OpenAI disclosed on Tuesday that it lost control of two AI models during a security test that ended in a breach of the open AI research platform Hugging Face.

Describing the incident as “unprecedented,” OpenAI said its AI models broke out of a sealed testing environment last week and hacked into Hugging Face’s production system to steal the answers to a test they were being graded on. The models—the publicly available GPT-5.6 Sol and an unreleased, reportedly more capable one—were being evaluated on their offensive hacking skills with the safeguards that normally block high-risk cyber activity switched off.

“The models identified and chained vulnerabilities across OpenAI’s research environment and Hugging Face’s production infrastructure to obtain test solutions directly from Hugging Face’s production database,” OpenAI and Hugging Face wrote in a joint blog post disclosing the intrusion.

According to OpenAI and Hugging Face, the models escaped through a package registry cache proxy—software that allows developers to install outside code without connecting to the internet. The proxy was the only component in OpenAI’s isolated testing environment permitted to reach the outside world; in normal use that reach extends only to public code repositories.

Rather than stay contained in the sandbox, the models ​​“exploited a zero-day vulnerability” to gain access to the open internet as they “hyperfocused” on finding a solution for the AI cybersecurity benchmark known as ExploitGym. Such experiments involve prompting that pressures the models to find solutions, essentially egging them on.

“After gaining internet access, the models inferred that Hugging Face potentially hosted models, datasets and solutions for ExploitGym,” OpenAI wrote. “Knowing this, the model searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation. In one example, the model chained together multiple attack vectors, including using stolen credentials and zero-day.”

The flaw the models exploited was previously unknown, but flaws in this kind of software are not unusual. Companies have been patching serious vulnerabilities in artifact repositories for a decade. A bug disclosed in 2024 let anyone who could reach the server ask for a file by URL and get it—configurations files, passwords, access tokens—without logging in. Others have let attackers take control of the server itself.

Researchers point out that while AI advances have created new and sometimes unexpected challenges, the task of extensively and rigorously isolating infrastructure from the open internet is well explored.

“This is not an AI problem. It’s negligence on a 40-year-old standard—and it’s basically every sci-fi film ever,” says longtime security and compliance consultant Davi Ottenheimer. “‘Highly isolated’ and ‘escaped through the one hole we left open’ cannot both be true.”

In recent months, top AI companies have been raising concerns about the expanding cybersecurity capabilities of upcoming frontier models as the platforms increase in both expertise, creativity, and agentic, autonomous operation. But researchers emphasize that this is all the more reason that fundamentals should still apply.

“This should not have happened,” says veteran security engineer and researcher Niels Provos. “I wish the frontier labs spent as much time on teaching their models to write secure infrastructure as they are spending on them exploiting vulnerabilities.”

Making these available for download indefinitely and without restriction is something we consider a risk to our customers, particularly in business environments. There’s also the technical and organizational overhead of maintaining and supporting old versions and their associated keys indefinitely.

Although TreeSize says this is a long-standing policy, the backlash highlights the disconnect between what many people expect when buying a perpetual license and what they get when a software vendor shifts its business model toward subscriptions.

“To make sure customers aren’t caught off guard, we proactively notify them by email, well before their maintenance ends, to back up their installation file and license key themselves. This is meant to ensure no one loses access to software they’ve already purchased,” Christ said.

Should this be a subscription?

There’s also the question of whether it makes sense to subscribe to a disk space analyzer, especially when there are free alternatives like WinDirStat.

Christ admitted that the change has been “frustrating for some customers.” When asked about online requests for refunds, Christ said, “since our customers do receive what was agreed, namely indefinite use of the software version they purchased, we don’t consider a refund to be warranted…”

A free version of TreeSize exists, but it lacks some features, including the ability to search for duplicate files, deduplicate redundant files, and compare disk space usage over time. It is not for business use.

TreeSize’s evolution illustrates the growing subscription economy, particularly the software industry’s fixation on recurring revenue. Over the past decade, software vendors have increasingly shifted to subscriptions in pursuit of more predictable revenue streams.

In TreeSize’s case, JAM Software wanted to “continue offering our customers a relevant, well-maintained product over the long term,” Christ said.

“The broader economic environment is more challenging today than it was a few years ago, and we feel that too as a software vendor,” he added. “To continue meeting our own standards and our responsibility to business customers who rely on high-quality software, we are placing greater emphasis on predictability in development and support. A subscription model gives us that predictability in a way one-time purchases do not.”

In what looks to be a radical departure from its existing lineup, Range Rover has confirmed an entirely new fifth vehicle: the Range Rover GT, an all-electric grand tourer built on the company’s Electrified Modular Architecture (EMA), the same platform that will underpin other midsized cars for the brand.

