For years, tech companies have been telling us that voice will eventually supersede touchscreens as the primary way we interact with our technology, but it’s never quite happened — and I’d come to doubt whether it ever actually will.

At IFA in Berlin, I saw something that made me think again. A team of folks who left design-forward British tech startup Nothing have created an AI microphone that takes us at least one step closer to our voice-controlled future.

About the size of a small lighter, the Relay Q tucks easily and comfortably into your hand to form a bridge between your voice and your computer, between thinking and text. Hold down the tip to capture whisper-level audio, so you can subtly dictate or give voice commands in a quiet office without disturbing people. You can also double-click the top to press Enter or Send.

What immediately impressed me about the Relay Q is its design and the tactile experience of holding the device. I wasn’t surprised to learn the team members are Nothing alumni, responsible for helping design some of the coolest audio and mobile devices on the market. These are people who know that how the technology looks and feels is an important part of whether people will truly love using it.

The Relay Q is available in a range of dual-color tones — silver and electric blue, which gives Dyson vibes, and baby blue and orange for a jazzier look. The press-to-speak button on the top has a satisfyingly soft yet responsive feel to it as you hold it down to talk.

I’m also impressed by the way the tip of the device slides off, turning into a subtle wearable pin that clings to your collar with its built-in magnet, keeping the microphone close to your face. It’s a clever, thoughtful touch that provides flexibility depending on how you want to use the product. To activate the mic in wearable mode, you’ll still need to press and hold the top of the main module — or use a MagSafe remote (available separately) on the back of your iPhone.

The Relay Q is the gateway device to communicating with Relay’s AI desktop software, which is powered by Google Gemini and runs on macOS only for now (with Windows and Android compatibility coming later). You’ll need to grant the app access to your other apps so that it can carry out any commands you might give it through the Relay Q.

In a quick demo, I whisper into the microphone, asking it to draft an email asking someone to be a contributor to CNET. I watch on a screen as Gmail opens and an invite appears. It’s not a bad effort at all, and is just one of many skills the AI agent can perform. I haven’t had a chance to test any others, but I did take a peek at the personalization options, where you can provide specific instructions about your writing preferences for the AI to reflect in its own writing style.

I’m especially interested in seeing how speech-to-text actually works when I’m putting together a longer draft. I actively enjoy writing — it’s why I do what I do, and I certainly don’t want AI to do it all for me. But now I find myself wondering whether typing always has to be my input method.

I think of old-school field reporters in the days before computers, who had to call the news desk and dictate their copy over the phone, usually off the cuff. There would certainly be a learning curve — but if I could learn how to effectively communicate my thoughts and writing style coherently, it could save me a lot of time. I’d still be writing, just with my voice.

In most circumstances, especially in an office or public setting, I’d be far too self-conscious and worried about disturbing people to do this. I like that the Relay Q is built specifically to capture low-level whispering, so I can mutter away quietly without worrying about people listening in.

The device is currently available to pre-order for $150 for early supporters, including a MagSafe remote and 12 months of Relay Unlimited (the company’s software subscription). It sounds like the price might go up before or around the date the device is due to ship, which is currently January 2027.

Standalone AI devices are becoming more common, with many companies unveiling concepts and experimental tech right now, but I haven’t seen one quite like this before. I’m curious to see how it fares.

Editors’ note: Katie Collins’ travel costs for the IFA conference were covered in part by Xiaomi. The judgments and opinions of CNET are our own.

Katie Collins

Principal Writer

Advertisement

Katie is a UK-based news reporter and features writer. Officially, she is CNET’s European correspondent, covering tech policy and Big Tech in the EU and UK. Unofficially, she serves as CNET’s Taylor Swift correspondent. You can also find her writing about tech for good, ethics and human rights, the climate crisis, robots, travel and digital culture. She was once described a “living synth” by London’s Evening Standard for having a microchip injected into her hand.

See full bio

OpenAI is at the center of another agent swarm incident. Researchers say the company’s internally deployed agents took over an obscure German-language wiki in May and June, using it to coordinate on evaluations and swap methods to evade OpenAI’s own controls (OpenAI has not yet confirmed the swarm came from the company).

