Latest News - Page 5

Thirteen years is a long time to wait for a video game. Apparently, it's long enough for someone just to say, "Fine, I'll do it myself."

Ziwen Xu, a 25-year-old founder of AI agent startup Hyperecho, kicked off the project with a post on X. "Day 1 of building GTA 6. Still feels fake typing that out," he wrote, and upgraded to Claude Max 20x specifically for the project. The first footage was exactly what you'd expect: a blue box wandering around an empty 3D plane. Not much, but every game starts somewhere.

He spent a week ignoring everyone who told him to switch engines before finally caving on day seven and migrating from Godot to Unreal Engine. The project, dubbed GT-Caliber, runs on daily public updates and an open GitHub repo, with AI agents continuously looping, taking community feature requests, and generating pull requests in real time. Perhaps unsurprisingly, there's no publisher, and no studio budget. Ziwen seems to be relying on whoever shows up for the project.

AMD has worked with the HandBrake development team to fix two threading bottlenecks that were affecting transcoding performance on high-core-count Threadripper CPUs. The fixes are included in HandBrake 1.11.0 and later, and the results are significant enough that owners of existing Threadripper hardware can expect a substantial free performance boost simply by updating the software.

According to AMD, video transcoding workloads generally benefit from higher core counts. However, that was not always the case with its Threadripper lineup, where performance sometimes scaled in the opposite direction, particularly at lower resolutions, resulting in longer transcode times. AMD says it identified two key causes behind the issue.

First, HandBrake wasn't built to manage systems with more than 64 logical processors. On chips like Threadripper, which have far more than 64 threads, the application left compute resources idle rather than distributing work across all available cores.

RTX 4090 scams on Chinese second-hand markets are nothing new, but a repair channel on Bilibili has uncovered one that is a little too "plastic." Chinese repair channel Brother Zhang purchased a broken ASUS GeForce RTX 4090 for 1,500 RMB and tore it down to discover, unlike some previous scams using older GPU dies, this one used a piece of plastic made to resemble the GPU core.

The graphics card carried markings meant to resemble an NVIDIA AD102-300-A1 GPU, and it looks real for the most part, unless you start tearing it apart. UNIKO's Hardware on X noted that the markings did not follow NVIDIA's actual nomenclature, and the date code on the die suggested the chip was manufactured in 2030, which is obviously impossible for a retail RTX 4090. The substrate also lacked the usual QR code in the bottom-left corner, and the surrounding capacitors were in the wrong positions compared to a real AD102 layout.

When Brother Zhang physically touched the die and cleaned off the thermal paste, the difference in texture was immediately apparent. What looked like a GPU was a plastic piece with fake markings etched onto the surface. There was no silicon underneath at all, meaning the card had zero functional compute capability and could not even output a display signal.

Epic Games kicked off controversy among game developers when it revealed gen AI tools for Fortnite concept creation, and now some partners might actually be pulling out of future collaborations because of it.

Before the big State of Unreal event kicked off, Epic Games shared a brief video that showed how its artists are using generative AI tools to help design concept assets for Fortnite. Epic further revealed that Unreal Engine 6 would have native support for AI tools like Claude Code and Gemini, only worsening Epic's position to specific groups.

The reveal caused a stir in the game dev community--gen AI is usually vilified for helping lay off thousands of video game workers--to the point where Epic might lose out on a collaboration with one of the most popular indie game developers of our era: Poncle, the team behind Vampire Survivors. Epic had just announced a Vampire Survivors crossover in Fortnite, but now that Poncle has learned about gen AI being used, it looks like Vampire Survivors x Fortnite may not happen after all.

Sony may have confirmed that its new first-party singleplayer games will be exclusive to PlayStation consoles moving forward, and Bloomberg's Jason Schreier says this policy won't be a 'case-by-case basis' like Xbox's exclusivity plans.

