1 844.742.7448

Email Us

Remote Support

Contact

Software

  1. Home
  2. Software

Google explains why AI Overviews couldn’t understand a joke and told users to eat one rock a day – and promises it’ll get better
  • No Comments

If you’ve been keeping up with the latest developments in the area of generative AI, you may have seen that Google has stepped up the rollout of its ‘AI Overviews’ section in Google Search to all of the US.

At Google I/O 2024, held on May 14, Google confidently presented AI Overviews as the next big thing in Search that it expected to wow users, and when the feature finally began rolling out the following week it received a less than enthusiastic response. This was mainly due to AI Overviews returning peculiar and outright wrong information, and now, Google has responded by explaining what happened and why AI Overviews performed the way it did (according to Google). 

The feature was intended to bring more complex and better-verbalized answers to user queries, synthesizing a pool of relevant information and distilling it into a few convenient paragraphs. This summary would then be followed by the listed blue links with brief descriptions of the websites that we’re used to. 

Unfortunately for Google, screenshots of AI Overviews that provided strange, nonsensical, and downright wrong information started circulating on social media shortly after the rollout. Google has since pulled the feature, and published an explanatory post on its ‘Keyword’ blog to explain why AI Overviews was doing this, as mentioned – being quick to point out that many of these screenshots were faked. 

What AI Overviews were intended to be

Keynote speech at Google i/o 2024

(Image credit: Future)

In the blog post, Google first explains that the AI Overviews were designed to collect and present information that you would have to dig further via multiple searches to find out otherwise, and to prominently include links to credit where the information comes from, so you could easily follow up from the summary. 

According to Google, this isn’t just its large language models (LLMs) assembling convincing-sounding responses based on existing training data. AI Overviews is powered by its own custom language model that integrates Google’s core web ranking systems, which are used to carry out searches and integrate relevant and high-quality information into the summary. Accuracy is one of the cornerstones that Google prides itself on when it comes to search, the company notes, saying that it built AI Overviews to show information that’s sourced only from the web results it deems the best. 

This means that AI Overviews are generally supposed to hallucinate less than other LLM products, and if things happen to go wrong, it’s probably for a reason that Google also faces when it comes to search, giving the possible issues as “misinterpreting queries, misinterpreting a nuance of language on the web, or not having a lot of great information available.”

What actually happened during the rollout

Windows 10 dual screen

(Image credit: Shutterstock / Dotstock)

Google goes on to state that AI Overviews was optimized for accuracy and tested extensively before its wider rollout, but despite these seemingly robust testing efforts, Google does admit that’s not the same as having millions of people trying out the feature with a flood of novel searches. It also points out that some people were trying to provoke its search engine into producing nonsensical AI Overviews by carrying out ridiculous searches. 

I find this part of Google’s explanation a bit odd, seeing as I’d imagine that when building such a feature as AI Overviews, the company would appreciate that folks are likely to try to break it, or send it off the rails somehow, and that it should therefore be designed to handle silly or nonsense searches in its stride.

At any rate, Google then goes on to call out fake screenshots of some of the nonsensical and humorous AI Overviews that made their way around the web, which is fair I think. It reminds us we shouldn’t believe everything we see online, of course, although the faked screenshots looked pretty good if you didn't scrutinize them too closely (and all this underscores the need to check AI-generated features, anyway).

Google does admit, though, that sometimes AI Overviews did produce some odd, inaccurate, or unhelpful responses. It elaborates by explaining that there are multiple reasons why these happened, and that this whole episode has highlighted specific areas where AI Overviews could be improved.

The tech company further observes that these questionable AI Overviews would appear on searches for queries that didn’t happen often. A Threads user, @crumbler, posted an AI Overviews screenshot that went viral after they asked Google: “how many rocks should i eat?” This returned an AI Overview that recommended eating at least one small rock per day. Google’s explanation is that before this screenshot circulated online, this question had rarely been asked in search (which is certainly believable enough). 

A screenshot of an AI Overview recommending that humans should eat one small rock a day

(Image credit: Google/@crumbler on Threads)

Google continues to explain that there isn’t a lot of quality source material to answer that question seriously, either, calling instances when this happens a “data void” or an “information gap.” Additionally, in the case of the query above, some of the only content that was available was satirical by nature, and was linked in earnest as one of the only websites that addressed the query. 

