If it’s not in your hands in an open format it’s not yours.

Indeed.

It’s standard distributed systems design to have a hierarchy of storage with different speeds whose contents is allocated based on the frequency with which certain data is accessed, and HDDs are really only good for bulk data which is seldom accessed (basically the speed category for long term storage with low wait times when it does get needed but not really meant to be constantly accessed, which is just above things like tapes and other backup storage methods).

So for example for a dynamic website with thousands of users most current data should be in SSDs and HDDs would maybe contain low access info such as historical data from the last couple of years and in front of those SSDs there would be a ton of memory to serve as a cache for the most accessed of all data (say, the CSS, JS and images of the home page) as in-memory data is even faster to access than data in an SSD.

The idea that SSDs aren’t useful for servers is hilarious ignorant.

There are a number of simultaneous bubbles at the moment, the AI one being a lot like the Internet bubble of the late 90s but possibly worse (bigger share of GDP and it seems there is actually less value in most of the tech invested in as “AI” than on the Internet-related tech) and at the same time there is a financial debt bubble like in 2007 (in the US mainly around loans for car purchase, but more in general overall consumer indebtness has reached the 2007 levels), a worldwide realestate bubble (measured in terms of house-price to income ratios) and a stockmarket bubble measured in terms of P/E ratios, just to mention the biggest ones.

The risk is that when one blows the rest blow by contagium: something the 2008 Crash showed us is that in modern markets when there are sudden large losses on a asset class it pulls money over to cover them from all other asset classes, in turn creating downwards price pressure in those other asset classes, which in turn might cause price collapses there with large losses and that will pull even more money out from other asset classes. IMHO assets classes with historically high valuation not backed by fundamentals (for example stocks with P/E which are 10+ times the historical average) are likely to be far more likely to collapse when money gets pulled away from them to cover losses elsewhere. Also there is the panic factor: fearing exactly what I describe, many investors will preemptivelly sell their assets in those assets classes they feel as more speculative - i.e. less supported by fundamentals - possibly creating the very problem they fear in those markets by starting a stampede to the exits.

All this to say that I expect this one when it blows up will be bigger than 2008 and 2000, possibly bigger than both of those combined.