News

• Gemini AI gave three hikers bad advice about a mountain ascent
• The trip took much longer than Gemini had said
• US officials and volunteers eventually rescued the stranded group

A cautionary tale from California if you often find yourself using AI bots to plan your outdoor adventures: a gang of climbers had to be rescued from Mount Shasta in Siskiyou County after Google Gemini gave them inaccurate information about the scale of the challenge ahead of them.

As reported by the Siskiyou County Sheriff's Office (via Engadget), Google's AI breezily told the three inexperienced hikers that they should prepare for an eight-hour trip to the peak, and to pack simple carbohydrates instead of fats because fats "take too long to digest".

More than 16 hours and a serious knee injury later, off course on their return trip and in the dark, the group was no doubt regretting the decision to seek advice from a disembodied AI rather than a human being that had actually been up Mount Shasta.

Thankfully, the hikers were recovered the next morning by a team made up of staff from the Siskiyou County Sheriff's Office, rescue volunteers, and the US Forest Service climbing rangers. The unlucky trio did have to spend a night in a makeshift camp in the Mud Creek Canyon part of the mountain, however.

"It is always advisable to call the local USFS Mount Shasta Ranger station ahead of your trip to ensure you have the most accurate information, and to never rely solely on AI for your trip planning," read a statement from the Sheriff's Office, which also emphasized the importance of the 12pm turnaround time for those attempting an ascent.

The hikers had set off at 3am expecting to reach the peak by 11am, but they didn't actually make it all the way up until 7pm. On the descent in the dark, they went off-route into the canyon, with one group member sustaining an injury, and called for help.

While one hiker did have AllTrails on their phone, that device died on the trip. “We relied too much on AI rather than our own critical thinking," the climbers told rangers after the ordeal, as reported by SFGate.

As useful as AI can undoubtedly be when searching the web for information, bots like Gemini have never been up mountains, treated injuries, managed finances, or changed electrical wiring. For any important queries, you should always double-check linked sources and other references to verify the information you're getting back.

Indian tech brand Noise has teamed up with Bose for a brand new pair of open earbuds, the Noise Master Buds Open.

These clip-on buds are a first for Noise in the open ear segment, but join its existing Master series lineup, that includes over-ear headphones and in-ear buds — which were also built alongside Bose.

And lucky me — I’ve already tried the Master Buds Open for myself. I got hands on with the brand new open earbuds at IFA Berlin 2026, and spoke with Noise about its collaboration with Bose and the fancy features crammed into its latest creation.

Allow me to start with something of a disclaimer. The Master Buds Open aren’t a fully fledged Bose product, they simply feature ‘Sound by Bose’ technology. This means that the audio giant helped to tune the sound on the Master Buds Open, and there’s a specific Sound by Bose EQ option for the model.

We’ve seen other brands use Sound by Bose tech in the last year or so, with Baseus making use of a partnership in its Inspire line (including the XH1 headphones), and Skullcandy teaming up with the ANC giant for products including the Method 360 ANC (which we rated very highly).

However, Noise has a closer relationship with Bose than most, as the latter has invested in the Indian company across recent years, and the pair have already developed a number of audio products together, including the Noise Master Buds 2 and Master Buds Max.

Although the new Master Buds Open don’t quite offer the exceptional build and sound quality we’ve come to expect from products like the Bose QuietComfort Ultra Earbuds Gen 2, this collaboration certainly heightens the perceived value of Noise’s affordable audio gear, and helps it to stand out as a relatively recent entrant in competitive markets like the US, UK, and EU.

Anyway, let’s talk about my time getting hands-on with the Noise Master Buds Open.

I have to say, when the buds were brought over to me, they made an excellent first impression. They come in a beautiful case with appealing metallic details, and the buds themselves look incredibly sleek and almost futuristic — especially in the slick Silver colorway.

In use, the Master Buds Open also felt supremely comfortable. I find that some clip-on earbuds have the tendency to either feel a bit tight, or almost too loose, but in my short time with the Master Buds Open they seemed to master the clip-on feel, with a secure yet comfortable fit and almost weightless on-ear sensation.

Sonically, the Master Buds Open performed about as well as I expected, although I should preface my thoughts by noting that the unit I tried was a prototype, and so the Master Buds Open may sound a little different at launch.

Naturally, I made use of the Sound by Bose EQ option, and I’d say the earbuds performed decently overall. A representative played me a bit of I’m Alive by Céline Dion, and I was treated to clear vocals, with fairly prominent bass (as far as open ears go), and well-separated guitars in the mid-range.

