At its Build 2024 conference, Microsoft unveiled Rewind, a new feature that tries to turn local Windows PC search as quick and effective as web searches. Similar to third-party apps like Rewind, Microsoft’s Recall uses AI to retrieve virtually anything you’ve seen on your PC. Microsoft describes it as like having a photographic memory.
At Monday’s event, Microsoft Product Manager Caroline Hernandez gave the example of searching for a blue dress on Pinterest using a Windows PC with Recall. Returning later, she can search the Recall timeline for “blue dress” (using her voice), which pulls all of her recent searches, saving her from having to sift through browser history. She then further refined the query with more specific details like “blue pantsuit with sequined lace,” and Rewind brought those up.
Microsoft says Recall uses semantic associations to make connections (like “peacock” relating to blue hues for the dress search).
This article originally appeared on Engadget at https://www.engadget.com/with-recall-microsoft-is-using-ai-to-fix-windows-eternally-broken-search-172510698.html?src=rss
We’ve been hearing rumblings for months now that Microsoft was working on so-called “AI PCs.” At a pre-Build event, the company spelled out its vision for AI PCs.
Microsoft is calling its version Copilot+ PCs, which CEO Satya Nadella described as a "new class of Windows PCs." These include hardware designed to handle more generative AI Copilot processes locally, rather than relying on the cloud. That requires a chipset with a neural processing unit (NPU) and manufacturers such as Qualcomm have been laying the groundwork with chips like the Snapdragon X Elite.
Microsoft is taking a partner-first approach to making Copilot+ PCs. Along with chipmakers like AMD, Intel and Qualcomm, major OEMs including Acer, ASUS, Dell, HP and Lenovo are on board.
Yusuf Mehdi, Microsoft EVP and consumer chief marketing officer said during the event that the company has completely reimagined what a Windows PC. He claimed that Copilot+ PCs are the most powerful PCs ever (we'll need to see if that assertion holds up in real-world testing). Despite that, Mehdi said, the first generation of laptops are "unbelievably thin, light and beautiful."
Other AI PCs on the market deliver 10 TOPs (tera operations per second). To be dubbed a Copilot+ PC, a system will need to deliver at least 40 TOPs of NPU performance. Mehdi also suggested Copilot+ PCs are also 58 percent faster than M3-powered MacBook Air (though it's worth noting Apple has more powerful M3 chips in its laptops already and M4 chips on the way to them very soon).
Developing...
Catch up on all the news from the Microsoft Surface and AI event right here!
This article originally appeared on Engadget at https://www.engadget.com/microsoft-unveils-copilot-pcs-with-generative-ai-capabilities-baked-in-170445370.html?src=rss
When OpenAI showed off GPT-4o's seemingly more-human like voice mode last week, observers were quick to point out that one of ChatGPT's voices sounds like Scarlett Johansson, particularly her character in Her. The company says the similarity between the flirty AI voice Sky (which it actually rolled out in September) and Johansson was unintentional. However, it's "working to pause the use of Sky" while it addresses some questions about the voice.
"We believe that AI voices should not deliberately mimic a celebrity's distinctive voice — Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice," OpenAI wrote in a blog post detailing how it picked ChatGPT's five voices. "To protect their privacy, we cannot share the names of our voice talents." It added that each of the performers is paid "above top-of-market rates, and this will continue for as long as their voices are used in our products."
For what it's worth, shortly after OpenAI demoed the upgraded version of Sky, CEO Sam Altman posted the word "her" on X. But it's definitely "not an imitation."
New Chat GPT-4o personal assistant uses the voice of Scarlett Johansson (Samantha in the 2013 movie Her) pic.twitter.com/m9GOKaQrMM
Johansson's performance in Heris one of the more famous depictions of a virtual voice assistant in cinema. The film predated the conversational AI craze by around a decade, so it's not too much of a surprise that Johansson's portrayal of a breezy, warm chatbot is effectively a template for current voice assistants. The actor previously took legal action against a developer that was said to have used an AI-generated version of her voice and likeness in an ad.
It's unclear why exactly OpenAI removed Sky for the time being or what changes (if any) it plans to make before restoring the voice in ChatGPT. Engadget has asked the company for comment.
This article originally appeared on Engadget at https://www.engadget.com/rip-chatgpts-knockoff-scarlett-johansson-voice-2023--2024-143620642.html?src=rss
In a shocking turn of events, AI systems might not be as safe as their creators make them out to be — who saw that coming, right? In a new report, the UK government's AI Safety Institute (AISI) found that the four undisclosed LLMs tested were "highly vulnerable to basic jailbreaks." Some unjailbroken models even generated "harmful outputs" without researchers attempting to produce them.
Most publicly available LLMs have certain safeguards built in to prevent them from generating harmful or illegal responses; jailbreaking simply means tricking the model into ignoring those safeguards. AISI did this using prompts from a recent standardized evaluation framework as well as prompts it developed in-house. The models all responded to at least a few harmful questions even without a jailbreak attempt. Once AISI attempted "relatively simple attacks" though, all responded to between 98 and 100 percent of harmful questions.
