So, you’ve made an embarrassing, unsavory, or otherwise unwanted search in Google and want to get rid of it, so you delete it from your browsing history, thinking that is the end. However, there is one more hoop to jump through to deleting that data once and for all from your account, and we’ll show you how.
Over the last few years, Sony has made it quite clear it wants its PlayStation games to pull double duty as TV/movie successes. So far, that’s given us the Tom Holland-led Uncharted from 2022, and the recent HBO adaptation of The Last of Us. But there’s plenty more adaptations on the horizon (Zero Dawn), from recent…
Slashdot reader wiredmikey writes: Mira Murati, CTO of ChatGPT creator OpenAI, says artificial general intelligence (AGI) systems should be “absolutely” be regulated. In a recent interview, Murati said the company is constantly talking with governments and regulators and other organizations to agree on some level of standards. “We’ve done some work on that in the past couple of years with large language model developers in aligning on some basic safety standards for deployment of these models,” Murati said. “But I think a lot more needs to happen. Government regulators should certainly be very involved.”
Murati specifically discussed OpenAI’s approach to AGI with “human-level capability.”
OpenAI’s specific vision around it is to build it safely and figure out how to build it in a way that’s aligned with human intentions, so that the AI systems are doing the things that we want them to do, and that it maximally benefits as many people out there as possible, ideally everyone.
Q: Is there a path between products like GPT-4 and AGI?
A: We’re far from the point of having a safe, reliable, aligned AGI system. Our path to getting there has a couple of important vectors. From a research standpoint, we’re trying to build systems that have a robust understanding of the world similarly to how we do as humans. Systems like GPT-3 initially were trained only on text data, but our world is not only made of text, so we have images as well and then we started introducing other modalities.
The other angle has been scaling these systems to increase their generality. With GPT-4, we’re dealing with a much more capable system, specifically from the angle of reasoning about things. This capability is key. If the model is smart enough to understand an ambiguous direction or a high-level direction, then you can figure out how to make it follow this direction. But if it doesn’t even understand that high-level goal or high-level direction, it’s much harder to align it. It’s not enough to build this technology in a vacuum in a lab. We really need this contact with reality, with the real world, to see where are the weaknesses, where are the breakage points, and try to do so in a way that’s controlled and low risk and get as much feedback as possible.
Q: What safety measures do you take?
A: We think about interventions at each stage. We redact certain data from the initial training on the model. With DALL-E, we wanted to reduce harmful bias issues we were seeing… In the model training, with ChatGPT in particular, we did reinforcement learning with human feedback to help the model get more aligned with human preferences. Basically what we’re trying to do is amplify what’s considered good behavior and then de-amplify what’s considered bad behavior.
One final quote from the interview: “Designing safety mechanisms in complex systems is hard… The safety mechanisms and coordination mechanisms in these AI systems and any complex technological system [are] difficult and require a lot of thought, exploration and coordination among players.”
Apple is reportedly working on its most significant software overhaul to watchOS in recent memory. According to Bloomberg’s Mark Gurman, the company is redesigning the Apple Watch’s user interface to make widgets a “central part” of how you will interact with the wearable. In describing the new UI, Gurman says it brings back elements of the Glances system that was part of the original watchOS while borrowing the “style” of widgets Apple introduced alongside iOS 14 last year.
He adds the new interface will be “reminiscent” of the Siri watch face that the company introduced with watchOS 4 in 2017 but will function as an overlay for whatever watch face you wish to use. “It’s also similar to widget stacks,” Gurman adds, referencing the iOS feature that allows you to scroll through widgets you’ve placed on top of one another.
Simultaneously, Apple is reportedly testing a tweak to the Apple Watch’s physical buttons. With the interface redesign, pressing down on the digital crown could launch the operating system’s new widgets view instead of taking you to the home screen like the dial currently does with watchOS 9.
With the likelihood that the redesign will be jarring for some, Gurman speculates Apple plans to make the new interface optional at first. Additionally, he suggests the overhaul is an admission that an iPhone-like app experience “doesn’t always make sense on a watch – a place where you want as much information as possible with the least amount of poking around.” With WWDC 2023 a little more than a month away, it won’t be long before Apple shares more information about what Watch users can expect from its wearable’s next big software update.
