Stevie Bonifield | The Verge

Tesla’s Optimus robot is going through growing pains

2026-09-25T13:01:36-04:00

Hitting its goal of making 20,000 Optimus robots per week is reportedly proving tricky for Tesla. The Information reports that Tesla produced “several hundred robots a week” last month, after it repurposed its Model S and Model X production lines for Optimus earlier this year. However, this strategy is reportedly creating manufacturing snags, like issues lining up parts precisely, as The Information notes: “Not only are the Optimus production lines new, but the Optimus parts are much smaller and have to fit together far more precisely than car parts.”

It reports that even in the “V3” Optimus robots it’s currently producing, the hands and forearms “have more than 100 screws and other small components that currently require a worker to assemble manually.” These robots are lighter and have more cameras than previous versions, but lack a “sensing glove” currently in development that will improve durability by making it easier to replace touch sensors without swapping out the entire hand.

Optimus has faced delays for years now, even as Tesla CEO Elon Musk has made bold promises about it, saying Optimus will be Tesla’s “biggest product of all time.” However, at the moment they’re still being used internally for specific tasks in limited areas, and not generally or with remote teleoperation. Tesla also reportedly plans to lease the robots initially rather than selling them, and use the training data collected to improve the robots, but it will have to get them out in the world first.

Meanwhile, other developers have begun shifting to less lifelike designs for their robots, and earlier this month, Chinese regulators reportedly started holding back IPO listings for humanoid robotics companies.

Gemini can now call businesses for you so you don’t have to wait on hold

2026-09-24T12:15:24-04:00

Google’s launching an “early experiment” feature on Pixel 11 that lets users delegate local business calls to Gemini, like making a reservation, checking if a product is in stock, or rescheduling an appointment. According to Google, you don’t even need to start the call to have Gemini handle it for you:

Instead of dialing yourself, just tell Gemini to call on your behalf using the Gemini app on your Pixel. Gemini will call the business, introduce itself, navigate automated phone menus, wait on hold, and handle the conversation with the person on the other end. You remain in control at all times with a live transcript view and the ability to take over the conversation at any point.

Google announced a similar “Call for me” feature last year that has its AI call local businesses to ask if they have a product you’re looking for, but it didn’t have an option for users to take over those phone calls. Meta is also working on a similar automated phone call feature for its Muse AI agent, although those calls reportedly might be made by a human behind the scenes. Google’s new AI calling feature is rolling out starting today as an “early preview” for “Gemini paid subscribers enrolled in the Phone by Google Public Beta on Pixel 11 in the US.”

Meta Connect 2026: The 7 biggest announcements

2026-09-23T20:00:33-04:00

Meta announced a slate of new wearables during its annual Meta Connect showcase on Wednesday. The star of the event was the new Meta VR Glasses, which pack virtual reality tech into a lightweight smart glasses-style frame with a tethered puck. Meta also announced new pairs of screen-free smart glasses, along with some new AI features and style and frame options.

You can catch up on all the highlights from this year’s Meta Connect below.

Meta VR Glasses bring virtual reality to a lighter form factor

Meta’s latest wearable offers full virtual reality in a smart glasses-style form factor. Unlike a typical VR headset, the VR Glasses don’t have any straps, and Meta says they’re “five times lighter than Meta Quest 3” at just 100g. Similar to Google’s Project Aura, the glasses are tethered to a pocket-sized puck with the compute hardware, battery, and storage, including a Qualcomm Snapdragon Reality Elite, 128GB of storage, and 12GB of RAM. The 5K micro OLED “Infinite Display” has 2412 x 2288 pixels per eye, with a refresh rate of up to 120Hz.

The Meta VR Glasses are launching in spring 2027 for $1,299, pricier than the Meta Ray-Ban Display but less expensive than the Apple Vision Pro, while offering similar features. Users can connect the glasses to a Mac or Windows laptop for use as a virtual external display, with an option for a virtual surface keyboard and touchpad.

The VR Glasses also have access to the full Meta Quest catalog of apps and games and support for Xbox Cloud Gaming. Meta says the glasses are “the first IMAX Enhanced certified VR device ever,” and users will be able to watch “immersive live sports” with them as well as 3D movies. They even have a “hologram” feature that creates a “photorealistic digital version of your face” for using in calls while wearing the glasses, reminiscent of the “personas” on the Apple Vision Pro.

Meta is launching camera-free smart glasses

The highlight of the Ray-Ban Meta Audio is the feature they don’t have: built-in cameras. You can use them for things like listening to music, taking calls, or chatting with an AI assistant, but they can’t take pictures or record video — features that have sparked privacy and harassment concerns around Meta’s other glasses.

