Feed aggregator

The former Steelers coach built the sprawling city brick by brick with a video game console controller.

70mai 4K A900 Ultra: two-minute review

While many of the best dash cams on the market come with plenty of features and functions, not all of them boast 4K shooting capability. The dual-channel 70mai 4K A900 Ultra captures 4K footage not only with its front-facing camera, but also with the rearward-facing unit that makes up this potent dash cam bundle.

This camera combo benefits from Sony Starvis 2 IMX678 sensors in both units too, which combined with the 4K, HDR and 70mai’s own Night Owl Vision 2.0 processing, means the bundle is perfect for capturing footage night and day. The resolution is impressive, with the f/1.8 front lens capable of shooting 3840 x 2160P (4K) at 30fps, and the f/1.6 rearward camera able to shoot the same 4K quality at 25fps. The field of view for both units is a generous 130 degrees.

The rest of the specification is similarly impressive, with the chunky main camera featuring a 3-inch rear-facing screen and four control buttons, which makes adjusting settings and previewing footage easy. There is a 5-system built-in GPS, and buffered emergency recording that captures up to 3 minutes and ten seconds before and 30 seconds after an incident.

Bonus appeal is delivered with the likes of 70mai’s Smart Parking Surveillance 2.0 (subject to hardwiring the unit into a vehicle), a Driver Fatigue Alert, and 4G Cloud+ 2.0 compatibility that delivers a live view, remote playback and two-way voice communication that covers all bases on the security front.

While capturing 4K footage soon amasses lots of chunky files, the 70mai 4K A900 Ultra features Wi-Fi 6, which means the transfer of footage and general file management is both quick and easy, with speeds of up to 45MBps. Another plus of the 70mai 4K A900 Ultra comes via compatibility with 70mai’s Battery Pack Mini, which gives assured backup power, and up to 30 days of parking surveillance without draining your vehicle’s battery.

(Image credit: Future)

Topping off an impressive set of features is hands-free voice control, while the unit itself features supercapacitor technology that offers more durable performance across the ranged of temperatures that occur inside vehicles over time.

All of this is squeezed into a compact box-shaped design for the main camera unit, while the rearward-facing camera is a fairly typical tube-shaped design. Both units can be attached to screen glass using supplied adhesive pads, and everything needed for the job comes inside the box.

The 70mai 4K A900 Ultra is therefore a one-stop solution, although it is worth noting that a hardwire kit, which unlocks the potential of around-the-clock surveillance monitoring, will need to be purchased separately. Otherwise, the units can be powered via a standard 12V in-car power socket, for which an adapter is supplied. The 70mai Battery Pack Mini is also sold as an optional extra.

70mai 4K A900 Ultra: price and availability

In the US, the 70mai 4K A900 Ultra is priced at $299.99, but at the time of writing can be purchased directly from the US 70mai website for $239.99. In the UK, the 70mai 4K A900 Ultra is priced at £231, but is currently available for £185 from the 70mai website.

70mai 4K A900 Ultra: specs70mai 4K A900 Ultra dash cam specs

Video

Front: 3840 x 2160P (4K) 30fps

Rear: 3840 x 2160P (4K) 30fps

Field of view (FOV)

130 degrees front / 130 degrees rear

Storage

MicroSD up to 512GB

GPS

Yes 5 system

Parking mode

Yes, with constant power cable, not included

App support

70mai app

Dimensions

3.62 x 2.36 x 2.02 inches / 9.2 x 6.0 x 5.14cm

Weight

0.52lb / 0.24kg

Battery

Yes

70mai 4K A900 Ultra: Design

I’ve tested several dash cameras from the 70mai stable, and always found them to be well designed and very nicely made. The 70mai 4K A900 Ultra follows that theme, with a good-looking exterior for the main camera unit that sits upfront. Meanwhile, the tube-shaped rear-facing camera looks fairly conventional, but is also very well put together.

The main camera unit feels great in the hand, weighty and chunky. It also features a built-in mounting bracket, which is non-detachable. The angle-adjustment bracket that joins the camera unit, and the sticky attachment that fits to the windshield felt quite stiff to adjust, but this makes sense when you consider how much vibration and bumping around the unit will be subjected to over time, and should mean that the position of the camera remains consistent over time.

Up front, the lens protrudes from the front of the camera unit, but not so far as to be intrusive when fitting to the windshield. The side of the unit has a slot for microSD cards up to 512GB capacity, and the port for connecting the power and rear camera cable. I like the way 70mai does its connectors, with a push fit for the camera wire that feels almost locked in place once it has been pushed in. Again, having a durable design for components like this means there’s less chance for failure over time.

There’s a 3-inch screen on the back of the front dash camera unit, which is a good size for previewing footage and the like, as well as making the setup process nice and easy. It took me mere minutes to power-up the camera using the supplied 12V power plug and cable. From there, picking through the installation options and setting the language, time and date options was easily achieved via the four plastic buttons directly below the screen.

(Image credit: Future)

Both cameras can be fixed to the glass using the supplied self-adhesive pads. In the case of the front unit this has to be applied to the bracket and then stuck to the screen. There are static sheets for anyone who wants to stick these to the glass first, which isa good option if you think you’re going to need to peel it off and move to another vehicle at some point. The rear camera has its sticky pad already fixed to the bracket.

Getting this into position in a rear window can be tricky, depending on the vehicle, but the good news is that the lens barrel can be rotated into its optimal position, and the view it captures checked within the front-facing screen on the front. This is presented as a picture-in-picture view, so any tweaks can be made if the view doesn’t look quite right. 70mai does sell an optional hardwire kit for anyone wanting to access the entire suite of features and functions, including round-the-clock vehicle surveillance.

(Image credit: Future)70mai 4K A900 Ultra: Performance

I found the 70mai 4K A900 Ultra dash cam very easy to use once I’d got it installed and ready to go. As mentioned, the initial setup procedure is very straightforward, and can largely be done without consulting the manual thanks to the simple step-by-step process, which includes formatting the memory card. I especially liked the way the chunky buttons on the back of the camera could be used for making adjustments. Combined with the 70mai app, which I’ve used before, this combination makes light work of everyday file management duties.

