsoft, in an interview with The Verge. “We’re setting the stage for the next wave of computing, and that stage is this idea that Microsoft 365 Copilot will begin to truly understand you through its memory, its understanding, and that it will start to bring you the right tools that you need.”Those tools include a new AI-powered enterprise search that can access third-party apps like ServiceNow, Google Drive, Slack, Confluence, and Jira to give businesses more relevant search results even if they don’t use everything Microsoft has to offer.The Microsoft 365 Copilot app shares the same styling as the consumer version now, and the Pages feature, which first debuted in the consumer app, is also available to businesses. Pages is a powerful way to work collaboratively with AI agents, as you can share Pages with colleagues and they can work on them as if they’re a co-edited Word document. “We took the same model as Loop has, it works in Outlook, in Teams, and everywhere in the ecosystem,” says Friedman.The new Notebooks feature lets you focus Copilot on individual projects. Image: MicrosoftMicrosoft has also added Notebooks to Microsoft 365 Copilot. “Notebooks are project-based notebooks that allow you to organize information around a set of data in the Office Graph,” explains Friedman. Think of Notebooks as a collection of pages, links, and files that you can share with Copilot so it’s focused directly on answering questions about a specific project.There’s also a new Copilot Create section of the app, which is essentially the evolution of Microsoft’s Designer product which lets you generate images, videos, surveys, and other content for Office documents. It utilizes OpenAI’s GPT‑4o model, so if you really want AI-generated Ghibli art in your PowerPoint presentation then that’s now possible.The Create section lets you generate images using OpenAI’s latest GPT‑4o model. Image: MicrosoftAt the heart of this updated app is a new agent store, where you can easily access AI agents from Microsoft or third parties. The researcher and analyst agents will be part of this store, and any AI agents will be immediately available from the left pane. “We’re trying to optimize the entire left pane for content, rather than applications,” explains Friedman.The consumer and commercial Copilot apps are still not all that similar yet, but I get the impression from Friedman that there needs to be some separation between the two. “In consumer you earn the right to go pretty far and be really personal, at work with all of the professionalism that exists there’s a healthy tension,” says Friedman. “In my mind Copilot can be one branded experience, and it goes from warm and personal to performant and professional, and that’s what we’ve been working on together.”Microsoft lays out its vision of a new Frontier Firm that puts AI first. Image: MicrosoftMicrosoft is also launching this refreshed Copilot for businesses at the same time it reveals its annual work trend index, that’s based on a survey of businesses, Microsoft 365 telemetry, and LinkedIn hiring and labor trends. Microsoft believes its data shows that there’s a new kind of organization emerging, what it calls the “Frontier Firm” that is built around AI tools and and humans becoming the bosses of AI agents.“As agents increasingly join the workforce, we’ll see the rise of the agent boss: someone who builds,delegates to, and manages agents to amplify their impact and take control of their career in the age ofAI,“ says Jared Spataro, chief marketing office of AI at work at Microsoft. ”From the boardroom to the frontline, every worker will need to think like the CEO of an agent-powered startup.“An “agent boss” certainly sounds like the latest version of an AI prompt engineer, but Microsoft also warns that businesses will have to look at how many AI agents are needed and “how many humans are needed to guide them.” Microsoft has been pitching AI agents as a way to automate tasks, but as these get increasingly more complex and capable, how long until we see the rise of even more intelligent agents that don’t need humans to boss them around?See More:

OpenAI’s upgraded image generator in ChatGPT brought a surge of users to the AI service thanks to its ability to create Studio Ghibli-style images and really dull dolls, and now it’s coming to other apps. The company says the same “natively multimodal model” powering the image generator will be accessible in its API via “gpt-image-1,” according to a blog post, and some major names have already signed up to use it.