Status Updates From The Singularity Is Nearer: ...

The other form of AI is neural networks, using a brute force approach with vast amounts of data to sit through and see patterns identify patterns that would’ve been in perceptible to humans in one form called unsupervised learning. Again weaknesses include air, accumulation, lack of knowledge of what the goal is or even what a correct answer is, takes vast machine, power, and vast data sets.

That’s called the complexity ceiling.

Ch 1
Rules based AI comma adds points for mistakes with each rule. Therefore, the more complex the algorithm the higher, the error rate.

Ch 1
The touring test, can be failed by the AI being too smart, and therefore recognizable as machine instead of human,? Is this wrong - the machine being too smart? The machine having achieved this goal and surpass it is probably a better explanation.

50