KEY TAKEAWAYS

With AI taking off, a lot of tech companies are developing their own chatbots to assist users in specific tasks. The latest addition to the market comes from Adobe, which is adding a new AI assistant to Acrobat. The tool, aptly named "AI Assistant in Acrobat," will assist people in breaking down large documents into summaries that are easier to digest, and will allow you to "talk" to the document.

The new AI Assistant in Acrobat

As reported by The Verge, Adobe is adding its AI chatbot to Acrobat PDF. You can invoke this chatbot at any time and ask it to summarise long documents for you. You can also ask the chatbot questions related to the content within the document, and it will do its best to find an answer and relay it back to you. The chatbot can even generate citations for the information it provides, so you know its answers are based on what the document actually says, and not something it hallucinated.

Right now, the chatbot is in a beta phase, with Adobe planning to add more features before its official release. If you want to give the chatbot a try, you'll need to be subscribed to Adobe's Standard, Pro, or Teams plan. Right now, the AI chatbot comes as part of the subscription fee, but Adobe stated that it wants to add a separate monthly subscription specifically for the AI chatbot once it comes out of beta.

The question is, will people warm to Adobe's chatbot? We've already seen a ton of competition enter the AI market, and a lot of them come with a subscription fee. We've already seen what happens when multiple big-name companies come out with separate monthly subscription plans; just ask anyone trying to juggle watching shows on Netflix, Disney Plus, and other media streaming sites.

The solution is usually to subscribe to one or two plans and leave the rest alone. With this in mind, will people feel it's worth shelling out for Acrobat's new AI assistant when they could just as easily pay for Copilot Pro and ask it to sum up their documents in Word, alongside other features? Time will tell; for now, however, we'll have to wait and see how the public responds to Adobe's new chatbot.

Source

In CPU manufacturing, Intel usually makes its own processors. But that’s not the same for the likes of AMD and others.

These days, AMD designs the CPUs and TSMC makes the CPUs for them. Based in Taiwan, TSMC is the world’s largest semiconductor manufacturer. TSMC also makes CPUs and GPUs for companies like Nvidia, Qualcomm, Broadcom and others.

In 2022, AMD announced the release of the Zen 4 based Ryzen 7000 series processors. These processors are built on the Zen 4 architecture and were made on a 5nm process.

Zen 4 is the same architecture that powers the conventional Ryzen 7000 CPUs, 3D V-Cache based X3D CPUs and also recently launched Ryzen 8000G APUs that come with relatively more powerful built-in GPUs.

Now it seems AMD is all set to release Zen 5 architecture based Ryzen 9000 series CPUs later this year.

3nm Wafers To Enter Production In Q2

Taiwan based UDN reports (translated), via Dan Nystedt on Twitter, that AMD is going to launch Zen 5 based CPUs this year.

This news comes from the information that TSMC is going to start mass production of the 3nm wafers to be used in Zen 5 processors from the second quarter this year. However, full mass production will take some time. Still, the production capacity is expected to increase month by month.

The report mentions that AMD is expected to introduce the Zen 5 architecture based processors in desktops, notebooks and even servers. This means that the Ryzen 9000 CPUs are included in it.

The report also how AMD’s launch of Zen 5 architecture will launch along with TSMC entering the mass production of 3nm process. 3nm process based chips are expected to power AMD Ryzen 9000 CPU series.

However, there are rumors floating around that 3nm process based chips will not be used in whole Ryzen 9000 CPUs. There are some suggestions that only Zen 5c cores are expected to be based on 3nm. Zen 5c cores are smaller and slower cores with same architecture that of larger and faster ones. The usual Zen 5 based big cores are expected to be based on 4nm process instead.

So far, nothing is confirmed. But we expect AMD to announce or even release Zen 5 based Ryzen 9000 CPU series in the third quarter this year. It’s interesting to see how big architectural and performance improvements it offers.

Source

Commenting on the possible upcoming console, Serkan Toto, CEO of Kantan Games, told CNBC:

“There seems to be a broad consensus in the game industry that Sony is indeed preparing a launch of a PS5 Pro in the second half of 2024. And Sony will want to make sure to have a great piece of hardware ready when GTA VI hits in 2025, a launch that will be a shot in the arm for the entire gaming industry.”

Since the PlayStation 5 launched in 2020, the company has already released the slim version of the console. According to CNBC, the Pro edition could mark a significant upgrade over previous versions.

The prediction of a new console update comes just a week after Sony cut its forecast for PS5 sales. The Japanese company said that for the fiscal year ending in March, 21 million units would be sold instead of 25 million units. During the company’s earnings call, one of its executives said that the firm was anticipating a gradual decline in unit sales from the next fiscal year.

The idea that shipments will fall over the coming years is hardly surprising. Eventually Sony will be launching a new console and that date is getting closer and closer so people might be delaying their purchase. Additionally, we have to consider the increasing number of people who have already made a purchase.

If Sony does indeed release the PS5 Pro this year, it will be interesting to see if it can reinvigorate the public’s interest in the PS5 line-up.

Source: CNBC

Source

On this week's episode of the Window's Central Podcast, Dan and Zac discuss the Windows 11 version 24H2 being confirmed, and expected to bring new AI features & quality of life improvements, Windows using AI to improve your video games, the new features on the Windows 11 'Moment 5' feature update, Microsoft Gaming CEO Phil Spencer confirming some Xbox games are heading to PlayStation 5 and Nintendo Switch, confirmation that new Xbox hardware will be announced this holiday season, Game Pass reaching 34 million subscribers, OpenAI's video generation breakthrough, and more!

Links

• Windows 11 version 24H2: New features, next-gen AI PCs, and everything we know so far - Windows Central
• The next version of Windows will use AI to improve your video games - Windows Central
• Windows 11 'Moment 5' feature update is coming this month — here are the new features - Windows Central
• Microsoft Gaming CEO Phil Spencer confirms some Xbox games are headed to PlayStation 5 and Nintendo Switch - Windows Central
• Xbox Game Pass has hit a new growth milestone - Windows Central
• Microsoft confirms new Xbox hardware will be announced this holiday season - Windows Central
• OpenAI's latest model generates lifelike minute-long AI videos - Windows Central

Hosts:

• Daniel Rubino
• Zac Bowden

Find us elsewhere:

• Download directly: Audio
• Listen via Spotify
• Subscribe via iTunes
• Subscribe via RSS
• Subscribe via Google Play Music
• Subscribe via Pocket Casts

Source

What you need to know

• An emerging report indicates that Reddit is in the middle of a megadeal with an unnamed large AI company worth $60 million yearly.
• The deal could potentially allow the company to use the content from Reddit to train its AI models. 
• More details about the deal, the company's identity, and what it intends to do with the content it gets from Reddit remain unclear.

Generative AI is a hot topic in the technology landscape as more companies continue to warm up to it and integrate it into their workflows. In the past year and change, we've seen people using AI to unlock new heights and tap into new opportunities across education, medicine, computing, and more.

While this is impressive, there's a growing concern revolving around the safety and privacy measures in place to prevent AI from spiraling out of control. Adding to this, there's also the issue of companies like Microsoft and OpenAI stealing using copyrighted information to train their models.

At the beginning of this year, Microsoft and OpenAI were slapped with lawsuits over intellectual property theft by The New York Times and two non-fictional authors. The companies argue that copyright law doesn't forbid the use of copyrighted material to train AI models. 

Interestingly, OpenAI CEO Sam Altman admitted that it's virtually impossible to create ChatGPT-like tools without using copyright material, further indicating that restricting the training of these tools to copyright-free material would create AI chatbots that cannot meet the average user's minimum requirements. 

The copyright restrictions explain the increase in reports of chatbots like OpenAI's ChatGPT getting dumber and Microsoft Copilot's (formerly Bing Chat) decline in accuracy. As it happens, a large AI company is reportedly in the middle of a new licensing deal with Reddit worth "about $60 million on an annualized basis," according to a spot by Bloomberg.