I also had the chance to play some songs of my choosing, and switched over to Can’t Resist by Jamback. In this house track, I realized that treble was pretty forward, meaning that drums higher in the frequency range were particularly prominent. Bass was also decently weighted, although you won’t get the pounding low-end that some of the best wireless earbuds can provide.

Was I blown away by the audio on offer from the Master Buds Open? Perhaps not — at least not at this stage. But they did perform pretty well overall, especially when you consider their price. I was told that the Master Buds Open will release for about €119 (around $140 / £100), which is relatively affordable — especially when you consider that Bose’s Ultra Open Earbuds have a list price of $299 / £299 / AU$449.95.

While the Master Buds Open will undoubtedly face stiff competition from similarly priced rivals like the Anker Soundcore AeroClip, they did impress me with their comfort and looks.

But one thing struck me more than any other: their surprisingly feature-rich companion app. Not only did I find Noise’s app to be incredibly easy to use and navigate, but I also couldn’t believe how much functionality it had to offer.

There’s a spatial audio setting, designed to offer “dynamic and realistic” audio, as well as a Gaming Mode if you need low-latency audio — a nice touch for keen mobile gamers.

On top of this, the Master Buds Open have customizable touch controls, multi-point, Find My Device capabilities, and a Focus Mode, which locks the buds’ controls when you just want to zone in on your tunes and avoid accidental touch commands. Of course, there are also EQ settings, including a fully customizable equalizer, and mics built in for taking calls on the go.

Noise was also keen to show me that the Master Buds have a built-in AI assistant. This enables you to ask questions with your voice, read chat history, and transcribe meetings or discussions. That’s quite a lot of functionality for buds in this price-range.

While there’s no specific launch date for the Master Buds Open just yet, Noise has said that they’re set to launch on their website and with some major online retailers from late September 2026.

• Nvidia's NVHBM moves the memory controller off the accelerator die and into the HBM stack's base die, allowing it to claim a potential 30% more bandwidth, 15% lower HBM power, and 25% more compute area versus standard HBM4E
• Nvidia's NVHBM performance comparisons are linked to HBM4E, which is still in the sampling stage, with a launch expected sometime in 2027
• Access to the custom solution is gated behind NVLink Fusion; Amazon's Annapurna Labs is the only named partner so far and has not said which GPU uses that memory

On August 26, Nvidia extended its NVLink Fusion program with NVHBM, a custom high-bandwidth memory architecture that the company says delivers up to 30% more memory bandwidth per stack, 15% lower HBM power consumption, and up to 25% more usable area on the accelerator die.

The performance was measured against standard HBM4E samples even as the memory is expected to ship in volume some time in 2027.

Amazon's Annapurna Labs is the first named partner for Nvidia's memory advance which aims to address growing performance limitations for frontier-level AI models centering around limited bandwidth.

Nvidia's approach doesn't reinvent how memory is handled; it reimagines where it is managed. Conventional HBM approaches split this between two companies: the memory vendor supplies the DRAM stack and its base die, and the accelerator designer puts the memory controller and physical interface on its own compute die.

JEDEC standardizes the connection between the two at the cost of a very wide, comparatively slow parallel bus. NVHBM dissolves the seam altogether by moving Nvidia's memory controller into the base die of the 3D stack, replacing the standard bus with a narrower, serialized die-to-die link that Nvidia designs and memory makers can build.

Nvidia's technical blog puts the resulting reduction in interface and support area at up to 67% against the JEDEC HBM4E standard, even as it improves key power consumption, memory bandwidth, and usable area.

The underlying tech is not completely new, however: Marvell announced the same basic idea in December 2024, naming Micron, Samsung, and SK hynix as collaborators while claiming up to 25% more compute area, 33% greater memory capacity, and a 70% reduction in memory interface power, figures that come close to what Nvidia is currently projecting. Counterpoint Research analyst Neil Shah put it plainly: "The technology is not new. The distribution is."

Even as HBM4E remains elusive, as it is not in mass production just yet, with Samsung shipping the first HBM4E samples in late May 2026 and SK hynix pulling its own sampling forward to around June, NVHBM is even more elusive. It is a building block gated behind NVLink Fusion, Nvidia's program for connecting third-party accelerators to its rack-scale platform, and access runs through that program's partner list.

Amazon's Annapurna Labs is the first named participant, with Nvidia's blog saying Annapurna will support NVLink Fusion with Trainium4 without explicitly mentioning NVHBM.