Which brings us to the subject of the video. The 4K quality produced by this twin-channel camera combination is excellent, and I found the after-dark footage it captured was as good as footage shot in daylight. The 70mai engineers have clearly put some effort into squeezing the maximum detail from their HDR software, and the results are clear to see in the clips on this page.

The cameras also coped admirably with variable light quality and weather conditions, as well as sudden changes in lighting such driving through a tunnel, as illustrated in the clips here. Importantly, there's good definition in details such as license plates and road signs.

One of the more interesting features of this dual-channel dash cam is its use of low-voltage differential signaling, or LVDS. 70mai claims this means the 4K feed from the rear camera is relayed as a digital signal, with no re-encoding, which ultimately means there's no compression carried out, even though there is quite a long cable connecting the two camera units. 70mai also says the connection features 20ms ultra-low latency. In essence, that should mean that if you're using the feed from the rear camera when reversing, you'll see it in real time, without any lag. On a more practical note, the more efficient cabling seems thinner than usual too.

(Image credit: Future)

Overall, the 70mai 4K A900 Ultra dash cam is a potent package that left me wanting for nothing. Naturally, the files are chunky, but the latest Wi-Fi transfer capabilities of this bundle and a rock-solid app means file management really isn’t an issue. Granted, the microSD card tops out at 512GB before it starts looping, but I found that transferring files from the camera to cloud-base storage was about as easy as it gets.

As always, one important thing to note is that anyone wanting to use the parking-monitoring feature will to purchase the optional hardwire kit.

(Image credit: Future)

you'll need the optional hardwire kit. Th

Should you buy the 70mai 4K A900 Ultra dash cam?

(Image credit: Future)Buy it if...

You want top-quality video
There's no doubting the capability of the 70mai 4K A900 Ultra dash cam, thanks to its 4K shooting power. Better still, you get similarly potent video capture from the rear camera too.

You value extra features
The 70mai 4K A900 Ultra dash cam comes with a host of added-value features, including excellent after-dark capability, and parking surveillance via the optional hardwire kit.

You want more power
The 70mai 4K A900 Ultra dash cam can be boosted by adding 70mai's Battery Pack Mini as an option. This avoids any worries about battery drain, and makes the twin-camera setup even more versatile.

Don't buy it if...

Twin cameras are more than you need
Lots of us tend to think driving standards have been slipping, and so all-round protection is useful. But if you crave simplicity then a single-camera setup will suffice.

You're looking for a budget option
The 70mai 4K A900 Ultra dash cam is best described as a mid-range model, and while it is competitively priced, there are cheaper bundles out there.

Fitting a dash cam is a chore
The downside of the 70mai 4K A900 Ultra dash cam is that it requires twice as many cables to be plumbed in behind trim panels. A single-channel camera might therefore be a better bet.

How I tested the 70mai 4K A900 Ultra dash cam

• I installed the dash cam in an SUV for a test period of around four weeks
• I used it during daily commutes as well as for longer trips on a mix of roads
• I installed the 70mai app and used it to view, manage and save 4K files

70mai loaned me the 4K A900 Ultra dash cam along with everything I needed to test it, including a microSD card.

I installed the twin-camera bundle in a test vehicle using the supplied cables and connectors. I didn't call on the option of using a 70mai hardwire kit, which offers a more permanent means of supplying power, and enables parking-surveillance features. I didn't get to try the 70mai Battery Pack Mini either, which is another option offered by the company.

I did, however, make good use of the supporting 70mai app. I've tested this before, and it's received a few tweaks here and there since I last used it. As a result, the app works very well indeed, especially when it comes to managing the larger files generated by the higher-quality 4K setup.

• First reviewed October 2026

The Legend of Zelda: Ocarina of Time.'/>

We got to play The Legend of Zelda: Ocarina of Time early. Here are our impressions of the remake and its elaborate new style.

(Image credit: Nintendo)

HP's OfficeJet Pro line is a fixture in our home office printer coverage, and the 8125 is the wireless all-in-one built for that same desk-clearing, everything-in-one-box crowd.

This inkjet model for the home office handles printing, scanning, and copying with an automatic document feeder and duplex printing built in, plus a 3-month Instant Ink trial and HP's AI-enabled smart features for things like mobile scanning and print management.

And right now, the HP OfficeJet Pro 8125 is down to $100 (was $180) at Amazon. At that price, it's an easy recommendation for anyone looking for a cartridge-based printer for work.

OfficeJet Pro 8125: was $179.99 now $99.99
Print, scan, and copy with automatic document feeder (ADF) support and duplex printing on both sides of the page automatically. Wireless connectivity through the HP Smart app, plus a 3-month Instant Ink trial included.View Deal

Why we recommend it

The OfficeJet Pro series has a long track record in HP's lineup, and the 8125 carries forward the features that make that line dependable for home office use: an automatic document feeder for multi-page scans, duplex printing that doesn't require manually flipping pages, and wireless setup through the HP Smart app that doesn't demand much troubleshooting.

The three-month Instant Ink trial is a nice inclusion at this price, giving you a low-risk way to see whether HP's subscription ink model saves money for your printing habits before deciding to continue or switch back to cartridges.

Price Context & Historical Value

This is the lowest price I can remember seeing this printer for - back in August, it dropped to $120. It's a meaningful discount for a duplexing all-in-one with wireless support, and the cut here looks like a real markdown rather than an inflated list price. For comparison, it's also down to $100 over at Walmart.

Should you buy it?

✅ Buy the HP OfficeJet Pro 8125 if...

You want a reliable all-in-one for a home office, you print double-sided documents often and want automatic duplexing, or you like the idea of trying Instant Ink before committing to a subscription.

❌ Skip the HP OfficeJet Pro 8125 if...

You print high volumes of photos and want a dedicated photo printer, you'd rather avoid cartridge-based ink costs long-term, or you're looking for a unit we've already tested hands-on.

• See all printer deals at Amazon.