The deal could potentially lead to the unnamed company using Reddit posts to train its AI models, though details revolving around this deal remain slim and left to speculation. Reddit harbors credible information across its subreddits coupled with comments and interactions from avid users. As such, it's a gold mine and credible tool that can be leveraged to further enhance and improve the capabilities of LLMs. 

Compensation at last, but at what cost?

(Image credit: Future)

For what seems like an eternity, companies like Microsoft and OpenAI simply lift information from websites and package it as their own in bite-size form, with little regard for referencing to the source or even compensation.

It's only last year in December when it was reported that OpenAI was in the middle of a megadeal with German publisher Axel Springer, which will see it part tens of millions of euros in 3 years (A first, if you ask me). In return, the tech company will have access to articles (archived and current) from the publisher to train its AI models.

However, it remains uncertain what kind of reception this will get. Reddit has had its own fair share of issues and challenges over the past few years. You might remember last year's fiasco when the company announced its plans to start charging for access to its APIs. A move that led to thousands of forums being shut down in protest of the move, which consequently led to the crashing of the site.

Source

With the shift to artificial intelligence away from traditional search engines, Gartner says that companies will have to adjust their marketing channels strategies. Alan Antin, Vice President Analyst at Gartner, said:

“Organic and paid search are vital channels for tech marketers seeking to reach awareness and demand generation goals. Generative AI (GenAI) solutions are becoming substitute answer engines, replacing user queries that previously may have been executed in traditional search engines. This will force companies to rethink their marketing channels strategy as GenAI becomes more embedded across all aspects of the enterprise.”

The analyst firm also said that search engine algorithms will favour quality content to help offset the growing amount of AI-generated content. In addition, it’s expected that watermarking will become more important as a means to highlight high-value content.

While Google will probably not be liking this prediction, it’s important to note that it and many of the other providers of search engines are the main providers of generative AI services; Google has Gemini and Microsoft has Copilot.

So while traditional search engine usage may decline in favour of AI, it will still be the likes of Google getting the traffic, however, it could mean that there will need to be some sort of replacement for sponsored links which Google relies on for revenue,

Gartner didn’t mention this in its forecast but the generative AI revolution also means smaller players could become significant competitors to Google in search. Most will know of Microsoft’s Copilot which is essentially ChatGPT with web access but there is also Perplexity, a startup that has attracted funding from Amazon founder Jeff Bezos and NVIDIA.

Source: Gartner

Source

For much of last year, knocking OpenAI off its perch atop the tech industry looked all but impossible, as the company rode a riot of excitement and hype generated by a remarkable, garrulous, and occasionally unhinged program called ChatGPT.

Google DeepMind CEO Demis Hassabis has recently at least given Sam Altman some healthy competition, leading the development and deployment of an AI model that appears both as capable and as innovative as the one that powers OpenAI’s barnstorming bot.

Ever since Alphabet forged DeepMind by merging two of its AI-focused divisions last April, Hassabis has been responsible for corralling its scientists and engineers in order to counter both OpenAI’s remarkable rise and its collaboration with Microsoft, seen as a potential threat to Alphabet’s cash-cow search business.

Google researchers came up with several of the ideas that went into building ChatGPT, yet the company chose not to commercialize them due to misgivings about how they might misbehave or be misused. In recent months, Hassabis has overseen a dramatic shift in pace of research and releases with the rapid development of Gemini, a ”multimodal” AI model that already powers Google’s answer to ChatGPT and a growing number of Google products. Last week, just two months after Gemini was revealed, the company announced a quick-fire upgrade to the free version of the model, Gemini Pro 1.5, that is more powerful for its size and can analyze vast amounts of text, video, and audio at a time.

A similar boost to Alphabet’s most capable model, Gemini Ultra, would help give OpenAI another shove as companies race to develop and deliver ever more powerful and useful AI systems.

Hassabis spoke to WIRED senior writer Will Knight over Zoom from his home in London. This interview has been lightly edited for length and clarity.

WIRED: Gemini Pro 1.5 can take vastly more data as an input than its predecessor. It is also more powerful, for its size, thanks to an architecture called mixture of experts. What do these things matter?

Demis Hassabis: You can now ingest a reasonable-sized short film. I can imagine that being super useful if there's a topic you're learning about and there's a one-hour lecture, and you want to find a particular fact or when they did something. I think there's going to be a lot of really cool use cases for that.

We invented mixture of experts—[Google DeepMind chief scientist] Jeff Dean did that—and we developed a new version. This new Pro version of Gemini, it’s not been tested extensively, but it has roughly the same performance as the largest of the previous generation of architecture. There’s nothing limiting us creating an Ultra-sized model with these innovations, and obviously that’s something we're working on.

In the last few years, increasing the amount of computer power and data used in training an AI model is the thing that has driven amazing advances. Sam Altman is said to be looking to raise up to $7 trillion for more AI chips. Is vastly more computer power the thing that will unlock artificial general intelligence?

Was that a misquote? I heard someone say that maybe it was yen or something. Well, look, you do need scale; that's why Nvidia is worth what it is today. That’s why Sam is trying to raise whatever the real number is. But I think we're a little bit different to a lot of these other organizations in that we've always been fundamental research first. At Google Research and Brain and DeepMind, we've invented the majority of machine learning techniques we're all using today, over the last 10 years of pioneering work. So that’s always been in our DNA, and we have quite a lot of senior research scientists that maybe other orgs don't have. These other startups and even big companies have a high proportion of engineering to research science.

Are you saying this won’t be the only way that AI advances from here on?

My belief is, to get to AGI, you’re going to need probably several more innovations as well as the maximum scale. There’s no let up in the scaling, we're not seeing an asymptote or anything. There are still gains to be made. So my view is you've got to push the existing techniques to see how far they go, but you’re not going to get new capabilities like planning or tool use or agent-like behavior just by scaling existing techniques. It’s not magically going to happen.

The other thing you need to explore is compute itself. Ideally you’d love to experiment on toy problems that take you a few days to train, but often you'll find that things that work at a toy scale don't hold at the mega scale. So there's some sort of sweet spot where you can extrapolate maybe 10X in size.

Does that mean that the competition between AI companies going forward will increasingly be around tool use and agents—AI that does things rather than just chats? OpenAI is reportedly working on this.

Probably. We’ve been on that track for a long time; that’s our bread and butter really, agents, reinforcement learning, and planning, since the AlphaGo days. [In 2016 DeepMind developed a breakthrough algorithm capable of solving complex problems and playing sophisticated games.] We’re dusting off a lot of ideas, thinking of some kind of combination of AlphaGo capabilities built on top of these large models. Introspection and planning capabilities will help with things like hallucination, I think.

It's sort of funny, if you say “Take more care” or “Line out your reasoning,” sometimes the model does better. What's going on there is you are priming it to sort of be a little bit more logical about its steps. But you'd rather that be a systematic thing that the system is doing.

This definitely is a huge area. We're investing a lot of time and energy into that area, and we think that it will be a step change in capabilities of these types of systems—when they start becoming more agent-like. We’re investing heavily in that direction, and I imagine others are as well.

Won’t this also make AI models more problematic or potentially dangerous?

I've always said in safety forums and conferences that it is a big step change. Once we get agent-like systems working, AI will feel very different to current systems, which are basically passive Q&A systems, because they’ll suddenly become active learners. Of course, they'll be more useful as well, because they'll be able to do tasks for you, actually accomplish them. But we will have to be a lot more careful.

I've always advocated for hardened simulation sandboxes to test agents in before we put them out on the web. There are many other proposals, but I think the industry should start really thinking about the advent of those systems. Maybe it’s going to be a couple of years, maybe sooner. But it’s a different class of systems.