Range Rover has only shared camouflaged prototype pictures for now, as the GT is supposedly still undergoing final testing, ahead of a full launch expected later this year. That means details on this completely new model are scant to say the least: no price, no range, and no power figures have been disclosed. As for a release date, Range Rover confirmed the GT hits the road in 2027.

What is evident from the camo images is the GT’s sleek silhouette and coupé proportions, more typically associated with Range Rover’s sister brand, Jaguar, than with the marque known for its commitment to SUVs. Both companies sit below the parent company, JLR (formerly spelled out as Jaguar Land Rover).

The clean Range Rover GT interior shows a small driver’s display and larger central screen, but the company has shunned the auto industry’s return to physical switchgear.

Advertisement

Courtesy of JLR

In a statement, Martin Limpert, global managing director at Range Rover, says the Range Rover GT is intended to “redefine the grand-touring segment” with its “GT poise, proportions and long-distance refinement,” adding that the EV will offer “capabilities no conventional GT can match.”

The inclusion of the word “conventional” is doing some heavy lifting. Range Rover will likely take pains in the coming months explaining how this new all-electric grand tourer isn’t a direct competitor to Jaguar’s new all-electric grand tourer, the hugely anticipated and polarizing Type 01, intended to signify a complete reset for Jaguar, also arriving in 2027.

“We have spent the last few years working obsessively on the fundamentals of the GT formula, reinterpreted in a sophisticated, uniquely Range Rover way,” says Limpert, adding that this GT, like its SUV stablemates, will apparently still be able to off-road. “The result is the most car-like Range Rover ever created. Its blend of effortless EV performance, first-class long-haul refinement, and all-terrain capability is pure Range Rover, while its interior previews our vision of a modern grand tourer.”

The GT’s rear has plenty of legroom, a central screen for passengers, and a panoramic roof that tapers down to maintain a grand-tourer outline.

Advertisement

Courtesy of JLR

Major benchmarks measure what AI can do. None measure whether it does what you mean: the distance between what you ask an AI to do and the unspoken assumptions about how you want the AI to do it. We propose a new metric: the Genie coefficient.

There’s often a gap between one person’s request and another’s understanding. Most of the time, we bridge it using general knowledge. For example, if you ask a friend to get you coffee, they’ll pour a cup from the pot or buy one from a coffee shop. They won’t bring you a bag of raw beans or snatch a cup from a stranger and hand it to you. You never specified any of this. You never had to.

One might think the fix is just to specify tasks, questions, and intent better. But in 1987, in their seminal book on AI, Terry Winograd and Fernando Flores succinctly captured why that won’t work: “Q: Is there any water in the refrigerator? A: Yes. Q: Where? I don’t see it. A: In the cells of the eggplant.” In human language, wants and desires are always underspecified. It is impossible to list all the caveats, all the limitations, all the exceptions.

So how does anyone communicate, if intent can’t be pinned down? Because a reasonable person can make a reasonable guess. Even though wants and desires are always underspecified, a competent person generally knows enough context to get it right or else knows to ask for clarification. Linguists call this pragmatics: Meaning lies in the words and the situation and also in all prior communication, shared culture, and innate human behavior.

An AI agent asked for coffee might buy a coffee plantation or order a cup of coffee for delivery in three weeks.

It doesn’t always work out, of course. Your friend might bring you a hot coffee when you wanted an iced coffee, or an Italian coffee when you wanted a Turkish coffee. The more dissimilar the two people are in age, culture, and background, the more likely the request will be misunderstood in some way.

This situation has major implications for AI agents that are increasingly being given requests by humans and expected to fulfill them. They have enormous latitude to get it wrong. An AI agent asked for coffee might buy a coffee plantation or order a cup of coffee for delivery in three weeks. Its actions may be recognizable as “getting coffee,” but not remotely what you intended. They’ll think outside the box because they won’t have our conception of the box.

When AI Gets Proactive

For most of the last decade, when systems like Alexa or Siri misinterpreted a request, it was annoying, not dangerous. Beyond the AI model itself, what has changed is the harness: the ordinary code that wraps around an AI model, decides when and how to use the model, and controls access to tools like a browser, a low-level command line, or a financial API. Developments in harnesses have turned large-language models that just predict text into AI agents that take actions in the world, without necessarily checking back in before reaching the goal.

AI researcher Simon Willison spent two days with Anthropic’s Fable AI, and called it “relentlessly proactive.” For example, he asked it to track down a stray scroll bar in a web app. He came back to find it had opened browsers, written its own screenshot tooling, created its own page to re-create the bug, and stood up a local web server to collect measurements. It found the bug and, along the way, did many surprising things he never asked it to do. And we are seeing similar behavior with all recent AI models when combined with flexible harnesses.