Other nonsensical and silly AI Overviews pulled details from sarcastic or humorous content sources, and the likes of troll posts from discussion forums.

Google's next steps and the future of AI Overviews

When explaining what it’s doing to fix and improve AI Overviews, or any part of its Search results, Google notes that it doesn’t go through Search results pages one by one. Instead, the company tries to implement updates that affect whole sets of queries, including possible future queries. Google claims that it’s been able to identify patterns when analyzing the instances where AI Overviews got things wrong, and that it’s put in a whole set of new measures to continue to improve the feature.

You can check out the full list in Google’s post, but better detection capabilities for nonsensical queries trying to provoke a weird AI Overview are being implemented, and the search giant is looking to limit the inclusion of satirical or humorous content.

Along with the new measures to improve AI Overviews, Google states that it’s been monitoring user feedback and external reports, and that it’s taken action on a small number of summaries that violate Google’s content policies. This happens pretty rarely – in less than one in seven million unique queries, according to Google – and it’s being addressed.

The final reason Google gives for why AI Overviews performed this way is just the sheer scale of the billions of queries that are performed in Search every day. I can’t say I fault Google for that, and I would hope it ramps up the testing it does on AI Overviews even as the feature continues to be developed.

As for AI Overviews not understanding sarcasm, this sounds like a cop-out at first, but sarcasm and humor in general is a nuance of human communication that I can imagine is hard to account for. Comedy is a whole art form in itself, and this is going to be a very thorny and difficult area to navigate. So, I can understand that this is a major undertaking, but if Google wants to maintain a reputation for accuracy while pushing out this new feature – it’s something that’ll need to be dealt with.

We’ll just have to see how Google’s AI Overviews perform when they are reintroduced – and you can bet there’ll be lots of people watching keenly (and firing up yet more ridiculous searches in an effort to get that viral screenshot).

YOU MIGHT ALSO LIKE…

TechRadar – All the latest technology news

Read More

Meta’s own store has leaked the Meta Quest 3S – which could be the cheap VR headset you’ve been waiting for
  • No Comments

We might know the name of Meta’s next VR headset, as its own store has leaked the existence of the Meta Quest 3S.

Rumors have been swirling for some time that Meta is working on a cheaper version of the Meta Quest 3 – which has been called both the Meta Quets 3S and Meta Quest 3 Lite by people claiming to be in the know. The most recent leak revealed that it’s essentially the Quest 3’s brain (the Snapdragon XR2 Gen 2 chipset) inside the Quest 2’s bulkier body.

It seemed likely Meta might want to launch a cheaper VR headset, as the Quest 3 is a lot pricier than the old Oculus Quest 2. We think it’s worth the added cost, but it comes in at $ 499 / £479 / AU$ 799 instead of the $ 299 / £299 / AU$ 479 the 2 was at launch (and the $ 199.99 / £199.99 / AU$ 359.99 it is currently) which is a fairly hefty price increase.

Now Meta has all but confirmed the Meta Quest 3S is on the way by listing it as one of its headsets in the Quest Store.

The Alo Moves Quest Store page with the Meta Quest 3S name listed

(Image credit: Future / Meta / Alo Moves)

First spotted by UploadVR (and we’ve verified it), if you head to the Alo Moves XR page you’ll see that the app “Supports Meta Quest 3S, Meta Quest 3, Meta Quest Pro and Meta Quest 2.” Upload VR says it also appears on several other store pages – suggesting it’s not simply an error on the Alo Moves developer’s part – but we’ve not been able to spot it on any of the ones we’ve looked through. 

It’s possible many of these pages have already spotted and fixed the mistake, and we expect the Alo Moves reference to the 3S will disappear soon too.

Either way, this is as close to an official confirmation as we’ll get until Meta makes a proper announcement – something we had expected would happen soon as part of this year’s Meta Quest Gaming Showcase, a gaming-focused event that usually lands in June or July.

That official announcement is still going to be important as while this leak does seemingly confirm a Meta Quest 3S is on its way, it doesn’t tell us any details about what we can expect from the device. Yes, leaks have strongly hinted at a more affordable XR device but we won’t know for sure until we hear the details from Meta directly.