You previously said that it took longer to test your most powerful model, Gemini Ultra. Is that just because of the speed of development, or was it because the model was actually more problematic?

It was both actually. The bigger the model, first of all, some things are more complicated to do when you fine-tune it, so it takes longer. Bigger models also have more capabilities you need to test.

Hopefully what you are noticing as Google DeepMind is settling down as a single org is that we release things early and ship things experimentally on to a small number of people, see what our trusted early testers are going to tell us, and then we can modify things before general release.

Speaking of safety, how are discussions with government organizations like the UK AI Safety Institute progressing?

It’s going well. I'm not sure what I'm allowed to say, as it's all kind of confidential, but of course they have access to our frontier models, and they were testing Ultra, and we continue to work closely with them. I think the US equivalent is being set up now. Those are good outcomes from the Bletchly Park AI Safety Summit. They can check things that we don’t have security clearance to check—CBRN [chemical, biological, radiological, and nuclear weapons] things.

These current systems, I don't think they are really powerful enough yet to do anything materially sort of worrying. But it's good to build that muscle up now on all sides, the government side, the industry side, and academia. And I think probably that agent systems will be the next big step change. We'll see incremental improvements along the way, and there may be some cool, big improvements, but that will feel different.

Source

Source

However, it looks like Spencer has made another fairly major announcement about a future Xbox feature to a fan this week as well. According to X (formerly Twitter) user @jhe007, a fan with the Gamertag "NASburg" asked Spencer when players will be able to access and play their currently owned Xbox games in the cloud. Spencer replied, "Should be this year".

This exchange has reportedly been confirmed as genuine by Windows Central. This should be a very big deal for Xbox gamers who may want to play games in the cloud via Xbox Game Pass Ultimate, but want to continue playing them in the cloud if they are removed from the service.

There are still a couple of questions about this upcoming feature. One is whether players will have to continue to pay for Xbox Game Pass Ultimate to access their owned games in the cloud or if they can continue to do so if they decide not to get a Game Pass subscription. If you don't have to pay for a Game Pass plan, Microsoft might find another business model for gamers who want to play their owned games in the cloud.

Microsoft has not made any major announcements about cloud gaming in a while, but perhaps in 2024 we will get some updates on this service, especially as the company continues to emphasize it wants people to play Xbox games on any screen.

Source

One of the main issues for the company is that it was forced to close its Chinese business as a result of US chip controls. Last year, the company announced that its revenues had fallen by a massive 46% and needed new funding as a result.

To date, the company has reduced staff numbers like very many other tech companies and it has closed offices that it has abroad.

If the information obtained by The Guardian is correct, the company could be worth $500 million after investors increased the value of their stakes. This valuation is a far cry from late 2020 when the company was valued at a much higher $2.8 billion.

As a bit of background if you’ve not heard of Graphcore, it has been on people’s radar since before the COVID-19 pandemic and has been developing Intelligent Processing Unit (IPU) chips to power machine learning and artificial intelligence jobs.

Speaking to ZDNET several years ago, Graphcore CEO Nigel Toon said that IPUs are able to outperform GPUs with massive parallelism and on-chip memory. They can also get AI workloads done with less energy usage which is cited as a significant problem with generative AI right now which relies on GPUs.

If OpenAI was to pick up Graphcore it could be pretty significant. It could potentially make its GPT-4 LLM much cheaper to run which could lead to lower subscription costs. It would also be helpful for the company if it wanted to host a larger next-generation LLM cost effectively.

Source: The Telegraph via City A.M.

Source

It's a reminder that not all of Apple's product launches are home runs. That was certainly true of the company's first venture into stand-alone digital cameras. This nearly forgotten Apple device, the QuickTake 100, was first announced by the company 30 years ago this week, on February 17, 1994, as part of the MacWorld trade show in Japan.

Apple started working on a digital camera, with the code name Venus, in 1992. The company decided to partner with the leading camera company at that time, Kodak, for the project. Kodak ended up creating the CCD sensor for the QuickTake 100, and Apple partnered with the Japan-based camera maker Chinon to actually build the product.

The QuickTake 100's design didn't really look like a normal film camera. Indeed, it looked more like a projector. The one-pound camera's sensor was just 0.3 MPs, and it could take and store up to 8 24-bit color 640 × 480 resolution images, or 32 320 × 240 resolution images. It did include a built-in flash, but there were no focus or zoom features.

The Apple QuickTake 100 was marketed by the company as the first digital camera for consumers, even though that wasn't really true (the Fuji DS-X was sold in Japan in 1989, and the Dycam Model 1/ Logitech FotoMan launched in the US in 1990). It launched a version that worked with Mac computers in May 1994 for US $749 (about $1,575 in 2024 dollars). A version for Windows PCs launched late in 1994.

Apple later released the QuickTake 150 in April 1995. It had the same hardware as the 100 model, but the new version incorporated better compression technology, allowing the camera to take 32 standard-quality images, or 16 best-quality images, both at 640 × 480 resolution.

On February 17, 1997, three years to the day that it announced the QuickTake 100 Apple announced the QuickTake 200 digital camera. This camera, which was manufactured by Fujifilm, went on sale in March 1997 for $600.

The QuickTake 200 had a design that looked much more like a standard film camera compared to the 100 and 150 models. It could take still images at 640 × 480 resolution, and in addition to its 2 MB onboard storage, Apple sold a custom 4 MB card that could increase its storage. It also had an LCD screen in the back where you could view the images you had just shot, and there were now ways you could adjust the camera's focus and aperture.

By the time the QuickTake 200 had launched, two forces came into play that basically doomed the camera line. One was that well-established film camera makers were coming out with digital cameras of their own, and consumers gravitated to those devices. The other big force was the return of Apple co-founder Steve Jobs as CEO in 1997. He gave the order to shut down Apple's non-computing businesses, and that included the QuickTake line.

Today, many people don't even know about the existence of the Apple QuickTake cameras. However, it is considered by others to be a major milestone in consumer digital photography. Even Time magazine put the Apple QuickTake 100 on its list of the 100 best gadgets ever back in 2010. By that time, Apple had launched its iPhone lineup where it has put a ton of emphasis on its digital camera software and hardware. The iPhone's camera certainly owe a debt to the QuickTake products, even though ultimately they were not embraced by the public.

Source

Table of contents:

1. Windows 11 news
2. Updates are available
3. Gaming news
4. A blast from Microsoft's past
5. 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.

On February 13, 2024, Microsoft released a new wave of Patch Tuesday updates for supported Windows versions:

Those updates move the Copilot button to the right in Windows 11 version 23H2, add more natural voices for Narrator, fix File Explorer bugs, and improve security. Note that Windows 11 version 22H2 no longer receives optional non-security updates starting this month. As for Windows 10, Microsoft will end its mainstream support in less than a year. Meanwhile, Google hopes those unable to upgrade to Windows 11 will switch to ChromeOS Flex.

Moving from cumulative updates to more substantial changes, here is an interesting finding in the recently released build 26052, the first build of the now-official Windows 11 version 24H2. Enthusiasts discovered that the operating system now requires a new instruction called POPCNT. Processors that do not support it cannot run Windows 11 version 24H2. You can check whether your old timer supports the POPCNT instruction using this short manual.

But you should not worry too much about this. Officially supported processors will continue running Windows 11 just fine. The change will cripple very old systems with decades-old processors. And if you use Windows 11 on unsupported hardware, this new utility will help you upgrade to the upcoming version 24H2. By the way, the updated Windows 11 setup app can now warn you about the missing instruction set.

Another broken thing in build 26052 is the ExplorerPatch app. You can no longer use its latest version with Windows 11 version 24H2. There is currently no information on what is the culprit, so we are waiting for news from the developers.