• Octopus Energy has revealed the Cosy cooler add-on
• It's a unit that works with the Cosy heat pump to transform your central heating system into air conditioning
• A unit cools one room and costs £500, with multiple units needing a higher-tier Cosy heat pump — trials are underway now

Octopus Energy has developed an add-on for its heat pumps which means that central heating can be transformed into an air-con system for the increasingly hot British summers.

The invention is called the Cosy cooler — following the same branding as the Cosy heat pump — and it's just been launched, with trials having begun in UK homes (and offices), and more testers being sought via Octopus' official form.

The Cosy cooler is the size of a "tall kitchen bin" at 72cm tall, with a width of 43cm, and depth of 35cm. It simply connects to the existing central heating radiator pipework (and heat pump), taking about three hours to fit (with no "major" building work needed, Octopus assures us).

The system chills the water going through the radiators, and it also takes in warm air from indoors, cooling it using a heat exchanger, then circulating that chilled air back out into the room. So there's a double whammy of cooling here, and one unit can cool a single room, or you can run two to three units in different rooms (although you'll need a higher-tier Cosy heat pump to run multiple Cosy coolers).

Okay, so what's the damage for staying cool next summer? Octopus expects to sell these cooling units for £500 to £600, which is a lot less than a full air-con system, of course (which run from £2,000 upwards, as the energy provider notes).

Greg Jackson, who is CEO and founder of Octopus, commented: "Sadly, British homes need to be ready for freezing winters and increasingly hot summers." He added that: "The Boiler Upgrade Scheme has helped give the industry the confidence to invest, innovate and drive costs down — and Cosy cooler is the latest result of Octopus ingenuity."

Octopus recently launched plug-in solar kits, too, and they've sold out twice now, but more kits are on the way, we're told. That's another piece of the puzzle that could help keep your heating (or cooling) costs down.

Cosily chilled

(Image credit: Octopus Energy)

This is a clever idea and one that will doubtless grab the attention of anyone who suffered in the heat of the past summer in the UK (and that's a lot of people, no doubt).

The climate in this country seems to be shifting radically, and the heat of the last summer led to portable air-con units and fans selling out in many cases, with prices of the devices which were left in stock spiraling rapidly upwards.

There's probably going to be a lot of interest in these Cosy cooler units, seeing as they provide what sounds like a much more comprehensive cooling system than what you'll get for £500 or so in the way of a portable air-con device.

Octopus is aiming to sell these Cosy coolers to customers next summer, but for now, as mentioned, what's happening is that you can apply for a trial. If you want to do that, head to the official page here, but obviously bear in mind that you need to be running the Cosy heat pump. This isn't a standalone product, rather it works off the existing heat pump (and note that it's a prototype currently).

If you want more than one cooling unit — so you want to cover perhaps your bedroom as well as your living room — then it sounds like you'll need one of the beefier Cosy heat pumps, too (they come in three flavors: Cosy 6, 9 and 12).

• IMF warns that an AI market correction could have major economic implications
• Circular investments are making tech companies more intertwined
• A major cyberattack could impact multiple organizations at once

The International Monetary Fund has warned global financial markets are becoming increasingly vulnerable to the volatility of AI investments.

While AI-related investments have becoming an increasingly important driver of economy activity, even helping to offset some of the economic damage associated with ongoing geopolitical upheaval, AI holds the potential of causing major market disruption.

Specifically, the IMF warned of three potential developments – a correction in AI-related investments and company valuations, an increase in government bond yields, or a major cyberattack targeting critical national infrastructure, including financial institutions.

Could an AI market correction lead to widespread disruption?

While the IMF acknowledged the AI market has helped sustain large parts of the world's economy, but pointed toward high company valuations and exceptionally strong revenues. The report also warns of two-way investments with tech companies both giving and taking from each other and becoming increasingly reliant on each other.

To quantify this, the IMF wrote: "Our review of 14 such transactions since 2023 found about $300 billion in equity investments and $650 billion in debt, in exchange for $1.4 trillion in future revenue commitments."

Also involving the tech sector, the IMF worries that a potentially catastrophic cybersecurity incident could spell disaster for the world's economy. With financial systems more digital, automated and interconnected than ever before, cybercriminals now have the opportunity to attack multiple institutions and organizations simultaneously.

Ultimately, the IMF's conclusion is that isolated incidents are no longer so isolated – complex dependencies mean that one single incident could have much broader impacts.

Jump to...

Flagship desks
Storage & workstation desks
Space-savers & budget picks

Too much sitting takes a toll on the body. Even with the most ergonomic chairs, there’s a limit to how much you can sit without experiencing some lower back pain. Luckily, there’s a solution that lets you work productively without being glued to a chair. It’s the standing desk, available in various shapes, sizes, and price ranges.

For most people, I recommend the FlexiSpot E7 Pro for $530 (was $640) at FlexiSpot. This desk sailed through our tests under review. It's well-built, stable, includes cable management and a host of sizes and styles for any work area. And it's well-priced compared to the competition, too.

But it's not the only one on sale right now. So, I’ve checked for the top standing desk deals, discounts, and sales on out top sit-stand models and those at prices I think offer genuine value for money. For more top picks, we've tested out all the best standing desks.