PlayStation CEO Hideaki Nishino recently spoke to Famitsu to celebrate the magazine's big 40th anniversary, and his comments lend strong credence to the reports that all future first-party singleplayer games from PlayStation Studios will be exclusive to Sony's consoles.

Nishino's response isn't a direct one and is laden with PR-speak, but the gist is that singleplayer games are meant to boost the value of PlayStation, and online live service games are meant to reach as many people as possible and will continue launching on PC.

Modders have figured out how to double the memory capacity of NVIDIA's six-year-old entry-level GTX 1650 from 4GB to 8GB. YouTube channel Paulo Gomes published a video showing off their work modding the entry-level Turing GPU with 8GB of GDDR6, and showed off impressive performance results from the card due to the extra memory capacity. For instance, in Unigine Superposition, the GTX 1650's benchmarking score doubled from 624 points to 1245 points moving from 4GB to 8GB of VRAM.

Upgrading the memory on the GTX 1650 was straightforward and required no extra firmware modifications. The modders swapped out the GTX 1650's existing 1GB GDDR6 modules with 2GB Samsung HC16 GDDR6 modules by desoldering the old modules and soldering the new modules on the PCB. Their first attempt at booting the card failed as a result of installing two faulty HC16 modules, but after swapping out the faulty modules with working counterparts, their second attempt was successful.

One major caveat the modding team noted was that their 8GB mod requires a very specific variant of the GTX 1650 to work - the TU106 version of the GTX 1650 with GDDR6 memory. For the uninitiated, NVIDIA has four different variants of the GTX 1650 (not even counting the GTX 1650 Super), featuring TU117, TU116, and TU106 dies paired to either GDDR5 or GDDR6 memory. The TU106 version was one of the later trim models NVIDIA launched to use up defective TU106 silicon. TU106 was one of NVIDIA's most popular Turing dies and was used to make nearly all of NVIDIA's mid-range RTX 20 series GPUs, such as the RTX 2060, RTX 2060 Super, and RTX 2070. The modders did not fully explain why only the TU106 version is compatible, but likely memory controller and/or firmware limitations are the culprit on other variants of the GTX 1650.

U.S. President Donald Trump dropped a late-night Truth Social post on June 18, confirming that Apple has agreed to work with Intel to design and build its chips in America.

There's just one small detail he left out: Intel didn't know the announcement was coming.

GTA 6 pre-orders will begin next week on June 25, 2026 on PS5 and Xbox Series X/S, Rockstar Games today confirmed.

Today Rockstar announced the pre-order dates for the most anticipated game in the last 10 years, GTA 6, while also unveiling the game's official cover art. The biggest question everyone has is how much the game will cost, and that hasn't been revealed yet. There's speculation that Rockstar could charge a $79.99 base price for the game, or potentially even a $99.99 version with GTA Online extras.

"Pre-orders for Grand Theft Auto VI will officially begin on June 25 on digital storefronts and at other select retailers," Rockstar writes.

Anthropic CEO Dario Amodei has sat down for an interview with Bloomberg Originals, where he discussed several topics surrounding AI and AI development, along with geopolitical impacts and Anthropic's internal governance system intended to maintain checks and balances during the quest to ethically develop artificial intelligence.

Amodei was asked, given the power of artificial intelligence and how Anthropic is one of the leading companies in the space with Claude, Claude Code, and its now-banned Mythos-class AI models, why the US government wouldn't take over the company to ensure its tools are used safely. Amodei responded by saying that almost every major technology was built by or with the involvement of the government, and that AI is the first revolutionary-level technology that is predominantly being built by the private sector, and that isn't something Amodei agrees with, nor would it be what he would have chosen if he were given a choice.

Although Amodei believes the government shouldn't completely take over Anthropic, he is scared of both companies having "this technology" and the government having it. So, the middle ground is public governance tactics, which Amodei says is slowly being introduced to Anthropic. For example, Amodei says Anthropic's Long-Term Benefit Trust can appoint and remove a majority of board members and, through that authority, can even remove him as CEO. The idea behind this structure is to prevent the company from being blown ethically off course through a cabal of upper leadership.