Finally, the recent Windows 11 preview builds made it impossible to bypass system requirements with the command that fools the OS into thinking it is a server SKU.

As for more positive findings, it appears that Windows 11 is getting a built-in AI upscaler. A feature ID in build 26052 lets you toggle on "Automatic Super Resolution" to "make supported games play more smoothly with enhanced details." Unfortunately, it does not appear to be working, so all we can do is feast our eyes on the new toggle in the Settings app and wait for more details from Microsoft.

In addition, Windows 11 will soon receive improvements for Bluetooth discoverability to make it easier to detect and connect to uncommon device types and a dedicated notification center for Widgets.

Here is some more news about the recently released preview of the sudo command. As it turned out, sudo is not coming to Windows Server 2025. Microsoft said that the toggle was added to server build 26052 by mistake, and developers will remove all the sudo references in the upcoming updates. Ironically, at the same time, enthusiasts managed to make sudo work not only on Windows 10 but all the way back to Windows 7. With that said, turning on sudo on Windows 7 requires slightly more effort than flicking a single toggle in the Settings app.

There was also an update to the story about stealing BitLocker keys using dirt-cheap hardware. As it turned out, not only old devices are susceptible to this attack. Researchers have found a way to sniff BitLocker keys on recent hardware. However, the method is quite tedious and requires physical access to the target device.

Windows Insider Program

Here is what Microsoft released for testing this week:

One of the most interesting changes in build 26058 is a PowerToys utility that is now a native Windows feature. The latest preview build lets you toggle on a crosshair cursor if you need more visual assistance while working with your computer.

There are also slightly redesigned context menus with labels for the common file actions, such as copy, cut, paste, rename, delete, and share. It's a pretty neat-looking change as long as your Windows 11 copy is set to the English language.

Finally, Microsoft published a new ad detailing why you should update from Windows 10 to 11, and we published a trick that can help you install Windows updates faster.

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.

Developers, get ready for the main event of 2024. Microsoft has confirmed that Build 2024, its flagship developers conference, is coming May 21-23. Look out for registration and more information about ticket prices.

Files, a popular third-party file manager for Windows 10 and 11, has been updated to version 3.2. The latest release includes List View, thumbnail rendering improvements, the ability to edit album covers in audio files, and plenty of under-the-hood changes that significantly reduce the likelihood of crashes, especially when working with large files. Version 3.2 is now available for download from the Microsoft Store, GitHub, and the project's official website.

Do you remember Wunderlist, that app Microsoft purchased nearly a decade ago and turned into what is now known as Microsoft To Do? Many will be glad to know that the app is back in a new shape and form. The founder of Wunderlist launched a new app called Superlist, which aims to be a middle ground between primitive task managers and overly complicated programs like Notion. Superlist is already available on the Web, macOS, iOS, and Android.

Skype Preview received another big update, proving once again that the service is far from dead. Version 8.113 introduced voice message transcriptions, a reworked reaction picker, and 25 new AR lenses made by Snap. It also fixed an annoying bug spawning multiple messages about missed calls in chats.

If you wonder what will happen to Outlook if you delete Microsoft Edge, here is the simple answer: nothing. Do not believe false reports on the Internet claiming that Microsoft's browser is required for the new Outlook app to run. What is needed is WebView2 Runtime, getting rid of which is not an easy task. Speaking of Microsoft Edge, there is a bug with the downloads UI, which forces you to make multiple clicks to open a downloaded file.

Microsoft is working on several improvements for Teams, its corporate messenger. One of the recent updates replaced the Microsoft 365 Chat app with Copilot. Also, Teams users will soon be able to join work meetings using personal accounts.

By the way, Microsoft announced plans to kill Publisher. The company will end its support in October 2026 and remove the app from the Microsoft 365 subscription. It says other apps in the Office suite can already do what Publisher does. What Microsoft is not killing is Azure IoT Central. As it turned out, that message about deprecation was fake news published by mistake.

Microsoft President Brad Smith published a post on the company's official blog to detail how Microsoft plans to fight abusive AI-generated content, such as the recent influx of explicit images of American pop singer Taylor Swift.

There is also an update to the European Union's investigation of whether Apple and Microsoft should be considered gatekeepers. The European Commission announced that iMessage, Bing, Edge, and Microsoft Advertising should not be labeled as gatekeepers.

Here are the new drivers released this week:

• Nvidia GeForce Game-Ready Driver 551.52 WHQL with Skull and Bones support and fixes for micro-stutters in games with vertical synchronization. In addition to the driver, Nvidia released Chat with RTX, a tool for building local chatbots powered by Nvidia RTX graphics cards.
• Intel 31.0.101.5330 non-WHQL with support for Nightingale and Pacific Drive, plus some notable performance improvements in existing games.
• AMD Radeon Software 24.2.1 with Nightingale, Helldivers 2, and Skull and Bones support.

Other notable releases and updates from this week include the following:

• The redesigned Sticky Notes app is now available as part of OneNote for Microsoft 365 Insiders.
• As promised, Microsoft now allows hiding the new Windows Backup app on certain Windows 10 and 11 editions.
• MSI Afterburner dropped Windows XP support. Windows XP fans are mourning.
• AMD has updated its Ryzen Master app with support for Ryzen 8000G and improved Curve Optimizer controls for easier tuning without frequent restarts.
• Microsoft Edge received a fix for the feature that brings your browsing data from other browsers to Edge.
• A new tool for removing AI-related stuff from Windows 11.
• Microsoft turned off the Bing Deep Search feature to fix caching and make it more robust.
• Microsoft launched the first preview of .NET 9 and announced its development goals.
• Microsoft 365 Insiders can try annotation improvements for OneNote and a new feature for PowerPoint on iPad.
• Microsoft Loop now lets you insert Figma links.
• Copilot commercial data protection is coming to more Microsoft 365 plans.
• Surface Pro 9 with Intel processors received the February 2024 firmware update.

Finally, here are the latest additions to Microsoft 365 Roadmap, a website where you can track upcoming features for Microsoft's productivity apps. The company is working on new capabilities for Teams and Outlook, plus upgrades for the file viewer in OneDrive for the web.

On the gaming side

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

Xbox fans can breathe a sigh of relief. No, Microsoft is not abandoning Xbox, nor it is porting its entire game library to PlayStation. The company held the promised Xbox Podcast, where its gaming CEO announced four games coming soon to other consoles (one such game could be Microsoft Flight Simulator) and denied rumors about Starfield or the upcoming Indiana Jones titles on PlayStation 5. There were also some hits about the future hardware, including promises about "the largest generation leap ever."

In addition, Microsoft announced Diablo IV for Xbox Game Pass and promised to release more Activision Blizzard games in its gaming service.

Microsoft released a new software update for the Xbox consoles, adding touch controls for remote play and the ability to calibrate thumbsticks.

Microsoft Flight Simulator is getting a new free DLC in collaboration with Dune. You will soon be able to pilot an Ornithopter across six new activities in Arrakis. In addition, Microsoft revealed a Dune-themed Xbox Series X and a levitating Xbox Wireless Controller. You can purchase them, but you can try your luck and win them in a giveaway.

Mojang has updated Minecraft and Minecraft Java Edition with a new feature that lets you quickly jump into your world right on the launcher screen. Quick Play support is now available in the latest public versions of the game.

IGN wants to try to replace the now-dead E3 with a new event called IGN Live. Later this year, the website will host a gaming show in Los Angeles where visitors can expect "a curated list of leading exhibitors across gaming and entertainment." Specific details remain unknown.

Finally, Nvidia added Halo Infinite and Skull and Bones to its cloud streaming service, allowing those without powerful hardware to play the latest and most demanding titles using the cloud infrastructure.

Deals and freebies

Dakar Desert Rally is free to claim in the Epic Game Store until the next Thursday. Get it while you can if racing all sorts of motor vehicles across sands is your thing. Also, you can still play No Man's Sky for free until February 19, 2024.

