zlacker

The free version of cyc has about a quarter-million terms alone, so they are likely wrong.

[edit]

According to this, it has "about seven million assertions" and notes that cyc can infer many more assertions from those.

http://www.cyc.com/kb/

>>aidenn+(OP)
I wonder how well Cyc could build assertions from fuzzy or only semi-trustworthy data from sources like Wikipedia.

>>cpeter+Tb
My guess is quite poorly. Remember the assertion that all humans have two arms and two legs isn't even 100% true, which is one reason of many why the majority of the AI field abandoned the formal logic approach for statistical methods.

The other side of the story would be that the majority of the AI field didn't want to spend 30 years formalizing the large body of general-purpose knowledge.