Meta says the new audio-only glasses are also its “lightest, slimmest AI glasses yet,” weighing just 43g. They offer up to 12 hours of battery life per charge, or 48 hours with the charging case. The Ray-Ban Meta Audio will launch on October 13th starting at $349 with 23 different color and lens combinations.

Refreshed Ray-Ban Meta Gen 3 arrive with battery and design upgrades

Meta launched an updated version of its main Ray-Ban Meta glasses, starting at $449. The Ray-Ban Meta Gen 3 offer up to nine hours of battery life, one hour more than the Gen 2. They also have “a new customizable action button for instant access to Meta AI” and a new six-mic array that Meta says “cuts more than 90% of background noise.”

The Ray-Ban Meta Gen 3 has a slimmer design than the Gen 2 as well, with 27 color and lens combinations, including new Zena and Aviator frame styles. Notably, one thing that hasn’t changed is the built-in LED light that indicates when the glasses are recording, which Meta rolled out a privacy fix for last month.

Budget glasses launch along with new glasses styles

Alongside its pricier Ray-Ban glasses, Meta also announced some updates to its in-house line of glasses. It’s launching a new lower-cost version of the Meta Adventurer, which will start at $249 for a basic design with black frames and gray lenses, launching October 23rd.

Meta is also introducing a new “Mix & Match” program for customizing its glasses with different frames, lenses, and case colors. Plus, its EssilorLuxottica collection is getting two new styles, Capri and Nova, and the Kylie Jenner Starfire frames are getting a new ivory colorway. Meta’s also launching a new collaboration collection with Lisa of Blackpink.

Meta Glasses are expanding to more countries starting today, too, including Singapore, South Korea, Mexico, and the United Arab Emirates, with availability in Brazil “coming soon.”

Meta Ray-Ban Display is coming to more regions with some new features

A year after they originally launched, the Meta Ray-Ban Display are coming to more regions. The glasses are on sale in the UK and Canada starting today, and they’ll be available in France, Italy, and Germany starting on October 13th. Meta’s also introducing online ordering for customers in those regions and the US.

Meta’s launching a new Explore app for Display glasses where users will be able to “browse and use a growing selection of experiences built for your glasses.” The Meta Ray-Ban Display is also getting photo editing, an improved calendar, and mirrored phone notifications. Plus, Meta says the “hologram” feature launching on its VR Glasses is rolling out on the Meta Ray-Ban Display “in Early Access on WhatsApp to Meta Ray-Ban Display users in the US later this fall.”

Meta’s Muse AI is coming to glasses and video calls

After launching its new Muse assistant earlier this month, Meta says it’s “working on” bringing the AI to its smart glasses. When it arrives on the glasses, Muse will be able to do things like guide your workouts, log meals, book appointments, and “help buy products you see.”

Meta also announced some new features for Muse, including video calls with a customizable live avatar of the AI. Plus, Muse agents are getting their own email addresses for completing tasks and communicating with users.

Meta is making an AI ‘charm’

Mark Zuckerberg teased one last device, Muse Charm, a small AI gadget that houses the Muse agent and looks a little bit like a supersized Tamagotchi. More details will be shared “soon,” he said, but the product is supposed to ship in time for the holiday season.

Muse is coming to Meta smart glasses

2026-09-23T19:40:56-04:00

Just a couple of weeks after launching Muse, Meta announced that it’s “working on” bringing the agent to its smart glasses, including the new glasses it unveiled at Meta Connect. Users will be able to activate Muse from their glasses by saying its name and ask it to handle tasks like guiding a workout, logging meals, or helping them shop for products they’re looking at.

Muse is just one of several major updates coming to Meta’s glasses. They’re also getting some new accessibility features, like a Hearing Enhancement feature Meta says is “FDA-cleared software for adults with perceived mild to moderate hearing loss.” It will be available later this year with a Meta One subscription or as a $149 add-on for compatible glasses. Meta AI will also be able to give “more context-rich answers” when users ask the glasses what they’re looking at, potentially making Meta’s glasses more effective for helping blind and low-vision users interpret their surroundings. A new audio customization feature will also allow users to shift their smart glasses’ volume between their left and right ears, including sending all audio to one side.

Meta’s introducing some new audio and photo features as well, including Dolby Atmos Capture, which will allow users to record videos with 360-degree spatial audio. A new Dynamic Photo feature automatically captures multiple frames when you take a photo with Meta’s glasses, then uses Meta AI to suggest the best frame from each batch.