As usual, you can find more deals and offers in our weekly Weekend PC Game Deals series.

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.

On February 17, 2024, Windows 2000 marked its 24th birthday. This week's look-back article is dedicated to this old piece of Microsoft-made software, its development, various editions, different capabilities, and more.

Random fact about Microsoft

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

This week's random fact about Microsoft comes from @wowstartsnow on X. Did you know that Microsoft experimented with mobile versions of Windows before Windows CE or Windows Mobile? In 1993, Microsoft was working on WinPad, an operating system for PDAs based on Windows 3.1. The idea was to let users sync emails, files, faxes, and other productivity data from a PC to a mobile device.

For better or for worse, Microsoft scrapped WinPad in 1994. The next year, Microsoft released Windows CE (codenamed Pegasus) for handheld computers.

Source

On Steam, the highest concurrent player count for Helldivers 2 was hit on Saturday with 333,827 gamers online at once. It remains the top-selling game on Steam at $39.99, and even its more expensive $59.99 "Super Citizen" bundle edition, with some extra in-game items, is in third place. There's no word on the player counts for the PS5 version.

However, Helldivers 2 continues to deal with overloaded servers, as Arrowhead clearly didn't anticipate the popularity of the game. On Saturday evening, the developer announced on the game's official Discord channel (via VG247) that a surge of new players online has resulted in "some mission payouts failing, some players being kicked to their ships, or being logged out."

While the team says it is fixing some of these issues, it also added that it is "struggling" to get its servers scaled up to deal with the influx of players. It added:

Therefore we've had to cap our concurrent players to around 450,000 to further improve server stability. We will continue to work with our partners to get the ceiling raised. If you have progression related issues, please restart the game in order for things to sync back up. Thank you for your continued patience

Arrowhead announced on X (formerly Twitter) earlier this week that it decided to increase the amount of player XP that can be earned by 50 percent this weekend. This was an attempt to make up for a bug that caused some players to not get their rewards properly at the end of missions. It also added some new and very hard Defense Missions this week where players are battling AI-based Automatons.

Source

When Covid-19 sent people home in early 2020, the computer scientist Tom Zahavy rediscovered chess. He had played as a kid and had recently read Garry Kasparov’s Deep Thinking, a memoir of the grandmaster’s 1997 matches against IBM’s chess-playing computer, Deep Blue. He watched chess videos on YouTube and The Queen’s Gambit on Netflix.

Despite his renewed interest, Zahavy wasn’t looking for ways to improve his game. “I’m not a great player,” he said. “I’m better at chess puzzles”—arrangements of pieces, often contrived and unlikely to occur during a real game, that challenge a player to find creative ways to gain the advantage.

The puzzles can help players sharpen their skills, but more recently they’ve helped reveal the hidden limitations of chess programs. One of the most notorious puzzles, devised by the mathematician Sir Roger Penrose in 2017, puts stronger black pieces (such as the queen and rooks) on the board, but in awkward positions. An experienced human player, playing white, could readily steer the game into a draw, but powerful computer chess programs would say black had a clear advantage. That difference, Zahavy said, suggested that even though computers could defeat the world’s best human players, they couldn’t yet recognize and work through every kind of tough problem. Since then, Penrose and others have devised sprawling collections of puzzles that computers struggle to solve.

Chess has long been a touchstone for testing new ideas in artificial intelligence, and Penrose’s puzzles piqued Zahavy’s interest. “I was trying to understand what makes these positions so hard for computers when at least some of them we can solve as humans,” he said. “I was completely fascinated.” It soon evolved into a professional interest: As a research scientist at Google DeepMind, Zahavy explores creative problem-solving approaches. The goal is to devise AI systems with a spectrum of possible behaviors beyond performing a single task.

A traditional AI chess program, trained to win, may not make sense of a Penrose puzzle, but Zahavy suspected that a program made up of many diverse systems, working together as a group, could make headway. So he and his colleagues developed a way to weave together multiple (up to 10) decisionmaking AI systems, each optimized and trained for different strategies, starting with AlphaZero, DeepMind’s powerful chess program. The new system, they reported in August, played better than AlphaZero alone, and it showed more skill—and more creativity—in dealing with Penrose’s puzzles. These abilities came, in a sense, from self-collaboration: If one approach hit a wall, the program simply turned to another.

Tom Zahavy helped design a computer system that plays chess more creatively by

combining the approaches and strategies of up to 10 different programs.

 Courtesy of Tom Zahavy

That approach fundamentally makes sense, said Allison Liemhetcharat, a computer scientist at DoorDash who has worked with multi-agent approaches to problem-solving in robotics. “With a population of agents, there’s a higher probability that the puzzles are in the domain that at least one of the agents was trained in.”

The work suggests that teams of diverse AI systems could efficiently tackle hard problems well beyond the game board. “This is a great example that looking for more than one way to solve a problem—like winning a chess game—provides a lot of benefits,” said Antoine Cully, an AI researcher at Imperial College London who was not involved with the DeepMind project. He compared it to an artificial version of human brainstorming sessions. “This thought process leads to creative and effective solutions that one would miss without doing this exercise.”

Chasing Failures

Before joining DeepMind, Zahavy was interested in deep reinforcement learning, an area of artificial intelligence in which a system uses neural networks to learn some task through trial and error. It’s the basis for the most powerful chess programs (and used in other AI applications like self-driving cars). The system starts with its environment. In chess, for example, the environment includes the game board and possible moves. If the task is to drive a car, the environment includes everything around the vehicle. The system then makes decisions, takes actions and evaluates how close it came to its goal. As it gets closer to the goal, it accumulates rewards, and as the system racks up rewards it improves its performance. The “deep” part of this approach describes the neural networks used to analyze and assess behaviors.

Reinforcement learning is how AlphaZero learned to become a chess master. DeepMind reported that during the program’s first nine hours of training, in December 2017, it played 44 million games against itself. At first, its moves were randomly determined, but over time it learned to select moves more likely to lead toward checkmate. After just hours of training, AlphaZero developed the ability to defeat any human chess player.

But as successful as reinforcement learning can be, it doesn’t always lead to strategies that reflected a general understanding of the game. Over the past half-decade or so, Zahavy and others noticed an uptick in the peculiar glitches that could happen on systems trained with trial and error. A system that plays video games, for example, might find a loophole and figure out how to cheat or skip a level, or it could just as easily get stuck in a repetitive loop. Penrose-style puzzles similarly suggested a kind of blind spot, or glitch, in AlphaZero—it couldn’t figure out how to approach a problem it had never seen before.

But maybe not all glitches are just errors. Zahavy suspected that AlphaZero’s blind spots might actually be something else in disguise—decisions and behaviors tied to the system’s internal rewards. Deep reinforcement learning systems, he said, don’t know how to fail—or even how to recognize failure. The ability to fail has long been linked to creative problem-solving. “Creativity has a human quality,” Kasparov wrote in Deep Thinking. “It accepts the notion of failure.”

Antoine Cully has built robots that can effectively brainstorm multiple different solutions to a given problem.

 Courtesy of Imperial College London

AI systems typically don’t. And if a system doesn’t recognize that it has failed to complete its task, then it may not try something else. Instead, it will just keep trying to do what it has already done. That’s likely what led to those dead ends in video games—or to getting stuck on some Penrose challenges, Zahavy said. The system was chasing “weird kinds of intrinsic rewards,” he said, that it had developed during its training. Things that looked like mistakes from the outside were likely the consequence of developing specific but ultimately unsuccessful strategies.

The system regarded these weird rewards as steps toward the greater goal, which it couldn’t actually achieve, and it didn’t know to try something new. “I was trying to make sense of them,” Zahavy said.

