Learn about how TweakTown tests and reviews hardware. As an Amazon Associate, we earn from qualifying purchases. TweakTown may also earn commissions from other affiliate partners at no extra cost to you.
Stay Updated
Follow TweakTown for breaking tech news, reviews, and daily updates.
NVIDIA caught scraping 'human lifetime' of YouTube videos per day to train AI
It was only last month we heard about Apple, NVIDIA and many other big name players in the AI race being caught up in an investigative report that found they all used a public data set containing YouTube video transcripts to train their respective AI products, which is a violation of YouTube's terms-of-service (TOS).
YouTube has said in the past that any "unauthorized scraping or downloading of YouTube content" is strictly prohibited, and it's especially prohibited when that data is then used for commercial projects. Last month, a Proof News investigation found NVIDIA, Apple, and other AI companies used an academic data set containing subtitles from more than 170,000 YouTube videos to train AI models, and now NVIDIA has been caught in the spotlight again with a report from 404 Media.
According to the publication that spoke with a former NVIDIA employee about the company's internal processes, employees were instructed to scrape videos from Netflix, YouTube, and other sources to add to the data sets that are being used to an AI model for NVIDIA's Omniverse 3D world generator, self-driving car systems, a "digital human" AI avatar product, and the Cosmos deep learning model.
Dell crushes its HR department with 12,500 layoffs in a single day
Dell announced its company is "getting leaner" with executives informing thousands of employees via a memo it has begun its effort into "streamlining layers of management."
President of global sales and customer operations Bill Scannell and global channels president John Byrne explained in the memo the company is attempting to grow faster in the market by concentrating its efforts on the development of "modern IT and AI" and that it can seemingly achieve that without 10% of it workforce, or approximately 12,500 employees. The layoffs are part of Dell's overall push to reduce its workforce to under 100,000 people.
These layoffs at Dell come only a matter of days after Intel announced it was laying off 15% of its workforce, amounting to 15,000 job cuts and a 26% stock drop. The massive stock drop is attributed to recently confirmed hardware issues with Raptor Lake chips, which is causing a fire internally for the CPU manufacturer. As for Dell, last year, the company planned on reducing its workforce by 5% but ended up reducing it by double that, or 13,000 people.
Continue reading: Dell crushes its HR department with 12,500 layoffs in a single day (full post)
World's most used operating system loses millions of users to Windows 11
Microsoft is currently pushing as many Windows 10 users as it possibly can to Windows 11 ahead of the announced end-of-support date in October 2025.
Microsoft announced it was ending support for Windows 10, the world's most popular operating system, in October 2025, which means the popular operating system will no longer receive security updates and support from Redmond. The announcement has caused the slow migration of millions of Windows 10 users over to Windows 11, but Windows 11 adoption isn't happening as fast as previous operating system generations.
It was only in July that Statcounter reported Windows 10 still doubles the market share of Windows 11, with Windows 10 commanding 65% and Windows 11 taking up 29.75%. Now, Statcounter has published new numbers that show Windows 10 has decreased to 64.99%, and Windows 11 has, for the first time, broken the 30% mark with 30.83% market share.
OpenAI creates ChatGPT content detector with 99.9% accuracy
Since the explosion of AI-powered models, academics, teachers, and many other professions have been rightfully concerned about tools such as ChatGPT being used to assist in the writing of essays and research papers.
The easy accessibility to powerful AI models such as ChatGPT has given rise to a substantial increase in the amount of AI-generated content across the internet. Unfortunately, it appears some of this content has already made its way into scientific journals within submitted papers, and classrooms with students taking advantage of these free tools to write assignments and papers.
To prevent the abuse of this technology in the academic setting, companies rolled out AI detection tools. Unfortunately, these detection tools proved unreliable, but according to a new article by The Washington Post, OpenAI, the creators of ChatGPT, have developed a new method that can detect ChatGPT-generated content with a 99.9% accuracy rate. The new system issues a watermark to ChatGPT-generated content that cannot be seen by the human user but detected by the detection too.
Continue reading: OpenAI creates ChatGPT content detector with 99.9% accuracy (full post)
Elon Musk files lawsuit against OpenAI CEO Sam Altman for tricking him
Elon Musk has filed a new lawsuit against OpenAI CEO Sam Altman for allegedly being tricked into funding what turned into a for-profit business.
