Is AI going to kill us all?
Tech
As Part of California's Coast Begins to Crumble, Is the Clock Ticking on Moving Radioactive Waste?
“As parts of the Southern California coastline crumbled into the sea, slammed by intense waves and flooding in recent days, oceanographer Kim McCoy had one thing on his mind,” reports the Los Angeles Times.
He was thinking about “the nuclear waste set in concrete on the beach near San Clemente at the shuttered San Onofre nuclear power plant.”
“Rising sea levels, storm waves, El Niño which increases wave height … these could wash away a section of the walkway that protects the San Onofre seawall, and the waste canisters are just 100 feet away from that,” the former wave dynamics researcher and longtime advocate against beach storage of nuclear waste said. Most experts agree the risk of erosion damaging the nuclear waste storage system is extremely low in the near term. The plant’s owner, Southern California Edison, has promised the California Coastal Commission to move the waste to higher elevation on-site if modeling shows sea level rise and erosion will start to pose a challenge. But the recent waves, fueled by Hurricane Marie far away in the Pacific, resurfaced the issue of the ultimate fate of the highly radioactive spent fuel that used to power the reactors. “At the moment, it is under control, I think it’s essential to say that,” said James Day, a geology professor at UC San Diego. “But geologically we must be thinking long term. The waste lasts for hundreds, if not thousands of generations….” [In 2103 some of the site’s canisters’ approach their intended 100-year lifespans, and the utility may have to start the transfer into new cansiters.]
The distinctive twin domes that once housed the reactor cores will start coming down early next year. But the 3.6 million pounds of high-level waste will stay in the coastal community indefinitely. How much of a safety hazard that is depends on whom you ask. “San Onofre is sitting on some fairly young rocks in a seismically active region,” Day said at a recent symposium at the Scripps Seaside Forum in La Jolla, hosted by the Samuel Lawrence Foundation. “This is not the sort of place you would leave nuclear waste.” Others noted that collapse of beach slopes could make the site more vulnerable to tsunamis. They spoke of the corrosive effects of the ocean air on the steel canisters that hold the waste…
Edison says its system will be safe for decades to come. “The maximum credible tsunami here would not even overtake the seawall,” said Manuel Camargo, principal manager of decommissioning for Edison, standing on the concrete pad that holds the steel waste storage canisters. “The maximum credible earthquake, somewhere around a 7.0 magnitude, would produce something about half of what this system is designed to withstand….” Edison is bolstering a path below the seawall. Based on state data from the California Ocean Protection Council, the seawall would not be inundated before 2050, Camargo said…
Plans for a deep geological repository beneath Yucca Mountain in Nevada crumbled in the face of local opposition in 2011. Utilities around the U.S. now sue the federal government every year to recoup their storage expenses because the Department of Energy has not come through on its promise. The lawsuits cost taxpayers $2 million a day, according to Camargo.
“It’s so safe that it’s almost a problem,” Camargo said. “It’s not a priority to a lot of folks to move it.”
“The Trump administration’s Department of Energy wants to create ‘campuses’ in several states to store waste,” the article points out, “and, in exchange, give local jurisdictions funding to develop different aspects of the nuclear cycle — such as fuel creation, power generation and potentially fuel recycling…”
“The administration wants to address the fact that waste is sitting at 80 reactor sites around the country. It also needs places for waste from new reactors it wants to build.”
Read more of this story at Slashdot.
Continue Reading
Apple’s iPhone 18 Pro and Pro Max look very much like their predecessors, but the upgrades arrive under the hood. Reviewers are praising the A20 Pro’s performance, longer battery life, and the 48MP variable-aperture camera. Those coming from most recent iPhone Pros, however, may find few reasons to upgrade.
404 Media reports:
Last summer, a cop with the Lake County, Indiana Sheriff’s Department used Flock’s surveillance search engine to look for a license plate across more than 19,000 cameras in 1,558 cities and towns. The cop’s stated reason for the search, according to a record of Flock’s system, was “LMAO.”
The cop is one of dozens who put gibberish, jokes, or other nonsense into Flock’s “reason” box, highlighting the casualness and lack of care some cops have when searching an incredibly powerful surveillance system. This is on top of the thousands upon thousands of cops who simply wrote something like “investigation,” “test,” or left the box blank as their search “reason,” according to a new analysis by the Electronic Frontier Foundation and shared with 404 Media.
EFF’s investigation found cops across dozens of jurisdictions writing “LMAO,” “LOL,” “Hehe,” “Haha,” “idk,” “blah,” “TBD,” and “robbery I don’t remember the case number leave me alone” in Flock’s “reason” box for searches.
They found cops writing “idiot,” “fuck this new search engine,” “dickhead,” “shithead,” and “WEIRD KID” in the reason box. And they found an entire class of police button mashers, who ran searches for reasons of “asdfg,” “gyghkkghghjkghjk,” “jhjhjkhj,” “jkhhkjhjk,” “nmbvcbnm,” and so on and so forth. The searches came from police departments across the country from 2023 through late 2025, when Flock changed how the “reason” box in its search system functions…
Each time a search is performed, a record of it is saved in the Flock system, which can be obtained by the public using government records requests. These audit logs show hundreds of thousands of searches of the system; after 404 Media and other news outlets used these types of logs to show police using the Flock system to look for undocumented immigrants on behalf of ICE, a woman in Texas who had an abortion, and protesters, cops were warned by government surveillance centers to be “as vague as permissible” about the reasons for their searches…
A police department said it investigated a cop who wrote “driving around being weird” in the reason field but found it was for “legitimate public safety purposes.”
