I know it's fun to snipe at #Crowdstrike and #Microsoft right now - and we should continue doing that. They suck.
But spare a moment to think that a system that collapses because a vendor messed up is a badly designed system.
If a public utility stops providing its services because a software product went offline, then that utility is badly designed.
No system that people depend on should be remotely possible to take out this easily and at this scale.
@chrisg So the halting problem was solved then?
@chrisg Dylan Beattie's talk, "Failure is Always an Option", is pretty relevant here
@chrisg Even as simple as "Why is this all in the Kernel, including all the loading and validation".
Linux isn't magically better but this kind of thing isn't encouraged to be loaded directly into the Linux Kernel.
That design we can safely critique regardless.
@chrisg
You know who noticed? All the US enemies out there. This definitely signals a weak spot and a great opportunity for them. It is like a trial scenario of how disruptive a cyber attack will be without even lifting a finger. How we react and solve this is really important.
@chrisg
It is a house of card where every piece of software stops development and ships as soon as it passes enough tests even when those tests are not relevant to the actual function of said software.
@chrisg
In a 'real' capitalist system, as envisioned in a platonic ideal state, this couldn't happen, because there would be dozens of vendors filling this same role.
Resiliency through chaos, and better designed competitive spaces.
The thing is, these aren't competing in an open market (and it's questionable if anything could).
And before this comes off as a 'do capitalism right' please understand I'm actually saying 'fuck all capitalism, but especially this faux awful one we have now'.