This is great fun, but did anyone else think the suggested insults for experimentalists were very weak? The Xerox Parc one is good but most experimentalists would love a question about bottlenecks! And who cares about testing on Intel's latest chip? I suggest a proper insult would suggest an easily avoided bottleneck or unnecessary point of failure.
I feel like experimentalists have changed quite a bit in 20 years as big distributed systems have become more common. A good question might be “Sure your performance scales better than other systems but how does it compare to a single core, or a single big machine?”