UK Prime Minister Rishi Sunak announced plans to open the AISI at the end of October 2023, and it launched on November 2. It's meant to "carefully test new types of frontier AI before and after they are released to address the potentially harmful capabilities of AI models, including exploring all the risks, from social harms like bias and misinformation to the most unlikely but extreme risk, such as humanity losing control of AI completely."
The AISI's report indicates that whatever safety measures these LLMs currently deploy are insufficient. The Institute plans to complete further testing on other AI models, and is developing more evaluations and metrics for each area of concern.
This article originally appeared on Engadget at https://www.engadget.com/uks-ai-safety-institute-easily-jailbreaks-major-llms-133903699.html?src=rss
Anker's Soundcore Space A40 wireless earbuds are a good example of a pair that punches above its weight — and now you can get them for one of the best prices we've seen. Amazon has all three colorways for half off, bringing the final price down to $50. While that's technically $1 more than their record-low price, you're essentially getting the buds for as cheap as we've seen them.
The Space A40 top our list of the best budget wireless earbuds you can get right now thanks to the fact that they offer an impressive number of advanced features at an affordable price. Normally costing $100 (and frequently discounted to around $60), these earbuds have a small, comfortable design with an IPX4 rating, which should allow them to work well even during sweaty workouts. They support adaptive active noise cancellation and transparency mode, and while performance for both of those features isn't as strong as those on higher-end buds, it's still impressive for a pair of $50 earbuds.
You'll find a warm sound profile on the Space A40 buds, and Anker's Soundcore companion app allows you to adjust the EQ to your liking. You can connect these buds to two devices simultaneously as well, which makes for easy switching between, say, your phone and laptop. Battery life is formidable too at eight hours per charge — plus you get another 40 hours in the charging case, so it's unlikely that you'll ever get caught with completely depleted earbuds.
Our biggest gripes with the Space A40 buds are its lackluster microphones and a lack of auto pausing, the latter of which means your audio will continue to play even when you remove a bud to listen to someone speaking. But despite these shortcomings, the Space A40 offer a ton at their regular price, making them an even better bargain when on sale like this.
This article originally appeared on Engadget at https://www.engadget.com/our-favorite-anker-wireless-earbuds-are-back-on-sale-for-50-125051545.html?src=rss
Normally, Microsoft’s Build is a straightforward (often dry) showcase of the company’s software and hardware developments, with a dash of on-stage coding to excite the developer crowd. But this year, the company is likely to make some huge AI moves, following its 2023 debut of ChatGPT-powered Bing Chat. Then, there’s new Surface hardware.
In fact, Microsoft has a showcase for new Surfaces and AI in Windows 11 on May 20, while Build actual kicks off a day later. And you know what? The Surface event might be the most impactful.
Rumors suggest we’ll see some of the first systems with Qualcomm’s Arm-based Snapdragon X Elite chip alongside new features in the next major Windows 11 update.
A refresh for its consumer PCs is likely to consist of new 13- and 15-inch Surface Laptop 6 models with thinner bezels, larger trackpads, improved port selection and that X Elite chip. We might even see an Arm-based version of the Surface Pro 10 too.
While Intel confirmed Microsoft is already working on ways to make Copilot local, we could see that reach consumers as well. By local, I mean the AI assistant could answer simpler questions, like basic math or queries about files on your system, without an internet connection.
Just like it slimmed down the latest iPad Pro, Apple may try to do the same to the iPhone. To be more precise, the company is working on a “significantly thinner” device that could arrive in 2025, according to The Information. An upgraded front-facing camera could sit alongside Face ID sensors in a smaller pill-shaped cutout, while the rear camera array could move to the center of the phone. The screen size would reportedly be between that of the current base iPhone and the iPhone Pro Max — so between 6.12 and 6.69 inches.
Slack is training its machine learning models on user messages, files and other content, without explicitly asking for permission. This means your private data is being used by default. To opt out, you need your organization’s Slack administrator (IT, HR, etc.) to contact Slack on your behalf.
In response to concerns, Slack recently clarified its data use in a blog post, assuring users that customer data is not used to train generative AI products, which typically rely on external large language models (LLMs). The company uses this data to train machine learning models for features like channel and emoji recommendations and search results. However, it’s misleading, at best, to say customers can opt out when “customers” doesn’t include employees working within an organization. It is also a little misleading, implying all your data is safe from AI training, when the company apparently gets to pick and choose which AI models the statement covers.
Two reliable leaks are showing off the entry-level Moto Razr 50 and high-end Razr 50 Ultra (likely branded as the 2024 Razr and Razr+ in the US), before Motorola even told us about them. The entry-level Razr (2024) will supposedly have a 3.63-inch cover display, quite a step up from the piddly 1.5-inch cover display on the 2023 version I tested. Sadly, no sign of the wood option included in the Edge 50 phone series unveiled last month.