The Meta Ray-Ban Display glasses are getting the new “Hologram” feature Meta announced for its new VR Glasses, which it says is “a photorealistic digital version of your face for calls from Meta Ray-Ban Display” that “infers” your expression using your voice. The feature’s “rolling out in Early Access on WhatsApp to Meta Ray-Ban Display users in the US later this fall.”

A couple new heads-up navigation features are coming to the Meta Ray-Ban Display as well. Starting later this month, users in the US and Canada will have an option to use their voice to ask for real-time cycling or public transit directions. The Meta Ray-Ban Display will also start using “the world in front of you rather than street signs in turn-by-turn walking directions,” like referencing a house, car, or landmark the glasses detect. A few other new features are on the way for the Display glasses, including an option for editing photos directly on the glasses’ display, a new “Explore” app for finding new apps, an improved calendar, and mirroring for phone notifications.

This $199 GPS sports watch is repairable, modular, and open-source

2026-09-23T12:28:42-04:00

The Una Watch costs less than most Garmin watches now that its price has been lowered from the Kickstarter offer of $350 to $199, but its modular, repairable design grabbed my attention first. It features a 1.2-inch MIP LCD display and up to 10 days of battery life, similar to the Garmin Forerunner 70 or the Pebble Round 2 I preordered earlier this year. Like Pebble’s revamped watches, the Una Watch is also open-source, including its SDK, housing, and circuit boards, but it has more fitness features out of the box than Pebble’s watches currently offer. Una is also encouraging users to build their own apps for it, including with vibe coding — it says users can “ask an AI coding assistant to interpret Una’s developer documentation, produce code, diagnose errors and refine the resulting application.”

Una says every component in the watch is “accessible with a standard tool,” so users can repair or upgrade it themselves. The open-source mechanical components also open the door for users to design and 3D print their own accessories or mods for it, like converting it into a cycling computer.

It features dual-frequency GPS, 24/7 heart rate tracking, a barometric altimeter, a six-axis accelerometer and gyroscope, a three-axis magnetometer and a replaceable 280mAh battery with USB-C charging. However, the Una Watch is only IPX5 splash- and showerproof, not fully waterproof.

These are the first five Googlebook laptops

2026-09-21T13:12:20-04:00

Google and partners are finally announcing the first laptops under the new Googlebook banner. There are five models from Acer, Asus, Dell, HP, and Lenovo to kick things off. They’re available for preorder starting today and arrive in stores on October 4th.

Starting prices range from $899 to $1,299, putting the Googlebooks in the same territory as the MacBook Air and a plethora of Windows laptops. Each Googlebook also comes with 12 months of Google AI Pro with 5TB of cloud storage, along with three-month trials of YouTube Premium, Adobe Photoshop, and CapCut. Google’s also promising up to 10 years of updates.

Here’s a look at the hardware each Googlebook has to offer for those prices, and how they compare to each other.

Acer Googlebook 14

The Acer Googlebook 14 is the only convertible 2-in-1 in the first round of Googlebooks, featuring a design a bit like Acer’s Chromebook Plus Spin 714. Acer says it offers up to 16 hours of battery life. It also currently has the lowest starting price at $899, although that might not last — Acer’s Kelly Odle said in a comment to The Verge that $899 is the laptop’s “initial launch price available for a limited time,” adding that Acer plans to “reevaluate pricing over time based on component supplies and other market pressures.”

Ongoing shortages of RAM and other components have been driving up prices on other laptops, like Apple’s MacBook Neo, and unfortunately it sounds like a price hike isn’t out of the question for the Acer Googlebook 14, either.

Intel Core Ultra 5 or Core Ultra 7 processor
16GB of RAM
256 or 512GB of storage
14-inch 2.8K OLED touch display
13.9mm thick, 2.5lbs (1.13kg)
1x Thunderbolt 4, 2x USB-C, 1x headphone jack
Fingerprint sensor

Asus Googlebook 14

Asus’ Googlebook 14 takes after its ExpertBook Ultra with its design, featuring a metallic gray-blue chassis and more ports than most of the other Googlebooks. It’s the lightest of the first five models at 2.2 pounds, although it isn’t the thinnest (that title goes to Dell). The Asus Googlebook 14 is one of only two that come with an HDMI port as well. It’s one of the pricier Googlebooks, starting at $1,299, but it features spec options similar to the others.

