Twitter at scale is far more complex. With Tesla's telemetry it's just writing to a datastore, with less fuss around latencies, data duplication or consistency. There's no need for millions of clients to then get updated, as there is on twitter. Musk tweets and 110M+ users (+thousands of API clients) need to receive that almost immediately. Twitter has figured out how to do this in a very scalable fashion. Once you know how it's done it's not all that impressive on its own but when you take into account the littany of other features twitter has to maintain at scale you start to realize how many unique problem domains there are within the org.