In his email announcing Panay's departure, as posted on The Verge, Rajesh Jha, Microsoft’s vice president of experience and devices, went over the new teams and leaders. Yusuf Mehdi, who previously was Microsoft's Corporate Vice President and its Consumer Chief Marketing Officer "will take on the responsibility of leading the Windows and Surface businesses with our OEM and Retail partners" according to Jha's email.

In addition, Pavan Davulury, who previously was the general manager for the Surface division as well as its Corporate Vice President for Windows Silicon and Systems Integration will now be the leader for Microsoft's Windows planning and release management, according to Jha.

He added that the company's "commitment to Surface and MR (mixed reality) remains unchanged." That means we can expect to see new Surface PC products beyond this year. Microsoft is set to show off a number of new Surface hardware product refreshes at its September 21 press event in New York City.

Finally, Jha's email stated that Mikhail Parakhin will be the leader of an all-new Windows and Web Experiences Team. Parakhin was previously Microsoft’s head of Advertising and Web Services, but in the past several months he has offered a lot of commentary and info via his X (formerly Twitter) account about the company's Bing Chat chatbot services.

It definitely will be interesting to see how Panay's departure influences Microsoft's Surface products going forward, along with its plans to improve the Windows OS, particularly as the company continues to push its AI products like Bing Chat and Windows Copilot.

Source

The big headline from this is that the next installment in The Elder Scrolls franchise, The Elder Scrolls VI, originally announced at E3 2018, is not coming to Sony's PlayStation 5 console. The other key bit of information around this title is that it will not be released for at least another 3 years, slated to be 2026 or later.

The above is very interesting given that only a couple of months ago, Phil Spencer stated that they were "unclear on what platforms it's launching on," even though the image above clearly shows that The Elder Scrolls VI is to launch only on Xbox and PC platforms.

Ever since Microsoft purchased ZeniMax Media, which includes Bethesda Game Studios, back in 2021 for $7.5 billion, it's been expected that the studio's upcoming games would end up as Microsoft-exclusive titles even when they were historically multi-platform, such as The Elder Scrolls V: Skyrim.

Upcoming titles from Bethesda Game Studios have been confirmed to be coming only to Microsoft's platforms in previous months, such as the untitled Indiana Jones game that is in development (not featured on the graphic above), which was originally announced in 2021 with no further details revealed since.

This news follows on from a few revelations from these documents, such as how Microsoft has been exploring the ability to stream PC games as well as Xbox titles in the future. Any developments into this would ultimately fall into Ubisoft's hands if the deal to buy Activision Blizzard was approved by the UK's Competition and Markets Authority.

Source: The Verge

Source

The announcement was met with a massive amount of negative comments online, particularly from current Unity game developers who felt that this policy change would put many smaller indie game creators in financial jeopardy. Unity has since tried to do damage control, stating some exceptions to the policy for certain platforms and for certain releases, such as games released for charity and also for subscription services like Microsoft's Xbox Game Pass.

However, those clarifications have not served to end the swarm of protests against Unity. Tonight, in an unusual late Sunday post on its X (formerly Twitter) account, Unity seems to have blinked at least partially,

The post stated that the company wanted to "apologize for the confusion and angst" about its revenue policy changes. It says that it has been talking and listening to "team members, community, customers, and partners" and that it will be "making changes to the policy". The post added that we can expect more details "in a couple of days."

Even if Unity does a complete 180 and reverses its per-install fee plans for its runtime, it may be too late for many current Unity team members. One of them is Garry Newman, the creator of Garry's Mod and the head of Facepunch, which made the popular survival game Rust with the Unity engine.

In a blog post, Newman stated that while the per-install fee policy would not affect Facepunch that much, he did state that was not the reason they were upset:

It hurts because we didn't agree to this. We used the engine because you pay up front and then ship your product. We weren't told this was going to happen. We weren't warned. We weren't consulted. We have spent 10 years making Rust on Unity's engine. We've paid them every year. And now they changed the rules.

Newman added that if a sequel to Rust is made, it "definitely won't be a Unity game."

Source

Table of contents:

1. Windows 11 news
2. Updates are available
3. Gaming news
4. Reviews are in
5. A blast from Microsoft's past
6. Random fact about Microsoft

Windows 11

Here we talk about everything happening around Microsoft's latest operating system in the Stable channel and preview builds: new features, removed features, controversies, bugs, interesting findings, and more. And of course, you may find a word or two about older but still supported versions.

This week, Microsoft released the latest Patch Tuesday updates with traditional fixes, improvements, and security patches. KB5030219 is now available for Windows 11 version 22H2; KB5030217 is for those still using Windows 11 version 21H2 (it will soon reach its end of life, by the way); and KB5030211 is here for Windows 10 users not amused with what Windows 11 has to offer so far.

We also got a major release on the unofficial side. NTDEV, the creator of the tiny10/11 project, released a new version of tiny11. Release 23H2 features gaming improvements and more unnecessary parts removed. However, despite carrying the "23H2" name, it is still based on Windows 11 version 22H2. You can download tiny11 version 23H2 x64 from the Internet Archive.

Finally, here are a few Windows quirks users discovered this week. Did you know Windows 11 have a dedicated shortcut key for launching Microsoft's LinkedIn? Hit Ctrl + Shift + Alt + Win + L and the OS will fire up your default browser and open LinkedIn. By the way, similar shortcuts exist for Office apps. Press Ctrl + Shift + Alt + Win + X to open Excel, Ctrl + Shift + Alt + Win + W to launch Word, and Ctrl + Shift + Alt + Win + P to open PowerPoint.

Windows Insider Program

Microsoft released plenty of new builds for each and every insider to test:

• Windows Server vNext: Build 25951.
• Canary Channel: Build 25951 with SMB NTLM blocking and Dialect Management.
• Dev Channel: Build 23545 with much-needed File Explorer performance improvements.
• Release Preview Channel: Build 22621.2359 (KB5020310) with a few new features coming down from the Beta and Dev Channel.
• Release Preview Channel: Build 22000.2479 (KB5030301) with various bug fixes.

As a reminder, builds older than 25931/23526 are now expired. That means those still running outdated releases will now get nagged with full-screen banners, prompting them to upgrade to something more recent.

Besides releasing new builds, Microsoft shipped two big updates for the Snipping Tool and Photos app. The company confirmed earlier reports about it adding new AI-powered features, namely OCR, Background Blur, content and location search, and Samsung Motion Photos support.

Updates are available

This section covers software, firmware, and other notable updates (released and coming soon) delivering new features, security fixes, improvements, patches, and more from Microsoft and third parties.

Microsoft updated a lot of Surface devices with new hardware this week:

• Surface Laptop 5 received network stability improvements and fixes for bugs causing problems when connecting to external monitors.
• Surface Pro X got an app performance boost, audio enhancements, fingerprint reliability improvements, and Wi-Fi stability fixes.
• Surface Pro 7 received charging stability improvements and fixes for a stuck Surface logo on boot.
• Surface Pro 9 5G was also updated with Windows Hello fixes and overall stability improvements.

The now-deceased Surface Duo is also worth mentioning since it got a new set of drivers for running Windows 11. Also, the DuoWOA team released ready-to-go/made dual boot images. Finally, speaking of Surface devices, a new report emerged this week with details about the pricing of the upcoming Surface devices.

On September 21, 2023, Microsoft will unveil the second-generation Surface Laptop Studio 2 with significantly upgraded hardware. According to leaks published on one German website, the computer will cost up to $1000 more than its predecessor. Oof!

This week was full of Microsoft Edge news. To begin with, Microsoft released version 117 with several new features, such as Edge for Business for non-managed devices, favorites recovery, and more. Those thinking the browser is getting too much bloat will be happy to learn that version 117 deprecates several niche features to simplify your browsing experience.

Alongside Edge 117 for Windows 10 and 11 (macOS and Linux too), developers pushed a small update for customers using Edge 109 on Windows 7, 8, and 8.1. There is nothing new but security fixes for a vulnerability exploiting a heap buffer overflow in WebP.

