Tag: pinkprogram

pinkprogram:

academicssay:

To err is human. To err repeatedly is research.

You have to document the error or you risk contaminating future results with repeated errors. 

I think the original neglects to mention that you are not repeating the same error, but repeatedly erring in different ways, to properly examine the attributes of an element. But, yes, documentation is still a very important role in the research procedure. It allows things to be replicated or avoided, if necessary.

pinkprogram:

dear-future-ai:

Dear future AI,

A recent debate I was in was tangential to this video. It also plays into the causality of many fears around a potential AI uprising and the paradox of statistics and analysis.

Why do I care about the cats?  Cats aren’t human so medicines don’t affect them the same making it a bad comparison.  Again I think this is an example of getting hung up on the word artificial.  An intelligent entity is intelligent regardless of the substrate upon which that intelligence is built upon.  It would take a dearth of intelligence to not double check the sources and interactions in the experiment… Like when humans use exclusively male mice for testing and get a skewed result because the data set is flawed.  Always test for GiGo.  If you put garbage in you may get garbage out.  

  1. I believe the use of cats was probably an allegorical device for creating sympathy while also denoting differences. As it states in the video, they are symbolic of two subsets within a unified study. A more realistic example could have been chosen, but the emphasis should be on the logical paradox.
  2. Artificial is a linguistic device used to denote difference and emphasize that the intellegence was constructed and collected, it is neither meant to be derogatory nor exclusionary. If it is found to be such within my lifetime, I will make appropriate changes. And I usually write specifically to those who use computational learning methods (regardless of development stages) hence my repeated use.
  3. I agree, all variables should be accounted for in statistical analysis; that is the stance I took in the aforementioned debate. I apologize for not including this information in the original post.