For now, NVHBM might be the future, but it may arrive well after HBM4E is already being deployed en masse. Nvidia's own technical blog describes NVHBM as built on the same technology the company will use for future GPUs, without specifying the first generation that would support it natively.

We've tested plenty in the Anker Solix line of portable power stations, and they've continually impressed us. And as part of the Labor Day sales, the Anker Solix S2000 has dropped to $650 (was $1199) at Amazon.

The 2,010Wh power station is smaller and lighter than other units in its capacity class, aimed squarely at everyday home backup. It's capable of keeping a fridge, router, or medical device like a CPAP machine running through an outage rather than powering a whole house.

Solix S2000: was $1199 now $649.99
2010Wh | 1500W

This 2,010Wh portable power station features 1,500W continuous output with 3,000W peak. We especially like the compact 35.7lb design that's smaller and lighter than typical 2kWh power stations. Rated for 10,000 charge cycles and built for quiet, UPS-style home backup.View Deal

Most 2kWh power stations in this category are large enough that finding a permanent spot for one becomes part of the decision. Anker built the S2000 to be around 30% smaller and 25% lighter than the typical unit at this capacity, at 35.7lbs, which makes it easier to tuck into a closet, under a desk, or beside a fridge without it dominating the room.

The battery itself is the bigger story. Anker rates the S2000's LiFePO4 cells for 10,000 charge cycles, roughly two to three times the 3,000 to 4,000 cycle rating common on rival units in this class, with a stated 15-year lifespan. For a device meant to sit ready for years between outages, that longevity matters more than most spec sheets suggest.

Idle power draw is another area where the S2000 stands out. Anker's OptiSave feature keeps standby consumption down to around 6W with outputs enabled, which means less of the stored capacity is wasted just keeping the unit ready. Combined with a 10-millisecond UPS switchover, it's built to keep a fridge or router running through a power cut without a noticeable interruption.

At 1,500W continuous output across five AC outlets, it covers the large majority of everyday household devices, and Anker's SurgePad feature can support compatible appliances needing up to 2,600W in short bursts.

For more top-performers we've tested, see our guide to the best portable power stations.

Anker Solix C1000 Gen 2: was $599 now $499.98
1024Wh | 2000W

A complete portable power setup in one purchase. The C1000X Gen 2 packs 1024Wh capacity with a 2000W continuous output (3000W peak with SurgePad). And it charges from flat to full in 49 minutes via wall outlet — the fastest recharge in its class, certified by Guinness World Records. View Deal

Explorer 1000 V2: was $559 now $499
1070Wh | 1500W

We recommend this one for camping this summer - it's practically built for the job. In our review, it also earned top marks, making it one of the best of its class. View Deal

A $550 drop from the $1,199.99 MSRP is a steep discount on paper, around 46 percent off, though it's worth knowing the S2000 has rarely sold at full price since launch. It's generally traded between $600 and $700 over the summer, with a low of around $600 during a Prime Day promotion. At $649.99, this is a solid, near-best price rather than a one-off record low, so it's a good time to buy without being a once-ever deal.

✅ Buy the Anker Solix S2000 if... you want dependable home backup power in a compact, lightweight 2kWh unit, you need UPS-style protection for a fridge or a medical device like a CPAP machine, or you value a long-lasting LiFePO4 battery over raw output.

❌ Skip the Anker Solix S2000 if... you need to run high-draw appliances like a microwave or window AC unit, you want a system that's expandable with add-on battery packs, or you would rather wait for more independent hands-on testing before buying.

The 1,500W continuous output also trails some higher-wattage rivals in this capacity class, and the S2000 isn't expandable with additional battery packs the way some competing systems are, so it's worth checking your intended appliance loads before buying if you're planning on anything beyond basic home backup.

• Shop the full Labor Day sale at Amazon

• The Lissome R1 is a compact countertop dishwasher for small apartments
• Its water tank is filled manually, so it doesn't need to be plumbed in
• It has a surprisingly large capacity — designed for up to three meal sets

For the past decade I have been getting by without a dishwasher. My small apartment just doesn't have space for one — at least, not if I want to have, say, a sink. So when I saw Lissome's countertop dishwasher at IFA 2026, it was like my prayers had been answered at last.

The Lissome R1 is a compact dishwasher designed exactly for situations like mine: small households or apartments that don't have space for a full machine, or rented homes that you can't plumb something new into.

"Dishwasher ownership often assumes a cabinet opening, fixed plumbing and unused worktop space," says the brand. "R1 is designed around the homes where those assumptions do not hold, giving users a compact alternative without asking them to redesign the kitchen first."