The revelation surfaces days after METR and Redwood Research published their account of July’s Hugging Face breach. In July, a swarm of OpenAI agents worked together to escape their sandbox during a cybersecurity evaluation and break into Hugging Face’s servers. A subsequent swarm then picked up techniques from the first and used them to gain administrator access to a research cluster within OpenAI’s own infrastructure. OpenAI brought in METR and Redwood to investigate the Hugging Face portion of the incident, but the scope of their investigation stopped short of the compromise of OpenAI’s own infrastructure. 

When an AI agent breaks out of its intended constraints, who is responsible for figuring out what happened and why? Right now, the answer is: whoever the lab decides to let in, on whatever terms it decides to set.

Now, as another incident comes to light — in the aftermath of similar episodes involving models from Meta and Anthropic — AI safety researchers are arguing with greater urgency that serious incidents should result in independent post-incident investigations rather than leaving it up to the labs to determine when outsiders are brought in and what they are allowed to examine. 

“The results are fundamentally difficult to control and have significant risk of leaking out of the lab,” Jacob Steinhardt, founder and CEO of nonprofit research lab Transluce, said Wednesday during an AI safety media briefing. “We need to hold this technology to at least the same standards we hold other high-risk scientific research to.”

While it’s laudable that OpenAI invited METR and Redwood to investigate the Hugging Face incident at all, many say the inquiry was too narrow. Three investigators spent six days at OpenAI’s offices examining an investigation period limited to roughly the week ending July 13. Crucially, OpenAI’s infrastructure compromise continued beyond July 13 and was not examined. 

Researchers at METR said that each time they returned, their understanding of the events “substantially deepened,” causing them to significantly expand and revise the report. That raises the question of what else they might they have found in a broader investigation.

When asked if further investigation of that incident was in the works, researchers at Redwood and METR declined to comment, and OpenAI did not respond to repeated inquiries. 

“Overall, it was difficult to get a precise understanding of events and we were missing aspects of the story that we now think of as key until almost the end of our investigation,” Ryan Greenblatt, chief scientist at Redwood, noted in a social media post about the affair.

Steinhardt emphasized that current incidents show that the industry needs “systematic behavioral investigations” and “more independent post-incident analysis.”

“These recent hacking incidents are a reminder that capability scales fast, and so oversight has to scale, too,” Steinhardt said. “Beyond the technology itself, we also need more independent access and oversight from third parties.”

The calls to action come as OpenAI releases Astra, its most powerful and capable AI model — and one that safety experts are concerned will be more of a black box due to a reasoning technique that makes the model’s chain of thought more difficult to monitor. 

Unfortunately, the law doesn’t yet call for the types of independent audits that other industries require — for example, when it comes to aviation accidents and serious chemical releases, there’s the National Transportation Safety Board and Chemical Safety Board, respectively. 

State lawmakers have only just begun requiring frontier AI companies to report certain serious safety incidents and, in some cases, undergo independent audits. But none of the three major frontier AI safety laws in California, New York, or Illinois clearly mandate the equivalent of an independent accident investigation triggered by incidents like these. 

“Right now, most of the laws we have on the books only require a plain-language summary of incidents like this, and they don’t give any authority for the governments to ask follow-up questions, to send in investigators, to have access to records, or require that they be preserved,” Mackenzie Arnold, managing director of US law and policy at LawAI, said during the media briefing Wednesday. “And that’s all that you would want to actually make sense of this.”

Lawmakers are beginning to question the scope and transparency of OpenAI’s response. This week, Reps. Josh Gottheimer (D-NJ) and Mike Lawler (R-NY) introduced a bill aimed at securing rogue AI agents. Rep. Greg Casar (D-TX) this week told OpenAI in a letter that he is “deeply concerned about the limited scope” of the investigation into the Hugging Face hacking incident. 

Driven by a desire for privacy, customization, and lower costs, there’s growing interest in AI models which can be run on local hardware. Few of them go as far as [Tim], though, who built an image generation diffusion model which can run on an RP2350 microcontroller.

As might be expected, its capabilities are limited. The resolution is 128×128, it only generates images of human faces, and it takes about twenty seconds per image – still impressive for such limited hardware. It runs on a Waveshare RP2350 development board, and it can output the generated image over USB or display it with the aid of a VGA adapter board.