“The model’s versatility allows it to create images across diverse styles, faithfully follow custom guidelines, leverage world knowledge, and accurately render text – unlocking countless practical applications across multiple domains,” OpenAI says.Companies like Adobe and Figma are already incorporating the model into their tools. Here’s how, per the blog post:Adobe’s leading ecosystem of creative tools including its Firefly and Express apps will provide access to OpenAI’s image generation capabilities, giving creators the choice and flexibility to experiment with different aesthetic styles – something business professionals, consumers and creators all value when generating new creative ideas.Figma is leveraging the latest model to bring advanced image generation and editing capabilities across its platform. Rolling out starting today, users can use ‘gpt-image-1’ in Figma Design to generate and edit images from simple prompt – adjusting styles, adding or removing objects, expanding backgrounds, and more. This new integration lets designers rapidly explore ideas and iterate visually, all in Figma.OpenAI says that it’s also “continuing to work with developers and businesses to uncover more ways image generation in the API can serve their use cases,” including with Canva, GoDaddy, and Instacart.The “gpt-image-1” model will initially be available via OpenAI’s Images API, and the company says support for the Responses API is “coming soon.”See More:

WhatsApp is launching a new “Advanced Chat Privacy” feature that aims to prevent people from taking conversations outside the app. When the setting is turned on, you can block others from exporting your chat history and automatically downloading photos and videos sent in the app. The feature will prevent people from using messages for Meta AI as well, which you can currently use to ask questions within a chat and generate images. By default, WhatsApp saves photos and videos in a chat to your phone’s local storage. It also lets you and your recipients export chats (with or without media) to your messages, email, or notes app. The Advanced Chat Privacy setting will prevent this in group and individual chats. This feature still doesn’t stop people from taking screenshots of your messages or manually downloading media from chats, WhatsApp spokesperson Zade Alsawah confirmed to The Verge. However, WhatsApp says this is its “first version” of the feature, and that it plans to add more protections down the line. “We think this feature is best used when talking with groups where you may not know everyone closely but are nevertheless sensitive in nature,” WhatsApp says in its announcement. WABetaInfo first spotted this feature earlier this month, and now it’s rolling out to the latest version of the app. You can turn on the setting by tapping the name of your chat and selecting Advanced Chat Privacy. Update, April 23rd: Added information from WhatsApp.

I find something weirdly captivating about tech products that come with transparent cases that let you see the works inside. We all know that there are all sorts of chips and wires in there that make the product do what it does, but unless we plan to break the thing open (and void the warranty), those items are usually hidden. If you also find this kind of thing fascinating, then Sharge’s charging gear might be right up your alley — and you may want to try out the Shargeek 170 Power Bank, which is currently available from Amazon (for Prime members) and Sharge for $99, a savings of a $100.Shargeek 170 power bank$99$19950% off$99A see-through case protects a 24,000mAh battery, 170W maximum output, and the ability to charge three devices simultaneously via a pair of 140W USB-C ports and a single USB-A port.$99 at Amazon (with Prime)$99 at SchargeThe Shargeek 170 doesn’t just look good, either. It features a 24,000mAh battery, 170W maximum output, and can charge three devices simultaneously via a pair of 140W USB-C ports and a single USB-A port. Meanwhile, a smart display keeps you up to date on recharging time and percentage, along with the Shargeek’s current output and input. The Verge’s Sean Hollister saw it at CES last year, and thought it looked “pretty sweet.” You might, too.RelatedMore sales to play and clean withA real classic, The Elder Scrolls IV: Oblivion Remastered is now available, and if you’ve always wanted to try out the 20-year-old game — but were reluctant to use an old game on new tech — now is your chance. The remastered version features overhauled graphics optimized for modern hardware and includes both the Knights of the Nine and Shivering Isles expansions. And it’s on sale at Fanatical: you can get the Standard PC version for $41.49 (about $9 off), while the Deluxe edition — which comes with new quests, a digital artbook, and a soundtrack app — can be had for $49.79 ($10 off).Let me be really honest here: I tend to eat at my desk a lot. That’s why my keyboard — and desktop — is often littered with crumbs, and why I could probably use something like the Hoto Compressed Air Capsule. The handheld vacuum cleaner provides 15,000Pa of suction power and includes five attachments, allowing it to handle crumbs, dust, and other detritus in tight corners. It can even be used to clean your coffee grinder! It’s now on sale at Amazon for $57.99 ($52 off), which, although not as low as its recent price of $31, is still a reasonable drop.If you’re in the market for the latest iPad with Apple’s A16 chip, the 11-inch base configuration is matching its all-time low of $319.99 at Amazon in blue and yellow, down around $29 from its usual price of $349. If color is important to you, you can also get it in silver for $327 and in pink for $339. Read our review.See More:

Tech evangelists have been yammering about “working smarter, not harder” for years. Now, two 21-year-old Columbia University dropouts are proposing a new $5.3 million twist on the concept: use their AI tool Cluely to “cheat on everything.”That’s what it literally says in Cluely’s online manifesto: “We want to cheat on everything.” Unlike the AI chatbots you’re familiar with, it describes Cluely as an “undetectable AI-powered assistant built for virtual meetings, sales calls, and more.” It claims to read your screen, listen to your audio, and let you discreetly prompt AI to find answers or whip out smart responses in real-time. Basically, the next time you’re in a team meeting, job interview, sales call, or online test, Cluely promises you’ll come off smarter thanks to AI — and no one will be the wiser.“Imagine you’re trying to sell someone something and you got this tool that knows every single detail about them, their professional lives, about you, and about your company. It’s as if you’ve done 10 hours of research and all of the sudden, every single question they ask, every single objection they face — you immediately have an answer,” Cluely cofounder Chungin “Roy” Lee tells me in a video call. Lee describes it as “true AI maximalism,” where in every possible use case AI can be helpful it is. Lee recently went viral for cheating his way to an Amazon internship with his last project, Interview Coder. Similar to Cluely, Interview Coder was pitched as an invisible app that helps programmers secretly use AI chatbots on technical tests in job interviews. Not only did Lee document and post the entire process, the stunt led to him getting suspended from Columbia. (He and his cofounder Neel Shamugan decided to drop out after disciplinary proceedings.)“The video was like a launch of our vision, not a launch of the product.”It’s a wild story. Even wilder is the six-figure ad Cluely dropped over the weekend. Lee stars in the ad, using Cluely to catfish his date into thinking he’s a 30-year-old senior software engineer. He can see an AR display that analyzes her speech in real time while providing visual references to his own dating profile and answers to her questions. When his date catches on to the ruse, Cluely tries to salvage the situation in real-time as if it were an AI Cyrano de Bergerac. It hints he should reference her artwork and quickly generates a script to convince her that despite the lies, he’s worth a second shot. This Black Mirror-esque ad is Lee’s elevator pitch for what “cheating on everything” looks like. After all, why stop at technical interviews when you could have an AI wingman?Apologies for the crappy photos but this doesn’t show up in screenshots. Image: Victoria Song / The VergeI’m a journalist. My job is asking smart people smart questions. Why not try “cheating” with Cluely to become a better interviewer? Who better to test this hypothesis on than Lee himself?Hopping onto a Zoom call with Lee, Cluely doesn’t work like I’d imagined. In the ad, Cluely works like magic. It instantly understands the situational context and the user doesn’t have to do anything. In reality, we spend the first couple minutes troubleshooting Cluely-related audio problems. The AI can’t intuit what I need to know even though I gave it some context ahead of the call. There’s no being discreet when you have to type prompts with a clacky mechanical keyboard. The few times I try, it’s obvious my eyes are wandering to the side of my screen. And whenever I shoot off a prompt, the AI takes forever to generate a response. These are all flaws that Lee acknowledges. “Right now the product is in its earliest possible stages. This is a bit more than a proof of concept that was developed in a few weeks,” Lee says. “The video was like a launch of our vision, not a launch of the product.” The problem with AI has never been a lack of vision. The fine print is in the execution. Poor execution almost always shatters the illusion of whatever future tech founders are peddling. Cluely is no exception. When I show my spouse Cluely, they lift a quizzical brow and ask, “Why not just use Google?” “The reason to use AI over Google is pretty obvious. AI will just give you better answers than Google does, and if people don’t think that, then they should just use Google,” says Lee. It’s a reasonable answer, if, like in the story of Cyrano, your AI pal is always smarter, faster, and wittier than you. But what if it isn’t? What if it’s boring, slow, or worse than you at comprehension? This isn’t a bad pitch but in our newsroom, I know my editors would push me to go for a more unique angle. Image: Victoria Song / The VergeI tried using Cluely with my editor and during one of my actual team meetings. Neither went smoothly.With my editor, I had many of the same