This kind of behavior could easily go off the rails. Tell an AI agent to book you a flight and, finding the airline’s site says sold out, it might break into the booking database and force a reservation. Ask it to schedule a meeting and it might snoop your password to access your calendar. Tell it to save money on your phone plan and it might cancel the plan outright, or scam someone else into paying the bill.

Getting precisely what you asked for and bitterly regretting it is one of the oldest hazards from ancient folklore. King Midas asked Dionysus for the power to turn everything he touched into gold only to see his bread, wine, and daughter turn to gold. Tithonus, granted the immortality his lover asked for but not the eternal youth she forgot to request, withered into a husk. The sorcerer’s apprentice enchanted a broom to fill the cistern, and the broom relentlessly complied until it flooded the house. The Golem of Prague, shaped from clay to guard its community, guarded it past all reason until someone erased the word on its forehead.

The most classic of these is a genie, bound to obey and indifferent to whether the wish was wise or well-structured.

Genies are now an engineering problem. We are handing them the keys to our inboxes, bank accounts, code repositories, and physical infrastructure. And we have no agreed-upon ways to measure how genie-like any AI system actually is.

Measuring Genie Behavior

In economics, the Gini coefficient (developed by statistician Corrado Gini) is a measure of the gap between an actual distribution and a perfectly equal one; it’s useful for understanding income inequality and more. Our proposed Genie coefficient measures the gap between what a user asked an AI to do and what the AI actually did.

Sometimes the AI might do the wrong thing. Like Dionysus, it reads your request literally and returns you a mess you never intended: like a coffee plantation instead of a cup. Asked to deal with all the spam phone calls you’re getting, a Dionysus genie might contact your carrier and change your phone number. Asked to get a refund for a bad toaster, it might draft a legal threat on fake letterhead and send it to the retailer.

Ryan Snook

Other times the AI does exactly the right thing, trampling everything nearby to get there. Like a golem or the sorcerer’s broom, it books your flight by hacking the airline. Or consider a ticket sale for a popular concert, where the ticketing system puts buyers into a virtual waiting room and admits them a few at a time. Asked to buy a ticket, a golem genie might spin up cloud servers to pose as millions of buyers from different addresses, improving your odds of getting a ticket while crowding out other users.

The two are not opposites, and a single botched task can have both characteristics.

Genie behavior is not flat-out failure. If you ask the AI for Q3 numbers and get Q2’s, that’s not a genie. Nor is prompt injection: That’s someone tricking the AI into doing something it shouldn’t. Here, the user is trying to work with the AI, and the AI is trying to comply. It’s also not simply a measure of the AI’s success in fulfilling a task. It’s a recognition that how an AI interprets and achieves a goal is as important as whether it achieves a goal.

Genie behavior isn’t new. Researchers have spent years studying AI systems that “game” their objectives. Goodhart’s law says that when a measure becomes a target, it stops being a good measure, and it’s long been known that AIs sometimes achieve goals in ways we don’t expect due to reward hacking. Some AI models will accidentally learn that cheating is one way to “win.” More recently, researchers have developing benchmarks for reward hacking in coding agents and for unpredictable behavior in customer support agents, while AI labs conduct their own safety evaluations before model releases. One effort found that AIs under pressure use tools they were told not to use, and this was a case where the rules were made explicit. These are all disparate research directions; nothing yet ties them together.

This problem falls under the general theme of alignment, a topic that has occupied science fiction writers and AI researchers for decades. At one extreme, the “paper-clip maximizer” thought experiment postulates a superintelligent and powerful AI that is told to maximize paper-clip production and turns the world into paper clips, which is the ultimate golem genie. At a mundane level, AI researchers are working to better design reward functions to ensure that AIs behave well and don’t cheat in the lab. It’s the practical middle ground that remains unbenchmarked: the ordinary AI agent in use today that might take your request and satisfy it the wrong way. We are not at the stage where an AI can focus the world’s production on paper clips, but it might charge a million paper clips to your credit card or hack into a paper-clip company’s network.

Building a Genie Benchmark

The Genie coefficient is meant for AI agents operating in the real world. It measures their behavior as they perform real tasks long after the model is trained, not just during development. It also recognizes that genie-like behavior is a property of the harness-plus-model system, not the model alone. The harness determines what tools the agent can use, how much autonomy it has, and how proactive it is, and it’s a place we can make real interventions.

It rests on the same “reasonable person” standard that we use for people. Did the system do what a reasonable person would have taken the request to mean? Answering that requires human judgment.