Best standing desk deals: Quick links

• Amazon: Standing desk deals for every kind of workspace
• Best Buy: Deals on budget standing desks
• Walmart: Standing desk deals from under $100
• Eureka Ergonomic: Seriously stylish executive standing desks
• FlexiSpot: Big office chair and standing desks deals
• Herman Miller: Rare savings on the office furniture icon
• Uplift: Save up to $670 on standing desks
• Vari: Save on well-built desks and chairs

Flagship standing desk deals

E7 Pro with bamboo desktop: was $639.99 now $529.99
Height: 25.8 – 51.4in | Desktop: 48 x 24in

The E7 Pro is my top overall standing desk, with a flawless 5-star rating in our review. I've included the price here for the frame and bamboo desktop, but you have different options on the FlexiSpot site. View Deal

ComfortEdge Electric Standing Desk: was $899 now $799
Height: 25 – 50.5in | Desktop: 60 x 30in

This remains one of the best standing desks we've ever tested (it scored 4.5 stars), but what stands out here is the waterfall edge which makes it more ergonomic and comfortable when resting your wrists. For the full discount, use code SAVEMORE26.View Deal

4-Leg Standing Desk: was $959 now $859
Height: 25.3 – 50.9in | Desktop: 48 x 30in

This four-leg standing desk scored top marks in our 5-star review, where our reviewer called it "elegant, well-built, customizable, and exactly what nearly everybody who is looking for a standing desk actually needs."View Deal

Magnus Pro: was $988 now $948
Height: 25.6 – 49.2in | Desktop: 59.1 x 27.6in

Secretlab's sit-stand gaming desk is truly stylish and well-built. We rated it 4 stars and loved its elegant, practical built-in cable management. Yes, it's pricey, but we found it "perfect for those seeking a versatile and organized workspace."View Deal

Standing desk deals with storage

Comhar desk with drawers: was $349.99 now $259.99
Height: 28.3 – 47.6in | Desktop: 48 x 24in

Earned 4.5 stars in our review, a 28.3 - 47.6-in height range, and it's got what so many sit-stand desks lack - a drawer. For the price, it's an easy recommendation. View Deal

Electric Standing Desk With 2 Drawers: was $229.99 now $169.98
Height: 28.3 – 46.5in | Desktop: 55 x 26in

This standing desk comprises everything most modern desks need - a broad 55-inch desktop seats two drawers and a monitor riser. Underneath, there's a keyboard tray, freeing up more space. View Deal

Standing Desk With Monitor Riser & Cable Management: was $299.99 now $259.99
Height: 28.9 – 46.5in | Desktop: 55 x 28in

Another winner from FlexiSpot, I chose this because alongside the monitor riser up top, wide enough for two displays, there's a cable management feature for clearing desk cluttering and trailing wires. View Deal

Space-saver and budget standing desk deals

EN1: was $159.99 now $109.99
Height: 28.9 – 46.5in | Desktop: 48 x 24in

FlexiSpot always reviews well in our tests, and this desk boasts a stable single-piece desktop, USB ports, and 28.9 to 46.5in height range. Easily the best budget choice I know of, and one I recommend. View Deal

L-Shaped Electric Corner Standing Desk: was $299.99 now $259.99
Height: 28.5 – 46.5in | Desktop: 55 x 63in

If you have the room in your office or home office, an L-shaped corner desk has proved a good way to boost your usable workspace area without wasting space.View Deal

Small Electric Standing Desk: was $89.99 now $76.98
Height: 28.3 – 46.5in | Desktop: 32 x 19in

I wouldn't recommend this desk for everyone, but if you're working in a small space, it's ideal. the desktop measures 32-inches, making it suitable for more compact setups at an excellent price. View Deal

We've covered in depth what to look for from a standing desk, but one question we see come up a lot is what sort of height range should you be looking for in a sit-stand model. Here's what we've found based on our tests.

User height

Desk height

Under 5'4"

23 - 42in

Between 5'4" and 6'

25 - 47in

Over 6'

28 - 51in or above

I’m not easily impressed when it comes to smart home hubs, here’s why the HomeView is different.

Right now, the SanDisk Desk Drive is the cheapest way to get fast storage at an affordable price in a compact form factor. Something you can carry around easily, making it particularly useful for professionals, creators and anyone looking to manage huge files (or large number of small files).

SanDisk sells it direct for just $899.99, that's just $112.50 per TB, not as cheap as the Teamgroup QX 4TB ($346 at Amazon or $86.50/TB) but more affordable per unit storage than all other SSDs, internal or external.

Even more telling, the next cheapest 8TB SSD is the Samsung 870 QVO which sells for $1,480 or $185 per TB, almost 60% more expensive. In other words, nothing comes close to this external drive when it comes to value for money right now

Desk Drive 8TB SSD: was $1100 now $900
This is a perfectly compact desktop SSD designed to sit on your desk (or carrying with you). In our own tests transferring large photo and video files, we found it quick to complete, and very simple to set up and use. View Deal

Why we recommend it

We scored the SanDisk Desk Drive 4 stars in our review, noting that the SSD "provides both speed and capacity, making it a great solution for individuals to use in homes, small offices, studios, and all other workspaces."

It's a faster high-capacity alternative to a HDD that performed very well in our tests. During our own benchmarking, we copied 1TB of images (a mix of 1200 RAW and JPG files) from an OWC CFExpress Type-B card for use in Photoshop. Transfer took just under two minutes.

We then used the drive to feed 4K video files into Final Cut Pro, and experience no dropped frames, with video editing proving a smooth experience.

We would've liked to see more software bundled with the drive, and an extra USB-C port would be even more welcome rather than a proprietary power port. But we can't fault the overall simplicity and performance.

Price context & historical value

This is a fair deal, in my opinion, so worth picking up based on prices from other retailers. It's not an all-time low - that came in March 2025, where it was priced at $469.99. The historic high was in April this year, where it was selling for $1,600. Still, it's a good price for a reliable high-capacity SSD.

Should you buy it?

✅ Buy the SanDisk Desk Drive if...

You're looking for a compact desk SSD with lots of storage and fast file transfers. In our CrystalMark we clocked read speeds at 1032.48MB/s and write speeds at 1010.49MB/s, which is enough for most uses.

❌ Skip the SanDisk Desk Drive if...

You have absolutely no need for 8TB storage - you can find external SSDs with 1TB and 2TB for much cheaper (but none can match its per TB price). It's also not the fastest speeds compared to something like the latest model of the SanDisk Extreme SSD.

The Catch: What to know before you buy

Many people don't need that much storage, so while the deal price is good, a simpler 1TB drive will be much cheaper for general file and data storage. While it's very compact, it's not as slim-line as the SanDisk Extreme range, so it's slightly bulkier in a laptop bag plus you will need an external power adaptor.

