Yeah but my point is, conceptually, if we have a general concept of compressing things in memory... I'd imagine there's enough people that could as well.

Not saying anything with regards to the odds of meta releasing it, or keeping it to themselves. I'd imagine it's a matter of business. If their goal is to get ahead in the datacenter race, that might be a key worth holding on to for a while, or maybe they think that it will render shitty phones able to keep viewing facebook longer so it's worth releasing to the public early. Point I was going at is simply the tech exists, it should be possible to replicate it by someone who isn't as profit driven.