Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.
The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.
Shion arrives when Season 3 starts on June 16.
After teasing the newest addition to the lineup of colorful characters in its team shooter earlier this week, Blizzard has released the gameplay trailer for another new damage hero. Shion will join Overwatch when its next season, the third since the Reign of Talon reboot, kicks off on June 16. This character won’t do anything to change the impression that Blizz is leaning hard on making hot heroes or that it’s starting to suffer same-face syndrome with its female lineup. But at the very least, Shion is one of the team’s most dramatic creations in a while, looking very stylish in an all-white suit paired with petite red horns and cybernetic red eyes.
Appearance-wise she might call to mind the moody Reaper character, but Shion’s kit actually feels more reminiscent of Tracer. She’s dual-wielding semi-automatic pistols as her base weapon, but can focus fire in a more powerful burst to finish off a low-health opponent. Mobility also seems core for this hero, since she has both a dash move and can ride actual motorcycle in-game. The bike seems especially satisfying, since it looks like there’s an ability for Shion to jump off and launch the vehicle forward into a rival. Her ultimate ability turns Shion into a damage-dealing vortex that can dash toward enemies.
Considering the big Overwatch reboot aims to emphasize a bigger narrative tying together the characters, it seems likely that the next lore drops will highlight story threads surrounding the Hashimoto clan where Shion is a boss. So expect more background on support heroes Kiriko and Mizuki as part of the larger drama about Vendetta’s efforts to take over the world as Talon’s new villain-in-charge.
Microsoft is accelerating updates to its Edge browser, switching from a monthly release schedule to a biweekly one. The change takes effect with Edge 152, due on August 27, and puts the browser on the same cadence as Google Chrome.
The shift does not mean users will get twice as many new features. As Microsoft explained in a recent blog post, each release under the new schedule will carry roughly half the content of the current monthly drops, keeping the overall volume of changes roughly constant. The practical effect for most users is a steadier, smaller stream of updates rather than a sudden jump in new functionality.

Microsoft framed the change as a benefit for both consumers and enterprise customers, noting that security and platform fixes will reach users faster, and that smaller change sets are easier for IT teams to validate before deployment.
The new cadence applies to users on the standard Stable channel. Those on the Stable Extended channel, a longer-term option aimed at organizations that prefer less frequent updates, will stay on the current every-two-months schedule.
Google Chrome moved to a two-week release cycle in March, and Edge’s realignment closes the gap between the two Chromium-based browsers. The biweekly releases kick in with Edge 152 on August 27, giving users and IT admins a couple of months to prepare before the new schedule takes hold. For everyday users, the transition should be largely invisible. Automatic updates will simply arrive more often, each with a smaller footprint than before.
Personal Tech
Casual IT team learns that building bespoke PCs can be a false economy
ON CALL 你好 Nǐ hǎo, dear reader, and welcome to another installment of On Call, The Register’s Friday column that shares your stories of translating technical trauma while delivering transcendent tech support.
This week, meet a reader we’ll Regomize as “Jackson” who
told us about his time providing tech support in a university’s biology
department.
“It was sometime in the mid-2000s and our IT group at the
time consisted of myself, my boss, and a part-timer,” he told On Call. “We were
a very casual IT group; nothing in the way of any formal policies or standards
for anything at all. If someone needed a new PC, we just ordered parts and
assembled them ourselves.”
The department’s PC fleet therefore had a diverse gene pool,
with no two machines possessing the same bill of materials.
“This was fine by me – I enjoyed building them and it never
really caused any issues that I couldn’t handle,” Jackson told On Call. “Until
one day we got a panicked support call from one of the secretaries who claimed
that her PC just rebooted and then started talking to her.”
Jackson and his colleagues didn’t believe a word of it until
the secretary stopped talking and placed her phone next to the talking PC.
“I could clearly hear a muffled voice repeating a message of
some sort,” Jackson told On Call.
