Latest News - Page 919

Microsoft founder Bill Gates has penned a new essay titled "The Age of AI has Begun," where he details why he believes that the emergence of artificial intelligence is "revolutionary".

A Twitter user posted the link to Gates' essay, writing that Gates being involved in OpenAI and being bullish on the overall development of artificial intelligence is a "big" thing while also mentioning that it speaks volumes that Microsoft has been tracking the developments being made in the sector for so long. Notably, Gates' essay argues that artificial intelligence has the potential to transform the workforce, healthcare, and education while also being beneficial for the world's inequities.

Additionally, Gates writes that he was meeting with the OpenAI team as early as 2016, and the last time he sat down with them was in September 2022. Elon Musk founded the Silicon Valley darling OpenAI alongside several other Silicon Valley executives in 2015 and then pledged $1 billion to the company over several years. Musk stepped down as a board member and severed ties with the company in 2018, citing a competing interest between Tesla and OpenAI, wanting both wanting the best AI developers working at their companies.

Microsoft has rolled out a new feature for its newly released artificial intelligence-powered chatbot Bing Chat, with users now able to convert their words into images via a simple prompt.

Bing Chat, which uses the underpinning software running OpenAI's ChatGPT, has been upgraded to convert natural language prompts into images through the power of the DALL-E model, also created by OpenAI. According to Microsoft's announcement, BingAI has integrated an advanced version of the DALL-E model with the assistance of OpenAI and developers have put in safeguards to stop users from getting the AI to generate unsafe or offensive images.

These safeguards come from OpenAI's own policies and principals that it has outlined in its AI guidelines. Furthermore, Microsoft's Corporate Vice President and Chief Marketing Officer Yusuf Mehdi explained that when the AI detects that a potentially harmful image would be generated from a prompt the system blocks the creation of that image and warns the user of what has happened and why. Mehdi adds that Microsoft has added a modified Bing icon at the bottom left corner of each image to distinguish which images were created using Bing's Image Creator.

Developing a ChatGPT client and interface for a modern platform or device is all the rage, but what if you transported the powerful bit of AI technology back to the MS-DOS powered 1980s?

That's precisely what Yeo Kheng Meng has managed to do, even though DOS does not have native networking capabilities. The machine in question is the vintage IBM 5155 Portable PC, first released in 1984, powered by a (not-so-powerful anymore) 4.77 MHz Intel 8088 CPU and 640KB of RAM.

The first major hurdle was finding a compiler that could support the roughly 40 years of history between this old PC and ChatGPT - with the Open Watcom C/C++ compiler fitting the bill. During development, testing was carried out via a virtual machine running MS-DOS 6.22, which makes sense. However, it would have been cooler if builds were stored on giant floppy disks and transferred to the old IBM PC every few hours.

The PowerColor Liquid Devil Radeon RX 7900 XTX is now available with a steep MSRP of USD 1499. But, with that, you've got AMD's flagship RDNA 3 offering with high-end water cooling designed for extreme overclocking.

And that's precisely what renowned overclocker Roman "der8auer" Hartung has done, with his latest video putting the PowerColor Liquid Devil Radeon RX 7900 XTX to the test. And yeah, as per the headline, the GPU clocks can be pushed as high as 3.2 GHz - though the result is a power draw of 650W.

As a custom GPU the PowerColor Liquid Devil Radeon RX 7900 XTX features three 8-pin connectors, a new 14-layer high TG PCB, a 17-phase VRM design, and a 2+2 phase for the 24GB VRAM. It's the sort of build designed not only for extreme overclocking without the need to go the liquid nitrogen route but also unlocking the full potential of Navi 31 and the flagship RDNA 3 GPU.

COLORFUL has announced a new line of memory with DD5 and DDR4 modules, with the Battle-Ax Redline series sporting funky deep red accents that look like the memory has survived an encounter with a Velociraptor. COLORFUL describes the look as "aggressive" and "beastly," so well done to the design team on that one.

Designed for the latest generation of Intel and AMD platforms, the memory modules feature a low-profile heat spreader with the Battle-Ax Redline DDR5 memory sporting a white and red finish and the Battle-Ax Redline DDR4 memory sporting a black and red finish.

COLORFUL Battle-Ax Redline DDR5 memory arrives with a high-performance DD5-6000 MHz frequency and CL30 latency. Available in 32GB kits (2 x 16GB), there's also support for both XMP 3.0 and AMD EXPO memory overclocking. Based on the looks and features, you can probably guess these kits are aimed at PC gamers.

Thanks to VideoCardz.com, we've got our first look at an entry-level A620 motherboard for AMD's new AM5 platform introduced with the new Ryzen 7000 CPU range. A620 will be the affordable entry-level option below B650 and X670 series offerings.

A620 boards will greatly interest those looking to create a new Ryzen build that doesn't break the bank.

With the first shots of the upcoming ASRock A620 motherboard (A620M-HDV/M.2), we've also got some idea of what's being cut to pave the way for a more affordable range of AMD-compatible boards.

Kingston FURY has announced some new additions to its DDR5 memory lineup sporting stylish white heat spreaders, which would pair nicely with builds that incorporate the latest color trend.

Kingston FURY BEAST DDR5 memory sports the new white heat spreader with Kingston FURY BEAST DDR5 RGB adding a dose of lighting with 18 built-in lighting effects that can be tweaked further via Kingston FURY CTRL software.

The main draw here is the additional speed that comes with making the jump to DDR5. Kingston FURY BEAST DDR5 supports plug-and-play speeds of 4800 MT/s, with Intel XMP 3.0 or AMD EXPO Certified kits offering overclocking up to 6000 MT/s. Kingston FURY BEAST DDR5 is available in 8GB, 16GB, and 32GB single modules, kits of two up to 64GB, with kits of four coming next month.

An administrative law judge has ordered Microsoft to comply with all Activision merger document requests made by the FTC's complaint counsel in a recent motion.

Recent filings indicate the Federal Trade Commission is having trouble getting specific documents from Microsoft and Activision. FTC complaint counsel (the lawyers representing the FTC that seek to block the merger) filed a motion to compel production of documents. Statements from the FTC say that Microsoft was largely uncooperative and "have refused to produce anything at all for 24 requests."

Microsoft and Activision-Blizzard counsel argued against the motion and said that the requests (known as RFPs in this context) are "duplicative and unduly burdensome" because they'd require the companies to "re-review millions of documents previously deemed least relevant." The respondents (Microsoft and Activision) argue that complaint counsel has received 17 million pages in a massive second request.

ChatGPT can be put to a lot of uses, as we've seen - such as its recent introduction to gaming chat platform Discord - and the latest one is the AI being recruited to help consumers build a PC over at major US retailer Newegg.

In fact, Newegg is putting the AI's skills to use for a raft of things across the retailer's site, but primarily to beef up the PC Builder tool, which is now labeled as 'Build with AI' - and a clear caveat that this is still a beta experience.

As you might expect, if you want to put together a PC, ChatGPT will help provide recommendations for various components that will go together well, or fit a certain budget.

Last week, ChatGPT went down in what first appeared to be a global outage, but was in fact, due to OpenAI shutting the service (temporarily) after finding a bug that made user chat histories visible to the public - and more besides, we've now found out, worryingly.

In what was later described as a "significant issue" pertaining to a bug present in an open source library, titles of chat histories, and possibly the initial messages, were in some cases seen by other users who were active at the same time as a person having a chat.

The bug is now fixed, but after fully investigating what happened here, OpenAI has admitted that some credit card details of a small subset of users were possibly exposed.

Microsoft has swiftly deployed emergency fixes for a security flaw in Windows 11 that affected the Snipping Tool (and the Snip & Sketch app in Windows 10, too).

Those screenshot-grabbing and editing utilities were blighted by an issue whereby cropped data in PNG image files wasn't being properly overwritten, playfully named the "acropalypse" bug.

In other words, when users crop a file, the part of the picture discarded could potentially be recovered and scrutinized by someone exploiting the flaw.

Earth was just smacked with a very powerful blast from the Sun that caused the spawning of auroras more south than they typically appear.

The event occurred on March 25 and was confirmed by the U.S. National Oceanic and Atmospheric Administration (NOAA), which issued a statement declaring a severe geomagnetic storm warning and an image showing where the impact would be focused. NOAA's map of Earth shows a large swath of North America being right in the impact zone, which coincides with the plethora of gorgeous aurora images that were posted online during and following the storm.

NOAA ranked the solar storm at G4 on the 5-grade scale that is used to define the intensity of a geomagnetic storm. Notably, the impact of the storm was so strong that there were reports of the Northern Lights appearing as far south as New Mexico in the U.S. So, how does this happen? Auroras such as the Northern and Southern Lights typically appear close to the Earth's poles as the planet's magnetic field deflects charged particles around the planet to its poles, where they reach the weakest points of Earth's magnetic field.

The Turtle Beach Stealth Pro is the latest premium gaming headset from the peripheral maker and audio company. It supports all gaming devices, Bluetooth, active noise-cancellation, and other notable features.

Regarding active noise cancellation, Turtle Beach claims that it's "the most effective, competitively benchmarked noise-cancellation available in a gaming headset," with dedicated internal microphones that lead to a reduction of up to 25dB. This makes the new Turtle Beach Stealth Pro gaming headset something you could effectively take on the go or use while traveling.

This tech extends to the noise-canceling TruSpeak boom microphone, which is also detachable. With support for Xbox Series X|S, Xbox One, PS5, PS4, Windows PC, Mac, Nintendo Switch, and mobile/tablet devices (via Bluetooth), it's certainly versatile.

