We were enamored with the Kindle Scribe when we reviewed it, but we weren't blind to its shortcomings, including its limited editing capabilities. Now, Amazon has rolled out an update for the device, which solves some of the issues we had with it and makes it a much better option for note-taking. One of the new features is the ability to write directly on page. It's still not available for every book, but the Kindle Store on the Scribe now has a new section called "Write-on Books" where you can find titles that support the feature. 

At the moment, the selection of books with direct on-page writing is mostly made up of guided journals, as well as games like crossword and sudoku. In other words, books that don't make sense on a device where writing on them is not supported. If you're checking out titles on a laptop or a phone, you can also see if they support the feature by looking for "On-page writing" under their product details section. For books that don't support on-page writing, you'll still have to create a sticky note and write within a designated rectangle. You'll have to keep your fingers crossed for the ability to doodle and take notes on the margin of a book if that's something you've been wishing for. 

If you want to share or upload notes you've written on the Scribe, the update also gives you the capability to easily convert them to text. You'll find a new option that reads "Convert to text and quick send" in the share menu and another that says "Convert to text and email." Both options share your handwritten notebooks as a .txt file, but for the latter, you'll be given the the chance to review and edit the converted text before sharing them via email to five addresses. 

There's also a new lasso select tool to move your notes and doodles around. You simply have to circle your handwritten text or drawings with it to resize them, drag them to another location or to cut, copy and paste them within a sticky note, a notebook or a PDF. Speaking of PDFs, files with the format uploaded through Send to Kindle are now more customizable. You can crop their margins to make their text easier to read, select text to highlight it and add text notes. You'll now also be able to look up the definitions, translations and Wikipedia results for text found in PDF files, as well as switch between portrait and landscape view. 

These features will roll out to all Kindle Scribes in the coming weeks, but you can manually download the latest software update to get them right now. 

BMW announced new EVs today as part of its summer 2023 lineup. The new models include the i4 xDrive 40 (an all-wheel-drive variant of the i4), the single-motor i7 eDrive50 and the hybrid 750e xDrive. In addition, the automaker revealed an updated infotainment operating system for some models.

The 2024 i4 xDrive40 is an all-wheel-drive, 396-horsepower variant of the popular Gran Coupe. The all-electric vehicle has dual motors that provide an estimated 307-mile range using the standard 18-inch tires (it drops to about 282 miles with optional 19-inch wheels.) In addition, the EV can accelerate from zero to 60 in 4.9 seconds. The i4 xDrive40 will start at $61,600 with an added $995 destination fee. BMW expects US-based deliveries to begin in the third quarter of 2023.

Meanwhile, the rear-wheel-drive i7 eDrive50 is powered by a single GEN5 motor, supplying 449 horsepower. BMW will announce range and performance details “closer to market launch” this fall, but we know the model will start at $105,700 (plus destination fee). Finally, the 750e xDrive combines a 308-horsepower six-cylinder internal combustion engine with a 194-horsepower electric motor. It also offers 483 horsepower and 516 lb-ft of torque. In addition, the plug-in hybrid’s purely electric range is rated at 35 miles. The 750e xDrive will start at $107,000 and the same $995 destination fee. It also launches in the US this fall.

BMW

The automaker is updating its infotainment operating system “in certain models.” BMW Operating System 8.5 gives the home screen “clearly arranged functions” designed to work better on the company’s curved display. Ridding itself of sub-menus, it uses a “zero-layer principle” that keeps all relevant controls and information on a single level, using widgets arranged vertically on the driver’s side. In addition, it includes symbols to quick-access the climate control menu, app library, navigation and Apple CarPlay / Android Auto.

Xbox games will hit cloud gaming platform Boosteroid for the first time next week. Users of the platform will be able to access Deathloop, Gears 5, Grounded and Pentiment starting on June 1st.

Boosteroid users in the US, UK, European Union and Ukraine will be able to stream eligible Microsoft-owned games that they buy from Steam or the Epic Games Store. Support for Microsoft Store purchases and other titles is on the way. "We’ll regularly add more hits and fan favorites from our extensive catalog of PC games," Sarah Bond, corporate vice president at Xbox, wrote in a blog post.

This will make Boosteroid the second external cloud service where Microsoft will make its games available (Game Pass Ultimate subscribers can stream all these titles via Xbox Cloud Gaming as well). NVIDIA GeForce Now users have been able to stream Gears 5 since May 18th. Deathloop, Grounded and Pentiment will land on GeForce Now later this week.

Microsoft signed 10-year deals with NVIDIA, Boosteroid and other cloud gaming providers to offer those platforms access to Xbox and Bethesda games, as well as PC versions of Activision Blizzard titles if its attempted acquisition of that company goes through. It struck those agreements to try and persuade regulators to approve the $68.7 billion deal.