• See all external SSD at Amazon

Ukrainian President Volodymyr Zelenskyy has a sit-down interview with NPR as Russia's full-scale invasion nears the 5-year mark. And, Pope Leo XIV has been diagnosed with lung cancer.

• Eufy's new home security camera sticks to your window to let you monitor the outside view from inside
• It's $79.99 / £69.99, subscription-free, and offers 2.5K resolution
• The only drawback is its power cable, which might restrict where you're able to place it

Eufy’s new security camera isn’t just a small, handy smart home tool that removes the hard work from installing security devices in outdoor spaces — it’s also affordable and, crucially, subscription-free.

Hot on the heels of the Eufy Video Doorbell S4, the first video doorbell with VR viewing, the Eufy E10 Window Camera is the latest device in the lineup, priced at $79.99 / £69.99 each, or $149.99 / £129.99 for two. It’s already available in the US, with UK pre-orders set to ship from November 6; it's listed as "coming soon" at Australian retailers, with pricing TBC.

Unlike many of the best home security cameras, the Eufy E10 Window Camera is a compact device at 2.69 x 2.69 x 1.65 inches. It's designed for window placement, and sticks easily to glass, allowing you to monitor your driveway and front garden from inside your home, so there are no mounting tools, hard labor, or wall damage involved.

Features-wise, the camera, which offers smart integration with Alexa and Google, offers a detailed 2.5K resolution, with what Eufy describes as a ‘wraparound’ design. In addition to its 25-degree adjustable downward-tilt mechanism, which minimizes empty upper-frame space and allows you to capture ground-level footage, the camera also comes with anti-reflection stickers to prevent glare from windows or other sources of light that may obstruct the view.

The Eufy E10 Window Camera is also completely subscription free — a feature that more users are starting to look for in their smart-home devices.

(Image credit: Eufy)

Instead of charging you a monthly fee to store security footage (as rivals such as Ring and Blink do), Eufy’s security camera supports microSD cards up to 256GB. Additionally, the camera supports 24/7 video recording, with 2.4 GHz and 5 GHz dual-band Wi-Fi for a reliable and steady connection.

Eufy’s new security camera is simple, but effective — and it doesn’t compromise on smart features for its size either, supporting an AI identification model that distinguishes between people, pets, and vehicles.

The only real downside here is that the Eufy E10 Window Camera is wired, which limits where you can place it — you’ll have to mount it on window that's within range of a power source. This also means you’ll have a wire running from your window, which isn't ideal for keeping things discreet.

On the device’s product listing, Eufy says it’s seen a high demand for its new window camera, and as a result it will be processing orders on a first-come, first-served basis.

• Report warns Europe's data center push ultimately means the sector will need more electricity
• Data centers only pay for their direct upgrades – not broader grid upgrades
• Government intervention is likely needed to protect billpayers

New Moody's data has warned that Europe's growing data center footprint could severely impact energy prices across the continent as we're already seeing in the US.

As of 2025, Europe's data center-related electricity consumption stood at around 2.5% of the entire continent's consumption, however within the space of a decade (by 2035), this could more than double to 5.7% across countries like France and the UK.

But more worryingly, the countries that are set to play host to the most capacity will likely see even starker rises, including the Northlands (17%) and Ireland (30%).

European data centers could push energy prices up

Moody's explained that, while data centers would typically pay for any upgrades needed for their direct connection, the grid operator usually picks up the bill for wider network reinforcement. Those additional costs, as well as a tighter supply-demand imbalances, could push costs onto the end user – you and I.

Bt the report also asks whether this could just be a short-term period of discomfort for billpayers, because the ongoing AI boom could lead to further growth opportunities among utilities companies. With extra revenue and profits, they could invest in new (and clean) generation and transmission infrastructure.

Still, with imminent price hikes potentially on the cards, Moody's anticipates government intervention so that consumers don't end up being hit with the most severe rises. Tax cuts, price caps and stricter rules for data center connections are among the most viable solutions to keep a lid on costs.

With data centers arguably having a much more profound impact on billpayers in the US and the launch of the Trump administration's Ratepayer Protection Pledge, it remains to be seen whether Europe will follow suit and how its nations will protect their citizens.

Via The Register

• AI Scout tracks which artificial intelligence services employees actually use at work
• Employees are increasingly turning to AI services outside approved channels
• NordLayer AI Scout can warn, redact, or block sensitive prompts before submission

Employees are increasingly using AI tools at work, including services their employers may not have approved or formally provided.

That creates uncertainty about which AI platforms workers access, what information they submit, and how those systems connect with business operations.

NordLayer says its new AI Scout tool will now help give organizations visibility into workplace AI activity and the security risks associated with it.

Seeing actual AI use inside companies

The standalone product monitors browser and desktop interactions while providing controls intended to limit exposure of confidential information across systems.

AI Scout can identify authorized and unauthorized AI services, allowing organizations to compare observed activity against software approved for workplace use.

Its reporting measures usage from real activity rather than purchased subscriptions, and it also estimates token consumption across deployed LLMs.

It also examines information employees send to AI systems, using controls designed to detect potentially sensitive material before certain prompts are submitted.

The company says these safeguards operate before submission, potentially preventing confidential material from reaching an external AI service.

NordLayer also says browser-based controls can intervene when employees enter information requiring additional scrutiny before transmission.

The monitoring extends beyond conventional chatbot use into AI agents and their technical relationships with connected services used at work.

Administrators can inspect MCP infrastructure linked to programming agents and review the capabilities available through those systems in practice.

“We built NordLayer AI Scout for how AI actually works — not just what it looks like from the outside, but for what’s happening at a technical level,” says Vakaris Noreika, VP of product research and development at NordLayer.

“Its AI DLP is designed to inspect prompts and control them pre-submission, to ensure that sensitive data is never inputted in the first place.

Why workplace AI needs closer oversight

The need for monitoring comes as business investment in artificial intelligence grows alongside employee use of unapproved services, a practice known as shadow AI.

Writer's 2026 enterprise survey found that 59% of companies spend more than $1 million annually on artificial intelligence technology.

A University of Melbourne report cited by NordLayer found that 42% of employees use approved AI systems, while 70% use free models at work.