There was nothing for it but to visit the PC, which he found
hung in the middle of a Power-On Self-Test, flashing an alphanumeric error code
and unmistakably playing a voice through its internal speaker.
In Chinese!
Jackson rebooted the machine and it ended up in the same
state, reciting the same message. Chinese isn’t a language in which Jackson is
fluent, so he had no idea what the PC was trying to tell him.
“After poking around in the BIOS, I found the culprit,”
Jackson revealed. “This particular model of motherboard had a ‘talking error
BIOS’ whereby certain POST codes triggered the playback of a friendly, spoken
error message, with Chinese set as the default language.”
Jackson found the relevant BIOS settings, changed the
default language to English, and the next time he rebooted the machine it
helpfully let him know: “Your floppy drive may not be connected
properly.”
In his mail to On Call, Jackson hypothesized that the PC’s
CMOS battery died, so the BIOS was unable to access its stored settings and
reverted to factory settings that assumed the presence of a nonexistent second
floppy drive.
“It triggered a feature I didn’t even know the motherboard
had!” Jackson told On Call.
Have you found yourself flummoxed by a feature you didn’t
know about? If so, click here to send On Call an email – we’ll assume that’s a feature you know well – so we
can tell your story on a future Friday. ®
If you’re looking for home and office furniture that will be part of your life for the long haul, Herman Miller is a go-to for well-designed, long-lasting furniture. Whether it’s upgrading your work-from-home setup with a better office chair or amping up your curb appeal with stylish new house numbers, Herman Miller can help. There are several pieces from Herman Miller that we love here at WIRED: the Herman Miller Embody is our “buy it for life” office chair, while Herman Miller’s Spout Sit-to-Stand standing desk is where I’m currently writing from. The brand is known for its thoughtful, high-quality pieces, but that quality level comes with a price tag to match. Luckily for you, you can pick up your own Herman Miller pieces for a little cheaper thanks to a Herman Miller promo code or a discounted Herman Miller bundle.
Refreshing your office? Say no more. Herman Miller has a bundle, aptly named the Better Home Office Bundle, that lets you bundle your choice of office chair, desk, and third item (a storage option, table lamp, or office accessory) to get 15% off the entire purchase. Use the Herman Miller promo code BUNDLE15 once you’ve put all three items in your cart.
Need it delivered? Get free delivery with the Herman Miller promo code FORYOU for any purchase over $2,000. You might worry you’ll need to purchase multiple items to hit that threshold, but many high-quality office furnishings (like our favorite office chairs from Herman Miller) will easily hit the minimum for you.
First-time shoppers at Herman Miller also get a special discount. You’ll need to sign up for Herman Miller’s newsletter, but a quick newsletter signup is well worth a Herman Miller discount code.
If you wish, you can get multiple discount coupons going at the same time on Herman Miller’s site. You can stack Herman Miller’s coupon codes—like the one you’ll get for signing up for the newsletter, or bundling three items together—with seasonal sales to get multiple discounts at the same time. Memorial Day is right around the corner, after all, and Herman Miller’s site already has sales going.
We named the Herman Miller Aeron office chair the “Best Office Chair for Long Hours,” because it’s durable, supportive, and airy’ plus, it’s also an iconic office chair that graces workspaces all over the world. A Herman Miller chair is an investment, which means it can get very pricey. That’s where Herman Miller refurbished chairs come in. They use genuine Herman Miller parts, including casters, tilt mechanisms, and Pellicle suspension material, which have been restored or replaced to original specifications. Each chair is backed by a five-year warranty, covering defects in materials and workmanship, with parts and labor included. And you can sit assured knowing that each Aeronchair has been rigorously tested to meet the original specifications for performance and ergonomic support.
Apple’s executives say the new-and-improved Siri won’t try to be your friend or your partner, because it’s only there to help you as efficiently as it can.
