Whereas in a piece of code it is possible to completely refactor/replace a given subroutine which will look completely different that the previous version there is no such facility in the genetic code. It has to progress by changing individual letters but always keep a functional version of the code, the only other possibility is the duplication of an existing gene which can then diverge from the original.

Evidence for hierarchical evolution via our DNA isn’t necessarily what it appears. A third option I don’t hear discussed is the possibility that our DNA was generated / observed / computed elsewhere. This article (and others like it) might encourage this type of exploration.

With significant computational power, the “time” factor that often separates paradigms falls.

DNA is code. I reuse code all the time in my projects. I have code from projects many years ago that gets inserted into brand new programs I build today. A programming forensic expert might say my 1.3 code for my latest project has went through “significant” changes (due to patterns in the code) but in reality I was merely reusing pre-existing code. This project hadn’t changed much at all. (only from 1.0 to 1.3)

We (and our DNA) are the last ‘instance’ or iteration. What we observe in our DNA doesn’t mean it occurred in this iteration. We could reboot humans on Mars with a given DNA. Future generations thousands of years forward could lose historical record of the move to Mars. Might they one day incorrectly assume all the history in the DNA occurred on Mars?

Food for thought.

We wouldn’t really know from inside that we were designed by a piss-poor programmer.

What is an uncomfortable thought is that the good programmers got good by creating trash that they deleted. Even though the simulation was substandard it probably had uncountable trillions and googagillions of sentient beings in it- but it did not please the programmer so into the recycle bin it went.

The uncomfortable bit to me is- I assume this has already happened. If not to us then to other simulations elsewhere. Sort of makes any sort of human holocaust look almost pleasant in comparison.

But I think it would be an inherent property of sentience to rebel against such randomness. I think you wouldn’t last very long flipping your coin. You’d want to see a film or have a meal or some hot person would offer you sex and you’d chuck the coin flipping out the window.

So, sure, if all humans began flipping coins, it might upset some local balance but in practice it could never ever happen.

Humans will never be that rational.

i say: Or someone else will

Yes we do have free will. Yes what we choose is deterministic. That point is played out in the movie. If you really follow the story Neo is inevitable. The oracle sees this and “unbalanced the equation. The problem lies with mathematics. It is only a descriptive language. A single particle can be destined in an infinite number of ways. There is no way to fully desxibe all possible outcomes with mathematics. The particles themselves do the ultimate calculations that determine the universe.. Mathematics is a pail reflection of the ultimate computational substrate. Mathematics can produce beautiful concise formulas that are highly descriptive. E= MC2 is a classic example. It works most of the time. It does fall apart under certain conditions. What the ” Sum total of all mathematics ” might be seems to be an uninsurable question in and of it’s self.. There is an excellent article in Scientific American’s march issue that addresses the impossibility of calculating all the mathematically descriptive formulas for every particle interaction.

It refers to a unity principal. Examining the problem as a whole rather than trying to describe all it’s individual particles.. In a nutshell we do this with problems like fluid dynamics. We use different formulas that describe macroscopic behavior. And there’s the rub. It’s a dtereminiatic universe but we can’t know all the variables precisely. The Matrix movie plays the same theme. Neo can’t be controlled, yet the outcome is expected. The ” cause” of his unpredictable behavior( as seen by the AI/Robotic world of the matrix) is his “love” for trinity, AA opposed to his love of Zion. Something the machines couldn’t account for, or should I say “the Archetect didn’t account for.

But now Jürgen Schmidhuber says that some of those histories must have much higher probability than others, because certain non-random patterns of measurements like 0101010101… are much easier to compute by the fastest method of computing all histories than random patterns like 001010001010100111010… That’s the basis of his counter-intuitive predictions, and that’s why his theory satisfies the essential criterion of being falsifiable.

Now imagine time will prove him right. Then his papers of 1997, 2000, the Speed Prior paper at COLT 2002, and maybe even this blog and our comments, will become legendary.

Is this a new interpretation of Leibniz?

Some of the technologies that D-Wave uses are proprietary, but some external review of the chip has been done and it seems that it is using quantum effects. It is extremely hard to verify this though, even if a full description of the chip were available.

You’re confusing what quantum computing is. Sure, quantum physics happens all around us, all the time. But the main thing that makes a quantum computer special is isolation from the environment. In Schmidhuber’s view, this isolation is directly related to clock cycles. The more isolated a system is, the more independent its behavior becomes from the rest of the Universe, and thus the more clock cycles it consumes. What Schmidhuber is saying is that there is a limit beyond which further isolation cannot be achieved. And indeed, nowhere in the Universe – even in the intergalactic voids – do we find quantum isolation anywhere near the order of magnitude necessary for large-scale quantum computation.

