This week at Lenovo’s Tech World ‘23, Motorola introduced its Adaptive Smartphone Concept, an Android-powered device that features flexible hardware. Could this be the next step in the foldable hardware evolution?
Building on Motorola’s existing work in the foldable and rollable device space, this concept phone features a 6.9-inch FHD+ pOLED display that can be bent and shaped into varying forms. When used in an upright position, shown below, you have 4.6-inches of usable display running a “more compact form of full Android.”
Motorola even showcases the ability to fully wrap the device around your wrist which then toggles a user experience similar to the one found on the external display of the Razr+ smartphone. If it stays well wrapped, I imagine this could be very nifty when hitting the running trails.
Given the conceptual nature of the device, there’s no guarantee we’ll ever see this hit the market, but dang it, it’s still awesome. Whether you think it practical or not, companies need to show that innovation and creativity is happening. Personally, I think this looks very cool.
The year is 1998. You clutch your Game Boy Color, focusing on your fourth run of the Elite Four in Pokemon Blue, back propped against the couch while your parents watch the season four finale of Frasier on the 32-inch “big-screen” CRT TV in your living room. It is a canon event, crystalized in your memory, as you beat your creature-catching rival while Kelsey Grammer’s eternally befuddled psychiatrist gets romantically crushed by a post-Terminator 2 Linda Hamilton.
With a questionably necessary reboot of Frasier arriving thirty years later, now’s the perfect time for an equally anachronistic Game Boy RPG for the original series. And indie developer Edward La Barbera has delivered it. Frasier Fantasyis a bite-sized RPG following the beloved radio host as he putters around Seattle, lovingly crafted in the style of the Game Boy Color.
The graphics and sound don’t just evoke the GBC, they’re actually made for the hardware — you can download and play it on one if you have the right flash cart. The rest of us can play it on an emulator, or even better, just load the entire game in a browser window (spotted by Rock Paper Shotgun).
In Frasier Fantasy, you’ll make your way across the famous locations of the show, including the geographically impossible Elliot Bay Towers apartment, the Cafe Nervosa, and the KACL radio station. Your goal: To do your radio psychiatry show, rid the apartment of your father and the real star, Eddie, and retrieve your antique grape scissors before hosting a dinner party. The game nails the strange fusion of these cultural artifacts, giving you a whirlwind tour of the show’s cast interspersed with minigames and turn-based RPG battles. Frasier can employ special moves like “Freudian Slap” or simply drone his enemies to sleep with faux-posh psych patter.
The whole game can be cleared in under an hour, and might just be worth it for the Frasier super-fan. Deep cuts, like a battle with Star Trek-obsessed coworker Noel and Niles’ never-seen wife Maris in her sensory deprivation tank, show that this is truly a labor of love. The final boss, featuring a pale, frigid monster that makes Sephiroth look like a pansy, is a particular highlight.
Amazon is splashing out on new vehicle inspectors to watch for damage or wear to its vast fleet of delivery vans—and they’re not human. The retailer is installing camera-studded inspection stations equipped with artificial intelligence-powered technology called AVI, or automated vehicle inspection, at hundreds of its distribution centers worldwide.
When a driver working out of any of the 20 delivery centers currently equipped with the tech returns their vehicle at the end of a shift, they slowly drive it through a sensor-laden archway made by startup UVeye, which has headquarters in the US and Israel.
The technology is made up of three separate high-res camera systems: One scans a vehicle’s undercarriage, another checks tire quality, and another focuses on the vehicle exterior. The data they gather is compiled into a 3D image of the vehicle and used by machine-learning software to identify whether the vehicle is damaged or needs maintenance. The algorithms should pick up every nail in a tire, fluid leak, dent on a fender, or crack in the windshield.
UVeye’s AVI technology scans an Amazon delivery van.Courtesy of UVeye
Aziz Makkiya, Amazon’s senior manager of last-mile products and services, declined to discuss company financials, but said in an interview at an Amazon event last week that the technology shaves about four minutes off what is usually a five-minute inspection process. That could add up to a lot when multiplied over Amazon’s roughly 100,000-strong global fleet. Makkiya said the technology should make the vehicles safer, in part by catching vehicle maintenance issues early. “The safety aspect is what we really care about,” he said.
Amazon says it’s been testing the automated vehicle inspection system for nearly two years, and has now rolled it out to 20 delivery stations in the US, Canada, Germany, and the UK, with the goal of installing hundreds of units in the next few years.
