Yes on brain repair & self-repair, no on AI?
from the vision:-two-yes,-one-no dept.
from the New Scientists' Next Generation Symposium site: "Welcome to the future…it's getting seriously strange out there as we head for the millennium. Below 24 young scientists working at the cutting edge bring you their thoughts and predictions. Check them out before you take your journey into the future… " Included:
Brain Repair in the 21st Century: "How much of the brain can be replaced before you require a new passport?"
Soft-condensed matter: "we could design desirable structures without actually having to build them, and if you break them, they will 'repair' themselves…we'd really like to have systems that completely self-assemble and produce hard bits and soft bits and valves and pistons and all the necessary things we need to make nanomachines, all from exploiting the properties of soft matter."
AI is possible, but AI won't happen: "there is no obvious way of getting from here to there–to human-level intelligence from the rather useless robots and brittle software programs that we have nowadays." [Yes, okay, it's not obvious.]
August 26th, 2000 at 6:47 PM
Way to AI
Well, we'll never have AI the way most research is being done now. We need Mark Tilden/biomorphic like algorithms running on computers to get some actual thinking, which will, hopefully, lead to intelligence. An expert system or the average neural net my have plentry of knowledge and be able to call it up quickly, but that cartainly won't make it think.
August 27th, 2000 at 9:46 PM
No AI but CAI maybe…
I agree, by the time we perfect AI, hopefully MNT will have provided a tool to augment human intelligence through some kind of neural link. Work in this field was reported on this forum last week. Why reinvent the wheel? No need for AI. Computer augmented intelligence will allow us to have processing and memory capabilities equal to a computer while retaining human creativity and intuition.
August 30th, 2000 at 1:15 PM
uh uh
There are no grounds whatsoever for claiming "creativity" and "intuition " (whatever intuition is it is not metaphysical but a right proper deterministic property of neural connections) as in any way solely "human" properties, whatever that means. These distinctions are largely meaningless. Meat or carbon/silicon – all obey physics. all obey its implications.
August 30th, 2000 at 4:51 PM
Re:uh uh
Nor is there any reason to think that creativity or intuition are necessarily emergent properties of every simulation of a creative, intuitive human brain. We do not presently know what level of detail will be required – neuronal group, neuron, organelle, chemical or quantum. If such simulations require much sub-neural accuracy, then cyb-org hybrid approaches will be most efficient until simulations catch up. We need MNT before we can know how difficult brain simulation will be. If quantum accuracy is required in any way, it may take quite a while to get a simulation that large numbers of people will be willing to use. It seems quite possible to me that there will be a period of years to decades when our capabilities will be too limited to allow near-perfect brain simulation at the required chemical or quantum level, but will be sufficient to permit safe brain-computer interfaces. This will not be true only if quantum effects are negligible or small, high-qbit, long-coherence quantum computers with practical environmental requirements are buildable and programmable prior to the development of MNT brain-computer interfaces.
Even so, nearly everyone will wait decades at least before they jump in with both hemispheres.
August 30th, 2000 at 5:09 PM
Penrose : out of his field
Brilliant guy , but he should stick to Mathematics. His attempts to cloak religious attitudes toward mind in quantum mumbo jumbo are damaging in the extreme to science and reason.
August 31st, 2000 at 11:35 AM
Re:Penrose : out of his field
I would have agreed with you completely until last week, when I read Dimtri Nanopoulos' paper: http://www.qedcorp.com/pcr/pcr/nanopoul.pdf (warning-long download) . Certainly Penrose's speculations were not at all persuasive, but Nanopoulos' paper moves the quantum-microtubule hypothesis into the long-shot maybe category. His sources have persuaded me that microtubules are the immediate cause of synapse formation/removal. If so, then some means of interaction between the firing of neurons and the microtubule's pattern of synapse formation is needed. A non-local quantum explanation is possible but not likely. Microtubules are also leading candidates for the complex, adaptive behavior of individual microorganisms. Regardless, microtubules are quite interesting because of their status as the most common self-assembling mechanical structures in biology.
My intent in referring to a possible need for quantum simulation in my previous post was not to get into any such murky questions. Quantum simulation could be necessary for much more mundane things, such as protein folding. Some single protein types fold up different ways depending on details of each molecule's local environment, serving different biological purposes for each state. There will be practical problems in determining just how every protein folds in every circumstance and what the significance of every state is in vivo even with MNT probes. Some states that depend on protein-protein interactions will be disturbed by probes or be too short-lived to observe directly. Even if such knowledge can be obtained through "conventional" MNT, if detailed protein behavior affects mental functioning (e.g. the evolution of synaptic patterns) , as I think is probable, then protein simulation on Turing machines would consume much more power than the simulation of the neurons themselves. Small, long-coherence, high-qbit quantum computers may be necessary to simulate brains, and are likely to be useful.
May 10th, 2001 at 5:45 PM
Re:No AI but CAI maybe…
Augmenting the human brain to allow it to work better, faster, etcetera is an admirable goal, but there is still a need to develop distinct entities (AIs) capable of reasoning from a new perspective.
As I like to tell my associates when the subject comes up, my goal is not to create artificial humanity. It is to create an adaptive and unique synthetic intelligence.
The true strength of such such machines is that they will be able to process information in ways that are fundamentally different from how we do it. This new perspective will offer fresh points of view when solving tough problems, and the time and effort needed to create them will be repaid in countless ways as the technology matures.
May 16th, 2009 at 3:08 AM
http://www.message_cabascacroll.com/