4 ms·

"The world is very confusing, and we end up only seeing a tiny sliver of it, but we need to make some sense of it in order to survive. Once the reduced stream o

by nojvek 10y ago

"The world is very confusing, and we end up only seeing a tiny sliver of it, but we need to make some sense of it in order to survive. Once the reduced stream of information comes in, we connect the dots, fill in the gaps with stuff we already think we know, and update our mental models of the world." So wonderfully said. I wonder what biases AI will develop in its models

acomar 10y ago
Likely a lot of the same biases... Look at the categorization, these are things any general system of intelligence is going to have to do at some level. Only a system with infinite computational power can afford to be without bias.
- g4nt1 10y ago
  I even think that this biases could be a base of a good AI. Just code this biases in and your almost good to go!
not_that_noob 10y ago
This is an artifact of evolution - mental shortcuts enabled faster decision making, and species that got to the conclusion fastest survived. Of course, there is an error rate, but as long as it was small, the faster decision makers won. Logic is simply a way to systematically eliminate these errors in wetware (and ultimately hardware).
davemel37 10y ago
I once heard that our senses primarily function as filters. I.e. our eyes only see by filtering out all sorts of other visual stimuli, our ears hear by filtering out other noises, etc... That means that we are only capturing a "tiny sliver" of whats actually happening around us. If we only keep this small detail in mind we can be sure that lack of evidence is never proof of anything and that any evidence we do see can only be used to find the most probable of competing hypotheses and its diagnosticity should always be considered. (i.e. fever might mean you are sick but has no diagnostic value to what particular ailment you have since its consistent with other possibilities.) Same goes for using bounce rate as a metric for web analysis...its diagnostic value on its own doesnt tell us much, but with other context, like time on site and conversion goal of site (i.e. a phone call conversion will register as a bounce if they dont navihate elsewhere.) It can tell us a lot.

by nojvek 10y ago

acomar 10y ago

Likely a lot of the same biases... Look at the categorization, these are things any general system of intelligence is going to have to do at some level. Only a system with infinite computational power can afford to be without bias.

g4nt1 10y ago
I even think that this biases could be a base of a good AI. Just code this biases in and your almost good to go!

not_that_noob 10y ago

This is an artifact of evolution - mental shortcuts enabled faster decision making, and species that got to the conclusion fastest survived. Of course, there is an error rate, but as long as it was small, the faster decision makers won. Logic is simply a way to systematically eliminate these errors in wetware (and ultimately hardware).

davemel37 10y ago

I once heard that our senses primarily function as filters. I.e. our eyes only see by filtering out all sorts of other visual stimuli, our ears hear by filtering out other noises, etc... That means that we are only capturing a "tiny sliver" of whats actually happening around us. If we only keep this small detail in mind we can be sure that lack of evidence is never proof of anything and that any evidence we do see can only be used to find the most probable of competing hypotheses and its diagnosticity should always be considered. (i.e. fever might mean you are sick but has no diagnostic value to what particular ailment you have since its consistent with other possibilities.) Same goes for using bounce rate as a metric for web analysis...its diagnostic value on its own doesnt tell us much, but with other context, like time on site and conversion goal of site (i.e. a phone call conversion will register as a bounce if they dont navihate elsewhere.) It can tell us a lot.