Thread by @fchollet

There's a big difference between what's important and what gets talked about

Violent individual anecdotes or the latest caprice of a powerful boor get unending front pages, but the trends that shape our history get virtually no coverage

This is true for AI research too -- the attention of the community is constantly being captured by shiny objects, but these are not the ideas and results that will be retrospectively understood as having shaped the field

The worst thing you can do as a researcher is to be constantly moving from one shiny object to the next, and never stop to consider the important questions

This is perhaps exacerbated by social media, which not only directs our attention to the wrong things, but also creates an incentive system where you get rewarded for "contributing" to the latest fad or controversy