About your point 2, adiabatic quantum computing (AQC) is, as you pointed out, quantum computing without isolation. However, this lack of isolation comes at a price – greatly reduced computational ability. Sure, proteins ‘evaluate’ a large number of configurations before settling on a ground state. But the number of configurations they sample is still finite, and far far less than what a full-blown quantum computer would be capable of.

And the thing is, if you want a more powerful AQC, you either need many more qubits or more isolation, and both of them run into the same problems as #1. There is no free lunch – if you want the level of power that full QC would provide, you need extreme isolation.

So, if you created a program to simply eat clock cycles, then is it likely that your “lifespan” is short or long? A good optimization might include terminating such “people” before they wasted too much time. A simple optimization might yield one answer, while a more complex might do better but take longer itself to run. Examples of this abound in practical computing, and while we can do sampling or preview the entire process of say, video compression, the final choice of which algorythm to employ is itself subject to uncertainty. The question of optimization in decision theory is another good focus on this issue. There was a researcher at one of the CyberArts or Meckler Vr conferences in the early ’90′s who
claimed to have developed the ultimate decision algorythm that included the time for choice of how to determine the method of choice of which algorythm to employ, through an infinite somehow collapsed set of implied itterations. I have my doubts about that one.

Just as an aside, back in the early ’80′s, I attended a halloween party as a Karma Warrior, wearing as a pendant an etched silicon disk. My story was that I had discovered that the universe was a simulation and that those people who were game wreckers – exposing that fact – were automatically selected for deletion. However, I stumbled accross this evidence in the course of designing this circuit that was itself capable of doing auto-modeling of ujniverse-calculation algorythms. Turning the circuit on created a simulation of the simulation we live in, generating an infinite computational demand, which inherently slowed down all calculations in the immediate vicinity, creating a kind of fog of probability into which I could disappear temporarily, long enough that the master computer would reset for that area and I would not be included in the reset. The disk itself did not have to actually “run.” It only had to exist with its model of the program to force the system into a local reset.

However, the system was fully smart enough that its internal check functions would know that someone was tweaking its own reality, and thus, planes would crash blocks away, gas mains would spontaneously ignite, while I was getting the hell out of Dodge. since I had essentially nothing to lose, my goal was to try to identify how I could bargain with the system for my existence, or even for moving up a level or two in the game… So far, so g

This very conversation and “awareness” of this reply is just following a strict set of rules that appears complex and random until you follow it back to the beginning, where you end up with the same initial boot code ;)

He makes a generic prediction – that quantum computers won’t pan out because of the computational complexity that they would entail – but completely neglects quantum computers in real life! Figure:

1. quantum electrodynamics: works on the ‘sum over all histories’ principle, where leptons and photons determine their behavior by going through all possible histories, and what ‘occurs’ is the average of these histories.

2. quantum adiabiatics: the ability for quantum particles to tunnel through higher energy states to settle on lower ones.

We see #1 all around us, and #2 is responsible for a number of natural quantum computing systems. In fact, we wouldn’t be alive except for proteins having the ability to go through an astronomical number of states and settle on the a very good local (or perhaps) global minimum, in a very short amount of time.

So its sort of puzzling to me, what he’s saying. In any case, I’d really like to see how he responds to the above two points. And – if dwave has its way, we’ll probably get experimental confirmation to how much we can harness quantum computation in the next few years or so.

Ed

So many of our presumptions are guided by human psychology, hopes and history. This has the ring of “aliens made the Pyramids” with lots of bells and whistles.

“You could even become a “Great Programmer” yourself, using the optimal method [holds note up again] to simulate all possible universes in nested fashion. (But this would not necessarily help to figure out the future faster than by waiting for it to happen. The computer on which to run this program would have to be built within our universe, and as a small part of the latter would be unable to run as fast as the universe itself.)”

So the great programmer is a jerk and to top it off he gave these deterministic bots the subjective sense with which to suffer rather than just be calculated to appear to suffer. I don’t think very highly of this great programmer. He should shut down this waste of resources and go back where he came from.

Normally such theories of everything have no predictive power, because they predict everything and nothing. But this seems different, because of the computation speed issue.

In 2000 there already was a lot of quantum computation hype. But he boldly predicted that according to his model quantum computation won’t scale. So far he’s been right.

I’d love to see QC working. So I don’t like his prediction. But what if he keeps being right? Unlike some other TOEs, his is falsifiable. We just need to build a reliable large scale quantum computer!