AI is 'not smart' so what's next in artificial intelligence? (www.bbc.com)

Yann LeCun, a prominent figure in AI and former chief AI scientist at Meta, has launched Advanced Machine Intelligence Labs (AMI Labs) to develop a new type of artificial intelligence that aims to surpass the limitations of existing systems like ChatGPT. LeCun argues that current AI models, particularly Large Language Models (LLMs), lack a true understanding of the physical world and struggle with complex, real-world tasks. His new approach, known as Joint Embedding Predictive Architecture (JEPA), seeks to create abstractions of reality that enable AI to make informed predictions and decisions about outcomes—instead of merely generating statistics based on past data. AMI Labs has successfully raised over $1 billion in funding, making it one of the largest seed funding rounds in Europe, with support from major investors like Nvidia and Jeff Bezos's fund. The significance of LeCun's work lies in its potential to enhance AI's applicability in robotics and other critical domains, where conventional LLMs are deemed inadequate. By developing systems that can organize and learn from complex, dynamic environments, AMI Labs hopes to pave the way for more capable AI that ultimately complements human creativity and problem-solving rather than simply mimicking knowledge.