4 ms·
Eh, it's worth understanding the scale and scope of the problem. I've been thinking about this a lot, recently. The human brain has on the order of 1e11 neuron
by JabavuAdams 10y ago
Eh, it's worth understanding the scale and scope of the problem. I've been thinking about this a lot, recently.
The human brain has on the order of 1e11 neurons. The galaxy has on the order of 1e11 stars.
Now, would we confidently predict that we could "solve" the galaxy in less than 100 to 500 years, say? And yet we make very, very aggressive estimates of how quickly we might unravel the secrets of the brain.
Biology is just monstrously, monstrously complex!
The hope is that it may actually be easier to build a useful thinking machine than to fully reverse-engineer the brain.
If we can really build a neural lace in the sense that Iain Banks originated the term, then we've probably also solved cancer and the rest of biology. Imagine the fine-grained control you'd need to manufacture a molecular-scale structure in-situ.
- throwaway91111 10y agoWell, the issue with the galaxy has more to do with general relativity than the number of stars. The issue right now isn't bottlenecked in any single place like that.
- JabavuAdams 10y agoCan you elaborate? I don't think I got your point. I'm intrigued by the difference between "understanding" a population, in aggregate, and having a fine-grained view of each individual while also knowing how each individual relates to the whole. So, what might we learn about the galaxy if we knew what / where / when every single bright spot of star-ish size is, that we would not learn from an aggregated simulation? Rhetorical question, obviously. I don't even know how many people live within 5km of me. How many are human-trafficked, held against their will? etc. We can make approximations, but there's a lot of interesting information at the margins.
- throwaway91111 10y agoWell, most of what you get is verification. When the problem is entirely local, you can verify separately calculated measurements with other measurements. This is why I think we'll be able to move much faster to work on problems of similar magnitudes locally than on relativistic scales. While you don't know how many humans are within 5km of you, that's certainly easier to verify with confidence N than it is to verify how many humans within 5 light years.
- ceejayoz 10y agoWhile I'm a bit more optimistic than you, I think you're raising a very important point. We just found out the lungs make half the platelets in the body (http://www.nature.com/nature/journal/vaop/ncurrent/full/nature21706.html http://www.nature.com/nature/journal/vaop/ncurrent/full/natu...) - the idea that the brain will be easy to sort out is definitely an overconfident one.
- aerovistae 10y agoI don't really understand this perspective. Are you arguing that we shouldn't try? Of course it'll be extremely difficult and the work will take place over decades, not months or years, just like Spacex and Tesla. Nobody's expecting Iain Banks's neural lace tomorrow. But if someone doesn't start working on it tomorrow, we won't have it in a hundred years, either.
- JabavuAdams 10y agoNo, I agree we should try. It was just an "aha" moment for me. It highlighted to me the inconsistency in answer these two questions: 1) How soon do you think we'll be able to tap every neuron? 2) How soon do you think we'll understand the galaxy in terms of each and every individual constituent star (not aggregates).
- JabavuAdams 10y agoEDIT> We can hand-wave estimates for the number of constituents in each of these systems. But it's interesting that we don't really know how many neurons there are in your individual brain, or in mine, etc. We can't yet look across the scales required to e.g. count and localize every single cell in your brain. Also, at what time? It changes. Just interesting questions when you get to a fine scale...
- hprotagonist 10y agoabout 80 billion. Fortunately, glial cells matter, too! (don't tell anyone)