Oddbean new post about | logout
 nostr:npub1vzq495gux2rtp8dy34hge37f348009a9zgaam7swszycra2cf02q3agn6z nostr:npub1xxemmfvctc02ug4ue2u8sr2z683wpprh3ntl5r5gc3cd7l3f02ass4zxth  There is a reasonable assumption that the data is too massive (100 trillion gigabytes of yearly data volume), so  we have moved to streams analysis/reporting/AI rather than starting at basic human analysis.  The catch is, that in most use, data is a form of waste stream.  It is the as-is byproduct of a system that doesn't work. We need to break down requirements first, determine constraints, then move on to a new system that will have an entirely different stream of data.