AGI will not happen in your lifetime: Grady Booch and Gary Marcus discuss

The article by Gary Marcus discusses how the development of artificial general intelligence (AGI) is not likely to happen in our lifetime. Marcus explains that while significant progress has been made in AI research, specifically in the areas of machine learning and deep learning, these technologies are still far from achieving AGI. He argues that there are still major hurdles to overcome before making a functioning AGI, such as teaching machines common sense, getting them to understand the full context of a situation, and giving them an ethical system for decision-making. Furthermore, AI researchers have yet to come up with a unified theory of intelligence, which would be necessary for AGI to fully function.

Marcus goes on to point out that AI development has been slowed by what he calls the "algorithm gap". This term refers to the fact that current algorithms are good at handling specific tasks, but they lack the ability to learn and transfer knowledge between tasks. He argues that this prevents AGI from becoming a reality, since it requires the ability to learn a variety of tasks and generalize its knowledge. To bridge this gap, researchers would need to develop more complex algorithms and approaches.

In conclusion, Marcus believes that AGI is still many years away, and is unlikely to emerge in our lifetime. While advancements in AI research have been impressive, there are still numerous challenges to overcome before AGI could become a reality. Until then, AI will continue to remain limited to performing specific tasks, with no ability to transfer knowledge across tasks or understand the broader context of a situation.

Read more here: External Link