Microsoft is testing a neat feature update for its browser, namely an improved Split Screen. Users discovered that besides placing webpages side-by-side, preview versions of the browser now let you split tabs horizontally and keep one on top of the other.

You can change the Split Screen layout by selecting the corresponding option from the Split Screen menu. Here is how the new vertical layout looks like in Microsoft Edge Canary:

Shortly after that, Microsoft released Edge Dev 118 with the new Split Screen layout.

PowerToys, everyone's favorite tool for Windows 10 and 11, will soon get a neat update for the Run utility. Microsoft revealed a redesigned launcher with an acrylic background, reworked icons, improved fonts, and new features. The new Run should arrive in an upcoming PowerToys update, albeit the exact timeframe is a mystery.

That is not everything Microsoft has in store for PowerToys. Shortly after announcing the Run redesign, developers unveiled a reworked colour Picker with a fresh coat of paint better resembling Windows 11 UI.

Last week's Microsoft Weekly covered plenty of news about OneDrive, and this one has more to add. An update in the Microsoft 365 roadmap revealed the company's plans to add an offline mode for OneDrive for the web.

According to Microsoft, this feature will allow you to launch OneDrive in a browser and view, sort, rename, move, copy, and delete files without internet access. Additionally, for locally stored OneDrive files (marked as "always available offline"), you will be able to open and work on these in your browser, even if you are offline.

And here are other notable updates released this week:

• Auto Dark Mode received a massive update to version 10.4.0. Our favorite tool for toggling dark and light modes in Windows 11 and 10 now supports Windows Spotlight, system accent colour, and mouse cursors. Besides, the UI now matches Windows 11 design principles.
• Logitech Options+ got fixes for high CPU usage and buggy desktop switching. The latest release also added support for the recently announced Pebble Mouse 2 and Pebble Keys 2.
• Microsoft Designer Beta is now available on Android. Microsoft's latest AI-powered tool for generative design is now available in Beta on Android. You can download the app from the Google Play Store and preview new features before Microsoft pushes them out to the public.
• Office apps received August 2023 security patches. Those are available for customers using Office 2013, Office 2016, and Office 2019.

Finally, here are the latest drivers from NVIDIA, AMD, Intel, and other companies:

• NVIDIA 537.34 WHQL with optimizations for Mortal Kombat 1, Starfield, Counter-Strike: Global Offensive, and Lies of P. There is also support for six new G-SYNC-Compatible monitors from various manufacturers.
• Intel 31.0.101.4824 WHQL with optimizations for The Crew Motorfest and Mortal Kombat 1. In addition, the driver fixes flickering issues in Fortnite and random fan ramp ups.

On the gaming side

Learn about upcoming game releases, Xbox rumors, new hardware, software updates, freebies, deals, discounts and more.

Age of Empires II: Definitive Edition, released on Microsoft's latest consoles a couple of weeks ago, received a big update and a new DLC with two new campaigns. The Return of to Rome DLC lets you follow the creation of the Old Kingdom of Egypt and "command the mighty Carthaginian empire in their war against the aggressive Roman Empire." The update with a massive list of fixed bugs and two DLC campaigns is now available on PC and consoles. In case you did not know, Age of Empire II: Definitive Edition is available in the Xbox and PC Game Pass catalogs.

Starfield, the new hot topic in the gaming press, continues generating headlines. This time, users discovered that the sun only shines on those using NVIDIA cards (you can try it without buying one thanks to NVIDIA's GeForce Now service, where Starfield is now available). If you use an AMD or Intel card, chances are that you will not see any sun in the sky. The situation is extra ironic if you consider the ongoing allegation of Bethesda making the title favor the "red team's" GPUs.

However, gamers and enthusiasts are already fixing bugs and adding missing features (like DLSS) with mods. Mods are a big part of Bethesda games, but the official mod support will not arrive until 2024.

Colossal Order, the studio behind Cities: Skylines, is on track to release Cities: Skylines II on October 24, 2023. And with only one month before the release, developers published the final dev diary entry, detailing a new in-game camera and a dedicated photo mode. In addition to taking photos, players will be able to record cinematic videos to share their cities with the community. Cities: Skylines II will be available in Xbox and PC Game Pass on day one.

The next month will also bring us the next installment of Forza Motorsport, Microsoft's sim-oriented racing games. With less than one month to release, Turn 10 Studios published a lengthy gameplay video showcasing some new cars, tracks, visuals, sound design, and more. Besides, developers revealed details about AI and accessibility features that ensure everyone, even the visually impaired, can enjoy the next-generation Forza Motorsport. The game will arrive on Xbox and PC (including Xbox and PC Game Pass) on October 10, 2023.

Another highly anticipated game coming soon to PC and modern consoles is S.T.A.L.K.E.R 2. On September 7, 2023, the building hosting the developer teams caught fire and sustained major damage on one of its three floors. Fortunately, no one was injured. Although the damaged floor now requires a complete renovation, GSC Game World said no anomaly would stop them.

S.T.A.L.K.E.R 2 is expected in the first quarter of 2024, and we hope the accident will not cause any unscheduled delays.

Later this month, CD Projekt Red will release Cyberpunk 2077: Phantom Liberty and the studio wants to ensure gamers' PCs are in the right shape to run the new content. Filip Pierściński, the lead-scene programmer, published a post on X (Twitter), claiming the game will utilize "all what you have," with CPU loads reaching almost 100% on eight-core systems. Therefore, users better ensure their processor coolers can handle the heat.

By now, we are used to game developers recommending expensive hardware to run their titles (right, Todd Howard?), but seeing a studio recommending CPU cooler maintenance is something new.

Cyberpunk 2077 Phantom Liberty will be available on September 21, 2023. You can check out more details about the update and its new trailer here.

In a somewhat surprising move, Microsoft unveiled the Xbox Mastercard—a special credit card for those frequently buying games and digital goods in the Microsoft Store. Issued by Barclays, the card will let you earn points when making everyday purchases or purchasing digital goods. Items from the Microsoft Store and streaming services will get you up to 5x card points for every dollar you spend. Microsoft says the new card will be available to Xbox Insiders in the United States beginning September 21, 2023.

Another Xbox-related announcement from Microsoft this week is on the hardware side. The company unveiled the new Astral Purple Xbox Wireless Controller, which will be available for $64.99 on September 19, 2023. If you fancy a purple controller, preorder it now from the Microsoft Store.

On September 14, 2023, Microsoft launched Xbox Game Pass Core, the new subscription to replace the good-old Xbox Live Gold. Xbox Game Pass Core costs $9.99 per month and it offers 36 games from various studios in addition to providing Xbox owners with access to multiplayer. If you plan to subscribe to Xbox Game Pass Core, check out the full list of benefits here.

Speaking of Xbox Game Pass, more titles from its catalog are now playable on NVIDIA GeForce Now. This week, Starfield, Amnesia: The Bunker, Gears Tactics, and other games joined the ever-growing list of games available on GeForce Now.

A blast from Microsoft's past

John Callaham's weekly "Look back" series provides throwbacks into the past, detailing the company's products, partnerships, mishaps, and successes from years ago.

This week's look back article tells you a story about Microsoft Mira, a smart display tech Sony decided to cosplay with its PlayStation Portal. The idea behind Mira was to give users a "portable" 15-inch display to connect to their home PCs and perform Windows XP-era computing tasks from anywhere in the house. Only one year after the initial release, Microsoft pulled the plug on Mira, realizing that people just do not want or need such a device, especially with a hefty $1,000+ price tag.

Random fact about Microsoft

And here is a randomly selected piece of trivia about the company, Windows, and other Microsoft-made things.

Did you know that Microsoft Adventure, a game ported from Mainframe PDP-10 to Apple II in 1979 and later IBM PC, was one of the earliest games to feature a built-in Digital Rights Management system? Microsoft Adventure was released on a 5.25-inch floppy drive with a quirky system to prevent pirates from illegally copying the game.