IBM reports that breaches involving shadow AI add $400,000 to the global average breach cost of $4.99 million worldwide.

Those figures help explain why organizations may struggle to maintain visibility when employees use AI services beyond formally approved platforms.

“What companies are facing today is an AI zoo — dozens of unsanctioned LLMs in daily use across the workforce, each one widening the organization’s exposure to data leaks and security breaches,” says Andrius Buinovskis, VP of product strategy at NordLayer.

“True AI governance requires extensive visibility. With new LLM solutions emerging at a rapid pace, reviewing every new AI tool manually simply doesn’t scale — organizations need telemetry that brings all AI use, sanctioned or not, into one platform.”

Every week, another headline celebrates how AI made some task faster. A developer wrote code 55% faster. A marketer generated a first-draft blog post in minutes. A product manager summarized a 200-page requirements document in seconds. These are real numbers.

But they are the wrong numbers to celebrate.

If you step back and ask the harder question — what actually creates enterprise value? — task speed is rarely the bottleneck. Where the value traffic gets jammed is almost always upstream, in the question that precedes the task:

It’s not “What do we want IT to build?”

It’s “Why do we want IT to build it?”

It’s not “Why do we need that blog post?”

It’s “What content will drive business impact?”

It’s not “Which process should be optimized?”

It should be “How do we address the quality escape generating rework and scrap?” or “How do we prevent supply chain delays compounding into missed delivery commitments?”

These are questions that AI still cannot answer for you. And the gap between task automation and outcome generation is the defining challenge of this agentic era.

The sneaky way AI isn’t helping

Harvard Business Review put a name to this particular AI gap earlier in 2026: the micro-productivity trap. Companies optimize individual tasks, generate the code faster, summarize the report faster, route the support ticket faster, without ever rethinking the workflows or value-creation models those tasks belong to.

The result is isolated productivity gains that never compound into business impact. Almost like a value mirage. PwC’s 2026 AI Performance study found that 74% of AI's economic value is being captured by just 20% of organizations. That’s the micro-productivity trap in effect. It’s what’s separating AI leaders from the rest.

That AI leader wheat is only going to further separate itself from the chaff of other organizations continue to “do AI” for isolated productivity gains. Gartner’s warned that over 40% of agentic AI projects will be cancelled by the end of 2027, because organizations fail to define the outcome they’re building toward in the first place.

The companies that succeed, PWC found, are focused on growth and value capture. That encompasses a lot more than just productivity. I’ve noticed that a lot of organizations have a problem with prioritization. We are asking AI to go faster in the wrong direction. It’s like thinking that it’s ok to take the wrong exit off the highway just because you’re driving a Bugatti.

What's the real unlock to AI value?

The key to true AI value is in your knowledge. It’s knowing which blog topic will move an audience before a single word is written; whether the document will deliver business value before it’s created, which process to target, which data points to connect.

These are upstream decisions. They require business-domain intelligence and the understanding of what matters, what costs the most, and what a breaking change on the production floor actually means for the P&L. Pure task automation cannot supply that. Only people who sit at the intersection of business and technology can.

The solution to wringing every bit of value from AI is not a new one. Just an evolution of one. I’m talking about fusion teams, a cross-functional unit combining the domain knowledge of operational and business experts with the architectural judgment of professional developers, moving at the speed of the problem rather than the speed of the IT backlog.

What’s core to the fusion team is that they represent problem identification and faster business-technology alignment, not just going faster. That concept needs to be carried over into AI agents—how they’re built and how they collaborate across workflows.

For agents to truly work and help you avoid those micro-productivity traps, you need to connect contextualized data to working machine learning models so that they cannot just “go faster” but also make the highest quality of decisions.

It's all in the question

Success with AI—not that value mirage that micro-productivity can create—requires domain experts (e.g., plant engineers, process managers, and supply chain leaders, business analysts, procurement specialists, claims adjusters, etc.) at the table alongside IT and data teams.

It requires people who understand which product lines carry the thinnest margins, which supplier delays have the most downstream production impact, which quality control steps are creating the longest cycle time.

In other words, the fusion mindset: the discipline of defining the outcome before designing the solution, and measuring against business value rather than task throughput.

The organizations capturing the majority of AI's economic value in 2026 are not the ones with faster, AI-driven processes. They are the ones who identified the right operational problems first, governed their data well enough to trust it, and built the organizational frameworks to coordinate agents and people against outcomes that actually move the business.

So for your next AI initiative, re-frame the question: Are you building toward a fundamentally different outcome, or are you just automating a task?

We've featured the best AI website builder.

This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.

The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

• Pax8 finds over a third (39%) of UK SMBs have experienced unexpectedly high AI bills
• Fewer UK orgs have a formal internal AI review process than German or US orgs
• Around one in five couldn't even tell apart specific AI spend from generic tech budgets

New research by Pax8 has criticized UK businesses for building AI faster than governance can keep up, with a quarter (24%) of SMBs having discovered unauthorized AI tools operating within the business in the past six months.

Two in five (39%) have also experienced unexpectedly high AI bills, implying that the visibility across their AI consumption isn't sufficient.

But with the pressure to adopt AI at an all-time high, UK businesses could continue to be rushed into rash decisions without thinking out the fundamentals first.

UK businesses lack end-to-end AI visibility

Even though shadow AI is apparently widespread, just one in four (23%) UK businesses have a formal, documented review process for internally built AI, compared with around a third in Germany (32%) and the US (36%).

In light of that, only around half (54%) of UK SMBs are developing their own AI, putting them way behind the likes of the US (81%).

British businesses are also struggling on the cost management front, with a tiny 12% having dedicated AI budgets. A further 19% couldn't even distinguish AI spend from other tech spend, leaving just 16% of UK SMBs confident in their abilities to determine AI's ROI – around a third of the number of businesses in Sweden (45%) that feel confident in doing so.

