Depends on where "there" is. Obviously we have different destinations. As for your other arguments, I would simply say, make a useful prototype. Or not. It's up to you. ~PM
From: [email protected] Subject: Re: [agi] Re: Could Brain Emulation be NP-Hard? Date: Thu, 25 Jun 2015 17:42:34 -0700 To: [email protected] On Jun 25, 2015, at 4:39 PM, Piaget Modeler <[email protected]> wrote:I don't see anything special about what you are attempting. I am not attempting anything, I am pointing out upper bounds on computation from naive assumptions. All of your examples are trivial, toy cases. They cannot scale to anything interesting. That’s just computer science. It is absolutely possible to do complex spatiotemporal reasoning and pattern learning around e.g. entity-level population behavior and dynamics at extremely large scales, in real-time even. Organizations are doing this on continental scales today. But these systems are obviously not based on any of the naive computational models that you think are relevant. Scale has a quality all its own. First of all, you are NEVER, EVER, going to observe ALL of reality. That is imposible. Strawman much? I was responding to your assertion that observing what can be observed is not useful or important. There is enormous amounts of value to fusing as many unrelated data sources as your system can handle it you care at all about the quality of the result. Finally you DO just look things up in reality. For example you take an image of a room and see a cat sitting on a bed. You take a later image of the same room and see the cat is no longer on the bed. That is "looking things up", reality checking, to ascertain your state, or current situation. Huh? You collected two samples at different times. You did not know if there was a cat there when the first sample was collected. You did not know if there was a cat there when the second sample was collected. But you can look for changes between the two samples because you arbitrarily collected two samples and compared them. A system does not contain facts about reality that have never been observed. In the real-world, you need to constantly be collecting as many “first samples” as you can so that you can notice that the metaphorical cat is missing in a later sample. I don't think you've fully baked what your learning algorithm(s) is(are) for an AGI. I suggest you do some more research before making sweeping statements that only reflect your lack of knowledge, or lack of an approach. Who is talking about AGI? I was putting design constraints on an AGI that has the ability to learn and reason about spatiotemporal relationships in the physical world. Unlike you apparently, I would have no problem designing software within those constraints, AGI or otherwise. If some things can't be done, then don't do them. Do what is feasible and computable and make strides there. Show a proof of concept, for God's sake. Let's see a prototype of what you CAN do. That was my point. It can be done and is being done, no need for a proof of concept. But you absolutely cannot get there with your concept of how one might go about this. AGI | Archives | Modify Your Subscription ------------------------------------------- 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