Each disk with Microsoft Adventure had a randomized set of track numbers (a portion of the magnetic tape inside a floppy disk) to thwart disk copying software, usually expecting a standard set of tracks from 0 to 34. Throwing in misnumbered tracks (the drive had an instruction set on track 0 to tell the system how to access data) would cause a copying program to "think" the disk it tries to read is corrupt.

Source

Millions of people strap smartwatches—like the Apple Watch or Fitbit—to their wrists for their health benefits. Users can monitor their heart rate, count their steps, and gauge their activities. But according to some new research, these devices can also, and somewhat ironically, be teeming with disgusting bacteria.

As The Messenger reports, a paper published in Advances in Infectious Diseases found that roughly 95 percent of 20 wearables tested at random harbored some level of grime, from the skin-loving Staphylococcus (found on 85 percent) to the diarrhea-inducing Escherichia coli (60 percent). And let’s not ignore the impressive Pseudomonas (30 percent), which is good for a hearty lung infection.

A person’s habits may influence what kind of smartwatch bacteria they invite. Subjects who were frequent gym-goers tended to have the highest staph counts; people who handled animals had more E. coli.

Wearables that were made of rubber or plastic also had more germs as opposed to bands made of metal. Because the former are porous, they make for better breeding grounds. Bands made of gold or silver had practically no contamination.

But are these germs enough to actually make you ill? In a relatively healthy individual, perhaps not. But there’s still cause for concern. “Even at relatively low numbers these pathogens are of public health significance,” the study’s senior author Dr. Nwadiuto Esiobu, a professor in the department of biological sciences at Florida Atlantic University, said in a press release. “Importantly, the ability of many of these bacteria to significantly affect the health of immunocompromised hosts indicates a special need for health care workers and others in hospital environments to regularly sanitize these surfaces.”

The best way to combat smartwatch pathogens, unsurprisingly, is to regularly clean it. You’ll want to research the specific instructions for the best way to clean your individual wearable, but many experts recommend Lysol, Clorox, and/or alcohol wipes. Esiobu also recommends that other wearables, like earbuds, be studied for similar contamination.

Source

NEW YORK – America doesn’t have the factories or skilled labor to replace Chinese imports that support defense contractors and basic infrastructure, leaving the US economy vulnerable to harm in the event of an all-out trade war with China, corporate and government officials told Asia Times.

That’s why Biden administration officials are unlikely to heed calls from China hawks to completely cut off China’s semiconductor sector from US technology.

A group of 10 prominent House Republicans wrote to the US Commerce Department on September 14 demanding a shutoff of US exports of chip technology to China, claiming that the export controls imposed in October 2022 were ineffective.

The Republicans’ letter cited “recent reports that Huawei Technologies Co. (Huawei) has developed a smartphone containing 7-nanometer (nm) chips, capable of supporting 5G, produced by the Chinese state-owned Semiconductor Manufacturing International Corp. (SMIC).”

“We are extremely troubled and perplexed about the Bureau of Industry and Security’s (BIS) inability to effectively write and enforce export control rules against violators, especially China,” the letter added.

Semianalysis.com, a prominent chip industry website, declared, “US sanctions have failed. It called Huawei’s 7-nanometer chip “technically incredible,” and “a better designed chip than most in the West realize,” with capabilities similar to Nvidia’s and Qualcomm’s best AI processers. It noted the chip was produced with high yields, “with no access to cutting edge US intellectual property, and intentionally hampered.”

Huawei’s Mate60 Pro has a high-end chip. Photo: Sohu.com

Nothing short of a complete export ban on every category of semiconductor equipment would halt China’s progress, the website concluded. “Half measures will not work, but a full-scale assault will make it so the cost of replicating the semiconductor supply chain domestically is neigh on impossible. While we aren’t advocating for any of these specifically, it is clear the west can still stop China’s rise if decisive action is taken,” it wrote.

The US can’t stop China from making high-end chips like the new Kirin 9000 processor unless it shuts down all semiconductor fabrication in China. That would entail massive disruption not only of the semiconductor industry but of dozens of industries that depend on it, with grave economic consequences.

It is far from clear that the US would be able to enlist allies like Japan, South Korea, and the Netherlands. The Biden administration gave into Korean demands to maintain their existing chip fabs in China.

Holland’s ASML, the leading maker of chip-making lithography machines, won’t sell its most advanced equipment to China, but continues to sell the Deep Ultraviolet (DUV) devices that SMIC used to make the new Huawei chip.

Even if the US could persuade other countries to join in a total boycott of chip-making equipment to China, the US is not the only country that can wage economic war. The disruptive impact on the world economy would be incalculable and one possible upshot would be paralysis in US critical infrastructure.

Public discussion about possible Chinese retaliation against additional US export controls has focused on a possible ban on the use of Apple handsets by Chinese government officials. However, American vulnerability is evident in the form of thousands of critical components used in critical infrastructure and the US defense industry.

Graphic: Asia Times

The US imported US$33 billion in capital goods from China for electricity generation and distribution in 2022, items that are no longer manufactured in the US.

Substituting domestic production for these items would entail long lead times and exorbitant costs, industrial officials say. In the event of a full-scale trade war, a Chinese ban on critical components could cripple basic US infrastructure.

“The vulnerability of supply chains for critical infrastructure is acute and self-inflicted. The US and its allies have allowed themselves to become captive to Chinese cartels that control production of electronic components, high-powered magnets, printed circuit boards, computers, drones, rare earth metals, wind turbines, solar cells, cellular phones and lithium batteries… In fact, nearly every element of the technology-based digital smart grid is dependent on Chinese-made components,” Brien Sheahan, a former top US energy regulatory official, wrote in April.

US defense contractors also depend heavily on China. In a June 19 interview with the Financial Times, Raytheon CEO Greg Hayes said his company had “several thousand suppliers in China and decoupling is impossible. We can de-risk but not decouple,” adding that he believed this to be the case “for everybody” in US manufacturing.

Hayes added, “Think about the $500 billion of trade that goes from China to the US every year. More than 95% of rare earth materials or metals come from, or are processed in, China. There is no alternative. If we had to pull out of China, it would take us many, many years to re-establish that capability either domestically or in other friendly countries.”

Raytheon makes Tomahawk cruise missiles, Maverick air-to-surface missiles, Javelin anti-tank missiles and other mainstays of the American arsenal.

Meanwhile, American attempts at reducing dependence on foreign supply chains for critical goods have stalled. Taiwan’s TSMC, the world’s top chip fabricator, accepted $15 billion in cash subsidies and tax credits from the Biden administration to build a fab in Arizona, but shortages of skilled labor have delayed the plant until 2025.

The same bottlenecks would hamper US efforts to replace Chinese components in critical infrastructure. The US faces an air pocket in skilled labor supply during the next two years.

“The fact is that 22% of existing skilled manufacturing workers will be retiring by the end of 2025. This could result in as many as 2 million to 3.5 million unfilled manufacturing jobs by 2025,” according to the consulting firm HBK.

America’s capital stock of manufacturing equipment has remained stagnant since 2000, according to Federal Reserve calculations.

Graphic: Asia Times

US orders for manufacturing equipment have been steady at between $1.5 billion and $2 billion a month, roughly half the level before the 2008 recession.

Graphic: Asia Times

As the growth rate of manufacturing capital stock collapsed, America’s trade deficit surged.

Graphic: Asia Times

It’s true that China still depends on the West for a wide variety of chip-making equipment, while the US depends on China for a vast number of capital goods inputs. Both are capable of hurting each other badly.

The question is whether they will. Biden administration officials are keenly aware of American vulnerabilities, and reluctant to push the tech war against China to the point of Chinese retaliation.

Even with an all-out mobilization, it would take the US several years to build out sufficient flexible manufacturing capacity to replace critical Chinese components.

Source

Many people have wondered what the audience is for this device. You can only use it on a local home wireless network, and you can only play the games you have purchased on your PS5 console. You cannot play those games while on the road, and the Portal can't be used for connecting to cloud gaming services.

Looking back, the PlayStation Portal is a little like what Microsoft tried to launch over 20 years ago. At CES in January 2002, the company first announced plans to help launch a new portable and connected series of "Smart Display" for the home, using Microsoft's "Mira" technology.

