3 ms·
What are the current limitations of 'just' creating this on a much (much) larger scale? For instance, 160.000 or 1.6 million cores and would it be a significant
by tluyben2 14y ago
What are the current limitations of 'just' creating this on a much (much) larger scale? For instance, 160.000 or 1.6 million cores and would it be a significant improvement? And for which other fields would that work? I do know about neural net theory and I worked with them in college assignments quite a bit 20 years ago, but I'm wondering if anyone here knows what differences this kind of scale will make or if it's just interesting for still images (in which i'm very much not interested).
- marshallp 14y agoThe amazing thing about this algorithm is that it works for any data type (my other comment talks about speech and nlp). With more computers and more data it will keep getting better - i've thought a lot about this and following this stuff for a couple of years - so it's not just some idle speculation. (I broke this recent story on hacker news a few days ago, and from here it got picked up by the new york times and spread everywhere else).