You might also like

TechRadar – All the latest technology news

Read More

Is this the return of Google Glass? Magic Leap and Google team up could be bad news for Meta and Apple
  • No Comments

We already knew Google was making at least a tentative return to the world of extended reality (XR, a catchall for VR, AR, and MR) with the announcement it’s helping to make the Samsung XR headset. But it could also be looking for another try and with one of its most high-profile flops: Google Glass.

This speculation follows an announcement from Magic Leap that it is partnering with Google to “bring Magic Leap Augmented Reality (AR) expertise and optics leadership together with Google’s technology platforms and collaborate on Extended Reality (XR) solutions and experiences.”

This is hardly a Google Glass confirmation, but it follows a few rumors that Google wants to have another crack at AR specs – including what might have been an accidental leak from its own Google I/O presentation. It also comes after Meta teased its AR glasses project, and with Apple testing the waters with the Vision Pro it would seem the entire industry is chasing an AR trend.

See more

Even though this partnership is seemingly set in stone we shouldn’t get our hopes up that Google Glass 2 is coming soon. LG and Meta announced plans to partner on XR technology, only for rumors to come out weeks later that they had already parted ways – rumors LG refused to dismiss.

Much like Meta and LG reportedly butted heads in several ways, Google and Magic Leap could also disagree on how best to create an AR device which could lead to their relationship breaking down. 

What could a ‘Google Glass 2’ look like? 

Assuming this partnership does bear fruit, what do we expect to see from Google Glass 2 – or whatever Google wants to call it?

Well design-wise we imagine it’ll look a lot more like a typical pair of specs. While Google Glass’ space-age design might have charmed some, it’s not at all subtle. The obvious camera freaked people out, and it painted a target on wearers as they were clearly wearing expensive tech that would-be thieves could rip from them. And when the battery dies, they’re useless.

Modern smart and AR glasses have some signs they’re more than meets the eye – like thicker arms, and a light that makes it clear when the camera is in use – but in general you wouldn’t know the specs were anything but regular shades unless you’re well-versed in tech. With prescription or shaded lenses, they’re also something you can wear all the time even when they run out of charge.

Orange RayBan Meta Smart Glasses in front of a wall of colorful lenses including green, blue, yellow and pink

The Ray-Ban Meta Smart Glasses (Image credit: Meta)

As for its features, obviously, AR aspects would be included. To us, this means a HUD with an overlay showing you things like directions pointing you towards your destination, and apps that have you interact with virtual objects as if they’re in the real world.

But the other big feature will most likely be AI. We’ve already seen how the Ray-Ban Meta smart glasses can leverage cameras and its AI to help you identify objects, find recipes, translate signs, and generally answer questions you might have by simply talking to your specs. Google also has a generative AI – Gemini – and while its recent attempts at AI search haven’t been the best, we’d be shocked if this tech wasn’t incorporated into Google Glass 2.

We’ll have to wait and see what Google’s next AR device has in store for us if and when it launches. You can be sure we’ll be ready to give you the lowdown as soon as we have it.

You might also like

TechRadar – All the latest technology news

Read More

Windows 11 update introduces a long-awaited feature – alongside something else that might not go down so well
  • No Comments

Microsoft has rolled out an optional update for Windows 11, KB5037853, that will let you try an array of new features which includes support for drag-and-drop functionality in File Explorer’s address bar. The arrival of the latter ability has been long-awaited by some folks, so there’ll be some pleased Windows 11 users out there today.

If you’re one of those keen to try this – or any of the other features packed into KB5037853 – you can now download and install this preview update. As it’s an optional update, it won’t be downloaded automatically, and you’ll have to click ‘Download & Install’ in Windows Update (in the Settings app) to kick off the installation process. 

Apparently this optional patch is a pretty sizeable update, and along with the revamped File Explorer, it also introduces an ‘Account Manager’ for the Start menu. When you’re in Windows 11 signed into your Microsoft account, the latter will appear to highlight account benefits and make it easier to manage settings, Microsoft explains – though some may regard this as unnecessary clutter.

It sounds suspiciously like treading that thin line between ‘help’ and ‘advertising’ that Microsoft has been prone to err on the wrong side of lately.

Person using a laptop in a coffeeshop

(Image credit: Shutterstock)

File this one under 'finally'…

The drag-and-drop feature for File Explorer has been introduced before and pulled back due to buggy performance, and then introduced for gradual rollout and testing again, so it’s great to see it finally about to arrive (barring any last-minute disaster).