A Better Game

Part of the reason these glitches can prove so consequential—and so useful—comes from what researchers recognize as a problem with generalization. While reinforcement learning systems can develop an effective strategy for connecting a given situation to a specific action—which researchers call a “policy”—they can’t apply it to different problems. “What normally tends to happen with reinforcement learning, almost regardless of the method, is that you get the policy that solves the particular instance of the problem you’ve been training on, but it doesn’t generalize,” said Julian Togelius, a computer scientist at New York University and research director at modl.ai.

Zahavy saw the Penrose puzzles as requiring just this sort of generalization. Maybe AlphaZero couldn’t solve most puzzles because it was so focused on winning entire games, start to finish. But that approach introduced blind spots exposed by the unlikely arrangements of pieces in Penrose puzzles. Maybe, he reasoned, the program could learn to beat the puzzle if it had enough creative room to brainstorm and access different training methods.

So he and his colleagues first collected a set of 53 Penrose puzzles and 15 additional challenge puzzles. On its own, AlphaZero solved less 4 percent of the Penrose puzzles and under 12 percent of the rest. Zahavy wasn’t surprised: Many of these puzzles were designed by chess masters to intentionally confuse computers.

As a test, the researchers tried training AlphaZero to play against itself using the Penrose puzzle arrangement as the starting position, instead of the full board of typical games. Its performance improved dramatically: It solved 96 percent of the Penrose puzzles and 76 percent of the challenge set. In general, when AlphaZero trained on a specific puzzle, it could solve that puzzle, just as it could win when it trained on a full game. Perhaps, Zahavy thought, if a chess program could somehow have access to all those different versions of AlphaZero, trained on those different positions, then that diversity could spark the ability to approach new problems productively. Perhaps it could generalize, in other words, solving not only the Penrose puzzles, but any broader chess problem.

His group decided to find out. They built the new, diversified version of AlphaZero, which includes multiple AI systems that trained independently and on a variety of situations. The algorithm that governs the overall system acts as a kind of virtual matchmaker, Zahavy said: one designed to identify which agent has the best chance of succeeding when it’s time to make a move. He and his colleagues also coded in a “diversity bonus”—a reward for the system whenever it pulled strategies from a large selection of choices.

When the new system was set loose to play its own games, the team observed a lot of variety. The diversified AI player experimented with new, effective openings and novel—but sound—decisions about specific strategies, such as when and where to castle. In most matches, it defeated the original AlphaZero. The team also found that the diversified version could solve twice as many challenge puzzles as the original and could solve more than half of the total catalog of Penrose puzzles.

“The idea is that instead of finding one solution, or one single policy, that would beat any player, here [it uses] the idea of creative diversity,” Cully said.

With access to more and different played games, Zahavy said, the diversified AlphaZero had more options for sticky situations when they arose. “If you can control the kind of games that it sees, you basically control how it will generalize,” he said. Those weird intrinsic rewards (and their associated moves) could become strengths for diverse behaviors. Then the system could learn to assess and value the disparate approaches and see when they were most successful. “We found that this group of agents can actually come to an agreement on these positions.”

And, crucially, the implications extend beyond chess.

Real-Life Creativity

Cully said a diversified approach can help any AI system, not just those based on reinforcement learning. He has long used diversity to train physical systems, including a six-legged robot that was allowed to explore various kinds of movement, before he intentionally “injured” it, allowing it to continue moving using some of the techniques it had developed before. “We were just trying to find solutions that were different from all previous solutions we have found so far.” Recently, he has also been collaborating with researchers to use diversity to identify promising new drug candidates and develop effective stock-trading strategies.

“The goal is to generate a large collection of potentially thousands of different solutions, where every solution is very different from the next,” Cully said. So—just as the diversified chess player learned to do—for every type of problem, the overall system could choose the best possible solution. Zahavy’s AI system, he said, clearly shows how “searching for diverse strategies helps to think outside the box and find solutions.”

Zahavy suspects that in order for AI systems to think creatively, researchers simply have to get them to consider more options. That hypothesis suggests a curious connection between humans and machines: Maybe intelligence is just a matter of computational power. For an AI system, maybe creativity boils down to the ability to consider and select from a large enough buffet of options. As the system gains rewards for selecting a variety of optimal strategies, this kind of creative problem-solving gets reinforced and strengthened. Ultimately, in theory, it could emulate any kind of problem-solving strategy recognized as a creative one in humans. Creativity would become a computational problem.

Liemhetcharat noted that a diversified AI system is unlikely to completely resolve the broader generalization problem in machine learning. But it’s a step in the right direction. “It’s mitigating one of the shortcomings,” she said.

More practically, Zahavy’s results resonate with recent efforts that show how cooperation can lead to better performance on hard tasks among humans. Most of the hits on the Billboard 100 list were written by teams of songwriters, for example, not individuals. And there’s still room for improvement. The diverse approach is currently computationally expensive, since it must consider so many more possibilities than a typical system. Zahavy is also not convinced that even the diversified AlphaZero captures the entire spectrum of possibilities.

“I still [think] there is room to find different solutions,” he said. “It’s not clear to me that given all the data in the world, there is [only] one answer to every question.”

Source

Source

Source

Hardware testing and benchmarks remain among the most important part in showing the capabilities of a computer product. It doesn’t matter whether the product is from AMD, Nvidia, Intel or others.

These benchmarks are done via various different software which test these hardware products efficiently. Testing hardware products ranges from CPU to graphics cards to RAM to even SSD and others.

One of the more commonly used benchmarking software for CPUs, especially for server CPUs, is SPEC CPU 2017.

In Intel’s own words, “SPEC CPU 2017 is one of those common, industry-standard benchmark suites that measures and scores general compute-intensive performance.”

Now it seems that Intel is found cheating in the same SPEC CPU 2017 benchmarks.

2,600 Official Intel CPU Benchmarks Invalidated

Phoronix and ServeTheHome reports (via Tom’s Hardware) that SPEC CPU 2017 has stopped publishing results of some of the Intel CPU benchmarks done via the SPEC CPU 2017.

It has also added notice to about 2,600 benchmarks while invalidating them. These benchmark invalidations mostly effect the Intel 4th Gen Sapphire Rapids CPUs. They are server and workstation CPUs and not consumer ones.

For what. Apparently, it is found that Intel CPUs were using a compiler which was found to be wrongly boosting performance specifically only in certain part of SPEC CPU 2017 benchmarks, rather than improve performance in actual workloads. This essentially means that Intel CPUs were cheating in the benchmarks.

Phoronix says that a these benchmark submissions now mention this notice on them:

“SPEC has ruled that the compiler used for this result was performing a compilation that specifically improves the performance of the 523.xalancbmk_r / 623.xalancbmk_s benchmarks using a priori knowledge of the SPEC code and dataset to perform a transformation that has narrow applicability.

 

In order to encourage optimizations that have wide applicability (see rule 1.4 https://www.spec.org/cpu2017/Docs/runrules.html#rule_1.4), SPEC will no longer publish results using this optimization.

 

This result is left in the SPEC results database for historical reference.”

Phoronix also says that performance boost in the benchmarks can be as much as 9% in some case and 4% in another.

It also mentions that the said compiler is an open source one and Intel may or may not have any role in this. It can also be oversight from Intel’s side during the review process.

Conclusion

Intel doing this should not come as a surprise. Two years ago, Intel published a whole PDF trying to essentially say that SPEC CPU 2017 isn’t good enough and everyone should move ahead from it.

It’s not the first time that a combination of Intel CPUs and compilers have found to be in the news for the wrong reasons.

Years ago, Intel was accused of making a compiler that made sure that any software created using that compiler would perform and run slower on AMD CPUs than on Intel ones.

Anyway. For now, this is a developing story. It would be interesting to see what Intel replies.

