David Shapiro’s Substack
David Shapiro
"Moore's Law Squared" - Jensen Huang of NVIDIA
6
0:00
-21:26

"Moore's Law Squared" - Jensen Huang of NVIDIA

What did Jensen mean? Also, let's unpack Jim Fan's comments about fully functional humanoid robots within 2 to 3 years.
6

Sources: https://www.perplexity.ai/search/jensen-huang-recently-said-ai-GQs128vTTGS5zuxjgpBPpg

🚀 The Acceleration of Technology

Jensen Huang, NVIDIA's CEO, recently stated that technology is moving faster than Moore's Law squared. While traditional Moore's Law focused on computational density doubling every 18-24 months, the total amount of compute has been increasing at a much faster rate. From 1959 to 2010, total computational power doubled every 17-29 months, but from 2010 to 2022, this accelerated to doubling every 4-9 months.

💻 Advancements in AI and Machine Learning

Significant improvements have been made in model performance through optimizations and efficiencies. For example, ImageNet classification accuracy improved dramatically in just a few years, with training time dropping from 13 days to just 88 seconds. The cost of compute has also decreased by many orders of magnitude, making powerful computational resources more accessible.

🤖 The Future of Robotics

NVIDIA researchers predict that robotic hardware matching human dexterity and flexibility will be available in two to three years. Additionally, a general-purpose foundation model for robots is expected within the same timeframe. This could lead to the development of highly capable humanoid robots that can perform a wide range of tasks.

🧠 Specialized vs. General-Purpose AI

The future may see a combination of specialized AI models for specific tasks and more general-purpose AI systems. Application-specific integrated circuits (ASICs) and neuromorphic chips could be integrated into robotic chassis for improved efficiency and security.

💼 Economic Implications

The emergence of highly capable humanoid robots could have significant economic impacts. A robot with human-equivalent or superior capabilities might cost around $50,000, potentially making it an attractive option for businesses and some households. This could lead to widespread adoption in various industries and potentially disrupt the job market.

📚 Additional Information

Discussion about this podcast