technical problems, albeit the latency is less of an issue in a relaxed conversation about shared interests. She asked me what I thought of K-pop group BlackPink’s solo careers — particularly Jennie’s recent performance at Coachella. Thankfully, that’s a topic I have many thoughts on but I prompted Cluely anyway. It spat out a generic, stiffly-worded answer about how it’s awesome to watch a celebrity express themselves creatively 90 seconds after I’d already shared my true opinion. That’s an eternity of silence in an interview. In my meeting, I had to ask my colleagues if they’d be okay with me using Cluely beforehand. Cheating, by definition, requires subterfuge — something that Cluely’s own terms of service and privacy policy frown upon. Due to recording consent laws, Cluely says you should ask for consent of parties present because to do so otherwise could be illegal. That feels like pulling back the curtain on the Wizard of Oz, not to mention, defeating the purpose of “cheating.” Do I sound smarter if people know there’s a chance it’s AI-generated thoughts coming out of my mouth? On the meetings call, Cluely seemed to cause mic issues resulting in lots of audio feedback. My colleagues asked me multiple times to mute myself. (All the audio problems disappeared once I stopped Cluely.) It’s hard to look smart when the AI can take two whole minutes to digest a conversation, you get distracted by four errors that pop up, and everyone shushes you because of messed-up audio. There’s a future in which a faster, smarter AI could be everyone’s personal Cyrano. For what it’s worth, Lee doesn’t see AI or Cluely’s mission quite in that way. Cheating is the metaphor because AI, Lee says, will inevitably become so powerful, using it will feel like cheating. He’s convinced that “AI is the lever that will let us experience the true extent of our humanity” by cutting out tedium and letting us pursue whatever it is we actually want to do. It’s an idea AI evangelists frequently preach. But that’s not where we are today. While testing Cluely, I put a lot of effort into making it work for me. I’d ended up working harder to be worse at my job than I usually am. I wondered, wouldn’t it have been easier to simply not cheat?See More:

Looking back, the original idea behind YouTube seems almost quaint. The mythic founding story goes like this: in January of 2005, two PayPal employees, Chad Hurley and Steve Chen, were at a party. People were taking photos and videos on their digital cameras. Sharing photos was easy, sharing video was anything but. “People have different video types, video codecs they have to download, video software,” Chen said on Charlie Rose in 2006. He and Hurley had a sense that as digital cameras and cameraphones became ubiquitous, more people were going to want to share their footage. “We tried to simplify the process, to make it as easy as possible to share these videos online.” By the end of the year, their simple platform was already a huge hit.Fast forward to today, the 20th anniversary of the first-ever YouTube video upload, and the numbers have become so big they’re basically meaningless. Three and a half billion people watch YouTube every month, according to one study. Google’s earnings show it brought in about $36 billion in ad revenue alone last year. YouTube gets 50 percent more viewership than Netflix, and about as much as Disney, Prime Video, Peacock, and Paramount Plus combined — and that only counts people watching YouTube on their TVs. YouTube is by most measurements the second most popular search engine on the internet (after Google), and the second most popular social network (after Facebook). It’s the most popular service for both music and podcast listening. It’s the second most popular page on Wikipedia, for some reason. It’s for cat videos and Oscar winners. YouTube knows no bounds.Ahead of the 20th anniversary, I talked to a number of people inside YouTube about the state of the platform. I asked them all a variant of the same question: what is YouTube? It’s not just a video-sharing platform anymore. It’s podcasts and videos and music and games and group chats and a thousand other things. What has YouTube become? And what’s the plan going forward?Over and over, I heard the same thing: YouTube is more complicated and more diffuse than ever, and it’s definitely no longer a singular platform. But the idea at the core of the thing hasn’t changed at all. “The secret of YouTube was never really a secret”“The secret of YouTube was never really a secret,” said Scott Silver, a product and engineering leader at YouTube who has been at Google since the days of Google Video. (Google Video, you definitely don’t need to remember, was Google’s attempt to build a video-sharing platform before it gave up and bought YouTube for $1.65 billion in 2006.) “It’s just that if there’s a giant collection of videos somewhere, and you figure out which ones to show people, and then they watch them, and you’re able to pay the people who make that collection of