Source

Source

Source

Today, the game's developer and publisher, Larian, is rolling out the game's next major patch. The notes for Patch 6 show two small but interesting new improvements to the game. The biggest improvement should appear to the many gamers who enjoy the more romantic aspects of Baldur's Gate 3.

As Larian states in the release notes:

All characters now have unique kisses that reflect their personality, with an emphasis on the plural. These kisses are randomised and vary from the incredibly romantic to… uh, a little more intense.

Larian also says that characters that are shorter or taller than normal have had changes made to their kissing animation as well. Also, when you enter into one of the game's camps, you can see some improvements in the idle animations for your companion characters with this new patch.

The extensive patch notes also include lots of bug fixes for various issues. It also adds some extra Legendary Actions for some of the boss enemies.

Patch 6 rolls out today to PC, Xbox, and PlayStation users. If you play Baldur's Gate 3 on the PC, expect a pretty hefty 150GB file to download. Mac gamers will have to wait a little longer to get Patch 6.

In related news, Baldur's Gate 3 was one of the big winners this week in the annual D.I.C.E. Awards. It received the top award for Game of the Year and also won for Role Playing Game of the Year. In addition, it picked up awards for Best Story, Game Direction, and Game Design categories.

Source

Remedy announced today that the award-winning entry is its fastest-selling release ever, with the game reaching 1.3 million copies sold across PC (Epic Games Store), Xbox Series X|S, and PlayStation 5.

The title had already sold a million copies by the start of 2024, and the latest 300,000 had joined in just the past two months. The studio also shared that its previous game, Control, has now sold over four million copies since its launch in 2019, bringing in approximately 100 million Euros in revenue. "A great game can generate excellent long tail sales, and we expect this to be the case with Alan Wake 2 as well," it adds.

"We are happy with the start of Alan Wake 2's sales," says Remedy CEO Tero Virtala. "The price point has also remained at a high level, and the game has already recouped a significant part of the development and marketing expenses. We will continue to develop the game to serve existing fans and attract new players and expect the game to continue selling well."

Remedy has already announced it is working on two story expansions for Alan Wake 2, which will be dropping sometime in 2024 as steady updates to the base game continue.

The studio also has other gaming projects in the works to expand on its collected universe of supernatural games. This includes Project Condor, a cooperative spin-off, as well as a Control sequel. Max Payne and its sequel, Max Payne 2: The Fall of Max Payne, are receiving remakes under Remedy as well, with Rockstar serving as publisher for this project.

While no launch windows have been announced for future Remedy projects, Virtala adds that the titles will be reaching "their next development stages during the first half of 2024" as Alan Wake 2's development team transitions over to the new games.

However, one of the studio's upcoming projects, Vanguard, was not a part of the announcements today. The Tencent-published game recently went through an internal reboot. It was originally set to be a free-to-play live service experience, but a major direction change has made it a paid entry with a focus on cooperative play.

Source

According to OpenAI's research paper, the new Sora model is trained on "text-conditional diffusion models jointly on videos and images of variable durations" to generate hyper-realistic videos from text-based prompts. The company noted that it is teaching AI "to understand and simulate the physical world in motion, with the goal of training models that help people solve problems that require real-world interaction."

It also shared a bunch of clips on both the announcement post as well as the research paper showcasing the model, with company's CEO Sam Altman even taking prompt requests from users to show the model's capabilities.

OpenAI further noted that Sora model is not available to everyone as the company is currently working with red teamers to test the limits of the model to ensure safety.

The company is also working with policy makers "to understand their concerns" surrounding the technology. While the model is not available to the public, OpenAI said that it will make the model available to select group of artists to get their feedback on the technology.

We are also granting access to a number of visual artists, designers, and filmmakers to gain feedback on how to advance the model to be most helpful for creative professionals.

 

We’re sharing our research progress early to start working with and getting feedback from people outside of OpenAI and to give the public a sense of what AI capabilities are on the horizon.

There is no word on when OpenAI will offer the new model to the public and also what would it cost to the end users. OpenAI has been leading the charge with its CEO Sam Altman rumoured to be looking for an investment to the tune of $7 Trillion for his endeavors.

Source

Source

Cores, VRAM, PCIe 5.0, or DDR5 — you might be guilty of believing in these 10 PC hardware myths.

KEY TAKEAWAYS

Building a gaming PC might not be rocket science, but it still involves multiple components working together. And that inevitably gives rise to several PC building myths — big or small. But this time, I want to dive deeper and talk about some of the misconceptions we have about nearly every PC component. Despite the explosion of online information about PC building and PC gaming, these misconceptions continue to exist in 2024.

Some of these can be attributed to legacy PC building, while some have cropped up more recently. Whether you're planning to upgrade your hardware for PC gaming in 2024 or simply confused by the deluge of PC content flooding your feeds, this list might prove helpful.

10  More CPU cores is all that matters

Platform architecture and frequency matter too

 

This one has to be one of my favorites. No matter how many CPU generations we see, there will be someone shouting from the rooftops that more cores are always better. The fact is that more cores are better, yes, but only when you're comparing apples to apples. For instance, an 8-core Ryzen 7 7700X is more powerful than a 6-core Ryzen 5 7600X from the same CPU generation. But, this Ryzen 5 7600X might obliterate the much older 8-core Ryzen 7 3700X.

A newer platform almost always improves the IPC (instructions per clock) and the CPU frequency, delivering better performance in almost every workload. And this doesn't even take into account the impact of CPU cores in gaming. Most games don't need an 8-core processor to maximize performance, especially at higher resolutions like 1440p and 4K. You'll be more than fine with a 6-core processor from the same family.

Additionally, gaming FPS also depends heavily on how the game engine is coded to take advantage of hyperthreading. Some games could be heavily CPU-dependent while others won't need more than 6 cores. The best CPUs for gaming are those that excel in multiple factors at the same time — clock speeds, platform improvements, power efficiency, and price. Focusing on the core count alone, without seeing other specifications of a CPU, is pointless.

9  More VRAM means more performance

It's a lot more nuanced than that

Back when GPUs used to be simpler, you could be forgiven for thinking that VRAM was the only indicator of GPU performance. Generally, more powerful GPUs used to ship with more VRAM than their inferior counterparts. But when that stopped happening, and we started seeing 12GB RTX 3060s next to 10GB RTX 3080s, that's when things became complicated.

The average user might conclude that the card with more VRAM is the faster card, but the story is incomplete without factors like the architecture, shading units (CUDA cores, Stream Processors), memory type (GDDR5, GDDR6), memory bus width, and, of course, clock speeds.

Similar to CPUs, the best GPUs from a newer generation are generally faster than their comparable counterparts from the previous generation. More than VRAM, other factors determine the performance of a GPU.

Even if you're running your games at 4K with high-resolution textures, you'll end up hitting the compute limits of your graphics card before saturating the framebuffer or VRAM. More VRAM is necessary for creative professionals, but it only makes sense on graphics cards that are powerful enough to sustain the heavy workloads that need that additional VRAM.

8  Blazing-fast DDR5 RAM is essential for gaming

The sweet spot is probably lower than you think

The DDR4 vs. DDR5 debate might have settled down, but many users seem to have concluded that they need the absolute fastest memory kit to maximize their gaming performance. The best DDR5 RAM for gaming in your specific use case will most often be a 6000 MT/s CL30 kit. That happens to be the sweet spot for the latest Ryzen 7000 processors, at least.

Even for Intel, the performance gains are minimal unless you reach or surpass 8000 MT/s. But those kits are expensive, and more importantly, require extremely high-end motherboards to even be stable. For the average user, a 6000 MT/s or 6400 MT/s kit should be the target to enjoy games at maximum performance. Considering the price premium and stability issues related to faster DDR5 kits, it's best to save the money and invest that in a faster graphics card, if gaming is your main concern.