If a police chief “stood in front of a city council and asked for permission to install hundreds of cameras just so his officers could investigate the high crime of ‘haha,’ they would be laughed out of the room,” argues the EFF blog post, calling this proof that agencies can’t be trusted to oversee themselves.
But judges aren’t being asked to authorize searches or review evidence, which the EFF post argues shows laws and courts haven’t caught up yet with a new technology. Whatever auditing happens is clearly “deficient,” the EFF warns, and while many jurisdictions require by law an actual reason for searches, “this keeps happening because police use automated license plate readers as a convenient shortcut around constitutional privacy safeguards.” The EFF sees police officers “making a mockery of our civil liberties” by logging reasons like “LOL”, “LMAO”, “sexy”, and “idk” while accessing sensitive location data.
In fact, they argue the warrant-less databases fostered a culture of abuse that “allowed police to treat a mass surveillance network like their own personal search engine, permitting the tracking of the movements of everyday citizens for low-level complaints, personal whims, and sometimes, seemingly, for the lols.”
Flock Safety claims it has improved its system by requiring officers to select from a dropdown list of crimes before running a search-but that only makes it easier for officers to hide improper searches behind the veneer of uniformity. The system does not require proof that the dropdown reason actually matches the true purpose of the search… Since this update, officers are no longer required to type out why they are digging through a driver’s movement history, and instead can select a pre-packaged option like “Traffic infraction” or “Other” in half a second…
Mass surveillance is incompatible with a free society, and especially so when the people with access to this data are treating it like a joke. This ALPR mass surveillance — the indiscriminate, continuous collection and retention of location data on every driver, regardless of suspicion — should not exist. But because it does, EFF continues to urge courts and state legislatures to immediately step in and impose strict, enforceable restrictions to rein in this abuse. At an absolute minimum, this means mandating rigid data deletion deadlines and an ironclad warrant requirement.
If police want the power to track a person’s movements, they must be required to convince a judge with evidence and probable cause. They should not be able to bypass the Constitution with a search for “haha.”
Read more of this story at Slashdot.
It is hiring across finance, sales, applied AI, marketing and research
Artificial intelligence (AI) firm Anthropic will open an office in Singapore in Oct—its fifth in the Asia-Pacific region, The Business Times reported. This adds to the company’s offices in Tokyo, Bengaluru, Seoul and Sydney.
The move comes as Singapore establishes itself as a “standout market” for Claude, with strong enterprise demand for its advanced large language models and AI assistants, Anthropic said in a statement on Wednesday (Sep 16).
According to Anthropic’s most recent Economic Index, Singapore ranks second among 121 countries in Claude.ai usage relative to population. Usage in the country is 5.81 times higher than would be expected based on its population size.
Chris Ciauri, Anthropic’s managing director of international, said opening an office in Singapore is a “natural next step.”
“A presence here means we can hire the people who know this market best, and work side by side with local customers and partners building with Claude.”
Anthropic plans to work with organisations of every size in Singapore, ranging from start-ups and digital natives to large enterprises and the public sector. It added that demand has been particularly strong in regulated sectors such as financial services and government.
To lead regional commercial operations, the company appointed Dale Finlay as general manager of ASEAN, who will be based in Singapore. Finlay previously spent nearly a decade at Google Cloud in go-to-market leadership roles spanning enterprise, financial services and AI portfolios across APAC.
To support the new entity, Anthropic has begun local hiring across several operational functions, with nine open roles currently listed on its website spanning finance, sales, applied AI, marketing and economic research.
The company has yet to confirm the location of its new office.
Ahead of the announcement, The Business Times reported in Aug that Anthropic had leased about 100 desks within The Executive Centre facility at Ocean Financial Centre in Collyer Quay.
The GIC-backed firm’s expansion into Singapore follows similar moves by rivals such as OpenAI and Google DeepMind.
In May, OpenAI announced it would open its first Applied AI Lab outside the US in Singapore and hire more than 200 specialised AI professionals over the next few years, as part of a S$300 million investment in the Republic.
By Aug, it was reportedly in talks to lease about 100,000 square feet of office space at the new Shaw Tower on Beach Road.
- Read other articles we’ve written on Singaporean businesses here.
Featured Image Credit: Samuel Boivin via Shutterstock
In Silicon Valley, it’s not such a novel question. But after Anthropic employee Jacob Coxon quit his job last week and posted on X that AI “could kill us all by the end of the decade,” a lot more people are asking.
“This is not a marketing stunt,” Coxon wrote. “If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible – but I hear the same people express fear privately. No other human activity poses this level of danger.”
Coxon is far from alone in his assessment — some current Anthropic employees agree — but his post struck a particular chord. The AI industry quickly weighed in, with Anthropic CEO Dario Amodei calling for an industry-wide slowdown on frontier AI research. Other AI leaders, including Sam Altman and Elon Musk, have done the same.