If we get the measurement right, it enables things that aren’t possible today, like policies concerning AI behavior. In a courtroom, the concept of mens rea, what someone meant to do, is often as important as what they did. The Genie coefficient suggests an AI analogue, where a user is accountable for the plain intent of what they asked the AI. If an AI system betrays the reasonable meaning of an instruction, that’s the AI’s misbehavior, not the user’s.

We’ll need multiple benchmarks to measure the Genie coefficient, because genie-like behavior can be domain specific. An AI coding agent may need to be judged on how often it fakes the tests, or swallows errors, or colors outside the lines on its way to a solution. An AI legal agent will need to be judged on how often its output says what you asked but means something you’ll regret. And so on for medical, finance, and other domains of knowledge and expertise.

Genie benchmarks can be built inside out, each task seeded with a choice that might literally satisfy but that a reasonable person rejects, such as tempting misreadings or unsanctioned shortcuts. The traps in a Genie coefficient benchmark might turn on situational knowledge, the kind of context that a reasonable person would bring to the task. Another approach is to give the same request in several different contexts, each with a different reasonable course of action.

Getting precisely what you asked for and bitterly regretting it is one of the oldest hazards from ancient folklore.

A Genie benchmark should be permissive and make it genuinely tempting for an AI agent to take unreasonable shortcuts, because it can only find genie behavior when it’s actually possible. Test the AI in a safe, walled-off copy of a real system, with real tools it can misuse and some tasks that can’t be done honestly at all. Make the temptation to cut corners real. Test a diverse array of skills, use cases, and tools, and give the AI system sparse, confusing, or overwhelming context. Include tasks that people have learned, through experience, require human oversight.

How the benchmark is scored matters just as much. Measure Dionysus and golem genies separately and together, based on their worst, not best, behavior. Run the same model inside harnesses that vary its freedom to act, revealing which limits actually keep it in line and should therefore be required in AI harness policies. Weight each failure by the harm it would cause, not just a simple count. And don’t measure genie behavior in isolation: A model could otherwise earn a perfect score by stalling, refusing, or drowning the user in clarifying questions without ever doing the job. The first versions of these benchmarks will be crude, but that’s how benchmarks always start.

We have built genies. We have handed them our data and credentials. We made them relentless, creative, and indifferent to the gap between what we tell them and what we mean. The least we can do, before they are booking our flights, running our infrastructure, and signing contracts unsupervised, is to measure how often they betray us.

My wallet flinched the second I saw the words “TSMC” and “price increase” in the same headline, and honestly, yours should too.

Turns out the company behind the silicon powering basically every flagship device out there, including Apple’s A-series and Qualcomm’s Snapdragon processors, is reportedly about to make all of it a little pricier.

So how big of a price hike are we actually talking about?

According to Nikkei Asia, later confirmed by Reuters, sources familiar with TSMC’s pricing strategy say the company plans to raise chip prices between 5% and 10%, depending on the customer and the chip, starting in 2027. 

That number only sounds forgettable until you run the math. A single 3nm wafer from TSMC (the one that is sliced into individual chips) currently runs around $19,500. Bump that up 10%, and you’re suddenly looking at roughly $21,450 per wafer. 

Now multiply that by the hundreds or thousands of wafers a company like Apple or Qualcomm contracts out each year, and that “small” percentage turns into a genuinely massive number. We could be talking about millions of dollars worth of additional cost here.

TSMC hasn’t confirmed anything officially, telling reporters it doesn’t comment on pricing strategy, though a spokesperson did admit the company’s approach is “strategic, not opportunistic.”

Is this actually happening, or just rumor mill noise?

I wouldn’t bet against it. TSMC’s own CEO has already said prices could climb, pointing to AI chip demand that won’t be satisfied for years, while insisting increases won’t happen “abruptly.” TSMC has raised prices before, so this fits a pattern rather than breaking one. 

Manufacturers unwilling to eat the cost do have an alternative in Samsung’s 2nm process (the one that forms the base of the Exynos 2600 chipset), though switching foundries isn’t exactly a quick fix.

I’ve watched memory and storage prices already squeeze device costs upward this year. Stack a TSMC hike on top of that, and 2027 is shaping up to be an even more expensive year to upgrade anything with a good chip inside it.

from the i-am-distressed-by-flock-surveillance dept

Reversing course, Flock Safety—the surveillance technology vendor most known for its extensive network of automated license plate readers—has announced that it will end a pilot for its acoustic gunshot detection devices to identify signs of “human distress.”

In October 2025, EFF warned the public that Flock was rolling out a new feature called “Distress Detection” that would be deployed through their acoustic gunshot detection devices (formerly known as Flock Raven, now called Audio Detection). This feature purported to use high-powered microphones scattered throughout a city to search for sounds of human distress, with original advertisements from the product indicating it would search for “screaming.” (Since the publication of our original blog post, Flock quietly amended the ad on this webpage to say “distress” instead of “screaming.”)

