Latest News - Page 415

AMD has provided some instructions on how to run DeepSeek's exciting new R1 distilled "reasoning" model on AMD Ryzen AI and Radeon products... yeah, you can run a local DeepSeek R1 model on your PC at home.

AMD explains: "Reasoning models add a "thinking" stage before the final output - which you can see by expanding the "thinking" window before the model gives its final answer. Unlike conventional LLMs, which one-shot the response, CoT LLMs perform extensive reasoning before answering. The assumptions and self-reflection the LLM performs are visible to the user and this improves the reasoning and analytical capability of the model - albeit at the cost of significantly longer time-to-first-(final output)token".

"A reasoning model may first spend thousands of tokens (and you can view this chain of thought!) to analyze the problem before giving a final response. This allows the model to be excellent at complex problem-solving tasks involving math and science and attack a complex problem from all angles before deciding on a response. Depending on your AMD hardware, each of these models will offer state-of-the-art reasoning capability on your AMD Ryzen™ AI processor or Radeon™ graphics cards".

ASUS has revealed the launch pricing of its new custom GeForce RTX 5090 and RTX 5080 graphics cards, with the PRIME RTX 5080 starting at $999 (special launch pricing) while the new flagship ROG Astral LC RTX 5090 costs and eye-watering $3099.

In a new report from VC, we're learning that ASUS has suggested retail pricing but "retailers may impose their own margins", adding that "typically, this isn't an issue immediately at launch but becomes one shortly after. This is because many board partners work directly with large retailers, who are politely asked not to sell cards at higher prices. However, these agreements usually don't last long".

ASUS has since confirmed the MSRP models and which ones will cost up to 50% more than NVIDIA's suggested pricing, with ASUS's new flagship ROG Astral LC GeForce RTX 5090 OC Edition costing a whopping $3099, down to the starting price of the PRIME RTX 5080 at $999 (special launch pricing) or $1199 post-launch.

Tesla has reported Q4 2024 results pulling in $25.7 billion in revenue for the three-month period, delivering over 495,000 cars, with Elon Musk saying this will be the "biggest year in Tesla's history".

In a new post on X, the official Tesla account said that the company expects Model Y to become the world's best-selling car of any kind for the 2nd year in a row. Not only that, but it's become even better with the introduction of its new Model Y launching in all markets, with Tesla promising production will begin ramping at all factories with deliveries following all regions later in Q1 2025.

Tesla says that its Semi factory construction continues, with the first Semi truck builds scheduled for the end of 2025, ramping up in early 2026. Cybercab lines are being prepared at Giga Texas with volume production planned for 2026, an exciting glimpse into the future.

TSMC founder Morris Chang hasn't had a long-form video interview in 17 years, when he was interviewed by NVIDIA CEO Jensen Huang, but now he's appeared in an awesome new interview with the guys at Acquired. It's a long watch, but totally worth it, check it out:

During the interview, the TSMC founder discusses a bunch of hot topics over the last couple of decades, including the relationship between TSMC and NVIDIA, its 40nm process node development back in 2009, resolving TSMC's incredible core business vs its expansions, the 2009 dispute with NVIDIA, and more.

Acquired's interview with Chang includes talking about the all-in move on the 28nm node, meeting Apple and the entire debacle between Apple and Intel -- as Apple was going to have Intel make its chips, but eventually decided on TSMC -- to which the TSMC founder had some fighting words to say about that. Chang said Apple paused its discussions with TSMC at the time, talking to Intel which was making chips for the company's Mac family, but eventually decided on TSMC.

Apple has been secretly working with SpaceX and T-Mobile to add support for Starlink internet connectivity on its iPhones, which provides an alternative to Apple's in-house satellite communication service.

In a new post on X, Tesla and SpaceX boss Elon Musk said "medium resolution images, music & audio podcasts should work with the current generation Starlink direct-to-phone constellation. Next generation constellation will do medium resolution video".

