So, what about Robot, dreaming in New York, like an undocumented immigrant from Mexico, of the status of his great-grandchildren?
What about his evolutionary chances?
I got a phone call from Hauser. I'd sent him an e-mail, telling that I was interested in the faculty of language, either broad or narrow. Hauser invited me to his place. So, I made the trip to Cambridge.
» Okay, you wanted to talk. Talk.
» Sir, can you tell me about FLB and FLN.
He nodded to me.
» Go on.
» I asked you a question.
» Just talk. I'll listen.
» Talk about what.
» Don't bother. If you want to talk nonsense, feel free. Next week I'll answer your question.
» Okay ...
And I talked and talked ...
Next week I went back.
He showed me a scheme.
» You say, this is what I'm doing?
It appeared that what we call chatting or a serious meeting isn't talking or discussing.
- we do grammatical computations
- our phonological converter tunes the results for our sensory-motor
- our semantic driver tunes the result for our conceptual-intentional layer
in order to be sure that the message we want to deliver is delivered by our mouth!
Did you know that?
I do not bother. It's our way of thinking since we were kids.
For the Greeks we were just water earth fire.
After Galileo and Keppler we consisted of solar systems and galaxies.
When we invented the watch the human body was like a watch.
We invent God: humankind is his image.
We build systems: we believe we are systems.
That's our fate.
Have you ever eaten spaghetti à la Leonardo? Can you imagine, having a plate in front of you, prepared by me?
Suppose you come along my place, just to say: nice post Leonardo!
And I invite you to have dinner with me.
But you are in a hurry.
No problem, I say, just a quick and dirty (think of the dirty rice of Autogrill, but not that dirty - I don't like chicken livers or giblets)
There it is: a tangled ball of spaghetti with egg and bacon, tomatoes and herbs.
And you do: hmmmmmm
(thanks for the compliment).
And I say: Nice system ay?
I understand your confusion.
But there exists something like spaghetti code.
Programs which flow looks like a bowl of spaghetti.
As we now and then made quick and dirty programs - temporary provisions in systems because there's not enough time or money.
(By the way: forget about temporary.)
So, where does that leave Robot, eating from the pizza tree.
Jan is still fervently supporting the idea of evolution.
Some husbands are looking forward to it, a nice robot in the kitchen ... leaving the wife with mixed feelings.
Ah, systems and subsystems.
But there are two problems with our metaphor.
Our systems are discrete and far from perfect.
Natural systems lack discreteness and are perfect ...
... sorry, no objection, my phonological converter is still tuning up for the grammatical computations of Greenpeace and the climate movement.
Of course, Hauser and Co are talking concepts. But, reading their article you can hear them thinking: well-ordered systems, and logical, reusable components.
But what kind of law tells us that nature provides us with well thought-out modular systems. Nature has also a lack of time. What if there's an impact of a meteorite, or something like the first industrial revolution in England? Nature has not always time for structured programming.
And one thing for sure. Quick and dirty and spaghetti code suffer a flaw in common: you can't use these techniques for a component in a body of functions, procedures and programs.
The problem is: seeing ourselves as systems makes also the reverse true. Because we have evolution behind our shoulders, we believe that these systems we are building have an evolutionary future waiting for them.
Those who hate elections hope that the theory proves right, which means you will be rid of presidential and other hopefuls.
But, can Jan, the writer, be sure that no robot will beat him in what a scholar perhaps likes to call the function of word-processing but what I like to see as a brilliant manipulation of characters words text form.
Once again: what about Robot's aspirations?
I know three arguments for the absence of evolution in robots - sorry Jan, I am what we call in Italy un testardo.
- it's robust programming, certainly no quick and dirty style; so you can trust the processor feels comfortable in his envelope
- we pamper them: our robot has nothing to complain about his "natural" environment - there's no incentive for changes
- if there's something to improve we take initiative, way before the robots can even think about it
» Sorry Robot, the first scenario was the best for you. Leave at once New York and return to that police barricade. Don't do what they ask you to do. Let them fire that anti-aircraft missile. It means an honour to you and it will be their shame. Jan and I will create a new robot. I'll call him Robot II.
