Artificial Intelligence News - Page 2

The insatiable demand for AI GPUs and all things AI throughout 2023 has seen HBM memory makers enjoying a gigantic 500% spike in the average selling price of HBM chips. Not bad at all for HBM makers Micron, Samsung, and SK hynix.

HBM chips are one of the most important parts of an AI GPU, with the likes of AMD and NVIDIA both using the bleeding edge of HBM memory on their respective AI GPUs. Market research firm Yole Group has now chimed in, predicting that the HBM supply will grow at a compound annual rate of 45% from 2023 to 2028, with HBM prices expected to remain high "for some time," considering how hard it is to scale up HBM production to keep up with the crazy demand.

Samsung and SK hynix are the dominant players in the HBM memory-making business; between the South Korean companies, they have 90% of the HBM market share, leaving Micron with the scraps. SK hynix and TSMC (Taiwan Semiconductor Manufacturing Company) are teaming together, so you've got some interesting developments that will continue to roll on as the months (and years) go on.

NVIDIA is arguably the leader in creating GPU hardware for AI, and it shouldn't come as a surprise that it's using AI to accelerate chip design and production. Staying ahead of the game is a part of the picture; NVIDIA is now the world's sixth-largest company thanks to the rise of AI.

NVIDIA's custom LLM (Large Language Model) is called ChipNeMo, adapted from Meta's Llama 2, and is trained on the company's vast amounts of architectural data, documentation, and source code. ChipNeMo was first unveiled in late 2023 and is also used as a tool for training engineers via its chatbot functionality.

According to Bryan Catanzaro, NVIDIA's vice president of applied deep learning research, designing a new GPU takes close to 1,000 people working together. Access to an AI chatbot with all that data is an invaluable tool. The question now becomes, how much - if any - of the design for upcoming GPUs is being designed by ChipNeMo?

Robocallers have been a problem for many years now, with some individuals receiving more calls than others. But with the power of artificial intelligence (AI), the frequency of these calls has increased and seemingly reached its boiling point when an AI-powered robocaller imitated President Joe Biden.

These robocalls are likely familiar to everyone who owns a phone, and with the exponentially increasing power of AI-powered tools, scammers are becoming more and more creative in devising ways to scam individuals out of their hard-earned dollars.

The Federal Communications Commission (FCC) is aware of this issue and was pushed to action when an AI-generated robocall imitating President Joe Biden made headlines as the voice on the end of the line was telling New Hampshire Democrats not to vote for Biden in the upcoming presidential primary.

NVIDIA CEO Jensen Huang has been elected to the National Academy of Engineering (NAE), which is one of the highest academic distinctions awarded to engineers. The NAE credits NVIDIA CEO Jensen Huang "for or high-powered graphics processing units, fueling the artificial intelligence revolution".

It wasn't too long ago that Jensen won the "Best CEO of 2023" by The Economist, but now his "outstanding contributions" to the world of engineering were for both the GPU and AI GPU industries. The NAE takes in a bunch of new members each year after they spend multiple months selecting them.

There are specific qualities, achievements and contributions that potential NAE members are judged by, with the academy only honoring people who have made outstanding contributions in the following fields:

I think by now we all know that the "AI PC" and AI in general is the future of technology, but we haven't been able to run generative AI tasks locally -- as the likes of ChatGPT is run in the cloud -- but that is slowly changing. The entire PC industry is being dragged into AI whether they like it, or not.

AMD and Intel both have processors on the market with NPUs (Neural Processing Units) that are dedicated parts of the chip that are built for AI workloads, but you can also use your CPU or GPU to run generative AI tasks, or a combination of them. CPUs and GPUs aren't exactly optimized for AI workloads, which is where the NPU comes in... the faster it gets, the better AI experiences we'll have on future-gen hardware.

Analyst firm IDC is making some predictions to the AI PC market, now that AI processors are here with NPUs inside of them from AMD and Intel.

Well, I didn't think I'd be writing this story at this stage of the AI game, but here we are: a Russian man has used ChatGPT-4 with a custom AI bot he wrote to scour Tinder for the best date... and he ended up getting engaged to one of the women, all thanks to AI.

Aleksandr Zhadan had a bad breakup with his then-girlfriend in 2021, shortly after he started dating women around Moscow and St. Petersburg in his native country of Russia. He said the dates ended in disaster each time, but that all changed in 2022 when he got access to OpenAI's then-new GPT-3 API.

He then developed an AI bot that would automatically swipe right on profiles that featured two pictures, later integrating GPT-3 into his bot and allowing the LLM to chat with women that met his parameters, all without him touching a keyboard or smartphone. Zhadan used the prompt, "You're a guy talking to a girl for the first time. Your task, not immediately but to invite the girl on a date".

Microsoft has announced improvements to Copilot in what's the latest step forward for the AI (no doubt the first of many this year).

In a blog post, the software giant told us that it has made Copilot more streamlined, with a cleaner and slicker look, as well as introducing a carousel of suggested prompts to stoke your curiosity (perhaps) to explore those topics.

What this means is that the interface for Copilot does indeed look tidier and neater, with that carousel occupying center stage in the middle of the page.

Generative AI and autonomous agents will make their way into government-run programs, including the military, so this little experiment with some of the most well-known LLMs is interesting... to say the least. Like the classic 1983 film WarGames, various AI models were pitted against each other in multiple wargame scenarios to see how they'd react and make decisions.

You can read the full results of the experiment in a new paper titled 'Escalation Risks from Language Models in Military and Diplomatic Decision-Making' from several high-profile universities and institutes. Eight different "autonomous nation agents" using the same LLM were put in a wargame scenario - with a separate AI model summarizing the simulated world's outcomes, consequences, and state.

Turn-based tabletop gaming, except to see what would happen if AI was in charge of every military asset, including nuclear weapons. And yes, a few of the LLMs went nuclear and started dropping bombs.

NVIDIA's new Grace Superchip has just been benchmarked, giving us a look at NVIDIA's first-ever in-house server CPU that's based on the Arm architecture.

The Barcelona Supercomputing Center and the State University of New York have benchmarked the NVIDIA Grace Superchip, which doesn't beat Intel's new Xeon "Sapphire Rapids" CPUs with 48 cores and 96 threads per processor. NVIDIA's new Grace Superchips will be a competitive data center and HPC processor, which comes from its efficiency.

A single Grace CPU packs 72 cores and 480GB of LPDDR5X memory, while the Grace Superchip features 2 x Grace CPUs combined on a single board that provides a huge 144 cores and an even bigger 960GB of LPDDR5X memory. Impressive stuff to see, and just to once again point out the fact that this is an Arm-based CPU that NVIDIA has built.

NVIDIA has been the fuel of the AI race since its inception, garnering an estimated 90% of the AI GPU market, and now Goldman Sachs has increased its share price target by a whopping 28%.

NVIDIA will publish its earnings report later this month, and thanks to its absolute dominance in the AI business, the company could post a whopping report, says Goldman Sachs. NVIDIA has pushed out countless H100 Hopper AI GPUs, with beefed-up H200 AI GPUs coming soon, and next-gen B100 Blackwell AI GPUs are due in 2024, too.

Goldman Sachs says that not only is demand for AI GPUs and AI servers healthy, but the supply constraints that were hurting NVIDIA AI GPUs are now easing. The firm increased its share target for NVIDIA from $625 to a huge $800, representing a 28% increase, while also increasing NVIDIA's earnings per share estimates for the next few years.