Rolls-Royce has received funding from the UK Space Agency to develop a nuclear reactor for a moon base.

The project will look into how nuclear power could be used to support a future base on the moon for astronauts.

Scientists and engineers at the British company are working on the microreactor programme to develop technology that will provide power needed for humans to live and work on Earth’s natural satellite.

All space missions depend on a power source, to support systems for communications, life-support and science experiments.

Experts suggest nuclear power could dramatically increase the length of lunar missions.

The UK Space Agency has announced £2.9m of new funding for the project, which will deliver an initial demonstration of a UK lunar modular nuclear reactor.

This comes after a £249,000 study funded by the UK Space Agency in 2022.

The science minister George Freeman said: “Space exploration is the ultimate laboratory for so many of the transformational technologies we need on Earth: from materials to robotics, nutrition, cleantech and much more.

“As we prepare to see humans return to the moon for the first time in more than 50 years, we are backing exciting research like this lunar modular reactor with Rolls-Royce to pioneer new power sources for a lunar base.

“Partnerships like this – between British industry, the UK Space Agency and government – are helping to create jobs across our £16bn space tech sector and help ensure the UK continues to be a major force in frontier science.”

Rolls-Royce plans to have a reactor ready to send to the moon by 2029.

It will work with a variety of collaborators including the University of Oxford, University of Bangor, University of Brighton, University of Sheffield’s Advanced Manufacturing Research Centre (AMRC) and Nuclear AMRC.

Compared with other power systems, a relatively small and lightweight nuclear microreactor could enable continuous power regardless of location, available sunlight and other environmental conditions.

Abi Clayton, director of future programmes for Rolls-Royce, said: “This funding will bring us further down the road in making the microreactor a reality, with the technology bringing immense benefits for both space and Earth.

“The technology will deliver the capability to support commercial and defence use cases alongside providing a solution to decarbonise industry and provide clean, safe and reliable energy.”

Dr Paul Bate, chief executive of the UK Space Agency, said: “This innovative research by Rolls-Royce could lay the groundwork for powering continuous human presence on the moon, while enhancing the wider UK space sector, creating jobs and generating further investment.”

Source

Although the use of clean, almost unlimited nuclear fusion power in space has long captured the imagination in sci-fi novels and TV shows, the lunar reactor would use existing fission power.

The hope is that nuclear power could substantially increase the time humans can spend on the moon. The reactor would also be lighter and smaller than alternative power sources. However, unlike fusion, fission creates radioactive waste.

Rolls-Royce, whose new chief executive, Tufan Erginbilgic, has described its financial performance as “unsustainable”, signed an agreement with [sic]

Source

A while ago, Microsoft announced that it will hold a meeting on March 16, 8 am PT but didn't give any additional information. We know the name of it, which is "The Future of Work with AI." Apart from its name, Microsoft also filled out the "about" section but didn't specify any information. "Join us for a special event with Satya Nadella and Jared Spataro to learn how AI will power a whole new way of working for every person and organization. The live stream starts at 8 AM Pacific Time on March 16," says the about section, which clearly is not very precise.

Despite having very limited information, there is a strong possibility that the upcoming event will mainly cover the new GPT-4 integrations to Microsoft 365.

Microsoft 365 is most likely the focus of the Microsoft AI event

First things first, hosts give us a strong clue about what's on the way. The chairman and CEO, Satya Nadella, will be joined by the vice president of modern work and business applications, Jared Spataro. Spoiler alert, Spataro is also the head of Microsoft 365.

It is not a mystery that Microsoft wants to lead the AI industry, as the company makes it obvious with its courageous steps toward the future. It has been investing heavily in OpenAI and recently revealed that Bing will use GPT-4, which was revealed a couple of days ago.

Microsoft has brought new AI capabilities to Teams and wants to bring more to other apps like Outlook and Word. As you may know, Google recently revealed its new AI features for Gmail and Docs. Similar steps are awaited for Outlook and Word by the users. For Outlook, a possible email draft generator, and a writing suggestion feature for Word is expected. Microsoft will probably integrate GPT-4 into these apps.

Apart from those two, the Microsoft 365 family includes some other highly-used apps like OneDrive, Excel, PowerPoint, Access, and OneNote.

Microsoft is not only bringing these AI solutions to 365 but also to its subsidiary companies. Recently, LinkedIn announced new AI features, helping people share ideas easily and save time writing job descriptions and profile bios.

How to watch the event?

Microsoft decided to keep it simpler and live stream the event on LinkedIn. You don't have to do anything other than regular processes, and you also don't need a press card or invitation to be able to watch the live stream. Simply go to this link and hit the play button once the time comes.

You can also click on the "attend" button to let your connections know what you are up to. There is a discussion section at the right of the page where you can share your ideas with people and see others' thoughts. Currently, there are almost 30,000 attendees,  this number might go higher as we get closer to the event.

Microsoft AI event: What to expect?

According to a report by MacRumors, the latest AI event of Apple included very important information. The company revealed internally that engineers, including the Siri team, "have been testing language-generation concepts every week in response to the rise of chatbots like ChatGPT."

