(In the context of Microsoft Research) Man, the history of how large corps fumble major wins should be studied.

I remember during my thesis how Microsoft Academic (a Google scholar/Scopus alternative) was ridiculously good at its job.

I'm talking, "input the key words and filter out the wrong fields and you've got 20 highly relevant papers that no other search term has found" level.

...Of course, they proceeded to close it down.

I can't even express in words how mad and disappointed I was and still am at Microsoft for this alone. It felt the closest to a genie or an oracle in being amazing. (The old Swype vs modern Gboard is probably a close example for those who'd used both.) It's just... crazy, frustrating, and crazy frustrating.

I used to work on the search index and query processing for Microsoft Academic. It was a great project, perhaps the most interesting thing I've ever worked on.

It's still available here (separate from the papers that Microsoft Academic collected and indexed): https://www.microsoft.com/en-us/research/project/knowledge-e...

I haven't worked on this since 2016 so I don't know if it still works.