I haven’t been following estimates lately, but how much do people think it would cost in GPUs to approximate a human brain at this point given all the GPU performance leaps lately? I note that deep learning researchers seem to be training networks with up to 10b parameters using a 4 GPU setup costing, IIRC, <$10k, and given the memory improvements NVIDIA & AMD are working on, we can expect continued hardware improvements for at least another year or two.
(Schmidhuber’s group is also now training networks with 100 layers using their new ‘highway network’ design; I have to wonder if that has anything to do with Schmidhuber’s new NNAISENSE startup, beyond just Deepmind envy… EDIT: probably not if it was founded in September 2014 and the first highway network paper was pushed to arxiv in May 2015, unless Schmidhuber et al set it up to clear the way for commercializing their next innovation and highway networks is it.)
I haven’t been following estimates lately, but how much do people think it would cost in GPUs to approximate a human brain at this point given all the GPU performance leaps lately?
I had some recent discussions with Jacob Cannell about this, where he estimated that (with the right software which we don’t yet have) you could build a human-level AGI with about 1000 modern GPUs. The amortized cost plus electricity (or if you rent from Amazon AWS) is roughly $0.1 per hour per GPU so the total would be around $100 per hour.
Schmidhuber’s new NNAISENSE startup
FLI just gave Bas Steunebrink $196,650 to work on something called “experience-based AI” or EXPAI, and Bas is one of the co-founders of NNAISENSE. This EXPAI sounds like a traditional hand-coded AI, not ANN based. Possibly they set up the startup without any specific plans, but just in case they wanted to commercialize something?
From a very uninformed perspective, this looks like an area of science where China is leading the way. Can anyone more informed comment on whether that is accurate, and whether there are other areas in which China leads?
So the head of BGI, famous for extremely ambitious & expensive genetics projects which are a Chinese national flagship, is stepping down to work on AI because genetics is just too boring these days: http://www.nature.com/news/visionary-leader-of-china-s-genomics-powerhouse-steps-down-1.18059
I haven’t been following estimates lately, but how much do people think it would cost in GPUs to approximate a human brain at this point given all the GPU performance leaps lately? I note that deep learning researchers seem to be training networks with up to 10b parameters using a 4 GPU setup costing, IIRC, <$10k, and given the memory improvements NVIDIA & AMD are working on, we can expect continued hardware improvements for at least another year or two.
(Schmidhuber’s group is also now training networks with 100 layers using their new ‘highway network’ design; I have to wonder if that has anything to do with Schmidhuber’s new NNAISENSE startup, beyond just Deepmind envy… EDIT: probably not if it was founded in September 2014 and the first highway network paper was pushed to arxiv in May 2015, unless Schmidhuber et al set it up to clear the way for commercializing their next innovation and highway networks is it.)
I had some recent discussions with Jacob Cannell about this, where he estimated that (with the right software which we don’t yet have) you could build a human-level AGI with about 1000 modern GPUs. The amortized cost plus electricity (or if you rent from Amazon AWS) is roughly $0.1 per hour per GPU so the total would be around $100 per hour.
FLI just gave Bas Steunebrink $196,650 to work on something called “experience-based AI” or EXPAI, and Bas is one of the co-founders of NNAISENSE. This EXPAI sounds like a traditional hand-coded AI, not ANN based. Possibly they set up the startup without any specific plans, but just in case they wanted to commercialize something?
From a very uninformed perspective, this looks like an area of science where China is leading the way. Can anyone more informed comment on whether that is accurate, and whether there are other areas in which China leads?