assuming that our compression methods are the same as the ones developed accidently through evolution (highly unlikely as most of these are open source and haven't really been worked on much in comparison). that means we need at least 10k*8b= 80 trillion terabytes of data.
we don't have enough data center to house them even if we had the data available to copy right now.
of course, much of this data doesn't look the same as what you we are used to in computers. I had this issue again and again while trying to automate accounting with Ai, much of the insights i get from accountant i don't know what to do with or how to feed it into an LLM. Hell, i can't really put it into words myself of intellectually put a framework on it (and i studied accounting long time ago).
I am not saying Ai can't actually get there, it already is automating accounting. accountants used to laugh about the mere mention of it in 2018, now campfire is eating SAP. What i am saying is that we have really really underestimated the storage capacity needed to replace entire industries.