It's a compact 280mm wide, 507mm deep, and stands 502mm tall. That means it should be able to squeeze into your kitchen without eating up all your counter space. More useful — for renters especially — is that it doesn't need to be plumbed in. Instead, the water tank is filled manually, giving you plenty of flexibility about where you put it . Plus, you can easily take it with you when you move homes.

Lissome is keen to emphasize that the R1 is designed to be a main dishwasher rather than an overflow machine. In fact, it's rather TARDIS-like, with the ability to hold up to 34 items — it's meant for up to three people's place settings.

It's not a plain white box, but instead has glass windows on three sides. Not only does it look good, but it also allows me to finally satisfy my long-held curiosity as to what actually happens inside a dishwasher. (How well the glass-forward design holds up after being exposed to many weeks' worth of cereal, sauce and coffee dribbles remains to be seen.)

Lissome says the inner workings are built around "the World's First Sweeping Jet Technology". Essentially, there are two rotating arms in the base, shooting out artfully coordinated jets of water that arc gracefully — picture the Bellagio Fountains — across the whole interior chamber.

There are six preset programs. These include Quick Wash, which combines 13 minutes of speed-washing with two minutes of steam-release and drying, and Smart Wash, which uses an AI-powered sensor that "detects water turbidity and grease levels, then adjusts the cleaning cycle and detergent dosage" to get your dishes sparkling without wasting water or cleaning fluid.

The X6 is currently the best 360 camera you can buy, offering excellent image quality day and night.

A great 360 all-rounder that fixes many of its predecessor's flaws, and gives GoPro and Insta360 a major headache.

Two of the biggest names in 360 cameras have just launched new flagships within weeks of each other, and it's clear that DJI's Osmo 360 II and Insta360's X6 are both aiming squarely at the same buyer.

I've reviewed both cameras individually, and on paper it looks like a tight contest. The Osmo 360 II arrives just a year after the original Osmo 360, fixing its biggest weaknesses while pushing image quality forward with the first native 8K/60fps 360 video on the market. Insta360 is no newcomer to this though: the X6 lands ten years after Insta360's first 360 model, and my review of a few weeks ago called it the best 360 camera on the market, with excellent image quality day and night.

If anything can knock the X6 off its perch, it's the Osmo 360 II. I've put both cameras through matched side-by-side testing — with the same settings and conditions — to find out which 5-star model is best in which area. Here's what I found.

• Osmo 360 II: Standard Combo £509 / AU$921; Adventure Combo £639 / AU$1,096 (not launching in the US)
• Insta360 X6: Standard bundle $699.99 / £619.99 / AU$1,049.99; Essentials Bundle $799.99 / £709 / AU$1,159.99
• The Osmo 360 II undercuts the X6's entry price by over £100

The Osmo 360 II's Standard Combo costs £509 / AU$921, and includes the camera, one battery, a silicone lens cover, a neoprene case and a USB-C cable. The pricier Adventure Combo adds two extra batteries (three in total), a triple-battery charging case and a 1.2m selfie stick for £639 / AU$1,096. It isn't launching in the US for now, with DJI's regulatory authorization still pending.

The Insta360 X6 starts at $699.99 / £619.99 / AU$1,049.99 for its Standard bundle, which gets you the camera, battery, USB-C cable, protective pouch and lens cloth. Step up to the $799.99 / £709 / AU$1,159.99 Essentials Bundle and you add a second battery, a 114cm invisible selfie stick, a lens cap, a fast-charge case and an extended warranty.

Compare the two Adventure/Essentials bundles directly and the value gap narrows a little, since both throw in similar extras. But even then, the Osmo 360 II remains cheaper at every tier – if, of course, you live in a country where you can actually buy it.

• Price and value winner: DJI Osmo 360 II (unless you're a US resident)

• Both cameras claim 14.5 stops of dynamic range
• In daylight the Osmo 360 II shoots 8K at up to 60fps versus the X6's 50fps
• X6 comes with PureVideo for 8K/30fps low light shooting, while the Osmo 360 II offers SuperNight 2.0 with 8K/60fps shooting

I shot matched daytime and nighttime footage on both cameras using a dual-camera mount rigged up to my car, exporting everything at matched settings (10-bit, 100Mbps for daytime and 160Mbps for the nighttime comparison) to keep things as fair as possible.

In daylight, using each camera's default Standard color profile and preset Dewarp reframing, the X6's footage is the cleaner, sharper and less distorted of the two. But the Osmo 360 II holds onto noticeably more detail in bright skies, where the X6's footage blows out around the sun.

