Artificial Intelligence - Page 53

NVIDIA is reportedly working on a new AI GPU dubbed "Skinny Joe," which could be on its way to China with a 700W TDP attached.

The new NVIDIA "Skinny Joe" AI GPU has been spotted with a reference Dev_ID listing, just above the new NVIDIA L20 AI GPU and NVIDIA GeForce RTX 4090 D graphics card. The GeForce RTX 4090 D is a new tweaked gaming GPU that uses the same AD102-based GPU as the GeForce RTX 4090 but tweaked to allow it into China with new US export restriction rules.

It would be funny if NVIDIA's new AI GPU for China were dubbed "Skinny Joe" because it has been cut down for the Joe Biden administration over all these US export restrictions. "Skinny Joe" could be named from it being a "skinny" aka cut-down version of a regular AI GPU like the H100, refreshed H200, or even the newer B100 AI GPU that's on the way in 2024.

AI is the theme of this year's upcoming CES 2024, with one of the biggest South Korean tech giants, Samsung Electronics, is set to showcase brand-new AI-powered home appliances.

Samsung's new AI-powered home appliances aim to make any lifestyle consumer more convenient and efficient. The world's biggest tech show will be held in Las Vegas from January 9th to January 12th. Samsung's newest kitchen products, applications, and features are powered by artificial intelligence (AI). These new products with SmartThings connectivity will enable consumers to create their own home-smart ecosystem that delivers better experiences in the kitchen and everyday life.

At the event, Samsung plans to display their new Bespoke Jet Bot Combo and Bespoke Jet AI vacuum cleaners, which will both feature enhanced AI capabilities. The Bespoke Jet Bot Combo will use precise AI technology to help detect different types of flooring, from hardwood to carpet and clean them accordingly. The Bespoke Jet AI, however, has a mode that will categorize and optimize to take cleaning one step further, adjusting power based on brush load and pressure.

NVIDIA has been forced to tweak its AI GPUs to meet constantly updated US sanctions, with its new H20 AI GPU set to enter mass production in Q2 2024, ready for the Chinese market.

We reported -- last year, now that we're in 2024 -- that NVIDIA was in the process of making new AI GPUs for the Chinese market. These AI GPUs included the HGX H20, L20 PCIe, and L2 PCIe AI GPUs, which acted as alternatives to the AI GPUs that NVIDIA has on the market already.

The new H20 AI GPU is a cut-down version of the H100, with 96GB of HBM3 memory with up to 4.0TB/sec of memory bandwidth. There are 296 TFLOPs of compute power, and the H100 AI GPU die with a performance density of 2.9 TFLOPs/die, compared to the 19.4 TFLOPs/die of the regular H100 AI GPU.

NVIDIA is splashing down cash on next-gen HBM3e memory, where the AI GPU leader is securing enough stock of the new High-Bandwidth Memory for its upcoming Hopper H200 AI GPU and its next-gen Blackwell H200 AI GPU, which both arrive in 2024.

According to Korean media outlet Biz.Chosun, NVIDIA has placed orders for massive quantities of new HBM3e memory from both SK hynix and Micron. It was only yesterday that SK hynix announced its next-gen HBM4 memory has entered production, ready for 2024, with NVIDIA scooping up as much HBM3e memory as it can right now.

NVIDIA has reportedly pre-paid for around 700 billion to 1 trillion won for its HBM3e orders, which works out to a staggering $775 million USD. Remember: this is just the pre-payment and not the full amount... so we're talking an easy $1 billion (probably much more) in early HBM3e memory orders from NVIDIA alone.

As part of LG's "Zero Labor Home" vision, the company has announced a new cutting-edge robotic smart home Artificial Intelligence (AI) agent that can move, learn, comprehend, and engage in complex conversations. Don't let the cute robotic animal look fool you; LG is collaborating with Qualcomm Technologies to develop this new AI agent as a smart solution to enhance "users' daily lives."

For something small and "cute," LG has opted not to refer to it as a robot or even give it a name. The press release refers to the new AI-powered technology as the 'LG smart home AI agent.' Perhaps this is a way to differentiate it from popular robotic pets like Sony's Aibo Robot Dog.

