On Sat, Jan 2, 2021 at 3:16 AM <[email protected]> wrote:
> Or maybe Colin means (i) is the human brain naturally, (ii) is us creating > a real artificial humans brain where the actual physical implementation may > have effects that a computer sim wouldn't have (unless worked hard to code > in the physics/ rules), and (iii) is a computer simulated human > brain........Well, although certainly we could simulate a real brain with > synergy, it may be less feasible whereas a real implementation may allow > for things we would not have realized would interact together > synergetically, and of course a native hardware accelerator for AI speeds > up the algorithm over using a general purpose computer as hardware........ > Now the question is is it harder to implement a real brain in a computer > than to in real life? I.e synergy/ the sauce, not efficiency. A 2nd > remaining question is of course (ii) may also shed light on synergy of AGI. > I mean a real brain hardware may do the things we never knew to code, and > maybe we can figure out what they are too. That's the 2 questions here then. > > Hmm. Is it harder then? If so, we could observe what a real implementation > is doing to see what we missed. I DID find in my work there's lots of > energy leaks needed, hello activates node hello, and hi, bonjour, etc by > surrounding context nodes, and also the cluster of similars since all > welcome greetings are similar more than, any other, they are together as > one sorta, and energy even stays in nodes to predict them while fades away > to forget short term. Of course I reallllly hate implementing chips, I > don't! Unless get huge money. I still don't want to either. I want to > simulate the playground physics. Of course now that sounds like your way > makes sense, in some regard. I could implement rules, or my own brain > physics that allows those rules and possibly unknown ones. Well, it's a new > thought for me now then. Colin may have a good idea. The algorithm for AI > we make may lack natural synergy. Yes in a computer we can allow natural > playout, like make a tunnel axon system and just let the energy leak, but > that doesn't tell it to make energy stick in cells. Though neither does our > hardware, no? Maybe a real brain chip needs to be made to do those rules > too. Hmm..... Maybe colin doesn't have a point then? How can we make > something that has rules that our sim hardware wouldn't? Perhaps as signals > go through axons, they slow in speed maybe, and we don't implement that in > sim, but in real life brain if we make a chip it would have this result? > I addressed all of these points at length in the paper. I can't convey the central messages any more clearly. This article is not about anyone's 'way' being better or worse. It is, from paragraph 1 onwards entirely and only about repairing a broken science. There is a missing part to the science that needs to be restored. (ii) Replication. It is AWOL. And apparently what has gone on in its absence for 65 years is so completely defined by the lack that its reintroduction can be, to some, literally impossible to comprehend. There doesn't appear to be anything further I can say. I appreciate your attempts to grapple with the ideas. :-) ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T2f2a092379e757d2-M5b7aa70c5c6a318da9500e00 Delivery options: https://agi.topicbox.com/groups/agi/subscription