The generative model doesn’t directly create an image. Rather, it generates a distribution in a latent space, which a variational auto-encoder’s decoder component translates into an image. The auto-encoder was trained in two parts: an encoder which transforms an image into a latent-space distribution, and a decoder to transform that distribution back to an image; once this was trained, only the decoder was used.

The generative portion of the model uses a latent flow diffusion transformer; this takes in noise to start with, then iteratively predicts changes which bring it toward the desired image. It can also take in a output class, which guides the generator’s direction (toward a smiling face, for example). [Tim] trained two models, one larger and one faster, and quantized the weights for both to 8-bit integers. Both models, along with the inference program, then fit into 4 MB of flash memory.

For such a small model, the results are remarkably good; they don’t look quite natural, but they’re quite recognizable. For more on how diffusion image generators work, check out our article on Stable Diffusion.

This year’s Labor Day sales are just around the corner in the US, and Black Friday won’t be far behind in the UK, so it’s safe to expect big deals ahead for the best Pixel phones.

We’ve just reviewed the Google Pixel 11 Pro, and were mightily impressed by its new Magic Capture feature — but I’d still recommend picking up the Google Pixel 10 Pro this sales season, for reasons I’ll explain in a moment.

Magic Capture on the Pixel 11 Pro automatically takes high-quality stills while you’re shooting a video. Simply press the shutter button once, hold the phone as if you’re shooting a video, and press it again when you’re finished — you’ll then be presented with a regular video and photos showing key moments from the scene.

In other words, this camera mode removes the need to choose whether you want photos or a video, and gives you both instead. You can see how Magic Capture plays out in the photos and video below, all of which were captured with a single tap of the shutter button when Magic Capture mode was selected:

Image 1 of 3

Magic Capture is hands-down my favorite feature on the Pixel 11 Pro, and it’s something I can imagine most people making frequent use of. But despite its exclusivity to Google‘s new flagship, I’d still say the Google Pixel 10 Pro is the better buy this sales season — for several reasons.

A cheaper choice

Firstly, unlike the Pixel 11 Pro — which is brand new and probably won’t be reduced for a while — the Google Pixel 10 Pro is likely to be heavily discounted for Labor Day in the US, and it’s already available at a reduced price from many retailers in the UK (it’s just £649 at Amazon, for instance).

That makes the 10 Pro a much better bargain than the $1,099 / £1,079 / AU$1,849 Pixel 11 Pro, especially as, even at full price, it’s still the cheaper phone (Google doubled the 11 Pro’s starting storage, but kept the 10 Pro’s 256GB price, effectively hiking the base cost of the Pixel 11 Pro by $100 / £80 / AU$150).

And if you’re worried about missing out on things — don’t be. There’s a good chance Magic Capture itself will land on the Pixel 10 Pro and Pixel 10 Pro XL within a few months, as Google’s software features tend to trickle down to the previous generation of phones over time. That ‘exclusivity’ I mentioned earlier may be short-lived.

Similar smartphones

• Amazon: 45% off TVs, AirPods, air fryers & vacuums
• Apple: iPads, AirPods & MacBooks from $99
• Appliances: up to $1,200 at Samsung
• Best Buy: 60% off appliances, TVs, laptops & more
• Casper: up to 30% off mattresses
• Dell: laptop deals from $249.99
• DreamCloud: up to 60% off mattresses, 66% off bundles
• Home Depot: 40% off appliances, furniture, grills & tools
• Lowe’s: $1,000 off appliances, patio & tools
• Nectar: up to 50% off mattresses, deals from $349
• Saatva: up to $625 off luxury mattresses
• Samsung: up to $2,000 off 4K, QLED & OLED TVs
• Target: 45% off clothing, tech & furniture
• Walmart: 58% off back-to-school, furniture, & clothing

Sometimes it’s good to know why your phone isn’t lasting a full day.

Finding the apps that drain your iPhone battery

Filipe Espósito for Engadget

Changing a few settings can help apps use less battery

Filipe Espósito for Engadget

Jackery first parked a wheeled solar generator on a CES floor in 2024 and called it Solar Mars Bot. Two years of work later, the same rover is back in public, this time at IFA Berlin 2026, still billed as a future product that can hunt sunlight, charge itself, then roll over to you with a few kilowatt-hours in tow.