The feature lets users drag a file into File Explorer’s address bar (at the top) to move it to a new location. It’s a simple feature and convenient shortcut that a fair few people are looking forward to seeing, and indeed they are likely wondering why this basic functionality wasn’t present in Windows 11 in the first place.

In this update, Microsoft has also addressed a problem with File Explorer where it could take as long as two minutes to open if you pinned a specific folder to Quick Access that happened to be shared on a network. 

At any rate, we can hope the drag-and-drop feature is here to stay in File Explorer, and Microsoft is considering making more changes like this that make users happy. Adding more AI features that don’t necessarily improve the core functionality of Windows 11, and pushing ads in as many parts of the operating system as possible, are moves that are turning some people off Windows 11 – so it’s good to see Microsoft focus on improving the core Windows user experience here (that more dubious Start menu addition aside).

If you don’t want to download an optional update – as it may have problems, because it’s still in testing, after all – then you can simply wait for the  June 2024 Patch Tuesday update, which should contain all these changes, but with any wrinkles ironed out (in theory anyway).

Via Windows Latest

YOU MIGHT ALSO LIKE…

TechRadar – All the latest technology news

Read More

Google’s Gemini Nano could launch on the Pixel 8a as early as next month
  • No Comments

Google promised back in March that it would eventually bring Gemini Nano to the Pixel 8 and Pixel 8a although no one knew exactly when. Until now, that is, as the update may arrive “very soon.” 

Android Authority recently did a deep dive into the Pixel 8 series’ AICore app and found a pair of toggle switches inside the settings menu. The second switch is the main focus because flipping it turns on “on-device GenAI features.” Activating the tool presumably allows a Pixel 8 smartphone to harness Gemini Nano in order to power the device’s generative AI capabilities. 

It doesn’t actually say “Nano” in the accompanying text, though; just “Gemini”. However, we believe it is that particular model and not some stripped-down version, after all this is what Google said it would. Plus, the Pixel 8 trio runs on the Tensor G3 chipset and it can support the AI with the right hardware adjustments.

No one knows exactly what the AI will actually do here, though. Gemini Nano on the Pixel 8 Pro powers the phone’s “multimodal capabilities,” including, but not limited to, the Summarize tool in the Record app and Magic Compose in Google Messages.

Imminent launch

The other toggle switch isn’t as noteworthy – a screenshot in the report reveals it enables “AICore Persistent.” This gives applications “permission… to use as many [device] resources as possible”. 

Android Authority states that the sudden appearance of these switches could mean Google is almost ready to “announce Nano support for the Pixel 8/8a ”—maybe within the next couple of days or weeks. The company typically releases major updates for its mobile platforms in June, so we expect to see Gemini roll out to the rest of the Pixel 8 line next month. 

According to the publication, the toggles will likely be found in the Developer Options section of the smartphone’s Settings menu. However, it's important to note that this could change at any time.

Technically-minded users can find the switches by digging around the latest AICore patch. The software is available for download from the Google Play Store; however, your Pixel 8 model may need to be running Android 15 Beta 2.1. 9To5Google, in their coverage, claims to have found the AICore toggles on a Pixel 8 with the beta but not on a phone running Android 14.

As for the Pixel 8 Pro, it's unknown if the high-end model is going to receive the same update although there's a chance it could. Android Authority points out it's currently not possible for Pro users to deactivate Gemini Nano, but this update could give them the option.

Be sure to check out TechRadar's list of the best Pixel phones for 2024.

You might also like

TechRadar – All the latest technology news

Read More

ChatGPT’s free tier just got a massive upgrade – so stop paying for ChatGPT Plus
  • No Comments

Following its GPT-4o announcement during its Spring Update event, OpenAI has finally made its new AI tools available for free to everyone, begging the question: Is there any point paying for ChatGPT?

With ChatGPT-4o, all users can now access more advanced tools like discussing files and photos you upload to ChatGPT. The generative AI can also conduct data analysis and create charts, and it can access the internet to inform its responses. 

However, with all of these features rolling out to everyone – even if they come with usage limits for non-paying users – there’s a big question of if people should stay subscribed to OpenAI’s premium tier for ChatGPT.

It’s not like ChatGPT Plus has become entirely obsolete. Subscribers still have exclusive features like the ability to create custom GPTs, higher usage rate limits with 4o, and first access to new features – this includes early access to Voice Mode when it launches “in the coming weeks.”