The new lawsuit comes after Musk dropped a previous lawsuit back in June, and much like the previous lawsuit, Musk's legal team alleges Altman and OpenAI deceived Musk by "preying on Musk's humanitarian concern about the existential dangers posed by artificial intelligence." The new lawsuit is much larger than the previous one and alleges Musk agreed to fund OpenAI with $44.6 million between 2016 and 2020 under the assumption that OpenAI was going to use the overall initiative of open-sourcing artificial intelligence development.
However, according to the suit the Altman and OpenAI made promises to Musk they intentionally knew they weren't going to keep, and that Musk's funding was misappropriated for purposes that would result in for-profit business ventures. Musk has previously stated it was very important for OpenAI's research to be publicly available, or open-source, hence the name - OpenAI. Additionally, the lawsuit points out OpenAI's articles describing its mission to make its technology available to the public.
Continue reading: Elon Musk files lawsuit against OpenAI CEO Sam Altman for tricking him (full post)
AMD reportedly planning to release entry-level Radeon RX 7400 and 7300 RDNA 3 GPUs
AMD is not done with the RDNA 3 generation of Radeon graphics cards. The latest rumor from GPU leaker Komachi_Ensaka on X states that Team Red has a pair of entry-level cards coming, the Radeon RX 7400 and Radeon RX 7300.
These new GPUs will use the Navi 33 chip, as seen in the Radeon RX 7600 and Radeon RX 7600 XT. Both cards feature 2048 Stream Processors with a 128-bit Memory Interface built on TSMC's 6nm process node. The XT variant bumps up the memory capacity to 16GB, making it a more attractive option for AI and VRAM-intensive gaming.
Performance-wise, outside of ray-tracing, the Radeon RX 7600 combo sits between the GeForce RTX 3060 and the GeForce RTX 4060. With the new Radeon RX 7400 and Radeon RX 7300 set to use a cutdown version of the same chip found in the Radeon RX 7600, these cards would be GeForce RTX 3050 competitors.
NVIDIA preps GeForce RTX 4070 with regular non-X GDDR6 memory, specs leaked
Last week, we reported on a GDDR6X memory shortage that could potentially impact the availability of NVIDIA's GeForce RTX 40 Series graphics cards, specifically the popular mid-range GeForce RTX 4070, which is one of the most used GPUs among PC gamers, according to Steam's latest hardware survey.
According to new information from a reliable NVIDIA source, MEGAsizeGPU over on X, NVIDIA already has a GeForce RTX 4070 using GDDR6 (the non-X variant found in Radeon GPUs and cards like the RTX 4060 and RTX 4060 Ti). Although it's not confirmed, the leaker claims the new RTX 4070's GDDR6 memory could be rated at 20Gbps, making it slightly slower than the current GeForce RTX 4070 with GDDR6X memory.
The current model's 21Gbps puts the overall memory bandwidth at 504 GB/s, which would decrease to 480 GB/s. This small decrease shouldn't have a notable impact on the GPU's overall performance.
Foxconn still plans to ship NVIDIA's new GB200 AI servers on schedule in Q4 2024 amid issues
NVIDIA's next-generation Blackwell AI GPU has been reportedly experiencing some issues, which have caused some major delays that will see shipments kicking off in 2025... except for Foxconn.
Foxconn is the only company that has received enough of NVIDIA's new Blackwell GB200 Superchips, so it will be the only company pumping out new GB200 AI servers. According to UDN, they will be available in Q4 2024.
NVIDIA's new GB200 AI servers were meant to be shipping in large quantities in Q4 2024, but the yield issues with B200 forcing a redesign for B200A this time next year, is causing some major headaches for some of the biggest companies on the planet. UDN reports that it was found that the problem could stem from "unsatisfactory" yield rates of the advanced packaging required for GB200.
Elon Musk wants to 'give people superpowers', 'outperform a pro gamer' with his brain implants
Elon Musk has said he wants to "give people superpowers" with his brain-computer interface (BCI) implants through Neuralink, which the SpaceX and Tesla founder said will let you "outperform a pro gamer".
Elon wants to increase humans' "output rate." With Neuralink's BCI, he said that there's the potential to have "three, maybe six, maybe more orders of magnitude" of how fast our brain processes signals to the chip.
The SpaceX and Tesla boss sat down for the latest episode of the Lex Fridman podcast, with Fridman saying there would be "hundreds of millions" of people getting Neuralink implants in the "next couple of decades" to which Elon agreed.
Microsoft delays RPG Avowed to 2025 because it has too many games coming later this year
Microsoft and the Xbox team have confirmed that Obsidian's new RPG, Avowed, will not be released in 2024 and has been delayed to February 18, 2025. The Xbox social media team confirmed the delay in an announcement that began with the phrase "so many games!" hinting that the delay comes down to too many first-party and key third-party releases on the way this holiday season.
