“Curre AI architectures, such as large language models (LLMs), will give way to new paradigms in the next few years,” Jan Lacan, Meta’s chief artificial ielligence scieist, announced at the Debating Technology conference in Davos. He emphasized that these changes will improve the ability of artificial ielligence systems far beyond what is possible today.
According to Techcrunch, Lacan believes that the curre limitations of artificial ielligence are the inability to deeply understand the physical world, lack of persiste memory, lack of reasoning, and complex programming. He stated that the curre artificial ielligence systems are more skilled in finding patterns and are not capable of real thinking.
Lacan meioned the idea of ”universal models”; A concept that aims to help artificial ielligence systems better understand real-world dynamics. This concept includes capabilities such as memory, logic, common sense and iuition. According to Lacan, these developmes can be realized within the next 3 to 5 years and iroduce a new generation of artificial ielligence systems.
A decade for robotic systems and artificial ielligence
In addition to the advances in AI architecture, Lacan predicted that the next decade could become the “decade of robotics.” He emphasized the importance of combining artificial ielligence and robotics to create systems that have more complex real-world capabilities.
He also poied out that currely robots cannot even reach the level of understanding of the physical world by cats, but the combination of advanced artificial ielligence and robotics can overcome these limitations in the next decade.
Finally, by emphasizing the importance of combining artificial ielligence and robotics, Yann Lacan poied to a future in which iellige systems with a deeper understanding of the real world and the ability to plan and reason will transform our world. These developmes can not only overcome the curre limitations, but also create new horizons in technology and everyday applications.