The automated inspections will give Amazon a new window into the operations of the independent companies known as DSPs that it contracts to make deliveries, and which lease Amazon-branded vehicles from the company. Drivers employed by DSPs are usually responsible for inspecting their own vehicles. Amazon pays for maintenance such as tires and brakes, but DSPs have to cover damage from collisions. Maya Vautier, a spokesperson for Amazon, says the inspection technology only scans the outside of vehicles and doesn’t collect data on vehicle performance or utilization.
Got a tip?
Do you work for an Amazon contractor? What do you think of the new inspection technology? Email the author at aarian_marshall@wired.com. WIRED protects the confidentiality of its sources.
Makkiya says the data can be used to inform wider company decisions. If vehicles driven on certain routes or roads show consistent patterns of damage, Amazon might let a city know that trees need cutting or potholes fixing, for example. Amazon also plans to start using automated vehicle scans to guide its vehicle purchases or provide terrain-specific feedback to manufacturers. At some point in the next year, Makkiya says, the company should be able to go to a vehicle manufacturer and say, “Hey, you’ve got a problem with the tires in this area, or the suspension of the vehicle in this area.”
The idea of using AI to inspect vans and other vehicles is not new, but the Amazon deal adds validation to a concept that has become more prominent amidst investor enthusiasm for generative AI. UVeye announced a $100 million funding round in the spring from investors including automaker GM and used vehicle retailer Carmax, which is also using the AI inspection technology. Its previous investors include Hyundai, Volvo, and Toyota.
Tractable, a London-based AI inspection company that has significant partnerships with insurers, recently raised $65 million of new investment, and another company, Monk, was acquired by online used car auction company ACV last year. In general, tech providers offer to speed up the process of evaluating vehicle damage or predicting when maintenance might be needed, a task once left to workers with experience in vehicle management and repair.
William Demaree, who directs fixed operations at the Tom Wood dealership network in Indiana, Kentucky, and Minnesota, likes the 10 UVeye units his company leases for another reason: They demonstrate to customers that they’re not getting fleeced. Every customer who comes in for a repair or trade-in at shops with the technology installed drives their vehicle through the portal, he says. “The automotive industry has a rough name sometimes,” he says.
Drivers might not always trust car dealers, but they seem more comfortable with the new, hulking, machine-learning-powered inspector. Demaree says the technology “shows that we can be more transparent with our customers.” People also like exploring images of their vehicles on the unit’s big screens, he says, and love to take photos of their vehicle’s undercarriages.
Automated inspections aren’t perfect, Demaree says. Workers occasionally have to flag that something labeled as a bump or scratch is just a normal feature of the car. The feedback is collected by UVeye to train its inspection algorithms for future vehicles. Amazon’s partnership with the company promises to provide a new flood of such feedback.
At 9:30 am on a Wednesday in late September, a hacker who asked to be called Tom Smith sent me a nonsensical text message: “query voltage recurrence.”
Those three words were proof of a remarkable feat—and potentially an extremely valuable one. A few days earlier, I had randomly generated those terms, set them as the passphrase on a certain model of encryptedUSB thumb drive known as an IronKey S200, and shipped the drive across the country to Smith and his teammates in the Seattle lab of a startup called Unciphered.
Unciphered’s staff in the company’s Seattle lab.
Photograph: Meron Menghistab
Smith had told me that guessing my passphrase might take several days. Guessing it at all, in fact, should have been impossible: IronKeys are designed to permanently erase their contents if someone tries just 10 incorrect password guesses. But Unciphered’s hackers had developed a secret IronKey password-cracking technique—one that they’ve still declined to fully describe to me or anyone else outside their company—that gave them essentially infinite tries. My USB stick had reached Unciphered’s lab on Tuesday, and I was somewhat surprised to see my three-word passphrase texted back to me the very next morning. With the help of a high-performance computer, Smith told me, the process had taken only 200 trillion tries.
Smith’s demonstration was not merely a hacker party trick. He and Unciphered’s team have spent close to eight months developing a capability to crack this specific, decade-old model of IronKey for a very particular reason: They believe that in a vault in a Swiss bank 5,000 miles to the east of their Seattle lab, an IronKey that’s just as vulnerable to this cracking technique holds the keys to 7,002 bitcoins, worth close to $235 million at current exchange rates.
