A.I.

[Craig]

Well, I'm of the camp that believes that it is impossible, in principle, for a machine to completely mimic a human mind. And I've put forward at least some plausibility arguments to support that. But Mike you cant take the (not unreasonable) view that it is impossible based on current technologies and the evil corporations. You need to put forward an argument that it is impossible in principle ever to do so, to the heat death of the universe.

In other words, is it possible, but just very difficult and needs new understandings and technologies? Or is it actually impossible, ever to do so?

[ppppenguin]

This is analagous to some of the problems in physics. There are some questions that we can see how to decide, but can't afford to do it now or the tech isn't sufficiently advanced yet. Then there are those where we can see in principle how to decide, but don't have a clue about how we might do it in practice. Finally there are those where we can't see how to decide. It is debated as to whether these fall within the domain of science at all.

[Craig]

Many physics discoveries are serendipitous. Like Penzias and Wilson. They were trying to increase the sensitivity of an instrument for radio astronomy with cryogenic detectors in the early 60's. They got a mysterious background noise, that they thought was either coming from New York (it wasn't) or from bat and pigeon dropping in the entrance of the detector horn (they removed them), and the white noise was still there. They had actually detected the cosmic microwave background, which decided the debate between continuous evolution on the one hand and big bang on the other as the mechanism behind the universe. Nobel Prizes.

And black hole mergers were thought to be rare to impossible until LIGO found they were actually quite common in the observable universe. Nobel Prizes

The two big unanswered questions are the mechanism behind dark matter, and dark energy. No one has any idea what these are. The observation that the universe is accelerating (So called Dark Energy) has already had Nobel Prizes. And there will be a bunch more coming out of these topics. Most of the universe is missing - watch this space.

[ppppenguin]

There are multiverse theories which look as if they are unverifiable by experiment. They are one of the ways proposed to resolve some weird aspects of quantum theory. There are MOND theories (Modified Newtonian Dynamics) that have been used to explain the problem of missing mass without requiring dark matter. Experiments have given very little evidence in this area but who know what might turn up. When somebody finds out what's really happening there will be Nobel prizes again.

Likewise in AI. We really don't know what's going to be found. By using more or less brute force with large data sets and self training algorithms we have had some limited successes. Could all be barking up the wrong tree. The critical breakthroughs might come from the computing side or the physiology side. Or possibly out of the blue from chemistry, physics or who knows. Ther may be slow incremental progress or some earth shattering discovery next year or in 100 years. The point is that we don't know how or when, but it's unwise to predict that it will never be possible.

[Craig]

I suspect that the electrochemical nature of the brain is an essential element of consciousness. Replacing it with a silicon based system is barking up the wrong tree. Quantum computers might have more success in that regard - with certain classes of problem the current generation can produce solutions in minutes that would take decades with the best supercomputers on the planet. Of course they need cryogenic temperatures to work using current QC's.

The bizarre nature of how they work is that a seemingly insignificant number or Q-bits confers massive power. 50 bits effectively addresses 2^50 states.

[BusyBee]

My thoughts are that we are maybe following the typically American look on things such that throwing in lots of money and building powerful systems will crack it. I prefer more abstract views. One thought is to completely divorce things from the human mind and try to see things from other approaches. Another thought is to forget about normal daytime thinking and go back to night and dreams. Random fluctuations in connections of neurons giving an apparently non-logical scenario. Sometimes I have found it works as things are linked to give inspiration. The unconcious giving intelligent behaviours whilst the concious gets on with living? Randomness is always a tricky thing with a strickly logical entity like a digital computer, in fact logically impossible although with billions of memory cells random failure could give it. Strict system management is contrary to this though countering cell failure and hence over-ruling with logic - counter AI! It reminds me of a homemade cell for a robotic 'animal', published, I remember many years ago. It's actions depended on silver 'whiskers' being randomly formed for determining it's actions. Maybe randomness is the root of invention. This is perhaps where my thinking may go if I were working on it.

Tracy

[reset]

There may be slow incremental progress or some earth shattering discovery next year or in 100 years. The point is that we don't know how or when, but it's unwise to predict that it will never be possible.

So you don't think we should close the patent office