Xbox head Phil Spencer chimed in, adding, "We're proud of what the Obsidian team have accomplished with Avowed and want to make sure they have the best launch possible. We look forward to you enjoying it along with the rest of our lineup of games and updates coming this year and beyond."
According to insiders, development on Avowed is wrapping up, and the delay is purely due to timing - so that the highly anticipated RPG doesn't get lost in a sea of Xbox games.
Lenovo teases a more compact Lite version of its Legion Go handheld coming soon
We reviewed the Lenovo Legion Go handheld earlier this year. The Windows device powered by AMD's Ryzen Z1 Extreme Processor proved itself to be a powerful little unit for portable PC gaming. It can play titles like Far Cry 6 and Starfield alongside indie games and anything else that runs on a Windows 11 machine.
Performance aside, what separates the Lenovo Legion Go from competitors like the ROG Ally from ASUS is that it sports a large 8.8" QHD (2560 x 1600) 144 Hz IPS display with a physical design that includes removable controllers ala the Nintendo Switch, a kickstand, and ample cooling.
Big and bulky is not for everyone, so it's great to see that Lenovo is apparently preparing to launch a more compact Lite version of the Lenovo Legion Go with a 7-inch display.
Stable Diffusion AI image generation optimized for GeForce RTX, generates images in real-time
Stability AI's Stable Diffusion 3 is the latest version of its popular and powerful AI image-generation tool. It can deliver photorealistic results thanks to what it describes as the "most advanced text-to-image open model yet." For those following the rise of AI image generators, one thing that Stable Diffusion 3 brings to the table - or solves - is the ability to recreate repeating patterns and, yes, human hands.
It's impressive, and the model runs on hardware ranging from an Apple M2 Ultra to a GeForce RTX 4090-powered rig. Naturally, with its 24GB of GDDR6X memory and unlocked Ada Lovelace power, the latter is several magnitudes faster than any other "at-home" option. Still, with TensorRT acceleration and optimization, the GPU can render or generate images in real-time.
You can see this in action in NVIDIA's video, which it presented at SIGGRAPH 2024. The video showcases SDXL Turbo generating an image of "a hot rod, racing in the desert at sunset" in real-time, adding details to the image like adding a canyon to the background as it's being typed in. Impressive!
Silicon Motion SM2508: best perf-per-watt Gen5 SSD controller ready for next-gen AI PCs
Silicon Motion has just unveiled its new SM2508 Gen5 SSD controller, the best performance-per-watt Gen5 SSD controller at FMS 2024.
The new Silicon Motion SM2508 Gen5 SSD controller is the world's first PCIe Gen5 client SSD controller using TSMC's 6nm EUV process node, which provides a 50% reduction in power consumption compared to other SSD controllers using the 12nm process node.
The company says that the entire SSD uses less than 7W of power, the new SM2508 Gen5 SSD controller has 1.7x better power efficiency than Gen4 SSDs and up to 70% better than competitive Gen5 SSDs already on the market.
Logitech MIXLINE is a new and free audio-mixing app for creators and gamers
Logitech G's latest isn't hardware but new software aimed at PC gamers and creators. MIXLINE is described as an easy-to-use audio mixing solution, and best of all, it's free. Available now for Windows 10 and 11 users, it's designed to make dealing with more than one audio source a breeze.
"If you have more than one audio source while streaming, gaming, or even just wearing your headphones, then MIXLINE is for you," said Daniel Bowen, Senior Global Product Manager for MIXLINE.
An audio source can be Spotify, Google Chrome, Counter-Strike, Discord, etc. With MIXLINE, audio inputs are on the left side of the UI, and outputs are on the right side. This gives you direct control over routing and choosing where each signal goes. For creators with speakers, headsets, and output streams, it's a great tool to keep all their audio on track.
Kick streamer Adin Ross gives President Trump custom Tesla Cybertruck, vibes to tunes with him
President Donald Trump jumped into a Tesla Cybertruck with Kick streamer Adin Ross, and vibed to some beats. Check it out:
President Trump remarked that he loved the Tesla Cybertruck and its huge 18.5-inch display, swiping through Spotify to pick out a song after Ross asked him his top 3 artists. The first one that President Trump chose was California Dreamin' and then switched over to some Beach Boys, before playing one of my favorites: The Beach Boys' "God Only Knows".