For years, Unciphered’s hackers and many others in the crypto community have followed the story of a Swiss crypto entrepreneur living in San Francisco named Stefan Thomas, who owns this 2011-era IronKey, and who has lost the password to unlock it and access the nine-figure fortune it contains. Thomas has said in interviews that he’s already tried eight incorrect guesses, leaving only two more tries before the IronKey erases the keys stored on it and he loses access to his bitcoins forever.
Varda Space Industries plans to eventually land its spacecraft in Australia, but the company is still waiting on approval to bring down its already-launched first vehicle in Utah next year, according to a media report.
Varda launched its debut mission on SpaceX‘s Transporter-8 mission in June, and the capsule remains operational in space. Varda will land its future spacecraft at the Koonibba Test Range northwest of Adelaide, Australia under a newly announced agreement with Southern Launch. The first missions to use the site will land as soon as 2024, SpaceNews said in a report.
But the company, which is developing systems that will allow customers to manufacture products (for example, pharmaceuticals) in orbit and bring them down to Earth, is still awaiting approval from the U.S. Federal Aviation Administration (FAA) and the Air Force to land its first spacecraft at the Utah Test and Training Range.
“We got very, very close,” Delian Asparouhov, co-founder of Varda, said in an Oct. 20 interview with SpaceNews. The situation was related to “a coordination problem amongst three different groups that had not worked through this operation before,” he added, referring to the Utah range, the Air Force and the FAA.
California-based Varda is the first company to apply for an FAA reentry license through Part 450, a new set of regulations that were put in place to make the approval process easier. Asparouhov declined to comment on whether his company would have secured approvals faster under older FAA rules.
The spacecraft is fully operational in orbit. Asparouhov said that onboard experiments are finished, and that the FAA and Air Force have expressed no concerns about safety.
Join our Space Forums to keep talking space on the latest missions, night sky and more! And if you have a news tip, correction or comment, let us know at: community@space.com.
While companies like Microsoft and Nvidia are all-in on the power of next-generation machine learning algorithms, some regulators are dreading what it might mean for our already-stressed communication networks. The chairwoman of the US Federal Communications Commission, for one, who’s just proposed an investigation into what “AI” could mean for even more spam calls and texts. The FCC will vote to adopt a multi-tiered action in November.
Chairwoman Rosencworcel, who’s served on the Commission since 2012 and as its executive since being confirmed late in 2021, is particularly concerned with how newly empowered AI tools could affect senior citizens. The FCC’s initial press release (PDF link) lists four main goals: determining whether AI technologies fall under the Comission’s jurisdiction via the Telephone Consumer Protection Act of 1991, if and when future AI tech might do the same, how AI impacts existing regulatory frameworks, and if the FCC should consider ways to verify the authenticity of auto-generated AI voice and text from “trusted sources.”
That last bullet point would seem to contain the potential for the most problems. Auto-generated text and natural-sounding voice algorithms are already fairly easy tools to use, albeit not quite as fast as necessary for real-time back-and-forth in a phone call setting. Combine it with some “big iron” data centers, whether wholly created for the purpose of mass calls and texts or merely rented from the likes of Amazon and Microsoft, and you have a recipe for disaster.
Replacing human-staffed call centers around the world in scammer hotbeds like India and Cambodia with fully automated AI systems could exponentially increase both the volume and the efficacy of scams, which are already being sent hundreds of billions of times every year. While filters and blocks exist, it’s estimated that billions of dollars are lost to scams each year in the US alone, many of which target senior citizens specifically.
The FCC’s brief does mention that AI technology could also be used to fight against spammers and scams, presumably with some kind of real-time scanning system alerting users that they’re talking to a computer. But the details of this, and the potential evolution of the threat posed by AI tools, will have to wait for the Commission’s November 15th session.
A new tool lets artists add invisible changes to the pixels in their art before they upload it online so that if it’s scraped into an AI training set, it can cause the resulting model to break in chaotic and unpredictable ways.
The tool, called Nightshade, is intended as a way to fight back against AI companies that use artists’ work to train their models without the creator’s permission. Using it to “poison” this training data could damage future iterations of image-generating AI models, such as DALL-E, Midjourney, and Stable Diffusion, by rendering some of their outputs useless—dogs become cats, cars become cows, and so forth. MIT Technology Review got an exclusive preview of the research, which has been submitted for peer review at computer security conference Usenix.
AI companies such as OpenAI, Meta, Google, and Stability AI are facing a slew of lawsuits from artists who claim that their copyrighted material and personal information was scraped without consent or compensation. Ben Zhao, a professor at the University of Chicago, who led the team that created Nightshade, says the hope is that it will help tip the power balance back from AI companies towards artists, by creating a powerful deterrent against disrespecting artists’ copyright and intellectual property. Meta, Google, Stability AI, and OpenAI did not respond to MIT Technology Review’s request for comment on how they might respond.
