LeCun also argued that AI models are far from reaching human-level intelligence, comparing their computing power to that of a common housecat's brain. He believes that there is a crucial conceptual element missing in current AI technology that prevents it from reaching the intelligence level of animals and humans. Meanwhile, Musk has been vocal about his belief in the potential dangers of AI, stating that it could develop super-intelligence that could overpower human systems.
Key takeaways:
- Elon Musk has predicted that AI will be smarter than all humans combined by 2029, a claim that was disputed by Meta's AI chief scientist, Yann LeCun.
- LeCun argued that if AI were that advanced, it would be able to teach itself to drive a car in 20 hours of practice, something that is currently not possible despite the availability of millions of hours of labeled training data.
- LeCun compared the computing power of AI models to that of a common housecat's brain, stating that AI is not nearly as clever and that there is a crucial gap in our understanding of the technology.
- Musk has been vocal about his belief that AI is so powerful it could be dangerous, stating that it could develop super-intelligence that could overpower human systems. He has also banked on putting AI at the center of the development of his self-driving Tesla cars.