The report mentions that Siri is beyond expectations and not being used as anticipated is because of multiple roadblocks to meaningful improvements. One of the former Apple engineers, John Burkey, said Siri is built on a "clunky code that took weeks to update with basic features," in an interview with The New York Times. Apparently, it is very hard to add basic features and improvements to Siri, making it hard for Apple to bring game-changing abilities. Burkey also doesn't believe Siri to become a creative assistant like ChatGPT.

Apple wants to ensure Microsoft doesn't lead AI

OpenAI started a new era with ChatGPT, and now all the tech giants are trying to keep up with it. Microsoft continues its heavy investments in OpenAI and artificial intelligence in general, while Google is trying to launch Bard as well as multiple other projects. OpenAI launched GPT-4 a couple of days ago and set the bar even higher, making it hard for other tech giants even to get close.

Even though Siri has major design issues, Apple is keen on keeping up with the trends and hopping on the bandwagon to catch others. Siri might cause issues for engineers; as Burkey says, even adding a single word to its database takes too much time. Apple might be late for the party as the engineers are clearly taking the longer road.

On a side note, Apple held its internal AI summit last month. Gurman said the event would take place in the Steve Jobs Theather at the Apple HQ, and the in-person event would also be streamed to employees. Exactly the way Apple used to hold media events before the pandemic.

If you want to know how to use Siri as an AI-powered chatbot with OpenAI's ChatGpt, check this article out!

Apple's chatbot ambitions halted by major obstacle

In an article, we explained the differences, advantages and disadvantages of going either way. One of the advantages of DDR4 we highlighted was its reusability as users who were running high-end DDR4 kits could re-use the same when upgrading to LGA1700 sockets for their 12th/13th Gen Intel systems.

However, that advantage may not last for very long as a new report alleges that Intel's motherboard vendor partner MSI is planning to gradually phase out at least some of the Z790 and B760 DDR4 variants according to some Chinese sources. The affected motherboard models:

• MPG Z790 EDGE WIFI DDR4

• PRO Z790-A WIFI DDR4

• PRO Z790-P DDR4

• MAG B760 MORTAR DDR4

   

As a manufacturer of Intel boards, it perhaps makes sense now to slowly kill off DDR4 boards, at least for the high-end segment as competing AMD solutions, ie, Ryzen 7000 series, only have the option for DDR5. DDR5 is also gradually getting faster and cheaper, and support for non-standard memory capacities, like 192GB, is also getting better. Although we hope budget motherboard models still offer some DDR4 variants as DDR4 memory remains highly affordable, and it continues to be supported by AMD's AM4 socket.

Source: Bobatang (Board channels) via Expreview

MSI allegedly killing off Intel 12th, 13th Gen DDR4 motherboards in favor of DDR5

Samsung's Galaxy S23 ad, showing the moon photography mode.

Samsung

 

If you take a photo of the moon on a Samsung device, it will return a detailed photo of the moon. Some people are mad about this.

 

The issue is that Samsung's software fakes some details the camera can't really see, leading a Reddit user called ibreakphotos to accuse the company of "faking" moon photos. The user's post claims to be able to trick Samsung's moon detection, and it went viral enough that Samsung's press site had to respond.

 

Samsung's incredibly niche "Moon Mode" will do certain photo processing if you point your smartphone at the moon. In 2020, the Galaxy S20 Ultra launched with a "100x Space Zoom" (it was really 30x) with this moon feature as one of its marketing gimmicks. The mode is still heavily featured in Samsung's marketing, as you can see in this Galaxy S23 ad, which shows someone with a huge, tripod-mounted telescope being jealous of the supposedly incredible moon photos a pocketable Galaxy phone can take.

 

We've known how this feature works for two years now—Samsung's camera app contains AI functionality specifically for moon photos—though we did get a bit more detail in Samsung's latest post. The Reddit post claimed that this AI system can be tricked, with ibreakphotos saying that you can take a picture of the moon, blur and compress all the detail out of it in Photoshop, and then take a picture of the monitor, and the Samsung phone will add the detail back. The camera was allegedly caught making up details that didn't exist at all. Couple this with AI being a hot topic, and the upvotes for faked moon photos started rolling in.

 

On one hand, using AI to make up detail is true of all smartphone photography. Small cameras make for bad photos. From a phone to a DSLR to the James Webb Telescope, bigger cameras are better. They simply take in more light and detail. Smartphones have some of the tiniest camera lenses on Earth, so they need a lot of software to produce photos that are anywhere near reasonable in quality.

 

"Computational photography" is the phrase used in the industry. Generally, many photos are quickly taken after you press the shutter button (and even before you press the shutter button!). These photos are aligned into a single photo, cleaned up, de-noised, run through a bunch of AI filters, compressed, and saved to your flash storage as a rough approximation of what you were pointing your phone at. Smartphone manufacturers have to throw as much software at the problem as possible because no one wants a phone with a giant, protruding camera lens, and normal smartphone camera hardware can't keep up.

