AMD's new Zen 5-based Ryzen AI 300 series "Strix Point" APUs are finally out in the wild, with some fresh leaks on the beefier "Strix Halo" APU that will be released in 2025.

The new Strix Halo APU will feature up to a higher 16 cores and 32 threads of Zen 5 processing power, and a massively upgraded RDNA 3.5 GPU system over the current Strix Point APUs that will deliver discrete GPU levels of gaming performance.

In the new leak, we're seeing the new Strix Halo APU huge package: 37.5 x 45.0 mm and using the same FP11 package. The APU itself will be much smaller, with the biggest die on the Strix Halo APU being the RDNA 3.5 GPU that will consume 307mm² die space, while the 2 x Zen 5 CCDs use 66mm² die space, with a total die space (Strix Halo is multi-chiplet) of 439mm².

NVIDIA has hit some major roadblocks with its new Blackwell AI GPUs, with redesigns in the works, and a new B200A on the way... but in the middle of this mess, we get some new details on the specs of Blackwell Ultra.

NVIDIA first teased its mid-generation refresh "Blackwell Ultra" earlier this year, as well as its next-gen Rubin R100 AI GPU that will feature ultra-fast HBM4 memory in 2025 (if there's no delays for Rubin, as there are with Blackwell).

SemiAnalysis has the scoop here, with Blackwell Ultra acting as a mid-generaiton enhancement of Blackwell, with the standard CoWoS-L Blackwell Ultra known as "B210" or "B200 Ultra". Inside, NVIDIA's new Blackwell Ultra features both a memory refresh -- up to 288GB of 12-Hi HBM3E memory -- and performance enhancements of up to 50% in FLOPS.

NVIDIA is reportedly experiencing some major issues with its new Blackwell AI GPUs, with the new B100 and B200 AI GPU shipments heavily impacted, with design flaws causing major headaches not just for NVIDIA.

In a new report published by SemiAnalysis, the outlet reports that there are issues with the Blackwell die needing a possible redesign or that the bridge dies inside TSMC's new CoWoS-L advanced packaging needs a redesign.

The result? NVIDIA is now reportedly working on the B200A AI GPU, which will be released towards the second half of 2025 and won't have the issues that are now plaguing the fastest AI GPU before it could hit the market.

NVIDIA's upcoming Blackwell AI GPU family is "encountering major issues" according to a new detailed breakdown by SemiAnalysis, which will impact production targets through into the first half of 2025.

The main problems are from the NVIDIA Blackwell GPU architecture and the chip packaging used: TSMC's new CoWoS-L. SemiAnalysis explains that one of the rumors swirling is that one of the top few global routing metal layers and bump out of the Blackwell die needs a redesign, while another is that the bridge dies inside the CoWoS-L packaging need a redesign.

SemiAnalysis explains that there has a been a warpage in the CoWoS-L packaging that was caused by heat, with TSMC's new CoWoS-L advanced packaging being more complex than the now more mainstream CoWoS-S packaging.

Elon Musk has officially dubbed Tesla's new supercomputer cluster at Giga Texas "Cortex" which will be powered by 100,000 x NVIDIA H100/H200 AI GPUs with "massive storage" for video training its FSD and Optimus projects.

Elon Musk revealed the new name of the Tesla Gigafactory Texas supercomputer cluster in a post on X, which comes as no surprise to the SpaceX and Telsa (and X) boss.

Elon did say that the new Cortex supercomputing cluster will feature around 100,000 x NVIDIA H100 and H200 AI GPUs that will be used to video train the neural networks for FSD (Full Self-Driving) and Optimus robotics.

AMD's new Ryzen AI 300 series "Strix Point" APUs have had their new RDNA 3.5-based Radeon 890M integrated GPU tested at 18W of power, compared to 28W+ power and the Arc GPU on Intel's current Meteor Lake processors.

The gaming tests between the integrated GPUs was conducted by TechEpiphany, comparing the 17W Strix Point APU against the 28W+ Meteor Lake CPU. On the AMD side of things, the Ryzen AI 9 HX 370 "Strix Point" flagship APU was used, with 16 compute units of RDNA 3.5-based GPU with clocks of up to 2900MHz.

It is inside of the ASUS VivoBook S16 OLED which is equipped with 32GB of LPDDR5X-7500 memory, and tuned down to 17W as the APU itself is set to 60W by default. For comparison, the Intel Core Ultra 7 155H "Meteor Lake" processor was used, with 8 Xe-Cores and a 2.25GHz GPU clock (which is 100MHz slower than the Core Ultra 9 185H integrated GPU). This chip also featured 32GB of RAM, but LPDDR5X-6400 (slightly slower).

The US Department of Justice is being pushed to investigate NVIDIA over anti-competitive concerns in its dominance in the AI GPU business.

Progressive groups backed by Senator Elizabeth Warren, have written a letter to anti-trust boss Jonathan Kanter, to open an investigation into NVIDIA. The push is coming from Demand Progress and 9 other entities that push for government scrutiny against monopolies in business, and now NVIDIA is in their crosshairs.