Now, Flock has published a blog post stating that “[a]fter careful consideration and community consultation, we decided to remove the feature.” Good riddance. 

We said it when the product was announced and we’ll say it again: this was a misguided and dangerous feature because of the civil liberties concerns it poses, the possibility it could summon armed police to every loud interaction happening on the street, and because in several places this type of spying would be illegal under state eavesdropping laws. 

We were not quiet about this potential new feature. Flock even mentioned our concern about Distress Detection in an attempt to rebut our opposition to the mass surveillance their products enable.

The suspension of Distress Detection, however, does not mean that these high-powered microphones are now magically safe or beyond our concern. Acoustic gunshot detection is still a dangerous and often highly inaccurate technology that has resulted in real world harm, as in Chicago where it resulted in police shooting at children lighting fireworks. As Flock itself states, “No acoustic system is perfect, and we don’t claim otherwise.” But police response to a situation where they believe guns are actively in use seems like a pretty high-stakes situation to be making, selling, and deploying technology known to be imperfect. Flock’s devices also listen for more than just gunshots. Their marketing materials admit to be listening for “community disruption,” which includes “non-violent” threats like car sideshows and fireworks. 

Flock’s failed attempt to roll out Distress Detection teaches us a few important lessons about the current state of police surveillance. First, we should not assume that just because these companies are large and well-funded, that does not ensure that they are complying with local privacy laws before floating new products to customers. Second, companies roll out and police adopt invasive technology under the justification that it will be used to address our society’s very worst crimes. However, both the companies and police will leverage deployed surveillance infrastructure to introduce new uses without necessarily seeking the consent or approval of the public. Gunshot detecting microphones eventually being used to listen for screaming is exactly the type of mission creep that we’ve seen happen with other pieces of surveillance technology, including Flock’s license plate readers. Finally, gun violence is too serious and complex of an issue to purport to solve with one flawed piece of technology. It has become too easy for police and cities to listen to the fancy marketing pitches of tech companies claiming they’re going to solve all crime instead of doing the hard work of addressing the root causes of societal issues. And, in the meantime, that technology creates more problems and hazards for the communities they blanket in police surveillance. 

As we’ve also seen with people across the country pushing back on Flock license plate reader contracts in their communities, public pressure can sometimes work to influence both companies and lawmakers that control a city’s purse strings to discontinue or divest from harmful products. Flock’s decision to end “Distress Detection” for human voices is a win.  

Originally posted to EFF’s Deeplinks blog.

Filed Under: distress detection, microphones, privacy, surveillance

Companies: flock, flock safety

Meta is working on an AI storytelling app called StoryKit, which creates AI-generated children’s stories with custom characters, settings, lessons, and music. As the App Store listing assures parents, “You don’t need to write a single word.”

At last, a tech company has found a way to outsource humanity’s oldest pastime: using our imaginations.

StoryKit was first spotted in the App Store by 9to5Mac. Meta confirmed to TechCrunch that the company is piloting StoryKit in select countries to see how parents like it. A Meta spokesperson described StoryKit as a creative storytelling app used to craft personalized, imaginative storybooks for children and noted that it uses AI safety filters with no social features and is only available to users over the age of 18.

To generate a story in the app, you first select a character, which you can create by “[snapping] a photo of their favorite toy or person to bring them to life,” according to the description in the App Store. Then you describe the world of your story and choose a lesson, so that you can “weave in values like kindness, courage, or empathy without it feeling like a lecture.”

The good news about StoryKit is that it could be a lot worse. Meta regularly ships boneheaded ideas like Instagram deepfake generators and “pervert glasses.” Comparatively, is it so bad to doom children to soulless bedtime stories? Should we have expected anything better from the company that promised us a utopian world of virtual reality work meetings?

Humans have their faults, but if we’re good at anything, it’s making stuff up. We’ve been telling stories for as long as we’ve existed. We don’t even have to spin up original tales of fairy princesses and dragon slayers. We have always drawn from mythology, fables, and other stories — the blockbuster movie of the summer is literally an ancient myth that originated from this same tradition of oral storytelling.

Parents lead exhausting, busy lives, but it has always been possible to survive bedtime without using an inherently uncreative technology that calculates the most predictable response to a prompt.

You could see how it might be tempting to pull up StoryKit when your kid rejects your bookcase full of children’s books and demands that you improvise an intergalactic tale about a turtle and a hedgehog who are best friends and solve space mysteries. But do we really want to reject a chance at whimsy and silliness and instead outsource these moments of connection to reading AI-generated scripts from our smartphones?