Apple’s latest attempt at modernizing its virtual assistant was the star of the show at WWDC 2026. Siri AI now supports natural language, contextual awareness, and advanced in-app actions, along with a chatbot-style experience.
However, even with its improved capabilities, Siri will always be different from existing AI assistants, thanks to Apple’s approach.
During an interview in the Mostly Human podcast, spotted by 9to5Mac, Apple’s software chief and head of worldwide marketing spoke about Siri AI. The two executives explained the role of Apple’s AI features and the ideas behind them.
Commenting on the significance of AI, Apple’s SVP of Software Engineering, Craig Federighi, likened it to the Industrial Revolution.
“I think that AI, in Apple’s view, can be used as this incredibly empowering thing. But there is a sense that, for one, things are changing very fast in ways that I think it’s hard for any normal person to keep up with,” said Federighi. “And while I think we can look at this as yet another thing of the scale of the Industrial Revolution, which changed our world in a giant way, that displaced a lot of people in the process and still occurred over like 80 years.”
He continues by saying he can relate to people who see AI as a big change. He believes it makes sense that people feel uncertainty about the potential impact of AI on jobs.
However, Apple remains optimistic. The company is looking for ways to make everyday tasks easier with AI.
“We don’t do AI for AI’s sake,” explained Greg Joswiak, “it’s ‘how does AI make everything better,’ and that makes our products better, our features better.” Joswiak is Apple’s Senior Vice President of Worldwide Marketing.
For Apple, the answer is to use AI to offer convenient, non-intrusive suggestion buttons in the Messages app. Users can add events to their Calendar app without engaging with a chatbot. In a sense, the company still views AI as a tool first and foremost, and not something that can replace social interaction.
Even with Siri’s new conversational abilities, Craig Federighi says the assistant never tries to make a connection with its users.
“If you use many of the existing chatbots, they’re really focused on engagement to a large degree, and sycophancy, right? They kind of want to pull you in,” says Federighi. “They might encourage you to reveal things about yourself and then use that as a basis to establish a connection.”
He suggests that Apple’s approach is quite the opposite. “The way that we have designed Siri, Siri really wants to say, ‘Listen, that’s not what I’m here for. I’m here to help you.’”
Siri won’t pursue romantic or emotional connections with its users, he said. “But if you try to engage Siri as a romantic partner, Siri’s not into that,” Federighi continued.
Greg Joswiak echoed Federighi’s statements regarding Apple’s approach to AI. In short, the company doesn’t prioritize user engagement with AI features.
“The motivations of Apple are different than some other companies,” said Joswiak. “Some people, their whole business model is ‘I need to keep you in what you’re in, I need to keep you in my app, my experience; that’s how I make my money’. That’s not us.”
Joz goes on to say that Apple’s motivation is to help the user, not keep them engaged. Delivering concise answers is the goal and it is up to the user to continue the conversation if needed.
In essence, the two Apple executives highlighted the company’s balanced idea of AI. Siri needs to understand natural language to help its users, not to build perceived emotional bonds with them.
Sometimes, as with the suggestions in the Messages app, Apple’s AI is a non-intrusive, optional utility. Simply put, Siri is not there to replace humans, their jobs, or their partners and friends.
Proof you can have it all? The JBL Xtreme 5 combines power with surprisingly coherence and insight. It’s a party speaker everyone can enjoy. Most probably would not want to make this a key feature of their living room but it’s proof of how much attention to detail goes into speakers not necessarily made for the most audio quality obsessed crowd
Powerful but tasteful audio
Highly water resistant
Portable, despite its size
Not a stylistic fit for an upmarket home
Limited codec support
Battery
Up to 28 hours of battery life
IP68 rating
Water and dustproof
Lossless audio
High quality audio through its USB-C input
The JBL Xtreme 5 is the 2026 entry to a long-standing speaker series. It’s JBL’s second largest mainstream Bluetooth speaker behind the Boombox, and is arguably the biggest one that primarily works as a portable unit.