7  PCIe 5.0 NVMe SSDs are indispensable

You'll rarely feel the difference

 

PCIe 5.0 SSDs have been available for some time now, and have even come down in price a lot (relatively). But do you really need one? If all you're doing is gaming, you don't. Loading times and overall performance don't improve much by moving from a Gen4 NVMe SSD to a Gen5 one. DirectStorage might change this, but we're yet to see that in action. Paying double the money for seemingly no improvement in gaming doesn't make a whole lot of sense, does it?

 

On the other hand, if you're a creative professional and frequently find yourself transferring huge media files, you could find value in a high-speed PCIe 5.0 NVMe SSD. You'll be able to actually use the increased read/write speeds to see a tangible difference in performance. The price premium might not feel prohibitive if you're accelerating your workflows and saving time for your business.

6  The motherboard doesn't matter

It does, but not in the way you think

You might already know that a fancy, high-end motherboard will not really boost your gaming FPS. But, that doesn't mean you can pick any low-end motherboard for your build and expect a top-tier experience. This is because the motherboard is responsible for a lot of the other components you put inside your system.

• Need to overclock a high-end processor? You need a board with capable VRMs
• Need enough M.2 ports for future SSDs? Get a premium motherboard with at least 3 high-speed M.2 slots.
• Want to run blazing-fast DDR5 RAM? You'll need a board that has your desired kit in its QVL (Qualified Vendor List)
• Packing a beefy graphics card? A board with a reinforced PCIe slot might help

Even things like lower boot time and bundled NVMe heatsinks can enhance the value of the motherboard you're buying. So, the best motherboards might not increase your FPS but they'll avoid crippling your PC experience in other ways.

5  Bottlenecking should be avoided at all costs

It doesn't matter as much as you think

Bottlenecking is a genuine concern among new builders and the vast majority of uninformed gamers. While it's true that your aim should be to minimize the bottlenecks in your build, you'll never be able to eliminate them completely. So, a slightly weaker CPU with a stronger GPU should not be a world-ending concern for you. Similarly, opting for an affordable memory kit, albeit one that's slightly slower, doesn't mean your performance will suffer. significantly.

Besides, the actual bottleneck might change from game to game, and even between different environments in the same game, as different areas might stress either your CPU or GPU more than the other. So, stressing about your CPU and GPU based on the games you want to play might also not work out in the end.

You should look at performance benchmarks, CPU scaling reviews, and other content before finalizing your configuration, keeping in mind your budget and preferences. As long as you match your CPU and GPU well, and pick other decent components with good price-to-performance, you'll end up with a great rig.

4  Bronze-rated PSUs are always bad

Quality of the unit matters more

Of late, I've seen a lot of focus on gold-tier or platinum-tier PSUs in the community, to the point that bronze-rated PSUs have been shunned almost entirely. It's always advisable to go for a better PSU if your budget permits, but you shouldn't overlook bronze PSUs simply due to the lower-rated efficiency. It's better to check the reviews of all the PSUs in your shortlist before making any decision.

You can also check reputable lists online to verify the reliability of a PSU. Quality bronze-rated units will almost always trump no-name gold-rated units. Sometimes, even some renowned brands end up producing sub-par PSUs, so keeping a lookout for reviews and news stories about the most popular PSUs is the best way to go.

3  Open-air cases are dust collectors

They're just not

Source:  Tech Tesseract (YouTube)

Cases used to come in a familiar form factor — panels on all sides with some mesh for air intake. But, with the growing popularity of open-air cases, your components don't need to be trapped inside a closed box. The case against the best open-air cases hinges on dust buildup and poor airflow. But, the best open-air PC cases make it easy to clean any dust buildup, and with the right configuration, airflow shouldn't be a concern.

PC cases, open or closed, will always gather dust. But, with open-air cases, you can at least clean the dust easily. Cleaning regular cases feels like an awful chore — I've been delaying cleaning my Lancool II Mesh for months now. So, you shouldn't pass up on some great-looking open-air cases simply due to fears of dust collection or airflow.

2  Gaming monitors are expensive

Not anymore

1080p gaming was, and still is, the predominant choice among PC gamers. Higher resolutions like 1440p require investing in not just a new monitor but, sometimes, a stronger graphics card too. The best 1440p monitors used to be significantly pricier than their 1080p brethren, but that isn't the case anymore. You needn't deprive yourself of 1440p gaming on account of the monitor, at least.

Some of the most popular 27-inch and 32-inch 1440p high-refresh gaming monitors can now be had for around $300. You can get models from brands like MSI, Gigabyte, Acer, and ViewSonic. Moreover, you also have curved options in this budget if that's something you prefer. I bought myself an LG 27GL850 for almost $400 around 3 years ago. The upgraded version of that model now costs less than $300.

4K monitors are still relatively expensive, considering even decent budget models cost around $400. But, prices for these models are also coming down all the time.

1  Future-proofing is important

A flawed endeavor

Future proofing is less important than ever before. At a time when PC components are getting outclassed by their next-gen versions at a rapid pace, it's a doomed exercise. Making your rig last longer by buying high-end and overkill components that you don't actually need is never advisable. Sure, you'll enjoy your RTX 4090 for a year or two. But you'll inevitably be replacing it with the next big thing in order to keep your rig in the same performance class.

Instead, you can invest in components that are within the price-to-performance sweet spot. That way, you'll be able to upgrade more frequently, and even get a better percentage of your initial investment back by selling your old parts (compared to top-of-the-line parts that lose more of their value every year).

PC components: The good, bad, and ugly

Building PCs, especially gaming PCs, comes with a fair bit of research and shopping around to arrive at the best bang-for-your-buck build. Whether you're going for a balls-out enthusiast gaming PC to enjoy path tracing in all its glory, or simply need a budget gaming PC, the work you need to put in is more or less the same. But, despite the less-than-ideal aspects of putting a custom build together, you can rest assured that you'll be delighted by the end result, no matter your budget.

Source

SpaceX wants to minimize this downside by proactively discarding a portion of its satellites:

“In the coming weeks and months, SpaceX will perform controlled descents of approximately 100 additional early-version 1 Starlink satellites. These satellites are currently maneuverable and serving users effectively, but the Starlink team identified a common issue in this small population of satellites that could increase the probability of failure in the future.”

The company specifies that the “lowering operation” of satellites will take approximately six months for most of the vehicles. SpaceX will also update other satellite operators on the trajectory of these devices and will take responsibility for any avoidance maneuvers if needed.

SpaceX communication is emphasizing its commitment to space sustainability. Understandably so, as the company has been under the scrutiny of regulators, experts, and the public for a long time.

Discarding satellites proactively is an important decision. As SpaceX points out, its satellites are placed in low Earth orbit (LEO) below the altitude of 600 kilometers, which means the atmospheric drag brings unresponsive devices down within five years – an industry standard. However, proactive deorbiting significantly shortens this period, preventing the risk of collisions with other spacecraft and making room for new, hopefully even more reliable replacements.

To this date, SpaceX has conducted a controlled deorbit on 406 Starlink satellites. Only 17 vehicles are non-maneuverable, although the company emphasizes they are “well-tracked” to prevent collision with other objects in orbit – not great, not terrible.

Even though the current “bad batch” is still operational, the company says the action would have no impact on the customer experience: “SpaceX has the capacity to build up to 55 satellites per week and launch more than 200 satellites per month, which allows us to continually improve our system and make it more resilient.”

With the currently deployed second generation of Starlink, SpaceX is not only improving its reliability but also adding new capabilities. Just in January, the company successfully tested direct communication between smartphones and Starlink satellites for the first time.

However, Elon Musk is not the only one with plans for smartphones communicating through the space infrastructure. Google, AT&T, and Vodafone only recently invested in Texas-headquartered company AST SpaceMobile, focused on building the world’s first space-based cellular broadband network.

Source