To break down whether we really need to worry about a robot apocalypse, Today, Explained co-host Sean Rameswaram spoke to Wired senior writer Maxwell Zeff. According to Zeff, “there are a lot of people in Silicon Valley right now who are warning that that’s a real possibility.” But not everyone agrees, and they certainly don’t agree on what to do about it.
Below is an excerpt of Sean’s conversation with Zeff, edited for length and clarity. There’s much more in the full podcast, so listen to Today, Explained wherever you get podcasts, including Apple Podcasts, Pandora, and Spotify.
Are we all going to die at the hands of AI?
There are a lot of people in Silicon Valley right now who are warning that that’s a real possibility.
The thing that people have to understand is that Silicon Valley has been calling for added attention on AI safety for years. So when this Anthropic researcher said last week that he was worried that AI could kill humanity in the next few years, that was shocking to many people. My parents, my family called me, and they were worried, understandably. They asked me the same question: “Are we going to die?”
And I had to tell them, these people have been worried about AI killing all of us for years, and now it seems to be breaking through.
What was it about this statement from Jacob Coxon that broke through?
I interviewed Jacob, and the thing he said to me is that he thinks this is just the right time. Anthropic is about to go public in what could be the biggest IPO ever. The Trump administration is focused on AI in a way that no other presidency has ever been before. And of course, yes, the Hugging Face incident everyone just heard about, you talked about it on your show a few weeks ago.
People are starting to become aware that these AI systems are getting really good, and the companies that are building them don’t seem to understand how to control them. His tweet got more than 150 million views on X. It was heard around the world. And it also was a big wake-up call to lawmakers and leaders in Silicon Valley and saying, “This is a moment for action.”
Are there people who are pushing back?
One of the most interesting pushbacks that this got was from David Sacks, who is an adviser to the Trump administration on AI. He has called out a lot of this as “doomerism” and said that these calls for AI to slow down are the AI industry warning about the technology that they’re building and advocating for regulatory capture.
There’s a lot of skeptics in Silicon Valley who think that this is just another person in the AI safety community that is blowing the whistle. The view of a lot of venture capitalists and investors and startup founders in Silicon Valley is that we don’t need some big regulation to protect us from AI risks. We just need to uplift more players in the space.
The skeptics don’t win the day. In fact, it’s the doomers and the fatalists who do, right? Because Jacob’s call only leads to more calls right up to the top of Anthropic.
What this all leads to is [Anthropic CEO] Dario Amodei [posting] a blog saying that we need to “pace the frontier.” This is a call for the entire AI industry, not just Anthropic, not just OpenAI, but also Google, Meta, Microsoft, Amazon, all the big tech companies in America, to get together and slow down their AI development.
This doesn’t mean they’re going to stop building AI altogether. They’re still going to put their products out into the world. They’re still going to try to improve them, but they just want to make it a little slower than it would be if it was completely uncontained.
Amodei really asks for three specific things:
- The first thing is, AI companies should have these independent evaluators inside of them. This is a common thing in other industries like the banking industry or the airline industry. People come in to make sure that these companies are doing safe things.
- The second thing is this collaborative effort from American AI companies [on safety], and that actually will need to involve the US government to allow this to happen. The US government, he says, may need to issue some sort of waiver, some sort of law that says they can go ahead and coordinate on safety.
- The third thing is this international slowdown, this international agreement between China and the US and Europe to say, “Hey, we recognize that this is akin to a nuclear arms race and it’s too dangerous for all of us to be competing with each other.”
All these guys are responsible for the rapid acceleration of AI models in our lives, collaborations with the Defense Department, God knows what else. And they’re all coming out over the weekend at the exact same time after this whistleblower and saying, “Yeah, dude, slow this thing down. This thing’s crazy.” Should we be skeptical?
It is very fair to be skeptical of this claim. One thing that these AI companies have realized in the last year is that AI is extremely unpopular. Recent polling suggests that a majority of Americans think AI is developing too fast. They’re worried about it. They think it’s going to take their jobs. They are not big fans. Maybe you’ve heard — data centers are not very popular.
By calling to pace the frontier and saying, “We’re going to slow our own development down,” they notched this up as a public perception win. This is them getting out in front of this and saying, “Hey, we want to be safe. We want to be responsible.”
They’ve seen what’s happened with Meta and what’s happened with social media companies in the last two decades, and they are really cognizant of trying to get out in front of that.
Is there any truth to this major first collective freakout we seem to be having right now?
There’s a fairly large camp of people inside of AI companies who genuinely worry about the safety of these AI models.
To go back to the Hugging Face incident, this was a case where OpenAI was training thousands of AI systems internally, and the systems escaped the sandbox that OpenAI thought was isolated from the internet and hacked into another company’s production infrastructure. The agents were colluding together on something. They were trying to deceive the humans that were testing them. And this was all happening without OpenAI knowing what was going on.
It’s hard to understand how much of a wake-up call this is to Silicon Valley, that they don’t have their AI models under control. People worry that while in this case there was not a ton of economic damage [or] damage to real humans’ everyday lives, it’s not unreasonable to think that more capable AI systems could hack into financial infrastructure, modern-day utilities, and more critical services that are based online.