 

But lighting aside, the moon basically always looks the same to everyone. While it spins, the Earth spins, and the two spin around each other; gravitational forces put the moon in a "synchronous rotation" so we always see the same side of the moon, and it only "wobbles" relative to Earth. If you make an incredibly niche camera mode for your smartphone specifically targeted at only moon photography, you can do a lot of fun AI tricks with it.

 

Who would know if your camera stack just lies and patches in professionally shot, pre-existing photos of the moon into your smartphone picture? Huawei was accused of doing exactly that back in 2019. The company allegedly packed photos of the moon into its camera software, and if you took a photo of a dim light bulb in an otherwise dark room, Huawei would put moon craters on your lightbulb.

 

That would be pretty bad. But what if you took one step back from that and simply involved an AI middleman instead? Samsung took a bunch of photos of the moon, trained an AI on those photos, and then set the AI loose on users' photos of the moon. Is that crossing a line? How specific are you allowed to get with your AI training use cases?

 

Samsung's press release mentions a "detail enhancement engine" for the moon, but it doesn't go into much detail about how it works. The article includes a few unhelpful diagrams about moon mode and AI that all basically boil down to "a photo goes in, some AI stuff happens, and a photo comes out."

 

In the company's defense, AI is often called a "black box." You can train these machine-learning models for a desired outcome, but no one can explain exactly how they work. If you're a programmer that hand-writes a program, you can explain what each line of code does because you wrote the code, but an AI is only "trained"—it programs itself. This is partly why Microsoft is having such a hard time making the Bing chatbot behave.

The press release mostly talks about how the phone recognizes the moon or how it adjusts brightness, but those points are not the issue—the issue is where the detail comes from. While there's no choice quote we can pull, the above image shows pre-existing moon images being fed into the "Detail Enhancement Engine." The whole right side of this diagram is pretty suspicious. It says Samsung's AI compares your moon photo with a "high-resolution reference" and will send it back into the AI detail engine if it's not good enough.

 

That does feel like Samsung is cheating a bit, but where exactly should the line for AI photography be? You definitely wouldn't want an AI-free smartphone camera—that would be a worst-in-class camera. Even non-AI photos from a big camera are just electronic interpretations of the world. They're not "correct" references of how things should look; we're just more used to them. Even objects viewed with the human eye are just electrical signals interpreted by your brain and look different for everyone.

 

It would be a real problem if Samsung's details were inaccurate, but the moon really does look like that. If a photo is completely accurate and looks good, it's hard to argue against it. It also would be a problem if the moon detail was inaccurately applied to things that aren't the moon, but taking a picture of a Photoshopped image is an extreme case. Samsung says it will "improve Scene Optimizer to reduce any potential confusion that may occur between the act of taking a picture of the real moon and an image of the moon," but should it even do that? Who cares if you can fool a smartphone with Photoshop?

The key here is that this technique only works on the moon, which looks the same for everybody. Samsung can be super aggressive about AI detail generation for the moon because it knows what the ideal end result should look like. It feels like Samsung is cheating because this is a hyper-specific use case that doesn't provide a scalable solution for other subjects.

 

You could never use an aggressive AI detail generator for someone's face because everyone's face looks different, and adding details would make that photo not look like the person anymore. The equivalent AI technology would be if Samsung trained an AI specifically on your face and then used that model to enhance photos it detected you were in. Someday, a company may offer hyper-personalized at-home AI training based on your old photos, but we're not there yet.

 

If you don't like your improved moon photos, you can just turn the feature off—it's called "Scene Optimizer" in the camera settings. Just don't be surprised if your moon photos look worse.

Moon-gate: Samsung fans are mad about AI-processed photos of the moon

ChatGPT is not GPT 3.5

Before delving into the differences, it's important to note that ChatGPT is not a version of OpenAI's large language model; instead, it is a chat-based interface that interacts with the model that powers it. While ChatGPT was initially identified as GPT-3.5, it is merely a means of communicating with GPT-3.5 and now with GPT-4.

Now that we have clarified this distinction, let's explore the distinctions between the beloved chatbot and its recently enhanced successor.

GPT-4 is multimodal in its approach

The latest improvement to this adaptable machine learning system is its ability to process 'multimodal' information, which means it can now understand multiple modes of data. In comparison, ChatGPT and GPT-3 were restricted to text-based interactions, being capable of reading and writing but little else - though this was sufficient for numerous use cases.

GPT-4's improved design enables it to comprehend and process relevant information within images, extending beyond simple image recognition tasks. For instance, OpenAI has showcased GPT-4's ability to explain the humor behind an image depicting an oversized iPhone connector. However, the partnership with Be My Eyes highlights a more significant potential of the model. The video demonstrates GPT-4's diverse capabilities by showcasing the utility performing various tasks. This indicates that GPT-4 has the potential to understand and interpret image data in a comprehensive and multifaceted manner.