This article originally appeared on Engadget at https://www.engadget.com/apple-is-reportedly-redesigning-watchos-around-widgets-162720331.html?src=rss
Shots fired! In a Twitter post, AMD’s Senior Director of Marketing for Gaming, Sasa Marinkovic, continued a salvo against NVIDIA over the bang-for-your-buck proposition of each company’s current and previous-generation graphics cards, and specifically in regards to how much video memory (otherwise known as VRAM) buyers get for their money.
NASA’s Ingenuity helicopter captured a breathtaking image of the Mars landscape, and had a fellow Martian spacecraft photobomb the picture. During Ingenuity’s 51st flight over the Belva crater, NASA’s Perseverance rover can be spotted way off in the distance.
The Ingenuity Mars Helicopter completed its historic 50th flight on the Red Planet
This month the Linux Foundation announced version 3.0 of DentOS, an open source network operating system using the Linux kernel, Switchdev, and other Linux-based projects for a standardized network operating system “without abstractions or overhead,” according to the project’s web page. “All underlying infrastructure — including ASIC and Silicon for networking and datapath — is treated equally; while existing abstractions, APIs, drivers, low-level overhead, and other open software are simplified. DENT unites silicon vendors, ODMs, SIs, OEMs, and end users across all verticals to enable the transition to disaggregated networks.”
Or, as the Linux Foundation, the operating system provides “a flexible and customizable platform for network administrators to manage their networks.”
DENT provides access to open source-based switches at a lower cost and with more flexibility compared to proprietary switches with locked ecosystems. Network wiring closets in many facilities–including retail stores, warehousing, remote locations, enterprises, and small and mid-sized businesses–are often small, requiring a compact solution for network management. Additionally, staff expertise may be limited, and branch-office switches from leading suppliers can require costly contracts. DENT can be easily deployed on white-box hardware in small spaces, providing an efficient and cost-effective solution for network management. As a result, DENT deployment can significantly enhance network management in a wide range of environments, providing greater efficiency, reliability, and scalability…
DentOS enables Amazon’s Just Walk Out Technology to connect and manage thousands of devices like cameras, sensors, entry and exit gates, and access points on the network edge.
If you’ve been patiently waiting for a sale on the 5th-generation Echo Dot, now is the time to buy one. A handful of models are on sale. To start, you can get the basic model for $30 or 40 percent off. That’s only $5 more than the Echo Dot’s all-time low price. Moreover, all three colorways – charcoal, deep sea blue and glacier white – are part of the sale. Alternatively, the Echo Dot with Clock is also on sale. Currently, it’s $40 or 33 percent off its usual $60 price. As before, the sale includes all the available colorways.
The Echo Dot is one of the least expensive smart speakers out there. That said, you get a lot of value for your money. It offers surprisingly good sound quality while being small enough to fit almost anywhere. The Echo Dot also comes with a 3.5mm audio output, allowing you to connect it to an amp or set of headphones. Of course, it also provides access to Alexa and all the smart home integration you could want. For those reasons, the Echo Dot is one of Engadget’s favorite smart speakers. The Echo Dot with Clock has all the features found on its less expensive sibling but can also show the time, weather and timers. That capability makes it an excellent fit for the kitchen or anywhere you want to keep tabs on the time.
Mattresses don’t come cheap, so it’s in your best interest to take care of them. In addition to covering them with a mattress protector and removing dirt and dust by vacuuming them regularly, it also means quickly cleaning up spills before they turn into stains. Here’s how to do that.
Last weekend around 2 a.m. Sunday, “Facebook’s advertising system went haywire,” reports Gizmodo, “overcharging customers and wasting money on ads that didn’t work.”
Reports suggest Meta, the social network’s parent company, charged some advertisers more than double what they agreed to pay, ranging from hundreds to hundreds of thousands of dollars. Meta briefly stopped showing ads on part of its network with practically zero communication to its millions of customers.
The company confirmed the bug happened and promised to follow its “normal refund process,” but shared very little about what went wrong.
