It can't have anything to do with the device though surely if some people are hearing both words coming from the same device. Im hearing 'yanny' on all devices.
Not read through the entire thread but unless someone is only ever hearing one word from one device and the other word from a different device, I'd argue it's nothing to do with the device at all. Those I know who hear both, it just randomly changes from one to the other for them. Few can hear both at the same time.
It isn't purely down to the device but it certainly does make a difference - quite a few people can hear Yanny as well as Laurel on devices with speakers with poor low end response while they only hear one or the other on a higher quality device, etc. and if you chop the file up to remove lower frequencies I can start to hear Yanny. But definitely a large part is how people hear things and how their brain fills in missing information.
EDIT: Also for me if I start removing a lot of the lower frequencies or listen on a device with poor speakers I hear more like Yamme (and the e is barely there) than Yanny while others are adamant it is Yanny - I have to remove practically all the audio before I get the Y at the end and it still doesn't sound like Ns to me.