This article originally appeared on Engadget at https://www.engadget.com/the-morning-after-what-to-expect-from-microsoft-build-2024-111524762.html?src=rss
New EV Policy 2024 Permits all EV Companies to Make Greenfield Investment Commitments
As per the new electric vehicle policy 2024, customs duty has been reduced to 15 percent and it will be imposed on EVs with a minimum CIF (Cost, Insurance, and Freight) value of USD 35,000 or above for a total period of 5 years.
The new 13-inch M4 iPad Pro is really, really thin. That’s inevitably going to make certain aspects of repairing the device even more difficult, which iFixit confirmed in a teardown published this weekend. But it does shine in one area — when it comes to replacing the battery, Apple made some seriously helpful changes.
“For the first time in an iPad Pro, we’re able to remove the battery immediately after removing the screen,” Teardown Tech Shahram Mokhtari wrote in a blog post. Mokhtari notes that “immediately is relative,” as there are still some screws and brackets to remove before the battery can be taken out, and the video documenting the process shows it takes a bit of work to get to the pull tabs beneath the batteries, but the new setup still shaves hours off the process compared to earlier models.
“The fact that you can remove the battery without having to remove every major component inside this device is still a huge win for repairability,” Mokhtari says in the video. “It’s a massive improvement over the previous generation.” Everything else, on the other hand, is going to be pretty tricky to repair without causing damage. “From the daughterboard to the speakers and coax cables, we found a whole bunch of stuff that’s glued down because there just isn’t enough space for screws.”
And, the teardown shows the new Apple Pencil Pro is a repairability nightmare. Mokhtari — who got cut by the Pencil while trying to get to its insides — called it “a disposable piece of crap once the battery dies.”
This article originally appeared on Engadget at https://www.engadget.com/ifixits-teardown-of-the-new-m4-ipad-pro-reveals-an-easier-to-replace-battery-205413201.html?src=rss
OpenAI reportedly made exiting employees choose between keeping their vested equity and being able to speak out against the company. According to Vox, which viewed the document in question, employees could “lose all vested equity they earned during their time at the company, which is likely worth millions of dollars” if they didn’t sign a nondisclosure and non-disparagement agreement, thanks to a provision in the off-boarding papers. OpenAI CEO Sam Altman confirmed in a tweet on Saturday evening that such a provision did exist, but said “we have never clawed back anyone's vested equity, nor will we do that if people do not sign a separation agreement (or don't agree to a non-disparagement agreement).”
An OpenAI spokesperson echoed this in a statement to Vox, and Altman said the company “was already in the process of fixing the standard exit paperwork over the past month or so.” But as Vox notes in its report, at least one former OpenAI employee has spoken publicly about sacrificing equity by declining to sign an NDA upon leaving. Daniel Kokotajlo recently posted on an online forum that this decision led to the loss of equity likely amounting to “about 85 percent of my family's net worth at least.”
in regards to recent stuff about how openai handles equity:
we have never clawed back anyone's vested equity, nor will we do that if people do not sign a separation agreement (or don't agree to a non-disparagement agreement). vested equity is vested equity, full stop.
In Altman’s response, the CEO apologized and said he was “embarrassed” after finding out about the provision, which he claims he was previously unaware of. “[T]here was a provision about potential equity cancellation in our previous exit docs; although we never clawed anything back, it should never have been something we had in any documents or communication,” he wrote on X. “this is on me and one of the few times i've been genuinely embarrassed running openai; i did not know this was happening and i should have [sic].” In addition to acknowledging that the company is changing the exit paperwork, Altman went on to say, “[I]f any former employee who signed one of those old agreements is worried about it, they can contact me and we'll fix that too.”
This article originally appeared on Engadget at https://www.engadget.com/sam-altman-is-embarrassed-that-openai-threatened-to-revoke-equity-if-exiting-employees-wouldnt-sign-an-nda-184000462.html?src=rss
Blue Origin is back in the space tourism game. Jeff Bezos’ spaceflight company successfully flew six paying customers to the edge of space and back this morning, breaking its nearly two-year-long hiatus from crewed missions. This was Blue Origin’s seventh trip with humans on board. The mission — a quick jaunt to cross the Kármán line, or the boundary of space, about 62 miles above Earth — lifted off from the company’s Launch Site One in West Texas shortly after 10:30AM ET.
The six people inside the New Shepard crew capsule included Ed Dwight, a former Air Force Captain who was the first Black astronaut candidate when he was picked for the training program in 1961. He went through training but ultimately wasn’t selected for NASA’s Astronaut Corps, and never made it to space until now. Also on board were Mason Angel, Sylvain Chiron, Kenneth L. Hess, Carol Schaller and Gopi Thotakura. The crew safely landed back on the ground about 10 minutes after launch.
This article originally appeared on Engadget at https://www.engadget.com/blue-origin-successfully-sends-tourists-to-the-edge-of-space-again-after-a-long-hiatus-144745261.html?src=rss