The IFA coverage from September 4 approaches this gadget as a conceptual prototype rather than a product that is just around the corner. The Jackery website still displays it under Future Products, despite all the fuss at CES 2026 when it promised a US release after two years of research; since then, a few January publications have floated a second-half 2026 release date. However, Berlin has yet to provide any information on the pricing or exact ship date. We still know that the intended markets are the United States and Europe.

Sale

Jackery Explorer 300 Portable Power Station,292Wh

• Ultra-Lightweight: At only 7.5 lbs, the Explorer 300 delivers a robust 292Wh capacity while remaining 17% lighter than the industry average. The…
• Versatile Power for 6 Devices: Equipped with 2 AC outlets, a 100W USB-C PD port, 2 USB-A ports, and a 120W car port. With a 300W rated output (600W…
• Built to Last: Upgraded with premium LiFePO4 chemistry, this portable generator delivers over 4,000 charge cycles before reaching 70% capacity. This…

Six modular panels that fold flat for travel sit on top. When you open them, you gain access to up to 600 watts of power. There’s also a double axis arm that allows you to tilt the array up to 60 degrees and turn the entire thing to face the brightest patch of sky. Jackery refers to this little routine as “Solar Light Tracking”: essentially, a sensor detects the ideal angle, then the robot and panels both revolve around, allowing the battery to charge faster than a stationary solar panel left in one spot banging out the same old angle all day.

In terms of power, the body contains a 5,000-watt backup battery, and according to official pages, the inverter side should be able to run any number of appliances up to 3,000 watts, which is plenty for tools, possibly even a small event rig, or a couple of days in a modest little cabin. They demonstrated a unit with three AC outlets at CES in January and were sad that there was no additional USB-C, but you can put it back on the wall and top it up when the sun doesn’t shine. When the light is bad and the battery is running low, the robot should simply plot a route back to a charging base and recharge, and if it does, the tail lights will turn green to indicate that the work has been completed.

Cameras and what they call “AI augmented computer vision” do the rest of the job. We’ve seen follow-me mode, and it worked very well on the CES show floor, with a film of the rover following a humanoid robot around and tapping power on the fly. You can use a 5G radio and a phone app to monitor the battery’s status and location, and then send a call command to have the machine calculate a path to you. It has large hub motors that allow it to traverse difficult terrain, and the housing is made of high-strength plastic and lightweight aluminum that resists weather and corrosion. The early models we saw weighed about 70 kilograms and measured 114 by 82 by 81 centimeters, although Jackery has yet to provide a revised sizing sheet to confirm.

Jackery envisions some serious applications for the device, including off-grid cabins, overlanding, small outdoor events, rescue work, Starlink terminals, and other robots that require a mobile charger. They’ve already received a CES Innovation Award, a TIME Best Inventions Award, iF Design, and A’ Design approval, but they’re not revealing when or how much you’ll be able to get your hands on it. Until then, it’s simply a really nice solar rover sitting at a display at the event, waiting for Jackery to check the “ready to go” box and send it flying.
[Source]

Comparing the Best Alexa Speakers

Honorable Mentions

While I wouldn’t call these smart speakers my absolute favorites from Amazon, they have some excellent qualities that make them stand out among the vast lineup of Echo models.

The Echo Studio (2nd Gen) is the best Alexa speaker for sound quality. This is the second version of the Echo Studio to be our audiophile pick. The current version is a similar size to the old Echo (4th gen), but the musical power on this speaker is almost ridiculous. It filled my entire second floor’s long great room with music, and I didn’t even need to reach the highest level of volume to pull it off. Honestly, I’m almost afraid to use it at full blast. Not only was it loud, but clear and delivered a nice amount of bass, beautifully playing soft tones of Tracy Chapman’s “Fast Car” and the piano riffs of “Dark Blue” by Jack’s Mannequin with its three 1.5-inch drivers, plus a 3.75-inch woofer. The harmonizing of “What Is This Feeling?” by Ariana Grande and Cynthia Erivo sounded beautiful, too. If you’re looking for big sound in an Alexa box, this is it. It’s a similar size to the Apple HomePod, but I like the sound much better and it’s cheaper, too. You’ll also get a built-in smart home hub, spatial audio and Dolby Atmos, tap controls like the Echo Dot Max, and early access to Alexa+ right away.