But it’s understandable why subscribers feel a little burned. They’re paying $ 20 (around £16 / AU$ 30) per month for a service that’s not that different from the free one. Unless you’re an AI power user, now’s seemingly a terrible time to sign up for ChatGPT Plus.

Thinking long term

A close up of ChatGPT on a phone, with the OpenAI logo in the background of the photo

(Image credit: Shutterstock/Daniel Chetroni)

So why would OpenAI want to make its premium service less appealing? Well, there are two prevailing theories.

The far-fetched one is that OpenAI will soon release an early version of GPT-5, or at least some kind of exciting new features that’ll be exclusive to its paid members beyond the voiced version of ChatGPT. It’s not out of the question, though this feels like something OpenAI would have mentioned during its Spring Update event on May 13, so color us skeptical.

The likely reason is that OpenAI is changing track to focus on bringing in as many users as possible, rather than paid ones, at least for now.

That’s because a report recently revealed that hardly any of us use ChatGPT and other AI tools in our day-to-day lives. If OpenAI wants people to get excited by its tools it can’t then lock the best features away behind a paywall.

What’s more, ChatGPT’s rivals – like the Meta AI and Google Gemini – are free to use and offer many of the same premium tools at no cost. If it’s already a struggle to get people to use AI when it’s free, you can bet it’s significantly harder with a paywall in the way.

We’ll have to wait and see if ChatGPT Plus gets any improvements in the coming weeks, but if you’re currently subscribed (or thinking of joining) you might want to hold off for now.

You might also like

TechRadar – All the latest technology news

Read More

Truecaller’s new feature can turn your voice into a personal secretary
  • No Comments

Caller ID service Truecaller is giving users the ability to create a digital assistant that has their voice and can respond on their behalf. If you’re unfamiliar with the app, Truecaller launched its AI Assistant feature in 2022 to screen phone calls and take messages, among other things. Up to this point, it utilized pre-made voices, but thanks to the power of Microsoft’s Azure AI Speech, you can now use your own.

Setting up your voice within Trucaller is quite easy; you just need a subscription to Truecaller Premium, which is $ 9.99 a month per account. Once that is set, the software will immediately ask you to select an AI assistant – but instead of picking one of the pre-made personalities, select “Add your Voice.”

You’ll then be asked to read a consent sentence and a brief training script out loud into your smartphone’s microphone. Doing so ensures the AI has a voice that mimics your “speaking style.” When done, Truecaller states that Microsoft’s Azure Custom Voice begins to process the recording to create a “high-quality digital replica.” The app will give you a demo sound bite to help you imagine what it’ll sound like when someone calls you. 

Truecaller's training script

(Image credit: Truecaller)

Robo-voice

Keep in mind the technology isn’t perfect. While the digital assistant may sound like you, it does come across as rather robotic. The company published a YouTube video on its official channel showing what the AI sounds like. Admittedly, the software does a decent job at mimicking a person’s vocal inflections; however, responses still sound stiff. That said, it is an interesting and interactive way to screen calls as they come in, especially when stopping spam ones. 

Keep an eye out for the patch when it arrives, as we tried to create our own digital secretary on our Android but couldn’t since we didn’t receive the feature as of yet. It’s unknown exactly when and where the update will be available. TechCrunch claims the tool will roll out “over the next few weeks” as a public beta across a small selection of countries. These include, but are not limited to, the US, Canada, Australia, and Sweden. Soon after, it’ll become widely available “to all users in the eligible markets. 

We also reached out to Truecaller with a couple of questions, including how recordings are stored, whether they are saved on the device or uploaded to company servers, and more. If we hear back, this story will be updated.

While we have you, check out TechRadar's round up of the best encrypted messaging apps on Android for 2024.

You might also like

TechRadar – All the latest technology news

Read More

Next Windows 11 update might include a new PC optimization app – and a sneaky way for Microsoft to introduce ads
  • No Comments

We’re expecting the next major upgrade for Windows 11, the 24H2 update, to roll out widely in September or October of this year, and it’s possible that it could include Microsoft’s system optimizing app for Windows 11 (and Windows 10) called ‘PC Manager.’

The app was originally made for Windows 10 and Windows 11 users in China but has since spread to some other regions. Where it’s available, users can download PC Manager from the Microsoft Store if they want, but now it appears that Microsoft is going one step further by including the app in the 24H2 update.