You can attach a shoulder strap, where the heavier Boombox 4’s handle may be a bit much for anything beyond carrying short distances.
This is an excellent speaker than basically enhances what previous generations offered. Yes, it’s good for parties, particularly this year thanks to improved bass depth, but the JBL Xtreme 5 is also as coherent and insightful as speakers that on the surface appear to have loftier aspirations.
The JBL Xtreme 5 costs £329.99, just like the Xtreme 4 from 2024. Despite this new version’s improvements, though, you may also want to have a quick scout for deals on the older models — they can at times be found for far less.
The Xtreme series is a JBL family I’ve always felt I shouldn’t like. Its design, that loud name and the whole concept of at outright party speaker don’t really chime with what I look for in a wireless speaker. And yet JBL has always done such a great job, I’ve loved all the JBL Xtreme generations I’ve tried.


JBL hasn’t changed too much of what I find superficially off-putting in the JBL Xtreme 5, though, and even leans into some other areas further.
The JBL Xtreme 5 design, and that of its predecessors, is based around two massive passive radiators — one on each end. These help max-out bass production in smaller speakers, and leave the Xtreme with a shape not unlike a conga drum laid on its side.


It’s a fairly large speaker, but is still designed to be carried around, using the tough metal shoulder strap mounts on the top. Water resistance is excellent too, rated at IP68. This makes it an ideal poolside companion.
Much of the JBL Xtreme 5’s outer is a tough nylon fabric weave covering a plastic sub-frame, while the top panel and bottom foot have a rubbery finish.


There’s not a huge arount of metal here, but the strap mounts are metallic, as is the lettering of the JBL logo and the mildly obnoxious exclamation mark signs of the passive radiators’ outer.
As I’ve already said, this sort of bold teenage-baiting style isn’t my thing, but JBL’s execution of the idea is good. Each end of the Xtreme 5 also has tough rubbery feet, should you not have the room to let it sit lengthwise.
This year’s big change is LED lighting. Two thin strips of LEDs live at the bottom of the rubbery control panel, and where the speaker’s foot meets the nylon weave. These can be as subtle or obnoxious as you like, and are controlled using the JBL Portable phone app.


They are multi-colour LED strips, so can be plain white if you like, or can switch through tones using one of a few animated presets. Most have some form of animation to the light if it’s turned on, but the Freeze setting lets you pick a single, solid colour, for some less attention-grabbing mood lighting.
One of the JBL Xtreme series characteristics I’ve admired is how JBL has kept this series a largely focused, simple speaker. It’s a Bluetooth unit, not one that tries to pack in all sorts of smart features or multi-room streaming extras. They have offered in the larger Boombox series in the past.
That’s true once more, but JBL has still managed to cram in some neat extras.
As in previous years, there’s stereo pairing, and you can use the battery as a power bank. JBL earns bonus points for covering the battery slot with a panel held in place with a couple of TX screws for easier replacement.


Newer parts include Bluetooth Auracast and USB audio. You can easily plug the Xtreme into a laptop or PC, for fuss-free lossless audio transmission. This could be great for, picking a scenario out of the air, using it as the audio source for a projector setup while on a camping holiday.
Auracast is a technoloy that lets multiple devices receive audio from the single transmitter. It’s probably of more use for a couple of people who want to listen to TV or music with their own pairs of headphones, but can also be used as a more robust party mode — where you fling the same track over to a bunch of speakers.


It basically means companies like JBL no longer have to use their own proprietary multi-speaker pairing system, which is a pro-consumer move.
Battery life is rated at 24 hours, and while that won’t stand if you’re using at maximum volume to annoy everyone at the local park, it’s a long-lasting speaker nevertheless.
Within about five seconds of playing music through the JBL Xtreme 5 I could tell this speaker is just as special and capable and its predecessors, if not more so.