Adin Ross presented the former President with some expensive, and beautiful gifts including a gold Rolex and a custom-wrapped Tesla Cybertruck. Ross said as he started his Kick stream on Monday after many doubted he landed an interview with President Trump, where he said: "Today is going to be the most important stream I've ever done".
Sad day for gamers: Intel's 36-core Beast Lake CPU, rumored to blow away AMD's X3D, is canceled
Remember the rumors about Intel's Beast Lake processors, silicon designed to be highly performant gaming chips, which sounded all very exciting? Well, that excitement has been well and truly quelled with the news that Beast Lake is apparently canceled.
That's the latest from YouTube leaker Moore's Law is Dead (MLID), who tells us that Beast Lake - which he first aired info on, and apparently represented the "full realization of Intel's Royal Core project" - is now canned.
The flagship for Beast Lake was, in theory, a chip with 4 'extra big' cores (very high performance) plus 32 efficiency cores, a mixture which could have been seriously potent in terms of driving gaming frame rates - or that was certainly the idea.
AOOSTAR's new external GPU dock with OCulink launched: built-in 400W PSU, costs just $149
AOOSTAR has just announced that its new AG01 OCulink external GPU dock is now available, with the external GPU dock rocking a built-in 400W power supply, with the GPU dock costing $149.
The new AOOSTAR AG01 is a GPU docking station that will connect to your laptop, Mini-PC, or desktop PC through an OCulink connection. There's no USB or M.2 SSD slots on this external GPU dock, so it's just purely an external GPU dock. But, what AOOSTAR is doing differently here is using a compact 400W built-in PSU.
The built-in 400W power supply has enough juice to support a graphics card that will use up to 250W of power, so you'll need to ensure that the GPU you're using in the AOOSTAR AG01 external GPU dock, uses less than 250W of power under load. AOOSTAR recommends the GeForce RTX 4070 or Radeon RX 7700 XT graphics cards for its new AG01 external GPU dock.
ASUS ROG Z13 Flow 2025 teased: new gaming tablet with AMD Strix Halo APU at 110W
ASUS is reportedly preparing a new ROG Z13 Flow 2025 compact gaming tablet, featuring AMD's heavily beefed-up Strix Halo APU.
The leaks are coming from thermal specialist and designer Sam Jiun-Wei Hu, who published some documents covering the ROG Flow design for 2025, which is when AMD's new Strix Halo APU will be released. The engineer was not shy in saying that the ASUS ROG Z13 Flow 2025 tablet was made for Strix Halo, with some huge details leaked.
The thermal design schematics show us how much space was saved by replacing the CPU and GPU with a single Strix Halo APU due, with both platforms using LPDDR5X memory, which already has a large vapor chamber covering it. Strix Halo on the new ASUS ROG Z13 Flow 2025 will have an even bigger one. We're to expect an improved cooling design that weighs less, but the overall design is thicker. Engineers have described this as less portable and holdable.
US recession scare wipes trillions from stock market, trading platforms are going down
The US woke up to over $2 trililon wiped away from the stock market, with virtually all companies impacted. But... a new development has occurred: trading platforms are going down as the market issues continue across the US.
We're talking about major players going down as the US stock market is falling, including Vanguard, Fidelity, and TD Ameritrade, as well as E-Trade, and Charles Schwab. Wall Street and the crypto market have seen a sea of red today, both bleeding trillions of dollars.
The Dow opened down over 1000 points, the S&P 500 was down over 4.2% this morning, and the Nasdaq Composite dropped over 6% according to CNN. The US stock market dropping so hard and so far, forced five trading platforms to go down as the markets were plunging into the red.
$2 trillion wiped from stock market implosion: Amazon, Meta, Google, Apple, NVIDIA all down
It's a hot August so far with the US stock market seeing a mammoth $2 trillion wiped away from stocks including Microsoft, Google, Apple, NVIDIA, Meta, and more.
There are fresh smells of recession fear in the air throughout the global financial markets, forcing the Federal Reserve to consider reducing interest rates before its upcoming meeting in September. The stock exchange isn't the only one in the red, with the cryptocurrency market seeing over $340 billion in outflows from crypto products, and that's just at the start of the day. The overall figure fells to $1.8 trillion, when it was over $2.1 trillion the day before.
Recession concerns have seen the Taiwan stock market plummet down into its worst losses since 1967, while the stock market in Japan has been rocked with its worst loss since 1987. The day has only started in the US, so we still have many hours to go and our eyes will be peeled to those digits.





