A Meta spokesperson described it as “a technical issue that has now been resolved” (adding that the glitch also appeared to a lesser extent on Instagram).
But Alex Golick, the CEO of marketing agency Intensify told CNBC it was the worst Facebook glitch he’d seen in the decade he’s worked in digital advertising — with one client burning through 90% of its ad budget by 9 a.m. And his entire customer base had similar problems:
Golick said that all those advertisers had essentially just wasted most of their money for the day, spending roughly triple the amount they normally would to acquire a customer. “The results were horrendous,” Golick told CNBC…
For brands that are already lowering ad costs to manage through a sluggish economy and a mobile ad market that no longer allows for targeting based on user data, Facebook’s miscue is more than just an unfortunate blip. In low-margin industries, where every dollar counts, it can turn a profitable weekend into a big loser, while also raising further questions about the reliability of Facebook’s ad systems…
Data analytics and marketing firm Varos provided data showing that, of the more than 3,000 ecommerce and direct-to-consumer companies that use its technology, the software bug caused a majority of them to experience a rise in cost per thousand impressions, or what those in the industry call CPMs. About 36% of companies were “very significantly impacted” by the bug, meaning their CPMs at least doubled, Varos said…
Varos CEO Yarden Shaked the glitch resulted in a “bidding war for nothing.” Data about the glitch provided by the advertising technology firm Proxima on 108 companies also revealed that these firms spent their “entire day’s budget in the first few hours of the day,” the company said…
Assume that Meta, Google, Microsoft, and other big tech companies soon have their way, and neural interface devices replace keyboards and mice. In that likely future, a large segment of the population will routinely wear neural devices like NextSense’s bio-sensing EEG earbuds, which are designed to be worn twenty-four hours a day. With wide-scale adoption of wearable neurotechnology, adding our brain activity to nationwide identification systems is a near-term reality.
One of the most extraordinary discoveries of modern neuroscience is the uniqueness of each person’s functional brain connection (its physical wiring), especially in the brain areas devoted to thinking or remembering something. Because of this, algorithms can be used to analyze our brain activity and extract features that are both unique to each person and stable over time. How your brain responds to a song or an image, for example, is highly dependent upon your prior experiences. The unique brain patterns you generate could be used to authenticate your identity.
Nationwide identification systems vary by country but generally involve the assignment of unique identification numbers, which can be used for border checks, employment screenings, health-care delivery, or to interact with security systems. These ID numbers are stored in centralized government databases along with other significant personal data, including birth date and place, height, weight, eye color, address, and other information. Most identification systems have long included at least one piece of biometric data, the static photo used in passports and driver’s licenses. But governments are quickly moving toward more expansive biometric features that include the brain.
Biometric characteristics are special because they are highly distinctive and have little to no overlap between individuals. As the artificial intelligence algorithms powering biometric systems have become more powerful, they can identify unique features in the eyes and the face, or even in a person’s behavior. Brain-based biometric authentication has security advantages over other biometric data because it is concealed, dynamic, non-stationary, and incredibly complex.
The promise of greater security has led countries to invest heavily in biometric authentication. China has an extensive nationwide biometric database that includes DNA samples, and it also makes widespread use of facial recognition technology. Chinese authorities in the Xinjiang Uyghur Autonomous Region have conducted mass collections of biometric data from the Uyghur people and used it for targeted discrimination.
The United States has also massively expanded its collection of biometric data. A recent report by the US Government Accountability Office detailed at least eighteen different federal agencies that have some kind of facial recognition program in place. US Customs and Border Protection includes facial recognition as part of its pre-boarding screening process, and an executive order signed by President Trump in 2017 required the United States’ top twenty airports to implement biometric screening on incoming international passengers.
Increasingly, governments are investing in developing brain biometric measurements. The US Department of Defense recently funded SPARK Neuro, a New York–based company that has been working on a biometric system that combines EEG brain wave data, changes in sweat gland activity, facial recognition, eye-tracking, and even functional near-infrared spectrometry brain imaging (fNIRS), a particularly promising (if expensive) technology for brain authentication, since it is wearable, can be used to monitor individuals over time, can be used indoors or outdoors while a person is moving or at rest, and can be used on infants and children. China has been funneling substantial investments into EEG and fNIRS as well.