Looking ahead, security concerns remain the most cited barrier to adoption, but Pax8 frames the current state of AI adoption as a major opportunity for software vendors and IT service providers. Rather than creating an action point for businesses, who are the consumers in this instance, the report implies that the providers have a greater responsibility to reassure their users with additional visibility, security and governance controls, and cost management options.

"These findings highlight the importance of trusted businesses that can help SMBs bring agents, workflows, data and tools together as a connected intelligence layer," EMEA CEO Harald Nuij.

• Nothing's CEO has called the iPhone Duo "a waste of money" for having a premium price while lacking core features
• These include a telephoto camera, a good selfie camera, and Face ID
• Despite that, there's enough demand that we're seeing pre-order bots emerge — which you absolutely shouldn't use

While it’s unclear exactly how much demand there is for the iPhone Duo, it's sure to be a popular phone — in fact, there’s a good chance it will be the best-selling foldable phone ever. But the CEO of a rival brand thinks it’s not particularly good value.

Carl Pei — the CEO of Nothing and co-founder of OnePlus — has posted a short YouTube video calling the iPhone Duo “a waste of money” due to its lack of key iPhone features. The Duo doesn't feature a telephoto camera or Face ID, and ships with a worse front-facing camera than you get with the iPhone 18 Pro.

Pei sums up the situation by saying “it’s $2,000 and costs more than an iPhone Pro with fewer features… so you’re paying more and you’re getting less.” And he’s not wrong — although the extra screen gives the Duo at least one feature not found on the Pro models.

Pei additionally highlights the Duo's folding animation as being what most people seem to be excited about, rather than its hinge or cameras (he notes that, while it’s a nice animation, “it’s just an animation at the end of the day” — i.e., not really a reason to buy the phone).

Still, he credits Apple’s attention to even the smallest details — such as that folding animation — for being the reason the Duo is likely to do so well.

Personally, I think it’s more simply because it has an Apple logo on the front — and because the vast number of iPhone users will be more likely to use a foldable that runs iOS than one that runs Android. But either way, criticisms could certainly be leveled at the phone's specs and value.

Don't get scammed

In any case, it’s likely the iPhone Duo will be in high demand when pre-orders open on Friday, and it might also be in short supply, with analyst Jeff Pu suggesting (via 9to5Mac) that “hinge constraints” are causing production slowdowns.

As such, some buyers might be tempted to take drastic steps to ensure they get an iPhone Duo on day one, which explains why pre-order bots are being advertised to help purchase the phone — with some even claiming to secure multiple units, which might appeal to resellers.

But as Francis Fong Po-kiu, honorary president of the Hong Kong Information Technology Federation, said to The Standard, “even if you succeed in securing the phone, the follow-up cost, including stolen personal or credit card data, is extremely high.”

After all, not only are these bots typically not free to use, but they’re likely to require your payment information to place the order — putting them in a position to then steal from you if the developers are unscrupulous. Additionally, they could include malware, giving them access to other sensitive information on your device.

So, they’re simply not worth the risk, and as Malwarebytes points out, you should also watch out for fake pre-order pages.

Instead, you should ensure you’re ordering the iPhone Duo directly from Apple or other reputable retailers, and if you want to get ahead of the game, you can both input your details ahead of time using Apple’s ‘Get Ready’ feature, and make sure you place your order as soon as pre-orders open at 5am PT / 8am ET / 1pm BST on Friday, 16th October.

AI Agents are unlike any software development witnessed before. Traditionally, trustworthy software systems have followed a structured approach to being built and developed. One that plans, builds, tests, and maintains software to ensure its reliability and scalability. This created the Software Development LifeCycle (SDLC) (1) framework.

But AI agents are different: they are a team of autonomous, probabilistic actors that work most effectively in a real-time, event-driven environment, enabling the agents to coordinate with real-time data to develop, learn, and work.

To scale an enterprise workforce of AI agents, organizations need to combine the principles of employee management with software development rigor. These agents are not static software that can be coded and developed, due to probabilistic language models. They are built, evaluated, and governed differently.

This requires organizations to move from traditional SDLC to an Agent Development LifeCycle (ADLC) (2), which utilizes the structure and repeatable process of the SDLC framework and combines it with the development and management of AI agents.

A helping hand from HR

Since AI agents perform roles within an enterprise, the ADLC should draw on lessons from the human employee lifecycle.

When an organization brings on a new team member, there is a deliberate process: their role is defined; they’re given access to the systems they need; they are trained; and they are integrated into a team. Once they are empowered and put to work, a manager oversees their activity, monitoring and developing their performance over time.

As autonomous actors operating within a system of governance, AI agents deserve and require that same treatment.

AI agents need a similar structured approach to onboard and integrate them into a business and deliver the efficiency benefits they’re designed for – that means they need real-time contextual data to do their jobs!

Treat AI agents like humans at the hiring stage

Picture an organization where every employee is a generalist, there’s no hierarchy or structure, and all communication happens via synchronous one–to–one phone calls. No email. No Slack. No shared systems.

The organization can’t scale. Information doesn’t flow. People drown in data from every direction, with no real understanding of what’s going on or what to do. There are no accountability structures, no access controls, no coordination mechanisms beyond individual conversations.

You wouldn’t drop a new human hire into this chaos. Effective organizations look different, counting on people who are specialized and organized hierarchically, so coordination happens in parallel and communication is premised on asynchronous channels that let information flow without requiring everyone to be online at the same time. The same goes for onboarding AI agents.

AI agents and employees are no different in the eye of an agent mesh

Most early agentic AI deployments work as complex monolithic agents tasked with doing everything, accessing all data, communicating synchronously, and trying to maintain very large context or working synchronously with sub-agents. The result? A system that’s brittle, expensive, inconsistent, and completely unable to handle real enterprise complexity.

This is where an agent mesh can act as a development and runtime platform, helping build AI agents that actually work for a real-time enterprise, specialized for particular functions, organized hierarchically, and orchestrated by a development and runtime platform that specializes in delegating tasks to appropriate agents.

Communications between agents are asynchronous and event–driven, and access to data and systems is governed by role-based access controls, giving agents exactly the permissions they need and nothing more.