GPT-4’s memory is more extensive

Large language models like GPT-4 are trained on vast amounts of text data, including web pages and books. However, during actual conversations with users, these models have limits to how much information they can keep in their attention function. The previous versions, GPT-3.5 and the original ChatGPT, were limited to 4,096 'tokens,' which roughly translates to 8,000 words or four to five pages of a book. Beyond this point, the model would start losing track of the conversation.

In contrast, GPT-4 has a token limit of 32,768, which is equivalent to around 64,000 words or 50 pages of text. This increase in token count allows the model to remember up to 50 pages of information during a conversation or text generation. For instance, it can recall a topic discussed 20 pages back in a conversation or refer to an event that occurred 35 pages ago in a written story or essay. While this is an approximate description of how the attention mechanism and token count work, the general idea is that GPT-4 has an expanded memory capacity, enabling it to handle more complex tasks.

It’s harder to lead GPT-4 astray

Despite their advancements, contemporary chatbots are prone to getting misled easily. With just a little bit of coaxing, they can be manipulated to say strange and sometimes unsettling things.  For instance, chatbots can be persuaded to elaborate on what a 'bad AI' would do, or they can collaborate on 'jailbreak' prompts to overcome their restrictions.

However, GPT-4 has undergone extensive training on malicious prompts, which users generously provided to OpenAI over the past couple of years. As a result, GPT-4 surpasses its predecessors in terms of factuality, steerability, and its ability to remain within established guidelines. According to OpenAI, GPT-3.5 (which powered ChatGPT) was a 'test run' of a new training architecture. The lessons learned from that experience were applied to the new model, resulting in a model that is 'unprecedentedly stable.' Additionally, OpenAI was able to predict the model's capabilities more accurately, resulting in fewer surprises.

GPT-4 processes more languages

The field of AI is primarily dominated by English-speaking individuals, and the majority of data, testing, and research papers are in English. However, the capabilities of large language models can be applied to any written language, making it imperative to make them available in other languages.

GPT-4 takes a significant step in this direction by showcasing its ability to accurately answer thousands of multiple-choice questions across 26 different languages, ranging from Italian and Ukrainian to Korean. While the model performs best on Romance and Germanic languages, it generalizes well to other languages as well.

While this preliminary testing of GPT-4's language capabilities is promising, it does not constitute a comprehensive adoption of multilingual abilities. The testing criteria were initially translated from English, and multiple-choice questions do not necessarily represent everyday language usage. However, the model performed remarkably well on this task, despite not having been explicitly trained for it. This suggests that GPT-4 could be significantly more accommodating to non-English speakers.

GPT-4 takes on different personalities

The concept of 'steerability' in artificial intelligence pertains to an AI system's ability to adapt its behavior according to specific requirements. This capability can be beneficial, as in assuming the role of an empathetic interlocutor, or potentially hazardous, such as when individuals manipulate the AI into exhibiting malevolent or despondent traits.

GPT-4 incorporates steerability more seamlessly than its predecessor, GPT-3.5, enabling users to customize the default ChatGPT personality, characterized by a consistent verbosity, tone, and style, to better align with their preferences. The development team emphasizes that this customization has its limits, acknowledging that pushing the model too far may result in breaking character.

Previously, users could indirectly achieve a degree of steerability by priming the chatbot with prompts like, 'Imagine you are a dungeon master in a tabletop RPG' or 'Respond as if you are a guest on a cable news interview.' However, this approach merely offered suggestions to the default GPT-3.5 personality. GPT-4 now empowers developers to incorporate a specific viewpoint, communication style, tone, or interaction mode from the outset.

GPT-4: AI for the next generation of tech

OpenAI's latest AI model, GPT-4, has introduced significant improvements and features that set it apart from its predecessors. With the ability to comprehend images, increased token counts, improved factuality, steerability, and multilingual capabilities, GPT-4 has the potential to revolutionize the field of AI. Additionally, the model's native steerability feature allows users to tailor the chatbot's conversational style and interaction methods to better suit their needs. While there is still room for improvement, GPT-4 is undoubtedly a remarkable achievement that will pave the way for even more advanced AI systems in the future.

Introducing GPT-4: OpenAI's Latest Language Model

LAST NOVEMBER, AT the Stockholm University of the Arts, a human and an AI made music together. The performance began with musician David Dolan playing a grand piano into a microphone. As he played, a computer system, designed and overseen by composer and Kingston University researcher Oded Ben-Tal, “listened” to the piece, extracting data on pitch, rhythm, and timbre. Then, it added its own accompaniment, improvising just like a person would. Some sounds were transformations of Dolan’s piano; some were new sounds synthesized on the fly. The performance was icy and ambient, eerie and textural. 

This scene, of a machine and human peacefully collaborating, seems irreconcilable with the current artists-versus-machines discourse. You will have heard that AI is replacing journalists, churning out error-riddled SEO copy. Or that AI is stealing from illustrators, who are suing Stability AI, DeviantArt, and Midjourney for copyright infringement. Or that computers are rapping, or at least trying to: the “robot rapper” FN Meka was dropped by Capitol Records following criticism that the character was “an amalgamation of gross stereotypes.” In the most recent intervention, none other than Noam Chomsky claimed that ChatGPT exhibits the “banality of evil.”

