Kerala Technology
AI all set to walk the talk

AI is ready to enter the next frontier, says Nvidia CEO Jensen Huang. Image courtesy: Nvidia

AI all set to walk the talk

Hari Kumar By Hari Kumar, on June 11, 2024
Hari Kumar By Hari Kumar, on June 11, 2024

Artificial intelligence has made remarkable strides, enabling machines to perform tasks that once seemed possible only for humans. AI can produce convincing text and images from simple text prompts.

Its capabilities now include the creation of videos based on a few lines of typed text, marking a significant leap forward in multimedia content creation.

OpenAI’s Sora, though not yet available to the public, has already showcased impressive advancements in AI-driven video production. A standout example that emerged last week is the Kling AI Model from Kuaishou, a Chinese platform similar to TikTok, which can transform text into video clips up to two minutes long in 1080p resolution.

Despite such advancements, AI’s reach has primarily been confined to generating text, images, and videos. There have been many predictions about the changes AI will bring, but all the wonders it has created so far remain on the screens of computers and mobile phones.

This is frustrating to many as a recent debate on the internet showed. It all started with Joanna Maciejewska, a Virginia-based science-fiction and fantasy author and artist, posting her take on X about AI’s current direction.

Maciejewska’s post reads: “You know what the biggest problem with pushing all-things-AI is? Wrong direction. I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”

This went viral, with her post getting almost 3 million views in less than a week and many people voicing similar concerns. Subsequently, it ignited debates in various tech media about AI performance and hype.

This could all change if you go by the predictions of Jensen Huang, CEO of Nvidia Corp. At Computex 2024 in Taiwan last week, Huang said AI is all set for the next frontier: entering the physical realm.

AI won’t be staying put in our phones and desktops. We are going to see AI that understands the laws of physics, AI that can work among us, he said during a two-hour-long speech at the National Taiwan University Sports Centre in Taipei.

He said "Physical AI" was already unfolding in Taiwan with major partners like Foxconn Technology Group and TSMC leading the charge. Citing advances made by companies like Tesla in their autonomous cars, Huang predicted that every single car will someday have autonomous capability.

Then he added: "One day, everything that moves will be autonomous."

Nvidia recently surpassed a market cap of 3 trillion US dollars, overtaking Apple to become the world’s second-most valuable public company, with only Microsoft ahead of it now.

This has placed Huang at the forefront of what the tech world is calling “Jensanity.” This term, coined by Taiwanese journalist Joyce Tseng, signals the key role that Huang holds in the tech world now.

Nvidia’s groundbreaking achievements were highlighted during Computex, as a virtual avatar and cloned voice were featured during Huang’s speech. This demonstrated the power of the company’s GPUs, which deliver Trillion Operations Per Second (TOPS) performance.

Huang said that computers are evolving from mere data processors to factories generating intelligence for every industry. Nvidia plans to upgrade its AI accelerators annually, with the next-generation Blackwell Ultra expected next year, following the Blackwell platform chips introduced in March 2024.

Nvidia’s Isaac AI robot development platform is already being used to improve efficiency in factories and warehouses, including those for BYD Electronics and Siemens.

Isaac is being used for the research, development, and production of “tens of millions” of AI-powered autonomous machines and robots, including industrial arms and humanoids, according to the company.

Nvidia recently introduced GR00T, a general-purpose foundation model for humanoid robots capable of understanding natural language and mimicking human movements. “Building foundation models for general humanoid robots is one of the most exciting problems to solve in AI today,” Huang said.

Intelligent robots have been an elusive dream for countries like Japan, which have spent millions of dollars developing such gadgets to take care of the elderly. However, the robots developed over the last two decades have failed to deliver the efficiency needed for such jobs.

But that may change quickly with the infusion of AI. Robotic Transformer 2 (RT-2), built by Google, demonstrated the ability to move a drink can to a picture of Taylor Swift, even though it was not specifically trained on the singer's image. The AI-enabled robot used its knowledge gained from the internet, reports Nature magazine.

“A lot of internet concepts just transfer,” says Keerthana Gopalakrishnan, an AI and robotics researcher at Google DeepMind in San Francisco. This radically reduces the amount of physical data that a robot needs to have absorbed to cope with different situations.

Maybe this is what underlies Huang’s prediction that the age of intelligent robots has arrived. Processing a million calculation every second can accelerate the development of projects like autonomous vehicles, at least in controlled environments.

China has been leading the development of robots, with many of its factories employing them. In the US, Amazon has deployed robots in its centers and plans to expand their use.

India has hardly made a mark in this sector, though firms like GenRobotics have gained significant attention. In Kerala, Malayala Manorama is organising a week-long event focusing on the Indian robotics sector that will start on June 12 in Cochin.

Despite advancements in AI and robotics, creating humanoids like those portrayed in movies remains a challenge for tech companies. Experience has shown that mundane human chores, such as opening a bottle, are still too complex for robots. So, author Maciejewska may have to continue doing her laundry and dishes for a while before AI comes to the rescue.

 


 

Space scientists are over the moon

It was a blockbuster week for the space scientists. First, China landed its Chang’e-6 on the dark side of the moon and collected two kilograms of rocks and dust. The samples were then placed into a metal vacuum container that will return in a re-entry capsule due to arrive on Earth around June 25. This complex operation, never before achieved by anyone else, was carried out by sending commands through a relay satellite, as that part of the moon does not face the Earth. The machinery used for collecting samples was designed at the Polytechnic University in Hong Kong.

Meanwhile, across in the US, Boeing finally managed to send its Starliner crew to the International Space Station, although technical glitches continued to concern ground controllers. Thruster malfunctions and helium leaks threatened the mission, but the craft managed to successfully dock at its destination. On another front, Boeing's rival SpaceX launched its Starship rocket for the fourth time and successfully tested its re-entry. This demonstrates the progress it has achieved since the initial launch on April 20, 2023, which had to be aborted seconds after lift-off.

 


 

Don’t believe in coincidences

Do you remember the team behind Devika AI, Mufeed H, Praveen R, and Asjith Kalam? The co-founders of Stition.ai had shared their dream of getting into Y Combinator during our interaction in April. When we learned they were recently in Chennai, we were intrigued. After all, the YC list for 2024 is about to be announced, and the US consulate is based in Chennai. The Stition trio remained tight-lipped about their trip, but we usually don’t believe in coincidences. We will have to wait until the YC 2024 list is officially announced to know what we think we know is correct.

 


 

Own a piece of your hero

The name Lucia Ciric Bagaric may not ring a bell even among hardcore tennis fans. However, the Croatian player received enthusiastic cheers from a group of investors as she advanced through two rounds at the French Open before eventually bowing out. Bagric is funded by a group called FANtium, which crowdsources financial support for tennis players in exchange for access to the players and a percentage of their future earnings. FANtium's founder, Jonathan Ludwig, describes athlete investment as a new asset class. The group intends to expand its business by adding sports clubs as another asset class.

 


 

Back to the future with AI

Almost everyone is eager to know what their future holds. Now a new AI platform is here to help you with that. A team of AI researchers in Thailand and the US has developed an AI chatbot that can provide you with a potential version of your future self. All you need to do is provide data about yourself, the people in your life, and a photo. Voila, your older version is ready. Wonder if the researchers tried feeding it data about their project to see how long it would survive.