Artificial Intelligence News - Page 1

With AI technology and tools advancing at a pace that very few can comprehend, especially regarding applications and possibilities, it's an area where safety and responsibility across all levels are paramount.

This is where the U.S. Artificial Intelligence Safety Institute Consortium (AISIC) comes in, a division of the U.S. Department of Commerce and the National Institute of Standards and Technology.

AISIC will focus on the safety and 'trustworthy' development and deployment of AI, creating tools, methodologies, and industry standards. As a pioneer in AI hardware, tools, and software, NVIDIA has announced that it will work with AISIC to foster an age of AI safety.

NVIDIA will host its GPU Technology Conference (GTC) 2024 event on March 18, returning to an in-person event after the virtual world took over during the pandemic.

The company will be hosting a keynote with NVIDIA CEO Jensen Huang to talk about AI, which shouldn't be a surprise by now, with NVIDIA now promoting GTC 2024 as the "#1 AI Conference for Developers". What should we expect at GTC 2024? NVIDIA will most likely unveil its next-gen Blackwell GPU architecture, and its new B100 AI GPU that will absolutely demolish AI benchmark charts.

NVIDIA's exciting GTC events -- which I used to attend year after year and truly, dearly miss, were one of my favorite events of the entire year in tech -- normally focusing on data centers, workstations, and high-end GPUs. Now, the focus is shifting to the world of AI, which NVIDIA absolutely dominates. It has the most AI GPUs in the market, has 90%+ of the AI GPU market share, and it's about to unleash a next-gen B100 AI GPU.

The delivery speed of NVIDIA AI GPU shipment has been "greatly accelerated," according to USB analysts; previously, it was taking 8-11 months for deliveries, but now that's been reduced to just 3-4 months.

NVIDIA and TSMC (Taiwan Semiconductor Manufacturing Company) have greatly increased production capacities, solving the bottleneck issue of advanced CoWoS packaging technology. NVIDIA can now not only clear out its H100 AI GPU inventories, but it can begin pushing more of its beefed-up H200 AI GPU shipments, and upcoming next-gen B100 "Blackwell" AI GPUs that are coming this year.

Another factor in AI GPU shipments increasing from NVIDIA is that recent US export restrictions into China have freed up production capacity for other orders. USB analysts also believe that the decreased delivery cycle implies that short-term growth of NVIDIA AI GPUs is "close to its peak," so continued growth may be weak... but NVIDIA is expecting its AI GPU shipments to increase 150% year-over-year. CEO Jensen Huang also said that the AI market could be worth $2 trillion over the next 5 years recently at the World Government Summit in Dubai.

NVIDIA has officially taken the wraps off of its new Eos supercomputer, which is an incredibly powerful new system that's aimed at AI applications. Check it out:

The new NVIDIA Eos supercomputer features 576 x NVIDIA DGX H100 systems; each of the individual DGX H100 systems has 8 x NVIDIA H100 AI GPUs, delivering a grand total of 4608 x NVIDIA H100 AI GPUs inside powering them. NVIDIA is using its in-house Quantum-2 InfiniBand architecture to connect them all, delivering some truly mind-blowing performance.

NVIDIA's powerful new H100-powered Eos AI supercomputer features 18.4 exaflops of FP8 AI performance, providing NVIDIA with a continued edge against the competition regarding AI performance. NVIDIA first revealed its new Eos supercomputer at the Supercomputing 2023 trade show in November 2023, but now it's here and ready to take on AI workloads.

OpenAI has just announced its latest tool -- Sora -- which will generate videos from text prompts, and it has truly incredible videos out before the new tool has been officially released. Check it out:

OpenAI posted on X about Sora, teasing: "Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions". The new model is named "Sora" after the Japanese word for "sky" and can produce impressive video results up to 60 seconds long just from a text prompt.

OpenAI explained on its website: "We're teaching AI to understand and simulate the physical world in motion, with the goal of training models that help people solve problems that require real-world interaction".

The United States Patent and Trademark Office (USPTO) has pushed out new guidance on inventorship for AI-powered inventions, stating that only humans can hold patents.

The USPTO said that AI systems can help out during the creative process, but only natural persons (human beings) who make significant contributions to the conception of an invention can be named as inventors. This means that AI models can't just pump out patent ideas without considerable human input, stopping patent hoarding by AIs.

An AI model can't be named as an inventor, or even a joint inventor, on a patent. But, an AI helping out in the creation of that invention doesn't mean that the human being working on it can't hold the patent. The USPTO explains: "While AI systems and other non-natural persons cannot be listed as inventors on patent applications or patents, the use of an AI system by a natural person(s) does not preclude a natural person(s) from qualifying as an inventor (or joint inventors) if the natural person(s) significantly contributed to the claimed invention".

OpenAI has just announced it's rolling out a new memory feature into ChatGPT so that the AI chatbot will remember your past conversations, commands, and requests.

The new memory feature being baked into ChatGPT is coming soon and has been a long time coming. One of my major faults with ChatGPT is that it doesn't remember your conversation, forcing you to re-prompt things. It would be fantastic to remember everything, like an artificial intelligence would... not a limited AI model.

OpenAI explained on its website: "We're testing memory with ChatGPT. Remembering things you discuss across all chats saves you from having to repeat information and makes future conversations more helpful. You're in control of ChatGPT's memory. You can explicitly tell it to remember something, ask it what it remembers, and tell it to forget conversationally or through settings. You can also turn it off entirely. We are rolling out to a small portion of ChatGPT free and Plus users this week to learn how useful it is. We will share plans for broader roll out soon".

NVIDIA CEO Jensen Huang made some big statements at the recent World Government Summit recently held in Dubai, where he said AI is the future of economics and that there'll be $2 trillion worth of data centers powering the future of AI.

Huang added: "We are in the beginning of this new era and what's gonna happen is there's a trillion dollars worth of installed base in data centers around the world and over the course of next 4-5 years, we'll have 2 trillion dollars worth of data centers that will be powering software around the world and all of that is going to be accelerated and this architecture for accelerating computing is ideal for this next generation of software called generative AI".

It's not surprising that Jensen believes the data center market will explode into a gigantic $2 trillion market in the next 4-5 years, which will be the driving force behind the future of everything AI. NVIDIA is leading the pack here, where they're not only commanding an estimated 90% of the AI GPU market, but the company has seen its market capitalization explode over the last few months... adding $650 billion to see NVIDIA with a commanding $1.78 trillion market cap at the time of writing.

NVIDIA's going big on AI - in case you hadn't noticed - and the latest brainwave from Team Green is a new 'Chat with RTX' demo app.

This software allows you to leverage the power of your RTX graphics card - only RTX 3000 and 4000 GPUs are supported, mind - to run a local chatbot tapping into a large language model (the choices are Mistral and Llama, open source efforts of course).

Then you can use your preferred option, hook it up to some files on your PC, and query this local generative AI, letting it comb through those files for answers to queries.

NVIDIA CEO Jensen Huang was recently at the World Government Summit in Dubai, where he called for all countries to build their own "sovereign AI."

Huang said: "This is the beginning of a new industrial revolution, is about the production -- not of energy, not of food -- but the production of intelligence, and every country needs to own the production of their own intelligence, which is the reason why there's this idea called"sovereign AI".

He continued: "You own your own data. nobody owns it. Your country owns the data; it codifies your culture, your society's intelligence, your common sense, your history. You own your own data. You, therefore, must take that data, refine that data, and own your own national intelligence. You cannot allow that to be done by other people".