It makes a very important reply to anyone who claims that e.g. you should stick with Occam's original Razor and not try to rephrase it in terms of Solomonoff Induction because SI is more complicated.
Humans and their silly ideas of what's complicated or not.
What I find ironic is that SI can be converted into a similarly terse commandment. "Shorter computable theories have more weight when calculating the probability of the next observation, using all computable theories which perfectly describe previous observations" -- Wikipedia.
Here's the new thread for posting quotes, with the usual rules: