Engineers who work with very high scale will try to limit the size of operations so that they’re not as resource-intensive as they would otherwise be. For instance, instead of running a data query on the entire database of the service, the data might be sharded (segmented) into several physical databases, so that only the relevant databases get queried, limiting what resources get used.

So, this is what Elon is doing here. Except nobody told him that engineers who are experimented with scale will implement resource limitations that the user doesn’t notice, so it always feels to them that they’re seamlessly interacting with a massive service even when they’re not.

But that only works if you still have engineers experimented with scale on staff…