Here is how Microsoft first described its plans in its own press release from January 7, 2002:

Harnessing the remote desktop and wireless networking features of Windows XP and Windows CE .NET, “Mira”-enabled smart displays will deliver to consumers the freedom of the complete experiences in Windows, including browsing the Web, sending or receiving e-mail messages, listening to music, and editing and displaying digital images, from any room in the home.

“Mira” will enable smart displays in a variety of instant-on and silent-running form factors ranging from a primary PC monitor that detaches to become a portable wireless touch-screen monitor to a large digital television that presents a complete Windows XP experience including music and photos from a PC.

The plan was for Microsoft to license the Mira Smart Display tech to third-party hardware companies, including Philips, NEC, Fujitsu, and TriGem, so they could launch these portable touchscreen products before the end of 2002, in time for the holiday shopping season. Unfortunately, the first of these products, the ViewSonic Airplane V150p, didn't make that release time frame, as it launched in early 2003.

The hardware specs for this device included a big 15-inch touchscreen with a resolution of 1,024 x 768. Inside, it included a 400 MHz Intel XScale processor, along with 64MB of RAM. The battery on board is said to have lasted about four hours on a single charge, and the display weighs about 6 pounds.

The problems with these kinds of Smart Displays quickly became evident with the release of this Viewsonic product. The biggest issue is that why would you use this when you could get a notebook that could connect to your home network and basically could do the same things at this "smart display"?

The initial price of the Viewsonic product was also pretty high at over $1,000. Again, you could buy a notebook for several hundred dollars less. Plus, the smart display couldn't be used without connecting it wirelessly to its host Windows XP PC, as opposed to a stand-alone notebook PC, which could work well on its own.

You also couldn't use multiple Smart Displays in your home; only one could connect to the host PC. Did we mention there was no support for streaming video from your host PC to the smart display?

In the end, most people realized that this Smart Display couldn't do a lot of basic functions as well or better than a standard Wi-Fi-enabled Windows notebook. The rise of touchscreen tablets several years later would prove to be the kind of device and platform that Microsoft's Mira technology was trying to reach.

Microsoft decided to shut down the development of Mira and Smart Display in January 2004, only a year after the launch of the ViewSonic device. While it was an early attempt to launch a big-screen tablet-like touchscreen device, there were simply too many limiting factors surrounding Microsoft's tech at that time.

Source

"Hello heisters, we want to inform you that Denuvo is no longer in PAYDAY 3," announced Starbreeze global brand director Almir Listo in an update. "We look forward to seeing you in New York City! Happy heisting!"

Keep in mind that Payday 3 will still remain an always-online title even with this change. Unlike previous entries in the series, players are no longer able to play solo in offline mode this time. A Starbreeze account and a constant internet connection is required for playing solo or in co-op.

Usage of Denuvo was first revealed in Payday 3's PC version back in July as the game's playtests were ongoing. The move caused an uproar within the community for including the tech that's widely believed by PC gamers to cause performance issues in games, though Irdeto denies this. Starbreeze has not revealed why was Denuvo removed from the title at the last minute, it could be to appease angry fans or for another reason altogether.

Denuvo recently introduced new pieces of anti-tamper technologies as well, aimed at new Unreal Engine titles. However, while the tech seems to be aimed at thwarting malicious actors, game modders may also be hit by the new protections if they are implemented.

Payday 3 launches September 21 on PC, Xbox Series X|S, and PlayStation 5 for $39.99. It will also be available across Microsoft's Xbox Game Pass subscriptions as a day-one release. Like with many games nowadays though, purchasing Payday 3's Silver or Gold special edition will let players jump in on September 18 to begin their heisting sessions early.

Source

In an Xbox Wire post, it was announced that UK-based Wushu Studios will continue to develop and support State of Decay 2, as most of Undead Labs works on the previously announced State of Decay 3. The first Wushu update will be released for free on September 18.

This update is adding what is being labeled as "curveballs" to State of Decay 2, and is meant for gamers who might find playing the game repetitive after a while. The blog post states:

Housed in a new menu, curveballs will periodically appear – each with their own narrative attached – offering new objectives, rewards, enemy types and much more. Curveballs can be positive (like finding increased loot) or negative (like buffing regular zombies in multiple ways); they can be timed or objective based; they can affect enemies, your survivors, or other NPCs. Effectively, any element of the game you may have gotten comfortable with can, and often will, change.

Those changes will also be extended to the game's zombies as well. "Curveballs" could cause the undead to generate stronger attacks, have improvements in eyesight and hearing, and could even cause them to explode or "carry a stench that suffocates and drains the stamina of any survivor up close." The curveballs could also weaken zombies as well.

Even as State of Decay 2 is getting updated, Undead Labs is still working on the next game in the series, State of Decay 3. It was first announced in 2020, but since then there's been little to no info on the game's status.

Source

What you need to know

• For a couple months, Undead Labs and Wushu Studios have been teasing and testing the next State of Decay 2 update.
• Update 34: Curveball is officially coming to all players across Xbox and Windows PC on Sept. 18, 2023.
• The update introduces Curveballs, a wide variety of random events that can have drastic effects on almost every aspect of SoD2's gameplay.
• Wushu Studios will continue to support State of Decay 2 with new content moving forward, as most of Undead Labs is now working on State of Decay 3.

I'm not shy about professing my love for State of Decay 2. The game launched in a rough state, but Undead Labs has been overwhelmingly persistent about updating the game with endless new content, quality-of-life improvements, and community-requested features. Over five years after release, SoD2 is still breathing, with its next brand-new content update coming... Next week.

Except, it doesn't hail from Undead Labs, as most of that studio is busy working on State of Decay 3. Instead, the Xbox Game Studio tagged Wushu Studio to develop State of Decay 2's next update, and the duo have been teasing and testing that update for the last two months. Now, we have a lot more information — and a release date — on State of Decay 2 Update 34: Curveball. As the name suggests, it's a doozy of an update.

You can watch the extended stream above with members from Wushu Studio for more information on State of Decay 2 Update 34: Curveball, or you can check out this Xbox Wire post. Either way, I'll be summarizing all the changes and features you should be aware of, if you're a State of Decay fan like I am.

• Curveballs are semi-random, dynamic events. State of Decay 2 now features Curveballs, which are highly diverse and dynamic events that can affect practically every aspect of SoD2's gameplay and world.
  • Curveballs can begin happening after five in-game days
  • Curveballs can happen as a direct result of player action, or can occur entirely unexpectedly
  • Multiple Curveballs can be in effect at any one time
  • Players can track current, upcoming, and ended Curveballs from a new dedicated menu
• Curveballs dramatically change how the game plays. A common community feedback was that State of Decay 2 could grow repetitive, especially in the late game. Curveballs directly target that critique to make the game drastically more unpredictable, exciting, and potentially dangerous even for long-term veterans.
  • Curveballs can affect almost anything, including your community and settlement, resources and loot, NPCs, characters, enemies, world events, and more
  • Curveballs are often attached to new narratives, helping build out community stories
  • They can be objective or time based, and may require subtle or major shifts in strategies to overcome
  • Curveballs don't have to be negative though, with many offering positive boons or active buffs
• There are countless modifiers, even just for zombies. Curveballs are so varied, with so many potential modifiers and effects, that no two games will ever be the same again.
  • Mutational outbreaks can occur in zombies as a result of in-world events or because of Plague Hearts actively fighting back
  • These mutational outbreaks can affect any of the zombie types, including Freaks
  • Negative effects may include major physiological changes like faster movement or stronger attacks, heightened senses like improved vision, hearing, and even smell, special effects like exploding in bile gunk after being killed or giving off toxic fumes that drain survivor stamina, and more
  • Positive effects are possible, too, like making zombies weaker, slower, easier to behead, or more clumsy
• Players can choose to play with Curveballs. This update isn't meant to make State of Decay 2 feel like a completely different game, it's just meant to complement existing gameplay and make it more dynamic. Still, players can absolutely opt to not play with Curveballs and continue to play the game exactly as it was.
• The feature isn't done yet. Wushu Studios has already said it is looking to improve this feature after its release to give players even more control over it.