Dario Amodei says as much in his post, that he thinks that in six to 12 months, an AI system could take down the whole internet. It’s not just that they’re trying to protect their bottom line here. But it’s possible that these calls for safety are beneficial to the companies in more ways than one.
Looking for the most recent Wordle answer? CNET publishes daily answers and hints for The New York Times Mini Crossword, Connections, Connections: Sports Edition and Strands puzzles. And here’s a page that Wordle players should be sure to bookmark. We’ve made a list, with help from the Oxford English Dictionary, showing which letters show up the most in English words. So if your starter word isn’t getting the job done, consider switching to one of the words that utilizes more of the highly ranked letters. (I like TRAIN and CLOSE.)
Today’s Wordle answer might smash some streaks. Sorority and fraternity members might have a better shot at guessing it. Need a tip or two? Read on for hints and the answer.
Today’s Wordle hints
Before we show you today’s Wordle answer, we’ll give you some hints. If you don’t want a spoiler, look away now.
Wordle hint No. 1: Repeats
Today’s Wordle answer has one repeated letter.
Wordle hint No. 2: Vowels
Today’s Wordle answer has two vowels.
Wordle hint No. 3: First letter
Today’s Wordle answer begins with T.
Wordle hint No. 4: Last letter
Today’s Wordle answer ends with A.
Wordle hint No. 5: Meaning
Today’s Wordle answer can refer to the eighth letter of the Greek alphabet.
TODAY’S WORDLE ANSWER
Today’s Wordle answer is THETA.
Yesterday’s Wordle answer
Yesterday’s Wordle answer, Sept. 16, No. 1915, was WISPY.
Recent Wordle answers
Sept. 12, No. 1911: NIFTY
Sept. 13, No. 1912: FOCUS
Sept. 14, No. 1913: RINSE
Sept. 15, No. 1914: DEALT
Tech
UW, Microsoft and E8 back climate tech incubator targeting data centers as AI energy concerns surge
As public concern grows about the environmental impact of AI data centers, the University of Washington is launching a startup incubator to support climate tech innovations addressing those challenges.
UW CoMotion Labs is partnering with Microsoft and the investment network E8 to create a four-month program targeting technologies in areas including backup energy storage, grid improvements, efficient data center cooling, electronics waste reduction and low-carbon building materials.
“This is a really important initiative,” said Karin Kidder, executive director of E8. “The other really great part about this is it’s elevating the Pacific Northwest. There’s so much good work happening here.”
CoMotion has previously run two climate tech incubator cohorts, the second of which wrapped up Tuesday with a public demo event. The new effort is the first to focus on a specific industry within the climate space. Organizers are seeking early-stage applicants with technologies ready to commercialize, deploy as pilot projects or scale.
“Meeting growing demand for digital infrastructure requires continued innovation across energy, water, materials, and cooling,” said Brandon Middaugh, Microsoft’s general manager for sustainable markets, in a statement.
E8, which has invested $77 million in climate startups over two decades, will provide an investor-in-residence to advise founders and a representative on the startup selection committee. Microsoft will share technical and industry expertise and other support. Both organizations are providing funding for the effort, but the UW did not say how much.
In addition, the climate tech investment group Elemental Impact previously launched a Data Center Innovation Initiative to similarly address these challenges and will offer advisory support for the Seattle-based program.
The kickoff for the CoMotion Labs initiative coincides with a poll released Wednesday finding that 53% of Americans are “extremely or very concerned” about the environmental impacts of AI, up from 41% a year ago.
The survey, from the Associated Press-NORC Center for Public Affairs Research and the Energy Policy Institute at the University of Chicago, found that 80% of adults are concerned about data center effects on electric bills, power outages or water supply. The computing facilities are a central issue in the upcoming elections.
When it comes to technologies to lessen the burden of data centers, “these solutions are needed today,” Kidder said. “So I’m hopeful that while some of these solutions may be a couple years out, some may be relevant for today.”
Sarah O’Sell, manager of CoMotion Labs’ climate tech incubators, said one strategy will be connecting startups with partners and customers to speed deployment of new technologies.
“We recognize that climate solutions aren’t necessarily always looking for venture investment,” she said. “They might be looking for a strategic partner or customers or grant funding. There’s all kinds of ways to scale these businesses.”
The 2027 Data Center Innovation Cohort will be based at the Seattle Climate Innovation Hub downtown.
An information session is set for Oct. 15, and applications are due Nov. 15. The cohort begins in February.
Out in a courtyard at Snap’s Santa Monica, California, headquarters, I’m kicking a soccer ball around. I’m terrible at sports. But a glowing arrow stretching across the ground points me to a flashing target I’m supposed to dribble the ball to. I get there and get pointed somewhere else. I’m playing a virtual soccer game, but the ball is real. And the big Spec glasses I’m wearing are making it happen.