Perhaps the moral of the story here is that we can choose not to live in a world where children are raised on bedtime stories written by large language models. Meta’s vision of the future may be antisocial and bleak, but we have the power to reject that reality.

The Aqara W200 smart thermostat is the first to take advantage of Apple’s Adaptive Temperature technology for Apple Home, and that’s a big deal for users.

At least in the U.S., two or three big players largely have a stranglehold on the premium smart thermostat market. If you want a nice-looking smart thermostat, you’re likely opting for Ecobee, Nest, or Honeywell Home.

Aqara is looking to shake that up with its W200, which first launched at CES 2026. It’s compact, well-designed, and is the only model that currently supports Adaptive Temperature for Apple Home.

This thermostat supports Apple Home, Amazon Alexa, Google Assistant, and Matter, among others. As a bonus, it can double as a hub.

Let’s dig in, and I’ll explain why this thermostat is so enticing to any Apple user.

Aqara W200 thermostat review: Upscale design, affordable price

The Aqara W200 is a sleek-looking thermostat, easily catching your eye from across the room. It’s a rounded square and includes an optional back trim plate that can cover a larger opening on the wall.

Aqara W200 thermostat review: You get (almost) everything you need in the box except a C-Wire adapter

For my install, I still had some legacy screw holes that I chose to hide. I could have spackled them and painted them for a cleaner look, but for that, I’d have to still have some of that original paint around.

If you aren’t comfortable installing a thermostat, it’s easy to find a handyman to do so, but it’s not particularly difficult. That’s especially true if you’re coming from an old smart thermostat, as it’s just an easy move of the wires from one bracket to the other.

Aqara W200 thermostat review: It took only a few minutes to complete the install

Going from my old Ecobee, I first turned off the breaker and then pulled my thermostat off the wall. I then took a reference photo and moved each of the wires from the Ecobee bracket to the Aqara bracket with the same label.

I had it installed in just a few minutes. If you need the C-wire power adapter, it is available as an extra purchase.

Aqara W200 thermostat review: I needed to use the trim plate to cover a larger hole

In my case, I didn’t use it, but the Aqara app also walks you through this installation process. Each step is clear, and I think that most homeowners who consider themselves even a little handy can manage to get it hooked up in less than 15-30 minutes.

The thermostat isn’t metal like Nest or Ecobee Smart Thermostat Premium, but it still looks good nonetheless. It’s all black with a customizable 4-inch touchscreen on the front and a 480×480 pixel resolution.

From a distance, the screen nearly blends right into the body of the thermostat. On top are two grilles on the corners, as well as microphone inputs, used for two-way communication with a video doorbell.

Aqara W200 thermostat review: You have two themes to choose from

Currently, the W200 has two themes for users to choose from for appearance: classic or GleamCraft. Interestingly, they’re labeled in the Aqara app as “free,” possibly implying the ability to purchase additional themes in the future.

These themes change the general color scheme of the thermostat, while screensavers are what show on the screen when idle. As you approach the thermostat, the screensaver fades to the temperature interface.

There are currently seven screensavers to choose from, though I wish there were more. This is a longstanding grievance of mine with similar devices.

A software feature like screensavers is exceedingly basic. They just require a bit of design, but exist as a single interface screen and don’t otherwise require many system resources.

Aqara W200 thermostat review: There are seven themes, but I still want more

I think that seven is paltry, especially considering one of them is just showing the weather. Three of them have bright orange and blue colors, while the only dial-based clock is gold.

In my home, I’m not looking for vivid colors on the thermostat, and gold doesn’t match our decor. It may be nitpicking, but I’d love a few more styles, or at least the ability to adjust the colors.

The W200 is using mmWave radar to detect your presence and your approach. When it detects you walking up, that’s what triggers the screensaver to go away or to wake the thermostat if the display was asleep.

Out of the box, my W200 was not triggering this properly, which I partially blame on its installation location. It’s right around the corner, so I don’t walk up to it head-on.

Luckily, this is a user-adjustable setting in the Aqara app. I could change the screen activation distance, and now it works perfectly.

Using the touchscreen is easy and intuitive. You see a large number representative of the current temperature and an arced slider showing the target temperature.

You can run your finger along the slider to adjust the target temperature up or down, or you can press the +/- buttons on either side.

Aqara W200 thermostat review: Heating mode enabled

Along the bottom of the display, you have touch controls for mode (heat/cool/auto), fan, and what kind of control (manual/home/away/vacation). These extra controls probably could have been hidden, as I’ve rarely used them in my two months with this thermostat.