Even though the two-legged design with wheels is in line with robotic pets or companions we've seen in the past, LG's AI agent has the intelligence to interact with users while expressing itself through movements.

NVIDIA's current-gen H100 AI GPU is deployed across the world in countless data centers, supercomputers, and AI platforms... with over 3.5 million units expected to ship in 2024.

Market analysis source, Stock Talk, has published an estimated figure on the power consumption of all of these NVIDIA H100 AI GPUs in the wild. In the tweet, it said: "By next year, NVIDIA's ~3.5 million units of deployed H100 GPUs will consume a whopping 13,000 GWh of electricity annually, which is greater than the power consumption of some entire countries like Guatamela and Lithuania".

This is something we have heard before, where crypto mining installations were using as much power as countries, where I reported back in 2020 that one particular Bitcoin mining facility in Kazakhstan was using enough power for 180,000 homes. Another story in 2021 pointed to the fact that Bitcoin mining was using more electricity than countries like the Netherlands and Argentina.

Apple's "legendary designer," Jony Ive, and OpenAI founder Sam Altman have now acquired another Apple design veteran: Tang Tan.

Tan will join LoveForm and will help the company shape the look and capabilities of its new products, according to Bloomberg's sources. Altman has been at the forefront of AI for a while now, with ChatGPT used by millions across the planet, where Altman has plans to provide the software side of things -- AI -- to new devices.

Ive worked with Steve Jobs at Apple and created some of the most iconic designs of our times, with devices like the iPhone, iPad, and Mac all getting that Ive touch. Tan will be leading a new hardware engineering product at LoveForm, leaving his former position as Apple's vice president of iPhone and Watch product designs.

MemryX has just announced the availability of production-level silicon for its new AI accelerator, the MX3.

You probably haven't heard of MemryX Inc. before, but they're a new startup that specializes in accelerating artificial intelligence (AI) processing for edge devices. The company received its new MX3 silicon from TSMC (Taiwan Semiconductor Manufacturing Company), and within a month of getting it, it's ready to show its new MX3 AI accelerator.

MemryX has announced that it will be showing off its new MX3 AI accelerator running "hundreds of unaltered AI models" at the CES 2024 show in Las Vegas, which will be running between January 9-12, 2024... just weeks away now.

Intel CEO Pat Gelsinger is really pushing the fighting words against NVIDIA and its utter AI dominance, where he's said that the company got "extraordinarily lucky" with its AI business.

In a recent chat with MIT, Gelsinger said: "Jensen worked super hard at owning throughput computing, primarily for graphics initially, and then got extraordinarily lucky. They didn't even want to support their first AI project".

The fun didn't stop there because NVIDIA's Vice President of Applied Deep Learning Research, Bryan Catanzaro, who used to work for Intel, tweeted: "I worked at Intel on Larrabee applications in 2007. Then I went to NVIDIA to work on ML in 2008. So I was there at both places at that time and I can say: NVIDIA's dominance didn't come from luck. It came from vision and execution. Which Intel lacked".

Microsoft Copilot, the AI assistant available for Windows and Edge users, just got a fun update in time for the holiday season: the ability to generate a song from your input. The ability arrives thanks to a new partnership with Suno.ai, and accessing the feature is as simple as enabling the Suno plug-in and then describing the type of song you want to hear.

Microsoft describes the ability to turn ideas into songs as something for everyone, regardless of musical background. Suno.ai is a leader in artificial intelligence-based music creation, and they have their own standalone AI-powered web app you can use to generate songs. So, if you're looking for a hip-hop song about your loved one for the holidays, be sure to give this a go.

Simplicity is the key, as Suno.ai can generate a song from a single sentence, like this little song called Pickles and Strings, a folksong about a man who loves pickles. What's cool is that you can specify the genre, tempo, the types of instruments you want to hear, and more.

A group of three graduate students at Stanford built an AI model that is capable of identifying the location of photographs by simply looking at the picture.