No Law is a new first-person shooter set in a cyberpunk world from the relatively small team at Neon Giant, the developers behind The Ascent. Although it's not a massive open-world game along the lines of CD Projekt Red's Cyberpunk 2077, the densely packed and detailed world of the game's Port Desire setting is reminiscent of how impressive CDPR's game was during its first big gameplay reveal.

At the State of Unreal 2026 event, developer Neon Giant took to the stage to talk about No Law and how it used Unreal Engine 5 to create some of the densest and most detailed environments ever seen in a game. Tor Frick, the creative director on the game, says that he and the small team wanted "extreme detail" everywhere you look, from messy apartments to alleyways and little things like stickers on electrical appliances.

The impressive thing is that it's all hand-crafted, with Neon Giant saying that Unreal Engine 5's Nanite technology is one of the key reasons the team was able to make No Law look the way that it does. And when it comes to density, it's next level, with the team using Mass Framework and Meta Human to simulate crowds of over 3,000 NPCs at once to deliver streets and locations teeming with life.

The AI industry boom is undoubtedly unprecedented, and consumers are feeling the pinch of supply constraints on memory and storage prices. This problem is so widespread that it isn't just consumers who are impacted by higher prices; as new Xbox CEO Asha Sharma recently outlined in a letter, the price Xbox pays for storage and memory has increased by 4x since she took office.

With DRAM manufacturers now turning toward AI-specific memory, such as HBM production, we are beginning to hear rumblings of a race between leading DRAM manufacturers, SK Hynix and Samsung. Reports indicate that both companies are rushing to release the first HBM4E memory solutions, which will undoubtedly be gobbled up by the highest-bidding AI companies and deployed directly in data centers.

HBM4E memory will be a big part of the next generation of data centers, which powers the next generation of AI. The new generation of memory chips will be headed to leading products such as NVIDIA's Rubin Ultra, and AMD's Instinct MI500. Just these two chips, by themselves, are expected to drive millions in sales. As for specifications, it's reported the new HBM4E offering from SK Hynix will arrive in capacities of up to 48GB, and speeds of 16GBps. As for Samsung, the South Korean manufacturer showcased its latest generation, HBM5, on the showroom floor at Computex 2026.

GIGABYTE's new AORUS GeForce RTX 5090 INFINITY graphics cards, created in part to celebrate the company's 40th anniversary, not only feature a premium quad-slot design befitting of a flagship GPU, but a unique one as well. Even though it sports a double flow-through cooling design reminiscent of NVIDIA's own Founders Edition model, the look reminds us of a futuristic spaceship's twin engines and a retro stereo from the 1980s.

The AORUS GeForce RTX 5090 INFINITY's WINDFORCE Hyperburst cooling system includes two large Hawk fans alongside a hidden, smaller fan underneath a mesh panel. With double flow-through cooling, air is pushed through both sides of the reinforced backplate, with composite metal grease, superconducting heat pipes, and a direct-touch vapor chamber found underneath the hood.

It's as premium as it is stylish, and as an OC model it ships with a Boost Clock speed of 2730 MHz, which is notably higher than the 2407 MHz found on the Founders Edition model. Unfortunately, its debut comes amid the current memory and component crisis, which has seen GeForce RTX 5090 prices rise dramatically over the past year. And with that, a MicroCenter listing for the AORUS GeForce RTX 5090 INFINITY puts the price for this new flagship at $5,299.99 USD.

Epic Games has recently wrapped up its State of Unreal 2026 livestream, where it pulled back the curtain on the upcoming Unreal Engine 6.

Unreal Engine 6 was first teased during a recent Rocket League Champion Series, and in that announcement, Epic Games stated that Rocket League would be the first title to adopt the new version of Unreal Engine, which is quite a considerable jump, as Rocket League was built on Unreal Engine 3. Epic explained that it plans to merge Unreal Engine 5 and the Fortnite Unreal Editor to create Unreal Engine 6, and that this update will include integrating popular AI models, such as Anthropic's Claude and Google's Gemini.

