On Thu, Apr 23, 2015 at 4:52 AM, Dennis Ochei <[email protected]> wrote:
> In the thread discussing comp the topic of whether uploading is possible > came up. While tangentially related to comp, objections on the grounds of > practical impossibility miss the point. But! The topic is still very > interesting. > > Is uploading possible? Yes according to computationalism, the currently-dominant theory of mind. > If so, when will we have it? > I estimate by 2045. It is ultimately a matter of computational power and storage capacity. Imagine several high resolution microscopes/video cameras taking high resolution images as a frozen/plastinated brain is ablated (by laser, thin slicing, etc.) layer by layer. If the video resolution is great enough then the video recordings of watching the entire brain be destroyed in this way would provide sufficient information to reconstruct every neuronal connection. Harvard has a project ATLUM which automates this kind of brain scanning. As a process subject to information processing, it will follow the same trajectory of our increasing computing power. We currently have uploaded worms (nematodes) with 302 neurons: openwork.org Fruit flies have about 100x as many neurons (around 100,000), we need roughly 7 more doublings in computing power until we have openfruitfly.org Mice have 100 times as many neurons as fruit flies, 10,000,000, we need 7 more doublings in computing power to get to openmouse.org Cats have 100 times as many neurons as mice, 1 billion, we need 7 more doublings in computing power to get to opencat.org Humans have 100 times as many neurons as cats 100 billion, we need 7 more doublings in computing power to get to openhuman.org So this is a total of 28 doublings away. If computers double in power every year, we'll have openhuman.org by 2043. > What fidelity is necessary? > > That's a good question. I would hope knowing every type of neuron and all the connections between them would be sufficient. If 100 bytes is sufficient to describe every neuronal connection in enough detail, then it would take 77 PB to store the roughly 770 trillion connections. > Will the upload still be you? > > As much as the man who wakes up in your bed next morning is you (assuming computationalism) > Would you sign up for a destructive upload? Conservative? > > Yes, if others had gone before me and demonstrated satisfaction with the results. > Feel free to toss any other questions into the mix. > > For the record, I think uploading is possible, that destructive uploading > will come way sooner, Way sooner is relative, when technological growth accelerates exponentially. If we're talking about post-singularity technological progress, it might come only a few minutes later. > I'm uncertain about fidelity, but I do think there could be a functional > isomorphism that doesn't depend on on a structural one, i.e. 100 simplified > neurons might be required to capture the behavior of one physical one. The > substitution level i think is subcellular. I think uploading perserves > identity and I might actually prefer a destructive upload, when I consider > the disappointment of the me that wakes up still flesh and bone after a > conservative upload. > > Nice. See you on the other side. (hopefully). Jason -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/d/optout.