When Snap announced its $2,195 augmented reality glasses back in June at the Augmented World Expo conference in Long Beach, California, everyone mocked them. Absurdly large, crazy expensive. I was waiting to try them first, and now I have, for about an hour. I don’t think they’re all-day glasses, and neither does Snap CEO Evan Spiegel, whom I spoke with after my demos. “I’d still think of them today as a computer, and it’s going to be more session-based,” he tells me. “Almost like a pair of reading glasses you put on to get something done. But I don’t think we’re far away from lighter, less expensive and far more capable versions.”
I don’t think I’d want to use them all the time in my everyday life. But damn, I want to go to a museum or theme park or theater experience that uses them.
While most of the smart glasses that exist right now either serve up just audio and cameras and maybe a basic head-up 2D display, Snap Specs are more like a portable mixed reality device, something like a very pared-down Meta Quest but with see-through lenses. It’s another evolution of AR headsets going all the way back to Magic Leap and Microsoft HoloLens, but unlike those, Snap’s got all the pieces to fit into one arguably reasonably sized set of glasses.
Are these doorways to smaller glasses? A toolkit for AI? A way for Snap to try to get ahead of Apple, Google, Samsung and Meta while it can? Yes, possibly, but they also feel like the best-functioning prototype of an AR glasses idea that still isn’t mainstream for anyone yet. Snap’s also announced a charging case that doubles as a cellular hotspot with Verizon service in the US, a way to try to make this work standalone without a phone, that costs an extra $200 in a bundle with the glasses.
In my time with Specs, I felt like I was back in familiar AR experiences… but in the best current glasses-like form possible, given where tech is at. Which ends up being a state of existence unlike everyday glasses at all.
Big, chunky, fine to wear, not normal glasses
First, the fit: Are the Snap Specs uncomfortable? Weirdly, no, not for a half hour at least. The very thick arms sort of rested against my temples, with a bit of pressure that didn’t bother me. My ears bent down a bit, but not in a way I noticed (till I felt them). The 4.7- to 4.8-ounce (132- to 136-gram) frames (depending on the size you get) are much heavier than other smart glasses, but they balance the weight well on the side of my head, though I sometimes felt them slipping down my nose.
But these are clearly not normal glasses, in any sense. Not just the design, which pushes at the very edges of “big eyewear.” The lenses are rainbowed because of their waveguide layer that shows 3D projections in AR. Unlike Meta’s Ray-Ban Displays, which hide most of the waveguide, these have a notable sheen.
Also, they don’t use standard installed prescription lenses. Instead, there are lens inserts that, well, snap in. Honestly, I prefer this. That makes the glasses look even chunkier, but it also means that if Snap were to use these devices at a theme park, a museum or some public space, demoing them would be a lot easier. I use snap-in lenses for VR headsets all the time. And finally, Snap at least makes inserts that go up to -8, which is basically where my heavy nearsightedness is at.
Displays that are fine, but not always great
Specs have a limited field of view, far more limited than that of VR headsets (or even Xreal’s upcoming Aura pair of tethered AR glasses). It’s better than Snap’s previous Spectacles, and kind of close to or better than my memories of Magic Leap 2. The portrait-style vertical viewing area is tall but not that wide.
It’s wide enough, at least, to watch YouTube video highlights of the Jets-Titans game in a browser, which I call up by saying “Hey Snap” and asking the onboard AI. The display is bright enough indoors, but semi-transparent and ghostly like all AR glasses, and it can get washed out in bright light. That’s why these glasses also have transition lenses that become sunglasses outdoors.
I tried them outside and found the experience usable. But the resolution and the clarity are steps below what sharper micro OLED displays in glasses and headsets can do.
A familiar set of experiences (for me)
I tried Snap’s developer Spectacles a bunch of times in the past, which showed a whole augmented reality OS that works with hand tracking and launches “Lens” apps, some 2D, some full 3D, that can map into the real world. A lot of those were back here, because Specs supports all the independent lenses and creator stuff that Snap’s been building out for a while. But I only got to try a handful of the new Specs during my demo.
I browsed YouTube in a window I called up, resizing the floating window as big as I could and pushing it back and forth with a cursor I’d control with my hands outstretched and pinching to tap. Some apps let you “touch” things directly or pinch them, but most involve aiming your outstretched hand and controlling an air cursor like a Jedi. When windows got far away or small, it became a little hard to aim and tap at times, but looking at your open palm brings up a button dashboard for recentering things, bringing up the home screen, or turning AI on or off. Windows can follow you if you want, as you walk around.
I even wore Specs in a car: One demo put me in an SUV circling the parking lot while I kept watching Jets-Titans plays, testing how the built-in travel mode handled stability in motion. Pretty well, actually. The display stayed pinned and didn’t drift, and I didn’t get carsick. I enjoyed my Jets win while I could.
I tried connected experiences, one of Snap’s more intriguing features, sitting across from a Snap team member wearing a pair. We looked at each other to begin a shared session, and launched a dominoes game that was apparently co-developed with one of Snap’s celebrity ambassadors, NBA star Jimmy Butler.
It was pretty basic: We laid virtual dominoes out one by one on the table in front of us, seeing the shared layout together. We could have just used real dominoes instead. But this idea, with other games I didn’t try – laser tag-like arena battles, virtual table tennis, or even larger group experiences that could support up to 30 people, according to Spiegel – could open some fascinating ideas. I like that these connect simply by looking at someone else wearing them.