The Flux Keyboard is a fascinating new peripheral that has blasted through its Kickstarter goal of AUD 200,000, where it was funded in eight minutes and hit AUD 1 million in four hours. And it's not hard to see why, as this modular keyboard integrates a fully dynamic display with the feel of a mechanical keyboard.

The display allows any icon to be placed on any key, paving the way for multi-language support and application-specific layouts and styles. Created by the Australian-based Flux Group, the high-definition IPS display sits below a modular, transparent layer that can display any image or video and can even showcase reactive animations as part of a singular background.

How does this translate to the precision and tactile feel you get from a mechanical keyboard? On that front, the Flux Keyboard uses custom Maglev switches that rely on magnets to deliver a tactile or linear feel similar to what you get from the spring in a traditional mechanical key switch.

Intel is releasing new Game On Drivers for its Intel Arc range of graphics cards at a steady clip, with the latest release adding support for the Resident Evil 4 Remake on PC.

With its low price point, the Intel Arc A750 is positioned as an affordable GeForce RTX 3060 competitor. With that, the new driver update Intel showcases the 1080p and 1440p performance in Resident Evil 4 with the Arc A750 using the game's "High" settings outperforming the RTX 3060.

Even though internal benchmarks should always be taken with a grain of salt, Intel adds that at 1440p, you're looking at an impressive 74% more performance per dollar with the Intel Arc A750, and it does so with a very solid 78 frames per second.

We've all lived through the GeForce RTX 30 Series generation, where the rise of cryptocurrencies and mining drove prices up and increased scarcity for GPUs. So this new interview with The Guardian is interesting in that it sees NVIDIA publicly distance itself from something that has negatively affected the PC gaming space.

Speaking with Michael Kagan, Chief Technology Officer at NVIDIA, he explicitly states that the whole crypto boom didn't "bring anything useful for society." Of course, this statement is also born from the current AI boom where NVIDIA hardware sits at the heart of AI technology like ChatGPT.

Even though crypto led to large volumes of sales for NVIDIA and increased scarcity and availability in most markets, the company attempted to minimize the impact on desktop sales by introducing LHR or "Lite Hash Rate" technologies in models like the GeForce RTX 3060.

When AMD announced its new flagship RDNA 3 GPUs last year, the Radeon 7900 XT and 7900 XTX, the company also briefly teased that it was developing FSR 3. Its next-generation FidelityFX Super Resolution 3 technology would, in effect, present an alternative to NVIDIA's DLSS 3 - specifically Frame Generation.

With FSR 2 being an open DLSS Super Resolution alternative, FSR 3 will follow suit - with the technology appearing at GDC 2023. For the still-in-development FSR 3, AMD aims to achieve a 2x performance increase over FSR 2 by creating new frames.

AMD outlined a high probability that each interpolated or generated pixel in a frame will have a sample. With no feedback loop, each interpolated frame will only be shown once. And with that, AMD notes that the recommendation for FSR 3 is that the input should be at least 60 frames per second.

Have you ever turned on your Android phone and been immediately blinded because the screen brightness was set too high? We're willing to bet that you have, but now it looks like Google is working on a fix for that infuriating problem.

According to a new report, Google has a fix in the works via a future Android software update but it will be a while before you can actually install it on your phone.

The problem comes when you first wake your phone because it doesn't yet know how bright the room is. The ambient light sensor doesn't kick in until after that point, meaning it doesn't have time to turn the brightness down if you're in a dark room in the middle of the night, for example.

WhatsApp user who wants to make sure that they stay fully up to speed on the latest goings on and new features? You could soon be getting all of that information right inside WhatsApp.

The Meta-owned instant messaging service has started to roll out its own group chat that will allow people to get information about WhatsApp including new features and more.

This feature has been in the works for a little while according to the folks at WABetaInfo, but some people have now started to see the chat appearing in their list of messages for the first time. Unfortunately, there doesn't currently appear to be any way to manually enter that chat, so it's a waiting game we're afraid.

Tesla, SpaceX, and Twitter CEO Elon Musk reportedly attempted to takeover OpenAI in 2018 but ended up cutting ties with the company after his plan was rejected.

The news comes from Semafor, which recently published an article that claims Musk, a founder and board member at the time, said to Sam Altman, current OpenAI CEO, and other fellow board members that OpenAI was losing to Google. Musk then followed up by offering to lead OpenAI himself, which was rejected by Altman and the other board members. According to the article Musk then stepped down from the board and pulled out his planned donation to the company.

At the time of Musk's departure, the Tesla CEO cited a conflict of interest between Tesla and OpenAI as the reason for his leaving. Musk said that Tesla was creating its own artificial intelligence system and would be competing for talent with OpenAI. However, according to the eight people familiar with the inside story that spoke to Semafor, a power struggle and rivalry started between Musk and Altman following Musk's rejection, which resulted in Musk stepping down from the board of directors and immediately stopping payments of his promised $1 billion donation that would have been rolled out over many years.