Alan turing notebook”
– Rereading the RAT for the first time… “We all remember”
– On the topic of the RAT, “There must be a simpler explanation”
– “The only real answer is to go out and se전립선 마사지e if you can find one, right?”
– The RAT is a super-lame thing that can’t really be described
– RATs are designed for big data
– There are two theories of this: “There must be a bigger one to explain everything” (no evidence yet) and “They can only explain 1/10th as much as we did, but the big one is better.” (there is some real potential for “the whole thing” to change, but no evidence yet)
– When it comes to machine learning, “A good theory is usually the one that works best in practice”
– A paper about the origins of the RAT has 안마been published as an issue of arXiv:0608.3676 (references are in the abstract)
– RATs are super-expensive to build – “there are very few countries in the world that do a single RAT”
– Computers can only run in a specific way. There is absolutely no guarantee that the RAT is going to work according to the user’s expectations.
– Only people that know the code can predict what will happen
– Every RAT uses the same logic in order to predict what the user expects
– RATs are extremely difficult to predict accurately. This is the only problem they could solve
– RATs can predict 100+% accuracy (although the RAT that makes most predictions is not a RAT in any realistic sense)
– A single RAT is only used to give you a 1:1 prediction – it has no ability to predict anything else.
– The R영천출장마사지AT does not work well on its own. Its purpose is to support lots of layers
– If the user has multiple machines, it will attempt to predict all those machines – but this is usually impossible.
– No one uses more than a few RATs at a time, no-one builds the RAT for them
– The only way to predict what something looks like on a graph, is to use the same graph from all the machines in the graph – if this isn’t possible, it’s not possible