Zhao’s team also developed Glaze, a tool that allows artists to “mask” their own personal style to prevent it from being scraped by AI companies. It works in a similar way to Nightshade: by changing the pixels of images in subtle ways that are invisible to the human eye but manipulate machine-learning models to interpret the image as something different from what it actually shows.
The team intends to integrate Nightshade into Glaze, and artists can choose whether they want to use the data-poisoning tool or not. The team is also making Nightshade open source, which would allow others to tinker with it and make their own versions. The more people use it and make their own versions of it, the more powerful the tool becomes, Zhao says. The data sets for large AI models can consist of billions of images, so the more poisoned images can be scraped into the model, the more damage the technique will cause.
A targeted attack
Nightshade exploits a security vulnerability in generative AI models, one arising from the fact that they are trained on vast amounts of data—in this case, images that have been hoovered from the internet. Nightshade messes with those images.
Artists who want to upload their work online but don’t want their images to be scraped by AI companies can upload them to Glaze and choose to mask it with an art style different from theirs. They can then also opt to use Nightshade. Once AI developers scrape the internet to get more data to tweak an existing AI model or build a new one, these poisoned samples make their way into the model’s data set and cause it to malfunction.
Poisoned data samples can manipulate models into learning, for example, that images of hats are cakes, and images of handbags are toasters. The poisoned data is very difficult to remove, as it requires tech companies to painstakingly find and delete each corrupted sample.
The researchers tested the attack on Stable Diffusion’s latest models and on an AI model they trained themselves from scratch. When they fed Stable Diffusion just 50 poisoned images of dogs and then prompted it to create images of dogs itself, the output started looking weird—creatures with too many limbs and cartoonish faces. With 300 poisoned samples, an attacker can manipulate Stable Diffusion to generate images of dogs to look like cats.
COURTESY OF THE RESEARCHERS
Generative AI models are excellent at making connections between words, which helps the poison spread. Nightshade infects not only the word “dog” but all similar concepts, such as “puppy,” “husky,” and “wolf.” The poison attack also works on tangentially related images. For example, if the model scraped a poisoned image for the prompt “fantasy art,” the prompts “dragon” and “a castle in The Lord of the Rings” would similarly be manipulated into something else.
COURTESY OF THE RESEARCHERS
Zhao admits there is a risk that people might abuse the data poisoning technique for malicious uses. However, he says attackers would need thousands of poisoned samples to inflict real damage on larger, more powerful models, as they are trained on billions of data samples.
“We don’t yet know of robust defenses against these attacks. We haven’t yet seen poisoning attacks on modern [machine learning] models in the wild, but it could be just a matter of time,” says Vitaly Shmatikov, a professor at Cornell University who studies AI model security and was not involved in the research. “The time to work on defenses is now,” Shmatikov adds.
Gautam Kamath, an assistant professor at the University of Waterloo who researches data privacy and robustness in AI models and wasn’t involved in the study, says the work is “fantastic.”
The research shows that vulnerabilities “don’t magically go away for these new models, and in fact only become more serious,” Kamath says. “This is especially true as these models become more powerful and people place more trust in them, since the stakes only rise over time.”
A powerful deterrent
Junfeng Yang, a computer science professor at Columbia University, who has studied the security of deep-learning systems and wasn’t involved in the work, says Nightshade could have a big impact if it makes AI companies respect artists’ rights more—for example, by being more willing to pay out royalties.
AI companies that have developed generative text-to-image models, such as Stability AI and OpenAI, have offered to let artists opt out of having their images used to train future versions of the models. But artists say this is not enough. Eva Toorenent, an illustrator and artist who has used Glaze, says opt-out policies require artists to jump through hoops and still leave tech companies with all the power.
Toorenent hopes Nightshade will change the status quo.
“It is going to make [AI companies] think twice, because they have the possibility of destroying their entire model by taking our work without our consent,” she says.
Autumn Beverly, another artist, says tools like Nightshade and Glaze have given her the confidence to post her work online again. She previously removed it from the internet after discovering it had been scraped without her consent into the popular LAION image database.
“I’m just really grateful that we have a tool that can help return the power back to the artists for their own work,” she says.
via Technology Review Feed – Tech Review Top Stories https://ift.tt/F16CbYD
