Hi Berick, Sorry about the previous glitch in your name. I'm responding in-line, if that's ok.
On Wed, Jul 3, 2019 at 10:56 AM Berick Cook <[email protected]> wrote: > If RIGHT produces something that is functionally equivalent (demonstrated > via empirical evidence) to MIDDLE, then is LEFT not irrelevant? > If someone did achieve full equivalence between RIGHT and MIDDLE, then yes, you could claim that the natural physics is irrelevant and peers may accept it. But that equivalence demonstrably never happens. What actually happens is superhuman narrow AI function that is learned at an abysmal million-fold slower rate than humans, lacking the autonomy humans (nature generally) have, and having terrible fragility to novelty of a kind not displayed by humans (natural general intelligence). It's the real, full equivalence that signifies you've achieved the result. This never happens. Not even close. It may not matter in a commerce context. You have a valuable narrow AI tool that functions well. But as *science* it completely misses the target ... Full literal equivalence with nature..... It learns autonomously at a rate comparable to natural brains, and accommodates novelty like natural brains. So yes, in theory the circumstance (an extremum in the conduct of the science) can occur. If you've been following the discussion, you'll see that assuming RIGHT is the only way to AGI, no matter how long it takes, and never knowing for sure it that is even possible or what computer resources and model-kinds are mandated if it was, ....is the current approach exceptionlessly adopted, without a discussion of the kind we are having. In AGI, if you want to know when brain-physics (LEFT) and a computed model/abstraction (RIGHT) part company, why, and the implication for AGI done under RIGHT, then you do both and see when and how they differ in the matters above: knowledge area, learning rate, fragility in the face of novelty ..... with MIDDLE as the gold standard. If someone makes a LEFT that is functionally equivalent (demonstrated via > empirical evidence) to MIDDLE, then is RIGHT not irrelevant? > In science, the function/purpose of building artificial versions of a natural phenomenon is to help validate or discover or test the predictions of theory on the RIGHT. Building LEFT can suggest novel experiments in MIDDLE. The whole framework is a mutually interacting single behaviour adapted by scientists. So, an artificial version of the MIDDLE, at some level of fidelity serves brilliantly in empirical science. It is the norm everywhere except here in AGI as currently conducted. Why does it matter if LEFT and RIGHT are equivalent if the goal is MIDDLE? Because this is science. Specifically neuroscience. Q. How does a natural brain physics deliver intelligent behaviour? And, most importantly here, where for the first time ever, there is potential for literal equivalence of brain physics, and a computed model abstraction of what it is, or what it does. Q. What are the characteristics of the model of a MIDDLE brain and it's hosting computer on the RIGHT, and intelligence of the kind in the middle? What deficit happens when which bit of brain physics is thrown out? Is there any aspect of natural general intelligence critically dependent on the use of what aspect of the natural physics. A science confined to RIGHT is fundamentally inpoverished and does not have a scientifically, empirically proved claim of the nature of the LEFT/RIGHT equivalence. That is the real current state of the science. It is walking blind into whatever computers can do with absolute zero evidence of the kind of trajectory it is on. The evidence so far is that human intelligence and artificial intelligence created on the RIGHT part company on day 2 and never even come close to each other. You can obviously produce really useful automation based on RIGHT. That is moot to the argument here. --------- So I hope you can see this discussion is about the conduct of a science of AGI, properly recognised as being within neuroscience, and insofar as it needs to create an artificial version of a natural intelligence, so that practitioners are properly informed by proved neuroscience. It's unique and privileged in being the only place in science where nature and a model of nature, can potentially reach functional equivalence. Doing the science of that equivalence is not optional ... And here, now we are all having the first real discussion of how that science must be conducted including an empirical branch discarded in 1956. I hope I am communicating this clearly. Happy to keep going. It's important. Cheers Colin > On Tue, Jul 2, 2019 at 3:54 PM Colin Hales <[email protected]> wrote: > >> BTW, Brick, welcome to the fray! >> It's hard work, but worth it, I think. >> :-) >> Colin >> >> >> On Wed, Jul 3, 2019 at 8:34 AM Berick Cook <[email protected]> wrote: >> >>> So what you're saying, Colin, is that the computation that occurs via >>> the binary transistors of a computer is fundamentally different than the >>> computation that occurs in the neural structures of the brain? And you take >>> issue with people assuming that both forms of computation "can create >>> equivalent levels (indistinguishable) of intelligent behaviour"? >>> >>> On Tue, Jul 2, 2019 at 2:18 PM Colin Hales <[email protected]> wrote: >>> >>>> >>>> >>>> On Wed., 3 Jul. 2019, 7:05 am Matt Mahoney, <[email protected]> >>>> wrote: >>>> >>>>> So if computation is not behind intelligence (based on 65 years of AGI >>>>> failure) and you have no idea what is, then what is the basis of your chip >>>>> design, and what do you hope to accomplish with it? >>>>> >>>> >>>> This just keeps on happening. >>>> >>>> I have explicitly said more times than I can remember that the brain is >>>> 100% computation. >>>> >>>> What said is that it is not a 'computer'. >>>> >>>> The basis of what computation happens in a computer is the physics >>>> (causality) of a computer. The computations performed by the brain are >>>> performed by brain physics(causality). >>>> >>>> This entire discussion has been about the assumed equivalence of these >>>> two things, and how it would be scientifically proved. That proof happens >>>> when the science does both (e)RIGHT (the computer) and the brain physics >>>> (e)LEFT are compared and contrasted..... As a matter of normalisation of a >>>> science that currently lacks the latter. >>>> >>>> The substrate independence hypothesis is a hypothesis that the two can >>>> create equivalent levels (indistinguishable) of intelligent behaviour. >>>> >>>> This is how you prove it: for the purposes you have to compare the two, >>>> not assume it true. >>>> >>>> Have I finally thrown the ball over the home plate? >>>> >>>> Colin >>>> >>>> >>> >>> -- >>> Berick Cook >>> Independent Developer >>> AI Research <http://airis-ai.com/> >>> Games / Software <http://berickcook.itch.io> >>> YouTube Channel <https://www.youtube.com/c/berickcook> >>> > > -- > Berick Cook > Independent Developer > AI Research <http://airis-ai.com/> > Games / Software <http://berickcook.itch.io> > YouTube Channel <https://www.youtube.com/c/berickcook> > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + delivery > options <https://agi.topicbox.com/groups/agi/subscription> Permalink > <https://agi.topicbox.com/groups/agi/T87761d322a3126b1-M943d9b4ab4a29067f15008d0> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T87761d322a3126b1-M509eef3eb5a39185e4e556cb Delivery options: https://agi.topicbox.com/groups/agi/subscription