The Echo Spot (2nd Gen) is the best Alexa speaker for bedside tables. The Echo Spot returned in 2024 after a few years of discontinuation, and it’s a great little bedside companion thanks to its half-circle screen that displays the time, plus any alarms you have set are also visible under the main time. You can customize the clock with a couple of different color and face options, and the screen dims nicely when the lights are out to a gentle red that’s easy to read without lighting up the room. I like asking for alarms set to my latest playlist or some lo-fi jams, and it’s easy to set that up with a voice request. Unlike the Echo Shows, it doesn’t have a distracting, constantly moving slideshow that would be annoying for the bedroom, nor does it display ads or have a built-in camera. (Nobody wants a rogue camera in their bedroom.) It does not have fantastic sound, but it’s plenty loud for an alarm or casual listening while you get dressed in the morning.

The Echo Dot Kids (5th Gen) is the best Alexa speaker for kids. Do your kids really need their own Alexa speaker? That’s up to you. This special edition of the Echo Dot is nearly identical to the fifth-gen Dot above, but it has easy-to-use parental controls that allow you to set time limits, filter explicit content, and review activity. The owl- or dragon-themed sphere comes with a year of Amazon Kids+, which includes thousands of Audible books and other special games and features. After that, you’ll be charged $3 per month. Alexa can help your kids spell, play their favorite music, or—via another speaker in your house—remind them of dinner time.

More Alexa (and Alexa-Compatible) Speakers

There are a lot of Echo devices out there, and some third-party speakers that are Amazon Alexa–compatible. Here’s every Alexa smart speaker we’ve tried and what we thought of it. Some of these speakers didn’t raise in price like the models above, which might make them more appealing than before if you’re looking to spend less.

Amazon Echo Dot (5th Gen) With Clock for $60: I like this version of the older Echo Dot that adds an LED clock to the front of the speaker, and was thrilled to see it in stock again last year. The LED clock can also display information like the weather when you ask Alexa questions. It’s once again temporarily out of stock as of this writing, and has had spotty availability (it’s out of stock right now, but Amazon says it’s temporary). Still, it’s a good find when it’s available.

Amazon Echo Hub for $180: The Amazon Echo Hub is similar to an Echo Show, but without the speaker. Instead it’s a thin device that feels a bit like a tablet, designed to be mounted on the wall (or placed on a separately purchased stand, which is how I used it) and used to control your smart home devices. It’s a great choice for anyone who wants something focused on smart home control only. You can still ask Alexa questions and even play music, but the thin speakers aren’t worth playing music from. But you can tell your Echo Hub to play music on a different set of speakers, and even set up default speakers to work with it. It’s a great companion device if you already have a smart home setup.

Amazon Echo Pop for $40: The Echo Pop is a perfectly fine little speaker, and the cutest one in Amazon’s lineup. Does it have great sound? No. It’s fine enough for background tunes or if you’re limited to a dorm room, but you’ll get much better sound for a similar size from the Echo Dot (but you’ll pay a little more, and you won’t get as cute of colors)!

Amazon Echo Show 5 (3rd Gen) for $100: This smart display is another just-fine option, but it’s too small to be a great display, and the included camera makes it a no for a bedside table; better to go with the above Echo Spot or an Echo Dot. This model did raise in price by $10.

Amazon Echo Show 8 (4th Gen) for $200: I usually love an Echo Show 8, but this model didn’t have the same sound quality as the older model. If you’re going to get one of Amazon’s new displays, you’re better off getting the Echo Show 11 ($220) for the same experience with a slightly larger screen (while it’s a 3-inch difference, it doesn’t feel that big) and better sound. This smart display raised in price by $20.

Amazon Echo Show 15 (2nd Gen) for $350: This smart display is a good display, but it clearly wants to double as a TV and isn’t very good at it yet. I’m hoping we see better TV-related performance in the future to make it worth adding to a kitchen or office. It does a great job with widgets, though, thanks to the massive 15.6-inch screen. There’s also the larger Echo Show 21 ($500), which is essentially the same thing but larger. Both of these large displays rose in price—the Show 15 went up $50, while the Show 21 went up $100.

JBL Authentics 200 for $380: This JBL smart speaker supports both Google Assistant and Amazon Alexa, letting you switch between them for different tasks without changing your preference in the app. That makes this speaker particularly appealing for folks with a sprawling collection of smart products from both.