A magnifying glass is being put over NVIDIA's bundling of hardware and software, which has created a tight-grip monopoly on the AI GPU market. The groups wrote: "This aggressively proprietary approach, which is strongly contrary to industry norms about collaboration and interoperability, acts to lock in customers and stifles innovation".

NVIDIA's next-gen Blackwell AI GPUs have reportedly been delayed, with The Information reporting that "design flaws" are stopping Blackwell from flooding the world of AI.

The Information reports on information from two unidentified people that helped produce the Blackwell AI GPU and its server hardware said that NVIDIA's new Blackwell AI GPUs could be delayed for 3 months or more, affecting major customers like Meta, Google, and Microsoft.

NVIDIA reportedly made Microsoft aware of the issue of Blackwell AI GPUs being delayed this week, at least its "most advanced" AI chip models, according to Bloomberg reporting on unidentified Microsoft staffers and "another person." The delays mean that big shipments aren't to occur until Q1 2025, according to The Information.

Samsung is completely dominating the foldable smartphone business, with a step further seeing its new foldable Galaxy Z Flip smartphones used as police body cameras.

Samsung is promoting the use of its Galaxy Z Flip smartphones as body-warn cameras for police officers, promoting the fact that some US police officers are using them as body cameras. The South Korean giant started a pilot program with the Kimberling City Police and the Indian Point Police in Missouri, where they started testing the Galaxy Z Flip as a body camera for police officers in the field.

The police officers in the pilot program wore the Samsung Galaxy Z Flip smartphones as body cameras in the field, to better maintain transparency. The program was a success, and expanded to other departments.

Intel is experiencing some pain right now. It has lost 30% of its share price in the last 48 hours and is now announcing extended warranty periods for its 13th- and 14th-generation Core CPUs because of their instability issues.

The company has issued a further two years of warranty for 13th Gen and 14th Gen Core CPUs, extending their 3-year warranty to a larger 5-year warranty. However, this is the for the retail boxed versions of the CPUs. There is STILL no recall on Intel's 13th Gen and 14th Gen Core CPUs, and there will be no stopping sales of these processors either.

Intel said it is "committed to making sure all customers who have or are currently experiencing instability symptoms on their 13th and/or 14th Gen desktop processors are supported in the exchange process. We stand behind our products, and in the coming days we will be sharing more details on two-year extended warranty support for our boxed Intel Core 13th and 14th Gen desktop processors".

Intel posted its Q2 2024 earnings report and within hours, saw its stock price absolutely TANK by over 30% and wiping away tens of billions of dollars in market cap in a near-instant.

The news comes on the announcement of Intel set to lay off over 15% of its global workforce, which will result in 17,500 people losing their jobs before the year is out. Q3 2024 forecast revenues were below market expectations, bleeding CPU market share to AMD all over the place (Ryzen for consumers, Threadripper for workstation users and content creators, and EPYC for servers and data centers).

Intel stock has dropped over 31% in the last couple of days, from $31 to $21 which wiped away over $30 billion in market cap, with large-scale lay offs to now occur.

Intel has been having some seriously bad issues with its 13th Gen and 14th Gen Core CPUs, with instability issues, crashes, and more... with no recall in sight and no real admittance to a gigantic issue by the company, Intel is now being slapped with a class action lawsuit.

Millions of customers are affected, so a class action lawsuit was inevitable. Now, Abington Cole + Ellery is tackling the case as the law firm specializes in class action lawsuits and intellectual property (IP). The law firm made a post about the issue, acknowledging the latest developments in the story, including 65W non-K processors being affected.

The law firm posted a summary of the issues in 13th Gen and 14th Gen Core CPUs from Intel, explaining: "Recent reports have highlighted instability issues with Intel's 13th and 14th Generation Core desktop processors. The primary cause of these instabilities has been identified as elevated operating voltages. Intel's analysis revealed that a microcode algorithm was incorrectly requesting higher voltages, leading to instability in the processors. This issue apparently does not stem from the earlier resolved via oxidation problem in Intel's manufacturing process​".

The rumors of Intel laying off thousands of employees were right, but it's even worse than that: Intel is laying off over 15% of its global workforce over a seriously bad Q2 2024 earnings report that has seen Intel stock sinking over 30% in 24 hours.

Intel stock has dropped from $31 down to $21 post-Q2 2024 earnings, which is a disaster for Intel shareholders, the brand itself, and the future of the company.

On top of that, their 13th Gen and 14th Gen Core CPUs are plagued with issues... no recalls announced... Arc is almost DOA apart from being thrown into Lunar Lake CPUs as GPU tiles... and now 30%+ of its stock price sliced away, and 15%+ of its global workforce to lose their jobs.

Sabrent has just teased its new Rocket XTRM 5, which uses Intel's new Thunderbolt 5 connectivity to create one of the fastest standards on the planet with up to 80Gbps of bandwidth.