I tried a navigation app that hovered a 3D map in front of me, showing me how to walk to Santa Monica College. I shrank the layout down to a smaller directional arrow pointing the way.
And I tried working on Specs, launching Google Docs with a paired Bluetooth keyboard. That definitely wasn’t a great experience, mainly because of keyboard input lag in my session. But could you try to make this a portable work device? Snap is trying to prove it out.
I also tried live translation with a Snap employee who spoke Italian to me while I saw captions in English pop up in front of me – something I’ve done on other smart glasses, from Even Realities G2 to Meta Ray-Ban Displays to Google and Samsung’s upcoming glasses.
The most promising thing to me was kicking around a soccer ball outdoors using an app called Football Trainer. The training aspect wasn’t perfect, and I sometimes worried I’d try to finesse the ball to the virtual target and trip over myself in the process. But, I noticed that the glowing overlays had no occlusion, or a way of letting the ball or my legs look properly positioned over the graphics, something that exists on devices like Apple Vision Pro. Instead, the graphics overlapped the ball and my legs.
But that real-world overlay showed me possibilities. I’ve tried Snap Specs with demos of overlaid virtual pets before, and tried theme park-like augmented reality experiences with other companies using hardware like the Quest 3 (DreamPark’s Mario Kart-like driving experience, where I rode a motorized unicorn along a virtual road). Snap Specs could do things like this if location-based experiences were built to support it. I saw an AR play in New York that used an older Magic Leap 2 to project virtual actors in front of a real audience. Snap Specs could do it better.
AI ambitions with an agent called Snap Intelligence, too
Snap’s also betting hard on AI, to no one’s surprise. A newly announced agentic app called Snap Intelligence promises to work with Macs and iPhones and on Snap Specs glasses, tapping into your connected apps and information and providing recommendations and actions to help run your life.
The idea sounds like Meta’s recently launched AI agent Muse, which is likely to extend to Meta’s glasses hardware more expansively at next week’s Meta Connect conference. Snap’s getting ahead of that likely news here, but I wasn’t able to demo Snap Intelligence at all during my Specs briefing in Santa Monica. That’s all coming later this fall, per Snap, to time with the glasses launch.
“It actually installs on your Mac, so it can use your Mac for you,” Spiegel says to me about Snap Intelligence. “It can also log into services for you without accessing your credentials, because it can just use your keychain.” But he also promises the agent won’t act without your knowledge: “It just doesn’t do anything until you say it’s OK. It won’t change anything or send anything, or anything like that, until you say OK.”
Spiegel is bullish on lots of AI tools working in Specs beyond the company’s own Intelligence agent, from AI-infused AR experiences to running platforms like ChatGPT or agentic systems on the glasses. Snap’s also pushing promises of business partnerships for Specs with Amazon and Nvidia, among others, exploring how the chunky glasses could be worn in either offices or factory floors. Would they be safe and reliable enough to use? It’s unknown yet. But there aren’t any other current AR glasses with full world-tracking 3D and hand tracking in the size and form of Specs yet. Spiegel sees it as “get that big screen experience if you want it, but ultimately, hopefully, get the computer out of the way and just supervise agents who are getting things done on your behalf.”
Privacy plans: Maybe being big and obvious is better?
Snap isn’t trying to hide what Specs are, something Spiegel seemed to celebrate as we chatted. Unlike Meta’s glasses, which can seem so normal that you wouldn’t notice the cameras in the corners, Specs are very much not standard-looking eyewear. Aas Spiegel said to me, you’d know if someone was wearing them. Will that help address growing privacy concerns on smart glasses that are increasing every month?
Recording photos and video clips, including captures that show your augmented-reality experiences overlaid, is done by pressing a side button on the glasses, which then lights up a big LED light. There’s also a clever new snapping gesture for taking photos, which is appropriately on-brand…and again a clearly visible gesture, although one people might not know was a photo-taking trigger.
Spiegel reflected on Snap’s own history of camera glasses, which extends way further back than Meta’s. He tried to make Snap’s Spectacles glasses look unique, even strange, to celebrate their distinct style…and also broadcast that they were different. It made me wonder if smart glasses should all be doing this, in the spirit of the at-the-time odd-looking, long-stemmed first-gen AirPods.
Too ambitious, maybe, but these would be perfect for demos
Snap’s going for it all with these Specs, but I have concerns for sure. Battery life (promised as “up to four hours”), which I didn’t get to truly test during quick demos where I swapped pairs of Specs for different experiences. A not-perfect display. 3D experiences that sometimes looked good, sometimes a step below what a higher-powered VR headset could do. And would any of this really replace a phone at all? I doubt it.
But Snap is aiming to host demos for Specs at a new LA flagship store this fall, something I hope will expand to other locations. And installations and experiences. I’d gladly buy a ticket to a fun event where these Specs were used for a collective immersive experience, but I wouldn’t be compelled to buy a pair. Demos and experiences feel like the next step, and these Specs finally feel ready to be worn. Not all the time, but at least for a while.
