I wrote a popular science article celebrating Claude Shannon's information theory for "Dimensión fractal", the joint outreach initiative of ICMAT and the Spanish online newspaper elDiario.es:
The article introduces the notion of Shannon entropy of a source as a measure of the average "surprise" that a given message will cause its receptor. It then touches on Shannon's noiseless coding theorem, his research on the entropy of language and its connection to the training process of modern large language models.