In a new report from Mark Gurman on Bloomberg, we're learning that that the first users of the new Starlink-powered service on iPhones received text messages from T-Mobile saying: "you're in the T-Mobile Starlink beta. You can now stay connected with texting via satellite from virtually anywhere. To start experiencing coverage beyond, please update to iOS 18.3".

NVIDIA has released a new graphics driver for its RTX 5090 and RTX 5080 GPUs which came out today - well, which sold out today - alongside a big update for the NVIDIA App that packs an interesting surprise.

Team Green announced support for its Blackwell graphics cards in the new driver, which is a given - those who were lucky enough to get one would be a bit stuck without that - and DLSS 4 officially came into play.

To begin with, there are 75 games (and apps) which support DLSS 4, and all RTX graphics cards can benefit from this - except for the MFG component, the new Multi Frame Generation feature, which is for RTX 5000 GPUs only.

No Man's Sky just got a huge patch which brings in all manner of goodies.

This is the 'Worlds Part II' update (which follows up, yes, you guessed it - 'Worlds Part I') and it expands the game's universe with new stars, planets, terrain and environments to dive into (literally, in the case of the deep oceans that have been introduced).

The developer tells us:

The official PC requirements for Clair Obscur: Expedition 33 have been released, following the developer announcing on X they are struggling to keep up with the demand for both the Collectors Edition and physical edition of the game.

Clair Obscur: Expedition 33 is a turn-based combat RPG that is powered by the graphically impressive Unreal Engine 5, and while the title may have slipped under some gamers' radars, the hype for it is certainly building momentum, with the developer Sandfall Interactive taking to the official Clair Obscur: Expedition 33 X account to announce they are struggling to keep up with pre-orders. Sanfall Interactive states in its post it has seen the "outpouring of support" and has requested restocks of certain editions and availability.

BioWare has put out a strangely opaque statement about the current developer situation at the company that has been coupled with a positive announcement regarding Mass Effect 5.

The developer took to its blog to announce that Mass Effect 5 has officially begun development, with the studio writing the "core team at BioWare is developing the next Mass Effect game" and that the project is headed by leadership from the original trilogy of games. That veteran leadership includes Mike Gamble, Preston Watamaniuk, Derek Watts, Parrish Ley, and others.

Moreover, BioWare writes that it's taking this time between the release of Dragon Age: Veilguard, which publisher EA confirmed underwhelmingly performed sales-wise, and the development of Mass Effect 5 to reimagine "how we work at BioWare" and to "think deeply about delivering the best experience to our fans."

A former Grand Theft Auto 5 and Red Dead Redemption 2 developer has said he expects Grand Theft Auto 6 to fall short of 60FPS on both consoles.

The former developer is GTA 6 and RDR2 animator Mike York who spoke on Kiwi Talkz and said he expects Rockstar will be aiming for 30FPS on both Xbox Series consoles and PlayStation 5. York said he doesn't know if Rockstar will be able to "pull off 60FPS," and instead, Rockstar will likely opt for locked 30FPS, meaning the game will never dip below that figure -- and if it does it will only be a few frames at most.

For those wanting 60FPS, York said, "But it can be bumped up to 40, 41, 52, whatever, and I bet you, like, later, once it's on PC, it'll probably get super optimized and changed and new graphics cards that come out, and you'll be able to run it at 60fps probably at that time." For comparison's sake, GTA 5 is currently running at 60FPS on PS5, but let's not forget that game released in 2013 and has had many years of optimization-focussed updates stabilizing its performance.

The release date for the highly anticipated Grand Theft Auto 6 is one of, if not the most sought-after date in the gaming industry, and now an online retailer has seemingly revealed it.

The online retailer is a Spanish store called XUruguay, which listed Grand Theft Auto 6's release for September 17, 2025. As you can probably imagine, gamers were quick to dissect the legitimacy of this "leak," with some suggesting it could be real, considering September lines up with Rockstar and publisher Take-Two Interactive's confirmed release window of Fall 2025.

Further conversation was sparked when the online retailer promptly removed the date from its website, and when queried about how it obtained the release date, the retailer said it works "directly with PlayStation, they let us know which products we can sell and on what dates."