It manages the unusual feat of delivering the kind of bass depth and punch needed to satisfy someone out for a party speaker, while doing a frankly unlikely job of maintaining more navel gazing standards in other areas. It’s mostly about the JBL Xtreme 5’s unusually good mid-range coherence and clarity, in a category where one of the standard tactics is just to rely on pronounced bass and treble to provide thrills.
The factor I noticed within the first few seconds was how the JBL Xtreme doesn’t have over-emphasised mid bass and upper bass. I find this common-used audio thickening agent hangs off vocal like a sack of sodden clothing, and makes podcast voice lines sound as if there’s a weird bass resonance running underneath them.


There’s very little of that effect in the JBL Xtreme 5, and it does wonders for the separation of vocals, and the coherence of the sound in general.
It is by no means a bass light speaker, though. The JBL Xtreme 5’s low bass is particularly fun and lively for a speaker of its size. Its bass floor is rated at 40Hz suggesting it’s capable of some sub-bass rendering, and that’s pretty much what I hear here.
In the last Xtreme I reviewed I moaned about how JBL’s increase of the bass over the Xtreme 3 was to its detriment overall — from an audio snob’s perspective at least. But in this generation I think JBL has come up with a formula that really works for this line. It’s a classier kind of bass-heavy and fun sound signature, and that’s just what the JBL Xtreme 5 should have.
I’m talking about the default JBL Signature mode here, and there are a few others. Chill calms down the low-bass so you won’t keep people awake. It means the sound pulls less focus too. It’s a good “background music” mode.


JBL’s Energetic setting amps up the sound a little, with a little more spark than JBL Signature, but is far more reserved and usable than it might have been. Vocal predictably focuses on the mids, which doesn’t initially sound great when directly comparing to other modes, but could be useful for non-music content. There’s also a 7-band custom EQ mode.
No major surprise here. The JBL Xtreme series has always sounded better than you’d expect given it looks largely like a dumb party speaker, and the Xtreme 5 is no different. Nice work, JBL.
My main complaint is that the volume ceiling on this one is so high, single clicks of your phone’s volume control can bump the output up or down more than you want. The Xtreme 5 even enters an “AC” mode when plugged in, allowing for handful of extra decibels of volume. I don’t think it’s necessarily for most occasions, though. Battery power alone is powerful enough.
It’s loud, it sounds great, and while it’s nowhere near JBL’s most portable speaker, the shoulder strap makes the Xtreme 5 comfortable to carry around.
While the sound quality stands up to its top peers, the Xtreme 5’s design is pretty loud, and in this generation LED-lit.
The JBL Xtreme 5 is one of the more unlikely-looking great-sounding wireless speakers. It can bring the good to the party when needed, but also has remarkably clean-sounding mids and great overall coherence.
It has been the calling card of this family for years now, but the JBL Xtreme 5 jazzes things up a little with a couple of customisable LED light strips and Bluetooth Auracast — which lets multiple speakers or headphones share the same stream.
Business as usual for the most part, then, but business is good.
We test every wireless speaker we review thoroughly over an extended period of time. We use industry standard tests to compare features properly. We’ll always tell you what we find.
We never, ever, accept money to review a product.
Find out more about how we test in our ethics policy.
It has IP68 water resistance, good for most weather conditions and submersion in water.
This is a Bluetooth only speaker, but it supports Auracast for multi-speaker streaming.
