Abstract: This talk will be centered around the concept of transfer entropy (TE), as introduced some 25 years ago. I will present an insider's recollection of its beginnings starting with the irritation that led to its definition in 2000.

When I left for an industry career, TE developed a life of its own. I will try to give my outsider's view of what happened in those years. There have been considerable theoretical advances and there is a broad range of data applications from fields as diverse as brain science, finance, social sciences and music. TE has also proven useful in AI model design and, surprisingly for me, for the understanding of emergence.

I will conclude this rather personal account by illustrating some remaining issues and limitations. I will also sketch some ideas for further research.

Passcode 164098