 

State of Decay 2 was already amazing, but it's about to get a lot more exciting. (Image credit: Xbox Game Studios)

 

Curveballs can mean positive, narrative-focused events that help your community and affect your survivors. (Image credit: Xbox Game Studios)

 

Of course, there are also negative effects, like making Juggernauts stronger and more perceptive... Because they needed that. (Image credit: Xbox Game Studios)

Yeah, that's a lot. It's still not the end for State of Decay 2, either, as Wushu Studios is now going to continue supporting SoD2 in the same way that Undead Labs have been since the game's launch. Wushu has plans for a lot more content and features to come, but it's also going to address community feedback on existing features and systems, even in cases when it may require overhauling. Over five years later, State of Decay 2 is still going strong and shows no signs of stopping.

Of course, it helps that State of Decay 2 still has an active and passionate player base, which makes sense given how good the game has become. If you've yet to be a part of the community (or if it has been a while), this is an excellent time to play. State of Decay 2's Update 34: Curveball officially releases for Xbox Series X|S, Xbox One, and Windows PC on Sept. 18, 2023.

Looking for the best way to play one of the best Xbox games? State of Decay 2: Juggernaut Edition is never expensive, it's often on sale, and it's an Xbox Play Anywhere title. SoD2 is a permanent member of the Xbox Game Pass library on console and PC, too, and through Xbox Cloud Gaming. In fact, it's one of the games included in the new Xbox Game Pass Core subscription. There are so many ways to play this excellent game, join the community, and enjoy its healthy post-launch support.

Source

Source

It's the Razer Viper V3 HyperSpeed wireless gaming mouse, and it offers a lot of high-end technology in a smaller and lightweight form factor. The mouse is also available for a price that, while not super cheap, is more affordable than many other Razer mice products, or ones made by its competitors.

The Viper V3 HyperSpeed weighs just 82 grams with its AA battery, which can power it up to 280 hours before it needs to be replaced. It includes a 30K DPI optical sensor for fast responses in games, and there's a DPI button on top that can be used to quickly switch from five different settings.

The mouse was designed in collaboration with eSports pros and was optimized for players that prefer claw or fingertip grip styles while using a gaming mouse.

The Razer Viper V3 HyperSpeed includes six programmable buttons, and they use mechanical mouse switches that the company claims will last up to 60 million button presses.

The mouse comes with a 2.4Hz dongle that supports Razer's HyperSpeed wireless technology, with a 1,000hz polling rate that offers low latency while playing games. You can also purchase the mouse with an optional Razer HyperPolling wireless dongle and cable which increases the polling rate from 1,000Hz to 4,000Hz (although at the expense of battery life, as it gets cut down to 75 hours).

The Razer Viper V3 HyperSpeed wireless gaming mouse is currently available at Amazon on its own for $69.99. If you want to get the mouse with the optional HyperPolling wireless dongle, it's available on Razer's site for $84.99.

As an Amazon Associate when you purchase through links on our site, we earn from qualifying purchases.

Source

Source

Source

Source

Source:  Columbia University

Summary:  The era of artificial-intelligence chatbots that seem to understand and use language the way we humans do has begun. Under the hood, these chatbots use large language models, a particular kind of neural network. But a new study shows that large language models remain vulnerable to mistaking nonsense for natural language. To a team of researchers, it's a flaw that might point toward ways to improve chatbot performance and help reveal how humans process language.

The era of artificial-intelligence chatbots that seem to understand and use language the way we humans do has begun. Under the hood, these chatbots use large language models, a particular kind of neural network. But a new study shows that large language models remain vulnerable to mistaking nonsense for natural language. To a team of researchers at Columbia University, it's a flaw that might point toward ways to improve chatbot performance and help reveal how humans process language.

In a paper published online today in Nature Machine Intelligence, the scientists describe how they challenged nine different language models with hundreds of pairs of sentences. For each pair, people who participated in the study picked which of the two sentences they thought was more natural, meaning that it was more likely to be read or heard in everyday life. The researchers then tested the models to see if they would rate each sentence pair the same way the humans had.

In head-to-head tests, more sophisticated AIs based on what researchers refer to as transformer neural networks tended to perform better than simpler recurrent neural network models and statistical models that just tally the frequency of word pairs found on the internet or in online databases. But all the models made mistakes, sometimes choosing sentences that sound like nonsense to a human ear.

"That some of the large language models perform as well as they do suggests that they capture something important that the simpler models are missing," said Dr. Nikolaus Kriegeskorte, PhD, a principal investigator at Columbia's Zuckerman Institute and a coauthor on the paper. "That even the best models we studied still can be fooled by nonsense sentences shows that their computations are missing something about the way humans process language."

Consider the following sentence pair that both human participants and the AI's assessed in the study:

That is the narrative we have been sold.

This is the week you have been dying.

People given these sentences in the study judged the first sentence as more likely to be encountered than the second. But according to BERT, one of the better models, the second sentence is more natural. GPT-2, perhaps the most widely known model, correctly identified the first sentence as more natural, matching the human judgments.

"Every model exhibited blind spots, labeling some sentences as meaningful that human participants thought were gibberish," said senior author Christopher Baldassano, PhD, an assistant professor of psychology at Columbia. "That should give us pause about the extent to which we want AI systems making important decisions, at least for now."

The good but imperfect performance of many models is one of the study results that most intrigues Dr. Kriegeskorte. "Understanding why that gap exists and why some models outperform others can drive progress with language models," he said.

Another key question for the research team is whether the computations in AI chatbots can inspire new scientific questions and hypotheses that could guide neuroscientists toward a better understanding of human brains. Might the ways these chatbots work point to something about the circuitry of our brains?

Further analysis of the strengths and flaws of various chatbots and their underlying algorithms could help answer that question.

"Ultimately, we are interested in understanding how people think," said Tal Golan, PhD, the paper's corresponding author who this year segued from a postdoctoral position at Columbia's Zuckerman Institute to set up his own lab at Ben-Gurion University of the Negev in Israel. "These AI tools are increasingly powerful but they process language differently from the way we do. Comparing their language understanding to ours gives us a new approach to thinking about how we think."

Source

A British judge has admitted using ChatGPT to write part of a case ruling.

Lord Justice Birss said he called on the "jolly useful" AI chatbot to summarise an area of law he wasn't familiar with, and then copy and pasted its response.

The Court of Appeal judge, who specialises in intellectual property law, said such tools had "great potential" to help people become more efficient at work.

Speaking at a law conference, reported by The Law Society Gazette and the Telegraph, Lord Justice Birss said: "I asked ChatGPT can you give me a summary of this area of law, and it gave me a paragraph.

"I know what the answer is because I was about to write a paragraph that said that.

"But it did it for me and I put it in my judgment.

"It's there and it's jolly useful."

However, the judge stressed the importance of not relying on AI for topics on which you know nothing about.

ChatGPT and similar tools have proved prone to something known as hallucination - or making stuff up.

Earlier this year, two New York lawyers were fined over a legal brief with fake case citations generated by ChatGPT.

They had used it to find legal precedents supporting a client's personal injury case against an airline, but many were shown up in court as being completely fake.

Lord Justice Birss said he felt confident to use it because he could recognise its answer "as being acceptable".

"I'm taking full personal responsibility for what I put in my judgment, I am not trying to give the responsibility to somebody else," he added.

Sky News has contacted the Judicial Office for comment.

Source

Source

What you need to know

• A seemingly AI-generated obituary for Brandon Hunter, a 42-year-old former NBA player who passed away unexpectedly, has gone viral after it referred to the recently deceased as 'useless.'
• The original article was posted by an unnamed "Editor" on a sports news site called "Race Track," which has since been deleted.
• The distasteful obituary gained attention after being distributed on Microsoft Start, a personalized news feed, and the MSN news portal aggregating content from third parties.
• MSN recently featured another AI-generated article recommending users visit a food bank in Ottawa as a tourist attraction.