*) this is a sequel to my post Robot
This comment may seem obvious, but probably Robot II will need some kind of ‘body’ with which he may interact with the environment. Why concentrate on this ‘brain’ and not to the extension of it, the corpse or ‘envelope’? This will make him complain and will create challenges. Robot II will need homeostasis, the biological fine-tuning with respect to the environment.
He might end up eating pizza’s all the same though.
Feedback is the keyword I guess. Feedback as an element of cybernetics. Well, homeostasis is kind of cybernetics, but cybernetics is not necessary homeostasis. You could say Robot is a multi-cause-and-effect-loop. And a very deterministic one, too. Given the same cause, actually a set of signals, you will see the same effect.
Of course, and probably that's the problem, there is no communication between the "brain" and the envelope. I suppose, John and Mary have dressed "him" to have a nice looking guy for all kind of errands. Robot doesn't have preferences, I guess ... well, he shouldn't have.
Under normal circumstances, Robot would have succeeded in his job. It's our fault, Jan's and mine, that we didn't foresee a policeman asking Robot for his driving license. Which event kick-started the catastrophic cause-and-effect-loop, and concluded with that anti-aircraft missile.
Quite obvious, Jan and I were sad, very sad. Jan especially. He needed comfort, so I told him, I swore to him, that our first guess - that police barricade was after all a guess, we didn't contact the police for obvious reasons -was wrong.
So, there was Robot, eating from the pizza tree.
Well, is it conceivable that this simple fact has changed Robot? That it started some communication between his - gorgeous! - body and his mind. Did it give him some information about homeostasis?
After all, where are the remains of all those pizzas?
The same happened with the Fall. Before the fall there was no need for homeostasis!
We can't ask him anymore. After I ordered him to go back to that police barricade, we lost contact - think of that Malaysian airliner.
The misconception is, perhaps, that nature is visualized with the personality of a scientist with some sort of goal - perhaps the goal, in your mind, of creating a thinking machine. But nature has no goals, it merely has possible processes with matter and energy and a hell of a lot more time than even you can imagine. It uses its time and matter and energy to try everything possible with these basic elements and when they don't work they disappear and when they work every possible variation of things that do work is fabricated. One of those things is a thinking mind and somehow somewhere in the universe it appears. Conditions, such as a planet that becomes to hot or too dry or something else impossible for life and even a thinking mind cannot manage to exist so it disappears. Then, perhaps, within a billion years with the right condition another thinking mind appears. But nature doesn't give a damn one way or another because nature is only jumbling together possibilities. It all is just an exploration of the possible.
I fully agree with you Jan. However, a thinking machine build by humans might be considered a sort of ‘extended phenotype’ of men’s brain activity and as such a ‘natural’ phenomenon, like the birds nests, the termites hills, cities and machines of course, also in case the robot is built by robots. If I understand well, leonardo thinks it is impossible to build a robot that will not be captured by the police (blocked by an unforeseen event) or become a ‘fallen’ robot so that the whole human story starts over again. It seems he thinks of the robot as a brain with an envelope, to comfortable so that it has no challenges to overcome obstacles. If you want to create a robot that responds to the environment, he should feel cold, pain, hunger, joy, so you will have to create almost a real body with all the possible feed-back or homeostasis. That might indeed be impossible. I just wanted to stress the importance of the body, as input and output for the ‘brain’ instead of only language input and output as is the case in the conversation with Hauser.
What should we expect from the idea of the superior robot if the human brain works (as seems to be the case) with quantum mechanics?
There are several things in this blog that I strongly disagree with:
1) giblets are not dirty!
2) the spaghetti recipe sounds pretty messy
3)block-structured programming is done for humans and if computers will program may do it very differently.
4) the world is full of very badly written programs. That in general don't work.
5) There is an active research area in CS that measures the behaviour of correct programs that are subject to random changes.
6) viewing things as systems is not a disease, but the only way we have to understand what is around us, including our selves.
7) about your 3 conclusions:
i) I don't really understand what it means that a processor feels comfortable, do you?
ii) There is no need to have fantasy to imagine a world in which there is competition among computer/robots producers and that this competition brings improvements in capability, energy consumption ecc. Maybe this is enough as an incentive for changes? Then if you want to think about robots producing other robots, it's easy to transpose this situations to robots producing robots competing with each other.
iii) today the human superiority with respect to computers is probably the ability to face unexpected events (whereas computers are much better in few specific tasks). However the number of neurons of the human brain (and all their connections) is much bigger that of chips of a processor. Thus, I would suggest a bit of patience.