The new Sabrent Rocket XTRM 5 delivers the dream of pumping desktop-class ultra-fast transfer speeds to an external SSD thanks to Sabrent's world-class SSDs, mixed with the new insane bandwidth capable through Thunderbolt 5.

Sabrent will make its upcoming Thunderbolt 5-powered Rocket XTRM 5 SSD in 1TB, 2TB, and 4TB capacities, so there's one for everyone. It would've been nice to see 8TB at launch, but I'm sure we'll see that down the track.

Qualcomm wants to offer more affordable devices with the Snapdragon X Plus chip as their engine next year.

The news comes from the top, Qualcomm's CEO Cristiano Amon, who made the revelation in an earnings call following the firm's Q3 2024 fiscal results.

The Verge reports that Amon said:

Star Wars Outlaws, a new open-world game set in the iconic universe where you take on the role of a smuggler from the studio behind The Division, sounds like a winner. However, with recent gameplay deep dives and previews being mostly lukewarm and Ubisoft asking for extra money to play the game a few days early, the hype for what should be this month's biggest release doesn't seem to be there.

On the plus side, the PC version sounds technically impressive, with real-time ray-tracing and the new DLSS 3.5 Ray Reconstruction tech dramatically improving image quality on a GeForce RTX rig. However, with Ubisoft updating the official storage page with full PC requirements and hardware specifications for Star Wars Outlaws, you're going to need some serious power to play this at 1440p or higher.

The recommended specs, which cover running the game at 1080, 60 FPS, with the 'High' graphics preset, require at least an NVIDIA GeForce RTX 3060 Ti or AMD Radeon RX 6700 XT GPU. But even then, DLSS or FSR needs to be set to 'Quality,' which means rendering the game at 720p before the upscaler does its thing.

Lots of stories are discussed in this week's episode of The TT Show, but before Jak and Kosta dive into the week that was, there's fallout from the CrowStrike Windows blue-screen and crash stuff to go through. As the dust settles, the billions of dollars in lost revenue and the number of PCs affected have become apparent.

Then there's the ongoing saga that is Intel's 13th and 14th Gen desktop CPU troubles. Thanks to an explosive new interview at The Verge, the size and scope of the problem are bigger than we thought and startling, to say the least. Irreparable damage, no product recall, and the entire lineup of CPUs, including 65W mainstream models, are all affected.

Onto the new stuff: NVIDIA's engineers released a paper on a brand-new AI-powered technology it calls Real-Time Neural Appearance Models. Without getting too technical, it will make objects based on real-world properties like metal and ceramic look more realistic in games and run up to 24 times faster - a true game-changer. Also, Logitech is thinking about creating what it calls a "forever mouse."

NVIDIA ChatRTX is an AI chatbot that is fully customizable, from the LLM used to the data. The company describes it as an AI that lets you "chat with your files" as you can customize the data set across different file formats, documents, images, and other data. The latest version of the application is available now, with a minimal and clean UI overhaul that highlights how powerful it can be.

As you can see in the screenshot, the new UI (built using the Electron + Material UI framework) highlights ChatRTX's two main components: the AI Model and the Data. Even better, users and developers can add their own UI elements and expand ChatRTX's functionality by using an open UI framework.

ChatRTx can tap into powerful AI models for customization and, as such, requires at least a GeForce RTX 30 or GeForce RTX 40 Series GPU with at least 8GB of VRAM to run.

Logitech is known for creating some of the highest-quality peripherals on the market and innovating in ways that have enabled it to capture large segments of consumers.

Its no surprise that Logitech Research and Development laboratories would be constantly producing new and interesting designs of established products, with the goal of inventing a fresh take that attracts more customers. Logitech's CEO Hanneke Faber gave a glimpse behind the curtain at Logitech's "innovation center" in Ireland in a recent interview with The Verge. Faber said they were shown a "forever mouse" that was designed to capture similar longevity as a watch.

Faber was asked what made the "forever mouse" a "forever mouse," to which Faber responded it was heavier, came with great software that users would constantly be updating and was "beautiful." Faber was further pressed about how Logitech would make money from selling a mouse to consumers that they would never have to replace. Faber said that such a product would need to offer value in the form of a service in the form of software updates.

At SIGGRAPH 2024, NVIDIA introduced and expanded on the future of digital humans by introducing its latest AI advancement - adding a path-traced realistic-looking human face (and voice) with emotion and detailed animation to the idea of interacting with an AI assistant. This is a massive leap forward compared to the usual text-based interface for AI chatbots and assistants, and it was brought to life by the latest in RTX technologies.

'James' is built using similar NVIDIA ACE technology that we went hands-on with at Computex 2024 as part of an interactive game demonstration where you interact with digital humans in real-time to solve a mystery. This time, however, it's being used in the context of a virtual assistant or customer service agent.

'James' is also built on NVIDIA NIM microservices; Jensen Huang explains how it works and confirms that the visual 'RTX On' front end can connect to existing AI chatbots like ChatGPT.