Microsoft users are scrutinizing the company's relationship with AI-generated content after a second snafu involving the company's personalized news feed, Microsoft Start, and the MSN news portal. Initially reported by Futurism, an article republished via MSN has gone viral for potentially using AI to write an obituary that called the recently deceased 'useless.'  The offensive article was spread on social media following its appearance on Microsoft Start, where users pondered whether it resulted from AI or a translation error mixed with plagiarism that resulted in the final piece.

A self-described sports news website called Race Track published the problematic obituary for Brandon Hunter, a former NBA player who passed away unexpectedly at the age of 42. The article included the headline "Brandon Hunter useless at 42".  Race Track's profile on Microsoft Start, which was previously Microsoft's MSN portal, included a link to a website called Autogear, which was deleted once the article about Hunter went viral.

The post on Race Track, attributed only to the "Editor," made multiple egregious errors in its obituary of Hunter and the distasteful headline. It described Hunter as having "performed in 67 video games over two seasons" during his NBA career and referred to his passing as being "handed away on [sic] the age of 42". Along with being error-laden, the article on Race Track appeared to be heavily plagiarized from an obituary published by TMZ Sports.

In August 2023, Microsoft came under fire when a travel guide appeared on the aggregated MSN news portal recommending a local food bank as a tourist attraction in Ottawa, Canada. Despite claims by MSN that the content aggregated for Microsoft Start and the news portal are from 'premium' partners that supposedly align with Microsoft's values, Microsoft fired its team of MSN journalists in 2020, just before the Microsoft Start rebrand. Microsoft blamed a human for the error, anyway.

Windows Central's take

Microsoft is only one of many content publishers that have decided to dive head-first into generative AI technology to replace human-produced content. This process has many issues, including low-quality content that has not been vetted for factual accuracy or just flat-out plagiarizes its sources. As more AI content replaces human-vetted pieces, it will lead to a content loop where misinformation produced by AI is used to train additional AI that produces more misinformation.

MSN and Microsoft Start have pushed poorly composed and factually inaccurate articles about everything from Bigfoot to anti-vaccine conspiracies from AI-generated sources without vetting. Microsoft shakes off these continued mishaps from AI content as little more than mild embarrassments. They should be regarded more as obvious red flags for the overuse of undercooked tech that will have serious real-world implications the more it is implemented.

Source

With higher download speeds, internet users in the UK can download larger files quicker including content such as games, films, apps, and streaming high-resolution content. Ofcom said that the percentage of people receiving at least 30 Mbit/s had increased to 88% compared to 83% the year before.

Those with actual download speeds of less than 10 Mbit/s represented 3% of broadband customers, down from 4% the year before and less than 1% of lines were connected to broadband services that actually have advertised download speeds of less than 10 Mbit/s.

Another important metric to look at is the upload throughput as this is the speed at which you can upload or send files to other people or online platforms. Ofcom revealed that there was a large increase in upload speed, rising to 18.4 Mbit/s in March 2023 – up 7.8 Mbit/s year-over-year.

One of the big issues the UK has been tackling, especially under former Prime Minister Boris Johnson’s levelling up campaign, was the issue of rural broadband and getting people in hard-to-reach areas connected. According to Ofcom, the average download speed in urban areas was 70.3 Mbit/s while in rural areas it was 56.0 Mbit/s. That’s a gap of 26%, an improvement over 2022 when the gap was a huge 58%.

Finally, Ofcom said that the network congestion issue had slightly improved over the last year. It said that during peak time from 8-10 p.m., broadband speeds were operating at 95% of their full speed compared to 94% in 2022. This is good news for anyone working 9-to-5 who goes online to relax during peak hours.

Source: Ofcom

Source

CD Projekt Red announced that Cyberpunk 2077's 2.0 patch will be released on September 21, bringing significant changes to the game. While the highly anticipated Phantom Liberty expansion lands just a few days later on September 26, many of the major new systems and features introduced in the paid DLC will also arrive in the base game free of charge.

The 2.0 update represents a genuine overhaul of how Cyberpunk 2077 is played, according to the developer. It brings fundamental adjustments and entirely new mechanics to improve all aspects of the open-world RPG experience.

Exact details are still scarce, but it seems that both new and returning players will discover an improved Night City with this major update. It must be a big update for the visuals, as Filip Pierściński, the game's lead scene programmer, advised players to "check the state of your cooling systems" before the expansion.

• Redesigned skill trees and perks
• Revamped cyberware and new capacity system
• Vehicle combat and car chases
• Combat AI improvements
• New police system
• UI and UX improvements
• Loot, items, and crafting changes
• New radio stations
• And much more

The update will be released for the game's PC, Xbox Series S/X and Playstation 5 platforms.

CD Project Red alsoreleased a new cinematic trailer for Phantom Liberty. One new addition players can look forward to in the spy-themed Phantom Liberty expansion is actor Idris Elba as solo Solomon Reed.

Set after the main story's events, it will entirely change how the game is played, according to CD Projekt Red. This expansion promises to take V's story in exciting new directions with an overhaul of systems and features.

The Phantom Liberty expansion will cost $29.99 and will also launch for th PC, Xbox Series S/X and the PlayStation 5.

Source

In 2020, Lenovo released the ThinkPad X1 Fold PC, but our review was mixed at best. The device had a lot of potential, but its low performance and its high price tag (starting at $2,499) were definitely not in its favor.

Today, HP announced the launch of its first foldable PC, and while it looks like an impressive product at first glance, the price for this device is pretty eye-watering.

The HP Spectre Foldable PC is being promoted as a 3-in-1 product. It can be used as a stand-alone 17-inch touchscreen tablet. It can also be used as a 17-inch desktop PC with its built-in kickstand and its included detachable Bluetooth keyboard. Finally, the PC can be turned into a 12.3-inch laptop again with the included detachable keyboard.

Here are the hardware specs for the HP Spectre Foldable PC:

• CPU: Intel Core i7-1250U (up to 4.7 GHz with Intel Turbo Boost Technology, 12 MB L3 cache, 10 cores, 12 threads); Intel Evo
• Memory: 16 GB LPDDR5-5200 MHz RAM (onboard)
• GPU: Intel Iris Xe Graphics
• OS: Windows 11 Home/Pro
• Storage: 1 TB PCIe Gen4 NVMe TLC M.2 SSD
• Wireless: Intel Wi-Fi 6E AX211 (2x2) and Bluetooth 5.3 wireless card
• Display: 3:4 43.2 cm (17 inch) diagonal, 2.5K (1920 x 2560), foldable OLED, touch, UWVA, BrightView, Low Blue Light, SDR 400 nits, HDR 500 nits, 99.5% DCI-P3
• Keyboard: Full-size, detachable rudy black soft cover Bluetooth keyboard
• Battery: 6-cell, 94.3 Wh Li-ion polymer; Up to 12:30 battery life (MM25 and Wireless Video
• Streaming Netflix)
• Audio: Bang & Olufsen Quad speakers
• Ports: Two Thunderbolt 4 ports with USB Type-C 40Gbps signaling rate (USB Power Delivery, DisplayPort 1.4, HP Sleep and Charge)
• Weight: 1354 g (SSD, without keyboard); 1624 g (SSD, with keyboard)
• Dimensions: 277.05 x 376.1 x 8.5 mm (Unfolded); 277.05 x 191.31 x 21.4 mm (Folded)

Last week, we got a virtual presentation of the HP Spectre Foldable PC. There's no doubt that a lot of impressive engineering went into the making of this product. The 17-inch foldable OLED display includes an integrated kickstand that allows it to stand on its own so you can work with it like a desktop PC with the included keyboard.

The display is also just 8.5mm thick or about as thick as a number 2 pencil. Its 1.35kg weight (just under 3 pounds) makes it very light, and with its foldable design, even more portable.

You can also use this display as a really big Windows 11 tablet for tasks like note-taking with its included stylus, along with art and presentations.

However, it's likely that it will be used most as a 12.3-inch laptop. You can place the Bluetooth keyboard on top of one of the two folded sections of the display, or you can use it detached from the display, with the bottom half serving as a kind of second 12.3-inch screen for some heavy multitasking.

