Mangled Thoughts

That is what the bogus science of AGW amounts too. Having compared AGW to it’s proper equivalents, the quackery of keynesians , sociology, pyschics and other mystic mikes. The believers continue to make ad hominem attacks on `skeptics’ and, the problem is, the `skeptics’ aren’t simply skeptics, they are scientists and they are shredding the AGW belief apart because it is not science and those engaged in it are not engaged in science, they are debasing science for supersitition and voodoo. It is rubbish, and so why has it been cultivated and proliferated. The culprits can be identified, politicians and bureaucrats. Consider Oz: apart from Greens and Dems m.p.s, and a couple of Earth worshippers in the Coalition and ALP, the rest in the major parties have embraced AGW only as an excuse to impose more exactions and controls. It can’t be, surely, they do believe it. There is no other reason and, later, we will back this up with evidence.

That scientists have demolished `evidence’ collected by boosters is something that is conveniently ignored. One strand of this `evidence’ one will recapitulate, `computer modelling’ and generated `scenarios’. The Federal Minister is fond of, or was, of proclaiming from the steps of parliament, `I have seen scenarios proving AGW’. Such statements,` I have….’, bye the bye, are alone a very crude indication of, what-ever the `experimenters’ are doing it does not include scientific experiments. The output of computer `models’, however, is most amusing. It has been found out, the generation of `scenarios’ `confirming AGW claims has a probability of no more than random chance. It has been found out, the out-put of one AGW type requires a large number of runs. So, what is never disclosed is the sheer number of `results’ which falsify claims. The most that can be said, if computer modelling is a true experiment, a warmer climate is one possibility but not the only one and that the probability of it is so low that, to the contrary of the claim of AGW boosters, there is no necessary determined relationship at all between human activity and climate warming . That there is no relationship at all, no cause and effect at all, is the force of those computer modelling results. Further, the results are such, it cannot be said there is a deterministic arrow of causation from human activity to climate at all, which is the force of, any possible scenario is no more than a random possibility and any one possibility has a very low probability attaching to it, so low that it barely satisifies one of two conditions of probability, it must be greater than zero.

The claim, however, computer modelling is a scientific experiment yeilding valid evidence, is entirely false. The problem is exactly the same as it is for A.I claims, a human mind can be instantiated in a computer. That is, what a brain is can be replicated in a machine. That has not been achieved and never can. All the `successes’ claimed by boosters of A.I. have been trivial tricks.
Rodeny Hide once posted a link on his website to one such trick.

The site, set up as world wide project in A.I., ran a programme, claiming the computer learns as you answer its questions -the first trick - anthropomorphism. Trick two. The programme run is intiated with what seem to be the most general of questions but in fact are narrow, e.g:

Do you have anything in your mouth?

The programme tree, contains lists of things which anyone might place in their mouth, e.g.cups cigarettes, lollies, biros.

Once the code ( the answer yes if yes) has been entered, the programme runs the next tree:

Question: Is it round /square/ oblong/ ( whatever shapes you care to code into the programme)?

Answer: Round

Q. Long, short

A. long

Q. What colour?

Now, readers can see what has been done, proceeding form the most general descriptions, each answer narrows the next set of possible descriptions so, carried to the last step, the computer arrives at the output, ` it is a cigarette’. And that is how such programmes are written. To amplify this, I did a run and to the questions:

Q. Does it taste nice?
A. Yes.
Q Is it enjoyable?
A. Yes.

And a number of such questions, the run reached the conclusion: `I give up, what is it.’

Answer , a cigarette.

The next output by the computer was;

That is incorrect because, a cigarrette does not taste nice, it smells, it leaves ash, it is unhealthy, it is unpleasant, smoke is unpleasant.

Computers don’t reason, imagine, make associations, just to name a few things the brain does. A computer is a machine, and a progamme is just an extended deductive argument. It replicates nothing. The range of possible output is bound, no ouput which does not run from the programme is possible at all. Thus, computers can never learn anything at all, which is only to point up the nonsense of A.I.

Thus the point to using AI as an illustration, what has been `replicated’ in computer form is not the brain. What has been replicated is to,use a simpler example from maths, addition, with the addition function coded as algorithmic progression.

So long as any programme is a valid code, that is, the code is, as it were, grammatically true, the programme will run. The output, therefore, which has meaning only for the user of a computer, is trivial, it can be anything, even `garbage in and garbage out’, it can be completely useless, is trivial - or, to put it another way, a badly written programme is no use at all to some-one who wants it for financial analysis, or an engineer running calculations for a bridge.

Computer programmes do not at all, in conclusion, instantiate reality, no matter how impressive the output might seem. The output of Deep Blue was impressive but more impressive was the chess masters who analysed Kasparov’s moves and the positions during each game, and the It team who translated the recommendations of the masters into code and thus added search trees and instructions. Kasparov did not play against a computer, he played against those who built the computer, the chess masters and an IT team.

[I will finish this tomorrow, I have to buzz of now, but readers have enough for now to guage some of the lines of argument]