Photon counting for the masses

I came across a quote by Marvin Minsky about AI. He said it was between three and eight years away. He made this statement in the mid-seventies, if I recall correctly.

I agree entirely. There are good reasons to suppose they will never be observed even if they did exist.

And so far relativity has resisted every extreme test we have tried.

The tricky bit is unifying gravity and all the other forces of nature. We are clearly missing something and it will only be obvious what about a century after the right mathematical formulation is discovered.

Even today you still get a bunch of loons and nutters insisting that Einstein was wrong based on their half baked misunderstanding of him.

And any such more complete theory will subsume general and special relativity and all its predictions in the weak field limiting case.

Agreed entirely. There are however a handful of cranks around that will clutch at any straw.

Yes, except that in a paramp, the signal and pump are inputs. Does a BBO crystal need to be pumped to effect down-conversion? If that is so, I understand how it works.

Jeroen Belleman

The difference is the homeopathy and dowsing claim to have results in the physical world. They don't confine themselves to the spiritual realm or domain.

I agree that anything that applies to the physical world has to have some observable evidence; otherwise, it's not "real." Or even if it is real, it's not objectively real.

I absolutely agree with you. Hence my objection to Fundamentalists (or anyone) who want to try to apply spiritual values to the physical world, contrary to what experience and observation show us.

This is why I say that religion needs to stick to its own domain, which is the spiritual; and leave the objective physical world alone.

No, all you can do is examine the behavior of someone, and try to infer what motivates them. The only reason we infer that a mother loves her child is that almost all of us have the experience of love.

We have this problem when observing animals, for instance. Some people deny that animals have emotions similar to humans. So when a mother gorilla protects her young, it's an "instinct." But when a human mom does it, it's love.

We don't know whether other animals experience love. All we can really say is that they behave much the way we do, when we operate out of love.

Exactly. I think maybe I didn't make myself clear about this. I agree with just about everything you say. When I say religion needs to stick to the spiritual realm, I mean that it should not be applied to the physical realm. Not that it should somehow be exempt from the rules of logic and observation, but still have a say in physical reality.

As for emotions - nobody doubts them, because almost everyone has them. Spiritual experiences are not as generally recognized. Hence, someone may have a spiritual experience, and others may assume it's untrue because they themselves have never had such an experience. Since it's a subjective experience, there is no way to demonstrate that it exists; since it's a *personal* experience, there is no need to demonstrate it. But it should be kept out of the physical world.

Yes, indeed...

Agreed. It's a personal, subjective thing.

Again, I totally agree with you. Even though I believe in ... something... there is absolutely no need for me to convince you of it. I don't get points with God if I make converts. If God wants converts, He can get someone's attention way better than I can.

My spirituality is a personal, subjective experience. It doesn't - and cannot - influence my physical world, except perhaps as it directs my behavior.

I think what you object to is the situation where people try to force their deity or their holy book onto reality, and try to suppress anything that conflicts with their religious beliefs. So you have Fundamentalist Christians who believe in a 6,000-year-old world. They've strayed from the proper domain, and tried to apply spiritual claims to the real world. Worse yet, some of these people are trying to get this taught in schools, and are also trying to suppress established science such as Darwin's theory of Evolution, or the Big Bang theory, etc.

This is a Bad Thing(tm). In case I was unclear before, religion has no place in describing the real world. It was once of some use in establishing social order, but those days are long past. As for describing physical reality, religion is entirely useless.

My sigs come up randomly, but in this case it's highly appropriate.

Minsky was wildly optimistic and obviously wrong even at the time.

David Levy made a bet with top AI proponents in 1968 that no chess program would be able beat him within 10 years. He was right and with a factor of two margin as well. Chess 4.7 came closest in 1978 4.5-1.5.

It was another decade before a dedicated LSI hardware assisted Deep Thought beat him in 1989 winning $1000 of the follow-on bet and $4000 donated by Omni magazine. The Deep Blue descendant went on to beat Kasparov in a match where the world human chess champion was in part out psyched by the machine. He thought the operators were cheating.

These days things have advanced to the point where you can buy shrink wrapped PC software for $10 that will easily beat Kasparov in a match. There are several free engines of similar strength.

Chess proved to be easier than was first thought yielding to a combination of brute force attack and sophisticated heuristic searching.

Go is still resisting AI research. The best humans can still trounce a computer by using deep strategic pattern matching skills.

I guess we have a disconnect - I don't see how it disallows it.

A purely mechanical universe can allow "a succession of states of the brain proceeding purely by physical causation". These states cause an organism to react to its environment. More sophisticated organisms can start to learn, predict events and make plans. Create internal models of others.

We can *see* parts of the mechanism in action, we can trace synapses in flatworms and measure potentials etc. From the other end of the spectrum we can do FMRI of working human brains. We can program computers to do logical reasoning in limited domains.

I can't give you the whole story, and neither I suppose could a neuroanatomist given our current state of knowledge. So yes it's handwaving from here, but I don't see where the huge gap is that proves your "logical thought is impossible in a mechanistic universe" claim.

That's absurd, tea-leaf reading voodoo science. I my experience, leftists are more neurotic, more dependant, less adventurous, and more afraid of life. And less generous.