The HP Spectre Foldable PC will be available to pre-order today in the US at Best Buy's website for a whopping $4,999.99. It will be available on HP's website sometime in October. It will also be available in other countries later this fall.

No, that price you just saw is not a typo, and that may be the biggest barrier for this product. The HP Spectre Foldable PC may have a ton of potential to be a multi-purpose computer, but the price is the same as a high-end desktop gaming PC with hardware specs that would blow away this product.

HP should be commended on going outside the box, but this is also clearly a first-gen product that we believe will be refined in later versions that will hopefully be priced much lower.

Source

Japan is building its own versions of ChatGPT — the artificial intelligence (AI) chatbot made by US firm OpenAI that became a worldwide sensation after it was unveiled just under a year ago.

The Japanese government and big technology firms such as NEC, Fujitsu and SoftBank are sinking hundreds of millions of dollars into creating AI systems that are based on the same underlying technology, known as large language models (LLMs), but that use the Japanese language, rather than translations of the English version.

“Current public LLMs, such as GPT, excel in English, but often fall short in Japanese due to differences in the alphabet system, limited data and other factors,” says Keisuke Sakaguchi, a researcher at Tohoku University in Japan who specializes in natural language processing.

English bias

LLMs typically use huge amounts of data from publicly available sources to learn the patterns of natural speech and prose. They are trained to predict the next word on the basis of previous words in a piece of text. The vast majority of the text that ChatGPT’s previous model, GPT-3, was trained on was in English.

ChatGPT’s eerie ability to hold human-like conversations, has both delighted and concerned researchers. Some see it as a potential labour-saving tool; others worry that it could be used fabricate scientific papers or data.

In Japan, there’s a concern that AI systems trained on data sets in other languages cannot grasp the intricacies of Japan’s language and culture. The structure of sentences in Japanese is completely different from English. ChatGPT must therefore translate a Japanese query into English, find the answer and then translate the response back into Japanese.

Whereas English has just 26 letters, written Japanese consists of two sets of 48 basic characters, plus 2,136 regularly used Chinese characters, or kanji. Most kanji have two or more pronunciations, and a further 50,000 or so rarely used kanji exist. Given that complexity, it is not surprising that ChatGPT can stumble with the language.

In Japanese, ChatGPT “sometimes generates extremely rare characters that most people have never seen before, and weird unknown words result”, says Sakaguchi.

Cultural norms

For an LLM to be useful and even commercially viable, it needs to accurately reflect cultural practices as well as language. If ChatGPT is prompted to write a job-application e-mail in Japanese, for instance, it might omit standard expressions of politeness, and look like an obvious translation from English.

To gauge how sensitive LLMs are to Japanese culture, a group of researchers launched Rakuda, a ranking of how well LLMs can answer open-ended questions on Japanese topics. Rakuda co-founder Sam Passaglia and his colleagues asked ChatGPT to compare the fluidity and cultural appropriateness of answers to standard prompts. Their use of the tool to rank the results was based on a preprint published in June that showed that GPT-4 agrees with human reviewers 87% of the time1. The best open-source Japanese LLM ranks fourth on Rakuda, while in first place, perhaps unsurprisingly given that it is also the judge of the competition, is GPT-4.

“Certainly Japanese LLMs are getting much better, but they are far behind GPT-4,” says Passaglia, a physicist at the University of Tokyo who studies Japanese language models. But there is no reason in principle, he says, that a Japanese LLM couldn’t equal or surpass GPT-4 in future. “This is not technically insurmountable, but just a question of resources.”

One large effort to create a Japanese LLM is using the Japanese supercomputer Fugaku, one of the world’s fastest, training it mainly on Japanese-language input. Backed by the Tokyo Institute of Technology, Tohoku University, Fujitsu and the government-funded RIKEN group of research centres, the resulting LLM is expected to be released next year. It will join other open-source LLMs in making its code available to all users, unlike GPT-4 and other proprietary models. According to Sakaguchi, who is involved in the project, the team hopes to give it at least 30 billion parameters, which are values that influence its output and can serve as a yardstick for its size.

However, the Fugaku LLM might be succeded by an even larger one. Japan’s Ministry of Education, Culture, Sports, Science and Technology is funding the creation of a Japanese AI program tuned to scientific needs that will generate scientific hypotheses by learning from published research, speeding up identification of targets for enquiry. The model could start off at 100 billion parameters, which would be just over half the size of GPT-3, and would be expanded over time.

“We hope to dramatically accelerate the scientific research cycle and expand the search space,” Makoto Taiji, deputy director at RIKEN Center for Biosystems Dynamics Research, says of the project. The LLM could cost at least ¥30 billion (US$204 million) to develop and is expected to be publicly released in 2031.

Expanding capabilities

Other Japanese companies are already commercializing, or planning to commercialize, their own LLM technologies. Supercomputer maker NEC began using its generative AI based on Japanese language in May, and claims it reduces the time required to create internal reports by 50% and internal software source code by 80%. In July, the company began offering customizable generative AI services to customers.

Masafumi Oyamada, senior principal researcher at NEC Data Science Laboratories, says that it can be used “in a wide range of industries, such as finance, transportation and logistics, distribution and manufacturing”. He adds that researchers could put it to work writing code, helping to write and edit papers and surveying existing published papers, among other tasks.

Japanese telecommunications firm SoftBank, meanwhile, is investing some ¥20 billion into generative AI trained on Japanese text and plans to launch its own LLM next year. Softbank, which has 40 million customers and a partnership with OpenAI investor Microsoft, says it aims to help companies digitize their businesses and increase productivity. SoftBank expects that its LLM will be used by universities, research institutions and other organizations.

Meanwhile, Japanese researchers hope that a precise, effective and made-in-Japan AI chatbot could help to accelerate science and bridge the gap between Japan and the rest of the world.

“If a Japanese version of ChatGPT can be made accurate, it is expected to bring better results for people who want to learn Japanese or conduct research on Japan,” says Shotaro Kinoshita, a researcher in medical technology at the Keio University School of Medicine in Tokyo. “As a result, there may be a positive impact on international joint research.”

doi: https://doi.org/10.1038/d41586-023-02868-z

Source

Modern cars are often described as “computers on wheels.” They come with automated driver assistance systems, large display screens, Internet connections and a multitude of ways to sync with smartphones.

Yet in the rush to innovate and one-up competitors with ever newer technology, things may have gone too far. Some developments have made driving safer, but others veer toward tech excess that can actually harm drivers. Cars in the current generation can be pricier to repair, harder to understand and operate and, some experts in the field say, more likely to cause distraction and driver disengagement.

And a report released last week now suggests that some new car tech also poses a major threat to data privacy. The new report was issued by the Mozilla Foundation, an Internet-focused nonprofit group, and concludes that when it comes to handling users’ personal information, today’s high-tech cars are the worst category of consumer products the organization has ever reviewed. Cars “have evolved into a privacy nightmare,” says Jen Caltrider, director of Mozilla’s privacy reporting program. Because they incorporate motion and pressure sensors, GPS, cameras, microphones and smartphone connectivity, modern vehicles can collect an astounding range of information on their drivers and passengers. The researchers assessed 25 popular brands’ privacy policies and found that 21 of these allow car companies to share or sell customer data with external service providers, data brokers and other businesses. Privacy policies from two brands—Kia and Nissan—even include a clause that notes that each company may collect and disclose data on users’ sexual orientation or sexual behavior.

It’s unclear if or how these automakers might be doing such things. “Kia does not and has never collected ‘sex life or sexual orientation’ information from vehicles or consumers in the context of providing the Kia Connect Services,” says Kia spokesperson James Bell. He adds that the brand includes the category in its privacy policy to define “sensitive personal information” under the California Consumer Privacy Act. Nissan spokesperson Brian Brockman similarly says, “Nissan does not knowingly collect or disclose consumer information on sexual activity or sexual orientation,” and adds that state laws such as California’s require the company to disclose inadvertent information that might be inferred from other data such as location tracking.

