crocabit:

animatedamerican:

asocratesgonemad:

ghoulishautomaton:

asocratesgonemad:

explodingbat:

aaaaa42:

holy fucking shit people simulated the 302 neurons in a worm’s brain with software, put it in a Lego robot and it behaves like a worm

http://www.smithsonianmag.com/smart-news/weve-put-worms-mind-lego-robot-body-180953399/?no-ist

upload the nematode

upload the nematode / lock it in a box / keep it there for fifty years / then see if it talks!

but, like, it’s a worm. its behavior isn’t that complex.

if it’s not accurately behaving like a worm, how would we tell?

^ an important point. IIRC they don’t even really have a complete characterization of the worm’s nervous system, so C. legogans here is at best a preliminary test. Still pretty cool, though.

I looked at the article, and to my mind the significant thing here is that the lego robot demonstrated consistent behavior with no programming.  They didn’t program any stimulus-response behavior into it.  They just gave it their best approximation of worm neurons, and it made up the behavior.

I may be misunderstanding some aspect of this but to me that feels kinda huge?

They made an artifical reactive set of neurons with consistent observable behaviors. Yeah, this is a big deal. It’s the first step towards ‘uploading’ minds.