At least Microsoft has introduced PC Manager in the latest Windows 11 preview build in the Beta channel, but only in China right now. This is quite a large step forward because it means the app is now one of the default inclusions in Windows 11 for Chinese users, as Neowin reports.

Logically, in the future, Microsoft may make PC Manager a default app in other regions, perhaps including the US. Indeed, the official page for the app in the US now allows Windows 11 and 10 users to download it – though of course, that’s a far cry from it becoming one of the stock apps in Microsoft’s operating systems.

woman working from home with child

(Image credit: Shutterstock)

Managing expectations

PC system optimizers have been around for a long time in the form of third-party apps, and it’s very interesting to see Microsoft offer up an official take on this kind of utility, built right into Windows, that many people think is pretty good.

The app promises to improve your system’s performance with typical actions like suggesting ways to free up memory and disk space, checking for potential threats to your PC like viruses or other issues, and reducing ads and pop-ups. The last aim is a little ironic seeing as a recent version of PC Manager was shown to suggest that users ‘fix’ their system by changing their default search engine to Bing. This hardly seems like a repair or a quality-of-life improvement, but just another shoehorned attempt at prompting users to switch to Microsoft products dressed up as a ‘suggestion.’ 

Neowin also tells us that in the past, PC Manager has been shown to inject affiliate links to products promoted by Microsoft, and recommend questionable actions in terms of optimizing your system. So, take that as something of a caution.

Conceptually, PC Manager sounds helpful or even necessary, offering assistance like quick system clean-up solutions, and protecting your default settings from unauthorized changes. In practice, it looks like Microsoft is experimenting with making some not-so-helpful (or even potentially harmful) additions.

I hope that Microsoft reconsiders this and puts the effort in to make it a great app, as that’s what’s best for Windows users and what will foster goodwill. It may also be the case that the changes made for China – including PC Manager becoming a default app in Windows 11 – may never be applied to the US market (or other regions).

YOU MIGHT ALSO LIKE

TechRadar – All the latest technology news

Read More

Windows 11’s Recall feature is already running on unsupported CPUs – and it shows why this is a bad idea
  • No Comments

Windows 11 enthusiasts are already playing with the new – and controversial – Recall feature the OS now has in preview (for 24H2), and have got it running on current Arm-based CPUs by fudging things.

While Recall is present in the recently released preview of the Windows 11 24H2 update, Microsoft makes it clear that the feature won’t work on current PCs, as it requires a Copilot+ PC (the new name for the ‘AI PC’).

In other words, Recall needs a device with a powerful enough NPU to run it (and other new AI features in 24H2), which is currently only the new Snapdragon X chips (and AMD plus Intel CPUs further down the line).

Even those Snapdragon laptops aren’t available just yet (they will be next month), but leaker Albacore has still managed to tinker under the hood of Windows 11 24H2 and get Recall working on a current Arm processor.

See more

You can see a video of Recall being summoned on a standard (non-Copilot+) Windows 11 PC in the above post on X (formerly Twitter). As Albacore says, it shows ‘screenray’ in action which refers to the context-sensitive mode entered when you find something using Recall and select it.

As you can see, if the search result you want is a text file, screenray presents options pertaining to what you’ll need to do – copy and paste text. Or if it’s an image, you’ll get choices to copy the picture or open it for editing in an app.

Analysis: Working as not intended

It’s pretty cool to see this feature working on a processor without the necessary strength in terms of a powerful NPU (like the new Snapdragon X silicon sports), but at the same time, it illustrates why that NPU is needed. As you probably noticed, the interaction with Recall and screenray therein looks a bit laggy here – what the NPU does is provide specific AI acceleration to ensure this process runs more smoothly.

Furthermore, the feature is still in testing within a preview build here, and that won’t help either.

Albacore even sounds hopeful about getting Recall working not just on current ARM chips, but also on existing AMD and Intel (x86) CPUs, which also can’t officially run the feature. (Again, even current-gen processors from Teams Red and Blue lack an NPU with enough raw grunt).

If that happens, we can expect a similar experience to what we see here – but it’s not possible yet anyway, as Microsoft has only provided the machine learning model bundles for Arm to laptop makers. These don’t exist for AMD or Intel CPUs yet (as there’s no need for them – Lunar Lake and Strix Point, which will drive Copilot+ PCs, are still some way off launching, but they are fully expected to debut before 2024 is out).