Finally I may propose a different ending for your story: it's not the computer that eats the human pizza, but may be the humans that eat whatever computers will kook according to their taste.
Humans always tend to be egocentric. Don't they ?
1)2) a matter of taste
3) I'm not aware of programming computers which are not instructed by human beings
4) the world is well organized - okay, in a reasonable way - so very badly written programs is a matter of uninformed discontent
5) no comment
6) I didn't say it's a disease; only that we have a lot of exits for our imagination
7i) I feel comfortable in my body, so, yes I do
7ii) ahem ... I do just not believe in robots producing robits like we produce robots
7iii) there's no need for patience: things just happen or don't happen in my lifetime
finally) no, you are not allowed to propose a different ending; you can write your own story
You seem to know me better than myself. I think you are right: I do not believe in creating a robot, which will not have problems with authorities.
Yes, the robot is a brain with an envelope - of course there's some communication: raise your arm, look to the left etc. But it's directive, not autonomous.
If you want to create a body that feels pain, cold, hungry - joy! very important, it's spiritual - then you are playing God. I do not mean the God of the creationist, not the Intelligent Designer - I'm talking about unknown powers.
And you are right, the ideas of Hauser & Co are just a little tiny thing in the complexity of the human body.
Considering quantum mechanics: I hope Jan will say something to that.
The problem you seem to be facing is that programming seems to be in a very elementary stage. Humans are also programmed by interaction with basic internal systems and also the environment, Current programmers seem not to be suficiently capable to write flexible programs that can modify themselves in context with the environment. It may take a while but we will get there if civilization survives. Considering the interactions of current politics and technical military proficiency civilization will not survive more than a few decades at best.
I think that you underestimate what computer can do already today:
a) there are programms capable of writing other programs that can cope with very complicated problems, such as the parsing of programs (including themselves)
b) There are robots that build other robots, for instance robots that build cars which soon will ride by themselves, so they will really be robots.
I think that this underestimation depends on the fact that men are still on control of the process. But there are programs that can produce results that cannot be forseen by people. For instance programs that invest in the stock markets.
I think that people tend to expect that computers behave like Hal of 2001 and, since they don't (yet), then think that they will never be capable of doing very original things. I think that there are two wrong points in such way of thinking:
i) computers have had an incredible progress in the last 30 years. Massa internet started some 20 years ago and it has transformed our lives! And who had forseen 20 years ago that all this was ready to happen? So why do we think that this progress is not astonishing enough (but computers are not human!) and that it is not going to continue in unexpected directions?
ii) people are disappointed because computers don't behave like humans. But computers are mainly useful for other, more practical, things. In those applications there has been enormous progress and enormous research (thus money) efforts. Adding feelings or consciousness or body to computers are probably fascinating problems for some people, but the effort behind those aims is enormously less. When computers with feelings will be proven to be more useful than emotionless computers, more effort will be put into producing them and, maybe, they will enter in production. Whereas, if this usefulness will not become evident (as I think is the case now) then their construction will remain a marginal effort in the hands of some curious people.
In other words, computers as such, don't need to go through natural selection, so why should they do it? They have their own history and development that for the moment is guided by the market.
First of all, I think you underestimate my esteem for the people who have brought the computer (pars pro toto) and its application at the level we experience today.
I think the stock market is not a good example: knowing the parameters and being fully aware of what is happening on the market, the result is foreseeable.
You write: " people are disappointed because computers don't behave like humans." Not this person, Gilberto, not this person. I'm very happy with the status quo. (Personally I believe that a lot of persons - and a hell of a lot for that matter - are glad the streets aren't filled with robots - drones! - or other electronic devices.)
I understand we disagree when it comes to computers with immaterial - which means to me: spiritual - feelings. No, I am not a believer, for a lot of reasons - and that was the starting point of my story: the controversy between Jan and me.
Remember, this comment is written by Leonardo, who tries to conform to sane thinking. The story is told by Leonardo the storyteller. Don't confuse the two!
Post a Comment