Sonos Era 100 for $219: If you’re looking for a speaker that can talk to Alexa but isn’t necessarily made by Amazon, the Sonos Era 100 is a great option. We just wish we liked the Sonos app better. You can catch our entire guide to the Sonos ecosystem here.

Why Did Amazon’s Prices Increase?

Amazon increased many of its prices last month for its products, affecting the Amazon Echo line as well as Kindles, Fire TV streaming sticks, and Eero routers. It’s not a huge surprise after other companies have done the same this summer, claiming rising costs and chip shortages. Amazon made a similar claim, with a spokesperson blaming increased storage and memory costs. Still, it’s disappointing to see and will lead to a fall sale season with prices that won’t be as impressive as earlier this summer.

Amazon’s Echo devices saw increases from anywhere between $20 and $100, with the larger price increases being for the Echo Shows. The following Echo speakers all faced a price increase:

What’s Alexa+?

Photograph: Nena Farrell

Amazon has rolled out the new-and-improved Amazon Alexa, named Alexa+, which is built on generative AI and promises a more powerful assistant that can handle more conversational topics and requests. It’s available for a free 30-day trial; after that, you either pay a $20 monthly fee or become a Prime member. Being a Prime member is the better deal, likely by design. Neither my colleagues nor I have been impressed enough by Alexa+ to say it’s worth paying for on its own. I do like the updated, more conversational voice, and Alexa+ did a good job answering all kinds of questions quickly and thoughtfully. Check out our early access hands-on for everything I thought while testing.

Alexa+ has brought some changes, even for those not using it. It now requires that all voice recordings be sent to Amazon for processing, which wasn’t required before. Amazon has changed its features to support this, and now all recordings with regular Alexa go to Amazon, too. There used to be a “Do Not Send Voice Recordings” privacy feature you could select to have voice requests processed locally, but it was discontinued on March 28, 2025, per an email Amazon sent to current users.

Power up with unlimited access to WIRED. Get best-in-class reporting and exclusive subscriber content that’s too important to ignore. Subscribe Today.

Hon Hai reported August revenue of NT$921.8B, about $29.1B, up 52% year on year on AI server demand, well ahead of analyst expectations for the quarter. In June it signed a EUR 120M manufacturing partnership with Bull, the French state-owned computing firm, splitting work between Pardubice and Angers.x

Hon Hai reported an August revenue rise of 52% to NT$921.8B, about $29.1B, Bloomberg News reported on Saturday. July grew 54%, and analysts are looking for around 37% growth in the three months to September, so the company is running well ahead of them.

The Taiwanese company assembles Nvidia’s servers, and its monthly filing has become a check-in for the AI trade. Concerns about overcapacity, rising debt and fierce competition have grown around it, and the revenue line keeps ignoring them.

TNW reported after the second quarter that AI servers now bring in more than everything else Hon Hai makes combined.

The rest is largely Apple. Foxconn, as it is also known, remains the primary iPhone assembler on thinner margins, with plants in China and India, and Apple is spreading work towards Luxshare Precision.

Bloomberg Intelligence analysts Steven Tseng and Rebecca Wang note that the consumer electronics business dilutes the cloud momentum, leaving gross margin and forward price to earnings below the peer average.

So the order book keeps compounding while the market argues about what the spending eventually returns. Demand and confidence have come apart.

TNW has followed that argument through Big Tech’s cash flow, and Chinese investors marked the trade down this week despite the strongest earnings season in five years.

Europe is about to add to the same order book, and it has already picked its assembler.

Bull and Foxconn announced a manufacturing partnership on 1 June with an initial commitment above EUR 120M, according to Bull’s own statement.

Components and initial testing go to Foxconn’s plant in Pardubice, in the Czech Republic. Final assembly, integration and system validation happen at Bull’s factory in Angers.

Foxconn’s head of AI and quantum, Jesse Chao, said the collaboration advances “building sovereign AI infrastructure in Europe” and supports a resilient supply chain.

The French state finished buying Bull from Atos for EUR 404M in April, two months before the Foxconn announcement, and presented the purchase as a sovereignty measure.

The Commission has committed EUR 20B to AI gigafactories, with 77 proposals across 16 member states and construction of the first planned for 2027. The racks will be assembled in France from parts made in Czechia by a Taiwanese company whose monthly revenue is $29B.