videos to make more of them, then there’s more stuff to choose from. It compounds on itself.”That’s the grand unified theory of YouTube, right there, and it has been compounding on itself for two decades. It has made YouTube enormous, and enormously powerful. But the job’s not done: there are always new formats, new content types, and new devices to reckon with. YouTube’s ambitions are only getting bigger, too — the company clearly aspires to be the size of the entertainment industry, or maybe even be the entertainment industry. That is going to require history’s largest video collection to get much, much larger.The many faces of YouTubeWhen I ask Brian Albert, a managing director on YouTube’s advertising team, to explain YouTube to me the way he explains it to clients, he breaks the platform into three separate categories. There’s streaming, the high-end stuff that competes with Netflix and the rest. There’s also social video, in YouTube’s case mostly meaning Shorts, up against TikTok and Reels. And there’s what Albert calls “straight online video,” the kind of creator-led mid- and long-form video you really only find on YouTube. “We have competitors across the board,” Albert said, “but there’s no single competitor who plays in each of those three lanes.”Albert delineates things this way because that’s how advertisers think about video. YouTube’s clients have budgets, and those budgets have categories — over the years, as YouTube has become a more sophisticated business, it has tried to convince advertisers of the unique value of YouTube but also to give them a place to put the dollars earmarked for live sports or prestige TV. The goal is to get advertisers in the door with products they already know, and then try to sell them on everything else. “Advertisers for a while now have been buying us like TV or like digital video,” said Tara Walpert Levy, YouTube’s vice president for the Americas. “Now they’re leaning in much more heavily on things like commerce or brand deals.” (YouTube, of course, gets a cut whenever you buy something through a shop or a sponsored video.) Advertisers have long been wary of putting too much emphasis on one thing, preferring to spread their spend around. YouTube is increasingly making the argument that it’s possible to hit all your budgets and all your target markets, all on one platform.If you look at YouTube as a series of products reverse-engineered from advertising budgets, the company’s many offerings start to make a kind of sense. Shorts look like TikToks and Reels because that’s what people like to watch, but also because creative agencies are already used to making short, vertical video ads. YouTube spent billions on NFL Sunday Ticket, and made a whole cable-replacement bundle in YouTube TV, to make pivoting from TV ads to digital ads as simple as possible. YouTube builds huge packages around awards shows, March Madness, and other big cultural moments, because that’s what the money is earmarked for.The simplest way to understand YouTube is as an insatiable content collector, constantly searching for anything that smells like time spentUltimately, though, the simplest way to understand YouTube is as an insatiable content collector, constantly searching for anything that smells like time spent. YouTube knows that its whole system — the sharp recommendation algorithm, the creator tools, the vast swaths of money — only works if it ultimately leads to someone creating the exact right thing to recommend to you every time you open the app. That’s one reason it pays creators better than other social platforms; it knows better than most that incentivizing them to make content makes the whole thing go. That’s why YouTube has spent so much time, money, and energy hoovering up every kind of content imaginable. YouTube was built on the back of pirated TV — you could argue the SNL sketch “Lazy Sunday” was the first viral YouTube video, and its huge popularity led to the billion-dollar Viacom lawsuit that briefly threatened to kill YouTube entirely, but ultimately both established it as a place safely full of copyright theft and taught YouTube the value of making deals. For years, YouTube tried to create its own Emmy-worthy shows and Oscar-winning movies, before eventually developing things like Primetime Channels that brought other streaming services onto the platform. (Now we have companies like Warner Bros. just dumping full movies onto YouTube, hoping they’ll get picked up by the algorithm.) The cable-like deals struck for YouTube TV bring all-important live events onto YouTube. By investing in podcasts and YouTube Music, YouTube turned itself into an audio-friendly service too. (YouTube Music is substantially smaller than Spotify or Apple Music, but if you include people consuming music on YouTube it’s the biggest music platform by a mile.) The goal is all the same: anytime you want content of any kind, you’ll open YouTube. I get the sense YouTube would happily start printing books if it thought people still liked to read.You