The AI industry was set alight when DeepSeek unveiled its R1 AI model, which the company claims is comparable in performance to OpenAI's latest AI model but was created for a fraction of the price.

The new R1 model was created with just $6 million, which is a tiny drop in the lake of money that is pouring into AI companies all throughout the world. For context, OpenAI said it spent approximately $100 million training its latest model, "o3", and it's estimated the company could spend as much as $500 million to create GPT-5, the company's AI model that's on the horizon.

Having a new AI model be released for a fraction of the costs of known AI models caused a massive pullback from investors, resulting in approximately $1 trillion being wiped away from AI companies and the companies that fuel their progression.

DeepSeek is in some deep trouble after cloud services provider Wiz Research did some digging, finding a publicly available database of the Chinese AI company's secret keys, messages, and more.

The database includes details like chat history, API secrets, and more, with Wiz researchers saying that the discovered database can be fully controlled. This means that they could execute code without any oversight, manipulating the data to serve their needs. Wiz reached out to DeepSeek to alert them, with the company quickly taking it down after restricting public access.

The team at Wiz Research set out to assess DeepSeek's external security posture, and find any potential vulnerabilities, with the team explaining: "Within minutes, we found a publicly accessible ClickHouse database linked to DeepSeek, completely open and unauthenticated, exposing sensitive data. It was hosted at oauth2callback.deepseek.com:9000 and dev.deepseek.com:9000".

NVIDIA's new GeForce RTX 5090 and GeForce RTX 5080 graphics cards are nearly here... but it looks like US launch supply is "basically non-existent" and that there could be less than 1000 of the flagship RTX 5090s available in the USA according to the latest leaks.

In a new video posted by leaker Moore's Law is Dead, who spoke with multiple sources that work at large US retailers, the situation seems pretty bleak for PC gamers. If you thought the stock situation of the RTX 5090 and RTX 5080 was bad, this would be really bad... less than 1000 of NVIDIA's new flagship GeForce RTX 5090 graphics cards, ouch.

Source 1: "At least as of this evening (1/29/25), we (major US retailer) have ZERO 5090s in store. Now that could change, we have had last minute deliveries on launch days before, but right now it almost seems we might not get any significant amount of 5090s until March".

When NASA announced it had collected samples from asteroid Bennu, a 4.5 billion-year-old asteroid, the recovered samples were much more than anticipated, approximately double the total planned weight.

The quantity of the samples actually proved to be a problem for researchers as NASA struggled to get the capsule open for nearly three months, eventually resulting in special tools being designed to remove the lid of the capsule. NASA set a mission goal of 60 grams to call the asteroid sample collection mission a success, but OSIRIS-REx gathered more than 120 grams. Those samples have now been analyzed by a large team of scientists from more than 40 institutions around the world, with the researchers publishing a new study in the scientific journal Nature.

The study examined the contents of the samples, and through analysis of the mineral grains' exact structure, along with a chemical composition analysis, the team discovered 14 of the 20 amino acids that create the proteins required for life as we know it on Earth. Additionally, the team found a high concentration of ammonia, and five nucleobases life on Earth uses to transmit genetic instructions within DNA and RNA.

It has been a tumultuous week in the AI industry after Chinese company DeepSeek unveiled its R1 model, which caused approximately $1 trillion to be wiped away from Silicon Valley AI companies as DeepSeek said it's model was on par with OpenAI's ChatGPT, but cost just $6 million to create, which is a fraction of the billions of dollars poured into ChatGPT.

DeepSeek has truly shaken up the AI industry and has caused AI heavyweights such as OpenAI, Microsoft, and others to reassess their own AI chatbots and processes. But now accusations are beginning to fly, with OpenAI and Microsoft alleging they have obtained evidence DeepSeek used stolen OpenAI data to train its R1 model. In a recent Financial Times article, OpenAI claimed DeepSeek trained its AI with OpenAI's models, while Microsoft told Bloomberg that it believes it has evidence of an OpenAI developer account being connected to DeepSeek.

