Konwersatorium im. J.Pniewskiego i L.Infelda

Artificial intelligence systems based on neural networks have led recently to the development of numerous text-to-image applications, capable of generating hyper-realistic images, videos, and music. Pre-trained transformer algorithms, such as ChatGPT and other Large Language Models with billions of parameters, are now able to answer questions on any topic, generate new ideas and pass final university examinations with high marks. I will show examples of a few such models, talk about their capabilities and emergent properties. The ability of Large Language Models to recruit specialized software, including Wolfram language for mathematics, has greatly expanded the power of ChatGPT's, bringing us closer to the Artificial General Intelligence.How will this development impact scientific research? Will human intelligence be overtaken by artificial intellignece, and if so, how soon? Is machine consciousness and sentience possible?

Pobierz plakat / Download the poster
Pobierz slajdy / Download the slides
Obejrzyj nagranie / Watch the video