Ultimately, this is an interesting fudge for now, but it’s likely a bad idea to be trying to get Recall up and running on a PC it’s not intended for. Simply because there may well be scenarios where it truly bogs down – such as when you have a larger, sprawling library of snapshots piled up – and there are doubtless good reasons why Microsoft has the mentioned NPU requirement in place.

Mind you, not everyone wants Recall, anyhow: certainly not those more privacy-conscious Windows 11 users out there who have already made their feelings clear. Indeed, a privacy watchdog in the UK is already investigating Recall before Microsoft even has the functionality officially live. The result of that enquiry will certainly be interesting, and Microsoft may be worried about another scenario where a big Windows 11 feature is blocked in Europe due to more stringent data regulations.

You might also like…

TechRadar – All the latest technology news

Read More

Microsoft’s trick for speeding up PC games in Windows 11 works with only 12 games to start with – but far more are actually supported
  • No Comments

We now know a lot more about how Microsoft’s Automatic Super Resolution (Auto SR) feature for speeding up gaming frame rates in Windows 11 will work, and what games it will initially support.

VideoCardz noticed a new entry in Microsoft’s support database on the topic of Auto SR, which underlines the requirements, as well as detailing what games will come as fully verified for the tech.

For those who missed it, Auto SR is an upscaling feature, meaning it runs a game at a lower resolution, upscaling to a higher one, so that you get a close-to-native-resolution image quality with a faster frame rate – using AI to pull off this trickery.

The notable catches are that you need a Copilot+ PC and indeed a Snapdragon X processor, one of the ARM-based chips that’ll power laptops launching next month. (You’ll also need Windows 11 24H2, which launches with those AI PCs).

As for the games which are verified and tested by Microsoft for Auto SR, the initial collection is as follows:

  • 7 Days to Die
  • BeamNG Drive
  • Borderlands 3
  • Control
  • Dark Souls III
  • God of War
  • Kingdom Come: Deliverance
  • Resident Evil 2
  • Resident Evil 3
  • Sekiro: Shadows Die Twice
  • Sniper Ghost Warrior Contracts 2
  • The Witcher 3: Wild Hunt

Analysis: Useful clarifications – and caveats

It’s interesting to see the fully verified games, and even if it’s only a small selection of a dozen right now, there are some big-name titles. However, the really interesting bit is the clarification that Automatic Super Resolution is a sweeping upscaling feature that can be applied to any game (DX11 or DX12).

We always assumed it would be a system-wide feature – after all, that was the whole point, compared to more targeted upscaling solutions that require support from the game dev such as Nvidia DLSS – and indeed this is the case. It’s just Microsoft worried us with its mention of Auto SR just applying to a “curated set of games” last week when it launched the feature, but these are just the verified games guaranteed to work well.

The majority of games should be fine with Auto SR in theory, but some may be wonky, or some may not work at all, and to that end, Microsoft is collaborating with the Worksonwoa.com website that lists games that can use the feature successfully – and also those that can’t use Auto SR for whatever reason. (This is the same website that also tells you whether your favorite PC game will run on Windows on ARM).

There are some nuances to note here, and the first is that verified games are set to work ‘out of the box’ with Auto SR, meaning the feature will be on by default. That could cause some confusion or conflict if a gamer is using another type of upscaling potentially – though you are told that by Windows that Auto SR is being enabled when the game is launched.

We guess Microsoft feels that less tech-savvy folks will benefit from having the feature automatically applied where it makes sense, in games that are fully tested to work well with Auto SR.

The Snapdragon X requirement is the other important point to note here, although we assume this will be widened to include future AMD and Intel laptop CPUs – those with a powerful enough NPU to qualify as the engine of a Copilot+ PC (as Auto SR will be for these PCs only).

However, we also noticed that Microsoft says Auto SR is only supported for games running on ARM64 natively or emulated x64 games (with the latter using Prism, the translation layer for running Windows games on ARM chips). Presumably that’s a reflection that currently (well, as of next month) only the new Snapdragon X can drive a Copilot+ PC, and that when AMD Strix Point or Intel Lunar Lake CPUs arrive for these AI-powered laptops, there’ll surely be fine with Auto SR.

You might also like…

TechRadar – All the latest technology news

Read More