Furthermore, the accusations don't stop there, as Microsoft said this OpenAI developer account is linked to stealing large amounts of data from OpenAI. These accusations were backed up by President Trump's artificial intelligence and cryptocurrency advisor David Sacks, who told Fox News, "There's substantial evidence that what DeepSeek did here is distilled the knowledge out of OpenAI's models." Sacks added it's "possible" DeepSeek has engaged in IP theft.

When Elon Musk purchased Twitter in 2022, his goal was to create the "Everything app," which would mimic China's WeChat, an app that has consolidated popular individual apps such as taxi hailing, food delivery, banking, bill payments, and communication into a singular app.

Musk's plan was to create this "Everything app," but saw the purchasing of then-Twitter and now X as an accelerant to achieving that goal, with the Tesla CEO writing on X in October 2022, "Buying Twitter is an accelerant to creating X, the everything app." In October 2023, we began hearing rumblings of X working on a seamless payment system that was in development, with Musk saying during an earnings call that he plans on turning X into a bank capable of storing "someone's entire financial life".

During that earnings call, Musk said it was likely the new monetary system would be rolled out on the platform "by the end of next year," which didn't seem to happen. However, X CEO Linda Yaccarino has now taken to her personal X account to tease that X has partnered with Visa for X Money, which will enable secure and instant funding to users' X wallets via Visa Direct, peer-to-peer payments (send/receive money between users), and the options to transfer funds to user bank accounts instantly.

ASUS is reportedly preparing a new family of NUC Mini-PCs that will be powered by AMD's new Ryzen AI Max "Strix Halo" APUs.

In a new post on X spotted by "Orlak29_" we're learning that ASUS's new Strix Halo APU-powered NUC has been tested by the BIS (Bureau of Indian Standards) for regulatory certification, with the manifest describing the NUC as the "NUC14LNS" which means it's from the NUC 14 family of systems.

The manifest details the NUC sample as the flagship Ryzen AI Max+ 395 "Strix Halo" APU that features 16 cores and 32 threads of Zen 5 processing power, with 80MB of cache (L2+l3), an RDNA 3.5-based Radeon 8060S GPU with 40 Compute Units (over twice the GPU cores as Strix Point), and an NPU with 50 TOPS of AI workload performance for Copilot+ systems.

At CES, we got to go hands-on with Indiana Jones and the Great Circle with Full Ray Tracing running in 5K on the new GeForce RTX 50 Series. From developer

MachineGames, the immersive first-person stealth, exploration, and adventure game, is not only one of the best-looking games for a high-end PC but also one of the best games of 2024.

Launching in December, it still feels new, and the build we played at CES 2025 implemented several DLSS 4 features, including the new Transformer model for Super Resolution, Ray Reconstruction, and Multi Frame Generation. The current build of Indiana Jones and the Great Circle on PC doesn't include Ray Reconstruction for its ray tracing options, so the jump to DLSS 4 is pretty mind-blowing in what it brings to the overall image quality.

German hardware maker be quiet! has announced a new line of power supplies designed for efficiency and stability for PC gamers. Power Zone 2 sees the return of the company's 'Power Zone' line-up after a twelve-year absence. It's the first series of PSUs from be quiet! to feature semi-passive cooling, and like the original Power Zone PSUs from 2013, the new 2025 models stick to the series' single-rail design (a single massive high-performance 12V-rail).

A modified Pure Wings 3 140mm fan cools the PSU during high-load scenarios (like gaming on a GeForce RTX 5090). The fan will remain off until a specific load or temperature threshold is reached. Interestingly, once on, the fan will always run for 2-5 minutes to avoid frequent changes in noise.

One key feature of the Power Zone 2 line-up, available in 1000W, 850W, and 740W models, is the low electrical noise from coil whine, which has led to a Cybenetics rating of A+ for the 1000W and 850W models and A++ for the 750W model. In addition, Power Zone 2 PSUs are the first from be quiet! to be Cybenetics Platinum certified for efficiency.