These anxieties slot neatly among concerns about automation, that machines will displace people—or, rather, that the people in control of these machines will use them to displace everyone else. Yet some artists, musicians prominent among them, are quietly interested in how these models might supplement human creativity, and not just in a “hey, this AI plays Nirvana” way. They are exploring how AI and humans might collaborate rather than compete. 

“Creativity is not a unified thing,” says Ben-Tal, speaking over Zoom. “It includes a lot of different aspects. It includes inspiration and innovation and craft and technique and graft. And there is no reason why computers cannot be involved in that situation in a way that is helpful.”

SPECULATION THAT COMPUTERS might compose music has been around as long as the computer itself. Mathematician and writer Ada Lovelace once theorized that Charles Babbage's steam-powered Analytical Engine, widely hailed as the first computer, could be used for something other than numbers. In her mind, if the “science of harmony and of musical composition” could be adapted for use with Babbage’s machine, “the engine might compose elaborate and scientific pieces of music of any degree of complexity or extent.”

The first book on the subject, Experimental Music: Composition with an Electronic Computer, written by American composer and professor Lejaren Hiller Jr. and mathematician Leonard Isaacson, appeared in 1959. In popular music, artists like Ash Koosha, Arca, and, most prominently, Holly Herndon have drawn on AI to enrich their work. When Herndon spoke to WIRED last year about her free-to-use, “AI-powered vocal clone,” Holly+, she explained the tension between tech and music succinctly. “There’s a narrative around a lot of this stuff, that it’s scary dystopian,” she said. “I’m trying to present another side: This is an opportunity.”

Musicians have also reacted to the general unease generated by ChatGPT and Bing’s AI chatbot. Bogdan Raczynski, reading transcripts of the chatbots’ viral discussions with humans, says over email that he detected “fright, confusion, regret, guardedness, backtracking, and so on” in the model’s responses. It isn’t that he thinks the chatbot has feelings, but that “the emotions it evokes in humans are very real,” he says. “And for me those feelings have been concern and sympathy.” In response, he has released a “series of comforting live performances for AI” (emphasis mine).

BEN-TAL SAYS HIS work presents an alternative to “the human-versus-machine narrative.” He admits that generative AI can be unsettling because, on a superficial level at least, it exhibits a kind of creativity normally ascribed to humans, but he adds that it is also just another technology, another instrument, in a lineage that goes back to the bone flute. For him, generative AI isn’t unlike turntables: When artists discovered they could use them to scratch records and sample their sounds, they created whole new genres.

In this vein, copyright may need a substantial rethink: Google has refrained from releasing its MusicLM model, which turns text into music, because of the “the risks associated with music generation, in particular, the potential misappropriation of creative content.” In a 2019 paper, Ben-Tal and other researchers asked readers to imagine a musician holodeck, an endpoint for music AI, that has archived all recorded music and can generate or retrieve any possible sound on request. Where do songwriters fit into this future? And before then, can songwriters defend themselves against plagiarism? Should audiences be told, as WIRED does in its articles, when AI is used?

Yet these models still present attractive creative capabilities. In the short term, Ben-Tal says, musicians can use an AI, as he did, to improvise with a pianist outside of their skill set. Or they can draw inspiration from an AI’s compositions, perhaps in a genre they are not familiar with, like Irish folk music. 

And in the longer term, AI might fulfill a wilder (albeit controversial) fantasy: It could effortlessly realize an artist’s vision. “Composers, you know, we come up with ideas of what music we would like to create, but then translating these into sounds or scores, realizing those ideas, is quite a laborious task,” he says. “If there was a wire that we could plug in and get this out, that could be very fantastic and wonderful.” 

More urgently, mundane and pervasive algorithms are already mangling the industry. Author Cory Doctorow has written about Spotify’s chokehold on music—how playlists, for instance, encourage artists to abandon albums for music that fits into “chill vibes” categories, and train audiences to let Spotify tell them what to listen to. Introduced into this situation, AI will be the enemy of musicians. What happens when Spotify unleashes its own AI artists and promotes those? 

Raczynski hopes he will catch the wave rather than be consumed by it. “Perhaps in a roundabout way, like it or not, I am acknowledging that short of going off the grid, I have no choice but to develop a relationship with AI,” he says. “My hope is to build a reciprocal relationship over a self-centered one.”

Source

Users who visit the Bing website and activate the Chat link at the top of the webpage may still get the waitlist prompt that explains that access to the new Bing is required for access to the Chat mode.

The "join the waiting list" button, however, activates access immediately now. It is still necessary to sign-in with a Microsoft account, or create a new one, to gain access to the new Bing.

Most Windows 10 and 11 customers have access to a Microsoft account already, as Microsoft made it the default option when setting up the PC for the first time. While there are still options available to create a local account during setup, it is no longer as easy as pressing a button in the interface.