really can’t overstate the importance of the sheer tonnage of YouTube, when you’re looking at why it has worked. Consider a counterexample: Netflix, which has somewhere in the range of 6,500 titles available. Total. “Netflix might have the algorithm knowledge of ‘we’ve seen what you like, and we can tell you what the next perfect movie would be for you,’” Pablo Lucio Paredes, head of engineering and data at the streaming guide company Reelgood, told me last year. “But does Netflix have that actual movie?” YouTube may not have Stranger Things (at least not officially), but it has a few billion other things you might like. And over the last 20 years, it has also managed to convince a lot of people that they’d rather watch MrBeast than the scrappy kids in Hawkins, Indiana.That’s content-biz, babyThere’s plenty of existing content left for YouTube to bring onto the platform, but the way YouTube wins is by getting creators to create. The company has always understood that its homegrown talent is its greatest asset. But every YouTuber eventually feels the platform’s constant desire for more content; if you don’t feed the algorithm, and tap into every new trend and format that bubbles up across the ecosystem, you might get left behind. YouTube seems to mint new creators faster than it burns them out, but that might not be true forever. And as the platform continues to grow, it’s harder and harder for new creators to find a big audience. Eventually, YouTube risks becoming too big for its own good.YouTube is betting on AI to solve a lot of its problems. The company has spent the last couple of years putting AI to work on practically every part of the creator experience, from replying to comments to coming up with ideas to making wholly generated videos. The company is also extremely bullish on using AI to automatically dub videos into other languages. If it all works, it could increase the YouTube library — and the platform’s chance of always having the right video to show you — like nothing ever has. If it doesn’t work, though, it could poison the well with AI slop and turn YouTube into a platform chock-full of content no one wants. One of these fates appears to be waiting for just about every content maker on the planet. The other ultra-ambitious plan for increasing content is coming in video games. Years ago, YouTube tried to compete directly with Twitch as a game-streaming platform, via a separate app called YouTube Gaming, which didn’t really work. Now gamers use YouTube as a platform for content about games, or just content that only makes sense in a world where everyone plays games. “If you had told me when we were building the gaming app,” said Katherine de León, who leads gaming at YouTube, “that in 2025, one of the biggest gaming creators on YouTube would be a mother of four in Texas who makes Minecraft roleplay videos for girls and women, I would have told you to go home.” She also points to huge hits like Skibidi Toilet as the kind of content that only makes sense when everyone’s a gamer.More recently, YouTube has started to offer what it calls “Playables.” They are essentially mini-games inside of YouTube, and feel a little like what you’d expect to find on Facebook circa the Zynga days. There are what appear to be official versions of Crossy Road and Angry Birds, multiple takes on solitaire and chess; and three identical-looking games called Bubble Pop Star, Bubble Shooter, and Bubble Tower 3D. Each one loads like a mini-app inside YouTube.“It’s consumption and creation, right?”At first glance, Playables make no sense. There are no built-in livestreaming tools or comment threads, no sign at all you’re even on YouTube. But de León makes two arguments in favor of Playables. One is that people like playing games, and that’s enough. They’re good content. But the long-term strategy hinges on games that are their own content-creation machines. “A lot of our top games on YouTube are sandbox games,” she said. “It’s consumption and creation, right?” In Fortnite, Roblox, and elsewhere, players are making content in the games, making content about the games, and making content with the tools of the game. I’m pretty sure Bubble Tower 3D isn’t going to turn into a content machine. So at some point, if you want to be the internet’s great content engine, you just build your own Roblox, right? At this, de León mostly just smiles. She was a game dev for years before coming to Google – her answer is clear. But all she’ll say is, it would definitely make sense. “As a game maker of 17 years, I’m excited about that whole loop,” she said. “You can watch, you can play, you can comment, and you can do it in Shorts, you can do it on YouTube TV, you can do it in a live-streaming channel.” Aping Roblox isn’t easy – just ask everyone who’s ever wasted millions or billions of dollars trying to build a live-service game people love — and it would be YouTube’s biggest structural change yet. But it’s content. So YouTube will try.The everywhere appWhile most of YouTube goes out and tries to corner the content market, the engineering team’s job is to take all of these disparate projects and make them something akin to universally accessible. “It’s one of the things YouTube has excelled at from the very early days,” said John Harding, a VP of engineering at YouTube. “We figure out how to get your media everywhere.” Once upon a time, that meant a web browser on your desktop computer. Now it’s much more than that. “I used to say we were trying to get YouTube on anything with a network connection and a screen,” Harding said. “Now we’re trying to get YouTube on things that don’t have network connections — and don’t have screens.”