When I was working at Intel, Boris Babayan, one of the original creators of Elbrus and an Intel fellow back then, came to visit the team I was working on. He was working very hard to sell the VLIW to us. That was post Sandy Bridge, which turned out to be a great success, and when Intel already realized what a disaster Itanium was. Yet Babayan was still strongly convinced that VLIW is superior if we could just make the compilers a little bit better. Sufficient to say that it is practically impossible to perfectly schedule VLIW at compilation time, and is a very much solved problem for OOO execution. We were all very much bewildered and just ignored him for most of the visit. I really have no clue why Intel is keeping him around.