Still, permissive policies are worrying regardless of how they’re currently applied, Caltrider says. Such information could be used against vulnerable people. For example, multiple vehicle brands’ policies advise that they might share customer data with U.S. law enforcement or government officials, even without warrants. In parts of the country with restrictive health care laws, those tracking data could be used as legal evidence against people seeking an abortion or gender-affirming care. Such sensitive customer data are also a target for illegal attacks: 17 out of the 25 car companies experienced some form of data leak, breach or hack in the past three years, according to the Mozilla report. A lack of data privacy is one of those things that often doesn’t seem like a problem until it is, Caltrider says. “It might not be impacting you now. But you might also not realize when suddenly it is, and you don’t have any choice or control over it,” she adds. Change will likely only come at the policy level, Caltrider says. There is no federal data privacy law, however, and only a handful of states grant residents the right to have their data deleted on request.

Privacy isn’t the only thing drivers need to keep an eye on. The Mozilla Foundation report’s findings are just the latest in a series of concerns that experts have voiced about the technology in our personal vehicles.

The true cost of high-tech cars is multifaceted, but one component comes down to dollars and cents in a very direct way. Repairing cars that come with advanced driver-assistance technology or other intensively computerized features can be up to three times as expensive as similar maintenance on vehicles without such tech, says Greg Brannon, director of automotive engineering and industry relations at the American Automobile Association (AAA). Additionally, not all shops are equipped to repair such vehicles. Windshield replacement, for instance, “used to be a pretty straightforward operation,” Brannon says. Many vehicles, however, now require specialized glass that allows the internal cameras to function.

In these cars, the cameras must be recalibrated after windshield installation, adding another step to the process. These cameras do enable some driver-assistance features with a clear safety benefit: automated emergency braking, which has been tested and refined over a decade, is now the industry standard on all new vehicles and can significantly reduce crash rates, Brannon says. “But there is a cost to that,” he adds, “and particularly a cost that comes with the repair of that vehicle.”

Beyond higher repair costs, drivers often end up dissatisfied with the high-tech features themselves and report frequent problems with components such as complicated door handles, infotainment interfaces and wireless phone chargers. In Brannon’s view, time will likely help defray some of the added expense and difficulties to both drivers and repair shops as these sorts of vehicles become more common. Plus, it’s hard to put a price on added safety. But Brannon contends that not all new features and vehicle designs are boosting overall safety.

“There is often a gap between what is marketed in those systems, both in the naming and in advertisements, and what the systems actually do in real life,” he says. “The gap between those two things makes it a very dangerous situation for someone [who] does not take the time to understand the system.” Brannon gives the example of partially automated features such as adaptive cruise control and lane-keeping assistance. Both require a driver to ultimately control a vehicle and to always watch the road. Yet Brannon says these features might be labeled with terms such as “traffic-aware cruise control” or “active steering assist,” which can imply that a human can leave much of the job up to the car. In the worst case, there’s Tesla’s suite of assistance features that are labeled “autopilot” and “full self-driving”—despite the features not actually functioning in accordance with those names. Last year California passed a law attempting to force Tesla to adjust these names to be less misleading. Tesla did not respond to repeated e-mails seeking comment for this article.

New car owners rarely receive training on when and how to properly deploy these partially automated features. As a result, “a lot of drivers and other road users don’t really have a good grasp of what this technology is for, how it works and its limitations,” says William Horrey, technical director of the AAA Foundation for Traffic Safety, AAA’s nonprofit arm. Some of these features are meant to be used only in certain scenarios, such as on highways and not in more complex city environments, for example.

That lack of training can have unintended consequences, Horrey says. As driver assistance advances, some studies suggest human drivers will become more disengaged. When people believe their cars can conduct much of the driving process automatically, they’re more likely to check out or multitask—and potentially cause accidents—both Horrey and Brannon say.

Other features, too, can disrupt safety via poor design. The biggest culprits, according to some experts, are the sprawling, multilayered, touch-screen-based interfaces that have taken over new vehicle consoles. In recent years these displays have grown larger and incorporated more components, says David Strayer, an applied cognitive psychologist at the University of Utah, who studies driver behavior. “There’s more information for the driver to look at,” Strayer explains—and with more information comes more distraction. Any glances away from the road that last longer than two seconds elevate crash risk, he says. Many cars allow drivers to input GPS directions or dictate texts and complex commands via Bluetooth or to navigate a touch screen to manage basic functions while the vehicle is in motion. Some consoles even display video. All of these features can absorb a driver’s attention for far longer than is safe.

Everyone generally knows that texting while driving is a bad idea. With built-in features, however, drivers tend to think, “If it’s in the car, it must have been put there because it’s safe, it’s been fully vetted and it’s going to work,” Strayer says. But that’s often not the case, he notes. “We have not kept safety as our priority,” Strayer adds.

Brannon agrees. “There is a point where you cross the threshold of benefit and cross into detriment. There’s no question about it,” he says. Some industry research has demonstrated that old-fashioned tactile knobs and buttons can be a much safer, simpler and easier to fix alternative to touch screens and voice commands. But going back to buttons would mean admitting that ever advancing technology isn’t always the best fix.

Source

You may remember that Bethesda made a statement saying that the Arc A770 was not compatible with Starfield, despite the GPU outperforming the RTX 3060 Ti and RX 6700 XT. Some gamers felt that Bethesda was unwilling to troubleshoot issues specific to Intel hardware, as the A770 is the company's highest-end current Arc GPU.

Anyway, on to the release notes of this driver:

GAMING HIGHLIGHTS:

Intel® Game On Driver support on Intel® Arc™ A-series Graphics for:

• The Crew Motorfest*

• Mortal Kombat 1*

FIXED ISSUES:

Intel® Arc™ Graphics Products:

• Fortnite* (DX12) may exhibit flickering corruptions during gameplay.

• Device fan may ramp up frequently on certain Intel® Arc™ Graphics products.

Intel® Core™ Processor Products:

• Tactics Ogre Reborn* (DX11) may exhibit flickering corruption during dialogue sequences

KNOWN ISSUES:

Intel® Arc™ Graphics Products:

• Starfield* (DX12) may experience sporadic instability in some areas of the game.

• Starfield* (DX12) may exhibit corruption when using Dynamic Resolution Scaling. A workaround is to change the Render Resolution Scale slider value.

• Starfield* (DX12) may exhibit low texture details on certain objects in the game.

• Starfield* (DX12) may exhibit texture flickering on light sources during gameplay.

• UNCHARTED: Legacy of Thieves Collection* (DX12) may exhibit texture corruption on characters.

• Halo Infinite* (DX12) campaign may experience an application crash on some system configurations.

• Dead by Daylight* (DX11) may experience an application crash during gameplay.

• Topaz Video AI* may experience errors when using some models for video enhancement.

• Adobe After Effects* may experience an application crash during render operations.

• Some Intel® Arc™ A-Series Graphics notebooks may incorrectly report a higher value than expected for default clock frequency value.

Intel® Iris™ Xe MAX Graphics Products:

• Driver installation may not complete successfully on certain notebook systems with both Intel® Iris™ Xe + Iris™ Xe MAX devices. A system reboot and re-installation of the graphics driver may be required for successful installation.

INTEL® ARC™ CONTROL FIXED ISSUES:

• Intel® Arc™ Control may fail to launch application on certain multi-GPU system configurations.

• Intel® Arc™ Control Performance Tuning settings may not retain after graphics driver update.

• Arc Control Studio Capture may fail to capture when multiple monitors are connected in clone mode.

INTEL® ARC™ CONTROL KNOWN ISSUES:

• Using Arc Control Studio capture with certain games may incorrectly generate multiple video files.

• The “Connector” type in the Display page may incorrectly show DP* when using an HDMI* display connection.

• May observe “stream has already ended” pop-up after the power events with Capture/Highlights/Broadcast toggle on.

• Performance Graphs may scale incorrectly after waking up from sleep.

To download the new Arc 31.0.101.4824 beta driver, head over to this page on Intel's official website.

Source