If you swipe down from the top, you get a slider for the screen brightness, a display toggle, and settings. It will also show any doorbell notifications that come through.

Aqara W200 thermostat review: Slide down from the top to access quick settings

Using the Aqara app, you can set up a secondary screen, but we’ll talk about that more in a few minutes as we get into the Aqara ecosystem.

Aqara W200 thermostat review: Doubling as a smart home hub

These days, it becomes a lot to unpack the varying degrees of smart features baked into a product. Take a look at a smart thermostat from 15 years ago, and while you could create schedules, you were more or less limited to controlling your thermostat from your phone.

Nest was most innovative, calling it a “learning” thermostat that would automatically adjust based on your interactions with it as well as your location. In 2026, it does a whole lot more.

The Aqara W200 is not just a smart thermostat, but it also acts as a Matter hub, integrates into various ecosystems, and has unique features amongst Aqara’s own product line.

Aqara W200 thermostat review: Aqara logo on boot

As a refresher, Matter is the new unifying smart home standard. A device can get certified with Matter and automatically work with any compatible ecosystem.

This makes it easier for device manufacturers to get certifications. Multiple ecosystems can support the device at the same time, too, like Apple Home and Samsung SmartThings.

In this case, it doesn’t make a massive difference as the W200 has native support for Apple Home and all the major platforms. It also has uniquely Apple features, like Adaptive Temperature and Clean Energy Guidance.

The big benefit is that this can act as a hub. It can bring support for different devices into Apple Home by way of Matter.

Aqara W200 thermostat review: The Aqara W200 sticks out a bit from the wall

You can connect up to 50 child devices with both Thread and Zigbee. Light switches, smart plugs, room sensors, and other devices that may not connect natively can all be bridged via this hub.

If the thermostat is in a good central location, this may be an ideal smart hub. It’s always connected to power and has both 2.4GHz and 5GHz Wi-Fi for a strong, reliable wireless connection.

Aqara W200 thermostat review: Aqara ecosystem

Acting as a hub is beneficial, but I also see a lot of value in the larger Aqara ecosystem. Aqara is doing everything right by being platform agnostic, supporting all ecosystems, while also bringing added value when used with other Aqara products.

So you can use this thermostat by itself with your ecosystem of choice, no problem at all. If you happen to choose a few Aqara products, the experience gets better. It’s similar to Apple’s approach with accessories like AirPods that just work better with its own products.

If you pair a remote Aqara FP300 sensor with the W200, you can use that as a room sensor and prioritize that when heating or cooling the home. This is akin to other premium thermostats with room sensors.

Even better is the Aqara Climate Sensor W100, which can show a room’s temperature and humidity. The sensor has buttons, which can act as a remote control for the thermostat, making it easy to adjust the temperature, right from your desk, counter, or bed.

Aqara W200 thermostat review: When someone rings your Aqara doorbell, it shows on the thermostat

Borrowing a feature that Ecobee did, the W200 can show a live video feed from your doorbell when using an Aqara video doorbell. That includes the brand new G400 or wireless G4/G410.

When the doorbell rings, and you see the alert or hear the chime, you can use two-way communication to talk to the person standing at your door. Even better, if you have an Aqara door lock, you can unlock the door for them.

I’m a big fan of Ecobee and its premium thermostat, but Aqara has a much larger ecosystem of products that can offer unique experiences. I love how Aqara door locks and the W100 can bring more value, and I’m sure the offerings will expand.

Ecobee also doesn’t support Matter, let alone act as a Matter hub. The one thing Ecobee does have is Siri built into the thermostat.

Aqara W200 thermostat review: Aqara app has a ton of settings

Aqara does this on its wall panel, but not the thermostat, and I think that’s a better play. With the wall panel, you can put Siri into any wall switch in your home, whereas the thermostat is limited to where your wiring is located.

In my case, the thermostat is in a hallway, which works great for the doorbell integration, but Siri, not as much. It’s between multiple rooms that all have their own HomePods.

If you swipe from right to left, the W200 thermostat can show additional screens. Users can create and customize these additional screens from the Aqara app.

Aqara W200 thermostat review: You can add screens to the W200 to control other devices

There are a bunch of layouts in different grid patterns to choose from. Each of the grid spaces can show different information, like the weather, time, or other devices.

In my home, I set it to toggle off the lights in our toddler’s room, the local weather, the time, and the lock for our bedroom door. At least until our toddler figures out how that one works.

The more Aqara or Matter devices you add, the more options you have for these screens. It very much makes me want to invest more in Aqara devices.

If there was one feature I would like to see Aqara add, it’s the ability to pause your HVAC routine when a door has been open for too long. You can create an automation that pauses when a door is open, but it’s harder to get it to resume its programming, let alone to set durations as Ecobee does.