There’s no classic 3.5mm aux input but you can plug in over USB for a cabled connection.
| JBL Xtreme 5 Review | |
|---|---|
| UK RRP | £329.99 |
| Manufacturer | JBL |
| IP rating | IP57 |
| Battery Hours | 24 |
| Size (Dimensions) | 346 x 165 x 155 MM |
| Weight | 2.9 KG |
| Release Date | 2026 |
| First Reviewed Date | 28/05/2026 |
| Audio Resolution | SBC, AAC, LC3 |
| Driver (s) | Two 0.75-inch tweeters, 3.9 x 5.7-inch woofer, two 20mm tweeters, 98 x 145mm woofer |
| Ports | USB-C |
| Audio (Power output) | 90 W |
| Connectivity | Bluetooth 6.0 |
| Colours | Squad, Black, Blue |
| Frequency Range | 40 20000 – Hz |
The big picture: Microsoft is easing one of the strict lines it previously drew around Copilot+ PCs, allowing more Windows 11 machines to run local AI workloads with the right GPU. An update shows that systems equipped with Nvidia RTX 30-series GPUs or newer, with at least 6GB of VRAM, can now support Windows’ local language model APIs. On paper, it’s a small, developer-only tweak, but it suggests Microsoft may be rethinking how tightly it ties on-device AI to Copilot+ branding.
When Copilot+ PCs launched on June 18, 2024, the messaging was clear: dedicated AI hardware was essential. These machines were defined in part by their neural processing units, along with baseline specs such as 16GB of RAM and solid-state storage. The NPU requirement, in particular, was positioned as the key to unlocking local AI features in Windows.
But NPUs are not the only chips capable of handling AI workloads. GPUs, especially modern ones, are built for heavy parallel processing and have long been used to run machine learning models. In practical terms, they can offer more raw throughput for many AI workloads than today’s NPUs, though typically at a higher power cost.
Until now, Microsoft had kept most of its built-in AI features tied to NPU-equipped devices. That left many powerful GPU-based PCs unable to access local text and image generation, as well as features like Windows Recall and other AI tools.
Now that gap is starting to close. In updated documentation and a GitHub post, Microsoft confirmed that developers can now run language model APIs on non-Copilot+ PCs using supported GPUs.
The company described the feature this way: “Language Model APIs on GPU [Experimental]. The Language Model APIs now run on non-Copilot+ PCs equipped with a supported GPU, bringing local language model capabilities to a broader range of Windows 11 devices.” It also specified that “supported hardware includes NVIDIA GeForce RTX 30 series and newer with 6+ GB VRAM.”
For now, this capability is still tucked inside the developer layer rather than exposed directly to everyday users. Running these APIs requires building or using applications that tap into the Windows AI framework. Still, it sets the stage for local AI to reach a much wider range of Windows machines.
At the center of this is a small on-device model called Phi Silica. Instead of being pre-installed on all systems, it can be downloaded through Windows Update when an app requires it. Once installed, the model runs locally on the machine’s hardware, using the GPU when available.
The current feature set is focused on text-based tasks. Through the Windows.AI.Text APIs, apps can summarize content, rewrite text, convert text into structured formats, and generate prompts. The functionality is similar to what users might expect from cloud-based AI tools, but it all runs locally on the device.
Running everything locally has practical benefits. It reduces reliance on cloud processing, which can improve responsiveness, and keeps data on the machine rather than sending it to external servers. For developers and enterprise users, that could make a meaningful difference in how AI features are adopted.
Still, the rollout is partial. Some of the more visible Copilot+ features, including Windows Recall and Click to Do, remain tied to systems with NPUs. GPU support, at least for now, is limited to the language model API layer rather than the broader suite of AI integrations.
Even with those limits, the trend is hard to miss. Microsoft is no longer treating NPUs as the only path to local AI on Windows. Letting GPUs take on these workloads broadens the pool of compatible hardware and makes Copilot+ PCs feel less exclusive than they did at launch.
Kyushu Electric Power Co., Inc. has disclosed a physical security incident that affects private data of more than 10 million customers.
In an official announcement, the company explains that the IT staff regularly performs backups to manage server storage. Due to capacity constraints, on April 27 an external storage device was used for the task.
The drive was then stored in a server room cabinet protected by multiple physical security layers. On May 26, when IT staff went to retrieve it, they found the cabinet had been left unlocked and the driver was missing.