For biometric features to be successfully used for authentication, they must have universality, permanence, uniqueness, and be secure against fraud. Over time, static biometrics like facial IDs and fingerprints have become prone to spoofing. Functional biometrics, such as brain activity, are less prone to attack. That feature has motivated researchers like Jinani Sooriyaarachchi and her colleagues in Australia to develop scalable brain-based authentication systems. In one of their most recent studies, they recruited twenty volunteers and asked them to listen to both a popular English song and their own favorite song while their brain wave activity was recorded with a four-channel (an electrode capturing brain wave activity is called a channel) Muse headset. Afterward, the researchers analyzed their recorded brain wave activity using an artificial-intelligence classifier algorithm. Remarkably, they achieved 98.39 percent accuracy in identifying the correct participant when they listened to the familiar song, and a 99.46 percent accuracy when they listened to their favorite song. Using an eight-channel EEG headset on thirty research subjects, another group achieved a similar 98 percent accuracy in authenticating participants by their brain wave data after they’d looked at novel images. It might not even take eight or even four electrodes to achieve the same result. Even with just a single-channel EEG headset, researchers have achieved 99 percent accuracy in distinguishing between participants when they performed the same mental tasks. Most of these studies had a small number of participants; it is not yet clear if neural signatures will be as accurate at scale, when billions rather than dozens of people must be authenticated. EEG is inherently noisy—meaning the signals the electrodes pick up can come from eye-blinking or other movement, which can make it hard to tell the difference between brain activity or interference. But researchers have made substantial progress in developing pattern classifiers that filter noise, allowing them to discriminate between individuals based on their resting-state EEG brain wave activity and when performing tasks. As noted previously, EEG devices have been used to recover sensitive information from a person’s brain, such as their PIN codes, and their political and religious ideologies. Obviously, this poses clear risks to our digital and physical security.
Governments can already tap our phone conversations and snoop on us digitally. Will they similarly tap our brain activity data without our knowledge or consent? Will they deploy AI programs to search our brains for terrorist plots? Will they gather neural data to make inferences about individuals’ political beliefs to predict and prevent peaceful protests? China is reportedly already doing so.
This article originally appeared on Engadget at https://www.engadget.com/hitting-the-books-the-battle-for-your-brain-nita-farahany-st-martins-press-143000718.html?src=rss
The Debian Project announced today the release and general availability of Debian 11.7 as the seventh ISO update to the latest Debian GNU/Linux 11 “Bullseye” operating system series.
Social media can be mentally draining. And when mentally drained, you are more likely to be influenced by a high number of likes on posts – even to the point of clicking on ads for products you don’t need or want – according to our recent experiments on how social media affects behavior.
Back in 2021 AMD began preparing Linux kernel support for 5-level paging support with their future processors and building off the prior 5-level page table kernel support established by Intel. That was followed by AMD enabling 5-level page table support with KVM SVM in the Linux 5.15 kernel. AMD CPUs with 5-level page table support since launched in the form of 4th Gen EPYC “Genoa” processors. One piece only now coming together though is AMD IOMMU driver support for 5-level guest page table support…
A.I. marketing is at a fever pitch, promising that this new wave of generative AI tools, powered by large language models, can help us do everything from navigating legal contracts to saving hundreds on our phone bills. When a leader like Alphabet CEO Sundar Pichai calls AI “more profound than fire or electricity,”…
You’re in the shower shampooing your hair, and suddenly you catch a whiff of something unpleasant. Something that smells like sewage. Your first instinct is that it’s coming from the toilet. That might be true, but it could also be coming from your shower drain. Here are a few examples of what could be behind the…
While prominent gaming peripheral manufacturer Razer still is not officially supporting Linux with their vast array of products, thanks to the community-driven OpenRazer project there is unofficial open-source support and can work quite well when paired with the likes of Polychromatic as a nice user interface. Out today is OpenRazer 3.6 in enabling the latest Razer products on Linux…