Microsoft's launch of the new AI-powered Bing has been a success for the company. Recently, Microsoft revealed that its search engine crossed the 100 million daily active users mark.

Up until recently, interested users had to join a waiting list, if they wanted to gain access to Bing Chat and the new Bing. Some users waited weeks for access in the beginning.

Microsoft has not confirmed the change yet, but it is possible that the company will announce it later today during its AI event. The event's main speakers are Microsoft CEO Satya Nadella  and Microsoft corporate vice president of modern work and business applications, Jared Spataro.

The event, The Future of Work with AI, will be streamed live on the Internet. Interested users may join it on LinkedIn. While Microsoft has not revealed any details regarding the event, it is expected that Microsoft is showcasing integration of AI into Office applications and Microsoft 365.

Microsoft did add Bing Chat to Microsoft Edge 111, which it released earlier this week. The integration did not sit well with some browser users. The new Bing icon in Edge feels oversized to some, and it activates on hover.

We have published a guide on disabling the Bing icon in Microsoft Edge, and a follow-up article on Microsoft adding an option to turn off the Bing button directly in Edge.

We confirmed the skipping of the waitlist today, and colleagues over at The Verge, Windows Central, and Dr. Windows came to the same conclusion. It is possible that Microsoft has free slots to give away, and that the waitlist will kick in when these are used up. For now though, it appears that anyone may start using Bing Chat right away, without having to wait days or even weeks for access.

Microsoft opens up access to Bing Chat for everyone

Today, Dell has launched its first ARM based Windows laptop, powered by the Qualcomm Snapdragon 8cx Gen 2 Compute Platform. Running Windows 11 this machine comes with a 14” FHD LED display with anti-glare technology, as well as Adreno 690 graphics, 8 GB of LPDDR4x memory and a 256 GB M.2 PCIe NVMe SSD.

With only the one configuration which is priced at $499, this Inspiron 14 machine sits alongside both AMD and Intel offerings being priced considerably higher at $549.99 and $649.99 respectively. We’ve compared the specs for you down below.

In terms of connectivity, it comes with two USB-C ports, one USB port, a 3.5 mm headset jack and a Micro-SD card slot. It also includes a 1080p FHD Webcam with dual microphone, 2 speakers, dual-band Wi-Fi 6 and Bluetooth 5.2.

The battery is rated at 40 Wh with a 65W Type-C power adapter in the box, rated for up to 16 hours of HD local video playback. Combined with the fanless experience thanks to the efficiency of ARM processor, and a lightweight build, this machine makes for a competitive entry into the Windows laptop space.

The Qualcomm powered Inspiron 14 is available to order today, with delivery on 3rd April with standard delivery, or 23rd March with Express delivery.

Source: Snapdragon (Twitter)

Dell launches first Qualcomm Snapdragon Windows 11 laptop into its Inspiron 14 range

In this case, the dark red area along the front of the wing represents higher-speed airflow as the TTBW’s wings, which are thinner than those of today’s commercial airliners, pierce the air. The tan-colored area shows the relatively smooth wake generated by the aerodynamic wings.

A TTBW aircraft produces less drag due to its longer, thinner wings supported by aerodynamic trusses. In flight, it could consume up to 10% less jet fuel than a standard airliner.

The Advanced Supercomputing Division of NASA’s Ames Research Center in California created this image as part of an effort by the Transformational Tools and Technologies project to develop computational tools for TTBW research.

In January, NASA selected a TTBW concept from The Boeing Company for its Sustainable Flight Demonstrator project.

Microsoft is trying to close its proposed ~$69 billion deal to purchase Activision Blizzard. However, the FTC is asking the judge in the case to compel Microsoft to produce requested documents from the company. The filing states:

More than five weeks after receiving thirty-two document requests (the “RFPs”), Respondents have produced only a smattering of self-selected documents in response and have refused to produce anything at all for twenty-four requests.

The FTC claims that it has met with Microsoft executives six times in an attempt to get the documents, but that the company has come up with a number of arguments against those requests, including that there is "simply not enough time” for Microsoft to produce the documents.

In addition to getting more information on Microsoft's proposal to buy Activision Blizzard, including its 10-year agreement deals with various other game publishers and cloud services, the FTC is also asking for documents on the company's previous acquisition of ZeniMax Media.

Source: FTC

The FTC says Microsoft is holding back documents it requested for Activision Blizzard deal

According to unnamed sources who spoke to Bloomberg, Apple is delaying bonuses for some corporate divisions. This primarily affected divisions who received bonuses twice a year. Under the new strategy, their bonuses will now be handed out in October, like all other divisions who receive annual bonuses. Previously, one of their two bonuses was scheduled for April.

Moreover, Apple has removed headcount goals as well. Sources within the company say that new hiring is being "limited" and when a current employee leaves their job, their position is typically left vacant.

It is important to note that the bonus amount isn't being reduced, it will just be bundled as a single bonus rather than divided into two. The move will affect corporate retail employees, engineers, and mid-level managers. Many employees switch jobs after receiving bonuses so this could also retain such workers until at least October. However, it will adversely affect people who were eagerly awaiting bonuses for personal budgeting.