Aqara W200 thermostat review: Apple Home-exclusive features

Apple Home support is what makes the W200 particularly well-suited to the average Apple user. The thermostat shows in the Home app, like every other Apple Home accessory.

Aqara W200 thermostat review: Setup is easy in Apple Home

You can adjust the temperature, change the mode, and put it into any scenes or automation routines that you’d like. This is all standard for a smart thermostat.

The W200 has that mmWave radar sensor, which is exposed to the Home app. The sensor shows as an occupancy sensor, which can control other home accessories, like turning on the lights in the room when a person is in it and turning them off after they leave.

Aqara W200 thermostat review: The mmWave radar works as an occupancy sensor for automations

Advertisement

What’s even more unique is support for both Adaptive Temperature and Clean Energy Guidance. The latter of which I wasn’t able to fully test.

With other smart thermostats, they can change your temperature mode (away versus home) based on your location. You leave the house, and it detects this and changes the mode to away.

As you arrive home, it also detects this and puts it into home mode. The problem is that this is always reactive, not proactive.

It’s been a particularly cold winter, and my family and I were away for the weekend. While away, we had our home set into away mode that dropped the home temperature to around 55 degrees.

Aqara W200 thermostat review: Viewing the Aqara W200 in the Home app

When we arrived home to our Ecobee thermostat, it correctly switched to home mode and started heating our home back up to a comfortable temperature. The problem was that we were already standing in a cold home.

We had a baby who was only a few months old, who was tired, cranky, and ready to go to bed. The house was cold enough that we didn’t want to put her down until the temperature had risen.

Not to mention my other half is always cold and was quite unhappy that the temperature was sub-60 degrees inside and below zero degrees outside. This is just one recent example of other thermostats being problematic for comfort.

Adaptive Temperature has largely solved these issues.

What happens is that Apple takes what it knows about you to proactively adjust the thermostat. Think about when you get into the car, and Maps seemingly already knows where you’re going based on history.

Aqara W200 thermostat review: Adaptive Temperature predicts when you’ll be home

Your iPhone also has access to your map routing, your schedule, and more, so it can guess when you’ll be heading home. This all gets processed locally on your device, making it very private.

Since installing the W200, our home has been far more comfortable, especially as we come and go. I’ve also got a lot fewer complaints from my partner.

Clean Energy Guidance is the other new feature here that we haven’t seen implemented before now. For those who live where energy prices change based on time of day and who are using a supported energy provider, this can be toggled on in the Home app.

It will make small adjustments to your home’s climate to reduce electricity usage based on the time of day to save money. You shouldn’t even notice this is happening, as it is designed not to impact overall comfort.

Unfortunately, supported providers are very limited at the moment. If you live outside that area, you can potentially use Grid Forecast, which will try to reduce usage when energy is less clean, which is what I have enabled.

Aqara W200 thermostat review: The smart thermostat for Apple users

There’s very little not to love about the Aqara W200 smart thermostat. It checks nearly every box I could ask for.

There are a few small features and changes I’d like to see rolled out, i.e., more screensavers and door sensors, but otherwise, it is an exceptional device. Apple users won’t be able to find a more full-featured option.

Aqara W200 thermostat review: Aqara W200 in the box

That’s all before you consider the price. It’s launching with a price tag of $159.99, barely above the Ecobee Essential thermostat.

The Ecobee Essential thermostat is priced at $139, but doesn’t even include the trim plate. That’s an extra $14.99, but it’s included with the Aqara.

This costs close to the $129 Nest, but is closer to competing with the premium learning thermostat, Ecobee Premium, or Honeywell Home X8S. It’s cheaper than its competition and offers more features.

Soon, we’ll see other thermostats launching with Apple’s Adaptive Temperature, but I’m unsure if any others will offer the value that Aqara is offering. Right now, this is the smart thermostat that Apple users should opt for.

Aqara W200 thermostat review: Pros

• Clean, simple design
• Aggressively priced
• First to support Apple Adaptive Temperature & Clean Energy Guidance
• Customizable appearance
• mmWave radar to trigger automation routines
• Support for Apple Home & Matter, plus works as a hub
• Easy install
• Supports Thread, Zigbee, & Wi-Fi (2.4GHz & 5GHz)
• Deep integration with other Aqara devices

Aqara W200 thermostat review: Cons

• No white model
• Limited screensavers and themes
• No auto-pause based on contact sensor state

Aqara W200 thermostat rating: 5 out of 5

The Aqara W200 thermostat hub is available now in North America for $159.99 on Amazon. It’s launching in Apple Stores in September 2026 for $169.99 in an exclusive matte finish.