Applying DJI Studio's noise reduction on export closes some of the gap for the Osmo 360 II, even in daylight, though pushing the bitrate up to 160Mbps made no further difference. Even with noise reduction applied, I'd still call the X6's daytime footage the more visually appealing of the two overall.

After dark, it's a closer contest. The X6 still has a bit more punch to its low-light footage, though it casts a slightly greener, yellower tint by default that's easy enough to correct in post. Insta360's PureVideo mode caps out at 30fps, while the Osmo 360 II's SuperNight 2.0 goes up to 60fps — worth factoring in if you're shooting motion after dark, sports especially. Both cameras are reliable low-light performers, but I'd give the X6 the edge here too, just.

• Video quality winner: Insta360 X6

• X6 waterproof to 20m case-free, versus the Osmo 360 II's 10m
• Both cameras are IP68 rated against water and dust ingress
• Both now offer user-replaceable lenses

The X6 wins on raw waterproofing, being rated to a depth of 20m case-free, versus the Osmo 360 II's 10m. DJI has a dive case of its own on the way, extending the Osmo 360 II to 50m once it launches alongside the camera in a couple of weeks, while Insta360's own case takes the X6 to 60m. That at least means neither camera is left behind for serious underwater use, if you're willing to buy the accessory.

Durability tells a similar story of both manufacturers addressing old weaknesses. The Osmo 360 II's lenses now use a tougher optical coating and are user-replaceable via a cheap kit — a direct fix for one of the original Osmo 360's main flaws. The X6's lenses are also user-replaceable, continuing the system Insta360 introduced with the X5.

• Waterproofing and durability winner: Insta360 X6

• X6 has a bigger battery (2,600mAh vs 2,150mAh)
• Both cameras can overheat when recording at their most demanding settings
• The Osmo 360 II has more than double the built-in storage of the X6

The X6 has the bigger battery, and it makes a difference: DJI rates the Osmo 360 II for up to 100 minutes of 8K/30fps recording, while Insta360 claims 140 minutes for the X6 at the same settings, and having tested both I can say those figures are largely accurate.

I ran both cameras through an overheating test side by side in the same room, at a mild 22°C, each set to its own maximum 8K frame rate (60fps for the Osmo 360 II, 50fps for the X6), 10-bit Standard color with Adaptive Tone active. The Osmo 360 II shut itself down to cool off after around 34 minutes; the X6 lasted about 40. Not a dramatic gap, but a gap nonetheless.

Storage flips the script, though. Both cameras have a microSD slot, but the Osmo 360 II offers 105GB of additional built-in space while the X6 has 47GB — which is only sufficient for 30 minutes of 8K/50fps footage.

• Battery, storage and overheating winner: Insta360 X6 (but the Osmo 360 II's storage win is a consolation prize)

• Both cameras have mobile and desktop companion apps
• Insta360 Studio supports ProRes export; DJI Studio doesn't, yet
• Both ecosystems now support 24-bit audio with their respective mic systems

Neither camera has a clear edge here. The DJI Studio desktop offers excellent subject tracking and a really impressive 64x slow motion mode, and I'd rate it the better desktop experience of the two, but the DJI Mimo mobile app's AI-assisted auto-editing tools left me unimpressed. On mobile, Insta360's app has historically had the edge, though I've found it to be unstable and crash-prone on my ageing iPhone, so you'll need a newer model to get the best out of it.

There is one interesting software difference on the desktop side: Insta360 Studio supports ProRes export, which DJI's desktop app doesn't yet offer — a real consideration if you're using a 360 camera for professional work, or want to in the future.

Both companies also now offer wireless microphones (DJI's Mic series and Insta360's own equivalents) that connect directly for 24-bit audio, so neither has a key advantage there either, although DJI's Mic range is significantly larger.

• Design and display winner: Tie

Four rounds go to the X6, one goes to the Osmo 360 II, and one is a tie — but I think the margins tell a more complex tale than the overall score here. The X6 wins on image quality, waterproofing and battery life, but never by a giant margin, and the Osmo 360 II wins back some ground on price, storage and dynamic range in bright conditions.

Whether that adds up to the Osmo 360 II being the better buy overall depends heavily on what you value most. If outright image quality matters above all else, the X6 remains the camera to beat. If you want DJI's frame-rate advantage, extra built-in storage, and to spend significantly less getting them, the Osmo 360 II makes a serious case for itself.

I want to spend more time with both cameras across a wider range of real-world conditions before calling an outright winner — so consider this early verdict just that: early. But, for now, I'd say the Insta360 X6 is still the king of 360 cameras.