While the acquisition has been rubberstamped in dozens of territories, most recently China, cloud gaming has been a sticking point for regulators in some countries. The UK's competition watchdog blocked the deal, claiming it would strengthen Microsoft's leading position in the nascent cloud gaming market. The US Federal Trade Commission cited concerns about cloud gaming (and other factors) in the lawsuit it filed in an attempt to block the merger. On the flip side, Microsoft's cloud gaming concessions have placated European Union officials, who gave the deal the green light last week.

TikTok filed a lawsuit Monday to challenge Montana’s ban of the social platform, as reported byThe Wall Street Journal. The case was brought against the state attorney general Austin Knudsen.

Montana’s governor signed the bill banning the app in the state last week, one month after the state’s legislature passed it. The law is set to take effect on January 1st, 2024. The following day after the signing, a group of creators sued the state along similar grounds as TikTok’s suit today.

The law prohibits the ByteDance-owned platform from operating in the state, as well as preventing app stores from listing the app for download. Although it isn't clear how Montana plans to enforce the ban, it states that violations will tally fines of $10,000 per day.

This is a developing story. Please check back for updates.

Multiple Twitter users are reporting that deleted tweets are reappearing on the site at an alarming rate, with one security expert noting that the social media app restored a whopping 34,000 messages, as reported by ZDNET. The bug also seems to be impacting deleted retweets.

The issue appears to be widespread, but all of the data’s anecdotal. It seems to mostly impact those who batch-deleted tweets via a third-party service like Redact or TweetDelete, though not always. In any event, check your feed for the reappearance of that tasteless Harambe tweet you wrote and then deleted back in 2016.

This is a trivial issue for many, but does speak to how the company handles and protects your private data. A former Twitter Site Reliability Engineer suggested on Mastodon that it “sounds a lot like they moved a bunch of servers between data centers and didn’t properly adjust the topology before reinserting them into the network.” There’s no way to find out anything official because Elon Musk disbanded the press team, replacing the press line with a bot that returns only poop emojis. In any event, we’ll update this story if the company issues a statement of any kind that addresses the issue.

This has been a bizarre few months for the social media platform. The city of San Francisco launched a probe into potential building code violations and the site’s been plagued with technical issues. All of this has led to a stark decline in usage numbers, with competitors waiting in the wings to scoop up those fleeing the service.

Amazon’s palm payment technology is moving from the world of convenience food to a booze-soaked baseball field. The company just announced that the Amazon One identification system is being implemented at Denver’s Coors Field, home of the Colorado Rockies Major League Baseball franchise, allowing customers to buy beer and hard alcohol with their palms.

This lets you stroll right into the field’s SandLot Brewery or Coors Light Silver Bullet Bar and order up some of the good stuff without having to waste precious moments digging through your pockets to find that dang ID. Over the course of an entire game, this could save 20 to 30 seconds, depending on how thirsty you are.

Of course, you have to be enrolled in the company’s biometric Amazon One service. The web giant’s setting up kiosks at the field for quick signups, but you need your ID, your Amazon sign-in information and you have to take a selfie. Once enrolled, however, just hover your palm over the reader and the bartender will know you are above 21 and ready to gulp. The system isn’t just for age verification, as it also lets you pay with your palm.

It’s not the easiest task in the world to grab a beer at a major sporting event, so this could be a useful bit of tech, if you’re willing to give Amazon even more access to your private life. The company hasn’t been the most reliable collector of biometrics, facing lawsuits in multiple states for violating biometric privacy laws. If you still want to give this a try and are nowhere near Colorado, Amazon One palm-scanning technology is also available at many Amazon-owned Whole Foods locations.

Meta has created an AI language model that (in a refreshing change of pace) isn’t a ChatGPT clone. The company’s Massively Multilingual Speech (MMS) project can recognize over 4,000 spoken languages and produce speech (text-to-speech) in over 1,100. Like most of its other publicly announced AI projects, Meta is open-sourcing MMS today to help preserve language diversity and encourage researchers to build on its foundation. “Today, we are publicly sharing our models and code so that others in the research community can build upon our work,” the company wrote. “Through this work, we hope to make a small contribution to preserve the incredible language diversity of the world.”

Speech recognition and text-to-speech models typically require training on thousands of hours of audio with accompanying transcription labels. (Labels are crucial to machine learning, allowing the algorithms to correctly categorize and “understand” the data.) But for languages that aren’t widely used in industrialized nations — many of which are in danger of disappearing in the coming decades — “this data simply does not exist,” as Meta puts it.

Meta used an unconventional approach to collecting audio data: tapping into audio recordings of translated religious texts. “We turned to religious texts, such as the Bible, that have been translated in many different languages and whose translations have been widely studied for text-based language translation research,” the company said. “These translations have publicly available audio recordings of people reading these texts in different languages.” Incorporating the unlabeled recordings of the Bible and similar texts, Meta’s researchers increased the model’s available languages to over 4,000.

