Robotics is gearing up for a giant leap in 2024, similar to what we saw with large language models (LLMs) in 2023, researchers in the field believe. While we’re still catching up with the latest progress, labs worldwide are crafting robot bodies and running them through their paces in virtual setups to soon give them a physical presence.
Jim Fan, who is a Senior Research Scientist and lead of AI Agents at Nvidia, sees robotics as the biggest thing in 2024.
“I've been asked what's the biggest thing in 2024 other than LLMs. It's Robotics. Period. We are ~3 years away from the ChatGPT moment for physical AI agents,” Fan posted on X.
He believes that humanity has long been plagued by Moravec's paradox, a counter-intuitive phenomenon where tasks that humans perceive to be very easy are extremely hard for AI, and vice versa.
Robotics researcher Hans Moravec, back in the 80s, noticed that sensorimotor and perception skills require enormous computational resources, while reasoning requires very little computation.
Now, Jim Fan expects 2024 to be remembered as the year that the AI community fought back big time against this curse.
“We will not win immediately, but we will be on the path of winning,” the post reads.
Fan listed a few of the future robot foundation models and platforms that were demonstrated in 2023. Those included multimodal large language models (LLMs) with robot arms as physical input/output devices, algorithms that bridge the gap between systems of high-level reasoning (LLMs) and low-level control, and “insane amounts of progress on robust hardware.”
“Data has always been the Achilles' heel of robotics. The research community is coming together to curate the next ImageNet, such as the Open X-Embodiment (RT-X) dataset. It's still not diverse enough, but a baby step is a major step,” the list continues.
Fan noted that simulation and synthetic data will play a critical role in solving robot dexterity and computer vision in general.
The researcher shared that NVIDIA Isaac, the accelerated platform for robotics and AI, which includes a virtual environment for simulating robotic scenarios, can already simulate reality 1000 times faster than in real-time. Photorealism can be enabled by hardware-accelerated raytracing technology. Simulators already multiply real-world data to create much larger datasets, reducing expensive human demonstration efforts.
“I'm all in, personally. The best is yet to come,” Fan concluded.
Other researchers share similar views
Bernt Bornich, CEO and founder of 1X, a company designing androids, expects 2024 to be the year when most of the tech world realizes how humanoids will completely change the economy.
“We will all disagree on timelines but agree it will happen. This will result in a similar craze as LLMs in 2023,” Bornich posted in his forecast.
Brett Adcock, technology entrepreneur and founder of AI robotics company Figure, posted a similar view.
“In 2022, I was hoping AI-powered humanoid robots were possible. In 2024, I know it’s possible,” Adcock said in his post on X.
He expects these robots to give rise to the greatest economic growth in human history, and 2024 “is the year everything changes.”
Adcock also says that for the last three months, he has witnessed industry-leading AI running in their lab, which has blown him away.
“I’m watching robots performing complex tasks entirely with neural nets. AI-trained tasks that I didn’t think were feasible until the end of the decade,” he said. “2024 will be the year of Embodied AI. We’ll be racing to get reliable hardware, vast training sets, and designing manufacturing processes for billions of units. I can’t think of a more exciting future.”
More from Cybernews:
Subscribe to our newsletter