Kyushu Electric Power Company is one of Japan’s major regional electric utilities, supplying electricity across the Kyushu region, which includes the prefectures of Fukuoka, Saga, Nagasaki, Kumamoto, Oita, Miyazaki, and Kagoshima.
The overall population of the Kyushu region is 12.6 million, and the company stated that the incident impacts up to 10.9 million accounts.
The data present on the now missing drive includes:
The firm has clarified that no bank account information or credit card data was stored in the drive. It also promised to notify impacted customers individually in the upcoming period.
Since the loss of the hard drive, the firm has interviewed all personnel who entered the server room and conducted investigations, but couldn’t locate it.
Media outlets report that 57 people had access to the said server room, and that Kyushu Electric filed a police report on June 4, suspecting someone had removed the drive.
NHK One reported that the Japanese Ministry of Economy, Trade, and Industry has given the firm until July 8 to report all the details about the incident and the preventative measures taken.
“The company is investigating all possibilities, including unauthorized removal of the device, but it has not yet been located,” reads the bulletin.
The incident has been reported to Japan’s Personal Information Protection Commission and the relevant government authorities.
Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.
The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.
An old algebra teacher used to say, “You have to take what you know and use it to get what you don’t know.” You might say the same thing about converting analog signals into digital. Computers know how to count and keep time. [Eric Explains] has a video purporting to explain “every type of analog-to-digital converter.” We aren’t sure he got every possible method, but there’s still a lot of information in the video, which you can see below.
From the flash ADC, using a ton of comparators to the successive approximation converter, which essentially plays a game of hi/lo, guessing the answer and figuring out if the real answer is higher or lower.
Those are pretty common, but the video also covers things like the Wilkinson ADC and other more exotic techniques. Each method, of course, has its advantages and disadvantages. For example, the flash ADC is fast, but requires a lot of components and power.
Sometimes, the method you use depends on how you are building. For example, you probably wouldn’t use a charge system on a breadboard since precision capacitors are finicky. But on an integrated circuit, capacitors made with photolithography may not be very precise, but the ratio between capacitors is super precise, making that a common technique in that domain.
Even if you never need to design your own converter, understanding the different architectures will let you make a better selection among alternatives. Then again, you can design your own. We’ve seen most of these architectures in past projects.
‘AI Overviews can no longer just be helpful summaries. Now, they must be legally defendable outputs,’ said Forrester principal analyst Nikhil Lai.
A landmark German ruling has determined that Google’s AI Overview are its own words, holding the company liable over statements Overview generated about two German publishers.
The Regional Court of Munich found that the AI Overview search tool was generating false claims about the two plaintiffs, including that they committed fraud and lured customers into subscription traps.
The court hit Google with a temporary injunction, banning the company from spreading false claims about the plaintiffs. Google has been ordered to pay 80pc of the legal costs, while the plaintiffs will each pay 10pc of the costs.
The court pinned responsibility on the search giant, reasoning that the AI used by Google independently compiles and summarises information, creating search results that go beyond just links.
It ruled that only Google has influence over the AI used in Overviews, as well as the algorithms with which it operates – meaning it must be held accountable for its results. It also said that the Overview search results about the plaintiffs included statements not even made in the search results.
“We invest deeply in the quality of AI Overviews to ensure that the overwhelming majority of responses provide accurate information, and they are designed to reflect the information that exists on the web. We’re carefully reviewing this decision, which is not yet final,” Google told Android Authority in a statement.
The Munich regional court’s ruling went a step further, examining existing rulings from Germany’s Federal Court of Justice (BGH), which gave search engines and autocomplete limited liability.
According to the BGH, search engines are only liable as indirect infringers because they disseminate information already created by third-party content publishers.
However, the Munich court said that this doesn’t apply to AI Overviews, as it “makes independent, new and substantive statements based on an evaluation and linking of content from various third-party websites” – rather than a traditional search engine which points to external sites.