Intel Core Ultra 5 or Core Ultra 7 Series 3 processor
16 or 32GB of RAM
256 or 512GB of storage
14-inch 3K OLED touch display
14.0mm thick, 2.2lbs (1.0kg)
2x USB-C, 2x USB-A, 1x HDMI, headphone jack
Fingerprint sensor

Dell XPS Googlebook

Dell’s first Googlebook, starting at $999 through a “special offer” at BestBuy, closely resembles the XPS 13 but with a light gold “shimmer” colorway, the Googlebook glowbar, and Google’s software. Unlike the Windows XPS 13, the XPS Googlebook also has a fingerprint reader. Unfortunately it’s just as sparse on ports as the original, with only two USB-C ports.

Qualcomm Snapdragon X Elite processor
16 or 32GB of RAM
512GB of storage
13.4-inch 2.5K touch display
12.7mm thick, 2.2lbs (1.0kg)
2x USB-C
Fingerprint sensor

HP Googlebook 14

The HP Googlebook 14 starts at $1,299 and notably includes both facial recognition and a fingerprint reader. It also currently has the highest storage capacity on a Googlebook, maxing out at 1TB. It sports a dark blue chassis similar to the HP Spectre, but it’s light on ports with just two USB-C ports and a headphone jack.

Qualcomm Snapdragon X Elite processor
16 or 32GB of RAM
256GB, 512GB, or 1TB of storage
14-inch 2.8K OLED touch display
14mm thick, 2.74lbs (1.24kg)
2x USB-C, headphone jack
Face authentication and fingerprint sensor

Lenovo Googlebook 15

Lenovo’s Googlebook starts at the higher end at $1,299, but its 15.3-inch display is the largest display out of the first five models. It’s one of the only ones with an HDMI port, and it’s also the only one with a white chassis, which has a satin-like finish to it, rather than metallic blue or gold. Lenovo’s also launching an AI mouse alongside the Googlebook 15 with a matching glowbar.

Intel Core Ultra 5 Series 3 processor
16GB of RAM
256 or 512GB of storage
15.3-inch 2.8K OLED touch display
15.95mm thick, 2.84lbs (1.28kg)
2x USB-C, 1x USB-A, 1x HDMI, headphone jack
Face authentication and fingerprint sensor

Update, September 21st: Added new starting price for Dell XPS Googlebook at BestBuy.

Security researchers used Claude to help them hack into OpenAI

2026-09-18T12:30:32-04:00

A team of three independent security researchers at Hacktron says it took less than 72 hours for them to hack into OpenAI employee accounts using Anthropic’s Claude Opus 4.8 and 5, The Wall Street Journal reports. They were able to access OpenAI’s GitHub repository, called “Monorepo,” which reportedly contains “OpenAI’s algorithmic secrets,” according to The Wall Street Journal’s sources.

They stopped short of accessing internal code in Monorepo themselves, but sent a pull request from an employee’s Codex account to prove they gained access. They were able to get in through Discourse, the third-party service that hosts OpenAI’s community forums, by exploiting an issue with the system it uses to process HEIF images. According to Hacktron, Claude Opus 5 launched in the evening on July 24th, and by 10AM the next day they had used it to achieve RCE on Discourse Cloud and accessed OpenAI’s instance.

The risk sits below the application layer in native C/C++ decoders such as libheif, often included indirectly through ImageMagick, libvips, Sharp, distro packages, or container images.

Crafted AVIF/HEIC uploads can reveal the version family, allowing a matching exploit to cause… pic.twitter.com/cB1y65ftTD
— Harsh Jaiswal (@rootxharsh) September 18, 2026

Their HEIF Heist project took “only one or two days” to adapt to different companies, including OpenAI, Slack, Meta, GitHub Ent, Rails, Next.js, ImageMagick, and others, using less than $3,000 in tokens, and to their knowledge, was only detected by one target, Shopify. The vulnerabilities Hacktron reported to Discourse and OpenAI have since been fixed, and Hacktron says OpenAI paid it $6,500 for finding the bug, but as Hacktron CTO Mohan Pedhapati said to the WSJ, “I don’t think we are as strong as Chinese threat actors… We’re just three guys with Claude and Codex subscriptions.”

Claude Code relaunches Projects to manage multiple AI agents in the cloud

2026-09-17T14:59:57-04:00

The revamped projects feature in Claude Code allows users to run multiple agents under the same roof, with a shared memory, goals, and library of files and artifacts. Similar to Grok Bot and other tools that manage groups of AI agents, each project has “threads” running different tasks in parallel, with a “coordinator” directing everything:

Under the hood, each thread is a Claude Code cloud session working on its own branch and copy of the repo. The coordinator keeps work organized, but if any threads work on the same code, the overlap is resolved as a merge conflict just like any other PR. Each thread can further split its delegated work into pieces using subagents, loops, and workflows when needed so large assignments finish faster.