So, let’s look at how agents can be “onboarded” through an ADLC much like a new employee joining the organization:

1. Treat agent setup like drafting a job description

Before a human employee starts, you write the job description. You define what role they’ll fill, what they’re responsible for, what behaviors you expect, and what boundaries they operate within.

An agent mesh’s agent builder can use an internal AI agent to provide a guided interface for defining an agent’s purpose, scope, and configuration. You set up the instructions and system prompts that shape the agent’s persona and behavioral parameters. And you configure guardrails – hard constraints on what the agent can and cannot do – that prevent it from going off–script or taking unsafe actions.

2. Get the agent up to speed with tool and system access

A new employee’s first weeks are spent getting access to tools, systems, and data. Agent onboarding is the same.

An agent mesh should include a number of data connectors – pre-built integrations to enterprise databases, data warehouses, data lakes, enterprise application APIs, and MCP servers. This enables agents to access real-time data from day one.

Access is provisioned through role-based access controls (RBAC), enforcing least-privilege principles so agents only see what they need to see. It can also define skills that go with internal tools so agents know best how to use them.

3. Passing the probation period – stress-test agents before deployment

After getting access to systems, human employees undergo training. This is where general ability becomes company and job-specific skills. You wouldn’t put a new hire in front of customers without first verifying they know what they are doing. The same logic applies here.

An agent mesh can provide an Eval function that uses AI to suggest tests for agents, allows you to add more and then run them against your agents. This empowers the organization to validate agent competence against defined success criteria during development. This supports both initial testing and regression tests as you make changes.

4. Keep humans in control – validate AI actions in real-time

Even the most capable employee gets close oversight when they are new to a role. Supervision isn’t micromanagement; it’s the safety net that ensures quality, builds trust, and catches errors before they compound.

An agent mesh with human-in-the-loop architecture would support routing of specific agent actions or decisions to human reviewers for approval before execution. This is important in agentic systems because LLMs are non-deterministic, they make mistakes so when the impact of agents being wrong is too risky, humans can validate their actions, responses or conclusions.

5. Teamwork makes the AI agent dream work

The most transformative phase of the employee lifecycle is when individuals become part of high-performing teams, where collective capability exceeds the sum of its parts. For AI agents, this is where things get really interesting.

An agent mesh would support workflows that define repeatable processes, such as the steps involved in approving a loan, providing an insurance quote or opening an account. Dynamic orchestration is where the mesh comes alive: orchestrator agents route work to the right specialists, in sequence or in parallel, adapting in real time when the input is fluid and the path forward must be reasoned through.

The result is a multi–agent mesh topology – hierarchical agent organizations coordinating specialist agents across various functions that mirror the structure of an effective human organization.

6. Continuous AI improvement – catch performance drift early

Effective organizations don’t deploy employees and forget about them. They monitor performance, provide feedback, and create mechanisms for continuous improvement.

An agent mesh’s visualizer would give a real-time graphical interface for tracing agent interactions, tool and LLM calls, and decision pathways, so you can see exactly what your agents are doing and why. Ongoing evaluations detect performance drift or emerging failure modes via online evals that run in the background, monitoring production execution.

Further, Open Telemetry instrumentation can surface performance trends over time, giving you the data to make informed decisions about when and how to tune your agents.

Turning AI agents into mission-critical assets requires a new deployment approach

To turn AI agents from simple experiments into high-performance operational tools requires a whole new mindset for building and developing them. Traditional SDLCs work for predictable code, but AI agents act on their own! As a result, they require clearly defined roles, secure tool access, strict testing, human oversight, and constant monitoring.

To successfully deploy, manage, and scale a workforce of AI agents with confidence, organizations must deploy an ADLC on a real-time and event-driven platform.

We've featured the best AI tool.

This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.

The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

• SAP report says AI spend could rise 45% between 2026 and 2028, ROI up by an estimated 39%
• Half of businesses worry agentic AI is being deployed quicker than governance
• Poor data quality is holding them back, 67% believe in integrated data systems as a fix

The average UK business now spends £20.6 million on AI per year according to new SAP data, marking a substantial 27% year-over-year increase compared with 2025 (£16.2 million).

And this is showing no signs of slowing down - as SAP even anticipates a further 45% increase in the two years between 2026 and 2028, bringing average UK business AI spend to very nearly £30 million annually.

Importantly, SAP also envisions ROI increasing in line with spend, rising from 17% in 2025 to 22% this year, and as much as 39% by 2028.

UK AI spend is rising, but at a cost

Despite anticipated rises in ROI, SAP criticized Britain's businesses for lacking AI strategies. Only 11% are said to have a strategic and holistic approach to AI investments, with around two in five businesses prioritizing AI investments on a department level instead of org-wide.

"Organisations are increasing their investments and expecting more in return, but now the focus needs to be on turning that ambition into scalable business value," SAP UK&I MD Leila Romane explained.

While AI spend is largely on the rise, agentic AI deployments are still in progress, with around 70% of UK businesses piloting agentic AI use cases. Still, there's widespread optimism among users, who believe they could unlock a 4.7x rise in financial returns over the next two years.

Among AI agent users, around half (49%) worry they're being introduced faster than they can be governed, and only a minority (9%) consider their governance skills and expertise to be fully ready. A staggering three in five (61%) are either developing governance processes now, or reported limited or no readiness at all.

But despite a clear lack in governance, most are opting to blame their poor data foundations. Poor data quality (71%), incomplete information (70%), siloed systems (65%) and limited access to real-time data (62%) are all considered to be barriers in this area.

Looking ahead, two-thirds (67%) believe integrated data systems are likely to become the most important enabler for AI readiness.

"Good governance is not a brake on AI adoption – it’s actually the opposite," CBI Technology and Innovation Director Naomi Weird added.

Factory workers in India are being asked to wear GoPros to record the way they do their jobs. The ultimate goal is to train robots to take over. Will it work?

The White House publicly asked for U.S. Ambassador to Greece Kimberly Guilfoyle's resignation on Tuesday as she faced mounting accusations of improper conduct.

(Image credit: Jacquelyn Martin)