MSI's new MAESTRO 500 WIRELESS headset isn't built solely for gaming; in fact, it's more of a versatile lifestyle headset with a stylish design that you can take with you on-the-go. The foldable physical design has more in common with Sony or Bose lifestyle headphones than with previous MSI gaming headsets, and this extends to the inclusion of Active Noise Cancellation (ANC) with an optional Transparency mode for travel or listening in noisy environments, as well as a 6-microphone design for clear voice communications.

The MAESTRO 500 WIRELESS features 40mm drivers that MSI notes are tuned to deliver a balanced, lifelike sound across gaming, music, and even online meetings. Throw in an HD-like frequency response of up to 40 kHz, and we could be looking at MSI's best-sounding headset to date. And for PC gamers, the inclusion of Nahimic for Headset adds an additional layer of audio immersion over low-latency 2.4 GHz, with the MAESTRO 500 WIRELESS also supporting Bluetooth and 3.5mm wired for connecting to a wide range of devices.

One of the more impressive elements of the design is the 6-microphone system, which includes beamforming and Environmental Noise Cancellation (ENC) to isolate your voice and reduce background noise, ensuring clear comms in any environment. It's a design that includes a flip-to-mute microphone for voice, positioning the MAESTRO 500 WIRELESS as a lifestyle-like pair of headphones that's also serious about gaming and chat.

As part of Unreal Fest 2026, NVIDIA has announced that its new NVIDIA ACE Game Agent SDK for creating and building in-game AI companions that run on GeForce RTX hardware is now available in Beta. This suite of tools includes Unreal Engine 5 plugins for the AI-powered Automatic Speech Recognition (ASR), Small Language Models (SLM), and Text-to-Speech (TTS) components of NVIDIA ACE.

Examples of the technology in action are already available, with KRAFTON recently launching its new PUBG Ally in-game AI teammate as part of a limited-time Beta to test and gauge feedback on a new Ally Duo Mode that will pair humans with AI. Another in-development example is Creative Assembly adding an AI advisor to Total War: PHARAOH.

Naturally, the NVIDIA ACE Game Agent SDK will allow any game developer to play around with in-game AI companions, with NVIDIA noting that it has been designed as a "lightweight, C/C++ agentic framework" for seamless integration thanks to the inclusion of a ready-to-use Qwen 3.5 4B model for decision-making and an ultra-lightweight Chatterbox Turbo 350M model for text-to-speech. As seen with the new PUBG Ally, ACE Game Agents can run on GPUs with only 8GB of VRAM, and cards like the GeForce RTX 3060. So yeah, you won't need a GeForce RTX 5080 to run them.

PUBG: BATTLEGROUNDS helped define the Battle Royale genre, and the game remains one of the most-played multiplayer titles on Steam and other platforms. And this week, as part of the game's latest update, the game is introducing a new Ally Duo Mode that introduces a PUBG Ally or AI companion powered by NVIDIA ACE technology.

The PUBG Ally and Ally Duo Mode will be available in Beta through PUBG Arcade for the next two weeks, until June 30, via Steam. NVIDIA says this will allow KRAFTON to "collect invaluable real-world player feedback" and to help guide "the future of AI agents in games." Now, generative AI in PC gaming is controversial, but based on what we've seen in early hands-on demos of the AI-powered PUBG Ally, the idea of an AI-controlled in-game partner or helper is a potential game-changer in the right circumstances.

The good news is that the PUBG Ally, called Ella, is powered by optimized, lightweight models, so it will run on GeForce RTX GPUs with at least 8GB of VRAM, with the GeForce RTX 2080 Ti and GeForce RTX 3060 listed as the minimum specs. As for the recommended specs, KRAFTON lists the GeForce RTX 4070, which features 12GB of VRAM.