Users can interact with each thread individually, or monitor and update things through the main project chat. At launch, threads run in the cloud, but Anthropic says support for local tools and code is coming “very soon.” The updated projects feature is available in beta starting today for “select Claude Pro and Max subscribers,” with access expanding to all Pro, Max, Team, and Enterprise users later, as well as Cowork and regular Claude chats.

Claude comes for Gemini with its own take on Docs and Slides

2026-09-16T10:24:02-04:00

Claude is getting a pair of new tools today: Docs and Slides. They’ll let you create documents and presentations through Claude chats, which you can export, edit, and share with other users.

As part of the announcement, Anthropic is also simplifying how Claude chats work, merging regular chats and Cowork into “one Claude,” with all of its AI productivity tools available from any chat. Artifacts and Claude Design capabilities will be available through the new single interface as well. According to Anthropic, Claude “can now figure out what a task needs, so what Cowork and Design can do is available from any conversation, with the context, skills, and connectors you already have.”

Claude Docs and Slides are launching in beta, joining Claude Design as part of the AI platform’s productivity tools. You can have Claude generate documents and presentations from any chat. You can ask Claude to make changes to whatever it creates, or make edits manually.

Like Google Docs, Claude Docs start as private documents, but you can share them with other editors who can work on the document at the same time as you. Collaborators, including Claude, can leave comments that look like those on a Google Doc. Anthropic says “Anything you make with Claude Design, Slides, or Docs lives at one shareable link you can open on your phone.”

Anthropic’s models are generally seen as more advanced than Google’s, but Google’s AI services have the advantage of being natively tied into its suite of productivity tools. Earlier this year, Google expanded Gemini’s role in Google Docs, Sheets, and Slides with features like a Gemini chat window in Docs, generating and editing entire documents with Gemini, and creating slides with Gemini. The addition of Docs and Slides to Claude starts to close the gap between Anthropic’s AI platform and Google’s productivity tools, giving users fewer reasons to leave a Claude chat for other tools.

The “one Claude” update is rolling out first to Pro and Max users on web, desktop, and mobile “over the next few weeks,” with access for Team and Free users coming later. Anthropic says there’s “nothing to turn on” for the update, and users can also keep using Claude chat normally.

Meta’s new One subscriptions put a price on social media and AI

2026-09-15T10:45:00-04:00

Shortly after launching its new do-everything AI assistant Muse, Meta’s launching subscription bundles that pair its standalone app subscriptions with extra AI usage. Some of the new Meta One bundles were in testing earlier this year, but are now available globally starting today, with several tiers for individual users, creators, and businesses. Meta says the “core experience” on its apps and Meta AI will still be free, and users can still get its subscriptions for Facebook, Instagram, and WhatsApp without a bundle. It also says it plans to expand the bundles to include “Edits, AI glasses, and more over time.”

There are two bundles for individual users, Core for $7.99 per month and Premium for $19.99 per month. They include Instagram Plus, WhatsApp Plus, and Facebook Plus, with “more media generation with Meta AI,” like generating images with Muse or using in-app AI features like Instagram’s Restyle. That’s a few dollars cheaper than paying $11 per month for all three of the standalone subscriptions.

The Meta One subscriptions aimed at creators and businesses range from $14.99 to $499 per month. They also include Meta’s individual app subscriptions, along with perks like a bold follow button, a verified badge, higher rankings in search results, and monitoring for fake impersonation accounts:

Essential ($14.99/month): Establish yourself and build confidence with new audiences. Get tools to manage your presence and more access to Meta Business Agent on WhatsApp to respond to customers 24/7. Plus, a verified badge, verified channel on WhatsApp Business app, and impersonation protection so people know you’re who you say you are (pending successful verification).
Advanced ($49.99/month): Invest in professional-grade tools. Advanced publishing features like scheduling stories up to 30 days in advance, links in organic posts and reels, exportable analytics, and deeper audience insights. Give team members access to your account without sharing your password and get more linked devices and business broadcast credits on WhatsApp. Plus, even more Meta Business Agent responses.
Expert ($149/month) and Max ($499/month): Scale and optimize. These plans include our highest levels of feature access and Meta Business Agent capacity for teams managing at scale.