4 ms·

Strongly believe that if anyone is going to solve human level intelligence in our lifetime, it's going to be Numenta. 'But we can't even figure out worms!' - W

Strongly believe that if anyone is going to solve human level intelligence in our lifetime, it's going to be Numenta. 'But we can't even figure out worms!' - Worms are made up of neurons but they perform a different function from what the neocortex does so studying them is not like studying a simpler problem, it's studying an entirely different one. 'But ML can do X better!' - Unlike industry or academia, Numenta's primary goal is to figure out how the neocortex works, it's not about profit or publications. 'Biology contains details we don't need!' - Numenta's approach is not biologically inspired, it's more like biologically constrained. They avoid implementations that are functionally different from how the neocortex works. I would highly suggest reading On Intelligence to learn more.

deepnotderp 9y ago
"Unlike industry or academia, Numenta's primary goal is to figure out how the neocortex works, it's not about profit or publications." Numenta is literally a corporation...
foldr 9y ago
>Worms are made up of neurons but they perform a different function from what the neocortex does so studying them is not like studying a simpler problem, it's studying an entirely different one. Huh? Take your pick of creatures that display highly intelligent behaviour and have less complex nervous systems than humans. Insect navigation would be a good example. If we can't figure out the biological basis of that -- when in many cases we already have a good idea of what calculations are being performed -- then our chance of "solving human level intelligence" is zero.
- shahbaby 9y ago
  The problem is in defining intelligence by behavior. What we see as intelligence in insects most likely originates from an entirely different process. Jeff talks about this exact issue in this speech he gave, 10 years go. He mentions it around the 11 min mark but the whole video is a great intro to what this is all about. https://www.youtube.com/watch?v=G6CVj5IQkzk https://www.youtube.com/watch?v=G6CVj5IQkzk
  - foldr 9y ago
    > What we see as intelligence in insects most likely originates from an entirely different process. That is pure speculation. As far as I can tell, what you're suggesting is that it might possibly be easier to figure out how human intelligence works than it would be to figure out how an insect performs a few calculations, in cases where we know almost exactly which calculations are being performed (see e.g. http://science.sciencemag.org/content/312/5782/1965 http://science.sciencemag.org/content/312/5782/1965). I don't see any reason to think that this is true other than wild optimism.

deepnotderp 9y ago

"Unlike industry or academia, Numenta's primary goal is to figure out how the neocortex works, it's not about profit or publications." Numenta is literally a corporation...

foldr 9y ago

>Worms are made up of neurons but they perform a different function from what the neocortex does so studying them is not like studying a simpler problem, it's studying an entirely different one. Huh? Take your pick of creatures that display highly intelligent behaviour and have less complex nervous systems than humans. Insect navigation would be a good example. If we can't figure out the biological basis of that -- when in many cases we already have a good idea of what calculations are being performed -- then our chance of "solving human level intelligence" is zero.

shahbaby 9y ago
The problem is in defining intelligence by behavior. What we see as intelligence in insects most likely originates from an entirely different process. Jeff talks about this exact issue in this speech he gave, 10 years go. He mentions it around the 11 min mark but the whole video is a great intro to what this is all about. https://www.youtube.com/watch?v=G6CVj5IQkzk https://www.youtube.com/watch?v=G6CVj5IQkzk
- foldr 9y ago
  > What we see as intelligence in insects most likely originates from an entirely different process. That is pure speculation. As far as I can tell, what you're suggesting is that it might possibly be easier to figure out how human intelligence works than it would be to figure out how an insect performs a few calculations, in cases where we know almost exactly which calculations are being performed (see e.g. http://science.sciencemag.org/content/312/5782/1965 http://science.sciencemag.org/content/312/5782/1965). I don't see any reason to think that this is true other than wild optimism.