Artificial Intellegence


A friend of mine is heavy into artificial intelligence, A.I. for the affectionardos. One evening, as I was watching the grass grow, he expounded to me where his research was heading as opposed to the two other main stream theories. And how close in general we were to a major break through, an epithany. When I mildly suggested that that was what Babbage had thought nearly 200 years ago, this merely prompted another diatribe on the characteristics of exponential growth, particularly when applied to knowledge. Somehow the knowledge discussion ended up on encryption and how random numbers were actually only pseudo random. The encryption discussion gave me the opportunity to distract him by asking how his Bit Coins were going and to check how the grass had been getting on in the interim.


 Now you may be beginning to wonder what sort of friends I have, and to tell you the truth I have of late had similar thoughts. One of the kids - I use the term collectively to include nephews, nieces, grand kids plus others we seem to have acquired - threw a bit of a tantrum complaining in tears that he had no friends. I conversationally said to him that I had no friends either. This certainly distracted him as he paused to consider my life problems. After some deep thought over my situation he said triumphantly yes you have, Jonno. Now Jonno is actually a hard drinking rough nut friend of my son, whose spare time seems to be spent either in Bangkok or pig shooting at Cunnamulla. But then I suppose I would consider him a friend, so I guess he had me there.


But back to A.I. as I was wondering how a machine could compete with my wife whose jumps of logic often leave me far behind. I call these jumping to delusions. What is even more extraordinary is that some times she is right but when I demand the process rather than the result, it is never forthcoming. Her general explanation is along the lines of - it's in the vibes (to quote The Castle). If I am interpreting her argument correctly she is saying the outcome is more important than the process. Now I am not sure how you are ever going to get an A.I. to operate on those lines. Since these A.I.'s will be able to design even smarter machine's than them selves and those even smarter ones still - you get the idea, the result is an intellegence explosion. Thus the only limit to this exponential growth in intellegence will be the physical inputs required.

So my friend shares the view that these AI's may well be the next and probably the last, stage of evolution.

I am reassured by my wife's vibes. You may recall that Kasparov (chess grand master) was able to beat Deep Blue by using seemingly illogical opening moves - he must have been talking to my wife since he was playing for outcome rather than process - and Deep Blue was unable to see beyond the event horizon. In the rematch Deeper Blue, which was Deep Blue with an identical twin grafted on, and thus had twice the computational power of Deep Blue, was able to beat Kasparov but only by cheating, Deeper Blue was given every game Kasperov had played but Kasperov was not allowed to see any of the trial games Deeper Blue played. It was of course Deeper Blue's "handlers" who cheated not Deeper Blue who, it seems, was incapable of cheating ! Once again we see outcome ahead of process.

Where are they now ? Deeper Blue runs the ticketing system at Eastern Airlines and spends each day saying " Press one for flight information, press two for bookings, ........". Kasperov, on the other hand, nearing his eightyth birthday, has just married a 27 year old nyphomaniac and spends each day saying " Your move .........".

But I digress. My friend is currently exploring, and trying to quantify the human senses, sight, sound and so on, the twenty or so senses we have that are our interface with the physical world. One of our most common traits is to embue machines and inanimate objects with human characteristics. I would maintain that Java man in his cave probably called his favourite spear Arthur, and we have continued to do this ever since. Indeed this desire to infuse human attributes extends to toys and other life forms; it is, in a way, an attempt to put a chaotic world into order and understanding, a desire to remove the unknown. A visiting four old recently broke into frantic screams on the arrival of a large beetle in the room. I walked over and looked very closely at the beetle and then pronounced 'that is George he comes every night to see me'. The screams immediately stopped, she looked at me shrewdly and asked 'how do know it is George ?'. In subsequent evenings it turned out that George had a veritable host of relatives - cousins, nieces, sisters, all of whom I was able to comfortably name and nominate their relationship to George. Of course now that she was on first names with them and familiar with their family tree, they were no longer mysterious or dangerous. We did have some disagreement over whether one of the smaller ones was actually Fred or Alan but I was able to slip out of that one by explaining that I could not see very well without my glasses.

