Hi Logan, On Sun, May 17, 2015 at 8:50 PM, Logan Streondj <[email protected]> wrote:
> On Sun, May 17, 2015 at 09:48:03AM +1000, colin hales wrote: > > Sorry about the previous empty. Phone issue. 10 thumbs. > > > > My particular flavour of the non-computer approach is irrelevant. I am > not pushing my own at all. > > > > Robot $ and kind irrelevant. I have the math you speak of. Wrong on both > counts. > > > > I do not care what kind of NC-AGI arises. All I know is that NC-AGI > important, neglected and needs a champion. > > okay so obviously you are the champion. > Why is it important? > > Because it's never been done. > you guys have been talking about it for a week now, > and I still have no idea why you think it has value. > > like lets be honest here, anything that isn't a computer or > technology is biology. so what you are really talking about > (seems to me) is biological-AGI, or connecting a vat of > brain-cells to a computer. > this has been done, and can play simple video games. > but so can deep neuronets on computers. > No it is not necessarily biology. H-AGI can use biological material or make inorganic versions of the biological substrate. Mine is totally inorganic. And while hooking bio material to other hardware has been done, it has not been done by anyone headed in the direction of an AGI. Pure wet neuroscience? Yes. Machine learning? Yes. Robot control? Yes. This particular approach is not what I intend. Dorian may be more interested in that. I don't want to stop anyone doing any of it just because it clashes with my own vision of it. > > > > So please set anything you think you know about me or my approach aside. > You actually know almost nothing and what little that is is irrelevant to > what is happening in this thread. > > okay so do you have some kind of proprietary secret approach? > > I was thinking you can Dorian can sign an NDA and then no one > will ever know about anything you guys do. > > personally I think that there are a lot of potential ethical > issues with using biological mediums for computation, also they > aren't particularly scalable or portable. > > > I can see scalability. I can see portability. I can see generativity. It will be clunky at first like all new ideas. It has an organic and an inorganic aspect. All untried as H-AGI. Both Dorian and I have written up and published everything that is needed to get your head round the fundamentals, which I know can be hard to see for those without the biophysics. We have both argued for a long time, one way or another, that the approach is novel. The one-liner explanation: *H-AGI is where the brain physics essential to an AGI is identified and included in an AGI substrate. This is achieved by actually replicating the physics (organic/inorganic, doesn't matter) and including that physics on the substrate then and testing its performance against alternatives that lack that physics (i.e. that might ignore it or model it, replacing it with the physics of the instantiation of the model, whatever that might be).* So it's rather simple. Both Dorian and I have identified candidate 'low-hanging fruit' physics. There may be others. That physics may be the crucial missing link that has dogged AGI for decades. If so, then all activity that did not include that essential physics was actually destined to underperform in mysterious ways that it is H-AGIs job to sort out. The IGI, if it existed, would do that. Yes there are heaps of ethical issues with any AGI approach. The H-AGI inorganic version will have one ethics/risk landscape, the H-AGI organic version another. Both of these will contrast with the C-AGI risk landscape. In what ways? I dunno! Let's find out! regards Colin Hales ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