If you’re like me, that approach may raise your eyebrows at first glance, as it sounds like a recipe for an AI model heavily biased toward Christian worldviews. But Meta says that isn’t the case. “While the content of the audio recordings is religious, our analysis shows that this does not bias the model to produce more religious language,” Meta wrote. “We believe this is because we use a connectionist temporal classification (CTC) approach, which is far more constrained compared with large language models (LLMs) or sequence-to-sequence models for speech recognition.” Furthermore, despite most of the religious recordings being read by male speakers, that didn’t introduce a male bias either — performing equally well in female and male voices.

After training an alignment model to make the data more usable, Meta used wav2vec 2.0, the company’s “self-supervised speech representation learning” model, which can train on unlabeled data. Combining unconventional data sources and a self-supervised speech model led to impressive outcomes. “Our results show that the Massively Multilingual Speech models perform well compared with existing models and cover 10 times as many languages.” Specifically, Meta compared MMS to OpenAI’s Whisper, and it exceeded expectations. “We found that models trained on the Massively Multilingual Speech data achieve half the word error rate, but Massively Multilingual Speech covers 11 times more languages.”

Meta cautions that its new models aren’t perfect. “For example, there is some risk that the speech-to-text model may mistranscribe select words or phrases,” the company wrote. “Depending on the output, this could result in offensive and/or inaccurate language. We continue to believe that collaboration across the AI community is critical to the responsible development of AI technologies.”

Now that Meta has released MMS for open-source research, it hopes it can reverse the trend of technology dwindling the world’s languages to the 100 or fewer most often supported by Big Tech. It sees a world where assistive technology, TTS and even VR / AR tech allow everyone to speak and learn in their native tongues. It said, “We envision a world where technology has the opposite effect, encouraging people to keep their languages alive since they can access information and use technology by speaking in their preferred language.”

Cadillac confirmed today that the first all-electric Escalade will arrive “later this year.” However, the automaker didn’t reveal any details about the Escalade IQ, a name first trademarked in 2021. The new model’s “IQ” branding aligns with the Celestiq luxury sedan and Lyriq mid-sized SUV.

Earlier this year, Cadillac VP Rory Harvey said the company would reveal three new EVs in 2023. If you add that to the company’s previous comments to Car and Driver, stating all three will arrive for the same model year, we can assume the Escalade IQ will be a 2024 model. It is also expected to use GM’s Ultium battery tech.

We’ll have to wait until later this year to learn more about the first Escalade EV. But as for its mid-sized counterpart, Engadget’s Roberto Baldwin found the Lyriq ($60,000 and up with over 300 miles of range) to have “the fit and finish you’d expect” from Cadillac with “a polished ride and almost eerily quiet interior.”

WhatsApp is rolling out a long-overdue feature to all users. Starting today, you'll be able to edit messages that you have already sent. You'll need to act fairly swiftly, as you'll have a 15-minute window to correct a mistake or drop in an emoji you initially forgot to include.

To modify a sent message, long press on it and choose the Edit option from the menu. Any message that you change after sending it will include the word "edited" next to the time stamp. No one else in the chat will be able to see the edit history, but they'll know that you've modified the message. WhatsApp says the feature will be available to everyone in the coming weeks.

WhatsApp is playing catchup to other messaging services here. Since the release of iOS 16, iPhone and iPad users have been able edit sent missives in iMessage. Twitter Blue subscribers can make changes to a tweet up to 30 minutes after sending it.

There are a lot of factors to consider when you’re looking for the best iPad to fit your needs. But, if you’re more interested in an entry-level model than something more powerful and pricey, the standard iPad might just fit the bill. Even better, the 2022 iPad is currently on sale for $399. That’s $50 off the regular price and it matches a record low.

This deal is for the base model with 64GB of storage. We gave this tablet a score of 85 in our review. We liked the updated design that takes a page out of the iPad Air's book, the improved performance and battery life compared with the previous generation and the switch to USB-C charging. Moving the front-facing camera to the landscape side is a welcome change too. That makes far more sense for those who use the tablet for video calls while propping it up with the kickstand or a keyboard folio. Speaking of which, the keyboard folio got an upgrade, since there are larger trackpad and function keys.

On the downside, the $250 folio is expensive; it's not far off the price of the tablet during this sale. We also felt that, while larger than previous models, the quality of the screen isn't as refined as those in higher-end iPads. For instance, it doesn't have an anti-reflective coating or support for the P3 wide color gamut. Perhaps most gallingly for those with a first-generation Apple Pencil, you'll need a USB-C cable and a Lightning to USB-C adapter to charge the peripheral from this iPad.

Still, there's more positives than negatives to be found here. This iPad has a lot of improvements over the previous generation. If you've been holding onto an older model for several years and don't want to break the bank on a replacement, this deal is certainly worth considering.

Follow @EngadgetDeals on Twitter and subscribe to the Engadget Deals newsletter for the latest tech deals and buying advice.