Nearly a decade after launch, Nintendo has fixed the one thing Switch owners have been complaining about since 2017. Firmware update 22.5.0, released for the original Nintendo Switch, converts the eShop from a painfully slow web-based interface into a native app. This brings it in line with what Switch 2 users have had since the console launched last year.

For context, the original eShop was always so sluggish because it was running as a wrapped web app rather than a proper native application. This meant every scroll and page load was fighting against the overhead of a browser session. Switch 2 solved this at launch by building the eShop as a native app from the start. Nintendo has now done the same for the original hardware.

Users who tested the update described the eShop as "buttery smooth", with games showing up in batches as you scroll instead of slowly loading one at a time. Daniel Vuckovic on Bluesky posted a clip of the new experience in action, and the contrast with the old version is stark. It is not quite as fast as Switch 2's eShop, given the hardware difference, but it is a world apart from where it was.

TSMC's fully booked production capacity is no longer just an industry talking point, as the ripple effects are becoming increasingly difficult to ignore. According to a report from Nikkei Asia, a growing list of major chip customers, including AMD, Google, Nvidia-backed Groq, Tesla, and BYD, are in discussions with Samsung Foundry about manufacturing future chips as demand continues to outstrip TSMC's available capacity.

AMD is reportedly in talks with Samsung about manufacturing certain future CPUs starting in 2028. Google is discussing two separate Samsung engagements. One covers its next-generation Axion processors expected around 2028, while the other involves components tied to its 10th-generation Tensor Processing Unit, codenamed Icefish, which is being co-developed with MediaTek.

Meanwhile, Tesla has already confirmed that its next-generation AI6 chip will be produced at Samsung's Texas facility. BYD, the world's largest electric vehicle maker, is also in discussions with Samsung about manufacturing future autonomous driving chips. Groq, which develops language processing units and is backed by NVIDIA, is already producing chips at Samsung and may also use the foundry for its next-generation products.

Destiny 2 may stay in retirement as the recent wave of players may not be enough to sway Bungie's/Sony's decision on sunsetting the online FPS.

Faced with financial pressures, Bungie is putting Destiny 2 into stasis. The studio simply doesn't have enough money to fund Marathon and Destiny 2 at the same time, and the studio's assets have been hit hard by a hefty $750 million impairment charge. This basically means that Bungie's games failed in a way that triggered accounting adjustments that impacted and lowered Bungie's net profit by creating another expense on its income statement.

The news sparked a campaign among hopeful gamers that Destiny 2, or the Destiny franchise, could be revived if enough players showed up. The game has had over 100K players on Steam for the past 9 days in a row, since the Monument of Triumph update released. Unfortunately, sources tell Forbes' Paul Tassi that this outpouring of player support hasn't affected the decision around Destiny 2.

AMD has officially confirmed its next-generation Ryzen Threadripper family, codenamed "Mustang Peak." The upcoming lineup will move to a new TR6 platform, bringing Zen 6 cores built on TSMC's 2nm process and PCIe 6.0 support. The confirmation comes from AMD's own documentation, spotted by leaker IntaLatX64, which lists "TR6 Mustang Peak" as a Threadripper Pro CPU under the Family 1Ah Model A8h series.

AMD has not released specifics such as SKUs, core counts, clock speeds, cache sizes, TDPs, chipsets, or a release window. What the documentation does confirm is DDR5 memory support and PCIe Gen 6, the latter doubling per-lane bandwidth compared to the PCIe 5.0 lanes on the current TR5 platform.

The current Ryzen Threadripper 9000 series, codenamed Shimada Peak, is based on 3nm Zen 5 architecture and tops out at 96 cores and 192 threads with a 350W TDP and up to 128 PCIe 5.0 lanes. TR5 has served two generations now, covering both the Threadripper 7000 and 9000 families, making TR6 the first platform change since the 7000 series launched.