“Now we’re trying to get YouTube on things that don’t have network connections — and don’t have screens.”In fact, Silver reckons, YouTube might be the Google application that can run on the most devices. “Except maybe Search,” he said, before reconsidering. “But then, you don’t really search on your TV, you don’t really search on your watch. But YouTube has to work on all of those devices, from set-top boxes to TVs to VR headsets to watches to car players. And then, of course, mobile phones and desktops and all those kinds of things.” It’s a hard job no matter what, and much harder when you have to reinvent the wheel on every new device. As much as possible, Silver said, “what we try to do is push stuff into our base platform.” No matter where you load a YouTube video, the goal is to have it run as much identical code as possible. Sometimes features might get developed for one part of the platform — like multi-view on YouTube TV, so people can watch four games at a time — but much of that is then brought back to the overall codebase. The goal, Silver said, is to build things as few times as possible.For 20 years, that’s how it has worked: get all the content, get it everywhere. The “all” and the “everywhere” in that plan have both expanded dramatically since the days of “Me at the zoo” in a desktop browser, but the job is still the job. And while YouTube’s competitors have occasionally beaten it in certain ways, particularly recently — people speak in reverent tones about how well TikTok’s algorithm understands them, and YouTube doesn’t have a capture-and-edit tool nearly as good as Instagram or CapCut — nobody has yet managed to copy the whole system. Get people to make videos; put those videos in front of the right people; pay the people who make the videos so they’ll make more. Somehow nobody else is doing that right. And the bigger YouTube gets, the more money comes into the ecosystem, the faster the flywheel turns.Now, though, YouTube is the established giant and no longer the cool upstart. It has convinced the world that creators are celebrities, that prank videos and documentaries can co-exist, that it is a mainstream entertainment business. As Walpert Levy put it, YouTube has reached the “nobody ever got fired for advertising on YouTube” phase. Now the company has its sights on everything from podcasts to gaming, with ideas about how to make them more YouTube-y. The YouTube-ification of the entertainment business is only just getting started. And it’s all a big bet that you’ll be there, watching, the whole time.See More:

Last night, Elon Musk did what he does best: promise “millions” of autonomous Tesla vehicles would be on the road by the end of next year.It’s become a familiar refrain. Every year since at least 2020, Musk has promised Tesla vehicles will finally achieve their full self-driving capabilities. And every year, those promises fail to come to fruition. Tesla has come under investigation for numerous fatalities involving its vehicles and its driver assist features, Autopilot and Full Self-Driving. His rivals continue to outpace him. And Musk has even acknowledged his overly optimistic outlook by labeling himself “the boy who cried FSD.”“I predict that there will be millions of Teslas operating autonomously, fully autonomously, in the second half of next year.”That fantasy of millions of autonomous Tesla vehicles is getting closer to reality, Musk promised again last night. You’ll be able to sleep in your car and wake up at your destination by the end of this year, he vowed. As Tesla’s financial problems mount, the billionaire CEO seems eager to shift focus away from the drab realities of running a car company amid a tumultuous global trade war, and toward a more idealized vision of a future-focused company producing millions of robots, both humanoid and car.But that fantasy is looking increasingly out of reach, especially as Musk confronts the limitations of Tesla’s technology, and the current political reality he helped usher in.During the call, it was clear that Musk is intent on continuing to pursue the goal of deploying fully autonomous vehicles using a camera-only sensor suite and an end-to-end neural network that’s been trained on billions of driving miles from the company’s fleet.Tesla doesn’t have a PR department and stopped responding to questions from reporters almost