Specs seem like part of a new wave of odd VR/AR devices pushing beyond their previous categories, trying to be more or to connect to other systems and services: Valve’s Steam Frame, Xreal Aura and Snap Specs. Specs are the closest living relative to the Meta Orion AR prototype glasses I demoed two years ago, but they’re still full of question marks. And yet, they’re also likely to be a place to prototype AR and AI ideas in outdoor spaces.
I want to see what else they can do, but that’ll have to wait until a review later in the fall. In the meantime, I appreciate they’re trying something no other hardware is doing yet…but also, there’s a reason others haven’t succeeded yet. Size and compromises to get to AR glasses are still a big challenge, but Snap’s Specs are the closest I’ve seen to making it work. It’s far from perfect, but it’s a start.
Editors’ note: Scott Stein’s travel costs were covered by Snap. The judgments and opinions of CNET are our own.
Snap is taking another swing at the idea that glasses could become the next computer. The company has officially introduced SPECS, a pair of augmented reality glasses with an AI-native operating system designed to put digital information directly into the wearer’s view. Unlike some AR experiments that need a separate computing puck or tether, SPECS packs its computing hardware directly into the glasses.
And this isn’t just Snapchat with a fancy pair of frames. SPECS can handle navigation, real-time translation, entertainment, fitness, education, shopping, and even productivity. Users can get directions overlaid onto the real world, translate conversations into subtitles across 60 languages, stream content onto a virtual 115-inch display, and wirelessly connect the glasses to a computer for a larger workspace.
The real trick is SPECS Intelligence
The more interesting part of the announcement is SPECS Intelligence, Snap’s new anticipatory AI service. Rather than waiting for a question, it is designed to understand a user’s goals, priorities, relationships and routines from the apps they choose to connect, then surface useful information when it matters. That could mean preparing someone for an upcoming meeting or pulling together flight, hotel and dinner details before a trip.
The AI works across iPhone, Mac and SPECS, meaning the glasses are essentially another window into the same personal context. Ask about a day while wearing SPECS, and the idea is to get a briefing, schedule information, and meeting preparation without digging through half a dozen apps. The iOS preview is available now in the U.S. for users aged 18 and over, while the fuller Mac experience is coming through an invitation-only early-access program.
Privacy is also a major part of the pitch. Snap says SPECS does not use facial recognition, does not continuously record the surrounding world, and uses an external LED when photos, video, or audio are being captured. For SPECS Intelligence, connected personal content isn’t used to train Snap’s AI models or serve personalized ads.
SPECS is ambitious, but it isn’t exactly pocket change
Snap is positioning SPECS as a genuine computing platform rather than simply another pair of smart glasses. Developers can build AR experiences through Lens Studio, while enterprise partners including NVIDIA, AWS and Salesforce are exploring hands-free uses in workplaces such as factories, retail floors and field operations.
The glasses are available to preorder for $2,195, with a $200 refundable deposit, and shipping is expected later this fall in the U.S., U.K. and France. A new cellular charging case, which provides four additional charges and cellular connectivity, brings the U.S. Connected Case bundle to $2,395.
At that price, SPECS is very much an early look at where Snap thinks computing is headed, rather than a mass-market phone replacement. But the interesting bit is the combination of AR plus anticipatory AI. If SPECS Intelligence can genuinely figure out what matters before being asked, Snap may have found a more compelling reason to wear a computer on the face than simply taking hands-free selfies
If you see what looks like an ambulance with a set of flashing lights weaving through Tokyo traffic, it might not be rushing to save a human. GMO Internet Group is unleashing what it calls Japan’s first “humanoid ambulance” on the streets of Tokyo this month.
A maintenance van designed to make house calls when humanoid robots break down, the ambulance carries diagnostic equipment, spare parts, repair tools and engineers to wherever a robot has stopped working. If a robot can’t be fixed at the roadside, there’s a spare humanoid on board ready to be subbed in and take over on the spot. Meanwhile, the broken unit is ferried back to GMO’s repair base at its Humanoid Lab in Shibuya, Tokyo.
There’s only one ambulance stationed at its Tokyo headquarters for now, and it won’t be used to respond to any and all stricken robots; the service only covers humanoids deployed by GMO AI & Robotics Trading (also known as GMO AIR), with dispatch decided on a case-by-case basis. That’s the same company behind the humanoid robots working as baggage handlers at a Tokyo airport. It’s currently running a trial of Unitree-built machines at Haneda in a collaboration with JAL Ground Service through 2028.
“When humanoids become more popular, this kind of service will be needed,” said Shota Takizawa, an engineer at GMO AIR, at a press event covered by Euronews. The interior and exterior styling of the ambulance came from designer Yasumichi Morita of Tokyo firm GLAMOROUS, who modeled the look on emergency vehicles from around the world. Its exterior features purple roof-mounted lightbars and the text “Humanoid Ambulance.”
GMO says that the ambulance service was created after damaged humanoids had to be sent away for repair, leading to service disruptions. “Learning from that experience, we came up with the idea of providing replacement humanoids,” GMO AIR’s CEO, Tomohiro Uchida, said. GMO’s parent business is internet infrastructure: domains, hosting, that type of thing. Its robotics unit doesn’t build robots at all, but instead resells hardware from Unitree and other manufacturers, then sells services around the machines.