The developers of the AI claim the model, which has been trained on Google Street View data, is capable of correctly predicting the location of where a photograph was taken with 95% accuracy and within 25 miles of the real location. The team noted that this new AI model has real-world applications rather than just being used to track the location of people, as the new model could be used to promptly identify roads with downed power lines or biological surveys of regions.

The trio of engineers got the idea for the new AI model by being fans of the online game GeoGuessr, which places a player somewhere around the globe using Google Street View. The player is then required to guess their location by dropping a pin on the map, and the closer the players get to their actual location, the higher the score they get. To create the new AI model officially called "Predicting Image Geolocations" (PIGEON), the team of engineers used OpenAI's CLIP neural network, which learns about images through text.

A group of scientists taught an artificial intelligence-powered algorithm to detect autism with photos of children's eyes.

A new study published in the JAMA Network Open earlier on December 15 details a team of researchers training an AI algorithm with photos of children's eyes to detect childhood autism. The paper explained that the AI model was trained on 1,890 retina images captured from 958 participants with an average age of 7.8 years.

Notably, half of the participants in the study were already diagnosed with Autism Spectrum Disorder (ASD), and once the model finished analyzing each of the retina images, it was asked if it could spot ASD within the images. According to the study, the AI was able to identify which study participants had autism and those that didn't with "100% accuracy". The team behind the study notes that it limited the participants to ages between 4 and 18, but they believe the algorithm could work on even younger children.

Intel has had a lot to talk about lately; at its recent AI Everywhere event, the company unveiled its new 5th Gen Xeon "Emerald Rapids" CPU platform, the new Core Ultra branding and their Meteor Lake CPUs, as well as its next-gen AI accelerator: Gaudi 3.

But now we've got reports from Korean media that Intel's upcoming Gaudi 3 AI accelerator could compete with NVIDIA's current-gen H100 AI GPU and AMD's just-announced Instinct MI300X AI accelerator. Intel CEO Pat Gelsinger said: "Our Gaudi roadmap remains on track with Gaudi3 out of the fab, now in packaging and expected to launch next year. In 2025, Falcon Shores brings our GPU and Gaudi capabilities into a single product".

We know that the new Intel Gaudi 3 AI accelerator will have 1.5x more performance than the previous-gen Gaudi 2, as well as 4x more BFloat16 performance, twice the compute power, and a 50% bump in memory capacity from 96GB on Gaudi 2 to a much better 144GB on Gaudi 3. We should expect Intel to be using HBM3 or the faster HBM3e memory standard, but that still needs to be confirmed.

Moore Threads has just announced its new MTT S4000 AI GPU that will work "seamlessly" with NVIDIA's CUDA framework, thanks to an in-house MUSIFY translation tool.

The new Moore Threads MTT S4000 AI GPU features MTLink 1.0, which is a new multi-GPU connector that we don't have photos of, but it sounds similar to NVIDIA's in-house NVLink connector. The new MTT S4000 AI GPU has 48GB of GDDR6 memory with up to 768GB/sec of memory bandwidth, compared to the 48GB of memory on NVIDIA's RTX 6000 GPU with 960GB/sec of memory bandwidth.

Moore Threads is passively cooling its new MTT S4000 AI GPU, with 25 TFLOPs of FP32 single-precision compute performance, 100 TFLOPs of FP16/BF16, and 200 TOPS in INT8. This is all found on a PCIe Gen5 x16 connector, too.

AMD launched its new flagship AI GPU and accelerator, the Instinct MI300X, earlier this month. During the launch event, AMD provided charts and data indicating that the MI300X outperformed NVIDIA's powerful H100 GPU in several tests.

It was up to 60% faster in a direct 8 to 8 server comparison; however, NVIDIA followed up with its own data. NVDIA's report shows (with data) that AMD did not use the latest NVIDIA TensorRT-LLM kernel optimizations for NVIDIA Hopper, nor did it show real-world MLPerf server scenarios where DGX H100 hardware delivers impressive "inferences per second" results.

Long story short, NVIDIA's response claims that if AMD benchmarked the H100 GPU properly, it would show that Team Green's flagship was 2X faster. So, where does that bring us to? AMD has reached out with an update showing that the MI300X's performance advantage has increased thanks to its own optimizations. And it's now accounting for latency.

A US news station has announced its plans to release the very first AI-generated news anchors, which will be free and streamed on ad-supported services such as Crackle, Tubi, or Pluto.

The announcement comes from LA-based news station Channel 1, who said that they will be using AI-powered news anchors instead of real human people in what will likely be a world's first for a syndicated news station. A new report from the Daily Mail revealed that Channel 1's news segments will use a combination of AI-generated humans and digital avatars that were created through the use of real actors. Notably, Channel 1 will still be using real humans for its most important news stories.

The founder of Channel 1, Adam Mosam, spoke to The Daily Mail and said that the intention behind releasing AI-generated news anchors was to "get out in front and create a responsible use of technology." As for concerns regarding fake news and the potential impact on society for falsifying the news through the power of AI, Mosam says that Channel 1 intends to be extremely transparent with viewers about what footage is original and what has been generated by artificial intelligence.

[UPDATE AMD has responded to the below with updated benchmark results and new optimizations of its own to show that the MI300X still has the performance advantage - read all about it here]. AMD, like all of the big players in the chip game, is going all in on AI hardware, and with the company's recent flagship MI300X GPU launch, it made some bold claims that compared performance between the MI300X and NVIDIA's powerful H100 GPU. Up to 20% faster than the H100 in a direct 1 to 1 comparison and up to 60% faster in an 8 to 8 server comparison.

Every slide in AMD's 'Advancing AI' presentation that covered the performance of the AMD Instinct MI300X to the NVIDIA H100 Tensor Core GPU shows the MI300X coming out on top or, at worse, performing on par. And with that, NVIDIA has taken the time to present its own results showing that the H100 GPU is 2X faster than the MI300X.

"At a recent launch event, AMD talked about the inference performance of the H100 GPU compared to that of its MI300X chip," NVIDIA's Dave Salvator and Ashraf Eassa write. "The results shared did not use optimized software, and the H100, if benchmarked properly, is 2X faster."

Intel CEO Pat Gelsinger has some fighting words to NVIDIA and its CUDA standard, where he said: "You know, the entire industry is motivated to eliminate the CUDA market".

Gelsinger claimed that inference technology will be more important than training for AI, just as it launches its new Intel Core Ultra and Intel 5th Gen Xeon "Emerald Rapids" CPUs at its "AI Everywhere" event in New York City. Gelsinger continued: "We think of the CUDA moat as shallow and small. Because the industry is motivated to bring a broader set of technologies for broad training, innovation, data science, et cetera".

Gelsinger used examples like MLIR, Google, and OpenAI, where he said they're moving towards a "Pythonic programming layer" that makes AI training more open.

NVIDIA is already dominating the AI GPU market, but now analyst Ming-Chi Kuo is expecting that the insatiable demand for AI GPUs will see NVIDIA AI GPU shipments surge 150% year-over-year in 2024.

NVIDIA has already reported absolutely stellar financials driven by its AI GPU dominance, where Team Green has a commanding 90%+ of the AI GPU market share. The company enjoyed an amazing 2023, but in the coming years the company is expecting to drive $300 billion from its AI business alone.

Tianfeng Securities analyst Ming-Chi Kuo says that NVIDIA plans on boosting the volumes of AI GPU shipments in 2024; here's what the analyst expects to see in 2024:

AMD's Chief Financial Officer Jean Hu has said that the AI PC will drive consumer sales in 2025, but we won't see the fruits of AI PC labor in 2024 as much as we will in the year after.

The AMD CFO was questioned during the Barclays Global Technology Conference, where she was asked about the future of PC markets and the upcoming trend of the "AI PC" and how much of an influence AI PCs will have in boosting sales. Hu said that the PC industry is moving towards recovery in the post-pandemic world, with revenue forecasts and consumer demand to increase in 2024-2025.

AMD CFO Jean Hu said during the Barclay Global Technology Conference: "Going forward, you will see the typical seasonality in Q1 and then, you know, next year, based on industry and it's economic-sensitive sector, too. So, we don't know what the macroeconomic situation will be. But in the end, the PC market is stabilized".