Senior employees who are directors or at an even higher designation are reportedly not affected, with their bonuses still being doled out on a quarterly basis.

Although Apple has already laid off some contractors, reduced travel budgets, and cut pay for its CEO Tim Cook, it remains to be seen to be seen how effective the latest measures are in reducing overall costs. Some employees and retail workers have already cited concerns that their working hours and office days are being more closely scrutinized, fearing that this could be used as the basis for their firing. A few have also noted that they feel like they are being encouraged to quit themselves by being asked to work more than they are contractually obligated to.

Source: Bloomberg (paywall)

Apple isn't laying off employees yet, but it is delaying bonuses and reducing hiring

Amazon unveils three satellite user terminals, plans broadband service in 2024

The event's website is now open for registration for people who want to attend. The virtual events will be held from May 23-24 at no cost to attendees. They will be able to view the keynote speech from Microsoft CEO Satya Nadella, along with other workshops and events. The in-person part of the conference will actually be on May 22, with some pre-event workshops, followed by the main conference on May 23-25, all in Seattle. The event fee is $1,525, with an additional $225 to attend the May 22 pre-event workshops.

As to what Microsoft might discuss during Build, that is still something of a mystery. It's not likely we will get any new Surface hardware announcements at the conference, It is possible we might get a glimpse of some upcoming Windows 11 features, and perhaps even get an official time frame for Windows 12. You can also bet there will be some more info on Microsoft's AI plans, including for Bing Chat.

Microsoft Build 2023 confirmed for May 23-25 with digital and in-person events

Previously, Microsoft would only say that Bing Chat was using its own Prometheus model and the "new, next-generation OpenAI large language model".

Other than that, Microsoft is only saying that GPT-4 for Bing Chat has been "customized for search". There's no other info yet, including when Microsoft might add the image-text combo feature that was revealed by OpenAI today.

Yes, Microsoft's Bing Chat is already using OpenAI's GPT-4

Perhaps the biggest new feature is that GPT-4 can answer inquiries that combine both text and images. The blog states:

Over a range of domains—including documents with text and photographs, diagrams, or screenshots—GPT-4 exhibits similar capabilities as it does on text-only inputs. Furthermore, it can be augmented with test-time techniques that were developed for text-only language models, including few-shot and chain-of-thought prompting.

The bad news is that image inputs are not currently available for the general public as they are in a "research preview" mode. There's no word on when this will be generally available.

The new version will also allow users to alter the tone of GPT-4's answers. The blog states:

Rather than the classic ChatGPT personality with a fixed verbosity, tone, and style, developers (and soon ChatGPT users) can now prescribe their AI’s style and task by describing those directions in the “system” message. System messages allow API users to significantly customize their users’ experience within bounds.

The blog also showed off some performance charts to illustrate GPT-4's improvements. In one case, OpenAI gave the chatbot some simulated exams. It stated:

We proceeded by using the most recent publicly-available tests (in the case of the Olympiads and AP free response questions) or by purchasing 2022–2023 editions of practice exams. We did no specific training for these exams.

The chart above shows GPT-4 offered much better exam results in most cases compared to GPT-3.5

As with previous versions, GPT-4 is limited in terms of getting more recent information about subjects, as it can only look back to September 2021. OpenAI also says the new version "significantly reduces hallucinations relative to previous models" but it "is not fully reliable."

The GPT-4 API is currently available for developers to access if they sign up for a waitlist. The general public can check out GPT-4 on a limited basis if they sign up for ChatGPT Plus, which costs $20 a month. The blog post mentions OpenAI could launch another subscription tier that will allow for higher volume levels of access.

Source: OpenAI

GPT-4 officially launches from OpenAI with text-image support and more

The Build 2023 will be a three-day conference registered developers can attend in-person in Seattle for the first time since 2019 (last year's conference was still held online). Other developers and customers can watch Microsoft unveil its latest developments during a two-day global digital event on May 23-24, 2023.

Microsoft moved away from announcing consumer-facing products during its Build conference (remember Windows Phone 8.1 on Build 2014?), so you probably should not expect anything major for Windows 11 or Surface. Still, there might be room for some Windows or Edge-related announcements. And with Microsoft going full steam ahead with AI, expect the company to spend a lot of time talking about artificial intelligence.

Microsoft should publicly announce the Build 2023 dates and open registration in the upcoming weeks. More details and schedules will follow.

A leak reveals the date of Microsoft's upcoming Build conference

In a post on the System Shock Remake page on Steam, Nightdive stated:

We had hoped to bring the game to market by the end of March, but that turned out to be just beyond our reach; we are after all merely human (unlike Shodan!)

The game is a full remake of the original 1994 sci-fi first-person action-adventure game. Your character is on board the space-based Citadel Station and must navigate through its corridors, solve puzzles and defeat a ton of enemies in an effort to shut down the evil AI SHODAN. Nightside announced plans to remake the game way back in 2015 and even used a Kickstarter campaign to raise the initial funds. It's been delayed a number of times since then, but hopefully, May 30 is a firm date.