six years ago. So often Musk’s claims about full self-driving go completely unchallenged, and rarely is he asked to defend his positions. Since his acquisition of Twitter, and its rebranding as X, Musk has been able to count on his legion of fans to amplify his promises about autonomous cars to the broader world, with little pushback.Tesla’s earnings calls remain the last place where Musk fields questions about the company’s technology — and often those questions are packaged with a lot of fawning and ingratiating remarks from financial analysts who have developed parasocial relationships with the billionaire CEO over the past decade. But last night, Musk faced several pointed queries about Tesla’s plans to roll out driverless cars in Austin in just a few short months, as investors continue to struggle to parse reality from fiction.As Tesla’s financial problems mount, the billionaire CEO seems eager to shift focus away from the drab realities of running a car company.Last year, Musk said that Tesla would rollout “unsupervised” robotaxis in Austin, Texas starting in June 2025. And in the call last night, Musk stuck to that deadline, but added a little more color about what to expect. He said the paid ridehailing service would consist of 10-20 Model Y vehicles with remote operators in case any of the cars get stuck. So unsupervised, but also teleoperated in case of emergency.Tesla’s robotaxis will have “audio input” to listen for approaching emergency vehicles — well, at least some of them will, according to Ashok Elluswamy, VP of Autopilot and AI software. “The customer facing versions don’t have audio input, but the version that’s going to be in Austin will have audio input,” he said.Musk has long promised Level 5 autonomy, which describes driverless vehicles that can travel anywhere, under any conditions, without limitations. But last night, he corrected himself: there will be some limitations. A “blizzard in Manhattan,” for example, may be off limits for Tesla’s self-driving taxis. In other words, geofencing, a practice used by every other operator of autonomous vehicles, but typically shunned by Musk because it conflicts with his vision of an unfettered, decentralized fleet of driverless cars free to roam anywhere they want.“It’s increasingly obvious that there’s some value to having a localized set of parameters for different regions and localities,” Musk said.One analyst asked about the reliability of Tesla’s cameras when confronting sun glare, fog, or dust. Musk claimed that the company’s vision system bypasses image processing and instead uses direct photon counting to account for “noise” like glare or dust.Other robotaxi operators have determined that cameras aren’t enough to reliably guide a driverless car through a complex environment. Companies like Waymo, which currently is doing 200,000 fully autonomous taxi trips each week, rely on a sensor suite that includes lidar, radar, and ultrasonic sensors for redundancy in case the vehicle runs into any difficulty. Musk calls lidar “a crutch.”During the call, Musk mocked Waymo’s approach as costing “way more money” than his company. Tesla builds its own cars, its own computers, and its own sensors. It’s more vertically integrated, which Musk argues give it the “leg up” over its rivals. Of course, he made no mention of the major hardware replacement that millions of Tesla vehicles will require in order to achieve their promise of full self-driving — a process he previously described as “painful” and expensive.“It’s increasingly obvious that there’s some value to having a localized set of parameters for different regions and localities.”“Teslas are… probably costing a quarter or 20 percent of what a Waymo costs and made in very high volume,” Musk said. “So, you know, ironically, we’re the ones that made the bet that a pure AI solution with cameras… is the right move.”Hurdles remain, Musk admitted. Tesla will have “10 million autonomous cars” on the road in a few years, unless “we’re blocked by regulatory situations.” What sort of regulatory situations, he didn’t enumerate. Right now, there’s nothing standing in Tesla’s way from deploying dozens or even hundreds of driverless cars if it chooses. The Trump administration, led by Musk’s Department of Government Efficiency, have slashed thousands of jobs, including roles in regulatory agencies investigating the safety of Tesla’s self-driving systems.Tariffs and the rising trade war could be a sticking point. Tesla’s Cybercab production could be disrupted after the company reportedly paused component imports from China thanks to President Trump’s 145 percent tariff on the county.Waymo’s fully driverless cars are currently operating in four cities, including Austin, with more on the way. Tesla has yet to demonstrate its cars can operate autonomously on public roads, preferring to stick to controlled demonstrations on closed lots or private roads.But the real proof is right around the corner — if you continue to believe in Elon Musk’s promises.See More: