It is much better than this. You can _directly_ enumerate all the objects, without any probabilities involved. There's nothing about probabilities in the interface of a PRNG, it's just non-determinism!

You could _implement_ non-determinism via probabilistic sampling, but you could also implement the same interface as exhaustive search.

▲

pfdietz 3 hours ago | parent [-]

Well, yes. But the point is that random sampling lets you do it without thinking. Even better, it can sample over multiple spaces at the same time, and over spaces we haven't even yet formalized. "Civilization advances by extending the number of important operations which we can perform without thinking of them." (Whitehead)

An example is something like "pairwise testing" of arguments to a function. Just randomly generating values will hit all possible pairs of values to arguments, again with a logarithmic penalty.

▲

AlotOfReading 34 minutes ago | parent [-]

The point is that you can exhaustively explore the space without logarithmic overhead. There's no benefits to doing it with random sampling and it doesn't even save thought.

▲

maxbond a few seconds ago | parent | next [-]

In my experience it is fairly easy to randomly generate a valid object but very difficult to write an algorithm that exhaustively searches the space of all possible objects, and usually not tractable anyway. The space is usually too big to exhaustively search.

▲

pfdietz 28 minutes ago | parent | prev [-]

I already explained what the benefit is. What is it with this focus on offloading work from computers to people? Let people do things more easily without thinking, even if it burns more increasingly cheap cycles.

	▲	AlotOfReading 9 minutes ago \| parent [-]
		You haven't explained what the benefit is. There aren't "spaces we haven't formalized" because of the pigeonhole principle. There are M bits. You can generate every one of those 2^M values with any max cycle permutation. What work is being offloaded from computers to people? It's exactly the same thing with more determinism and no logarithmic overhead.