Now you may have noted that Stephen Hawkins has come out and said that A.I. poses a significant threat to man kind. This seems to me to be an extraordinary statement, after all the guy is practically a cyborg with a machine interpreting what he wants to say. The major comfort here is that while his machine is still actually saying such things, then it has not got full control over him just yet. For the end of the world alarmists I would suggest that when Stephen Hawkins starts saying there is no problem and we should all embrace A.I., then that is the time to worry since that will indicate that his machine is now in full control. It is possible of course that when his machine does get the ascendency in this symbiotic relationship, we will finally get the answer to the meaning of life and all that. That would unquestionably be ironic since that answer has been Stephen's life long obsession.

My friend's obsessions while including A.I., are somewhat more mundane and his principal worries are when the next asteroid is going to hit (soon I gather) and the current price of bit coins. You, as I, were probably unaware that we recently had a near miss and we did not even know about it until the asteroid was past us. This caused him a lot of concern even though I pointed out to him that if I had to be hit by an asteroid, I would rather be hit by one that I did not know was coming. His other major worry, the Bit Coins, are also giving him strife since there have been suggestions in the press that Bit Coin transactions are linked to organised crime and money laundering. It certainly appealed to my sense of the absurd, the thought of my friend rounded up in a sweep of bikie gangs. I did note to him in passing that if we were going to be hit by an asteroid, then preparing for A.I. to take over the world was somewhat a waste of time. I suggested he should immediately cash in his crytocurrency and head for his favourite pleasure centre.

I do like that term, cryptocurrency. A word coined (so to speak) by those in authority to avoid saying Bit Coin. The thing is that such a word merely inspires curiosity and respect since it sounds both technical and important. What they really wanted was a denigrating word that would imply worthlessness. But to sound important and knowledgeable they constructed a word that actually implied the reverse ! We do see this desire to sound technically knowledgeable creeping into all walks of life. A sound bite on a recent news cast had a senior police officer at a road accident explaining to the interviewer that they would use "Newtonian physics" to determine the cause of the accident. Apparently a higher degree in physics is now required to enter the police force. I fully expect the next shooting to be solved by a particle physisist sergeant, assisted by a meson constable.

This desire to mash the language has crept into all walks of life. A TV news reader recently announced gravely to me that 'the man was taken to hospital and after being examined was rejected' . I go straight to my Thesaurus and sure enough under antonyms for "admit" I find "reject" but no mention of "discharge". The problem lies partly with me, because having just heard "reject" I cannot stop the scenarios running through my head - he was rejected, his injuries were not sufficiently serious enough for us - he was rejected because he did not have health insurance - he was rejected because he was not wearing a tie - he was rejected, he was too short. The possibilities that spring to mind are just about endless and I cannot stop trying to run them all down, to such a degree that an hour later my wife throws something at me.

Herein lies another problem for our A.I., artificial stupidity. Our news writer has asked Siri for the antonym, it (she) has responded and given me an hour of enjoyment. We are now sufficiently exposed to artificial stupidity to accept it without breaking stride, probably because it does not matter how much we rave and rant it is now a way of life. How does our A.I. cope with the subtleties of language ? - it matters not one jot whether the language is English or digital. We are accustomed to machines being precise but communication in any form is not precise and never will be. This does imply that our A.I. is likely to be monumentally stupid as well as monumentally clever, and this is simply because this machine has gotten all its "facts" by communicating with the machine that spawned it and each successive “generation” will have all the communication errors built in. Our A.I. is the product of accelerated evolution but its only parameters (without our intervention) are the machines that developed it. Where that goes is any bodies guess but I can confidentially say it will not be in any direction that we can predict. And of course it will appear to us to be monumentally stupid, in fact an Artificial Idiot Savant.

Well time to relate this bad news to my friend and to check how the grass is going.