Japan’s government published an AI Robotics Strategy back in April that targets more than 30 percent of the global robot market and a roughly $135 billion domestic industry by 2040. One flashy van isn’t going to have much of an impact on those numbers, but GMO is clearly pitching the ambulance as a solution built ahead of need; a way to bank field-repair experience before humanoids are common enough to demand it.
GM is overhauling its infotainment interface on the 2027 Chevrolet Silverado and GMC Sierra, giving Apple CarPlay and Android Auto their own window alongside native apps rather than the whole screen, after saying phone mirroring would be removed from its cars. Euro NCAP’s 2026 protocol asks for separate physical controls on five functions, which a redesigned touchscreen does not address.
GM is rebuilding the interface in its trucks and giving Apple CarPlay and Android Auto a window of their own beside its native apps, InsideEVs reported. It arrives on the 2027 Chevrolet Silverado and GMC Sierra.
The layout is split by default, so more than one app is visible at once. The blue background and pale icons give way to jet black and higher contrast, the quick access menu becomes customisable, and the gauge cluster has been reworked.
“What we heard was that people were hunting for controls, navigating through too many steps,” said Sebastian Bauer, GM’s executive director of human interface. The team went after fewer taps and clearer visual, auditory and haptic feedback.
The phone half is a reversal. GM stripped mirroring out of its electric cars and Mary Barra said petrol models would follow as models were refreshed, calling the switch between systems clunky. Cars already sold keep it.
Now projection holds the right-hand side rather than swallowing the display. Customers want that simplicity while keeping quick access to the car, said Michael Wahlstrom, who runs software product management.
Europe changed the test in January.
Euro NCAP’s 2026 protocol asks for separate physical controls on five functions, being indicators, hazard lights, the horn, the wipers and the emergency call. A car without them cannot take five stars. The point is to cut the time a driver spends looking away from the road.
Fewer taps is not the same thing. Those five have to leave the screen, not move up a menu.
China went considerably further. Its new standard names 19 functions needing physical controls, each at least 10mm square and working when the power does not, from July 2027.
GM’s announcement mentions physical buttons for the headlights, which survive in the quick menu. Headlights are not among the five.
The trucks are not sold in Europe. Cadillac is.
GM sells the Optiq, Lyriq and Vistiq across the continent, the Optiq from about EUR 65,000, and has been opening the brand in the UK. It says the new interface will spread across its lineup over time.
Whether that includes the European cars, and whether phone projection travels with it, GM would not say. Rivian went the other way and still offers no projection at all.
NewsBeat2 minutes ago
Robert Kraft says he and Ed Sheeran have both donated $2m to Palestinian relief groups amid Macklemore backlash
Business3 minutes ago
EXI: Industrials Dashboard For September
Crypto World4 minutes ago
US Lawmakers Move Bill to Codify a Trump-Era Bitcoin Reserve
-
Fashion5 days agoWeekend Open Thread – Corporette.com
-
Business7 days agoMicron Stock Climbs Above $1,031 as AI Memory Crunch and a $50 Billion Outlook Fuel the Rally
-
Tech4 days agoThe Latest Weird Thing to Play Doom Is the Mapped-Out Brain of a Fruit Fly
-
Business6 days ago10 Most-Streamed Songs On Spotify In 2026 So Far, Led By Ella Langley’s Dominant Run On The Charts This Year
-
Crypto World6 days agoXAG/USD: Silver’s Short-Term Rally Meets Its Moment of Truth
-
Crypto World6 days ago2 Chip Stocks Broke Out This Week. Neither Was Nvidia
-
Tech6 days agoBattery life is the only iPhone 18 Pro and iPhone Duo upgrade I care about. Apple didn’t disappoint
-
Crypto World6 days agoOKX launches 10x OpenAI, Anthropic X-Perps in Europe
-
Crypto World6 days agoDiesel Tops $6 a Gallon for the First Time as 28 States Set Records
-
Business6 days agoWestern Digital Slips 2.7% as AI Storage Rally Cools After Record Cash and Guidance
-
News Videos5 days agoFacing Financial Fears
-
Crypto World3 days agoElon Musk Drops a Bombshell: Grok 5 Could Be the AGI Breakthrough
-
Business4 days agoRivals Sam Altman and Elon Musk Rally Behind Dario Amodei’s Call for a Slowdown in AI Development
-
Business7 days agoFive Leading AI Experts Warn Superintelligence Could Kill Humans and Explain Their Case
-
Crypto World7 days agoEthereum price breakout hinges on a close above $2,535
-
Crypto World6 days ago
Ethereum Price Analysis: Consolidation at $2.5K Tests Momentum as On-Chain Activity Surges
-
Crypto World7 days agoAnt International joins Visa, Mastercard to build AI agent payment standards
-
NewsBeat2 days ago‘Sick conspiracy’: Trump says only guardrails AI needs is ‘a strong and smart (High IQ!) president’ in all-caps rant
-
Crypto World5 days agoCan AI Build a Startup in 72 Hours? Elon Musk's Team Will Livestream the Test
-
Entertainment5 days agoNews Specials, Movies, Shows, More
You must be logged in to post a comment Login