There are also a smaller

A lot of supposed kids' allergies are actually parent neurosis.

This morning my wife, a speech pathologist, cooked a bunch of black foods for a kids' feeding clinic. She's taking this group through foods by color, to make eating seem like a fun game. She gets lots of kids whose parents are neurotic loonies, and make their kids food-adverse by forcing them to eat stuff they don't want to eat. I help her out and give her occasional advice, since I have pretty much the appetites of a six-year-old.

She did just cook some wild rice, because it's black. I predict the kids will hate it. Stuff tastes awful.

Enormous Kleenex might have saved them.

If you mean "rational thought" to be the ability to observe and act to better survive, that could be purely mechanistic evolution. If you mean "thought" to be consciousness, *that* is the great mystery.

I suspect that consciousness and quantum uncertainty are deeply related, but that can probably never be proven either.

Indeed. Quite simple simulated digital organisms with trivial genomes rapidly evolve to match their environment or else die out.

Although it could still be an emergent behaviour in very large networks that are as complicated as the internal wiring of the brain which is roughly 10 billion neurons each with around 10,000 synapses.

Penrose makes the same conjecture in his various books. I am not so sure. I think we will be able to understand consciousness once we can completely model the brain in digital/analogue electronic hardware.

Some parts of the brain subsystems have already been modelled well enough to demonstrate the right sorts of patterns. IBM prematurely claimed to have modelled a cats brain but that was quickly debunked.

No. I always thought it was a rather beautiful result.

In practice you only have to worry about the path of least time give or take a coherence length or several - everything else integrates to zero.

OK. This is a handwaving explanation of what happens in the time gated photon version. You in effect have taken a snapshot of the photons in that narrow time window and in doing so crystallise the energies that they happen to have during that short period. The perfect vacuum isn't truly perfect or empty and there is a sea of virtual photons everywhere with energy being borrowed and returned at short timescales.

You can borrow energy according the Heisenberg uncertainty principle dE.dt >= h

In a very real sense your experimental setup rather neatly defines the hard limits on dt and so forces energy dispersion on the photons.

But what better way to survive than to create ever-more sophisticated responses to situations? The ability to outwit rivals would seem to be a strong driver of this. The ability to predict events, create models of what is going to happen... models of others, including a model of yourself. What is "consciousness" other than that? If I act exactly *as if* I am conscious, but am not really according to your (or Phils) definition.. would you be able to tell? Would *I*?

I don't see it is necessary, although perhaps it could be.

Suppose we did push a high-power, very narrow light pulse through an e/o modulator. Then pulse the eom to gate the pulse, but with some off-center waveform. I'm thinking you could net change the wavelength of the pulse (think SSB modulation) up or down. Assume up. Wouldn't the eom have to extract energy from the light pulse? Shouldn't that be visible in the eom drive waveform?

With detection being probabilistic, how can one be sure about the "exactly one, no more and no less" and "disappearance of the rest of the wave" for any single event?

Jeroen Belleman

Is there ever a case where two or more photons are simultaneously absorbed by an atom, to jump an electron by the sum of the photon energies? That would be nonlinear on beam intensity.

.... YES!

No, actually there is a basis for it. To some extent, it does appear that our political leanings are determined by our brains' wiring. That doesn't invalidate any political leaning, nor does it say one group is "healthier" or "more generous" than another. It just suggests that we don't necessarily base our political beliefs wholly on rational decisions, despite what we may tell ourselves.

Feh. Very disappointingly so, too. For one thing, he did a poor job of labeling it as conjecture. He said "'taint one, 'taint 'tother, must be quantum" as I recall. Essentially argument by exhaustion when exhaustion has not been shown...

Skyhook.

Nor am I, but it sounds like a productive thing to study.

I don't know that, either. We can't really explain language, at least not human levels of language.

We'll see. If/when this occurs, I predict a massive disruption of the social order. Robin Hanson talks about this frequently, and what he calls "ems" ( emulations ) would require a lot of ethical and economic adjustment. The politics of it are modestly horrifying.

-- Les Cargill

We all know that none of those people die from flu, because they are old or have a poor immune system. We also know that not one cent is spent to prevent ot treat people for flu.

I don't think chess has much to do with AI. We humans used to think it did, because chess is hard for us and the people who are good at it tend to be highly intelligent people. However, it can in theory be turned into a set of automatic actions and responses. Because of this, I don't see it as being more "AI" than a tic-tac-toe game. Just way, *WAY* more complicated.

But of course this raises the question, what *is* AI, or for that matter, what is human intelligence? I haven't come across anything better than the Turing test. If a computer can respond to a human in such a way that the human couldn't reliably say whether it was a computer or a human, then the computer could be said to be intelligent.

That's not a really satisfactory answer, in my opinion. But I haven't yet seen anything better.

It also seems that every time computers manage to accomplish something we once thought required intelligence, we move the goalposts. It can play chess? Well, OK, but can it write a symphony?

BTW - I remember reading how a computer that could play chess would take up as much room as the Empire State Building, and use more electricity than could be produced by Niagara Falls. My, how things have changed...