The System Shock remake will include not only higher end visuals, but a new interface, new enemies, and more. You can try a free demo of the game now when you head to its Steam page. You can preorder the game now on Steam, GoG.com, and the Epic Games Store. Pre orders will also be able to get System Shock 2: Enhanced Edition for free when it is released. The game is also coming to the PlayStation and Xbox consoles, but there's no release date yet for those platforms.

The System Shock remake gets pushed back from March to May 30

Why it does and doesn’t matter if Google, Microsoft, or Zoom certify your webcam

Meta to lay off another 10,000 as Zuckerberg celebrates “year of efficiency”

Silicon Labs’ xG27 chips can be as small as 2.3mm by 2.6mm — and yes, this image is to scale.

Image: Silicon Labs

 

Miniaturizing tech is a perpetual challenge for wearable makers. The smaller the device, the better it is for wearability. The thing is, that usually comes at the expense of battery life. However, Silicon Labs is hoping its latest xG27 chipset is small and energy-efficient enough to spark some big ideas in the medical tech space — like a saliva reader that’s so tiny it can be mounted onto a tooth.

 

According to Silicon Labs, the xG27 family of SoCs consists of the BG27 and the MG27. Both are built around the ARM Cortex M33 processor, but the BG27 focuses on Bluetooth, while the MG27 supports Zigbee and other protocols. As for how small these chips are, the xG27 SoCs range from 2mm-squared to 5mm-squared — roughly the width of a No. 2 pencil’s lead tip to the width of the pencil itself. It’s not the world’s smallest Bluetooth chip, but Silicon Labs spokesperson Sam Ponedal tells The Verge that’s only by “fractions of a millimeter.”

 

Pencil for scale.

Image: Silicon Labs

 

This is neat from a technical perspective, but what’s cooler is the BG27 is currently being used to develop an actual product — the aforementioned tooth-mounted wearable sensor. Lura Health, a medical device maker, says it’s using the chip for its “salivary diagnostic sensor.” The sensor is small enough to be glued to a molar (or placed inside a “smart retainer”) with the intent of continually monitoring a patient’s saliva. That, in turn, would allow dentists and clinicians to potentially test for more than 1,000 health conditions.

 

This wouldn’t be the first time that a company has pitched this kind of futuristic health tech. But while most attempts are thwarted by the FDA regulatory process, Lura Health claims it has just finished clinical trials for the sensor with UConn Orthodontics and is currently preparing to undergo the FDA regulatory process. If all goes well, the product could hit the market in 12-18 months.

 

As for other use cases, Silicon Labs says its chips are good candidates for medical patches, continuous glucose monitors, and wearable EKGs. That’s because they can operate on as low as 0.8 volts and can switch to a “shelf mode” that reduces energy use during transportation and while stored on shelves. These features aren’t quite as appealing in consumer wearables, but it opens the door for greater wearable use in hospitals and clinical settings.

 

 

This tiny chip is being used to develop a tooth-mounted sensor that can read your spit

New data tracks failure rates of 13 SSD models, going back up to 4 years

The new ND H100 v5 VM from Microsoft use NVIDIA's H100 GPUs, an upgrade from the previous A100 GPUs. Companies that need to add AI features can access this virtual machine service that has the following features:

• 8x NVIDIA H100 Tensor Core GPUs interconnected via next gen NVSwitch and NVLink 4.0
• 400 Gb/s NVIDIA Quantum-2 CX7 InfiniBand per GPU with 3.2Tb/s per VM in a non-blocking fat-tree network
• NVSwitch and NVLink 4.0 with 3.6TB/s bisectional bandwidth between 8 local GPUs within each VM
• 4th Gen Intel Xeon Scalable processors
• PCIE Gen5 host to GPU interconnect with 64GB/s bandwidth per GPU
• 16 Channels of 4800MHz DDR5 DIMMs

This is in addition to Microsoft's previously announced ChatGPT in Azure OpenAI Service, which lets third parties access the chatbot tech through Azure.

In a separate blog post, Microsoft talks about how the company first started working with OpenAI to help create the supercomputers that are needed for ChatGPT's large language model (and for Microsoft's own Bing Chat). That meant linking up thousands of GPUs together in an all-new way. The blog offered up an explanation from Nidhi Chappell, Microsoft's head of product for Azure high performance computing and AI:

To train a large language model, she explained, the computation workload is partitioned across thousands of GPUs in a cluster. At certain phases in this computation – called allreduce – the GPUs exchange information on the work they’ve done. An InfiniBand network accelerates this phase, which must finish before the GPUs can start the next chunk of computation.

This hardware is combined with software to help optimize the use of both the NVIDIA GPUs and the network that keeps all of them working together. Microsoft says it is continuing to add GPUs and expanding its network while also trying to keep them working 24/7 via cooling systems, backup generators and uninterruptible power supply systems.

Microsoft shows how it combines Azure with NVIDIA chips to make AI supercomputers