Google argued in proceedings that users should not blindly trust information generated in AI Overviews. While the court agreed that users can check links and ensure the validity of the information they receive, it said that that shouldn’t relieve the company of liability.
“AI Overviews can no longer just be helpful summaries,” said Nikhil Lai, a principal analyst at Forrester. “Now, they must be legally defendable outputs.
“I think we’ll see fewer assertive, highly confident claims and more hedging, including language like ‘according to…’ and ‘some sources suggest…’.”
Lai also expects that fewer queries searching for sensitive information such as financial, health or legal advice would result in AI Overviews.
“This is not a Google-specific problem. I think this will lead to the value of defensible AI, where information’s verifiability and traceability become more valuable than its polish.”
Last year, a group of independent publishers filed an antitrust complaint with the European Commission, arguing that Google’s AI Overview diverts traffic away from independent publishers, resulting in less readership and advertising revenue.
Last week, the UK’s competition regulator ordered Google to let publishers opt out of having their content used to power AI features in search, including AI Overviews.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
It costs $200 and integrates with the $46 InkSense Plus.
Boox just announced a refresh of its popular Go 6 ereader, and this one offers stylus support for note-taking. The 6-inch E Ink reader integrates with the company’s pre-existing InkSense Plus stylus. The tablet costs $200 and the stylus costs around $45, so that totals $245 for the whole package. Preorders are open now and shipments go out on June 17.
There’s more here beyond stylus support, though that’s pretty useful. The updated Go 6 is available in four new colors and ships with a native note-taking app, which can be used with the stylus to “mark up books, underline passages and capture handwritten notes and to-do lists.”
There’s also been an uptick in RAM, up to 3GB, which should make everything a bit more responsive. The ereader features adjustable front lights, a 300 PPI monochrome ePaper display, 32GB of storage and a microSD slot. It’s on the smaller side, with a thin form factor and a weight of just over five ounces.
It runs on Android, so users have access to the Play Store. However, running apps on an ereader can often be frustrating, due to the low frame rate. Results will vary depending on the app. Like most modern ereaders, the Go 6 should last days on a single charge with regular use.
This isn’t the first time Boox has introduced stylus support to one of its products. The Go 7 E Ink tablet got the feature last year.
Weekend Open Thread: Evereve – Corporette.com
Jensen Huang Approves Samsung, SK Hynix, and Micron for NVIDIA (NVDA) HBM4 Memory Supply
The Best Mystery Series of All Time Is Surging on Streaming 30 Years After It Ended
Alexander Zverev wins the French Open to finally earn a 1st Grand Slam title
Anatomy of the June crypto crash: Fed, Iran, Saylor
Suspicious Polyfill login prompts pop up on Toshiba, Muji websites
Senator Cynthia Lummis Calls CLARITY Act the Most Consequential Financial Legislation of This Generation
Microsoft launches MXC, an OS-level sandbox for AI agents, with OpenAI and Nvidia already on board
Microsoft unveils seven homegrown AI models in new bid for ‘long term self-sufficiency’
(VIDEO) Justin Bieber Delivers Surprise Happy Birthday Serenade to Diners at Los Angeles Mexican Restaurant
Alexander Zverev conquers demons and outlasts Flavio Cobolli to win French Open for first major title
Von der Leyen’s AI envoy pick draws conflict-of-interest fire
High Stakes for Wembanyama as New York Pushes for 3-0 Lead
The Pain Points Taking a Fragile Tech Rally Down a Notch
Eli Lilly (LLY) Stock Surges 4% Following Breakthrough Sleep Apnea Trial Results
Meta steals a tactic from Tesla and builds data centers in tents
Hackers now exploit SolarWinds Serv-U flaw to crash servers
LBank Surpasses 25 Million Users Worldwide as AFA Partnership Continues to Drive Global Growth
Trump’s AI Ownership Plan Could Benefit Anthropic at OpenAI’s Expense
Notion restores access to Anthropic after service disruption
You must be logged in to post a comment Login